* [Qemu-devel] [PATCH v3 00/35] target/riscv: Convert to decodetree
@ 2018-10-31 13:19 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:19 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Hi,
this patchset converts the RISC-V decoder to decodetree in four major steps:
1) Convert 32-bit instructions to decodetree [Patch 1-16]:
Many of the gen_* functions are called by the decode functions for 16-bit
and 32-bit functions. If we move translation code from the gen_*
functions to the generated trans_* functions of decode-tree, we get a lot of
duplication. Therefore, we mostly generate calls to the old gen_* function
which are properly replaced after step 2).
Each of the trans_ functions are grouped into files corresponding to their
ISA extension, e.g. addi which is in RV32I is translated in the file
'trans_rvi.inc.c'.
2) Convert 16-bit instructions to decodetree [Patch 17-19]:
All 16 bit instructions have a direct mapping to a 32 bit instruction. Thus,
we convert the arguments in the 16 bit trans_ function to the arguments of
the corresponding 32 bit instruction and call the 32 bit trans_ function.
3) Remove old manual decoding in gen_* function [Patch 20-30]:
this move all manual translation code into the trans_* instructions of
decode tree, such that we can remove the old decode_* functions.
4) Simply RVC by reusing as much as possible from the RVG decoder as suggested
by Richard. [Patch 31-35]
this series depends on the decodetree patches by Richard:
https://github.com/rth7680/qemu/tree/decodetree
full tree available at
https://github.com/bkoppelmann/qemu/tree/riscv-dt-v3
Cheers,
Bastian
v2 -> v3:
- ex_shift_amount returns int
- dropped insn argument of trans_foo functions
- removal of AUIPC moved to 0002
- &branch -> &b
- split 0004 into two patches for RV32 and RV64
- moved 64-bit only insn to insn64.decode
- removed %pred/%succ
- dropped TARGET_RISCV64 requirement for fclass_d
- illegal instruction call replaced with return false
- trans_c_flw_ld and trans_c_fsw_sd now return false for now, as we
cannot use the insn argument anymore for manual decoding. We fix this
in a later patch when rvc is properly split up into insn16-32.decode
and insn16-64.decode.
- special case of trans_c_addi4spn() returns false in this patch
- simplified trans_c_srai by Richard's suggestion
- Since trans_c_flw_ld and trans_c_fsw_sd still rely on the old decoder
we need to keep gen_load(). Thus we renamed it to gen_load_c.
- Since trans_c_flw_ld and trans_c_fsw_sd still rely on the old decoder
we need to keep gen_store(). Thus we renamed it to gen_store_c.
- tcg_memop_lookup is now only used by TARGET_RISCV64 insn so we wrapped
it in a ifdef
- trans_srli/srai now use tcg_gen_shri/srai_tl
- trans_addiw uses its own gen_addiw function which properly extends the
result
- &arith_imm -> &i
- &arith -> &r
- trans_mulw now uses gen_mulw which properly sign extends the result
- gen_arith_w -> gen_arith_div_w
- gen_arith_div_w properly sign extends the resul
Bastian Koppelmann (35):
target/riscv: Move CPURISCVState pointer to DisasContext
target/riscv: Activate decodetree and implemnt LUI & AUIPC
target/riscv: Convert RVXI branch insns to decodetree
target/riscv: Convert RV32I load/store insns to decodetree
target/riscv: Convert RV64I load/store insns to decodetree
target/riscv: Convert RVXI arithmetic insns to decodetree
target/riscv: Convert RVXI fence insns to decodetree
target/riscv: Convert RVXI csr insns to decodetree
target/riscv: Convert RVXM insns to decodetree
target/riscv: Convert RV32A insns to decodetree
target/riscv: Convert RV64A insns to decodetree
target/riscv: Convert RV32F insns to decodetree
target/riscv: Convert RV64F insns to decodetree
target/riscv: Convert RV32D insns to decodetree
target/riscv: Convert RV64D insns to decodetree
target/riscv: Convert RV priv insns to decodetree
target/riscv: Convert quadrant 0 of RVXC insns to decodetree
target/riscv: Convert quadrant 1 of RVXC insns to decodetree
target/riscv: Convert quadrant 2 of RVXC insns to decodetree
target/riscv: Remove gen_jalr()
target/riscv: Remove manual decoding from gen_branch()
target/riscv: Remove manual decoding from gen_load()
target/riscv: Remove manual decoding from gen_store()
target/riscv: Move gen_arith_imm() decoding into trans_* functions
target/riscv: make ADD/SUB/OR/XOR/AND insn use arg lists
target/riscv: Remove shift and slt insn manual decoding
target/riscv: Remove manual decoding of RV32/64M insn
target/riscv: Rename trans_arith to gen_arith
target/riscv: Remove gen_system()
target/riscv: Remove decode_RV32_64G()
target/riscv: Convert @cs_2 insns to share translation functions
target/riscv: Convert @cl_d, @cl_w, @cs_d, @cs_w insns
target/riscv: Splice fsw_sd and flw_ld for riscv32 vs riscv64
target/riscv: Splice remaining compressed insn pairs for riscv32 vs
riscv64
target/riscv: Remaining rvc insn reuse 32 bit translators
target/riscv/Makefile.objs | 22 +
target/riscv/insn16-32.decode | 31 +
target/riscv/insn16-64.decode | 33 +
target/riscv/insn16.decode | 114 ++
target/riscv/insn32.decode | 203 ++
target/riscv/insn64.decode | 72 +
.../riscv/insn_trans/trans_privileged.inc.c | 108 +
target/riscv/insn_trans/trans_rva.inc.c | 203 ++
target/riscv/insn_trans/trans_rvc.inc.c | 149 ++
target/riscv/insn_trans/trans_rvd.inc.c | 388 ++++
target/riscv/insn_trans/trans_rvf.inc.c | 388 ++++
target/riscv/insn_trans/trans_rvi.inc.c | 584 ++++++
target/riscv/insn_trans/trans_rvm.inc.c | 107 +
target/riscv/translate.c | 1771 ++---------------
14 files changed, 2615 insertions(+), 1558 deletions(-)
create mode 100644 target/riscv/insn16-32.decode
create mode 100644 target/riscv/insn16-64.decode
create mode 100644 target/riscv/insn16.decode
create mode 100644 target/riscv/insn32.decode
create mode 100644 target/riscv/insn64.decode
create mode 100644 target/riscv/insn_trans/trans_privileged.inc.c
create mode 100644 target/riscv/insn_trans/trans_rva.inc.c
create mode 100644 target/riscv/insn_trans/trans_rvc.inc.c
create mode 100644 target/riscv/insn_trans/trans_rvd.inc.c
create mode 100644 target/riscv/insn_trans/trans_rvf.inc.c
create mode 100644 target/riscv/insn_trans/trans_rvi.inc.c
create mode 100644 target/riscv/insn_trans/trans_rvm.inc.c
--
2.19.1
^ permalink raw reply [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 00/35] target/riscv: Convert to decodetree
@ 2018-10-31 13:19 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:19 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Hi,
this patchset converts the RISC-V decoder to decodetree in four major steps:
1) Convert 32-bit instructions to decodetree [Patch 1-16]:
Many of the gen_* functions are called by the decode functions for 16-bit
and 32-bit functions. If we move translation code from the gen_*
functions to the generated trans_* functions of decode-tree, we get a lot of
duplication. Therefore, we mostly generate calls to the old gen_* function
which are properly replaced after step 2).
Each of the trans_ functions are grouped into files corresponding to their
ISA extension, e.g. addi which is in RV32I is translated in the file
'trans_rvi.inc.c'.
2) Convert 16-bit instructions to decodetree [Patch 17-19]:
All 16 bit instructions have a direct mapping to a 32 bit instruction. Thus,
we convert the arguments in the 16 bit trans_ function to the arguments of
the corresponding 32 bit instruction and call the 32 bit trans_ function.
3) Remove old manual decoding in gen_* function [Patch 20-30]:
this move all manual translation code into the trans_* instructions of
decode tree, such that we can remove the old decode_* functions.
4) Simply RVC by reusing as much as possible from the RVG decoder as suggested
by Richard. [Patch 31-35]
this series depends on the decodetree patches by Richard:
https://github.com/rth7680/qemu/tree/decodetree
full tree available at
https://github.com/bkoppelmann/qemu/tree/riscv-dt-v3
Cheers,
Bastian
v2 -> v3:
- ex_shift_amount returns int
- dropped insn argument of trans_foo functions
- removal of AUIPC moved to 0002
- &branch -> &b
- split 0004 into two patches for RV32 and RV64
- moved 64-bit only insn to insn64.decode
- removed %pred/%succ
- dropped TARGET_RISCV64 requirement for fclass_d
- illegal instruction call replaced with return false
- trans_c_flw_ld and trans_c_fsw_sd now return false for now, as we
cannot use the insn argument anymore for manual decoding. We fix this
in a later patch when rvc is properly split up into insn16-32.decode
and insn16-64.decode.
- special case of trans_c_addi4spn() returns false in this patch
- simplified trans_c_srai by Richard's suggestion
- Since trans_c_flw_ld and trans_c_fsw_sd still rely on the old decoder
we need to keep gen_load(). Thus we renamed it to gen_load_c.
- Since trans_c_flw_ld and trans_c_fsw_sd still rely on the old decoder
we need to keep gen_store(). Thus we renamed it to gen_store_c.
- tcg_memop_lookup is now only used by TARGET_RISCV64 insn so we wrapped
it in a ifdef
- trans_srli/srai now use tcg_gen_shri/srai_tl
- trans_addiw uses its own gen_addiw function which properly extends the
result
- &arith_imm -> &i
- &arith -> &r
- trans_mulw now uses gen_mulw which properly sign extends the result
- gen_arith_w -> gen_arith_div_w
- gen_arith_div_w properly sign extends the resul
Bastian Koppelmann (35):
target/riscv: Move CPURISCVState pointer to DisasContext
target/riscv: Activate decodetree and implemnt LUI & AUIPC
target/riscv: Convert RVXI branch insns to decodetree
target/riscv: Convert RV32I load/store insns to decodetree
target/riscv: Convert RV64I load/store insns to decodetree
target/riscv: Convert RVXI arithmetic insns to decodetree
target/riscv: Convert RVXI fence insns to decodetree
target/riscv: Convert RVXI csr insns to decodetree
target/riscv: Convert RVXM insns to decodetree
target/riscv: Convert RV32A insns to decodetree
target/riscv: Convert RV64A insns to decodetree
target/riscv: Convert RV32F insns to decodetree
target/riscv: Convert RV64F insns to decodetree
target/riscv: Convert RV32D insns to decodetree
target/riscv: Convert RV64D insns to decodetree
target/riscv: Convert RV priv insns to decodetree
target/riscv: Convert quadrant 0 of RVXC insns to decodetree
target/riscv: Convert quadrant 1 of RVXC insns to decodetree
target/riscv: Convert quadrant 2 of RVXC insns to decodetree
target/riscv: Remove gen_jalr()
target/riscv: Remove manual decoding from gen_branch()
target/riscv: Remove manual decoding from gen_load()
target/riscv: Remove manual decoding from gen_store()
target/riscv: Move gen_arith_imm() decoding into trans_* functions
target/riscv: make ADD/SUB/OR/XOR/AND insn use arg lists
target/riscv: Remove shift and slt insn manual decoding
target/riscv: Remove manual decoding of RV32/64M insn
target/riscv: Rename trans_arith to gen_arith
target/riscv: Remove gen_system()
target/riscv: Remove decode_RV32_64G()
target/riscv: Convert @cs_2 insns to share translation functions
target/riscv: Convert @cl_d, @cl_w, @cs_d, @cs_w insns
target/riscv: Splice fsw_sd and flw_ld for riscv32 vs riscv64
target/riscv: Splice remaining compressed insn pairs for riscv32 vs
riscv64
target/riscv: Remaining rvc insn reuse 32 bit translators
target/riscv/Makefile.objs | 22 +
target/riscv/insn16-32.decode | 31 +
target/riscv/insn16-64.decode | 33 +
target/riscv/insn16.decode | 114 ++
target/riscv/insn32.decode | 203 ++
target/riscv/insn64.decode | 72 +
.../riscv/insn_trans/trans_privileged.inc.c | 108 +
target/riscv/insn_trans/trans_rva.inc.c | 203 ++
target/riscv/insn_trans/trans_rvc.inc.c | 149 ++
target/riscv/insn_trans/trans_rvd.inc.c | 388 ++++
target/riscv/insn_trans/trans_rvf.inc.c | 388 ++++
target/riscv/insn_trans/trans_rvi.inc.c | 584 ++++++
target/riscv/insn_trans/trans_rvm.inc.c | 107 +
target/riscv/translate.c | 1771 ++---------------
14 files changed, 2615 insertions(+), 1558 deletions(-)
create mode 100644 target/riscv/insn16-32.decode
create mode 100644 target/riscv/insn16-64.decode
create mode 100644 target/riscv/insn16.decode
create mode 100644 target/riscv/insn32.decode
create mode 100644 target/riscv/insn64.decode
create mode 100644 target/riscv/insn_trans/trans_privileged.inc.c
create mode 100644 target/riscv/insn_trans/trans_rva.inc.c
create mode 100644 target/riscv/insn_trans/trans_rvc.inc.c
create mode 100644 target/riscv/insn_trans/trans_rvd.inc.c
create mode 100644 target/riscv/insn_trans/trans_rvf.inc.c
create mode 100644 target/riscv/insn_trans/trans_rvi.inc.c
create mode 100644 target/riscv/insn_trans/trans_rvm.inc.c
--
2.19.1
^ permalink raw reply [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 01/35] target/riscv: Move CPURISCVState pointer to DisasContext
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:19 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:19 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
CPURISCVState is rarely used, so there is no need to pass it to every
translate function. This paves the way for decodetree which only passes
DisasContext to translate functions.
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
---
target/riscv/translate.c | 15 ++++++++-------
1 file changed, 8 insertions(+), 7 deletions(-)
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 18d7b6d147..e81b9f097e 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -52,6 +52,7 @@ typedef struct DisasContext {
to any system register, which includes CSR_FRM, so we do not have
to reset this known value. */
int frm;
+ CPURISCVState *env;
} DisasContext;
/* convert riscv funct3 to qemu memop for load/store */
@@ -1789,19 +1790,19 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
}
}
-static void decode_opc(CPURISCVState *env, DisasContext *ctx)
+static void decode_opc(DisasContext *ctx)
{
/* check for compressed insn */
if (extract32(ctx->opcode, 0, 2) != 3) {
- if (!riscv_has_ext(env, RVC)) {
+ if (!riscv_has_ext(ctx->env, RVC)) {
gen_exception_illegal(ctx);
} else {
ctx->pc_succ_insn = ctx->base.pc_next + 2;
- decode_RV32_64C(env, ctx);
+ decode_RV32_64C(ctx->env, ctx);
}
} else {
ctx->pc_succ_insn = ctx->base.pc_next + 4;
- decode_RV32_64G(env, ctx);
+ decode_RV32_64G(ctx->env, ctx);
}
}
@@ -1846,10 +1847,10 @@ static bool riscv_tr_breakpoint_check(DisasContextBase *dcbase, CPUState *cpu,
static void riscv_tr_translate_insn(DisasContextBase *dcbase, CPUState *cpu)
{
DisasContext *ctx = container_of(dcbase, DisasContext, base);
- CPURISCVState *env = cpu->env_ptr;
+ ctx->env = cpu->env_ptr;
- ctx->opcode = cpu_ldl_code(env, ctx->base.pc_next);
- decode_opc(env, ctx);
+ ctx->opcode = cpu_ldl_code(ctx->env, ctx->base.pc_next);
+ decode_opc(ctx);
ctx->base.pc_next = ctx->pc_succ_insn;
if (ctx->base.is_jmp == DISAS_NEXT) {
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 01/35] target/riscv: Move CPURISCVState pointer to DisasContext
@ 2018-10-31 13:19 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:19 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
CPURISCVState is rarely used, so there is no need to pass it to every
translate function. This paves the way for decodetree which only passes
DisasContext to translate functions.
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
---
target/riscv/translate.c | 15 ++++++++-------
1 file changed, 8 insertions(+), 7 deletions(-)
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 18d7b6d147..e81b9f097e 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -52,6 +52,7 @@ typedef struct DisasContext {
to any system register, which includes CSR_FRM, so we do not have
to reset this known value. */
int frm;
+ CPURISCVState *env;
} DisasContext;
/* convert riscv funct3 to qemu memop for load/store */
@@ -1789,19 +1790,19 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
}
}
-static void decode_opc(CPURISCVState *env, DisasContext *ctx)
+static void decode_opc(DisasContext *ctx)
{
/* check for compressed insn */
if (extract32(ctx->opcode, 0, 2) != 3) {
- if (!riscv_has_ext(env, RVC)) {
+ if (!riscv_has_ext(ctx->env, RVC)) {
gen_exception_illegal(ctx);
} else {
ctx->pc_succ_insn = ctx->base.pc_next + 2;
- decode_RV32_64C(env, ctx);
+ decode_RV32_64C(ctx->env, ctx);
}
} else {
ctx->pc_succ_insn = ctx->base.pc_next + 4;
- decode_RV32_64G(env, ctx);
+ decode_RV32_64G(ctx->env, ctx);
}
}
@@ -1846,10 +1847,10 @@ static bool riscv_tr_breakpoint_check(DisasContextBase *dcbase, CPUState *cpu,
static void riscv_tr_translate_insn(DisasContextBase *dcbase, CPUState *cpu)
{
DisasContext *ctx = container_of(dcbase, DisasContext, base);
- CPURISCVState *env = cpu->env_ptr;
+ ctx->env = cpu->env_ptr;
- ctx->opcode = cpu_ldl_code(env, ctx->base.pc_next);
- decode_opc(env, ctx);
+ ctx->opcode = cpu_ldl_code(ctx->env, ctx->base.pc_next);
+ decode_opc(ctx);
ctx->base.pc_next = ctx->pc_succ_insn;
if (ctx->base.is_jmp == DISAS_NEXT) {
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 02/35] target/riscv: Activate decodetree and implemnt LUI & AUIPC
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:19 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:19 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
for now only LUI & AUIPC are decoded and translated. If decodetree fails, we
fall back to the old decoder.
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- ex_shift_amount returns int
- dropped insn argument of trans_foo functions
target/riscv/Makefile.objs | 10 +++++++
target/riscv/insn32.decode | 30 +++++++++++++++++++++
target/riscv/insn_trans/trans_rvi.inc.c | 35 +++++++++++++++++++++++++
target/riscv/translate.c | 31 ++++++++++++----------
4 files changed, 92 insertions(+), 14 deletions(-)
create mode 100644 target/riscv/insn32.decode
create mode 100644 target/riscv/insn_trans/trans_rvi.inc.c
diff --git a/target/riscv/Makefile.objs b/target/riscv/Makefile.objs
index fcc5d34c1f..ee995b3fc7 100644
--- a/target/riscv/Makefile.objs
+++ b/target/riscv/Makefile.objs
@@ -1 +1,11 @@
obj-y += translate.o op_helper.o cpu_helper.o cpu.o fpu_helper.o gdbstub.o pmp.o
+
+DECODETREE = $(SRC_PATH)/scripts/decodetree.py
+
+target/riscv/decode_insn32.inc.c: \
+ $(SRC_PATH)/target/riscv/insn32.decode $(DECODETREE)
+ $(call quiet-command, \
+ $(PYTHON) $(DECODETREE) -o $@ --decode decode_insn32 $<, \
+ "GEN", $(TARGET_DIR)$@)
+
+target/riscv/translate.o: target/riscv/decode_insn32.inc.c
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
new file mode 100644
index 0000000000..44d4e922b6
--- /dev/null
+++ b/target/riscv/insn32.decode
@@ -0,0 +1,30 @@
+#
+# RISC-V translation routines for the RVXI Base Integer Instruction Set.
+#
+# Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
+# Bastian Koppelmann, kbastian@mail.uni-paderborn.de
+#
+# This program is free software; you can redistribute it and/or modify it
+# under the terms and conditions of the GNU General Public License,
+# version 2 or later, as published by the Free Software Foundation.
+#
+# This program is distributed in the hope it will be useful, but WITHOUT
+# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+# FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+# more details.
+#
+# You should have received a copy of the GNU General Public License along with
+# this program. If not, see <http://www.gnu.org/licenses/>.
+
+# Fields:
+%rd 7:5
+
+# immediates:
+%imm_u 12:s20 !function=ex_shift_12
+
+# Formats 32:
+@u .................... ..... ....... imm=%imm_u %rd
+
+# *** RV32I Base Instruction Set ***
+lui .................... ..... 0110111 @u
+auipc .................... ..... 0010111 @u
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
new file mode 100644
index 0000000000..9885a8d275
--- /dev/null
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -0,0 +1,35 @@
+/*
+ * RISC-V translation routines for the RVXI Base Integer Instruction Set.
+ *
+ * Copyright (c) 2016-2017 Sagar Karandikar, sagark@eecs.berkeley.edu
+ * Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
+ * Bastian Koppelmann, kbastian@mail.uni-paderborn.de
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2 or later, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+static bool trans_lui(DisasContext *ctx, arg_lui *a)
+{
+ if (a->rd != 0) {
+ tcg_gen_movi_tl(cpu_gpr[a->rd], a->imm);
+ }
+ return true;
+}
+
+static bool trans_auipc(DisasContext *ctx, arg_auipc *a)
+{
+ if (a->rd != 0) {
+ tcg_gen_movi_tl(cpu_gpr[a->rd], a->imm + ctx->base.pc_next);
+ }
+ return true;
+}
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index e81b9f097e..ba57cb29f5 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -1667,6 +1667,19 @@ static void decode_RV32_64C(CPURISCVState *env, DisasContext *ctx)
}
}
+#define EX_SH(amount) \
+ static int ex_shift_##amount(int imm) \
+ { \
+ return imm << amount; \
+ }
+EX_SH(12)
+
+bool decode_insn32(DisasContext *ctx, uint32_t insn);
+/* Include the auto-generated decoder for 32 bit insn */
+#include "decode_insn32.inc.c"
+/* Include insn module translation function */
+#include "insn_trans/trans_rvi.inc.c"
+
static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
{
int rs1;
@@ -1687,19 +1700,6 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
imm = GET_IMM(ctx->opcode);
switch (op) {
- case OPC_RISC_LUI:
- if (rd == 0) {
- break; /* NOP */
- }
- tcg_gen_movi_tl(cpu_gpr[rd], sextract64(ctx->opcode, 12, 20) << 12);
- break;
- case OPC_RISC_AUIPC:
- if (rd == 0) {
- break; /* NOP */
- }
- tcg_gen_movi_tl(cpu_gpr[rd], (sextract64(ctx->opcode, 12, 20) << 12) +
- ctx->base.pc_next);
- break;
case OPC_RISC_JAL:
imm = GET_JAL_IMM(ctx->opcode);
gen_jal(env, ctx, rd, imm);
@@ -1802,7 +1802,10 @@ static void decode_opc(DisasContext *ctx)
}
} else {
ctx->pc_succ_insn = ctx->base.pc_next + 4;
- decode_RV32_64G(ctx->env, ctx);
+ if (!decode_insn32(ctx, ctx->opcode)) {
+ /* fallback to old decoder */
+ decode_RV32_64G(ctx->env, ctx);
+ }
}
}
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 02/35] target/riscv: Activate decodetree and implemnt LUI & AUIPC
@ 2018-10-31 13:19 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:19 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
for now only LUI & AUIPC are decoded and translated. If decodetree fails, we
fall back to the old decoder.
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- ex_shift_amount returns int
- dropped insn argument of trans_foo functions
target/riscv/Makefile.objs | 10 +++++++
target/riscv/insn32.decode | 30 +++++++++++++++++++++
target/riscv/insn_trans/trans_rvi.inc.c | 35 +++++++++++++++++++++++++
target/riscv/translate.c | 31 ++++++++++++----------
4 files changed, 92 insertions(+), 14 deletions(-)
create mode 100644 target/riscv/insn32.decode
create mode 100644 target/riscv/insn_trans/trans_rvi.inc.c
diff --git a/target/riscv/Makefile.objs b/target/riscv/Makefile.objs
index fcc5d34c1f..ee995b3fc7 100644
--- a/target/riscv/Makefile.objs
+++ b/target/riscv/Makefile.objs
@@ -1 +1,11 @@
obj-y += translate.o op_helper.o cpu_helper.o cpu.o fpu_helper.o gdbstub.o pmp.o
+
+DECODETREE = $(SRC_PATH)/scripts/decodetree.py
+
+target/riscv/decode_insn32.inc.c: \
+ $(SRC_PATH)/target/riscv/insn32.decode $(DECODETREE)
+ $(call quiet-command, \
+ $(PYTHON) $(DECODETREE) -o $@ --decode decode_insn32 $<, \
+ "GEN", $(TARGET_DIR)$@)
+
+target/riscv/translate.o: target/riscv/decode_insn32.inc.c
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
new file mode 100644
index 0000000000..44d4e922b6
--- /dev/null
+++ b/target/riscv/insn32.decode
@@ -0,0 +1,30 @@
+#
+# RISC-V translation routines for the RVXI Base Integer Instruction Set.
+#
+# Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
+# Bastian Koppelmann, kbastian@mail.uni-paderborn.de
+#
+# This program is free software; you can redistribute it and/or modify it
+# under the terms and conditions of the GNU General Public License,
+# version 2 or later, as published by the Free Software Foundation.
+#
+# This program is distributed in the hope it will be useful, but WITHOUT
+# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+# FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+# more details.
+#
+# You should have received a copy of the GNU General Public License along with
+# this program. If not, see <http://www.gnu.org/licenses/>.
+
+# Fields:
+%rd 7:5
+
+# immediates:
+%imm_u 12:s20 !function=ex_shift_12
+
+# Formats 32:
+@u .................... ..... ....... imm=%imm_u %rd
+
+# *** RV32I Base Instruction Set ***
+lui .................... ..... 0110111 @u
+auipc .................... ..... 0010111 @u
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
new file mode 100644
index 0000000000..9885a8d275
--- /dev/null
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -0,0 +1,35 @@
+/*
+ * RISC-V translation routines for the RVXI Base Integer Instruction Set.
+ *
+ * Copyright (c) 2016-2017 Sagar Karandikar, sagark@eecs.berkeley.edu
+ * Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
+ * Bastian Koppelmann, kbastian@mail.uni-paderborn.de
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2 or later, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+static bool trans_lui(DisasContext *ctx, arg_lui *a)
+{
+ if (a->rd != 0) {
+ tcg_gen_movi_tl(cpu_gpr[a->rd], a->imm);
+ }
+ return true;
+}
+
+static bool trans_auipc(DisasContext *ctx, arg_auipc *a)
+{
+ if (a->rd != 0) {
+ tcg_gen_movi_tl(cpu_gpr[a->rd], a->imm + ctx->base.pc_next);
+ }
+ return true;
+}
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index e81b9f097e..ba57cb29f5 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -1667,6 +1667,19 @@ static void decode_RV32_64C(CPURISCVState *env, DisasContext *ctx)
}
}
+#define EX_SH(amount) \
+ static int ex_shift_##amount(int imm) \
+ { \
+ return imm << amount; \
+ }
+EX_SH(12)
+
+bool decode_insn32(DisasContext *ctx, uint32_t insn);
+/* Include the auto-generated decoder for 32 bit insn */
+#include "decode_insn32.inc.c"
+/* Include insn module translation function */
+#include "insn_trans/trans_rvi.inc.c"
+
static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
{
int rs1;
@@ -1687,19 +1700,6 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
imm = GET_IMM(ctx->opcode);
switch (op) {
- case OPC_RISC_LUI:
- if (rd == 0) {
- break; /* NOP */
- }
- tcg_gen_movi_tl(cpu_gpr[rd], sextract64(ctx->opcode, 12, 20) << 12);
- break;
- case OPC_RISC_AUIPC:
- if (rd == 0) {
- break; /* NOP */
- }
- tcg_gen_movi_tl(cpu_gpr[rd], (sextract64(ctx->opcode, 12, 20) << 12) +
- ctx->base.pc_next);
- break;
case OPC_RISC_JAL:
imm = GET_JAL_IMM(ctx->opcode);
gen_jal(env, ctx, rd, imm);
@@ -1802,7 +1802,10 @@ static void decode_opc(DisasContext *ctx)
}
} else {
ctx->pc_succ_insn = ctx->base.pc_next + 4;
- decode_RV32_64G(ctx->env, ctx);
+ if (!decode_insn32(ctx, ctx->opcode)) {
+ /* fallback to old decoder */
+ decode_RV32_64G(ctx->env, ctx);
+ }
}
}
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 03/35] target/riscv: Convert RVXI branch insns to decodetree
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:19 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:19 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- dropped insn argument of trans_foo functions
- removal of AUIPC moved to 0002
- &branch -> &b
target/riscv/insn32.decode | 19 ++++++++++
target/riscv/insn_trans/trans_rvi.inc.c | 49 +++++++++++++++++++++++++
target/riscv/translate.c | 12 +-----
3 files changed, 69 insertions(+), 11 deletions(-)
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index 44d4e922b6..81f56c16b4 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -17,14 +17,33 @@
# this program. If not, see <http://www.gnu.org/licenses/>.
# Fields:
+%rs2 20:5
+%rs1 15:5
%rd 7:5
# immediates:
+%imm_i 20:s12
+%imm_b 31:s1 7:1 25:6 8:4 !function=ex_shift_1
+%imm_j 31:s1 12:8 20:1 21:10 !function=ex_shift_1
%imm_u 12:s20 !function=ex_shift_12
+# Argument sets:
+&b imm rs2 rs1
+
# Formats 32:
+@i ............ ..... ... ..... ....... imm=%imm_i %rs1 %rd
+@b ....... ..... ..... ... ..... ....... &b imm=%imm_b %rs2 %rs1
@u .................... ..... ....... imm=%imm_u %rd
+@j .................... ..... ....... imm=%imm_j %rd
# *** RV32I Base Instruction Set ***
lui .................... ..... 0110111 @u
auipc .................... ..... 0010111 @u
+jal .................... ..... 1101111 @j
+jalr ............ ..... 000 ..... 1100111 @i
+beq ....... ..... ..... 000 ..... 1100011 @b
+bne ....... ..... ..... 001 ..... 1100011 @b
+blt ....... ..... ..... 100 ..... 1100011 @b
+bge ....... ..... ..... 101 ..... 1100011 @b
+bltu ....... ..... ..... 110 ..... 1100011 @b
+bgeu ....... ..... ..... 111 ..... 1100011 @b
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
index 9885a8d275..eaeab20282 100644
--- a/target/riscv/insn_trans/trans_rvi.inc.c
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -33,3 +33,52 @@ static bool trans_auipc(DisasContext *ctx, arg_auipc *a)
}
return true;
}
+
+static bool trans_jal(DisasContext *ctx, arg_jal *a, uint32_t insn)
+{
+ gen_jal(ctx->env, ctx, a->rd, a->imm);
+ return true;
+}
+
+static bool trans_jalr(DisasContext *ctx, arg_jalr *a, uint32_t insn)
+{
+ gen_jalr(ctx->env, ctx, OPC_RISC_JALR, a->rd, a->rs1, a->imm);
+ return true;
+}
+
+static bool trans_beq(DisasContext *ctx, arg_beq *a, uint32_t insn)
+{
+ gen_branch(ctx->env, ctx, OPC_RISC_BEQ, a->rs1, a->rs2, a->imm);
+ return true;
+}
+
+static bool trans_bne(DisasContext *ctx, arg_bne *a, uint32_t insn)
+{
+ gen_branch(ctx->env, ctx, OPC_RISC_BNE, a->rs1, a->rs2, a->imm);
+ return true;
+}
+
+static bool trans_blt(DisasContext *ctx, arg_blt *a, uint32_t insn)
+{
+ gen_branch(ctx->env, ctx, OPC_RISC_BLT, a->rs1, a->rs2, a->imm);
+ return true;
+}
+
+static bool trans_bge(DisasContext *ctx, arg_bge *a, uint32_t insn)
+{
+ gen_branch(ctx->env, ctx, OPC_RISC_BGE, a->rs1, a->rs2, a->imm);
+ return true;
+}
+
+static bool trans_bltu(DisasContext *ctx, arg_bltu *a, uint32_t insn)
+{
+ gen_branch(ctx->env, ctx, OPC_RISC_BLTU, a->rs1, a->rs2, a->imm);
+ return true;
+}
+
+static bool trans_bgeu(DisasContext *ctx, arg_bgeu *a, uint32_t insn)
+{
+
+ gen_branch(ctx->env, ctx, OPC_RISC_BGEU, a->rs1, a->rs2, a->imm);
+ return true;
+}
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index ba57cb29f5..909f7cd013 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -1672,6 +1672,7 @@ static void decode_RV32_64C(CPURISCVState *env, DisasContext *ctx)
{ \
return imm << amount; \
}
+EX_SH(1)
EX_SH(12)
bool decode_insn32(DisasContext *ctx, uint32_t insn);
@@ -1700,17 +1701,6 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
imm = GET_IMM(ctx->opcode);
switch (op) {
- case OPC_RISC_JAL:
- imm = GET_JAL_IMM(ctx->opcode);
- gen_jal(env, ctx, rd, imm);
- break;
- case OPC_RISC_JALR:
- gen_jalr(env, ctx, MASK_OP_JALR(ctx->opcode), rd, rs1, imm);
- break;
- case OPC_RISC_BRANCH:
- gen_branch(env, ctx, MASK_OP_BRANCH(ctx->opcode), rs1, rs2,
- GET_B_IMM(ctx->opcode));
- break;
case OPC_RISC_LOAD:
gen_load(ctx, MASK_OP_LOAD(ctx->opcode), rd, rs1, imm);
break;
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 03/35] target/riscv: Convert RVXI branch insns to decodetree
@ 2018-10-31 13:19 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:19 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- dropped insn argument of trans_foo functions
- removal of AUIPC moved to 0002
- &branch -> &b
target/riscv/insn32.decode | 19 ++++++++++
target/riscv/insn_trans/trans_rvi.inc.c | 49 +++++++++++++++++++++++++
target/riscv/translate.c | 12 +-----
3 files changed, 69 insertions(+), 11 deletions(-)
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index 44d4e922b6..81f56c16b4 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -17,14 +17,33 @@
# this program. If not, see <http://www.gnu.org/licenses/>.
# Fields:
+%rs2 20:5
+%rs1 15:5
%rd 7:5
# immediates:
+%imm_i 20:s12
+%imm_b 31:s1 7:1 25:6 8:4 !function=ex_shift_1
+%imm_j 31:s1 12:8 20:1 21:10 !function=ex_shift_1
%imm_u 12:s20 !function=ex_shift_12
+# Argument sets:
+&b imm rs2 rs1
+
# Formats 32:
+@i ............ ..... ... ..... ....... imm=%imm_i %rs1 %rd
+@b ....... ..... ..... ... ..... ....... &b imm=%imm_b %rs2 %rs1
@u .................... ..... ....... imm=%imm_u %rd
+@j .................... ..... ....... imm=%imm_j %rd
# *** RV32I Base Instruction Set ***
lui .................... ..... 0110111 @u
auipc .................... ..... 0010111 @u
+jal .................... ..... 1101111 @j
+jalr ............ ..... 000 ..... 1100111 @i
+beq ....... ..... ..... 000 ..... 1100011 @b
+bne ....... ..... ..... 001 ..... 1100011 @b
+blt ....... ..... ..... 100 ..... 1100011 @b
+bge ....... ..... ..... 101 ..... 1100011 @b
+bltu ....... ..... ..... 110 ..... 1100011 @b
+bgeu ....... ..... ..... 111 ..... 1100011 @b
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
index 9885a8d275..eaeab20282 100644
--- a/target/riscv/insn_trans/trans_rvi.inc.c
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -33,3 +33,52 @@ static bool trans_auipc(DisasContext *ctx, arg_auipc *a)
}
return true;
}
+
+static bool trans_jal(DisasContext *ctx, arg_jal *a, uint32_t insn)
+{
+ gen_jal(ctx->env, ctx, a->rd, a->imm);
+ return true;
+}
+
+static bool trans_jalr(DisasContext *ctx, arg_jalr *a, uint32_t insn)
+{
+ gen_jalr(ctx->env, ctx, OPC_RISC_JALR, a->rd, a->rs1, a->imm);
+ return true;
+}
+
+static bool trans_beq(DisasContext *ctx, arg_beq *a, uint32_t insn)
+{
+ gen_branch(ctx->env, ctx, OPC_RISC_BEQ, a->rs1, a->rs2, a->imm);
+ return true;
+}
+
+static bool trans_bne(DisasContext *ctx, arg_bne *a, uint32_t insn)
+{
+ gen_branch(ctx->env, ctx, OPC_RISC_BNE, a->rs1, a->rs2, a->imm);
+ return true;
+}
+
+static bool trans_blt(DisasContext *ctx, arg_blt *a, uint32_t insn)
+{
+ gen_branch(ctx->env, ctx, OPC_RISC_BLT, a->rs1, a->rs2, a->imm);
+ return true;
+}
+
+static bool trans_bge(DisasContext *ctx, arg_bge *a, uint32_t insn)
+{
+ gen_branch(ctx->env, ctx, OPC_RISC_BGE, a->rs1, a->rs2, a->imm);
+ return true;
+}
+
+static bool trans_bltu(DisasContext *ctx, arg_bltu *a, uint32_t insn)
+{
+ gen_branch(ctx->env, ctx, OPC_RISC_BLTU, a->rs1, a->rs2, a->imm);
+ return true;
+}
+
+static bool trans_bgeu(DisasContext *ctx, arg_bgeu *a, uint32_t insn)
+{
+
+ gen_branch(ctx->env, ctx, OPC_RISC_BGEU, a->rs1, a->rs2, a->imm);
+ return true;
+}
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index ba57cb29f5..909f7cd013 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -1672,6 +1672,7 @@ static void decode_RV32_64C(CPURISCVState *env, DisasContext *ctx)
{ \
return imm << amount; \
}
+EX_SH(1)
EX_SH(12)
bool decode_insn32(DisasContext *ctx, uint32_t insn);
@@ -1700,17 +1701,6 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
imm = GET_IMM(ctx->opcode);
switch (op) {
- case OPC_RISC_JAL:
- imm = GET_JAL_IMM(ctx->opcode);
- gen_jal(env, ctx, rd, imm);
- break;
- case OPC_RISC_JALR:
- gen_jalr(env, ctx, MASK_OP_JALR(ctx->opcode), rd, rs1, imm);
- break;
- case OPC_RISC_BRANCH:
- gen_branch(env, ctx, MASK_OP_BRANCH(ctx->opcode), rs1, rs2,
- GET_B_IMM(ctx->opcode));
- break;
case OPC_RISC_LOAD:
gen_load(ctx, MASK_OP_LOAD(ctx->opcode), rd, rs1, imm);
break;
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 04/35] target/riscv: Convert RV32I load/store insns to decodetree
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:19 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:19 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- split into two patches for RV32 and RV64
- dropped insn argument of trans_foo functions
target/riscv/insn32.decode | 10 ++++
target/riscv/insn_trans/trans_rvi.inc.c | 64 +++++++++++++++++++++----
2 files changed, 66 insertions(+), 8 deletions(-)
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index 81f56c16b4..076de873c4 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -23,6 +23,7 @@
# immediates:
%imm_i 20:s12
+%imm_s 25:s7 7:5
%imm_b 31:s1 7:1 25:6 8:4 !function=ex_shift_1
%imm_j 31:s1 12:8 20:1 21:10 !function=ex_shift_1
%imm_u 12:s20 !function=ex_shift_12
@@ -33,6 +34,7 @@
# Formats 32:
@i ............ ..... ... ..... ....... imm=%imm_i %rs1 %rd
@b ....... ..... ..... ... ..... ....... &b imm=%imm_b %rs2 %rs1
+@s ....... ..... ..... ... ..... ....... imm=%imm_s %rs2 %rs1
@u .................... ..... ....... imm=%imm_u %rd
@j .................... ..... ....... imm=%imm_j %rd
@@ -47,3 +49,11 @@ blt ....... ..... ..... 100 ..... 1100011 @b
bge ....... ..... ..... 101 ..... 1100011 @b
bltu ....... ..... ..... 110 ..... 1100011 @b
bgeu ....... ..... ..... 111 ..... 1100011 @b
+lb ............ ..... 000 ..... 0000011 @i
+lh ............ ..... 001 ..... 0000011 @i
+lw ............ ..... 010 ..... 0000011 @i
+lbu ............ ..... 100 ..... 0000011 @i
+lhu ............ ..... 101 ..... 0000011 @i
+sb ....... ..... ..... 000 ..... 0100011 @s
+sh ....... ..... ..... 001 ..... 0100011 @s
+sw ....... ..... ..... 010 ..... 0100011 @s
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
index eaeab20282..f3b88ebb69 100644
--- a/target/riscv/insn_trans/trans_rvi.inc.c
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -34,51 +34,99 @@ static bool trans_auipc(DisasContext *ctx, arg_auipc *a)
return true;
}
-static bool trans_jal(DisasContext *ctx, arg_jal *a, uint32_t insn)
+static bool trans_jal(DisasContext *ctx, arg_jal *a)
{
gen_jal(ctx->env, ctx, a->rd, a->imm);
return true;
}
-static bool trans_jalr(DisasContext *ctx, arg_jalr *a, uint32_t insn)
+static bool trans_jalr(DisasContext *ctx, arg_jalr *a)
{
gen_jalr(ctx->env, ctx, OPC_RISC_JALR, a->rd, a->rs1, a->imm);
return true;
}
-static bool trans_beq(DisasContext *ctx, arg_beq *a, uint32_t insn)
+static bool trans_beq(DisasContext *ctx, arg_beq *a)
{
gen_branch(ctx->env, ctx, OPC_RISC_BEQ, a->rs1, a->rs2, a->imm);
return true;
}
-static bool trans_bne(DisasContext *ctx, arg_bne *a, uint32_t insn)
+static bool trans_bne(DisasContext *ctx, arg_bne *a)
{
gen_branch(ctx->env, ctx, OPC_RISC_BNE, a->rs1, a->rs2, a->imm);
return true;
}
-static bool trans_blt(DisasContext *ctx, arg_blt *a, uint32_t insn)
+static bool trans_blt(DisasContext *ctx, arg_blt *a)
{
gen_branch(ctx->env, ctx, OPC_RISC_BLT, a->rs1, a->rs2, a->imm);
return true;
}
-static bool trans_bge(DisasContext *ctx, arg_bge *a, uint32_t insn)
+static bool trans_bge(DisasContext *ctx, arg_bge *a)
{
gen_branch(ctx->env, ctx, OPC_RISC_BGE, a->rs1, a->rs2, a->imm);
return true;
}
-static bool trans_bltu(DisasContext *ctx, arg_bltu *a, uint32_t insn)
+static bool trans_bltu(DisasContext *ctx, arg_bltu *a)
{
gen_branch(ctx->env, ctx, OPC_RISC_BLTU, a->rs1, a->rs2, a->imm);
return true;
}
-static bool trans_bgeu(DisasContext *ctx, arg_bgeu *a, uint32_t insn)
+static bool trans_bgeu(DisasContext *ctx, arg_bgeu *a)
{
gen_branch(ctx->env, ctx, OPC_RISC_BGEU, a->rs1, a->rs2, a->imm);
return true;
}
+
+static bool trans_lb(DisasContext *ctx, arg_lb *a)
+{
+ gen_load(ctx, OPC_RISC_LB, a->rd, a->rs1, a->imm);
+ return true;
+}
+
+static bool trans_lh(DisasContext *ctx, arg_lh *a)
+{
+ gen_load(ctx, OPC_RISC_LH, a->rd, a->rs1, a->imm);
+ return true;
+}
+
+static bool trans_lw(DisasContext *ctx, arg_lw *a)
+{
+ gen_load(ctx, OPC_RISC_LW, a->rd, a->rs1, a->imm);
+ return true;
+}
+
+static bool trans_lbu(DisasContext *ctx, arg_lbu *a)
+{
+ gen_load(ctx, OPC_RISC_LBU, a->rd, a->rs1, a->imm);
+ return true;
+}
+
+static bool trans_lhu(DisasContext *ctx, arg_lhu *a)
+{
+ gen_load(ctx, OPC_RISC_LHU, a->rd, a->rs1, a->imm);
+ return true;
+}
+
+static bool trans_sb(DisasContext *ctx, arg_sb *a)
+{
+ gen_store(ctx, OPC_RISC_SB, a->rs1, a->rs2, a->imm);
+ return true;
+}
+
+static bool trans_sh(DisasContext *ctx, arg_sh *a)
+{
+ gen_store(ctx, OPC_RISC_SH, a->rs1, a->rs2, a->imm);
+ return true;
+}
+
+static bool trans_sw(DisasContext *ctx, arg_sw *a)
+{
+ gen_store(ctx, OPC_RISC_SW, a->rs1, a->rs2, a->imm);
+ return true;
+}
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 04/35] target/riscv: Convert RV32I load/store insns to decodetree
@ 2018-10-31 13:19 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:19 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- split into two patches for RV32 and RV64
- dropped insn argument of trans_foo functions
target/riscv/insn32.decode | 10 ++++
target/riscv/insn_trans/trans_rvi.inc.c | 64 +++++++++++++++++++++----
2 files changed, 66 insertions(+), 8 deletions(-)
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index 81f56c16b4..076de873c4 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -23,6 +23,7 @@
# immediates:
%imm_i 20:s12
+%imm_s 25:s7 7:5
%imm_b 31:s1 7:1 25:6 8:4 !function=ex_shift_1
%imm_j 31:s1 12:8 20:1 21:10 !function=ex_shift_1
%imm_u 12:s20 !function=ex_shift_12
@@ -33,6 +34,7 @@
# Formats 32:
@i ............ ..... ... ..... ....... imm=%imm_i %rs1 %rd
@b ....... ..... ..... ... ..... ....... &b imm=%imm_b %rs2 %rs1
+@s ....... ..... ..... ... ..... ....... imm=%imm_s %rs2 %rs1
@u .................... ..... ....... imm=%imm_u %rd
@j .................... ..... ....... imm=%imm_j %rd
@@ -47,3 +49,11 @@ blt ....... ..... ..... 100 ..... 1100011 @b
bge ....... ..... ..... 101 ..... 1100011 @b
bltu ....... ..... ..... 110 ..... 1100011 @b
bgeu ....... ..... ..... 111 ..... 1100011 @b
+lb ............ ..... 000 ..... 0000011 @i
+lh ............ ..... 001 ..... 0000011 @i
+lw ............ ..... 010 ..... 0000011 @i
+lbu ............ ..... 100 ..... 0000011 @i
+lhu ............ ..... 101 ..... 0000011 @i
+sb ....... ..... ..... 000 ..... 0100011 @s
+sh ....... ..... ..... 001 ..... 0100011 @s
+sw ....... ..... ..... 010 ..... 0100011 @s
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
index eaeab20282..f3b88ebb69 100644
--- a/target/riscv/insn_trans/trans_rvi.inc.c
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -34,51 +34,99 @@ static bool trans_auipc(DisasContext *ctx, arg_auipc *a)
return true;
}
-static bool trans_jal(DisasContext *ctx, arg_jal *a, uint32_t insn)
+static bool trans_jal(DisasContext *ctx, arg_jal *a)
{
gen_jal(ctx->env, ctx, a->rd, a->imm);
return true;
}
-static bool trans_jalr(DisasContext *ctx, arg_jalr *a, uint32_t insn)
+static bool trans_jalr(DisasContext *ctx, arg_jalr *a)
{
gen_jalr(ctx->env, ctx, OPC_RISC_JALR, a->rd, a->rs1, a->imm);
return true;
}
-static bool trans_beq(DisasContext *ctx, arg_beq *a, uint32_t insn)
+static bool trans_beq(DisasContext *ctx, arg_beq *a)
{
gen_branch(ctx->env, ctx, OPC_RISC_BEQ, a->rs1, a->rs2, a->imm);
return true;
}
-static bool trans_bne(DisasContext *ctx, arg_bne *a, uint32_t insn)
+static bool trans_bne(DisasContext *ctx, arg_bne *a)
{
gen_branch(ctx->env, ctx, OPC_RISC_BNE, a->rs1, a->rs2, a->imm);
return true;
}
-static bool trans_blt(DisasContext *ctx, arg_blt *a, uint32_t insn)
+static bool trans_blt(DisasContext *ctx, arg_blt *a)
{
gen_branch(ctx->env, ctx, OPC_RISC_BLT, a->rs1, a->rs2, a->imm);
return true;
}
-static bool trans_bge(DisasContext *ctx, arg_bge *a, uint32_t insn)
+static bool trans_bge(DisasContext *ctx, arg_bge *a)
{
gen_branch(ctx->env, ctx, OPC_RISC_BGE, a->rs1, a->rs2, a->imm);
return true;
}
-static bool trans_bltu(DisasContext *ctx, arg_bltu *a, uint32_t insn)
+static bool trans_bltu(DisasContext *ctx, arg_bltu *a)
{
gen_branch(ctx->env, ctx, OPC_RISC_BLTU, a->rs1, a->rs2, a->imm);
return true;
}
-static bool trans_bgeu(DisasContext *ctx, arg_bgeu *a, uint32_t insn)
+static bool trans_bgeu(DisasContext *ctx, arg_bgeu *a)
{
gen_branch(ctx->env, ctx, OPC_RISC_BGEU, a->rs1, a->rs2, a->imm);
return true;
}
+
+static bool trans_lb(DisasContext *ctx, arg_lb *a)
+{
+ gen_load(ctx, OPC_RISC_LB, a->rd, a->rs1, a->imm);
+ return true;
+}
+
+static bool trans_lh(DisasContext *ctx, arg_lh *a)
+{
+ gen_load(ctx, OPC_RISC_LH, a->rd, a->rs1, a->imm);
+ return true;
+}
+
+static bool trans_lw(DisasContext *ctx, arg_lw *a)
+{
+ gen_load(ctx, OPC_RISC_LW, a->rd, a->rs1, a->imm);
+ return true;
+}
+
+static bool trans_lbu(DisasContext *ctx, arg_lbu *a)
+{
+ gen_load(ctx, OPC_RISC_LBU, a->rd, a->rs1, a->imm);
+ return true;
+}
+
+static bool trans_lhu(DisasContext *ctx, arg_lhu *a)
+{
+ gen_load(ctx, OPC_RISC_LHU, a->rd, a->rs1, a->imm);
+ return true;
+}
+
+static bool trans_sb(DisasContext *ctx, arg_sb *a)
+{
+ gen_store(ctx, OPC_RISC_SB, a->rs1, a->rs2, a->imm);
+ return true;
+}
+
+static bool trans_sh(DisasContext *ctx, arg_sh *a)
+{
+ gen_store(ctx, OPC_RISC_SH, a->rs1, a->rs2, a->imm);
+ return true;
+}
+
+static bool trans_sw(DisasContext *ctx, arg_sw *a)
+{
+ gen_store(ctx, OPC_RISC_SW, a->rs1, a->rs2, a->imm);
+ return true;
+}
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 05/35] target/riscv: Convert RV64I load/store insns to decodetree
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:19 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:19 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
this splits the 64-bit only instructions into its own decode file such
that we generate the decoder for these instructions only for the RISC-V
64 bit target.
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
target/riscv/Makefile.objs | 8 +++++---
target/riscv/insn64.decode | 25 +++++++++++++++++++++++++
target/riscv/insn_trans/trans_rvi.inc.c | 20 ++++++++++++++++++++
target/riscv/translate.c | 7 -------
4 files changed, 50 insertions(+), 10 deletions(-)
create mode 100644 target/riscv/insn64.decode
diff --git a/target/riscv/Makefile.objs b/target/riscv/Makefile.objs
index ee995b3fc7..b9b8152cc2 100644
--- a/target/riscv/Makefile.objs
+++ b/target/riscv/Makefile.objs
@@ -2,10 +2,12 @@ obj-y += translate.o op_helper.o cpu_helper.o cpu.o fpu_helper.o gdbstub.o pmp.o
DECODETREE = $(SRC_PATH)/scripts/decodetree.py
-target/riscv/decode_insn32.inc.c: \
- $(SRC_PATH)/target/riscv/insn32.decode $(DECODETREE)
+decode32-y = $(SRC_PATH)/target/riscv/insn32.decode
+decode32-$(TARGET_RISCV64) += $(SRC_PATH)/target/riscv/insn64.decode
+
+target/riscv/decode_insn32.inc.c: $(decode32-y) $(DECODETREE)
$(call quiet-command, \
- $(PYTHON) $(DECODETREE) -o $@ --decode decode_insn32 $<, \
+ $(PYTHON) $(DECODETREE) -o $@ --decode decode_insn32 $(decode32-y), \
"GEN", $(TARGET_DIR)$@)
target/riscv/translate.o: target/riscv/decode_insn32.inc.c
diff --git a/target/riscv/insn64.decode b/target/riscv/insn64.decode
new file mode 100644
index 0000000000..439d4e2c58
--- /dev/null
+++ b/target/riscv/insn64.decode
@@ -0,0 +1,25 @@
+#
+# RISC-V translation routines for the RV Instruction Set.
+#
+# Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
+# Bastian Koppelmann, kbastian@mail.uni-paderborn.de
+#
+# This program is free software; you can redistribute it and/or modify it
+# under the terms and conditions of the GNU General Public License,
+# version 2 or later, as published by the Free Software Foundation.
+#
+# This program is distributed in the hope it will be useful, but WITHOUT
+# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+# FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+# more details.
+#
+# You should have received a copy of the GNU General Public License along with
+# this program. If not, see <http://www.gnu.org/licenses/>.
+
+# This is concatenated with insn32.decode for risc64 targets.
+# Most of the fields and formats are there.
+
+# *** RV64I Base Instruction Set (in addition to RV32I) ***
+lwu ............ ..... 110 ..... 0000011 @i
+ld ............ ..... 011 ..... 0000011 @i
+sd ....... ..... ..... 011 ..... 0100011 @s
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
index f3b88ebb69..39a20a70e8 100644
--- a/target/riscv/insn_trans/trans_rvi.inc.c
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -130,3 +130,23 @@ static bool trans_sw(DisasContext *ctx, arg_sw *a)
gen_store(ctx, OPC_RISC_SW, a->rs1, a->rs2, a->imm);
return true;
}
+
+#ifdef TARGET_RISCV64
+static bool trans_lwu(DisasContext *ctx, arg_lwu *a)
+{
+ gen_load(ctx, OPC_RISC_LWU, a->rd, a->rs1, a->imm);
+ return true;
+}
+
+static bool trans_ld(DisasContext *ctx, arg_ld *a)
+{
+ gen_load(ctx, OPC_RISC_LD, a->rd, a->rs1, a->imm);
+ return true;
+}
+
+static bool trans_sd(DisasContext *ctx, arg_sd *a)
+{
+ gen_store(ctx, OPC_RISC_SD, a->rs1, a->rs2, a->imm);
+ return true;
+}
+#endif
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 909f7cd013..244855c82d 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -1701,13 +1701,6 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
imm = GET_IMM(ctx->opcode);
switch (op) {
- case OPC_RISC_LOAD:
- gen_load(ctx, MASK_OP_LOAD(ctx->opcode), rd, rs1, imm);
- break;
- case OPC_RISC_STORE:
- gen_store(ctx, MASK_OP_STORE(ctx->opcode), rs1, rs2,
- GET_STORE_IMM(ctx->opcode));
- break;
case OPC_RISC_ARITH_IMM:
#if defined(TARGET_RISCV64)
case OPC_RISC_ARITH_IMM_W:
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 05/35] target/riscv: Convert RV64I load/store insns to decodetree
@ 2018-10-31 13:19 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:19 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
this splits the 64-bit only instructions into its own decode file such
that we generate the decoder for these instructions only for the RISC-V
64 bit target.
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
target/riscv/Makefile.objs | 8 +++++---
target/riscv/insn64.decode | 25 +++++++++++++++++++++++++
target/riscv/insn_trans/trans_rvi.inc.c | 20 ++++++++++++++++++++
target/riscv/translate.c | 7 -------
4 files changed, 50 insertions(+), 10 deletions(-)
create mode 100644 target/riscv/insn64.decode
diff --git a/target/riscv/Makefile.objs b/target/riscv/Makefile.objs
index ee995b3fc7..b9b8152cc2 100644
--- a/target/riscv/Makefile.objs
+++ b/target/riscv/Makefile.objs
@@ -2,10 +2,12 @@ obj-y += translate.o op_helper.o cpu_helper.o cpu.o fpu_helper.o gdbstub.o pmp.o
DECODETREE = $(SRC_PATH)/scripts/decodetree.py
-target/riscv/decode_insn32.inc.c: \
- $(SRC_PATH)/target/riscv/insn32.decode $(DECODETREE)
+decode32-y = $(SRC_PATH)/target/riscv/insn32.decode
+decode32-$(TARGET_RISCV64) += $(SRC_PATH)/target/riscv/insn64.decode
+
+target/riscv/decode_insn32.inc.c: $(decode32-y) $(DECODETREE)
$(call quiet-command, \
- $(PYTHON) $(DECODETREE) -o $@ --decode decode_insn32 $<, \
+ $(PYTHON) $(DECODETREE) -o $@ --decode decode_insn32 $(decode32-y), \
"GEN", $(TARGET_DIR)$@)
target/riscv/translate.o: target/riscv/decode_insn32.inc.c
diff --git a/target/riscv/insn64.decode b/target/riscv/insn64.decode
new file mode 100644
index 0000000000..439d4e2c58
--- /dev/null
+++ b/target/riscv/insn64.decode
@@ -0,0 +1,25 @@
+#
+# RISC-V translation routines for the RV Instruction Set.
+#
+# Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
+# Bastian Koppelmann, kbastian@mail.uni-paderborn.de
+#
+# This program is free software; you can redistribute it and/or modify it
+# under the terms and conditions of the GNU General Public License,
+# version 2 or later, as published by the Free Software Foundation.
+#
+# This program is distributed in the hope it will be useful, but WITHOUT
+# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+# FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+# more details.
+#
+# You should have received a copy of the GNU General Public License along with
+# this program. If not, see <http://www.gnu.org/licenses/>.
+
+# This is concatenated with insn32.decode for risc64 targets.
+# Most of the fields and formats are there.
+
+# *** RV64I Base Instruction Set (in addition to RV32I) ***
+lwu ............ ..... 110 ..... 0000011 @i
+ld ............ ..... 011 ..... 0000011 @i
+sd ....... ..... ..... 011 ..... 0100011 @s
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
index f3b88ebb69..39a20a70e8 100644
--- a/target/riscv/insn_trans/trans_rvi.inc.c
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -130,3 +130,23 @@ static bool trans_sw(DisasContext *ctx, arg_sw *a)
gen_store(ctx, OPC_RISC_SW, a->rs1, a->rs2, a->imm);
return true;
}
+
+#ifdef TARGET_RISCV64
+static bool trans_lwu(DisasContext *ctx, arg_lwu *a)
+{
+ gen_load(ctx, OPC_RISC_LWU, a->rd, a->rs1, a->imm);
+ return true;
+}
+
+static bool trans_ld(DisasContext *ctx, arg_ld *a)
+{
+ gen_load(ctx, OPC_RISC_LD, a->rd, a->rs1, a->imm);
+ return true;
+}
+
+static bool trans_sd(DisasContext *ctx, arg_sd *a)
+{
+ gen_store(ctx, OPC_RISC_SD, a->rs1, a->rs2, a->imm);
+ return true;
+}
+#endif
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 909f7cd013..244855c82d 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -1701,13 +1701,6 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
imm = GET_IMM(ctx->opcode);
switch (op) {
- case OPC_RISC_LOAD:
- gen_load(ctx, MASK_OP_LOAD(ctx->opcode), rd, rs1, imm);
- break;
- case OPC_RISC_STORE:
- gen_store(ctx, MASK_OP_STORE(ctx->opcode), rs1, rs2,
- GET_STORE_IMM(ctx->opcode));
- break;
case OPC_RISC_ARITH_IMM:
#if defined(TARGET_RISCV64)
case OPC_RISC_ARITH_IMM_W:
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 06/35] target/riscv: Convert RVXI arithmetic insns to decodetree
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
we cannot remove the call to gen_arith() in decode_RV32_64G() since it
is used to translate multiply instructions.
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- moved 64-bit only insn to insn64.decode
- dropped insn argument of trans_foo functions
target/riscv/insn32.decode | 25 ++++
target/riscv/insn64.decode | 13 ++
target/riscv/insn_trans/trans_rvi.inc.c | 168 ++++++++++++++++++++++++
target/riscv/translate.c | 9 --
4 files changed, 206 insertions(+), 9 deletions(-)
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index 076de873c4..4fd88f48d2 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -21,6 +21,8 @@
%rs1 15:5
%rd 7:5
+%sh6 20:6
+
# immediates:
%imm_i 20:s12
%imm_s 25:s7 7:5
@@ -30,14 +32,18 @@
# Argument sets:
&b imm rs2 rs1
+&shift shamt rs1 rd
# Formats 32:
+@r ....... ..... ..... ... ..... ....... %rs2 %rs1 %rd
@i ............ ..... ... ..... ....... imm=%imm_i %rs1 %rd
@b ....... ..... ..... ... ..... ....... &b imm=%imm_b %rs2 %rs1
@s ....... ..... ..... ... ..... ....... imm=%imm_s %rs2 %rs1
@u .................... ..... ....... imm=%imm_u %rd
@j .................... ..... ....... imm=%imm_j %rd
+@sh6 ...... ...... ..... ... ..... ....... &shift shamt=%sh6 %rs1 %rd
+
# *** RV32I Base Instruction Set ***
lui .................... ..... 0110111 @u
auipc .................... ..... 0010111 @u
@@ -57,3 +63,22 @@ lhu ............ ..... 101 ..... 0000011 @i
sb ....... ..... ..... 000 ..... 0100011 @s
sh ....... ..... ..... 001 ..... 0100011 @s
sw ....... ..... ..... 010 ..... 0100011 @s
+addi ............ ..... 000 ..... 0010011 @i
+slti ............ ..... 010 ..... 0010011 @i
+sltiu ............ ..... 011 ..... 0010011 @i
+xori ............ ..... 100 ..... 0010011 @i
+ori ............ ..... 110 ..... 0010011 @i
+andi ............ ..... 111 ..... 0010011 @i
+slli 000000 ...... ..... 001 ..... 0010011 @sh6
+srli 000000 ...... ..... 101 ..... 0010011 @sh6
+srai 010000 ...... ..... 101 ..... 0010011 @sh6
+add 0000000 ..... ..... 000 ..... 0110011 @r
+sub 0100000 ..... ..... 000 ..... 0110011 @r
+sll 0000000 ..... ..... 001 ..... 0110011 @r
+slt 0000000 ..... ..... 010 ..... 0110011 @r
+sltu 0000000 ..... ..... 011 ..... 0110011 @r
+xor 0000000 ..... ..... 100 ..... 0110011 @r
+srl 0000000 ..... ..... 101 ..... 0110011 @r
+sra 0100000 ..... ..... 101 ..... 0110011 @r
+or 0000000 ..... ..... 110 ..... 0110011 @r
+and 0000000 ..... ..... 111 ..... 0110011 @r
diff --git a/target/riscv/insn64.decode b/target/riscv/insn64.decode
index 439d4e2c58..9a35f2aa19 100644
--- a/target/riscv/insn64.decode
+++ b/target/riscv/insn64.decode
@@ -19,7 +19,20 @@
# This is concatenated with insn32.decode for risc64 targets.
# Most of the fields and formats are there.
+%sh5 20:5
+
+@sh5 ....... ..... ..... ... ..... ....... &shift shamt=%sh5 %rs1 %rd
+
# *** RV64I Base Instruction Set (in addition to RV32I) ***
lwu ............ ..... 110 ..... 0000011 @i
ld ............ ..... 011 ..... 0000011 @i
sd ....... ..... ..... 011 ..... 0100011 @s
+addiw ............ ..... 000 ..... 0011011 @i
+slliw 0000000 ..... ..... 001 ..... 0011011 @sh5
+srliw 0000000 ..... ..... 101 ..... 0011011 @sh5
+sraiw 0100000 ..... ..... 101 ..... 0011011 @sh5
+addw 0000000 ..... ..... 000 ..... 0111011 @r
+subw 0100000 ..... ..... 000 ..... 0111011 @r
+sllw 0000000 ..... ..... 001 ..... 0111011 @r
+srlw 0000000 ..... ..... 101 ..... 0111011 @r
+sraw 0100000 ..... ..... 101 ..... 0111011 @r
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
index 39a20a70e8..01f751650a 100644
--- a/target/riscv/insn_trans/trans_rvi.inc.c
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -150,3 +150,171 @@ static bool trans_sd(DisasContext *ctx, arg_sd *a)
return true;
}
#endif
+
+static bool trans_addi(DisasContext *ctx, arg_addi *a)
+{
+ gen_arith_imm(ctx, OPC_RISC_ADDI, a->rd, a->rs1, a->imm);
+ return true;
+}
+
+static bool trans_slti(DisasContext *ctx, arg_slti *a)
+{
+ gen_arith_imm(ctx, OPC_RISC_SLTI, a->rd, a->rs1, a->imm);
+ return true;
+}
+
+static bool trans_sltiu(DisasContext *ctx, arg_sltiu *a)
+{
+ gen_arith_imm(ctx, OPC_RISC_SLTIU, a->rd, a->rs1, a->imm);
+ return true;
+}
+
+static bool trans_xori(DisasContext *ctx, arg_xori *a)
+{
+ gen_arith_imm(ctx, OPC_RISC_XORI, a->rd, a->rs1, a->imm);
+ return true;
+}
+static bool trans_ori(DisasContext *ctx, arg_ori *a)
+{
+ gen_arith_imm(ctx, OPC_RISC_ORI, a->rd, a->rs1, a->imm);
+ return true;
+}
+static bool trans_andi(DisasContext *ctx, arg_andi *a)
+{
+ gen_arith_imm(ctx, OPC_RISC_ANDI, a->rd, a->rs1, a->imm);
+ return true;
+}
+static bool trans_slli(DisasContext *ctx, arg_slli *a)
+{
+ gen_arith_imm(ctx, OPC_RISC_SLLI, a->rd, a->rs1, a->shamt);
+ return true;
+}
+
+static bool trans_srli(DisasContext *ctx, arg_srli *a)
+{
+ gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_I, a->rd, a->rs1, a->shamt);
+ return true;
+}
+
+static bool trans_srai(DisasContext *ctx, arg_srai *a)
+{
+ gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_I, a->rd, a->rs1, a->shamt | 0x400);
+ return true;
+}
+
+static bool trans_add(DisasContext *ctx, arg_add *a)
+{
+ gen_arith(ctx, OPC_RISC_ADD, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_sub(DisasContext *ctx, arg_sub *a)
+{
+ gen_arith(ctx, OPC_RISC_SUB, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_sll(DisasContext *ctx, arg_sll *a)
+{
+ gen_arith(ctx, OPC_RISC_SLL, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_slt(DisasContext *ctx, arg_slt *a)
+{
+ gen_arith(ctx, OPC_RISC_SLT, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_sltu(DisasContext *ctx, arg_sltu *a)
+{
+ gen_arith(ctx, OPC_RISC_SLTU, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_xor(DisasContext *ctx, arg_xor *a)
+{
+ gen_arith(ctx, OPC_RISC_XOR, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_srl(DisasContext *ctx, arg_srl *a)
+{
+ gen_arith(ctx, OPC_RISC_SRL, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_sra(DisasContext *ctx, arg_sra *a)
+{
+ gen_arith(ctx, OPC_RISC_SRA, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_or(DisasContext *ctx, arg_or *a)
+{
+ gen_arith(ctx, OPC_RISC_OR, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_and(DisasContext *ctx, arg_and *a)
+{
+ gen_arith(ctx, OPC_RISC_AND, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+#ifdef TARGET_RISCV64
+static bool trans_addiw(DisasContext *ctx, arg_addiw *a)
+{
+ gen_arith_imm(ctx, OPC_RISC_ADDIW, a->rd, a->rs1, a->imm);
+ return true;
+}
+
+static bool trans_slliw(DisasContext *ctx, arg_slliw *a)
+{
+ gen_arith_imm(ctx, OPC_RISC_SLLIW, a->rd, a->rs1, a->shamt);
+ return true;
+}
+
+static bool trans_srliw(DisasContext *ctx, arg_srliw *a)
+{
+ gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_IW, a->rd, a->rs1, a->shamt);
+ return true;
+}
+
+static bool trans_sraiw(DisasContext *ctx, arg_sraiw *a)
+{
+ gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_IW , a->rd, a->rs1,
+ a->shamt | 0x400);
+ return true;
+}
+
+static bool trans_addw(DisasContext *ctx, arg_addw *a)
+{
+ gen_arith(ctx, OPC_RISC_ADDW, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_subw(DisasContext *ctx, arg_subw *a)
+{
+ gen_arith(ctx, OPC_RISC_SUBW, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_sllw(DisasContext *ctx, arg_sllw *a)
+{
+ gen_arith(ctx, OPC_RISC_SLLW, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_srlw(DisasContext *ctx, arg_srlw *a)
+{
+ gen_arith(ctx, OPC_RISC_SRLW, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_sraw(DisasContext *ctx, arg_sraw *a)
+{
+ gen_arith(ctx, OPC_RISC_SRAW, a->rd, a->rs1, a->rs2);
+ return true;
+}
+#endif
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 244855c82d..855c241e97 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -1701,15 +1701,6 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
imm = GET_IMM(ctx->opcode);
switch (op) {
- case OPC_RISC_ARITH_IMM:
-#if defined(TARGET_RISCV64)
- case OPC_RISC_ARITH_IMM_W:
-#endif
- if (rd == 0) {
- break; /* NOP */
- }
- gen_arith_imm(ctx, MASK_OP_ARITH_IMM(ctx->opcode), rd, rs1, imm);
- break;
case OPC_RISC_ARITH:
#if defined(TARGET_RISCV64)
case OPC_RISC_ARITH_W:
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 06/35] target/riscv: Convert RVXI arithmetic insns to decodetree
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
we cannot remove the call to gen_arith() in decode_RV32_64G() since it
is used to translate multiply instructions.
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- moved 64-bit only insn to insn64.decode
- dropped insn argument of trans_foo functions
target/riscv/insn32.decode | 25 ++++
target/riscv/insn64.decode | 13 ++
target/riscv/insn_trans/trans_rvi.inc.c | 168 ++++++++++++++++++++++++
target/riscv/translate.c | 9 --
4 files changed, 206 insertions(+), 9 deletions(-)
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index 076de873c4..4fd88f48d2 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -21,6 +21,8 @@
%rs1 15:5
%rd 7:5
+%sh6 20:6
+
# immediates:
%imm_i 20:s12
%imm_s 25:s7 7:5
@@ -30,14 +32,18 @@
# Argument sets:
&b imm rs2 rs1
+&shift shamt rs1 rd
# Formats 32:
+@r ....... ..... ..... ... ..... ....... %rs2 %rs1 %rd
@i ............ ..... ... ..... ....... imm=%imm_i %rs1 %rd
@b ....... ..... ..... ... ..... ....... &b imm=%imm_b %rs2 %rs1
@s ....... ..... ..... ... ..... ....... imm=%imm_s %rs2 %rs1
@u .................... ..... ....... imm=%imm_u %rd
@j .................... ..... ....... imm=%imm_j %rd
+@sh6 ...... ...... ..... ... ..... ....... &shift shamt=%sh6 %rs1 %rd
+
# *** RV32I Base Instruction Set ***
lui .................... ..... 0110111 @u
auipc .................... ..... 0010111 @u
@@ -57,3 +63,22 @@ lhu ............ ..... 101 ..... 0000011 @i
sb ....... ..... ..... 000 ..... 0100011 @s
sh ....... ..... ..... 001 ..... 0100011 @s
sw ....... ..... ..... 010 ..... 0100011 @s
+addi ............ ..... 000 ..... 0010011 @i
+slti ............ ..... 010 ..... 0010011 @i
+sltiu ............ ..... 011 ..... 0010011 @i
+xori ............ ..... 100 ..... 0010011 @i
+ori ............ ..... 110 ..... 0010011 @i
+andi ............ ..... 111 ..... 0010011 @i
+slli 000000 ...... ..... 001 ..... 0010011 @sh6
+srli 000000 ...... ..... 101 ..... 0010011 @sh6
+srai 010000 ...... ..... 101 ..... 0010011 @sh6
+add 0000000 ..... ..... 000 ..... 0110011 @r
+sub 0100000 ..... ..... 000 ..... 0110011 @r
+sll 0000000 ..... ..... 001 ..... 0110011 @r
+slt 0000000 ..... ..... 010 ..... 0110011 @r
+sltu 0000000 ..... ..... 011 ..... 0110011 @r
+xor 0000000 ..... ..... 100 ..... 0110011 @r
+srl 0000000 ..... ..... 101 ..... 0110011 @r
+sra 0100000 ..... ..... 101 ..... 0110011 @r
+or 0000000 ..... ..... 110 ..... 0110011 @r
+and 0000000 ..... ..... 111 ..... 0110011 @r
diff --git a/target/riscv/insn64.decode b/target/riscv/insn64.decode
index 439d4e2c58..9a35f2aa19 100644
--- a/target/riscv/insn64.decode
+++ b/target/riscv/insn64.decode
@@ -19,7 +19,20 @@
# This is concatenated with insn32.decode for risc64 targets.
# Most of the fields and formats are there.
+%sh5 20:5
+
+@sh5 ....... ..... ..... ... ..... ....... &shift shamt=%sh5 %rs1 %rd
+
# *** RV64I Base Instruction Set (in addition to RV32I) ***
lwu ............ ..... 110 ..... 0000011 @i
ld ............ ..... 011 ..... 0000011 @i
sd ....... ..... ..... 011 ..... 0100011 @s
+addiw ............ ..... 000 ..... 0011011 @i
+slliw 0000000 ..... ..... 001 ..... 0011011 @sh5
+srliw 0000000 ..... ..... 101 ..... 0011011 @sh5
+sraiw 0100000 ..... ..... 101 ..... 0011011 @sh5
+addw 0000000 ..... ..... 000 ..... 0111011 @r
+subw 0100000 ..... ..... 000 ..... 0111011 @r
+sllw 0000000 ..... ..... 001 ..... 0111011 @r
+srlw 0000000 ..... ..... 101 ..... 0111011 @r
+sraw 0100000 ..... ..... 101 ..... 0111011 @r
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
index 39a20a70e8..01f751650a 100644
--- a/target/riscv/insn_trans/trans_rvi.inc.c
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -150,3 +150,171 @@ static bool trans_sd(DisasContext *ctx, arg_sd *a)
return true;
}
#endif
+
+static bool trans_addi(DisasContext *ctx, arg_addi *a)
+{
+ gen_arith_imm(ctx, OPC_RISC_ADDI, a->rd, a->rs1, a->imm);
+ return true;
+}
+
+static bool trans_slti(DisasContext *ctx, arg_slti *a)
+{
+ gen_arith_imm(ctx, OPC_RISC_SLTI, a->rd, a->rs1, a->imm);
+ return true;
+}
+
+static bool trans_sltiu(DisasContext *ctx, arg_sltiu *a)
+{
+ gen_arith_imm(ctx, OPC_RISC_SLTIU, a->rd, a->rs1, a->imm);
+ return true;
+}
+
+static bool trans_xori(DisasContext *ctx, arg_xori *a)
+{
+ gen_arith_imm(ctx, OPC_RISC_XORI, a->rd, a->rs1, a->imm);
+ return true;
+}
+static bool trans_ori(DisasContext *ctx, arg_ori *a)
+{
+ gen_arith_imm(ctx, OPC_RISC_ORI, a->rd, a->rs1, a->imm);
+ return true;
+}
+static bool trans_andi(DisasContext *ctx, arg_andi *a)
+{
+ gen_arith_imm(ctx, OPC_RISC_ANDI, a->rd, a->rs1, a->imm);
+ return true;
+}
+static bool trans_slli(DisasContext *ctx, arg_slli *a)
+{
+ gen_arith_imm(ctx, OPC_RISC_SLLI, a->rd, a->rs1, a->shamt);
+ return true;
+}
+
+static bool trans_srli(DisasContext *ctx, arg_srli *a)
+{
+ gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_I, a->rd, a->rs1, a->shamt);
+ return true;
+}
+
+static bool trans_srai(DisasContext *ctx, arg_srai *a)
+{
+ gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_I, a->rd, a->rs1, a->shamt | 0x400);
+ return true;
+}
+
+static bool trans_add(DisasContext *ctx, arg_add *a)
+{
+ gen_arith(ctx, OPC_RISC_ADD, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_sub(DisasContext *ctx, arg_sub *a)
+{
+ gen_arith(ctx, OPC_RISC_SUB, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_sll(DisasContext *ctx, arg_sll *a)
+{
+ gen_arith(ctx, OPC_RISC_SLL, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_slt(DisasContext *ctx, arg_slt *a)
+{
+ gen_arith(ctx, OPC_RISC_SLT, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_sltu(DisasContext *ctx, arg_sltu *a)
+{
+ gen_arith(ctx, OPC_RISC_SLTU, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_xor(DisasContext *ctx, arg_xor *a)
+{
+ gen_arith(ctx, OPC_RISC_XOR, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_srl(DisasContext *ctx, arg_srl *a)
+{
+ gen_arith(ctx, OPC_RISC_SRL, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_sra(DisasContext *ctx, arg_sra *a)
+{
+ gen_arith(ctx, OPC_RISC_SRA, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_or(DisasContext *ctx, arg_or *a)
+{
+ gen_arith(ctx, OPC_RISC_OR, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_and(DisasContext *ctx, arg_and *a)
+{
+ gen_arith(ctx, OPC_RISC_AND, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+#ifdef TARGET_RISCV64
+static bool trans_addiw(DisasContext *ctx, arg_addiw *a)
+{
+ gen_arith_imm(ctx, OPC_RISC_ADDIW, a->rd, a->rs1, a->imm);
+ return true;
+}
+
+static bool trans_slliw(DisasContext *ctx, arg_slliw *a)
+{
+ gen_arith_imm(ctx, OPC_RISC_SLLIW, a->rd, a->rs1, a->shamt);
+ return true;
+}
+
+static bool trans_srliw(DisasContext *ctx, arg_srliw *a)
+{
+ gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_IW, a->rd, a->rs1, a->shamt);
+ return true;
+}
+
+static bool trans_sraiw(DisasContext *ctx, arg_sraiw *a)
+{
+ gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_IW , a->rd, a->rs1,
+ a->shamt | 0x400);
+ return true;
+}
+
+static bool trans_addw(DisasContext *ctx, arg_addw *a)
+{
+ gen_arith(ctx, OPC_RISC_ADDW, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_subw(DisasContext *ctx, arg_subw *a)
+{
+ gen_arith(ctx, OPC_RISC_SUBW, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_sllw(DisasContext *ctx, arg_sllw *a)
+{
+ gen_arith(ctx, OPC_RISC_SLLW, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_srlw(DisasContext *ctx, arg_srlw *a)
+{
+ gen_arith(ctx, OPC_RISC_SRLW, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_sraw(DisasContext *ctx, arg_sraw *a)
+{
+ gen_arith(ctx, OPC_RISC_SRAW, a->rd, a->rs1, a->rs2);
+ return true;
+}
+#endif
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 244855c82d..855c241e97 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -1701,15 +1701,6 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
imm = GET_IMM(ctx->opcode);
switch (op) {
- case OPC_RISC_ARITH_IMM:
-#if defined(TARGET_RISCV64)
- case OPC_RISC_ARITH_IMM_W:
-#endif
- if (rd == 0) {
- break; /* NOP */
- }
- gen_arith_imm(ctx, MASK_OP_ARITH_IMM(ctx->opcode), rd, rs1, imm);
- break;
case OPC_RISC_ARITH:
#if defined(TARGET_RISCV64)
case OPC_RISC_ARITH_W:
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 07/35] target/riscv: Convert RVXI fence insns to decodetree
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- removed %pred/%succ
- dropped insn argument of trans_foo functions
target/riscv/insn32.decode | 2 ++
target/riscv/insn_trans/trans_rvi.inc.c | 21 +++++++++++++++++++++
target/riscv/translate.c | 14 --------------
3 files changed, 23 insertions(+), 14 deletions(-)
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index 4fd88f48d2..6d750b4c5a 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -82,3 +82,5 @@ srl 0000000 ..... ..... 101 ..... 0110011 @r
sra 0100000 ..... ..... 101 ..... 0110011 @r
or 0000000 ..... ..... 110 ..... 0110011 @r
and 0000000 ..... ..... 111 ..... 0110011 @r
+fence ---- pred:4 succ:4 ----- 000 ----- 0001111
+fence_i ---- ---- ---- ----- 001 ----- 0001111
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
index 01f751650a..a149e913b1 100644
--- a/target/riscv/insn_trans/trans_rvi.inc.c
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -318,3 +318,24 @@ static bool trans_sraw(DisasContext *ctx, arg_sraw *a)
return true;
}
#endif
+
+static bool trans_fence(DisasContext *ctx, arg_fence *a)
+{
+#ifndef CONFIG_USER_ONLY
+ /* FENCE is a full memory barrier. */
+ tcg_gen_mb(TCG_MO_ALL | TCG_BAR_SC);
+#endif
+ return true;
+}
+
+static bool trans_fence_i(DisasContext *ctx, arg_fence_i *a)
+{
+#ifndef CONFIG_USER_ONLY
+ /* FENCE_I is a no-op in QEMU,
+ * however we need to end the translation block */
+ tcg_gen_movi_tl(cpu_pc, ctx->pc_succ_insn);
+ tcg_gen_exit_tb(NULL, 0);
+ ctx->base.is_jmp = DISAS_NORETURN;
+#endif
+ return true;
+}
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 855c241e97..80f18fb6aa 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -1740,20 +1740,6 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
gen_fp_arith(ctx, MASK_OP_FP_ARITH(ctx->opcode), rd, rs1, rs2,
GET_RM(ctx->opcode));
break;
- case OPC_RISC_FENCE:
-#ifndef CONFIG_USER_ONLY
- if (ctx->opcode & 0x1000) {
- /* FENCE_I is a no-op in QEMU,
- * however we need to end the translation block */
- tcg_gen_movi_tl(cpu_pc, ctx->pc_succ_insn);
- tcg_gen_exit_tb(NULL, 0);
- ctx->base.is_jmp = DISAS_NORETURN;
- } else {
- /* FENCE is a full memory barrier. */
- tcg_gen_mb(TCG_MO_ALL | TCG_BAR_SC);
- }
-#endif
- break;
case OPC_RISC_SYSTEM:
gen_system(env, ctx, MASK_OP_SYSTEM(ctx->opcode), rd, rs1,
(ctx->opcode & 0xFFF00000) >> 20);
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 07/35] target/riscv: Convert RVXI fence insns to decodetree
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- removed %pred/%succ
- dropped insn argument of trans_foo functions
target/riscv/insn32.decode | 2 ++
target/riscv/insn_trans/trans_rvi.inc.c | 21 +++++++++++++++++++++
target/riscv/translate.c | 14 --------------
3 files changed, 23 insertions(+), 14 deletions(-)
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index 4fd88f48d2..6d750b4c5a 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -82,3 +82,5 @@ srl 0000000 ..... ..... 101 ..... 0110011 @r
sra 0100000 ..... ..... 101 ..... 0110011 @r
or 0000000 ..... ..... 110 ..... 0110011 @r
and 0000000 ..... ..... 111 ..... 0110011 @r
+fence ---- pred:4 succ:4 ----- 000 ----- 0001111
+fence_i ---- ---- ---- ----- 001 ----- 0001111
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
index 01f751650a..a149e913b1 100644
--- a/target/riscv/insn_trans/trans_rvi.inc.c
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -318,3 +318,24 @@ static bool trans_sraw(DisasContext *ctx, arg_sraw *a)
return true;
}
#endif
+
+static bool trans_fence(DisasContext *ctx, arg_fence *a)
+{
+#ifndef CONFIG_USER_ONLY
+ /* FENCE is a full memory barrier. */
+ tcg_gen_mb(TCG_MO_ALL | TCG_BAR_SC);
+#endif
+ return true;
+}
+
+static bool trans_fence_i(DisasContext *ctx, arg_fence_i *a)
+{
+#ifndef CONFIG_USER_ONLY
+ /* FENCE_I is a no-op in QEMU,
+ * however we need to end the translation block */
+ tcg_gen_movi_tl(cpu_pc, ctx->pc_succ_insn);
+ tcg_gen_exit_tb(NULL, 0);
+ ctx->base.is_jmp = DISAS_NORETURN;
+#endif
+ return true;
+}
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 855c241e97..80f18fb6aa 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -1740,20 +1740,6 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
gen_fp_arith(ctx, MASK_OP_FP_ARITH(ctx->opcode), rd, rs1, rs2,
GET_RM(ctx->opcode));
break;
- case OPC_RISC_FENCE:
-#ifndef CONFIG_USER_ONLY
- if (ctx->opcode & 0x1000) {
- /* FENCE_I is a no-op in QEMU,
- * however we need to end the translation block */
- tcg_gen_movi_tl(cpu_pc, ctx->pc_succ_insn);
- tcg_gen_exit_tb(NULL, 0);
- ctx->base.is_jmp = DISAS_NORETURN;
- } else {
- /* FENCE is a full memory barrier. */
- tcg_gen_mb(TCG_MO_ALL | TCG_BAR_SC);
- }
-#endif
- break;
case OPC_RISC_SYSTEM:
gen_system(env, ctx, MASK_OP_SYSTEM(ctx->opcode), rd, rs1,
(ctx->opcode & 0xFFF00000) >> 20);
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 08/35] target/riscv: Convert RVXI csr insns to decodetree
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- dropped insn argument of trans_foo functions
target/riscv/insn32.decode | 8 +++
target/riscv/insn_trans/trans_rvi.inc.c | 79 +++++++++++++++++++++++++
target/riscv/translate.c | 43 +-------------
3 files changed, 88 insertions(+), 42 deletions(-)
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index 6d750b4c5a..b67d3a9437 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -22,6 +22,7 @@
%rd 7:5
%sh6 20:6
+%csr 20:12
# immediates:
%imm_i 20:s12
@@ -43,6 +44,7 @@
@j .................... ..... ....... imm=%imm_j %rd
@sh6 ...... ...... ..... ... ..... ....... &shift shamt=%sh6 %rs1 %rd
+@csr ............ ..... ... ..... ....... %csr %rs1 %rd
# *** RV32I Base Instruction Set ***
lui .................... ..... 0110111 @u
@@ -84,3 +86,9 @@ or 0000000 ..... ..... 110 ..... 0110011 @r
and 0000000 ..... ..... 111 ..... 0110011 @r
fence ---- pred:4 succ:4 ----- 000 ----- 0001111
fence_i ---- ---- ---- ----- 001 ----- 0001111
+csrrw ............ ..... 001 ..... 1110011 @csr
+csrrs ............ ..... 010 ..... 1110011 @csr
+csrrc ............ ..... 011 ..... 1110011 @csr
+csrrwi ............ ..... 101 ..... 1110011 @csr
+csrrsi ............ ..... 110 ..... 1110011 @csr
+csrrci ............ ..... 111 ..... 1110011 @csr
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
index a149e913b1..09e7a0052a 100644
--- a/target/riscv/insn_trans/trans_rvi.inc.c
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -339,3 +339,82 @@ static bool trans_fence_i(DisasContext *ctx, arg_fence_i *a)
#endif
return true;
}
+
+#define RISCV_OP_CSR_PRE do {\
+ source1 = tcg_temp_new(); \
+ csr_store = tcg_temp_new(); \
+ dest = tcg_temp_new(); \
+ rs1_pass = tcg_temp_new(); \
+ gen_get_gpr(source1, a->rs1); \
+ tcg_gen_movi_tl(cpu_pc, ctx->base.pc_next); \
+ tcg_gen_movi_tl(rs1_pass, a->rs1); \
+ tcg_gen_movi_tl(csr_store, a->csr); \
+ gen_io_start();\
+} while (0)
+
+#define RISCV_OP_CSR_POST do {\
+ gen_io_end(); \
+ gen_set_gpr(a->rd, dest); \
+ tcg_gen_movi_tl(cpu_pc, ctx->pc_succ_insn); \
+ tcg_gen_exit_tb(NULL, 0); \
+ ctx->base.is_jmp = DISAS_NORETURN; \
+ tcg_temp_free(source1); \
+ tcg_temp_free(csr_store); \
+ tcg_temp_free(dest); \
+ tcg_temp_free(rs1_pass); \
+} while (0)
+
+
+static bool trans_csrrw(DisasContext *ctx, arg_csrrw *a)
+{
+ TCGv source1, csr_store, dest, rs1_pass;
+ RISCV_OP_CSR_PRE;
+ gen_helper_csrrw(dest, cpu_env, source1, csr_store);
+ RISCV_OP_CSR_POST;
+ return true;
+}
+
+static bool trans_csrrs(DisasContext *ctx, arg_csrrs *a)
+{
+ TCGv source1, csr_store, dest, rs1_pass;
+ RISCV_OP_CSR_PRE;
+ gen_helper_csrrs(dest, cpu_env, source1, csr_store, rs1_pass);
+ RISCV_OP_CSR_POST;
+ return true;
+}
+
+static bool trans_csrrc(DisasContext *ctx, arg_csrrc *a)
+{
+ TCGv source1, csr_store, dest, rs1_pass;
+ RISCV_OP_CSR_PRE;
+ gen_helper_csrrc(dest, cpu_env, source1, csr_store, rs1_pass);
+ RISCV_OP_CSR_POST;
+ return true;
+}
+
+static bool trans_csrrwi(DisasContext *ctx, arg_csrrwi *a)
+{
+ TCGv source1, csr_store, dest, rs1_pass;
+ RISCV_OP_CSR_PRE;
+ gen_helper_csrrw(dest, cpu_env, rs1_pass, csr_store);
+ RISCV_OP_CSR_POST;
+ return true;
+}
+
+static bool trans_csrrsi(DisasContext *ctx, arg_csrrsi *a)
+{
+ TCGv source1, csr_store, dest, rs1_pass;
+ RISCV_OP_CSR_PRE;
+ gen_helper_csrrs(dest, cpu_env, rs1_pass, csr_store, rs1_pass);
+ RISCV_OP_CSR_POST;
+ return true;
+}
+
+static bool trans_csrrci(DisasContext *ctx, arg_csrrci *a)
+{
+ TCGv source1, csr_store, dest, rs1_pass;
+ RISCV_OP_CSR_PRE;
+ gen_helper_csrrc(dest, cpu_env, rs1_pass, csr_store, rs1_pass);
+ RISCV_OP_CSR_POST;
+ return true;
+}
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 80f18fb6aa..be03acd066 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -1277,16 +1277,11 @@ static void gen_fp_arith(DisasContext *ctx, uint32_t opc, int rd,
static void gen_system(CPURISCVState *env, DisasContext *ctx, uint32_t opc,
int rd, int rs1, int csr)
{
- TCGv source1, csr_store, dest, rs1_pass, imm_rs1;
+ TCGv source1, dest;
source1 = tcg_temp_new();
- csr_store = tcg_temp_new();
dest = tcg_temp_new();
- rs1_pass = tcg_temp_new();
- imm_rs1 = tcg_temp_new();
gen_get_gpr(source1, rs1);
tcg_gen_movi_tl(cpu_pc, ctx->base.pc_next);
- tcg_gen_movi_tl(rs1_pass, rs1);
- tcg_gen_movi_tl(csr_store, csr); /* copy into temp reg to feed to helper */
#ifndef CONFIG_USER_ONLY
/* Extract funct7 value and check whether it matches SFENCE.VMA */
@@ -1349,45 +1344,9 @@ static void gen_system(CPURISCVState *env, DisasContext *ctx, uint32_t opc,
break;
}
break;
- default:
- tcg_gen_movi_tl(imm_rs1, rs1);
- gen_io_start();
- switch (opc) {
- case OPC_RISC_CSRRW:
- gen_helper_csrrw(dest, cpu_env, source1, csr_store);
- break;
- case OPC_RISC_CSRRS:
- gen_helper_csrrs(dest, cpu_env, source1, csr_store, rs1_pass);
- break;
- case OPC_RISC_CSRRC:
- gen_helper_csrrc(dest, cpu_env, source1, csr_store, rs1_pass);
- break;
- case OPC_RISC_CSRRWI:
- gen_helper_csrrw(dest, cpu_env, imm_rs1, csr_store);
- break;
- case OPC_RISC_CSRRSI:
- gen_helper_csrrs(dest, cpu_env, imm_rs1, csr_store, rs1_pass);
- break;
- case OPC_RISC_CSRRCI:
- gen_helper_csrrc(dest, cpu_env, imm_rs1, csr_store, rs1_pass);
- break;
- default:
- gen_exception_illegal(ctx);
- return;
- }
- gen_io_end();
- gen_set_gpr(rd, dest);
- /* end tb since we may be changing priv modes, to get mmu_index right */
- tcg_gen_movi_tl(cpu_pc, ctx->pc_succ_insn);
- tcg_gen_exit_tb(NULL, 0); /* no chaining */
- ctx->base.is_jmp = DISAS_NORETURN;
- break;
}
tcg_temp_free(source1);
- tcg_temp_free(csr_store);
tcg_temp_free(dest);
- tcg_temp_free(rs1_pass);
- tcg_temp_free(imm_rs1);
}
static void decode_RV32_64C0(DisasContext *ctx)
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 08/35] target/riscv: Convert RVXI csr insns to decodetree
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- dropped insn argument of trans_foo functions
target/riscv/insn32.decode | 8 +++
target/riscv/insn_trans/trans_rvi.inc.c | 79 +++++++++++++++++++++++++
target/riscv/translate.c | 43 +-------------
3 files changed, 88 insertions(+), 42 deletions(-)
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index 6d750b4c5a..b67d3a9437 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -22,6 +22,7 @@
%rd 7:5
%sh6 20:6
+%csr 20:12
# immediates:
%imm_i 20:s12
@@ -43,6 +44,7 @@
@j .................... ..... ....... imm=%imm_j %rd
@sh6 ...... ...... ..... ... ..... ....... &shift shamt=%sh6 %rs1 %rd
+@csr ............ ..... ... ..... ....... %csr %rs1 %rd
# *** RV32I Base Instruction Set ***
lui .................... ..... 0110111 @u
@@ -84,3 +86,9 @@ or 0000000 ..... ..... 110 ..... 0110011 @r
and 0000000 ..... ..... 111 ..... 0110011 @r
fence ---- pred:4 succ:4 ----- 000 ----- 0001111
fence_i ---- ---- ---- ----- 001 ----- 0001111
+csrrw ............ ..... 001 ..... 1110011 @csr
+csrrs ............ ..... 010 ..... 1110011 @csr
+csrrc ............ ..... 011 ..... 1110011 @csr
+csrrwi ............ ..... 101 ..... 1110011 @csr
+csrrsi ............ ..... 110 ..... 1110011 @csr
+csrrci ............ ..... 111 ..... 1110011 @csr
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
index a149e913b1..09e7a0052a 100644
--- a/target/riscv/insn_trans/trans_rvi.inc.c
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -339,3 +339,82 @@ static bool trans_fence_i(DisasContext *ctx, arg_fence_i *a)
#endif
return true;
}
+
+#define RISCV_OP_CSR_PRE do {\
+ source1 = tcg_temp_new(); \
+ csr_store = tcg_temp_new(); \
+ dest = tcg_temp_new(); \
+ rs1_pass = tcg_temp_new(); \
+ gen_get_gpr(source1, a->rs1); \
+ tcg_gen_movi_tl(cpu_pc, ctx->base.pc_next); \
+ tcg_gen_movi_tl(rs1_pass, a->rs1); \
+ tcg_gen_movi_tl(csr_store, a->csr); \
+ gen_io_start();\
+} while (0)
+
+#define RISCV_OP_CSR_POST do {\
+ gen_io_end(); \
+ gen_set_gpr(a->rd, dest); \
+ tcg_gen_movi_tl(cpu_pc, ctx->pc_succ_insn); \
+ tcg_gen_exit_tb(NULL, 0); \
+ ctx->base.is_jmp = DISAS_NORETURN; \
+ tcg_temp_free(source1); \
+ tcg_temp_free(csr_store); \
+ tcg_temp_free(dest); \
+ tcg_temp_free(rs1_pass); \
+} while (0)
+
+
+static bool trans_csrrw(DisasContext *ctx, arg_csrrw *a)
+{
+ TCGv source1, csr_store, dest, rs1_pass;
+ RISCV_OP_CSR_PRE;
+ gen_helper_csrrw(dest, cpu_env, source1, csr_store);
+ RISCV_OP_CSR_POST;
+ return true;
+}
+
+static bool trans_csrrs(DisasContext *ctx, arg_csrrs *a)
+{
+ TCGv source1, csr_store, dest, rs1_pass;
+ RISCV_OP_CSR_PRE;
+ gen_helper_csrrs(dest, cpu_env, source1, csr_store, rs1_pass);
+ RISCV_OP_CSR_POST;
+ return true;
+}
+
+static bool trans_csrrc(DisasContext *ctx, arg_csrrc *a)
+{
+ TCGv source1, csr_store, dest, rs1_pass;
+ RISCV_OP_CSR_PRE;
+ gen_helper_csrrc(dest, cpu_env, source1, csr_store, rs1_pass);
+ RISCV_OP_CSR_POST;
+ return true;
+}
+
+static bool trans_csrrwi(DisasContext *ctx, arg_csrrwi *a)
+{
+ TCGv source1, csr_store, dest, rs1_pass;
+ RISCV_OP_CSR_PRE;
+ gen_helper_csrrw(dest, cpu_env, rs1_pass, csr_store);
+ RISCV_OP_CSR_POST;
+ return true;
+}
+
+static bool trans_csrrsi(DisasContext *ctx, arg_csrrsi *a)
+{
+ TCGv source1, csr_store, dest, rs1_pass;
+ RISCV_OP_CSR_PRE;
+ gen_helper_csrrs(dest, cpu_env, rs1_pass, csr_store, rs1_pass);
+ RISCV_OP_CSR_POST;
+ return true;
+}
+
+static bool trans_csrrci(DisasContext *ctx, arg_csrrci *a)
+{
+ TCGv source1, csr_store, dest, rs1_pass;
+ RISCV_OP_CSR_PRE;
+ gen_helper_csrrc(dest, cpu_env, rs1_pass, csr_store, rs1_pass);
+ RISCV_OP_CSR_POST;
+ return true;
+}
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 80f18fb6aa..be03acd066 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -1277,16 +1277,11 @@ static void gen_fp_arith(DisasContext *ctx, uint32_t opc, int rd,
static void gen_system(CPURISCVState *env, DisasContext *ctx, uint32_t opc,
int rd, int rs1, int csr)
{
- TCGv source1, csr_store, dest, rs1_pass, imm_rs1;
+ TCGv source1, dest;
source1 = tcg_temp_new();
- csr_store = tcg_temp_new();
dest = tcg_temp_new();
- rs1_pass = tcg_temp_new();
- imm_rs1 = tcg_temp_new();
gen_get_gpr(source1, rs1);
tcg_gen_movi_tl(cpu_pc, ctx->base.pc_next);
- tcg_gen_movi_tl(rs1_pass, rs1);
- tcg_gen_movi_tl(csr_store, csr); /* copy into temp reg to feed to helper */
#ifndef CONFIG_USER_ONLY
/* Extract funct7 value and check whether it matches SFENCE.VMA */
@@ -1349,45 +1344,9 @@ static void gen_system(CPURISCVState *env, DisasContext *ctx, uint32_t opc,
break;
}
break;
- default:
- tcg_gen_movi_tl(imm_rs1, rs1);
- gen_io_start();
- switch (opc) {
- case OPC_RISC_CSRRW:
- gen_helper_csrrw(dest, cpu_env, source1, csr_store);
- break;
- case OPC_RISC_CSRRS:
- gen_helper_csrrs(dest, cpu_env, source1, csr_store, rs1_pass);
- break;
- case OPC_RISC_CSRRC:
- gen_helper_csrrc(dest, cpu_env, source1, csr_store, rs1_pass);
- break;
- case OPC_RISC_CSRRWI:
- gen_helper_csrrw(dest, cpu_env, imm_rs1, csr_store);
- break;
- case OPC_RISC_CSRRSI:
- gen_helper_csrrs(dest, cpu_env, imm_rs1, csr_store, rs1_pass);
- break;
- case OPC_RISC_CSRRCI:
- gen_helper_csrrc(dest, cpu_env, imm_rs1, csr_store, rs1_pass);
- break;
- default:
- gen_exception_illegal(ctx);
- return;
- }
- gen_io_end();
- gen_set_gpr(rd, dest);
- /* end tb since we may be changing priv modes, to get mmu_index right */
- tcg_gen_movi_tl(cpu_pc, ctx->pc_succ_insn);
- tcg_gen_exit_tb(NULL, 0); /* no chaining */
- ctx->base.is_jmp = DISAS_NORETURN;
- break;
}
tcg_temp_free(source1);
- tcg_temp_free(csr_store);
tcg_temp_free(dest);
- tcg_temp_free(rs1_pass);
- tcg_temp_free(imm_rs1);
}
static void decode_RV32_64C0(DisasContext *ctx)
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 09/35] target/riscv: Convert RVXM insns to decodetree
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- moved 64-bit only insn to insn64.decode
- dropped insn argument of trans_foo functions
target/riscv/insn32.decode | 10 +++
target/riscv/insn64.decode | 7 ++
target/riscv/insn_trans/trans_rvm.inc.c | 100 ++++++++++++++++++++++++
target/riscv/translate.c | 10 +--
4 files changed, 118 insertions(+), 9 deletions(-)
create mode 100644 target/riscv/insn_trans/trans_rvm.inc.c
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index b67d3a9437..b37fece633 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -92,3 +92,13 @@ csrrc ............ ..... 011 ..... 1110011 @csr
csrrwi ............ ..... 101 ..... 1110011 @csr
csrrsi ............ ..... 110 ..... 1110011 @csr
csrrci ............ ..... 111 ..... 1110011 @csr
+
+# *** RV32M Standard Extension ***
+mul 0000001 ..... ..... 000 ..... 0110011 @r
+mulh 0000001 ..... ..... 001 ..... 0110011 @r
+mulhsu 0000001 ..... ..... 010 ..... 0110011 @r
+mulhu 0000001 ..... ..... 011 ..... 0110011 @r
+div 0000001 ..... ..... 100 ..... 0110011 @r
+divu 0000001 ..... ..... 101 ..... 0110011 @r
+rem 0000001 ..... ..... 110 ..... 0110011 @r
+remu 0000001 ..... ..... 111 ..... 0110011 @r
diff --git a/target/riscv/insn64.decode b/target/riscv/insn64.decode
index 9a35f2aa19..008f100546 100644
--- a/target/riscv/insn64.decode
+++ b/target/riscv/insn64.decode
@@ -36,3 +36,10 @@ subw 0100000 ..... ..... 000 ..... 0111011 @r
sllw 0000000 ..... ..... 001 ..... 0111011 @r
srlw 0000000 ..... ..... 101 ..... 0111011 @r
sraw 0100000 ..... ..... 101 ..... 0111011 @r
+
+# *** RV64M Standard Extension (in addition to RV32M) ***
+mulw 0000001 ..... ..... 000 ..... 0111011 @r
+divw 0000001 ..... ..... 100 ..... 0111011 @r
+divuw 0000001 ..... ..... 101 ..... 0111011 @r
+remw 0000001 ..... ..... 110 ..... 0111011 @r
+remuw 0000001 ..... ..... 111 ..... 0111011 @r
diff --git a/target/riscv/insn_trans/trans_rvm.inc.c b/target/riscv/insn_trans/trans_rvm.inc.c
new file mode 100644
index 0000000000..ec3197ede8
--- /dev/null
+++ b/target/riscv/insn_trans/trans_rvm.inc.c
@@ -0,0 +1,100 @@
+/*
+ * RISC-V translation routines for the RV64M Standard Extension.
+ *
+ * Copyright (c) 2016-2017 Sagar Karandikar, sagark@eecs.berkeley.edu
+ * Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
+ * Bastian Koppelmann, kbastian@mail.uni-paderborn.de
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2 or later, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+
+static bool trans_mul(DisasContext *ctx, arg_mul *a)
+{
+ gen_arith(ctx, OPC_RISC_MUL, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_mulh(DisasContext *ctx, arg_mulh *a)
+{
+ gen_arith(ctx, OPC_RISC_MULH, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_mulhsu(DisasContext *ctx, arg_mulhsu *a)
+{
+ gen_arith(ctx, OPC_RISC_MULHSU, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_mulhu(DisasContext *ctx, arg_mulhu *a)
+{
+ gen_arith(ctx, OPC_RISC_MULHU, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_div(DisasContext *ctx, arg_div *a)
+{
+ gen_arith(ctx, OPC_RISC_DIV, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_divu(DisasContext *ctx, arg_divu *a)
+{
+ gen_arith(ctx, OPC_RISC_DIVU, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_rem(DisasContext *ctx, arg_rem *a)
+{
+ gen_arith(ctx, OPC_RISC_REM, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_remu(DisasContext *ctx, arg_remu *a)
+{
+ gen_arith(ctx, OPC_RISC_REMU, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+#ifdef TARGET_RISCV64
+static bool trans_mulw(DisasContext *ctx, arg_mulw *a)
+{
+ gen_arith(ctx, OPC_RISC_MULW, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_divw(DisasContext *ctx, arg_divw *a)
+{
+ gen_arith(ctx, OPC_RISC_DIVW, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_divuw(DisasContext *ctx, arg_divuw *a)
+{
+ gen_arith(ctx, OPC_RISC_DIVUW, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_remw(DisasContext *ctx, arg_remw *a)
+{
+ gen_arith(ctx, OPC_RISC_REMW, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_remuw(DisasContext *ctx, arg_remuw *a)
+{
+ gen_arith(ctx, OPC_RISC_REMUW, a->rd, a->rs1, a->rs2);
+ return true;
+}
+#endif
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index be03acd066..8c876187b1 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -1639,6 +1639,7 @@ bool decode_insn32(DisasContext *ctx, uint32_t insn);
#include "decode_insn32.inc.c"
/* Include insn module translation function */
#include "insn_trans/trans_rvi.inc.c"
+#include "insn_trans/trans_rvm.inc.c"
static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
{
@@ -1660,15 +1661,6 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
imm = GET_IMM(ctx->opcode);
switch (op) {
- case OPC_RISC_ARITH:
-#if defined(TARGET_RISCV64)
- case OPC_RISC_ARITH_W:
-#endif
- if (rd == 0) {
- break; /* NOP */
- }
- gen_arith(ctx, MASK_OP_ARITH(ctx->opcode), rd, rs1, rs2);
- break;
case OPC_RISC_FP_LOAD:
gen_fp_load(ctx, MASK_OP_FP_LOAD(ctx->opcode), rd, rs1, imm);
break;
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 09/35] target/riscv: Convert RVXM insns to decodetree
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- moved 64-bit only insn to insn64.decode
- dropped insn argument of trans_foo functions
target/riscv/insn32.decode | 10 +++
target/riscv/insn64.decode | 7 ++
target/riscv/insn_trans/trans_rvm.inc.c | 100 ++++++++++++++++++++++++
target/riscv/translate.c | 10 +--
4 files changed, 118 insertions(+), 9 deletions(-)
create mode 100644 target/riscv/insn_trans/trans_rvm.inc.c
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index b67d3a9437..b37fece633 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -92,3 +92,13 @@ csrrc ............ ..... 011 ..... 1110011 @csr
csrrwi ............ ..... 101 ..... 1110011 @csr
csrrsi ............ ..... 110 ..... 1110011 @csr
csrrci ............ ..... 111 ..... 1110011 @csr
+
+# *** RV32M Standard Extension ***
+mul 0000001 ..... ..... 000 ..... 0110011 @r
+mulh 0000001 ..... ..... 001 ..... 0110011 @r
+mulhsu 0000001 ..... ..... 010 ..... 0110011 @r
+mulhu 0000001 ..... ..... 011 ..... 0110011 @r
+div 0000001 ..... ..... 100 ..... 0110011 @r
+divu 0000001 ..... ..... 101 ..... 0110011 @r
+rem 0000001 ..... ..... 110 ..... 0110011 @r
+remu 0000001 ..... ..... 111 ..... 0110011 @r
diff --git a/target/riscv/insn64.decode b/target/riscv/insn64.decode
index 9a35f2aa19..008f100546 100644
--- a/target/riscv/insn64.decode
+++ b/target/riscv/insn64.decode
@@ -36,3 +36,10 @@ subw 0100000 ..... ..... 000 ..... 0111011 @r
sllw 0000000 ..... ..... 001 ..... 0111011 @r
srlw 0000000 ..... ..... 101 ..... 0111011 @r
sraw 0100000 ..... ..... 101 ..... 0111011 @r
+
+# *** RV64M Standard Extension (in addition to RV32M) ***
+mulw 0000001 ..... ..... 000 ..... 0111011 @r
+divw 0000001 ..... ..... 100 ..... 0111011 @r
+divuw 0000001 ..... ..... 101 ..... 0111011 @r
+remw 0000001 ..... ..... 110 ..... 0111011 @r
+remuw 0000001 ..... ..... 111 ..... 0111011 @r
diff --git a/target/riscv/insn_trans/trans_rvm.inc.c b/target/riscv/insn_trans/trans_rvm.inc.c
new file mode 100644
index 0000000000..ec3197ede8
--- /dev/null
+++ b/target/riscv/insn_trans/trans_rvm.inc.c
@@ -0,0 +1,100 @@
+/*
+ * RISC-V translation routines for the RV64M Standard Extension.
+ *
+ * Copyright (c) 2016-2017 Sagar Karandikar, sagark@eecs.berkeley.edu
+ * Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
+ * Bastian Koppelmann, kbastian@mail.uni-paderborn.de
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2 or later, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+
+static bool trans_mul(DisasContext *ctx, arg_mul *a)
+{
+ gen_arith(ctx, OPC_RISC_MUL, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_mulh(DisasContext *ctx, arg_mulh *a)
+{
+ gen_arith(ctx, OPC_RISC_MULH, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_mulhsu(DisasContext *ctx, arg_mulhsu *a)
+{
+ gen_arith(ctx, OPC_RISC_MULHSU, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_mulhu(DisasContext *ctx, arg_mulhu *a)
+{
+ gen_arith(ctx, OPC_RISC_MULHU, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_div(DisasContext *ctx, arg_div *a)
+{
+ gen_arith(ctx, OPC_RISC_DIV, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_divu(DisasContext *ctx, arg_divu *a)
+{
+ gen_arith(ctx, OPC_RISC_DIVU, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_rem(DisasContext *ctx, arg_rem *a)
+{
+ gen_arith(ctx, OPC_RISC_REM, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_remu(DisasContext *ctx, arg_remu *a)
+{
+ gen_arith(ctx, OPC_RISC_REMU, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+#ifdef TARGET_RISCV64
+static bool trans_mulw(DisasContext *ctx, arg_mulw *a)
+{
+ gen_arith(ctx, OPC_RISC_MULW, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_divw(DisasContext *ctx, arg_divw *a)
+{
+ gen_arith(ctx, OPC_RISC_DIVW, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_divuw(DisasContext *ctx, arg_divuw *a)
+{
+ gen_arith(ctx, OPC_RISC_DIVUW, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_remw(DisasContext *ctx, arg_remw *a)
+{
+ gen_arith(ctx, OPC_RISC_REMW, a->rd, a->rs1, a->rs2);
+ return true;
+}
+
+static bool trans_remuw(DisasContext *ctx, arg_remuw *a)
+{
+ gen_arith(ctx, OPC_RISC_REMUW, a->rd, a->rs1, a->rs2);
+ return true;
+}
+#endif
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index be03acd066..8c876187b1 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -1639,6 +1639,7 @@ bool decode_insn32(DisasContext *ctx, uint32_t insn);
#include "decode_insn32.inc.c"
/* Include insn module translation function */
#include "insn_trans/trans_rvi.inc.c"
+#include "insn_trans/trans_rvm.inc.c"
static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
{
@@ -1660,15 +1661,6 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
imm = GET_IMM(ctx->opcode);
switch (op) {
- case OPC_RISC_ARITH:
-#if defined(TARGET_RISCV64)
- case OPC_RISC_ARITH_W:
-#endif
- if (rd == 0) {
- break; /* NOP */
- }
- gen_arith(ctx, MASK_OP_ARITH(ctx->opcode), rd, rs1, rs2);
- break;
case OPC_RISC_FP_LOAD:
gen_fp_load(ctx, MASK_OP_FP_LOAD(ctx->opcode), rd, rs1, imm);
break;
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 10/35] target/riscv: Convert RV32A insns to decodetree
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- dropped insn argument of trans_foo functions
target/riscv/insn32.decode | 17 +++
target/riscv/insn_trans/trans_rva.inc.c | 145 ++++++++++++++++++++++++
target/riscv/translate.c | 1 +
3 files changed, 163 insertions(+)
create mode 100644 target/riscv/insn_trans/trans_rva.inc.c
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index b37fece633..235f4a031c 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -34,6 +34,7 @@
# Argument sets:
&b imm rs2 rs1
&shift shamt rs1 rd
+&atomic aq rl rs2 rs1 rd
# Formats 32:
@r ....... ..... ..... ... ..... ....... %rs2 %rs1 %rd
@@ -46,6 +47,9 @@
@sh6 ...... ...... ..... ... ..... ....... &shift shamt=%sh6 %rs1 %rd
@csr ............ ..... ... ..... ....... %csr %rs1 %rd
+@atom_ld ..... aq:1 rl:1 ..... ........ ..... ....... &atomic rs2=0 %rs1 %rd
+@atom_st ..... aq:1 rl:1 ..... ........ ..... ....... &atomic %rs2 %rs1 %rd
+
# *** RV32I Base Instruction Set ***
lui .................... ..... 0110111 @u
auipc .................... ..... 0010111 @u
@@ -102,3 +106,16 @@ div 0000001 ..... ..... 100 ..... 0110011 @r
divu 0000001 ..... ..... 101 ..... 0110011 @r
rem 0000001 ..... ..... 110 ..... 0110011 @r
remu 0000001 ..... ..... 111 ..... 0110011 @r
+
+# *** RV32A Standard Extension ***
+lr_w 00010 . . 00000 ..... 010 ..... 0101111 @atom_ld
+sc_w 00011 . . ..... ..... 010 ..... 0101111 @atom_st
+amoswap_w 00001 . . ..... ..... 010 ..... 0101111 @atom_st
+amoadd_w 00000 . . ..... ..... 010 ..... 0101111 @atom_st
+amoxor_w 00100 . . ..... ..... 010 ..... 0101111 @atom_st
+amoand_w 01100 . . ..... ..... 010 ..... 0101111 @atom_st
+amoor_w 01000 . . ..... ..... 010 ..... 0101111 @atom_st
+amomin_w 10000 . . ..... ..... 010 ..... 0101111 @atom_st
+amomax_w 10100 . . ..... ..... 010 ..... 0101111 @atom_st
+amominu_w 11000 . . ..... ..... 010 ..... 0101111 @atom_st
+amomaxu_w 11100 . . ..... ..... 010 ..... 0101111 @atom_st
diff --git a/target/riscv/insn_trans/trans_rva.inc.c b/target/riscv/insn_trans/trans_rva.inc.c
new file mode 100644
index 0000000000..d42d36ebf7
--- /dev/null
+++ b/target/riscv/insn_trans/trans_rva.inc.c
@@ -0,0 +1,145 @@
+/*
+ * RISC-V translation routines for the RV64A Standard Extension.
+ *
+ * Copyright (c) 2016-2017 Sagar Karandikar, sagark@eecs.berkeley.edu
+ * Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
+ * Bastian Koppelmann, kbastian@mail.uni-paderborn.de
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2 or later, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+static inline bool gen_lr(DisasContext *ctx, arg_atomic *a, TCGMemOp mop)
+{
+ TCGv src1 = tcg_temp_new();
+ /* Put addr in load_res, data in load_val. */
+ gen_get_gpr(src1, a->rs1);
+ if (a->rl) {
+ tcg_gen_mb(TCG_MO_ALL | TCG_BAR_STRL);
+ }
+ tcg_gen_qemu_ld_tl(load_val, src1, ctx->mem_idx, mop);
+ if (a->aq) {
+ tcg_gen_mb(TCG_MO_ALL | TCG_BAR_LDAQ);
+ }
+ tcg_gen_mov_tl(load_res, src1);
+ gen_set_gpr(a->rd, load_val);
+
+ tcg_temp_free(src1);
+ return true;
+}
+
+static inline bool gen_sc(DisasContext *ctx, arg_atomic *a, TCGMemOp mop)
+{
+ TCGv src1 = tcg_temp_new();
+ TCGv src2 = tcg_temp_new();
+ TCGv dat = tcg_temp_new();
+ TCGLabel *l1 = gen_new_label();
+ TCGLabel *l2 = gen_new_label();
+
+ gen_get_gpr(src1, a->rs1);
+ tcg_gen_brcond_tl(TCG_COND_NE, load_res, src1, l1);
+
+ gen_get_gpr(src2, a->rs2);
+ /* Note that the TCG atomic primitives are SC,
+ so we can ignore AQ/RL along this path. */
+ tcg_gen_atomic_cmpxchg_tl(src1, load_res, load_val, src2,
+ ctx->mem_idx, mop);
+ tcg_gen_setcond_tl(TCG_COND_NE, dat, src1, load_val);
+ gen_set_gpr(a->rd, dat);
+ tcg_gen_br(l2);
+
+ gen_set_label(l1);
+ /* Address comparion failure. However, we still need to
+ provide the memory barrier implied by AQ/RL. */
+ tcg_gen_mb(TCG_MO_ALL + a->aq * TCG_BAR_LDAQ + a->rl * TCG_BAR_STRL);
+ tcg_gen_movi_tl(dat, 1);
+ gen_set_gpr(a->rd, dat);
+
+ gen_set_label(l2);
+ tcg_temp_free(dat);
+ tcg_temp_free(src1);
+ tcg_temp_free(src2);
+ return true;
+}
+
+static bool gen_amo(DisasContext *ctx, arg_atomic *a,
+ void(*func)(TCGv, TCGv, TCGv, TCGArg, TCGMemOp),
+ TCGMemOp mop)
+{
+ TCGv src1 = tcg_temp_new();
+ TCGv src2 = tcg_temp_new();
+
+ gen_get_gpr(src1, a->rs1);
+ gen_get_gpr(src2, a->rs2);
+
+ (*func)(src2, src1, src2, ctx->mem_idx, mop);
+
+ gen_set_gpr(a->rd, src2);
+ tcg_temp_free(src1);
+ tcg_temp_free(src2);
+ return true;
+}
+
+static bool trans_lr_w(DisasContext *ctx, arg_lr_w *a)
+{
+ return gen_lr(ctx, a, (MO_ALIGN | MO_TESL));
+}
+
+static bool trans_sc_w(DisasContext *ctx, arg_sc_w *a)
+{
+ return gen_sc(ctx, a, (MO_ALIGN | MO_TESL));
+}
+
+static bool trans_amoswap_w(DisasContext *ctx, arg_amoswap_w *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_xchg_tl, (MO_ALIGN | MO_TESL));
+}
+
+static bool trans_amoadd_w(DisasContext *ctx, arg_amoadd_w *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_add_tl, (MO_ALIGN | MO_TESL));
+}
+
+static bool trans_amoxor_w(DisasContext *ctx, arg_amoxor_w *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_xor_tl, (MO_ALIGN | MO_TESL));
+}
+
+static bool trans_amoand_w(DisasContext *ctx, arg_amoand_w *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_and_tl, (MO_ALIGN | MO_TESL));
+}
+
+static bool trans_amoor_w(DisasContext *ctx, arg_amoor_w *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_or_tl, (MO_ALIGN | MO_TESL));
+}
+
+static bool trans_amomin_w(DisasContext *ctx, arg_amomin_w *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_smin_tl, (MO_ALIGN | MO_TESL));
+}
+
+static bool trans_amomax_w(DisasContext *ctx, arg_amomax_w *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_smax_tl, (MO_ALIGN | MO_TESL));
+}
+
+static bool trans_amominu_w(DisasContext *ctx, arg_amominu_w *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_umin_tl, (MO_ALIGN | MO_TESL));
+}
+
+static bool trans_amomaxu_w(DisasContext *ctx, arg_amomaxu_w *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_umax_tl, (MO_ALIGN | MO_TESL));
+}
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 8c876187b1..d8fe319f62 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -1640,6 +1640,7 @@ bool decode_insn32(DisasContext *ctx, uint32_t insn);
/* Include insn module translation function */
#include "insn_trans/trans_rvi.inc.c"
#include "insn_trans/trans_rvm.inc.c"
+#include "insn_trans/trans_rva.inc.c"
static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
{
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 10/35] target/riscv: Convert RV32A insns to decodetree
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- dropped insn argument of trans_foo functions
target/riscv/insn32.decode | 17 +++
target/riscv/insn_trans/trans_rva.inc.c | 145 ++++++++++++++++++++++++
target/riscv/translate.c | 1 +
3 files changed, 163 insertions(+)
create mode 100644 target/riscv/insn_trans/trans_rva.inc.c
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index b37fece633..235f4a031c 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -34,6 +34,7 @@
# Argument sets:
&b imm rs2 rs1
&shift shamt rs1 rd
+&atomic aq rl rs2 rs1 rd
# Formats 32:
@r ....... ..... ..... ... ..... ....... %rs2 %rs1 %rd
@@ -46,6 +47,9 @@
@sh6 ...... ...... ..... ... ..... ....... &shift shamt=%sh6 %rs1 %rd
@csr ............ ..... ... ..... ....... %csr %rs1 %rd
+@atom_ld ..... aq:1 rl:1 ..... ........ ..... ....... &atomic rs2=0 %rs1 %rd
+@atom_st ..... aq:1 rl:1 ..... ........ ..... ....... &atomic %rs2 %rs1 %rd
+
# *** RV32I Base Instruction Set ***
lui .................... ..... 0110111 @u
auipc .................... ..... 0010111 @u
@@ -102,3 +106,16 @@ div 0000001 ..... ..... 100 ..... 0110011 @r
divu 0000001 ..... ..... 101 ..... 0110011 @r
rem 0000001 ..... ..... 110 ..... 0110011 @r
remu 0000001 ..... ..... 111 ..... 0110011 @r
+
+# *** RV32A Standard Extension ***
+lr_w 00010 . . 00000 ..... 010 ..... 0101111 @atom_ld
+sc_w 00011 . . ..... ..... 010 ..... 0101111 @atom_st
+amoswap_w 00001 . . ..... ..... 010 ..... 0101111 @atom_st
+amoadd_w 00000 . . ..... ..... 010 ..... 0101111 @atom_st
+amoxor_w 00100 . . ..... ..... 010 ..... 0101111 @atom_st
+amoand_w 01100 . . ..... ..... 010 ..... 0101111 @atom_st
+amoor_w 01000 . . ..... ..... 010 ..... 0101111 @atom_st
+amomin_w 10000 . . ..... ..... 010 ..... 0101111 @atom_st
+amomax_w 10100 . . ..... ..... 010 ..... 0101111 @atom_st
+amominu_w 11000 . . ..... ..... 010 ..... 0101111 @atom_st
+amomaxu_w 11100 . . ..... ..... 010 ..... 0101111 @atom_st
diff --git a/target/riscv/insn_trans/trans_rva.inc.c b/target/riscv/insn_trans/trans_rva.inc.c
new file mode 100644
index 0000000000..d42d36ebf7
--- /dev/null
+++ b/target/riscv/insn_trans/trans_rva.inc.c
@@ -0,0 +1,145 @@
+/*
+ * RISC-V translation routines for the RV64A Standard Extension.
+ *
+ * Copyright (c) 2016-2017 Sagar Karandikar, sagark@eecs.berkeley.edu
+ * Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
+ * Bastian Koppelmann, kbastian@mail.uni-paderborn.de
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2 or later, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+static inline bool gen_lr(DisasContext *ctx, arg_atomic *a, TCGMemOp mop)
+{
+ TCGv src1 = tcg_temp_new();
+ /* Put addr in load_res, data in load_val. */
+ gen_get_gpr(src1, a->rs1);
+ if (a->rl) {
+ tcg_gen_mb(TCG_MO_ALL | TCG_BAR_STRL);
+ }
+ tcg_gen_qemu_ld_tl(load_val, src1, ctx->mem_idx, mop);
+ if (a->aq) {
+ tcg_gen_mb(TCG_MO_ALL | TCG_BAR_LDAQ);
+ }
+ tcg_gen_mov_tl(load_res, src1);
+ gen_set_gpr(a->rd, load_val);
+
+ tcg_temp_free(src1);
+ return true;
+}
+
+static inline bool gen_sc(DisasContext *ctx, arg_atomic *a, TCGMemOp mop)
+{
+ TCGv src1 = tcg_temp_new();
+ TCGv src2 = tcg_temp_new();
+ TCGv dat = tcg_temp_new();
+ TCGLabel *l1 = gen_new_label();
+ TCGLabel *l2 = gen_new_label();
+
+ gen_get_gpr(src1, a->rs1);
+ tcg_gen_brcond_tl(TCG_COND_NE, load_res, src1, l1);
+
+ gen_get_gpr(src2, a->rs2);
+ /* Note that the TCG atomic primitives are SC,
+ so we can ignore AQ/RL along this path. */
+ tcg_gen_atomic_cmpxchg_tl(src1, load_res, load_val, src2,
+ ctx->mem_idx, mop);
+ tcg_gen_setcond_tl(TCG_COND_NE, dat, src1, load_val);
+ gen_set_gpr(a->rd, dat);
+ tcg_gen_br(l2);
+
+ gen_set_label(l1);
+ /* Address comparion failure. However, we still need to
+ provide the memory barrier implied by AQ/RL. */
+ tcg_gen_mb(TCG_MO_ALL + a->aq * TCG_BAR_LDAQ + a->rl * TCG_BAR_STRL);
+ tcg_gen_movi_tl(dat, 1);
+ gen_set_gpr(a->rd, dat);
+
+ gen_set_label(l2);
+ tcg_temp_free(dat);
+ tcg_temp_free(src1);
+ tcg_temp_free(src2);
+ return true;
+}
+
+static bool gen_amo(DisasContext *ctx, arg_atomic *a,
+ void(*func)(TCGv, TCGv, TCGv, TCGArg, TCGMemOp),
+ TCGMemOp mop)
+{
+ TCGv src1 = tcg_temp_new();
+ TCGv src2 = tcg_temp_new();
+
+ gen_get_gpr(src1, a->rs1);
+ gen_get_gpr(src2, a->rs2);
+
+ (*func)(src2, src1, src2, ctx->mem_idx, mop);
+
+ gen_set_gpr(a->rd, src2);
+ tcg_temp_free(src1);
+ tcg_temp_free(src2);
+ return true;
+}
+
+static bool trans_lr_w(DisasContext *ctx, arg_lr_w *a)
+{
+ return gen_lr(ctx, a, (MO_ALIGN | MO_TESL));
+}
+
+static bool trans_sc_w(DisasContext *ctx, arg_sc_w *a)
+{
+ return gen_sc(ctx, a, (MO_ALIGN | MO_TESL));
+}
+
+static bool trans_amoswap_w(DisasContext *ctx, arg_amoswap_w *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_xchg_tl, (MO_ALIGN | MO_TESL));
+}
+
+static bool trans_amoadd_w(DisasContext *ctx, arg_amoadd_w *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_add_tl, (MO_ALIGN | MO_TESL));
+}
+
+static bool trans_amoxor_w(DisasContext *ctx, arg_amoxor_w *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_xor_tl, (MO_ALIGN | MO_TESL));
+}
+
+static bool trans_amoand_w(DisasContext *ctx, arg_amoand_w *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_and_tl, (MO_ALIGN | MO_TESL));
+}
+
+static bool trans_amoor_w(DisasContext *ctx, arg_amoor_w *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_or_tl, (MO_ALIGN | MO_TESL));
+}
+
+static bool trans_amomin_w(DisasContext *ctx, arg_amomin_w *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_smin_tl, (MO_ALIGN | MO_TESL));
+}
+
+static bool trans_amomax_w(DisasContext *ctx, arg_amomax_w *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_smax_tl, (MO_ALIGN | MO_TESL));
+}
+
+static bool trans_amominu_w(DisasContext *ctx, arg_amominu_w *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_umin_tl, (MO_ALIGN | MO_TESL));
+}
+
+static bool trans_amomaxu_w(DisasContext *ctx, arg_amomaxu_w *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_umax_tl, (MO_ALIGN | MO_TESL));
+}
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 8c876187b1..d8fe319f62 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -1640,6 +1640,7 @@ bool decode_insn32(DisasContext *ctx, uint32_t insn);
/* Include insn module translation function */
#include "insn_trans/trans_rvi.inc.c"
#include "insn_trans/trans_rvm.inc.c"
+#include "insn_trans/trans_rva.inc.c"
static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
{
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 11/35] target/riscv: Convert RV64A insns to decodetree
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- moved 64-bit only insn to insn64.decode
- dropped insn argument of trans_foo functions
target/riscv/insn64.decode | 13 +++
target/riscv/insn_trans/trans_rva.inc.c | 58 ++++++++++
target/riscv/translate.c | 140 ------------------------
3 files changed, 71 insertions(+), 140 deletions(-)
diff --git a/target/riscv/insn64.decode b/target/riscv/insn64.decode
index 008f100546..0bee95c984 100644
--- a/target/riscv/insn64.decode
+++ b/target/riscv/insn64.decode
@@ -43,3 +43,16 @@ divw 0000001 ..... ..... 100 ..... 0111011 @r
divuw 0000001 ..... ..... 101 ..... 0111011 @r
remw 0000001 ..... ..... 110 ..... 0111011 @r
remuw 0000001 ..... ..... 111 ..... 0111011 @r
+
+# *** RV64A Standard Extension (in addition to RV32A) ***
+lr_d 00010 . . 00000 ..... 011 ..... 0101111 @atom_ld
+sc_d 00011 . . ..... ..... 011 ..... 0101111 @atom_st
+amoswap_d 00001 . . ..... ..... 011 ..... 0101111 @atom_st
+amoadd_d 00000 . . ..... ..... 011 ..... 0101111 @atom_st
+amoxor_d 00100 . . ..... ..... 011 ..... 0101111 @atom_st
+amoand_d 01100 . . ..... ..... 011 ..... 0101111 @atom_st
+amoor_d 01000 . . ..... ..... 011 ..... 0101111 @atom_st
+amomin_d 10000 . . ..... ..... 011 ..... 0101111 @atom_st
+amomax_d 10100 . . ..... ..... 011 ..... 0101111 @atom_st
+amominu_d 11000 . . ..... ..... 011 ..... 0101111 @atom_st
+amomaxu_d 11100 . . ..... ..... 011 ..... 0101111 @atom_st
diff --git a/target/riscv/insn_trans/trans_rva.inc.c b/target/riscv/insn_trans/trans_rva.inc.c
index d42d36ebf7..fbfca7f835 100644
--- a/target/riscv/insn_trans/trans_rva.inc.c
+++ b/target/riscv/insn_trans/trans_rva.inc.c
@@ -143,3 +143,61 @@ static bool trans_amomaxu_w(DisasContext *ctx, arg_amomaxu_w *a)
{
return gen_amo(ctx, a, &tcg_gen_atomic_fetch_umax_tl, (MO_ALIGN | MO_TESL));
}
+
+#ifdef TARGET_RISCV64
+
+static bool trans_lr_d(DisasContext *ctx, arg_lr_d *a)
+{
+ return gen_lr(ctx, a, MO_ALIGN | MO_TEQ);
+}
+
+static bool trans_sc_d(DisasContext *ctx, arg_sc_d *a)
+{
+ return gen_sc(ctx, a, (MO_ALIGN | MO_TEQ));
+}
+
+static bool trans_amoswap_d(DisasContext *ctx, arg_amoswap_d *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_xchg_tl, (MO_ALIGN | MO_TEQ));
+}
+
+static bool trans_amoadd_d(DisasContext *ctx, arg_amoadd_d *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_add_tl, (MO_ALIGN | MO_TEQ));
+}
+
+static bool trans_amoxor_d(DisasContext *ctx, arg_amoxor_d *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_xor_tl, (MO_ALIGN | MO_TEQ));
+}
+
+static bool trans_amoand_d(DisasContext *ctx, arg_amoand_d *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_and_tl, (MO_ALIGN | MO_TEQ));
+}
+
+static bool trans_amoor_d(DisasContext *ctx, arg_amoor_d *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_or_tl, (MO_ALIGN | MO_TEQ));
+}
+
+static bool trans_amomin_d(DisasContext *ctx, arg_amomin_d *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_smin_tl, (MO_ALIGN | MO_TEQ));
+}
+
+static bool trans_amomax_d(DisasContext *ctx, arg_amomax_d *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_smax_tl, (MO_ALIGN | MO_TEQ));
+}
+
+static bool trans_amominu_d(DisasContext *ctx, arg_amominu_d *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_umin_tl, (MO_ALIGN | MO_TEQ));
+}
+
+static bool trans_amomaxu_d(DisasContext *ctx, arg_amomaxu_d *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_umax_tl, (MO_ALIGN | MO_TEQ));
+}
+#endif
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index d8fe319f62..a1b949ad0a 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -711,143 +711,6 @@ static void gen_fp_store(DisasContext *ctx, uint32_t opc, int rs1,
tcg_temp_free(t0);
}
-static void gen_atomic(DisasContext *ctx, uint32_t opc,
- int rd, int rs1, int rs2)
-{
- TCGv src1, src2, dat;
- TCGLabel *l1, *l2;
- TCGMemOp mop;
- bool aq, rl;
-
- /* Extract the size of the atomic operation. */
- switch (extract32(opc, 12, 3)) {
- case 2: /* 32-bit */
- mop = MO_ALIGN | MO_TESL;
- break;
-#if defined(TARGET_RISCV64)
- case 3: /* 64-bit */
- mop = MO_ALIGN | MO_TEQ;
- break;
-#endif
- default:
- gen_exception_illegal(ctx);
- return;
- }
- rl = extract32(opc, 25, 1);
- aq = extract32(opc, 26, 1);
-
- src1 = tcg_temp_new();
- src2 = tcg_temp_new();
-
- switch (MASK_OP_ATOMIC_NO_AQ_RL_SZ(opc)) {
- case OPC_RISC_LR:
- /* Put addr in load_res, data in load_val. */
- gen_get_gpr(src1, rs1);
- if (rl) {
- tcg_gen_mb(TCG_MO_ALL | TCG_BAR_STRL);
- }
- tcg_gen_qemu_ld_tl(load_val, src1, ctx->mem_idx, mop);
- if (aq) {
- tcg_gen_mb(TCG_MO_ALL | TCG_BAR_LDAQ);
- }
- tcg_gen_mov_tl(load_res, src1);
- gen_set_gpr(rd, load_val);
- break;
-
- case OPC_RISC_SC:
- l1 = gen_new_label();
- l2 = gen_new_label();
- dat = tcg_temp_new();
-
- gen_get_gpr(src1, rs1);
- tcg_gen_brcond_tl(TCG_COND_NE, load_res, src1, l1);
-
- gen_get_gpr(src2, rs2);
- /* Note that the TCG atomic primitives are SC,
- so we can ignore AQ/RL along this path. */
- tcg_gen_atomic_cmpxchg_tl(src1, load_res, load_val, src2,
- ctx->mem_idx, mop);
- tcg_gen_setcond_tl(TCG_COND_NE, dat, src1, load_val);
- gen_set_gpr(rd, dat);
- tcg_gen_br(l2);
-
- gen_set_label(l1);
- /* Address comparion failure. However, we still need to
- provide the memory barrier implied by AQ/RL. */
- tcg_gen_mb(TCG_MO_ALL + aq * TCG_BAR_LDAQ + rl * TCG_BAR_STRL);
- tcg_gen_movi_tl(dat, 1);
- gen_set_gpr(rd, dat);
-
- gen_set_label(l2);
- tcg_temp_free(dat);
- break;
-
- case OPC_RISC_AMOSWAP:
- /* Note that the TCG atomic primitives are SC,
- so we can ignore AQ/RL along this path. */
- gen_get_gpr(src1, rs1);
- gen_get_gpr(src2, rs2);
- tcg_gen_atomic_xchg_tl(src2, src1, src2, ctx->mem_idx, mop);
- gen_set_gpr(rd, src2);
- break;
- case OPC_RISC_AMOADD:
- gen_get_gpr(src1, rs1);
- gen_get_gpr(src2, rs2);
- tcg_gen_atomic_fetch_add_tl(src2, src1, src2, ctx->mem_idx, mop);
- gen_set_gpr(rd, src2);
- break;
- case OPC_RISC_AMOXOR:
- gen_get_gpr(src1, rs1);
- gen_get_gpr(src2, rs2);
- tcg_gen_atomic_fetch_xor_tl(src2, src1, src2, ctx->mem_idx, mop);
- gen_set_gpr(rd, src2);
- break;
- case OPC_RISC_AMOAND:
- gen_get_gpr(src1, rs1);
- gen_get_gpr(src2, rs2);
- tcg_gen_atomic_fetch_and_tl(src2, src1, src2, ctx->mem_idx, mop);
- gen_set_gpr(rd, src2);
- break;
- case OPC_RISC_AMOOR:
- gen_get_gpr(src1, rs1);
- gen_get_gpr(src2, rs2);
- tcg_gen_atomic_fetch_or_tl(src2, src1, src2, ctx->mem_idx, mop);
- gen_set_gpr(rd, src2);
- break;
- case OPC_RISC_AMOMIN:
- gen_get_gpr(src1, rs1);
- gen_get_gpr(src2, rs2);
- tcg_gen_atomic_fetch_smin_tl(src2, src1, src2, ctx->mem_idx, mop);
- gen_set_gpr(rd, src2);
- break;
- case OPC_RISC_AMOMAX:
- gen_get_gpr(src1, rs1);
- gen_get_gpr(src2, rs2);
- tcg_gen_atomic_fetch_smax_tl(src2, src1, src2, ctx->mem_idx, mop);
- gen_set_gpr(rd, src2);
- break;
- case OPC_RISC_AMOMINU:
- gen_get_gpr(src1, rs1);
- gen_get_gpr(src2, rs2);
- tcg_gen_atomic_fetch_umin_tl(src2, src1, src2, ctx->mem_idx, mop);
- gen_set_gpr(rd, src2);
- break;
- case OPC_RISC_AMOMAXU:
- gen_get_gpr(src1, rs1);
- gen_get_gpr(src2, rs2);
- tcg_gen_atomic_fetch_umax_tl(src2, src1, src2, ctx->mem_idx, mop);
- gen_set_gpr(rd, src2);
- break;
-
- default:
- gen_exception_illegal(ctx);
- break;
- }
-
- tcg_temp_free(src1);
- tcg_temp_free(src2);
-}
-
static void gen_set_rm(DisasContext *ctx, int rm)
{
TCGv_i32 t0;
@@ -1669,9 +1532,6 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
gen_fp_store(ctx, MASK_OP_FP_STORE(ctx->opcode), rs1, rs2,
GET_STORE_IMM(ctx->opcode));
break;
- case OPC_RISC_ATOMIC:
- gen_atomic(ctx, MASK_OP_ATOMIC(ctx->opcode), rd, rs1, rs2);
- break;
case OPC_RISC_FMADD:
gen_fp_fmadd(ctx, MASK_OP_FP_FMADD(ctx->opcode), rd, rs1, rs2,
GET_RS3(ctx->opcode), GET_RM(ctx->opcode));
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 11/35] target/riscv: Convert RV64A insns to decodetree
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- moved 64-bit only insn to insn64.decode
- dropped insn argument of trans_foo functions
target/riscv/insn64.decode | 13 +++
target/riscv/insn_trans/trans_rva.inc.c | 58 ++++++++++
target/riscv/translate.c | 140 ------------------------
3 files changed, 71 insertions(+), 140 deletions(-)
diff --git a/target/riscv/insn64.decode b/target/riscv/insn64.decode
index 008f100546..0bee95c984 100644
--- a/target/riscv/insn64.decode
+++ b/target/riscv/insn64.decode
@@ -43,3 +43,16 @@ divw 0000001 ..... ..... 100 ..... 0111011 @r
divuw 0000001 ..... ..... 101 ..... 0111011 @r
remw 0000001 ..... ..... 110 ..... 0111011 @r
remuw 0000001 ..... ..... 111 ..... 0111011 @r
+
+# *** RV64A Standard Extension (in addition to RV32A) ***
+lr_d 00010 . . 00000 ..... 011 ..... 0101111 @atom_ld
+sc_d 00011 . . ..... ..... 011 ..... 0101111 @atom_st
+amoswap_d 00001 . . ..... ..... 011 ..... 0101111 @atom_st
+amoadd_d 00000 . . ..... ..... 011 ..... 0101111 @atom_st
+amoxor_d 00100 . . ..... ..... 011 ..... 0101111 @atom_st
+amoand_d 01100 . . ..... ..... 011 ..... 0101111 @atom_st
+amoor_d 01000 . . ..... ..... 011 ..... 0101111 @atom_st
+amomin_d 10000 . . ..... ..... 011 ..... 0101111 @atom_st
+amomax_d 10100 . . ..... ..... 011 ..... 0101111 @atom_st
+amominu_d 11000 . . ..... ..... 011 ..... 0101111 @atom_st
+amomaxu_d 11100 . . ..... ..... 011 ..... 0101111 @atom_st
diff --git a/target/riscv/insn_trans/trans_rva.inc.c b/target/riscv/insn_trans/trans_rva.inc.c
index d42d36ebf7..fbfca7f835 100644
--- a/target/riscv/insn_trans/trans_rva.inc.c
+++ b/target/riscv/insn_trans/trans_rva.inc.c
@@ -143,3 +143,61 @@ static bool trans_amomaxu_w(DisasContext *ctx, arg_amomaxu_w *a)
{
return gen_amo(ctx, a, &tcg_gen_atomic_fetch_umax_tl, (MO_ALIGN | MO_TESL));
}
+
+#ifdef TARGET_RISCV64
+
+static bool trans_lr_d(DisasContext *ctx, arg_lr_d *a)
+{
+ return gen_lr(ctx, a, MO_ALIGN | MO_TEQ);
+}
+
+static bool trans_sc_d(DisasContext *ctx, arg_sc_d *a)
+{
+ return gen_sc(ctx, a, (MO_ALIGN | MO_TEQ));
+}
+
+static bool trans_amoswap_d(DisasContext *ctx, arg_amoswap_d *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_xchg_tl, (MO_ALIGN | MO_TEQ));
+}
+
+static bool trans_amoadd_d(DisasContext *ctx, arg_amoadd_d *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_add_tl, (MO_ALIGN | MO_TEQ));
+}
+
+static bool trans_amoxor_d(DisasContext *ctx, arg_amoxor_d *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_xor_tl, (MO_ALIGN | MO_TEQ));
+}
+
+static bool trans_amoand_d(DisasContext *ctx, arg_amoand_d *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_and_tl, (MO_ALIGN | MO_TEQ));
+}
+
+static bool trans_amoor_d(DisasContext *ctx, arg_amoor_d *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_or_tl, (MO_ALIGN | MO_TEQ));
+}
+
+static bool trans_amomin_d(DisasContext *ctx, arg_amomin_d *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_smin_tl, (MO_ALIGN | MO_TEQ));
+}
+
+static bool trans_amomax_d(DisasContext *ctx, arg_amomax_d *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_smax_tl, (MO_ALIGN | MO_TEQ));
+}
+
+static bool trans_amominu_d(DisasContext *ctx, arg_amominu_d *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_umin_tl, (MO_ALIGN | MO_TEQ));
+}
+
+static bool trans_amomaxu_d(DisasContext *ctx, arg_amomaxu_d *a)
+{
+ return gen_amo(ctx, a, &tcg_gen_atomic_fetch_umax_tl, (MO_ALIGN | MO_TEQ));
+}
+#endif
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index d8fe319f62..a1b949ad0a 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -711,143 +711,6 @@ static void gen_fp_store(DisasContext *ctx, uint32_t opc, int rs1,
tcg_temp_free(t0);
}
-static void gen_atomic(DisasContext *ctx, uint32_t opc,
- int rd, int rs1, int rs2)
-{
- TCGv src1, src2, dat;
- TCGLabel *l1, *l2;
- TCGMemOp mop;
- bool aq, rl;
-
- /* Extract the size of the atomic operation. */
- switch (extract32(opc, 12, 3)) {
- case 2: /* 32-bit */
- mop = MO_ALIGN | MO_TESL;
- break;
-#if defined(TARGET_RISCV64)
- case 3: /* 64-bit */
- mop = MO_ALIGN | MO_TEQ;
- break;
-#endif
- default:
- gen_exception_illegal(ctx);
- return;
- }
- rl = extract32(opc, 25, 1);
- aq = extract32(opc, 26, 1);
-
- src1 = tcg_temp_new();
- src2 = tcg_temp_new();
-
- switch (MASK_OP_ATOMIC_NO_AQ_RL_SZ(opc)) {
- case OPC_RISC_LR:
- /* Put addr in load_res, data in load_val. */
- gen_get_gpr(src1, rs1);
- if (rl) {
- tcg_gen_mb(TCG_MO_ALL | TCG_BAR_STRL);
- }
- tcg_gen_qemu_ld_tl(load_val, src1, ctx->mem_idx, mop);
- if (aq) {
- tcg_gen_mb(TCG_MO_ALL | TCG_BAR_LDAQ);
- }
- tcg_gen_mov_tl(load_res, src1);
- gen_set_gpr(rd, load_val);
- break;
-
- case OPC_RISC_SC:
- l1 = gen_new_label();
- l2 = gen_new_label();
- dat = tcg_temp_new();
-
- gen_get_gpr(src1, rs1);
- tcg_gen_brcond_tl(TCG_COND_NE, load_res, src1, l1);
-
- gen_get_gpr(src2, rs2);
- /* Note that the TCG atomic primitives are SC,
- so we can ignore AQ/RL along this path. */
- tcg_gen_atomic_cmpxchg_tl(src1, load_res, load_val, src2,
- ctx->mem_idx, mop);
- tcg_gen_setcond_tl(TCG_COND_NE, dat, src1, load_val);
- gen_set_gpr(rd, dat);
- tcg_gen_br(l2);
-
- gen_set_label(l1);
- /* Address comparion failure. However, we still need to
- provide the memory barrier implied by AQ/RL. */
- tcg_gen_mb(TCG_MO_ALL + aq * TCG_BAR_LDAQ + rl * TCG_BAR_STRL);
- tcg_gen_movi_tl(dat, 1);
- gen_set_gpr(rd, dat);
-
- gen_set_label(l2);
- tcg_temp_free(dat);
- break;
-
- case OPC_RISC_AMOSWAP:
- /* Note that the TCG atomic primitives are SC,
- so we can ignore AQ/RL along this path. */
- gen_get_gpr(src1, rs1);
- gen_get_gpr(src2, rs2);
- tcg_gen_atomic_xchg_tl(src2, src1, src2, ctx->mem_idx, mop);
- gen_set_gpr(rd, src2);
- break;
- case OPC_RISC_AMOADD:
- gen_get_gpr(src1, rs1);
- gen_get_gpr(src2, rs2);
- tcg_gen_atomic_fetch_add_tl(src2, src1, src2, ctx->mem_idx, mop);
- gen_set_gpr(rd, src2);
- break;
- case OPC_RISC_AMOXOR:
- gen_get_gpr(src1, rs1);
- gen_get_gpr(src2, rs2);
- tcg_gen_atomic_fetch_xor_tl(src2, src1, src2, ctx->mem_idx, mop);
- gen_set_gpr(rd, src2);
- break;
- case OPC_RISC_AMOAND:
- gen_get_gpr(src1, rs1);
- gen_get_gpr(src2, rs2);
- tcg_gen_atomic_fetch_and_tl(src2, src1, src2, ctx->mem_idx, mop);
- gen_set_gpr(rd, src2);
- break;
- case OPC_RISC_AMOOR:
- gen_get_gpr(src1, rs1);
- gen_get_gpr(src2, rs2);
- tcg_gen_atomic_fetch_or_tl(src2, src1, src2, ctx->mem_idx, mop);
- gen_set_gpr(rd, src2);
- break;
- case OPC_RISC_AMOMIN:
- gen_get_gpr(src1, rs1);
- gen_get_gpr(src2, rs2);
- tcg_gen_atomic_fetch_smin_tl(src2, src1, src2, ctx->mem_idx, mop);
- gen_set_gpr(rd, src2);
- break;
- case OPC_RISC_AMOMAX:
- gen_get_gpr(src1, rs1);
- gen_get_gpr(src2, rs2);
- tcg_gen_atomic_fetch_smax_tl(src2, src1, src2, ctx->mem_idx, mop);
- gen_set_gpr(rd, src2);
- break;
- case OPC_RISC_AMOMINU:
- gen_get_gpr(src1, rs1);
- gen_get_gpr(src2, rs2);
- tcg_gen_atomic_fetch_umin_tl(src2, src1, src2, ctx->mem_idx, mop);
- gen_set_gpr(rd, src2);
- break;
- case OPC_RISC_AMOMAXU:
- gen_get_gpr(src1, rs1);
- gen_get_gpr(src2, rs2);
- tcg_gen_atomic_fetch_umax_tl(src2, src1, src2, ctx->mem_idx, mop);
- gen_set_gpr(rd, src2);
- break;
-
- default:
- gen_exception_illegal(ctx);
- break;
- }
-
- tcg_temp_free(src1);
- tcg_temp_free(src2);
-}
-
static void gen_set_rm(DisasContext *ctx, int rm)
{
TCGv_i32 t0;
@@ -1669,9 +1532,6 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
gen_fp_store(ctx, MASK_OP_FP_STORE(ctx->opcode), rs1, rs2,
GET_STORE_IMM(ctx->opcode));
break;
- case OPC_RISC_ATOMIC:
- gen_atomic(ctx, MASK_OP_ATOMIC(ctx->opcode), rd, rs1, rs2);
- break;
case OPC_RISC_FMADD:
gen_fp_fmadd(ctx, MASK_OP_FP_FMADD(ctx->opcode), rd, rs1, rs2,
GET_RS3(ctx->opcode), GET_RM(ctx->opcode));
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 12/35] target/riscv: Convert RV32F insns to decodetree
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- dropped insn argument of trans_foo functions
target/riscv/insn32.decode | 35 +++
target/riscv/insn_trans/trans_rvf.inc.c | 334 ++++++++++++++++++++++++
target/riscv/translate.c | 1 +
3 files changed, 370 insertions(+)
create mode 100644 target/riscv/insn_trans/trans_rvf.inc.c
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index 235f4a031c..57efe0dcff 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -17,12 +17,14 @@
# this program. If not, see <http://www.gnu.org/licenses/>.
# Fields:
+%rs3 27:5
%rs2 20:5
%rs1 15:5
%rd 7:5
%sh6 20:6
%csr 20:12
+%rm 12:3
# immediates:
%imm_i 20:s12
@@ -50,6 +52,11 @@
@atom_ld ..... aq:1 rl:1 ..... ........ ..... ....... &atomic rs2=0 %rs1 %rd
@atom_st ..... aq:1 rl:1 ..... ........ ..... ....... &atomic %rs2 %rs1 %rd
+@r4_rm ..... .. ..... ..... ... ..... ....... %rs3 %rs2 %rs1 %rm %rd
+@r_rm ....... ..... ..... ... ..... ....... %rs2 %rs1 %rm %rd
+@r2_rm ....... ..... ..... ... ..... ....... %rs1 %rm %rd
+@r2 ....... ..... ..... ... ..... ....... %rs1 %rd
+
# *** RV32I Base Instruction Set ***
lui .................... ..... 0110111 @u
auipc .................... ..... 0010111 @u
@@ -119,3 +126,31 @@ amomin_w 10000 . . ..... ..... 010 ..... 0101111 @atom_st
amomax_w 10100 . . ..... ..... 010 ..... 0101111 @atom_st
amominu_w 11000 . . ..... ..... 010 ..... 0101111 @atom_st
amomaxu_w 11100 . . ..... ..... 010 ..... 0101111 @atom_st
+
+# *** RV32F Standard Extension ***
+flw ............ ..... 010 ..... 0000111 @i
+fsw ....... ..... ..... 010 ..... 0100111 @s
+fmadd_s ..... 00 ..... ..... ... ..... 1000011 @r4_rm
+fmsub_s ..... 00 ..... ..... ... ..... 1000111 @r4_rm
+fnmsub_s ..... 00 ..... ..... ... ..... 1001011 @r4_rm
+fnmadd_s ..... 00 ..... ..... ... ..... 1001111 @r4_rm
+fadd_s 0000000 ..... ..... ... ..... 1010011 @r_rm
+fsub_s 0000100 ..... ..... ... ..... 1010011 @r_rm
+fmul_s 0001000 ..... ..... ... ..... 1010011 @r_rm
+fdiv_s 0001100 ..... ..... ... ..... 1010011 @r_rm
+fsqrt_s 0101100 00000 ..... ... ..... 1010011 @r2_rm
+fsgnj_s 0010000 ..... ..... 000 ..... 1010011 @r
+fsgnjn_s 0010000 ..... ..... 001 ..... 1010011 @r
+fsgnjx_s 0010000 ..... ..... 010 ..... 1010011 @r
+fmin_s 0010100 ..... ..... 000 ..... 1010011 @r
+fmax_s 0010100 ..... ..... 001 ..... 1010011 @r
+fcvt_w_s 1100000 00000 ..... ... ..... 1010011 @r2_rm
+fcvt_wu_s 1100000 00001 ..... ... ..... 1010011 @r2_rm
+fmv_x_w 1110000 00000 ..... 000 ..... 1010011 @r2
+feq_s 1010000 ..... ..... 010 ..... 1010011 @r
+flt_s 1010000 ..... ..... 001 ..... 1010011 @r
+fle_s 1010000 ..... ..... 000 ..... 1010011 @r
+fclass_s 1110000 00000 ..... 001 ..... 1010011 @r2
+fcvt_s_w 1101000 00000 ..... ... ..... 1010011 @r2_rm
+fcvt_s_wu 1101000 00001 ..... ... ..... 1010011 @r2_rm
+fmv_w_x 1111000 00000 ..... 000 ..... 1010011 @r2
diff --git a/target/riscv/insn_trans/trans_rvf.inc.c b/target/riscv/insn_trans/trans_rvf.inc.c
new file mode 100644
index 0000000000..b101593ac4
--- /dev/null
+++ b/target/riscv/insn_trans/trans_rvf.inc.c
@@ -0,0 +1,334 @@
+/*
+ * RISC-V translation routines for the RV64F Standard Extension.
+ *
+ * Copyright (c) 2016-2017 Sagar Karandikar, sagark@eecs.berkeley.edu
+ * Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
+ * Bastian Koppelmann, kbastian@mail.uni-paderborn.de
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2 or later, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#define REQUIRE_FPU do {\
+ if (!(ctx->flags & TB_FLAGS_FP_ENABLE)) \
+ return false; \
+} while (0)
+
+static bool trans_flw(DisasContext *ctx, arg_flw *a)
+{
+ TCGv t0 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+ REQUIRE_FPU;
+ tcg_gen_addi_tl(t0, t0, a->imm);
+
+ tcg_gen_qemu_ld_i64(cpu_fpr[a->rd], t0, ctx->mem_idx, MO_TEUL);
+ /* RISC-V requires NaN-boxing of narrower width floating point values */
+ tcg_gen_ori_i64(cpu_fpr[a->rd], cpu_fpr[a->rd], 0xffffffff00000000ULL);
+
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_fsw(DisasContext *ctx, arg_fsw *a)
+{
+ TCGv t0 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+
+ REQUIRE_FPU;
+ tcg_gen_addi_tl(t0, t0, a->imm);
+
+ tcg_gen_qemu_st_i64(cpu_fpr[a->rs2], t0, ctx->mem_idx, MO_TEUL);
+
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_fmadd_s(DisasContext *ctx, arg_fmadd_s *a)
+{
+ REQUIRE_FPU;
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fmadd_s(cpu_fpr[a->rd], cpu_env, cpu_fpr[a->rs1],
+ cpu_fpr[a->rs2], cpu_fpr[a->rs3]);
+ return true;
+}
+
+static bool trans_fmsub_s(DisasContext *ctx, arg_fmsub_s *a)
+{
+ REQUIRE_FPU;
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fmsub_s(cpu_fpr[a->rd], cpu_env, cpu_fpr[a->rs1],
+ cpu_fpr[a->rs2], cpu_fpr[a->rs3]);
+ return true;
+}
+
+static bool trans_fnmsub_s(DisasContext *ctx, arg_fnmsub_s *a)
+{
+ REQUIRE_FPU;
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fnmsub_s(cpu_fpr[a->rd], cpu_env, cpu_fpr[a->rs1],
+ cpu_fpr[a->rs2], cpu_fpr[a->rs3]);
+ return true;
+}
+
+static bool trans_fnmadd_s(DisasContext *ctx, arg_fnmadd_s *a)
+{
+ REQUIRE_FPU;
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fnmadd_s(cpu_fpr[a->rd], cpu_env, cpu_fpr[a->rs1],
+ cpu_fpr[a->rs2], cpu_fpr[a->rs3]);
+ return true;
+}
+
+static bool trans_fadd_s(DisasContext *ctx, arg_fadd_s *a)
+{
+ REQUIRE_FPU;
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fadd_s(cpu_fpr[a->rd], cpu_env,
+ cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+ return true;
+}
+
+static bool trans_fsub_s(DisasContext *ctx, arg_fsub_s *a)
+{
+ REQUIRE_FPU;
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fsub_s(cpu_fpr[a->rd], cpu_env,
+ cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+ return true;
+}
+
+static bool trans_fmul_s(DisasContext *ctx, arg_fmul_s *a)
+{
+ REQUIRE_FPU;
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fmul_s(cpu_fpr[a->rd], cpu_env,
+ cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+ return true;
+}
+
+static bool trans_fdiv_s(DisasContext *ctx, arg_fdiv_s *a)
+{
+ REQUIRE_FPU;
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fdiv_s(cpu_fpr[a->rd], cpu_env,
+ cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+ return true;
+}
+
+static bool trans_fsqrt_s(DisasContext *ctx, arg_fsqrt_s *a)
+{
+ REQUIRE_FPU;
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fsqrt_s(cpu_fpr[a->rd], cpu_env, cpu_fpr[a->rs1]);
+ return true;
+}
+
+static bool trans_fsgnj_s(DisasContext *ctx, arg_fsgnj_s *a)
+{
+ REQUIRE_FPU;
+ if (a->rs1 == a->rs2) { /* FMOV */
+ tcg_gen_mov_i64(cpu_fpr[a->rd], cpu_fpr[a->rs1]);
+ } else { /* FSGNJ */
+ tcg_gen_deposit_i64(cpu_fpr[a->rd], cpu_fpr[a->rs2], cpu_fpr[a->rs1],
+ 0, 31);
+ }
+ return true;
+}
+
+static bool trans_fsgnjn_s(DisasContext *ctx, arg_fsgnjn_s *a)
+{
+ REQUIRE_FPU;
+ if (a->rs1 == a->rs2) { /* FNEG */
+ tcg_gen_xori_i64(cpu_fpr[a->rd], cpu_fpr[a->rs1], INT32_MIN);
+ } else {
+ TCGv_i64 t0 = tcg_temp_new_i64();
+ tcg_gen_not_i64(t0, cpu_fpr[a->rs2]);
+ tcg_gen_deposit_i64(cpu_fpr[a->rd], t0, cpu_fpr[a->rs1], 0, 31);
+ tcg_temp_free_i64(t0);
+ }
+ return true;
+}
+
+static bool trans_fsgnjx_s(DisasContext *ctx, arg_fsgnjx_s *a)
+{
+ REQUIRE_FPU;
+ if (a->rs1 == a->rs2) { /* FABS */
+ tcg_gen_andi_i64(cpu_fpr[a->rd], cpu_fpr[a->rs1], ~INT32_MIN);
+ } else {
+ TCGv_i64 t0 = tcg_temp_new_i64();
+ tcg_gen_andi_i64(t0, cpu_fpr[a->rs2], INT32_MIN);
+ tcg_gen_xor_i64(cpu_fpr[a->rd], cpu_fpr[a->rs1], t0);
+ tcg_temp_free_i64(t0);
+ }
+ return true;
+}
+
+static bool trans_fmin_s(DisasContext *ctx, arg_fmin_s *a)
+{
+ REQUIRE_FPU;
+
+ gen_helper_fmin_s(cpu_fpr[a->rd], cpu_env, cpu_fpr[a->rs1],
+ cpu_fpr[a->rs2]);
+ return true;
+}
+
+static bool trans_fmax_s(DisasContext *ctx, arg_fmax_s *a)
+{
+ REQUIRE_FPU;
+
+ gen_helper_fmax_s(cpu_fpr[a->rd], cpu_env, cpu_fpr[a->rs1],
+ cpu_fpr[a->rs2]);
+ return true;
+}
+
+static bool trans_fcvt_w_s(DisasContext *ctx, arg_fcvt_w_s *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_w_s(t0, cpu_env, cpu_fpr[a->rs1]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+
+ return true;
+}
+
+static bool trans_fcvt_wu_s(DisasContext *ctx, arg_fcvt_wu_s *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_wu_s(t0, cpu_env, cpu_fpr[a->rs1]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+
+ return true;
+}
+
+static bool trans_fmv_x_w(DisasContext *ctx, arg_fmv_x_w *a)
+{
+ /* NOTE: This was FMV.X.S in an earlier version of the ISA spec! */
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+
+#if defined(TARGET_RISCV64)
+ tcg_gen_ext32s_tl(t0, cpu_fpr[a->rs1]);
+#else
+ tcg_gen_extrl_i64_i32(t0, cpu_fpr[a->rs1]);
+#endif
+
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+
+ return true;
+}
+
+static bool trans_feq_s(DisasContext *ctx, arg_feq_s *a)
+{
+ REQUIRE_FPU;
+ TCGv t0 = tcg_temp_new();
+ gen_helper_feq_s(t0, cpu_env, cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_flt_s(DisasContext *ctx, arg_flt_s *a)
+{
+ REQUIRE_FPU;
+ TCGv t0 = tcg_temp_new();
+ gen_helper_flt_s(t0, cpu_env, cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_fle_s(DisasContext *ctx, arg_fle_s *a)
+{
+ REQUIRE_FPU;
+ TCGv t0 = tcg_temp_new();
+ gen_helper_fle_s(t0, cpu_env, cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_fclass_s(DisasContext *ctx, arg_fclass_s *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+
+ gen_helper_fclass_s(t0, cpu_fpr[a->rs1]);
+
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+
+ return true;
+}
+
+static bool trans_fcvt_s_w(DisasContext *ctx, arg_fcvt_s_w *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_s_w(cpu_fpr[a->rd], cpu_env, t0);
+
+ tcg_temp_free(t0);
+
+ return true;
+}
+
+static bool trans_fcvt_s_wu(DisasContext *ctx, arg_fcvt_s_wu *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_s_wu(cpu_fpr[a->rd], cpu_env, t0);
+
+ tcg_temp_free(t0);
+
+ return true;
+}
+
+static bool trans_fmv_w_x(DisasContext *ctx, arg_fmv_w_x *a)
+{
+ /* NOTE: This was FMV.S.X in an earlier version of the ISA spec! */
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+
+#if defined(TARGET_RISCV64)
+ tcg_gen_mov_i64(cpu_fpr[a->rd], t0);
+#else
+ tcg_gen_extu_i32_i64(cpu_fpr[a->rd], t0);
+#endif
+
+ tcg_temp_free(t0);
+
+ return true;
+}
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index a1b949ad0a..63c281b883 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -1504,6 +1504,7 @@ bool decode_insn32(DisasContext *ctx, uint32_t insn);
#include "insn_trans/trans_rvi.inc.c"
#include "insn_trans/trans_rvm.inc.c"
#include "insn_trans/trans_rva.inc.c"
+#include "insn_trans/trans_rvf.inc.c"
static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
{
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 12/35] target/riscv: Convert RV32F insns to decodetree
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- dropped insn argument of trans_foo functions
target/riscv/insn32.decode | 35 +++
target/riscv/insn_trans/trans_rvf.inc.c | 334 ++++++++++++++++++++++++
target/riscv/translate.c | 1 +
3 files changed, 370 insertions(+)
create mode 100644 target/riscv/insn_trans/trans_rvf.inc.c
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index 235f4a031c..57efe0dcff 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -17,12 +17,14 @@
# this program. If not, see <http://www.gnu.org/licenses/>.
# Fields:
+%rs3 27:5
%rs2 20:5
%rs1 15:5
%rd 7:5
%sh6 20:6
%csr 20:12
+%rm 12:3
# immediates:
%imm_i 20:s12
@@ -50,6 +52,11 @@
@atom_ld ..... aq:1 rl:1 ..... ........ ..... ....... &atomic rs2=0 %rs1 %rd
@atom_st ..... aq:1 rl:1 ..... ........ ..... ....... &atomic %rs2 %rs1 %rd
+@r4_rm ..... .. ..... ..... ... ..... ....... %rs3 %rs2 %rs1 %rm %rd
+@r_rm ....... ..... ..... ... ..... ....... %rs2 %rs1 %rm %rd
+@r2_rm ....... ..... ..... ... ..... ....... %rs1 %rm %rd
+@r2 ....... ..... ..... ... ..... ....... %rs1 %rd
+
# *** RV32I Base Instruction Set ***
lui .................... ..... 0110111 @u
auipc .................... ..... 0010111 @u
@@ -119,3 +126,31 @@ amomin_w 10000 . . ..... ..... 010 ..... 0101111 @atom_st
amomax_w 10100 . . ..... ..... 010 ..... 0101111 @atom_st
amominu_w 11000 . . ..... ..... 010 ..... 0101111 @atom_st
amomaxu_w 11100 . . ..... ..... 010 ..... 0101111 @atom_st
+
+# *** RV32F Standard Extension ***
+flw ............ ..... 010 ..... 0000111 @i
+fsw ....... ..... ..... 010 ..... 0100111 @s
+fmadd_s ..... 00 ..... ..... ... ..... 1000011 @r4_rm
+fmsub_s ..... 00 ..... ..... ... ..... 1000111 @r4_rm
+fnmsub_s ..... 00 ..... ..... ... ..... 1001011 @r4_rm
+fnmadd_s ..... 00 ..... ..... ... ..... 1001111 @r4_rm
+fadd_s 0000000 ..... ..... ... ..... 1010011 @r_rm
+fsub_s 0000100 ..... ..... ... ..... 1010011 @r_rm
+fmul_s 0001000 ..... ..... ... ..... 1010011 @r_rm
+fdiv_s 0001100 ..... ..... ... ..... 1010011 @r_rm
+fsqrt_s 0101100 00000 ..... ... ..... 1010011 @r2_rm
+fsgnj_s 0010000 ..... ..... 000 ..... 1010011 @r
+fsgnjn_s 0010000 ..... ..... 001 ..... 1010011 @r
+fsgnjx_s 0010000 ..... ..... 010 ..... 1010011 @r
+fmin_s 0010100 ..... ..... 000 ..... 1010011 @r
+fmax_s 0010100 ..... ..... 001 ..... 1010011 @r
+fcvt_w_s 1100000 00000 ..... ... ..... 1010011 @r2_rm
+fcvt_wu_s 1100000 00001 ..... ... ..... 1010011 @r2_rm
+fmv_x_w 1110000 00000 ..... 000 ..... 1010011 @r2
+feq_s 1010000 ..... ..... 010 ..... 1010011 @r
+flt_s 1010000 ..... ..... 001 ..... 1010011 @r
+fle_s 1010000 ..... ..... 000 ..... 1010011 @r
+fclass_s 1110000 00000 ..... 001 ..... 1010011 @r2
+fcvt_s_w 1101000 00000 ..... ... ..... 1010011 @r2_rm
+fcvt_s_wu 1101000 00001 ..... ... ..... 1010011 @r2_rm
+fmv_w_x 1111000 00000 ..... 000 ..... 1010011 @r2
diff --git a/target/riscv/insn_trans/trans_rvf.inc.c b/target/riscv/insn_trans/trans_rvf.inc.c
new file mode 100644
index 0000000000..b101593ac4
--- /dev/null
+++ b/target/riscv/insn_trans/trans_rvf.inc.c
@@ -0,0 +1,334 @@
+/*
+ * RISC-V translation routines for the RV64F Standard Extension.
+ *
+ * Copyright (c) 2016-2017 Sagar Karandikar, sagark@eecs.berkeley.edu
+ * Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
+ * Bastian Koppelmann, kbastian@mail.uni-paderborn.de
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2 or later, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#define REQUIRE_FPU do {\
+ if (!(ctx->flags & TB_FLAGS_FP_ENABLE)) \
+ return false; \
+} while (0)
+
+static bool trans_flw(DisasContext *ctx, arg_flw *a)
+{
+ TCGv t0 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+ REQUIRE_FPU;
+ tcg_gen_addi_tl(t0, t0, a->imm);
+
+ tcg_gen_qemu_ld_i64(cpu_fpr[a->rd], t0, ctx->mem_idx, MO_TEUL);
+ /* RISC-V requires NaN-boxing of narrower width floating point values */
+ tcg_gen_ori_i64(cpu_fpr[a->rd], cpu_fpr[a->rd], 0xffffffff00000000ULL);
+
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_fsw(DisasContext *ctx, arg_fsw *a)
+{
+ TCGv t0 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+
+ REQUIRE_FPU;
+ tcg_gen_addi_tl(t0, t0, a->imm);
+
+ tcg_gen_qemu_st_i64(cpu_fpr[a->rs2], t0, ctx->mem_idx, MO_TEUL);
+
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_fmadd_s(DisasContext *ctx, arg_fmadd_s *a)
+{
+ REQUIRE_FPU;
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fmadd_s(cpu_fpr[a->rd], cpu_env, cpu_fpr[a->rs1],
+ cpu_fpr[a->rs2], cpu_fpr[a->rs3]);
+ return true;
+}
+
+static bool trans_fmsub_s(DisasContext *ctx, arg_fmsub_s *a)
+{
+ REQUIRE_FPU;
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fmsub_s(cpu_fpr[a->rd], cpu_env, cpu_fpr[a->rs1],
+ cpu_fpr[a->rs2], cpu_fpr[a->rs3]);
+ return true;
+}
+
+static bool trans_fnmsub_s(DisasContext *ctx, arg_fnmsub_s *a)
+{
+ REQUIRE_FPU;
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fnmsub_s(cpu_fpr[a->rd], cpu_env, cpu_fpr[a->rs1],
+ cpu_fpr[a->rs2], cpu_fpr[a->rs3]);
+ return true;
+}
+
+static bool trans_fnmadd_s(DisasContext *ctx, arg_fnmadd_s *a)
+{
+ REQUIRE_FPU;
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fnmadd_s(cpu_fpr[a->rd], cpu_env, cpu_fpr[a->rs1],
+ cpu_fpr[a->rs2], cpu_fpr[a->rs3]);
+ return true;
+}
+
+static bool trans_fadd_s(DisasContext *ctx, arg_fadd_s *a)
+{
+ REQUIRE_FPU;
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fadd_s(cpu_fpr[a->rd], cpu_env,
+ cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+ return true;
+}
+
+static bool trans_fsub_s(DisasContext *ctx, arg_fsub_s *a)
+{
+ REQUIRE_FPU;
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fsub_s(cpu_fpr[a->rd], cpu_env,
+ cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+ return true;
+}
+
+static bool trans_fmul_s(DisasContext *ctx, arg_fmul_s *a)
+{
+ REQUIRE_FPU;
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fmul_s(cpu_fpr[a->rd], cpu_env,
+ cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+ return true;
+}
+
+static bool trans_fdiv_s(DisasContext *ctx, arg_fdiv_s *a)
+{
+ REQUIRE_FPU;
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fdiv_s(cpu_fpr[a->rd], cpu_env,
+ cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+ return true;
+}
+
+static bool trans_fsqrt_s(DisasContext *ctx, arg_fsqrt_s *a)
+{
+ REQUIRE_FPU;
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fsqrt_s(cpu_fpr[a->rd], cpu_env, cpu_fpr[a->rs1]);
+ return true;
+}
+
+static bool trans_fsgnj_s(DisasContext *ctx, arg_fsgnj_s *a)
+{
+ REQUIRE_FPU;
+ if (a->rs1 == a->rs2) { /* FMOV */
+ tcg_gen_mov_i64(cpu_fpr[a->rd], cpu_fpr[a->rs1]);
+ } else { /* FSGNJ */
+ tcg_gen_deposit_i64(cpu_fpr[a->rd], cpu_fpr[a->rs2], cpu_fpr[a->rs1],
+ 0, 31);
+ }
+ return true;
+}
+
+static bool trans_fsgnjn_s(DisasContext *ctx, arg_fsgnjn_s *a)
+{
+ REQUIRE_FPU;
+ if (a->rs1 == a->rs2) { /* FNEG */
+ tcg_gen_xori_i64(cpu_fpr[a->rd], cpu_fpr[a->rs1], INT32_MIN);
+ } else {
+ TCGv_i64 t0 = tcg_temp_new_i64();
+ tcg_gen_not_i64(t0, cpu_fpr[a->rs2]);
+ tcg_gen_deposit_i64(cpu_fpr[a->rd], t0, cpu_fpr[a->rs1], 0, 31);
+ tcg_temp_free_i64(t0);
+ }
+ return true;
+}
+
+static bool trans_fsgnjx_s(DisasContext *ctx, arg_fsgnjx_s *a)
+{
+ REQUIRE_FPU;
+ if (a->rs1 == a->rs2) { /* FABS */
+ tcg_gen_andi_i64(cpu_fpr[a->rd], cpu_fpr[a->rs1], ~INT32_MIN);
+ } else {
+ TCGv_i64 t0 = tcg_temp_new_i64();
+ tcg_gen_andi_i64(t0, cpu_fpr[a->rs2], INT32_MIN);
+ tcg_gen_xor_i64(cpu_fpr[a->rd], cpu_fpr[a->rs1], t0);
+ tcg_temp_free_i64(t0);
+ }
+ return true;
+}
+
+static bool trans_fmin_s(DisasContext *ctx, arg_fmin_s *a)
+{
+ REQUIRE_FPU;
+
+ gen_helper_fmin_s(cpu_fpr[a->rd], cpu_env, cpu_fpr[a->rs1],
+ cpu_fpr[a->rs2]);
+ return true;
+}
+
+static bool trans_fmax_s(DisasContext *ctx, arg_fmax_s *a)
+{
+ REQUIRE_FPU;
+
+ gen_helper_fmax_s(cpu_fpr[a->rd], cpu_env, cpu_fpr[a->rs1],
+ cpu_fpr[a->rs2]);
+ return true;
+}
+
+static bool trans_fcvt_w_s(DisasContext *ctx, arg_fcvt_w_s *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_w_s(t0, cpu_env, cpu_fpr[a->rs1]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+
+ return true;
+}
+
+static bool trans_fcvt_wu_s(DisasContext *ctx, arg_fcvt_wu_s *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_wu_s(t0, cpu_env, cpu_fpr[a->rs1]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+
+ return true;
+}
+
+static bool trans_fmv_x_w(DisasContext *ctx, arg_fmv_x_w *a)
+{
+ /* NOTE: This was FMV.X.S in an earlier version of the ISA spec! */
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+
+#if defined(TARGET_RISCV64)
+ tcg_gen_ext32s_tl(t0, cpu_fpr[a->rs1]);
+#else
+ tcg_gen_extrl_i64_i32(t0, cpu_fpr[a->rs1]);
+#endif
+
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+
+ return true;
+}
+
+static bool trans_feq_s(DisasContext *ctx, arg_feq_s *a)
+{
+ REQUIRE_FPU;
+ TCGv t0 = tcg_temp_new();
+ gen_helper_feq_s(t0, cpu_env, cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_flt_s(DisasContext *ctx, arg_flt_s *a)
+{
+ REQUIRE_FPU;
+ TCGv t0 = tcg_temp_new();
+ gen_helper_flt_s(t0, cpu_env, cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_fle_s(DisasContext *ctx, arg_fle_s *a)
+{
+ REQUIRE_FPU;
+ TCGv t0 = tcg_temp_new();
+ gen_helper_fle_s(t0, cpu_env, cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_fclass_s(DisasContext *ctx, arg_fclass_s *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+
+ gen_helper_fclass_s(t0, cpu_fpr[a->rs1]);
+
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+
+ return true;
+}
+
+static bool trans_fcvt_s_w(DisasContext *ctx, arg_fcvt_s_w *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_s_w(cpu_fpr[a->rd], cpu_env, t0);
+
+ tcg_temp_free(t0);
+
+ return true;
+}
+
+static bool trans_fcvt_s_wu(DisasContext *ctx, arg_fcvt_s_wu *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_s_wu(cpu_fpr[a->rd], cpu_env, t0);
+
+ tcg_temp_free(t0);
+
+ return true;
+}
+
+static bool trans_fmv_w_x(DisasContext *ctx, arg_fmv_w_x *a)
+{
+ /* NOTE: This was FMV.S.X in an earlier version of the ISA spec! */
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+
+#if defined(TARGET_RISCV64)
+ tcg_gen_mov_i64(cpu_fpr[a->rd], t0);
+#else
+ tcg_gen_extu_i32_i64(cpu_fpr[a->rd], t0);
+#endif
+
+ tcg_temp_free(t0);
+
+ return true;
+}
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index a1b949ad0a..63c281b883 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -1504,6 +1504,7 @@ bool decode_insn32(DisasContext *ctx, uint32_t insn);
#include "insn_trans/trans_rvi.inc.c"
#include "insn_trans/trans_rvm.inc.c"
#include "insn_trans/trans_rva.inc.c"
+#include "insn_trans/trans_rvf.inc.c"
static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
{
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 13/35] target/riscv: Convert RV64F insns to decodetree
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- moved 64-bit only insn to insn64.decode
- dropped insn argument of trans_foo functions
target/riscv/insn64.decode | 6 +++
target/riscv/insn_trans/trans_rvf.inc.c | 54 +++++++++++++++++++++++++
2 files changed, 60 insertions(+)
diff --git a/target/riscv/insn64.decode b/target/riscv/insn64.decode
index 0bee95c984..6319f872ac 100644
--- a/target/riscv/insn64.decode
+++ b/target/riscv/insn64.decode
@@ -56,3 +56,9 @@ amomin_d 10000 . . ..... ..... 011 ..... 0101111 @atom_st
amomax_d 10100 . . ..... ..... 011 ..... 0101111 @atom_st
amominu_d 11000 . . ..... ..... 011 ..... 0101111 @atom_st
amomaxu_d 11100 . . ..... ..... 011 ..... 0101111 @atom_st
+
+# *** RV64F Standard Extension (in addition to RV32F) ***
+fcvt_l_s 1100000 00010 ..... ... ..... 1010011 @r2_rm
+fcvt_lu_s 1100000 00011 ..... ... ..... 1010011 @r2_rm
+fcvt_s_l 1101000 00010 ..... ... ..... 1010011 @r2_rm
+fcvt_s_lu 1101000 00011 ..... ... ..... 1010011 @r2_rm
diff --git a/target/riscv/insn_trans/trans_rvf.inc.c b/target/riscv/insn_trans/trans_rvf.inc.c
index b101593ac4..b667c576d4 100644
--- a/target/riscv/insn_trans/trans_rvf.inc.c
+++ b/target/riscv/insn_trans/trans_rvf.inc.c
@@ -332,3 +332,57 @@ static bool trans_fmv_w_x(DisasContext *ctx, arg_fmv_w_x *a)
return true;
}
+
+#ifdef TARGET_RISCV64
+static bool trans_fcvt_l_s(DisasContext *ctx, arg_fcvt_l_s *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_l_s(t0, cpu_env, cpu_fpr[a->rs1]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_fcvt_lu_s(DisasContext *ctx, arg_fcvt_lu_s *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_lu_s(t0, cpu_env, cpu_fpr[a->rs1]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_fcvt_s_l(DisasContext *ctx, arg_fcvt_s_l *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_s_l(cpu_fpr[a->rd], cpu_env, t0);
+
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_fcvt_s_lu(DisasContext *ctx, arg_fcvt_s_lu *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_s_lu(cpu_fpr[a->rd], cpu_env, t0);
+
+ tcg_temp_free(t0);
+ return true;
+}
+#endif
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 13/35] target/riscv: Convert RV64F insns to decodetree
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- moved 64-bit only insn to insn64.decode
- dropped insn argument of trans_foo functions
target/riscv/insn64.decode | 6 +++
target/riscv/insn_trans/trans_rvf.inc.c | 54 +++++++++++++++++++++++++
2 files changed, 60 insertions(+)
diff --git a/target/riscv/insn64.decode b/target/riscv/insn64.decode
index 0bee95c984..6319f872ac 100644
--- a/target/riscv/insn64.decode
+++ b/target/riscv/insn64.decode
@@ -56,3 +56,9 @@ amomin_d 10000 . . ..... ..... 011 ..... 0101111 @atom_st
amomax_d 10100 . . ..... ..... 011 ..... 0101111 @atom_st
amominu_d 11000 . . ..... ..... 011 ..... 0101111 @atom_st
amomaxu_d 11100 . . ..... ..... 011 ..... 0101111 @atom_st
+
+# *** RV64F Standard Extension (in addition to RV32F) ***
+fcvt_l_s 1100000 00010 ..... ... ..... 1010011 @r2_rm
+fcvt_lu_s 1100000 00011 ..... ... ..... 1010011 @r2_rm
+fcvt_s_l 1101000 00010 ..... ... ..... 1010011 @r2_rm
+fcvt_s_lu 1101000 00011 ..... ... ..... 1010011 @r2_rm
diff --git a/target/riscv/insn_trans/trans_rvf.inc.c b/target/riscv/insn_trans/trans_rvf.inc.c
index b101593ac4..b667c576d4 100644
--- a/target/riscv/insn_trans/trans_rvf.inc.c
+++ b/target/riscv/insn_trans/trans_rvf.inc.c
@@ -332,3 +332,57 @@ static bool trans_fmv_w_x(DisasContext *ctx, arg_fmv_w_x *a)
return true;
}
+
+#ifdef TARGET_RISCV64
+static bool trans_fcvt_l_s(DisasContext *ctx, arg_fcvt_l_s *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_l_s(t0, cpu_env, cpu_fpr[a->rs1]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_fcvt_lu_s(DisasContext *ctx, arg_fcvt_lu_s *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_lu_s(t0, cpu_env, cpu_fpr[a->rs1]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_fcvt_s_l(DisasContext *ctx, arg_fcvt_s_l *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_s_l(cpu_fpr[a->rd], cpu_env, t0);
+
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_fcvt_s_lu(DisasContext *ctx, arg_fcvt_s_lu *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_s_lu(cpu_fpr[a->rd], cpu_env, t0);
+
+ tcg_temp_free(t0);
+ return true;
+}
+#endif
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 14/35] target/riscv: Convert RV32D insns to decodetree
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- dropped TARGET_RISCV64 requirement for fclass_d
- dropped insn argument of trans_foo functions
target/riscv/insn32.decode | 28 +++
target/riscv/insn_trans/trans_rvd.inc.c | 315 ++++++++++++++++++++++++
target/riscv/translate.c | 1 +
3 files changed, 344 insertions(+)
create mode 100644 target/riscv/insn_trans/trans_rvd.inc.c
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index 57efe0dcff..b933c349a4 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -154,3 +154,31 @@ fclass_s 1110000 00000 ..... 001 ..... 1010011 @r2
fcvt_s_w 1101000 00000 ..... ... ..... 1010011 @r2_rm
fcvt_s_wu 1101000 00001 ..... ... ..... 1010011 @r2_rm
fmv_w_x 1111000 00000 ..... 000 ..... 1010011 @r2
+
+# *** RV32D Standard Extension ***
+fld ............ ..... 011 ..... 0000111 @i
+fsd ....... ..... ..... 011 ..... 0100111 @s
+fmadd_d ..... 01 ..... ..... ... ..... 1000011 @r4_rm
+fmsub_d ..... 01 ..... ..... ... ..... 1000111 @r4_rm
+fnmsub_d ..... 01 ..... ..... ... ..... 1001011 @r4_rm
+fnmadd_d ..... 01 ..... ..... ... ..... 1001111 @r4_rm
+fadd_d 0000001 ..... ..... ... ..... 1010011 @r_rm
+fsub_d 0000101 ..... ..... ... ..... 1010011 @r_rm
+fmul_d 0001001 ..... ..... ... ..... 1010011 @r_rm
+fdiv_d 0001101 ..... ..... ... ..... 1010011 @r_rm
+fsqrt_d 0101101 00000 ..... ... ..... 1010011 @r2_rm
+fsgnj_d 0010001 ..... ..... 000 ..... 1010011 @r
+fsgnjn_d 0010001 ..... ..... 001 ..... 1010011 @r
+fsgnjx_d 0010001 ..... ..... 010 ..... 1010011 @r
+fmin_d 0010101 ..... ..... 000 ..... 1010011 @r
+fmax_d 0010101 ..... ..... 001 ..... 1010011 @r
+fcvt_s_d 0100000 00001 ..... ... ..... 1010011 @r2_rm
+fcvt_d_s 0100001 00000 ..... ... ..... 1010011 @r2_rm
+feq_d 1010001 ..... ..... 010 ..... 1010011 @r
+flt_d 1010001 ..... ..... 001 ..... 1010011 @r
+fle_d 1010001 ..... ..... 000 ..... 1010011 @r
+fclass_d 1110001 00000 ..... 001 ..... 1010011 @r2
+fcvt_w_d 1100001 00000 ..... ... ..... 1010011 @r2_rm
+fcvt_wu_d 1100001 00001 ..... ... ..... 1010011 @r2_rm
+fcvt_d_w 1101001 00000 ..... ... ..... 1010011 @r2_rm
+fcvt_d_wu 1101001 00001 ..... ... ..... 1010011 @r2_rm
diff --git a/target/riscv/insn_trans/trans_rvd.inc.c b/target/riscv/insn_trans/trans_rvd.inc.c
new file mode 100644
index 0000000000..a7e2335ffa
--- /dev/null
+++ b/target/riscv/insn_trans/trans_rvd.inc.c
@@ -0,0 +1,315 @@
+/*
+ * RISC-V translation routines for the RV64D Standard Extension.
+ *
+ * Copyright (c) 2016-2017 Sagar Karandikar, sagark@eecs.berkeley.edu
+ * Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
+ * Bastian Koppelmann, kbastian@mail.uni-paderborn.de
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2 or later, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+static bool trans_fld(DisasContext *ctx, arg_fld *a)
+{
+ TCGv t0 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+ REQUIRE_FPU;
+ tcg_gen_addi_tl(t0, t0, a->imm);
+
+ tcg_gen_qemu_ld_i64(cpu_fpr[a->rd], t0, ctx->mem_idx, MO_TEQ);
+
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_fsd(DisasContext *ctx, arg_fsd *a)
+{
+ TCGv t0 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+ REQUIRE_FPU;
+ tcg_gen_addi_tl(t0, t0, a->imm);
+
+ tcg_gen_qemu_st_i64(cpu_fpr[a->rs2], t0, ctx->mem_idx, MO_TEQ);
+
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_fmadd_d(DisasContext *ctx, arg_fmadd_d *a)
+{
+ REQUIRE_FPU;
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fmadd_d(cpu_fpr[a->rd], cpu_env, cpu_fpr[a->rs1],
+ cpu_fpr[a->rs2], cpu_fpr[a->rs3]);
+ return true;
+}
+
+static bool trans_fmsub_d(DisasContext *ctx, arg_fmsub_d *a)
+{
+ REQUIRE_FPU;
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fmsub_d(cpu_fpr[a->rd], cpu_env, cpu_fpr[a->rs1],
+ cpu_fpr[a->rs2], cpu_fpr[a->rs3]);
+ return true;
+}
+
+static bool trans_fnmsub_d(DisasContext *ctx, arg_fnmsub_d *a)
+{
+ REQUIRE_FPU;
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fnmsub_d(cpu_fpr[a->rd], cpu_env, cpu_fpr[a->rs1],
+ cpu_fpr[a->rs2], cpu_fpr[a->rs3]);
+ return true;
+}
+
+static bool trans_fnmadd_d(DisasContext *ctx, arg_fnmadd_d *a)
+{
+ REQUIRE_FPU;
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fnmadd_d(cpu_fpr[a->rd], cpu_env, cpu_fpr[a->rs1],
+ cpu_fpr[a->rs2], cpu_fpr[a->rs3]);
+ return true;
+}
+
+static bool trans_fadd_d(DisasContext *ctx, arg_fadd_d *a)
+{
+ REQUIRE_FPU;
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fadd_d(cpu_fpr[a->rd], cpu_env,
+ cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+
+ return true;
+}
+
+static bool trans_fsub_d(DisasContext *ctx, arg_fsub_d *a)
+{
+ REQUIRE_FPU;
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fsub_d(cpu_fpr[a->rd], cpu_env,
+ cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+
+ return true;
+}
+
+static bool trans_fmul_d(DisasContext *ctx, arg_fmul_d *a)
+{
+ REQUIRE_FPU;
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fmul_d(cpu_fpr[a->rd], cpu_env,
+ cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+
+ return true;
+}
+
+static bool trans_fdiv_d(DisasContext *ctx, arg_fdiv_d *a)
+{
+ REQUIRE_FPU;
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fdiv_d(cpu_fpr[a->rd], cpu_env,
+ cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+
+ return true;
+}
+
+static bool trans_fsqrt_d(DisasContext *ctx, arg_fsqrt_d *a)
+{
+ REQUIRE_FPU;
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fsqrt_d(cpu_fpr[a->rd], cpu_env, cpu_fpr[a->rs1]);
+
+ return true;
+}
+
+static bool trans_fsgnj_d(DisasContext *ctx, arg_fsgnj_d *a)
+{
+ if (a->rs1 == a->rs2) { /* FMOV */
+ tcg_gen_mov_i64(cpu_fpr[a->rd], cpu_fpr[a->rs1]);
+ } else {
+ tcg_gen_deposit_i64(cpu_fpr[a->rd], cpu_fpr[a->rs2],
+ cpu_fpr[a->rs1], 0, 63);
+ }
+ return true;
+}
+
+static bool trans_fsgnjn_d(DisasContext *ctx, arg_fsgnjn_d *a)
+{
+ REQUIRE_FPU;
+ if (a->rs1 == a->rs2) { /* FNEG */
+ tcg_gen_xori_i64(cpu_fpr[a->rd], cpu_fpr[a->rs1], INT64_MIN);
+ } else {
+ TCGv_i64 t0 = tcg_temp_new_i64();
+ tcg_gen_not_i64(t0, cpu_fpr[a->rs2]);
+ tcg_gen_deposit_i64(cpu_fpr[a->rd], t0, cpu_fpr[a->rs1], 0, 63);
+ tcg_temp_free_i64(t0);
+ }
+ return true;
+}
+
+static bool trans_fsgnjx_d(DisasContext *ctx, arg_fsgnjx_d *a)
+{
+ REQUIRE_FPU;
+ if (a->rs1 == a->rs2) { /* FABS */
+ tcg_gen_andi_i64(cpu_fpr[a->rd], cpu_fpr[a->rs1], ~INT64_MIN);
+ } else {
+ TCGv_i64 t0 = tcg_temp_new_i64();
+ tcg_gen_andi_i64(t0, cpu_fpr[a->rs2], INT64_MIN);
+ tcg_gen_xor_i64(cpu_fpr[a->rd], cpu_fpr[a->rs1], t0);
+ tcg_temp_free_i64(t0);
+ }
+ return true;
+}
+
+static bool trans_fmin_d(DisasContext *ctx, arg_fmin_d *a)
+{
+ REQUIRE_FPU;
+
+ gen_helper_fmin_d(cpu_fpr[a->rd], cpu_env,
+ cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+
+ return true;
+}
+
+static bool trans_fmax_d(DisasContext *ctx, arg_fmax_d *a)
+{
+ REQUIRE_FPU;
+
+ gen_helper_fmax_d(cpu_fpr[a->rd], cpu_env,
+ cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+
+ return true;
+}
+
+static bool trans_fcvt_s_d(DisasContext *ctx, arg_fcvt_s_d *a)
+{
+ REQUIRE_FPU;
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_s_d(cpu_fpr[a->rd], cpu_env, cpu_fpr[a->rs1]);
+
+ return true;
+}
+
+static bool trans_fcvt_d_s(DisasContext *ctx, arg_fcvt_d_s *a)
+{
+ REQUIRE_FPU;
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_d_s(cpu_fpr[a->rd], cpu_env, cpu_fpr[a->rs1]);
+
+ return true;
+}
+
+static bool trans_feq_d(DisasContext *ctx, arg_feq_d *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_helper_feq_d(t0, cpu_env, cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+
+ return true;
+}
+
+static bool trans_flt_d(DisasContext *ctx, arg_flt_d *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_helper_flt_d(t0, cpu_env, cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+
+ return true;
+}
+
+static bool trans_fle_d(DisasContext *ctx, arg_fle_d *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_helper_fle_d(t0, cpu_env, cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+
+ return true;
+}
+
+static bool trans_fclass_d(DisasContext *ctx, arg_fclass_d *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_helper_fclass_d(t0, cpu_fpr[a->rs1]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_fcvt_w_d(DisasContext *ctx, arg_fcvt_w_d *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_w_d(t0, cpu_env, cpu_fpr[a->rs1]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+
+ return true;
+}
+
+static bool trans_fcvt_wu_d(DisasContext *ctx, arg_fcvt_wu_d *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_wu_d(t0, cpu_env, cpu_fpr[a->rs1]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+
+ return true;
+}
+
+static bool trans_fcvt_d_w(DisasContext *ctx, arg_fcvt_d_w *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_d_w(cpu_fpr[a->rd], cpu_env, t0);
+ tcg_temp_free(t0);
+
+ return true;
+}
+
+static bool trans_fcvt_d_wu(DisasContext *ctx, arg_fcvt_d_wu *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_d_wu(cpu_fpr[a->rd], cpu_env, t0);
+ tcg_temp_free(t0);
+
+ return true;
+}
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 63c281b883..0f30ccbfbf 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -1505,6 +1505,7 @@ bool decode_insn32(DisasContext *ctx, uint32_t insn);
#include "insn_trans/trans_rvm.inc.c"
#include "insn_trans/trans_rva.inc.c"
#include "insn_trans/trans_rvf.inc.c"
+#include "insn_trans/trans_rvd.inc.c"
static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
{
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 14/35] target/riscv: Convert RV32D insns to decodetree
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- dropped TARGET_RISCV64 requirement for fclass_d
- dropped insn argument of trans_foo functions
target/riscv/insn32.decode | 28 +++
target/riscv/insn_trans/trans_rvd.inc.c | 315 ++++++++++++++++++++++++
target/riscv/translate.c | 1 +
3 files changed, 344 insertions(+)
create mode 100644 target/riscv/insn_trans/trans_rvd.inc.c
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index 57efe0dcff..b933c349a4 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -154,3 +154,31 @@ fclass_s 1110000 00000 ..... 001 ..... 1010011 @r2
fcvt_s_w 1101000 00000 ..... ... ..... 1010011 @r2_rm
fcvt_s_wu 1101000 00001 ..... ... ..... 1010011 @r2_rm
fmv_w_x 1111000 00000 ..... 000 ..... 1010011 @r2
+
+# *** RV32D Standard Extension ***
+fld ............ ..... 011 ..... 0000111 @i
+fsd ....... ..... ..... 011 ..... 0100111 @s
+fmadd_d ..... 01 ..... ..... ... ..... 1000011 @r4_rm
+fmsub_d ..... 01 ..... ..... ... ..... 1000111 @r4_rm
+fnmsub_d ..... 01 ..... ..... ... ..... 1001011 @r4_rm
+fnmadd_d ..... 01 ..... ..... ... ..... 1001111 @r4_rm
+fadd_d 0000001 ..... ..... ... ..... 1010011 @r_rm
+fsub_d 0000101 ..... ..... ... ..... 1010011 @r_rm
+fmul_d 0001001 ..... ..... ... ..... 1010011 @r_rm
+fdiv_d 0001101 ..... ..... ... ..... 1010011 @r_rm
+fsqrt_d 0101101 00000 ..... ... ..... 1010011 @r2_rm
+fsgnj_d 0010001 ..... ..... 000 ..... 1010011 @r
+fsgnjn_d 0010001 ..... ..... 001 ..... 1010011 @r
+fsgnjx_d 0010001 ..... ..... 010 ..... 1010011 @r
+fmin_d 0010101 ..... ..... 000 ..... 1010011 @r
+fmax_d 0010101 ..... ..... 001 ..... 1010011 @r
+fcvt_s_d 0100000 00001 ..... ... ..... 1010011 @r2_rm
+fcvt_d_s 0100001 00000 ..... ... ..... 1010011 @r2_rm
+feq_d 1010001 ..... ..... 010 ..... 1010011 @r
+flt_d 1010001 ..... ..... 001 ..... 1010011 @r
+fle_d 1010001 ..... ..... 000 ..... 1010011 @r
+fclass_d 1110001 00000 ..... 001 ..... 1010011 @r2
+fcvt_w_d 1100001 00000 ..... ... ..... 1010011 @r2_rm
+fcvt_wu_d 1100001 00001 ..... ... ..... 1010011 @r2_rm
+fcvt_d_w 1101001 00000 ..... ... ..... 1010011 @r2_rm
+fcvt_d_wu 1101001 00001 ..... ... ..... 1010011 @r2_rm
diff --git a/target/riscv/insn_trans/trans_rvd.inc.c b/target/riscv/insn_trans/trans_rvd.inc.c
new file mode 100644
index 0000000000..a7e2335ffa
--- /dev/null
+++ b/target/riscv/insn_trans/trans_rvd.inc.c
@@ -0,0 +1,315 @@
+/*
+ * RISC-V translation routines for the RV64D Standard Extension.
+ *
+ * Copyright (c) 2016-2017 Sagar Karandikar, sagark@eecs.berkeley.edu
+ * Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
+ * Bastian Koppelmann, kbastian@mail.uni-paderborn.de
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2 or later, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+static bool trans_fld(DisasContext *ctx, arg_fld *a)
+{
+ TCGv t0 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+ REQUIRE_FPU;
+ tcg_gen_addi_tl(t0, t0, a->imm);
+
+ tcg_gen_qemu_ld_i64(cpu_fpr[a->rd], t0, ctx->mem_idx, MO_TEQ);
+
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_fsd(DisasContext *ctx, arg_fsd *a)
+{
+ TCGv t0 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+ REQUIRE_FPU;
+ tcg_gen_addi_tl(t0, t0, a->imm);
+
+ tcg_gen_qemu_st_i64(cpu_fpr[a->rs2], t0, ctx->mem_idx, MO_TEQ);
+
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_fmadd_d(DisasContext *ctx, arg_fmadd_d *a)
+{
+ REQUIRE_FPU;
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fmadd_d(cpu_fpr[a->rd], cpu_env, cpu_fpr[a->rs1],
+ cpu_fpr[a->rs2], cpu_fpr[a->rs3]);
+ return true;
+}
+
+static bool trans_fmsub_d(DisasContext *ctx, arg_fmsub_d *a)
+{
+ REQUIRE_FPU;
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fmsub_d(cpu_fpr[a->rd], cpu_env, cpu_fpr[a->rs1],
+ cpu_fpr[a->rs2], cpu_fpr[a->rs3]);
+ return true;
+}
+
+static bool trans_fnmsub_d(DisasContext *ctx, arg_fnmsub_d *a)
+{
+ REQUIRE_FPU;
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fnmsub_d(cpu_fpr[a->rd], cpu_env, cpu_fpr[a->rs1],
+ cpu_fpr[a->rs2], cpu_fpr[a->rs3]);
+ return true;
+}
+
+static bool trans_fnmadd_d(DisasContext *ctx, arg_fnmadd_d *a)
+{
+ REQUIRE_FPU;
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fnmadd_d(cpu_fpr[a->rd], cpu_env, cpu_fpr[a->rs1],
+ cpu_fpr[a->rs2], cpu_fpr[a->rs3]);
+ return true;
+}
+
+static bool trans_fadd_d(DisasContext *ctx, arg_fadd_d *a)
+{
+ REQUIRE_FPU;
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fadd_d(cpu_fpr[a->rd], cpu_env,
+ cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+
+ return true;
+}
+
+static bool trans_fsub_d(DisasContext *ctx, arg_fsub_d *a)
+{
+ REQUIRE_FPU;
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fsub_d(cpu_fpr[a->rd], cpu_env,
+ cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+
+ return true;
+}
+
+static bool trans_fmul_d(DisasContext *ctx, arg_fmul_d *a)
+{
+ REQUIRE_FPU;
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fmul_d(cpu_fpr[a->rd], cpu_env,
+ cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+
+ return true;
+}
+
+static bool trans_fdiv_d(DisasContext *ctx, arg_fdiv_d *a)
+{
+ REQUIRE_FPU;
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fdiv_d(cpu_fpr[a->rd], cpu_env,
+ cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+
+ return true;
+}
+
+static bool trans_fsqrt_d(DisasContext *ctx, arg_fsqrt_d *a)
+{
+ REQUIRE_FPU;
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fsqrt_d(cpu_fpr[a->rd], cpu_env, cpu_fpr[a->rs1]);
+
+ return true;
+}
+
+static bool trans_fsgnj_d(DisasContext *ctx, arg_fsgnj_d *a)
+{
+ if (a->rs1 == a->rs2) { /* FMOV */
+ tcg_gen_mov_i64(cpu_fpr[a->rd], cpu_fpr[a->rs1]);
+ } else {
+ tcg_gen_deposit_i64(cpu_fpr[a->rd], cpu_fpr[a->rs2],
+ cpu_fpr[a->rs1], 0, 63);
+ }
+ return true;
+}
+
+static bool trans_fsgnjn_d(DisasContext *ctx, arg_fsgnjn_d *a)
+{
+ REQUIRE_FPU;
+ if (a->rs1 == a->rs2) { /* FNEG */
+ tcg_gen_xori_i64(cpu_fpr[a->rd], cpu_fpr[a->rs1], INT64_MIN);
+ } else {
+ TCGv_i64 t0 = tcg_temp_new_i64();
+ tcg_gen_not_i64(t0, cpu_fpr[a->rs2]);
+ tcg_gen_deposit_i64(cpu_fpr[a->rd], t0, cpu_fpr[a->rs1], 0, 63);
+ tcg_temp_free_i64(t0);
+ }
+ return true;
+}
+
+static bool trans_fsgnjx_d(DisasContext *ctx, arg_fsgnjx_d *a)
+{
+ REQUIRE_FPU;
+ if (a->rs1 == a->rs2) { /* FABS */
+ tcg_gen_andi_i64(cpu_fpr[a->rd], cpu_fpr[a->rs1], ~INT64_MIN);
+ } else {
+ TCGv_i64 t0 = tcg_temp_new_i64();
+ tcg_gen_andi_i64(t0, cpu_fpr[a->rs2], INT64_MIN);
+ tcg_gen_xor_i64(cpu_fpr[a->rd], cpu_fpr[a->rs1], t0);
+ tcg_temp_free_i64(t0);
+ }
+ return true;
+}
+
+static bool trans_fmin_d(DisasContext *ctx, arg_fmin_d *a)
+{
+ REQUIRE_FPU;
+
+ gen_helper_fmin_d(cpu_fpr[a->rd], cpu_env,
+ cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+
+ return true;
+}
+
+static bool trans_fmax_d(DisasContext *ctx, arg_fmax_d *a)
+{
+ REQUIRE_FPU;
+
+ gen_helper_fmax_d(cpu_fpr[a->rd], cpu_env,
+ cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+
+ return true;
+}
+
+static bool trans_fcvt_s_d(DisasContext *ctx, arg_fcvt_s_d *a)
+{
+ REQUIRE_FPU;
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_s_d(cpu_fpr[a->rd], cpu_env, cpu_fpr[a->rs1]);
+
+ return true;
+}
+
+static bool trans_fcvt_d_s(DisasContext *ctx, arg_fcvt_d_s *a)
+{
+ REQUIRE_FPU;
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_d_s(cpu_fpr[a->rd], cpu_env, cpu_fpr[a->rs1]);
+
+ return true;
+}
+
+static bool trans_feq_d(DisasContext *ctx, arg_feq_d *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_helper_feq_d(t0, cpu_env, cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+
+ return true;
+}
+
+static bool trans_flt_d(DisasContext *ctx, arg_flt_d *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_helper_flt_d(t0, cpu_env, cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+
+ return true;
+}
+
+static bool trans_fle_d(DisasContext *ctx, arg_fle_d *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_helper_fle_d(t0, cpu_env, cpu_fpr[a->rs1], cpu_fpr[a->rs2]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+
+ return true;
+}
+
+static bool trans_fclass_d(DisasContext *ctx, arg_fclass_d *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_helper_fclass_d(t0, cpu_fpr[a->rs1]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_fcvt_w_d(DisasContext *ctx, arg_fcvt_w_d *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_w_d(t0, cpu_env, cpu_fpr[a->rs1]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+
+ return true;
+}
+
+static bool trans_fcvt_wu_d(DisasContext *ctx, arg_fcvt_wu_d *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_wu_d(t0, cpu_env, cpu_fpr[a->rs1]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+
+ return true;
+}
+
+static bool trans_fcvt_d_w(DisasContext *ctx, arg_fcvt_d_w *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_d_w(cpu_fpr[a->rd], cpu_env, t0);
+ tcg_temp_free(t0);
+
+ return true;
+}
+
+static bool trans_fcvt_d_wu(DisasContext *ctx, arg_fcvt_d_wu *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_d_wu(cpu_fpr[a->rd], cpu_env, t0);
+ tcg_temp_free(t0);
+
+ return true;
+}
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 63c281b883..0f30ccbfbf 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -1505,6 +1505,7 @@ bool decode_insn32(DisasContext *ctx, uint32_t insn);
#include "insn_trans/trans_rvm.inc.c"
#include "insn_trans/trans_rva.inc.c"
#include "insn_trans/trans_rvf.inc.c"
+#include "insn_trans/trans_rvd.inc.c"
static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
{
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 15/35] target/riscv: Convert RV64D insns to decodetree
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- moved 64-bit only insn to insn64.decode
- dropped insn argument of trans_foo functions
target/riscv/insn64.decode | 8 +
target/riscv/insn_trans/trans_rvd.inc.c | 73 ++++
target/riscv/translate.c | 484 +-----------------------
3 files changed, 82 insertions(+), 483 deletions(-)
diff --git a/target/riscv/insn64.decode b/target/riscv/insn64.decode
index 6319f872ac..380bf791bc 100644
--- a/target/riscv/insn64.decode
+++ b/target/riscv/insn64.decode
@@ -62,3 +62,11 @@ fcvt_l_s 1100000 00010 ..... ... ..... 1010011 @r2_rm
fcvt_lu_s 1100000 00011 ..... ... ..... 1010011 @r2_rm
fcvt_s_l 1101000 00010 ..... ... ..... 1010011 @r2_rm
fcvt_s_lu 1101000 00011 ..... ... ..... 1010011 @r2_rm
+
+# *** RV64D Standard Extension (in addition to RV32D) ***
+fcvt_l_d 1100001 00010 ..... ... ..... 1010011 @r2_rm
+fcvt_lu_d 1100001 00011 ..... ... ..... 1010011 @r2_rm
+fmv_x_d 1110001 00000 ..... 000 ..... 1010011 @r2
+fcvt_d_l 1101001 00010 ..... ... ..... 1010011 @r2_rm
+fcvt_d_lu 1101001 00011 ..... ... ..... 1010011 @r2_rm
+fmv_d_x 1111001 00000 ..... 000 ..... 1010011 @r2
diff --git a/target/riscv/insn_trans/trans_rvd.inc.c b/target/riscv/insn_trans/trans_rvd.inc.c
index a7e2335ffa..2e45c8ed7e 100644
--- a/target/riscv/insn_trans/trans_rvd.inc.c
+++ b/target/riscv/insn_trans/trans_rvd.inc.c
@@ -313,3 +313,76 @@ static bool trans_fcvt_d_wu(DisasContext *ctx, arg_fcvt_d_wu *a)
return true;
}
+
+#ifdef TARGET_RISCV64
+
+static bool trans_fcvt_l_d(DisasContext *ctx, arg_fcvt_l_d *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_l_d(t0, cpu_env, cpu_fpr[a->rs1]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_fcvt_lu_d(DisasContext *ctx, arg_fcvt_lu_d *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_lu_d(t0, cpu_env, cpu_fpr[a->rs1]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_fmv_x_d(DisasContext *ctx, arg_fmv_x_d *a)
+{
+ REQUIRE_FPU;
+
+ gen_set_gpr(a->rd, cpu_fpr[a->rs1]);
+ return true;
+}
+
+static bool trans_fcvt_d_l(DisasContext *ctx, arg_fcvt_d_l *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_d_l(cpu_fpr[a->rd], cpu_env, t0);
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_fcvt_d_lu(DisasContext *ctx, arg_fcvt_d_lu *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_d_lu(cpu_fpr[a->rd], cpu_env, t0);
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_fmv_d_x(DisasContext *ctx, arg_fmv_d_x *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+
+ tcg_gen_mov_tl(cpu_fpr[a->rd], t0);
+ tcg_temp_free(t0);
+ return true;
+}
+#endif
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 0f30ccbfbf..8e91fa8284 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -180,44 +180,6 @@ static void gen_mulhsu(TCGv ret, TCGv arg1, TCGv arg2)
tcg_temp_free(rh);
}
-static void gen_fsgnj(DisasContext *ctx, uint32_t rd, uint32_t rs1,
- uint32_t rs2, int rm, uint64_t min)
-{
- switch (rm) {
- case 0: /* fsgnj */
- if (rs1 == rs2) { /* FMOV */
- tcg_gen_mov_i64(cpu_fpr[rd], cpu_fpr[rs1]);
- } else {
- tcg_gen_deposit_i64(cpu_fpr[rd], cpu_fpr[rs2], cpu_fpr[rs1],
- 0, min == INT32_MIN ? 31 : 63);
- }
- break;
- case 1: /* fsgnjn */
- if (rs1 == rs2) { /* FNEG */
- tcg_gen_xori_i64(cpu_fpr[rd], cpu_fpr[rs1], min);
- } else {
- TCGv_i64 t0 = tcg_temp_new_i64();
- tcg_gen_not_i64(t0, cpu_fpr[rs2]);
- tcg_gen_deposit_i64(cpu_fpr[rd], t0, cpu_fpr[rs1],
- 0, min == INT32_MIN ? 31 : 63);
- tcg_temp_free_i64(t0);
- }
- break;
- case 2: /* fsgnjx */
- if (rs1 == rs2) { /* FABS */
- tcg_gen_andi_i64(cpu_fpr[rd], cpu_fpr[rs1], ~min);
- } else {
- TCGv_i64 t0 = tcg_temp_new_i64();
- tcg_gen_andi_i64(t0, cpu_fpr[rs2], min);
- tcg_gen_xor_i64(cpu_fpr[rd], cpu_fpr[rs1], t0);
- tcg_temp_free_i64(t0);
- }
- break;
- default:
- gen_exception_illegal(ctx);
- }
-}
-
static void gen_arith(DisasContext *ctx, uint32_t opc, int rd, int rs1,
int rs2)
{
@@ -724,418 +686,6 @@ static void gen_set_rm(DisasContext *ctx, int rm)
tcg_temp_free_i32(t0);
}
-static void gen_fp_fmadd(DisasContext *ctx, uint32_t opc, int rd,
- int rs1, int rs2, int rs3, int rm)
-{
- switch (opc) {
- case OPC_RISC_FMADD_S:
- gen_set_rm(ctx, rm);
- gen_helper_fmadd_s(cpu_fpr[rd], cpu_env, cpu_fpr[rs1],
- cpu_fpr[rs2], cpu_fpr[rs3]);
- break;
- case OPC_RISC_FMADD_D:
- gen_set_rm(ctx, rm);
- gen_helper_fmadd_d(cpu_fpr[rd], cpu_env, cpu_fpr[rs1],
- cpu_fpr[rs2], cpu_fpr[rs3]);
- break;
- default:
- gen_exception_illegal(ctx);
- break;
- }
-}
-
-static void gen_fp_fmsub(DisasContext *ctx, uint32_t opc, int rd,
- int rs1, int rs2, int rs3, int rm)
-{
- switch (opc) {
- case OPC_RISC_FMSUB_S:
- gen_set_rm(ctx, rm);
- gen_helper_fmsub_s(cpu_fpr[rd], cpu_env, cpu_fpr[rs1],
- cpu_fpr[rs2], cpu_fpr[rs3]);
- break;
- case OPC_RISC_FMSUB_D:
- gen_set_rm(ctx, rm);
- gen_helper_fmsub_d(cpu_fpr[rd], cpu_env, cpu_fpr[rs1],
- cpu_fpr[rs2], cpu_fpr[rs3]);
- break;
- default:
- gen_exception_illegal(ctx);
- break;
- }
-}
-
-static void gen_fp_fnmsub(DisasContext *ctx, uint32_t opc, int rd,
- int rs1, int rs2, int rs3, int rm)
-{
- switch (opc) {
- case OPC_RISC_FNMSUB_S:
- gen_set_rm(ctx, rm);
- gen_helper_fnmsub_s(cpu_fpr[rd], cpu_env, cpu_fpr[rs1],
- cpu_fpr[rs2], cpu_fpr[rs3]);
- break;
- case OPC_RISC_FNMSUB_D:
- gen_set_rm(ctx, rm);
- gen_helper_fnmsub_d(cpu_fpr[rd], cpu_env, cpu_fpr[rs1],
- cpu_fpr[rs2], cpu_fpr[rs3]);
- break;
- default:
- gen_exception_illegal(ctx);
- break;
- }
-}
-
-static void gen_fp_fnmadd(DisasContext *ctx, uint32_t opc, int rd,
- int rs1, int rs2, int rs3, int rm)
-{
- switch (opc) {
- case OPC_RISC_FNMADD_S:
- gen_set_rm(ctx, rm);
- gen_helper_fnmadd_s(cpu_fpr[rd], cpu_env, cpu_fpr[rs1],
- cpu_fpr[rs2], cpu_fpr[rs3]);
- break;
- case OPC_RISC_FNMADD_D:
- gen_set_rm(ctx, rm);
- gen_helper_fnmadd_d(cpu_fpr[rd], cpu_env, cpu_fpr[rs1],
- cpu_fpr[rs2], cpu_fpr[rs3]);
- break;
- default:
- gen_exception_illegal(ctx);
- break;
- }
-}
-
-static void gen_fp_arith(DisasContext *ctx, uint32_t opc, int rd,
- int rs1, int rs2, int rm)
-{
- TCGv t0 = NULL;
-
- if (!(ctx->flags & TB_FLAGS_FP_ENABLE)) {
- goto do_illegal;
- }
-
- switch (opc) {
- case OPC_RISC_FADD_S:
- gen_set_rm(ctx, rm);
- gen_helper_fadd_s(cpu_fpr[rd], cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- case OPC_RISC_FSUB_S:
- gen_set_rm(ctx, rm);
- gen_helper_fsub_s(cpu_fpr[rd], cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- case OPC_RISC_FMUL_S:
- gen_set_rm(ctx, rm);
- gen_helper_fmul_s(cpu_fpr[rd], cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- case OPC_RISC_FDIV_S:
- gen_set_rm(ctx, rm);
- gen_helper_fdiv_s(cpu_fpr[rd], cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- case OPC_RISC_FSQRT_S:
- gen_set_rm(ctx, rm);
- gen_helper_fsqrt_s(cpu_fpr[rd], cpu_env, cpu_fpr[rs1]);
- break;
- case OPC_RISC_FSGNJ_S:
- gen_fsgnj(ctx, rd, rs1, rs2, rm, INT32_MIN);
- break;
-
- case OPC_RISC_FMIN_S:
- /* also handles: OPC_RISC_FMAX_S */
- switch (rm) {
- case 0x0:
- gen_helper_fmin_s(cpu_fpr[rd], cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- case 0x1:
- gen_helper_fmax_s(cpu_fpr[rd], cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- default:
- goto do_illegal;
- }
- break;
-
- case OPC_RISC_FEQ_S:
- /* also handles: OPC_RISC_FLT_S, OPC_RISC_FLE_S */
- t0 = tcg_temp_new();
- switch (rm) {
- case 0x0:
- gen_helper_fle_s(t0, cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- case 0x1:
- gen_helper_flt_s(t0, cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- case 0x2:
- gen_helper_feq_s(t0, cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- default:
- goto do_illegal;
- }
- gen_set_gpr(rd, t0);
- tcg_temp_free(t0);
- break;
-
- case OPC_RISC_FCVT_W_S:
- /* also OPC_RISC_FCVT_WU_S, OPC_RISC_FCVT_L_S, OPC_RISC_FCVT_LU_S */
- t0 = tcg_temp_new();
- switch (rs2) {
- case 0: /* FCVT_W_S */
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_w_s(t0, cpu_env, cpu_fpr[rs1]);
- break;
- case 1: /* FCVT_WU_S */
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_wu_s(t0, cpu_env, cpu_fpr[rs1]);
- break;
-#if defined(TARGET_RISCV64)
- case 2: /* FCVT_L_S */
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_l_s(t0, cpu_env, cpu_fpr[rs1]);
- break;
- case 3: /* FCVT_LU_S */
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_lu_s(t0, cpu_env, cpu_fpr[rs1]);
- break;
-#endif
- default:
- goto do_illegal;
- }
- gen_set_gpr(rd, t0);
- tcg_temp_free(t0);
- break;
-
- case OPC_RISC_FCVT_S_W:
- /* also OPC_RISC_FCVT_S_WU, OPC_RISC_FCVT_S_L, OPC_RISC_FCVT_S_LU */
- t0 = tcg_temp_new();
- gen_get_gpr(t0, rs1);
- switch (rs2) {
- case 0: /* FCVT_S_W */
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_s_w(cpu_fpr[rd], cpu_env, t0);
- break;
- case 1: /* FCVT_S_WU */
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_s_wu(cpu_fpr[rd], cpu_env, t0);
- break;
-#if defined(TARGET_RISCV64)
- case 2: /* FCVT_S_L */
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_s_l(cpu_fpr[rd], cpu_env, t0);
- break;
- case 3: /* FCVT_S_LU */
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_s_lu(cpu_fpr[rd], cpu_env, t0);
- break;
-#endif
- default:
- goto do_illegal;
- }
- tcg_temp_free(t0);
- break;
-
- case OPC_RISC_FMV_X_S:
- /* also OPC_RISC_FCLASS_S */
- t0 = tcg_temp_new();
- switch (rm) {
- case 0: /* FMV */
-#if defined(TARGET_RISCV64)
- tcg_gen_ext32s_tl(t0, cpu_fpr[rs1]);
-#else
- tcg_gen_extrl_i64_i32(t0, cpu_fpr[rs1]);
-#endif
- break;
- case 1:
- gen_helper_fclass_s(t0, cpu_fpr[rs1]);
- break;
- default:
- goto do_illegal;
- }
- gen_set_gpr(rd, t0);
- tcg_temp_free(t0);
- break;
-
- case OPC_RISC_FMV_S_X:
- t0 = tcg_temp_new();
- gen_get_gpr(t0, rs1);
-#if defined(TARGET_RISCV64)
- tcg_gen_mov_i64(cpu_fpr[rd], t0);
-#else
- tcg_gen_extu_i32_i64(cpu_fpr[rd], t0);
-#endif
- tcg_temp_free(t0);
- break;
-
- /* double */
- case OPC_RISC_FADD_D:
- gen_set_rm(ctx, rm);
- gen_helper_fadd_d(cpu_fpr[rd], cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- case OPC_RISC_FSUB_D:
- gen_set_rm(ctx, rm);
- gen_helper_fsub_d(cpu_fpr[rd], cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- case OPC_RISC_FMUL_D:
- gen_set_rm(ctx, rm);
- gen_helper_fmul_d(cpu_fpr[rd], cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- case OPC_RISC_FDIV_D:
- gen_set_rm(ctx, rm);
- gen_helper_fdiv_d(cpu_fpr[rd], cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- case OPC_RISC_FSQRT_D:
- gen_set_rm(ctx, rm);
- gen_helper_fsqrt_d(cpu_fpr[rd], cpu_env, cpu_fpr[rs1]);
- break;
- case OPC_RISC_FSGNJ_D:
- gen_fsgnj(ctx, rd, rs1, rs2, rm, INT64_MIN);
- break;
-
- case OPC_RISC_FMIN_D:
- /* also OPC_RISC_FMAX_D */
- switch (rm) {
- case 0:
- gen_helper_fmin_d(cpu_fpr[rd], cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- case 1:
- gen_helper_fmax_d(cpu_fpr[rd], cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- default:
- goto do_illegal;
- }
- break;
-
- case OPC_RISC_FCVT_S_D:
- switch (rs2) {
- case 1:
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_s_d(cpu_fpr[rd], cpu_env, cpu_fpr[rs1]);
- break;
- default:
- goto do_illegal;
- }
- break;
-
- case OPC_RISC_FCVT_D_S:
- switch (rs2) {
- case 0:
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_d_s(cpu_fpr[rd], cpu_env, cpu_fpr[rs1]);
- break;
- default:
- goto do_illegal;
- }
- break;
-
- case OPC_RISC_FEQ_D:
- /* also OPC_RISC_FLT_D, OPC_RISC_FLE_D */
- t0 = tcg_temp_new();
- switch (rm) {
- case 0:
- gen_helper_fle_d(t0, cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- case 1:
- gen_helper_flt_d(t0, cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- case 2:
- gen_helper_feq_d(t0, cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- default:
- goto do_illegal;
- }
- gen_set_gpr(rd, t0);
- tcg_temp_free(t0);
- break;
-
- case OPC_RISC_FCVT_W_D:
- /* also OPC_RISC_FCVT_WU_D, OPC_RISC_FCVT_L_D, OPC_RISC_FCVT_LU_D */
- t0 = tcg_temp_new();
- switch (rs2) {
- case 0:
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_w_d(t0, cpu_env, cpu_fpr[rs1]);
- break;
- case 1:
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_wu_d(t0, cpu_env, cpu_fpr[rs1]);
- break;
-#if defined(TARGET_RISCV64)
- case 2:
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_l_d(t0, cpu_env, cpu_fpr[rs1]);
- break;
- case 3:
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_lu_d(t0, cpu_env, cpu_fpr[rs1]);
- break;
-#endif
- default:
- goto do_illegal;
- }
- gen_set_gpr(rd, t0);
- tcg_temp_free(t0);
- break;
-
- case OPC_RISC_FCVT_D_W:
- /* also OPC_RISC_FCVT_D_WU, OPC_RISC_FCVT_D_L, OPC_RISC_FCVT_D_LU */
- t0 = tcg_temp_new();
- gen_get_gpr(t0, rs1);
- switch (rs2) {
- case 0:
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_d_w(cpu_fpr[rd], cpu_env, t0);
- break;
- case 1:
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_d_wu(cpu_fpr[rd], cpu_env, t0);
- break;
-#if defined(TARGET_RISCV64)
- case 2:
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_d_l(cpu_fpr[rd], cpu_env, t0);
- break;
- case 3:
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_d_lu(cpu_fpr[rd], cpu_env, t0);
- break;
-#endif
- default:
- goto do_illegal;
- }
- tcg_temp_free(t0);
- break;
-
-#if defined(TARGET_RISCV64)
- case OPC_RISC_FMV_X_D:
- /* also OPC_RISC_FCLASS_D */
- switch (rm) {
- case 0: /* FMV */
- gen_set_gpr(rd, cpu_fpr[rs1]);
- break;
- case 1:
- t0 = tcg_temp_new();
- gen_helper_fclass_d(t0, cpu_fpr[rs1]);
- gen_set_gpr(rd, t0);
- tcg_temp_free(t0);
- break;
- default:
- goto do_illegal;
- }
- break;
-
- case OPC_RISC_FMV_D_X:
- t0 = tcg_temp_new();
- gen_get_gpr(t0, rs1);
- tcg_gen_mov_tl(cpu_fpr[rd], t0);
- tcg_temp_free(t0);
- break;
-#endif
-
- default:
- do_illegal:
- if (t0) {
- tcg_temp_free(t0);
- }
- gen_exception_illegal(ctx);
- break;
- }
-}
static void gen_system(CPURISCVState *env, DisasContext *ctx, uint32_t opc,
int rd, int rs1, int csr)
@@ -1509,11 +1059,8 @@ bool decode_insn32(DisasContext *ctx, uint32_t insn);
static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
{
- int rs1;
- int rs2;
- int rd;
+ int rs1, rd;
uint32_t op;
- target_long imm;
/* We do not do misaligned address check here: the address should never be
* misaligned at this point. Instructions that set PC must do the check,
@@ -1522,38 +1069,9 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
op = MASK_OP_MAJOR(ctx->opcode);
rs1 = GET_RS1(ctx->opcode);
- rs2 = GET_RS2(ctx->opcode);
rd = GET_RD(ctx->opcode);
- imm = GET_IMM(ctx->opcode);
switch (op) {
- case OPC_RISC_FP_LOAD:
- gen_fp_load(ctx, MASK_OP_FP_LOAD(ctx->opcode), rd, rs1, imm);
- break;
- case OPC_RISC_FP_STORE:
- gen_fp_store(ctx, MASK_OP_FP_STORE(ctx->opcode), rs1, rs2,
- GET_STORE_IMM(ctx->opcode));
- break;
- case OPC_RISC_FMADD:
- gen_fp_fmadd(ctx, MASK_OP_FP_FMADD(ctx->opcode), rd, rs1, rs2,
- GET_RS3(ctx->opcode), GET_RM(ctx->opcode));
- break;
- case OPC_RISC_FMSUB:
- gen_fp_fmsub(ctx, MASK_OP_FP_FMSUB(ctx->opcode), rd, rs1, rs2,
- GET_RS3(ctx->opcode), GET_RM(ctx->opcode));
- break;
- case OPC_RISC_FNMSUB:
- gen_fp_fnmsub(ctx, MASK_OP_FP_FNMSUB(ctx->opcode), rd, rs1, rs2,
- GET_RS3(ctx->opcode), GET_RM(ctx->opcode));
- break;
- case OPC_RISC_FNMADD:
- gen_fp_fnmadd(ctx, MASK_OP_FP_FNMADD(ctx->opcode), rd, rs1, rs2,
- GET_RS3(ctx->opcode), GET_RM(ctx->opcode));
- break;
- case OPC_RISC_FP_ARITH:
- gen_fp_arith(ctx, MASK_OP_FP_ARITH(ctx->opcode), rd, rs1, rs2,
- GET_RM(ctx->opcode));
- break;
case OPC_RISC_SYSTEM:
gen_system(env, ctx, MASK_OP_SYSTEM(ctx->opcode), rd, rs1,
(ctx->opcode & 0xFFF00000) >> 20);
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 15/35] target/riscv: Convert RV64D insns to decodetree
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- moved 64-bit only insn to insn64.decode
- dropped insn argument of trans_foo functions
target/riscv/insn64.decode | 8 +
target/riscv/insn_trans/trans_rvd.inc.c | 73 ++++
target/riscv/translate.c | 484 +-----------------------
3 files changed, 82 insertions(+), 483 deletions(-)
diff --git a/target/riscv/insn64.decode b/target/riscv/insn64.decode
index 6319f872ac..380bf791bc 100644
--- a/target/riscv/insn64.decode
+++ b/target/riscv/insn64.decode
@@ -62,3 +62,11 @@ fcvt_l_s 1100000 00010 ..... ... ..... 1010011 @r2_rm
fcvt_lu_s 1100000 00011 ..... ... ..... 1010011 @r2_rm
fcvt_s_l 1101000 00010 ..... ... ..... 1010011 @r2_rm
fcvt_s_lu 1101000 00011 ..... ... ..... 1010011 @r2_rm
+
+# *** RV64D Standard Extension (in addition to RV32D) ***
+fcvt_l_d 1100001 00010 ..... ... ..... 1010011 @r2_rm
+fcvt_lu_d 1100001 00011 ..... ... ..... 1010011 @r2_rm
+fmv_x_d 1110001 00000 ..... 000 ..... 1010011 @r2
+fcvt_d_l 1101001 00010 ..... ... ..... 1010011 @r2_rm
+fcvt_d_lu 1101001 00011 ..... ... ..... 1010011 @r2_rm
+fmv_d_x 1111001 00000 ..... 000 ..... 1010011 @r2
diff --git a/target/riscv/insn_trans/trans_rvd.inc.c b/target/riscv/insn_trans/trans_rvd.inc.c
index a7e2335ffa..2e45c8ed7e 100644
--- a/target/riscv/insn_trans/trans_rvd.inc.c
+++ b/target/riscv/insn_trans/trans_rvd.inc.c
@@ -313,3 +313,76 @@ static bool trans_fcvt_d_wu(DisasContext *ctx, arg_fcvt_d_wu *a)
return true;
}
+
+#ifdef TARGET_RISCV64
+
+static bool trans_fcvt_l_d(DisasContext *ctx, arg_fcvt_l_d *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_l_d(t0, cpu_env, cpu_fpr[a->rs1]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_fcvt_lu_d(DisasContext *ctx, arg_fcvt_lu_d *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_lu_d(t0, cpu_env, cpu_fpr[a->rs1]);
+ gen_set_gpr(a->rd, t0);
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_fmv_x_d(DisasContext *ctx, arg_fmv_x_d *a)
+{
+ REQUIRE_FPU;
+
+ gen_set_gpr(a->rd, cpu_fpr[a->rs1]);
+ return true;
+}
+
+static bool trans_fcvt_d_l(DisasContext *ctx, arg_fcvt_d_l *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_d_l(cpu_fpr[a->rd], cpu_env, t0);
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_fcvt_d_lu(DisasContext *ctx, arg_fcvt_d_lu *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+
+ gen_set_rm(ctx, a->rm);
+ gen_helper_fcvt_d_lu(cpu_fpr[a->rd], cpu_env, t0);
+ tcg_temp_free(t0);
+ return true;
+}
+
+static bool trans_fmv_d_x(DisasContext *ctx, arg_fmv_d_x *a)
+{
+ REQUIRE_FPU;
+
+ TCGv t0 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+
+ tcg_gen_mov_tl(cpu_fpr[a->rd], t0);
+ tcg_temp_free(t0);
+ return true;
+}
+#endif
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 0f30ccbfbf..8e91fa8284 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -180,44 +180,6 @@ static void gen_mulhsu(TCGv ret, TCGv arg1, TCGv arg2)
tcg_temp_free(rh);
}
-static void gen_fsgnj(DisasContext *ctx, uint32_t rd, uint32_t rs1,
- uint32_t rs2, int rm, uint64_t min)
-{
- switch (rm) {
- case 0: /* fsgnj */
- if (rs1 == rs2) { /* FMOV */
- tcg_gen_mov_i64(cpu_fpr[rd], cpu_fpr[rs1]);
- } else {
- tcg_gen_deposit_i64(cpu_fpr[rd], cpu_fpr[rs2], cpu_fpr[rs1],
- 0, min == INT32_MIN ? 31 : 63);
- }
- break;
- case 1: /* fsgnjn */
- if (rs1 == rs2) { /* FNEG */
- tcg_gen_xori_i64(cpu_fpr[rd], cpu_fpr[rs1], min);
- } else {
- TCGv_i64 t0 = tcg_temp_new_i64();
- tcg_gen_not_i64(t0, cpu_fpr[rs2]);
- tcg_gen_deposit_i64(cpu_fpr[rd], t0, cpu_fpr[rs1],
- 0, min == INT32_MIN ? 31 : 63);
- tcg_temp_free_i64(t0);
- }
- break;
- case 2: /* fsgnjx */
- if (rs1 == rs2) { /* FABS */
- tcg_gen_andi_i64(cpu_fpr[rd], cpu_fpr[rs1], ~min);
- } else {
- TCGv_i64 t0 = tcg_temp_new_i64();
- tcg_gen_andi_i64(t0, cpu_fpr[rs2], min);
- tcg_gen_xor_i64(cpu_fpr[rd], cpu_fpr[rs1], t0);
- tcg_temp_free_i64(t0);
- }
- break;
- default:
- gen_exception_illegal(ctx);
- }
-}
-
static void gen_arith(DisasContext *ctx, uint32_t opc, int rd, int rs1,
int rs2)
{
@@ -724,418 +686,6 @@ static void gen_set_rm(DisasContext *ctx, int rm)
tcg_temp_free_i32(t0);
}
-static void gen_fp_fmadd(DisasContext *ctx, uint32_t opc, int rd,
- int rs1, int rs2, int rs3, int rm)
-{
- switch (opc) {
- case OPC_RISC_FMADD_S:
- gen_set_rm(ctx, rm);
- gen_helper_fmadd_s(cpu_fpr[rd], cpu_env, cpu_fpr[rs1],
- cpu_fpr[rs2], cpu_fpr[rs3]);
- break;
- case OPC_RISC_FMADD_D:
- gen_set_rm(ctx, rm);
- gen_helper_fmadd_d(cpu_fpr[rd], cpu_env, cpu_fpr[rs1],
- cpu_fpr[rs2], cpu_fpr[rs3]);
- break;
- default:
- gen_exception_illegal(ctx);
- break;
- }
-}
-
-static void gen_fp_fmsub(DisasContext *ctx, uint32_t opc, int rd,
- int rs1, int rs2, int rs3, int rm)
-{
- switch (opc) {
- case OPC_RISC_FMSUB_S:
- gen_set_rm(ctx, rm);
- gen_helper_fmsub_s(cpu_fpr[rd], cpu_env, cpu_fpr[rs1],
- cpu_fpr[rs2], cpu_fpr[rs3]);
- break;
- case OPC_RISC_FMSUB_D:
- gen_set_rm(ctx, rm);
- gen_helper_fmsub_d(cpu_fpr[rd], cpu_env, cpu_fpr[rs1],
- cpu_fpr[rs2], cpu_fpr[rs3]);
- break;
- default:
- gen_exception_illegal(ctx);
- break;
- }
-}
-
-static void gen_fp_fnmsub(DisasContext *ctx, uint32_t opc, int rd,
- int rs1, int rs2, int rs3, int rm)
-{
- switch (opc) {
- case OPC_RISC_FNMSUB_S:
- gen_set_rm(ctx, rm);
- gen_helper_fnmsub_s(cpu_fpr[rd], cpu_env, cpu_fpr[rs1],
- cpu_fpr[rs2], cpu_fpr[rs3]);
- break;
- case OPC_RISC_FNMSUB_D:
- gen_set_rm(ctx, rm);
- gen_helper_fnmsub_d(cpu_fpr[rd], cpu_env, cpu_fpr[rs1],
- cpu_fpr[rs2], cpu_fpr[rs3]);
- break;
- default:
- gen_exception_illegal(ctx);
- break;
- }
-}
-
-static void gen_fp_fnmadd(DisasContext *ctx, uint32_t opc, int rd,
- int rs1, int rs2, int rs3, int rm)
-{
- switch (opc) {
- case OPC_RISC_FNMADD_S:
- gen_set_rm(ctx, rm);
- gen_helper_fnmadd_s(cpu_fpr[rd], cpu_env, cpu_fpr[rs1],
- cpu_fpr[rs2], cpu_fpr[rs3]);
- break;
- case OPC_RISC_FNMADD_D:
- gen_set_rm(ctx, rm);
- gen_helper_fnmadd_d(cpu_fpr[rd], cpu_env, cpu_fpr[rs1],
- cpu_fpr[rs2], cpu_fpr[rs3]);
- break;
- default:
- gen_exception_illegal(ctx);
- break;
- }
-}
-
-static void gen_fp_arith(DisasContext *ctx, uint32_t opc, int rd,
- int rs1, int rs2, int rm)
-{
- TCGv t0 = NULL;
-
- if (!(ctx->flags & TB_FLAGS_FP_ENABLE)) {
- goto do_illegal;
- }
-
- switch (opc) {
- case OPC_RISC_FADD_S:
- gen_set_rm(ctx, rm);
- gen_helper_fadd_s(cpu_fpr[rd], cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- case OPC_RISC_FSUB_S:
- gen_set_rm(ctx, rm);
- gen_helper_fsub_s(cpu_fpr[rd], cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- case OPC_RISC_FMUL_S:
- gen_set_rm(ctx, rm);
- gen_helper_fmul_s(cpu_fpr[rd], cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- case OPC_RISC_FDIV_S:
- gen_set_rm(ctx, rm);
- gen_helper_fdiv_s(cpu_fpr[rd], cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- case OPC_RISC_FSQRT_S:
- gen_set_rm(ctx, rm);
- gen_helper_fsqrt_s(cpu_fpr[rd], cpu_env, cpu_fpr[rs1]);
- break;
- case OPC_RISC_FSGNJ_S:
- gen_fsgnj(ctx, rd, rs1, rs2, rm, INT32_MIN);
- break;
-
- case OPC_RISC_FMIN_S:
- /* also handles: OPC_RISC_FMAX_S */
- switch (rm) {
- case 0x0:
- gen_helper_fmin_s(cpu_fpr[rd], cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- case 0x1:
- gen_helper_fmax_s(cpu_fpr[rd], cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- default:
- goto do_illegal;
- }
- break;
-
- case OPC_RISC_FEQ_S:
- /* also handles: OPC_RISC_FLT_S, OPC_RISC_FLE_S */
- t0 = tcg_temp_new();
- switch (rm) {
- case 0x0:
- gen_helper_fle_s(t0, cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- case 0x1:
- gen_helper_flt_s(t0, cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- case 0x2:
- gen_helper_feq_s(t0, cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- default:
- goto do_illegal;
- }
- gen_set_gpr(rd, t0);
- tcg_temp_free(t0);
- break;
-
- case OPC_RISC_FCVT_W_S:
- /* also OPC_RISC_FCVT_WU_S, OPC_RISC_FCVT_L_S, OPC_RISC_FCVT_LU_S */
- t0 = tcg_temp_new();
- switch (rs2) {
- case 0: /* FCVT_W_S */
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_w_s(t0, cpu_env, cpu_fpr[rs1]);
- break;
- case 1: /* FCVT_WU_S */
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_wu_s(t0, cpu_env, cpu_fpr[rs1]);
- break;
-#if defined(TARGET_RISCV64)
- case 2: /* FCVT_L_S */
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_l_s(t0, cpu_env, cpu_fpr[rs1]);
- break;
- case 3: /* FCVT_LU_S */
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_lu_s(t0, cpu_env, cpu_fpr[rs1]);
- break;
-#endif
- default:
- goto do_illegal;
- }
- gen_set_gpr(rd, t0);
- tcg_temp_free(t0);
- break;
-
- case OPC_RISC_FCVT_S_W:
- /* also OPC_RISC_FCVT_S_WU, OPC_RISC_FCVT_S_L, OPC_RISC_FCVT_S_LU */
- t0 = tcg_temp_new();
- gen_get_gpr(t0, rs1);
- switch (rs2) {
- case 0: /* FCVT_S_W */
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_s_w(cpu_fpr[rd], cpu_env, t0);
- break;
- case 1: /* FCVT_S_WU */
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_s_wu(cpu_fpr[rd], cpu_env, t0);
- break;
-#if defined(TARGET_RISCV64)
- case 2: /* FCVT_S_L */
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_s_l(cpu_fpr[rd], cpu_env, t0);
- break;
- case 3: /* FCVT_S_LU */
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_s_lu(cpu_fpr[rd], cpu_env, t0);
- break;
-#endif
- default:
- goto do_illegal;
- }
- tcg_temp_free(t0);
- break;
-
- case OPC_RISC_FMV_X_S:
- /* also OPC_RISC_FCLASS_S */
- t0 = tcg_temp_new();
- switch (rm) {
- case 0: /* FMV */
-#if defined(TARGET_RISCV64)
- tcg_gen_ext32s_tl(t0, cpu_fpr[rs1]);
-#else
- tcg_gen_extrl_i64_i32(t0, cpu_fpr[rs1]);
-#endif
- break;
- case 1:
- gen_helper_fclass_s(t0, cpu_fpr[rs1]);
- break;
- default:
- goto do_illegal;
- }
- gen_set_gpr(rd, t0);
- tcg_temp_free(t0);
- break;
-
- case OPC_RISC_FMV_S_X:
- t0 = tcg_temp_new();
- gen_get_gpr(t0, rs1);
-#if defined(TARGET_RISCV64)
- tcg_gen_mov_i64(cpu_fpr[rd], t0);
-#else
- tcg_gen_extu_i32_i64(cpu_fpr[rd], t0);
-#endif
- tcg_temp_free(t0);
- break;
-
- /* double */
- case OPC_RISC_FADD_D:
- gen_set_rm(ctx, rm);
- gen_helper_fadd_d(cpu_fpr[rd], cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- case OPC_RISC_FSUB_D:
- gen_set_rm(ctx, rm);
- gen_helper_fsub_d(cpu_fpr[rd], cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- case OPC_RISC_FMUL_D:
- gen_set_rm(ctx, rm);
- gen_helper_fmul_d(cpu_fpr[rd], cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- case OPC_RISC_FDIV_D:
- gen_set_rm(ctx, rm);
- gen_helper_fdiv_d(cpu_fpr[rd], cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- case OPC_RISC_FSQRT_D:
- gen_set_rm(ctx, rm);
- gen_helper_fsqrt_d(cpu_fpr[rd], cpu_env, cpu_fpr[rs1]);
- break;
- case OPC_RISC_FSGNJ_D:
- gen_fsgnj(ctx, rd, rs1, rs2, rm, INT64_MIN);
- break;
-
- case OPC_RISC_FMIN_D:
- /* also OPC_RISC_FMAX_D */
- switch (rm) {
- case 0:
- gen_helper_fmin_d(cpu_fpr[rd], cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- case 1:
- gen_helper_fmax_d(cpu_fpr[rd], cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- default:
- goto do_illegal;
- }
- break;
-
- case OPC_RISC_FCVT_S_D:
- switch (rs2) {
- case 1:
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_s_d(cpu_fpr[rd], cpu_env, cpu_fpr[rs1]);
- break;
- default:
- goto do_illegal;
- }
- break;
-
- case OPC_RISC_FCVT_D_S:
- switch (rs2) {
- case 0:
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_d_s(cpu_fpr[rd], cpu_env, cpu_fpr[rs1]);
- break;
- default:
- goto do_illegal;
- }
- break;
-
- case OPC_RISC_FEQ_D:
- /* also OPC_RISC_FLT_D, OPC_RISC_FLE_D */
- t0 = tcg_temp_new();
- switch (rm) {
- case 0:
- gen_helper_fle_d(t0, cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- case 1:
- gen_helper_flt_d(t0, cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- case 2:
- gen_helper_feq_d(t0, cpu_env, cpu_fpr[rs1], cpu_fpr[rs2]);
- break;
- default:
- goto do_illegal;
- }
- gen_set_gpr(rd, t0);
- tcg_temp_free(t0);
- break;
-
- case OPC_RISC_FCVT_W_D:
- /* also OPC_RISC_FCVT_WU_D, OPC_RISC_FCVT_L_D, OPC_RISC_FCVT_LU_D */
- t0 = tcg_temp_new();
- switch (rs2) {
- case 0:
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_w_d(t0, cpu_env, cpu_fpr[rs1]);
- break;
- case 1:
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_wu_d(t0, cpu_env, cpu_fpr[rs1]);
- break;
-#if defined(TARGET_RISCV64)
- case 2:
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_l_d(t0, cpu_env, cpu_fpr[rs1]);
- break;
- case 3:
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_lu_d(t0, cpu_env, cpu_fpr[rs1]);
- break;
-#endif
- default:
- goto do_illegal;
- }
- gen_set_gpr(rd, t0);
- tcg_temp_free(t0);
- break;
-
- case OPC_RISC_FCVT_D_W:
- /* also OPC_RISC_FCVT_D_WU, OPC_RISC_FCVT_D_L, OPC_RISC_FCVT_D_LU */
- t0 = tcg_temp_new();
- gen_get_gpr(t0, rs1);
- switch (rs2) {
- case 0:
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_d_w(cpu_fpr[rd], cpu_env, t0);
- break;
- case 1:
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_d_wu(cpu_fpr[rd], cpu_env, t0);
- break;
-#if defined(TARGET_RISCV64)
- case 2:
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_d_l(cpu_fpr[rd], cpu_env, t0);
- break;
- case 3:
- gen_set_rm(ctx, rm);
- gen_helper_fcvt_d_lu(cpu_fpr[rd], cpu_env, t0);
- break;
-#endif
- default:
- goto do_illegal;
- }
- tcg_temp_free(t0);
- break;
-
-#if defined(TARGET_RISCV64)
- case OPC_RISC_FMV_X_D:
- /* also OPC_RISC_FCLASS_D */
- switch (rm) {
- case 0: /* FMV */
- gen_set_gpr(rd, cpu_fpr[rs1]);
- break;
- case 1:
- t0 = tcg_temp_new();
- gen_helper_fclass_d(t0, cpu_fpr[rs1]);
- gen_set_gpr(rd, t0);
- tcg_temp_free(t0);
- break;
- default:
- goto do_illegal;
- }
- break;
-
- case OPC_RISC_FMV_D_X:
- t0 = tcg_temp_new();
- gen_get_gpr(t0, rs1);
- tcg_gen_mov_tl(cpu_fpr[rd], t0);
- tcg_temp_free(t0);
- break;
-#endif
-
- default:
- do_illegal:
- if (t0) {
- tcg_temp_free(t0);
- }
- gen_exception_illegal(ctx);
- break;
- }
-}
static void gen_system(CPURISCVState *env, DisasContext *ctx, uint32_t opc,
int rd, int rs1, int csr)
@@ -1509,11 +1059,8 @@ bool decode_insn32(DisasContext *ctx, uint32_t insn);
static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
{
- int rs1;
- int rs2;
- int rd;
+ int rs1, rd;
uint32_t op;
- target_long imm;
/* We do not do misaligned address check here: the address should never be
* misaligned at this point. Instructions that set PC must do the check,
@@ -1522,38 +1069,9 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
op = MASK_OP_MAJOR(ctx->opcode);
rs1 = GET_RS1(ctx->opcode);
- rs2 = GET_RS2(ctx->opcode);
rd = GET_RD(ctx->opcode);
- imm = GET_IMM(ctx->opcode);
switch (op) {
- case OPC_RISC_FP_LOAD:
- gen_fp_load(ctx, MASK_OP_FP_LOAD(ctx->opcode), rd, rs1, imm);
- break;
- case OPC_RISC_FP_STORE:
- gen_fp_store(ctx, MASK_OP_FP_STORE(ctx->opcode), rs1, rs2,
- GET_STORE_IMM(ctx->opcode));
- break;
- case OPC_RISC_FMADD:
- gen_fp_fmadd(ctx, MASK_OP_FP_FMADD(ctx->opcode), rd, rs1, rs2,
- GET_RS3(ctx->opcode), GET_RM(ctx->opcode));
- break;
- case OPC_RISC_FMSUB:
- gen_fp_fmsub(ctx, MASK_OP_FP_FMSUB(ctx->opcode), rd, rs1, rs2,
- GET_RS3(ctx->opcode), GET_RM(ctx->opcode));
- break;
- case OPC_RISC_FNMSUB:
- gen_fp_fnmsub(ctx, MASK_OP_FP_FNMSUB(ctx->opcode), rd, rs1, rs2,
- GET_RS3(ctx->opcode), GET_RM(ctx->opcode));
- break;
- case OPC_RISC_FNMADD:
- gen_fp_fnmadd(ctx, MASK_OP_FP_FNMADD(ctx->opcode), rd, rs1, rs2,
- GET_RS3(ctx->opcode), GET_RM(ctx->opcode));
- break;
- case OPC_RISC_FP_ARITH:
- gen_fp_arith(ctx, MASK_OP_FP_ARITH(ctx->opcode), rd, rs1, rs2,
- GET_RM(ctx->opcode));
- break;
case OPC_RISC_SYSTEM:
gen_system(env, ctx, MASK_OP_SYSTEM(ctx->opcode), rd, rs1,
(ctx->opcode & 0xFFF00000) >> 20);
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 16/35] target/riscv: Convert RV priv insns to decodetree
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- illegal instruction call replaced with return false
- dropped insn argument of trans_foo functions
target/riscv/insn32.decode | 15 +++
.../riscv/insn_trans/trans_privileged.inc.c | 108 ++++++++++++++++++
target/riscv/translate.c | 49 +-------
3 files changed, 124 insertions(+), 48 deletions(-)
create mode 100644 target/riscv/insn_trans/trans_privileged.inc.c
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index b933c349a4..379a5d5438 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -57,6 +57,21 @@
@r2_rm ....... ..... ..... ... ..... ....... %rs1 %rm %rd
@r2 ....... ..... ..... ... ..... ....... %rs1 %rd
+@sfence_vma ....... ..... ..... ... ..... ....... %rs2 %rs1
+@sfence_vm ....... ..... ..... ... ..... ....... %rs1
+
+
+# *** Privileged Instructions ***
+ecall 000000000000 00000 000 00000 1110011
+ebreak 000000000001 00000 000 00000 1110011
+uret 0000000 00010 00000 000 00000 1110011
+sret 0001000 00010 00000 000 00000 1110011
+hret 0010000 00010 00000 000 00000 1110011
+mret 0011000 00010 00000 000 00000 1110011
+wfi 0001000 00101 00000 000 00000 1110011
+sfence_vma 0001001 ..... ..... 000 00000 1110011 @sfence_vma
+sfence_vm 0001000 00100 ..... 000 00000 1110011 @sfence_vm
+
# *** RV32I Base Instruction Set ***
lui .................... ..... 0110111 @u
auipc .................... ..... 0010111 @u
diff --git a/target/riscv/insn_trans/trans_privileged.inc.c b/target/riscv/insn_trans/trans_privileged.inc.c
new file mode 100644
index 0000000000..7a4f0c7846
--- /dev/null
+++ b/target/riscv/insn_trans/trans_privileged.inc.c
@@ -0,0 +1,108 @@
+/*
+ * RISC-V translation routines for the RISC-V privileged instructions.
+ *
+ * Copyright (c) 2016-2017 Sagar Karandikar, sagark@eecs.berkeley.edu
+ * Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
+ * Bastian Koppelmann, kbastian@mail.uni-paderborn.de
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2 or later, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+static bool trans_ecall(DisasContext *ctx, arg_ecall *a)
+{
+ /* always generates U-level ECALL, fixed in do_interrupt handler */
+ generate_exception(ctx, RISCV_EXCP_U_ECALL);
+ tcg_gen_exit_tb(NULL, 0); /* no chaining */
+ ctx->base.is_jmp = DISAS_NORETURN;
+ return true;
+}
+
+static bool trans_ebreak(DisasContext *ctx, arg_ebreak *a)
+{
+ generate_exception(ctx, RISCV_EXCP_BREAKPOINT);
+ tcg_gen_exit_tb(NULL, 0); /* no chaining */
+ ctx->base.is_jmp = DISAS_NORETURN;
+ return true;
+}
+
+static bool trans_uret(DisasContext *ctx, arg_uret *a)
+{
+ return false;
+}
+
+static bool trans_sret(DisasContext *ctx, arg_sret *a)
+{
+#ifndef CONFIG_USER_ONLY
+ tcg_gen_movi_tl(cpu_pc, ctx->base.pc_next);
+ CPURISCVState *env = current_cpu->env_ptr;
+ if (riscv_has_ext(env, RVS)) {
+ gen_helper_sret(cpu_pc, cpu_env, cpu_pc);
+ tcg_gen_exit_tb(NULL, 0); /* no chaining */
+ ctx->base.is_jmp = DISAS_NORETURN;
+ } else {
+ return false;
+ }
+ return true;
+#else
+ return false;
+#endif
+}
+
+static bool trans_hret(DisasContext *ctx, arg_hret *a)
+{
+ return false;
+}
+
+static bool trans_mret(DisasContext *ctx, arg_mret *a)
+{
+#ifndef CONFIG_USER_ONLY
+ tcg_gen_movi_tl(cpu_pc, ctx->base.pc_next);
+ gen_helper_mret(cpu_pc, cpu_env, cpu_pc);
+ tcg_gen_exit_tb(NULL, 0); /* no chaining */
+ ctx->base.is_jmp = DISAS_NORETURN;
+ return true;
+#else
+ return false;
+#endif
+}
+
+static bool trans_wfi(DisasContext *ctx, arg_wfi *a)
+{
+#ifndef CONFIG_USER_ONLY
+ tcg_gen_movi_tl(cpu_pc, ctx->pc_succ_insn);
+ gen_helper_wfi(cpu_env);
+ return true;
+#else
+ return false;
+#endif
+}
+
+static bool trans_sfence_vma(DisasContext *ctx, arg_sfence_vma *a)
+{
+#ifndef CONFIG_USER_ONLY
+ gen_helper_tlb_flush(cpu_env);
+ return true;
+#else
+ return false;
+#endif
+}
+
+static bool trans_sfence_vm(DisasContext *ctx, arg_sfence_vm *a)
+{
+#ifndef CONFIG_USER_ONLY
+ gen_helper_tlb_flush(cpu_env);
+ return true;
+#else
+ return false;
+#endif
+}
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 8e91fa8284..885e48881c 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -690,22 +690,8 @@ static void gen_set_rm(DisasContext *ctx, int rm)
static void gen_system(CPURISCVState *env, DisasContext *ctx, uint32_t opc,
int rd, int rs1, int csr)
{
- TCGv source1, dest;
- source1 = tcg_temp_new();
- dest = tcg_temp_new();
- gen_get_gpr(source1, rs1);
tcg_gen_movi_tl(cpu_pc, ctx->base.pc_next);
-#ifndef CONFIG_USER_ONLY
- /* Extract funct7 value and check whether it matches SFENCE.VMA */
- if ((opc == OPC_RISC_ECALL) && ((csr >> 5) == 9)) {
- /* sfence.vma */
- /* TODO: handle ASID specific fences */
- gen_helper_tlb_flush(cpu_env);
- return;
- }
-#endif
-
switch (opc) {
case OPC_RISC_ECALL:
switch (csr) {
@@ -720,46 +706,12 @@ static void gen_system(CPURISCVState *env, DisasContext *ctx, uint32_t opc,
tcg_gen_exit_tb(NULL, 0); /* no chaining */
ctx->base.is_jmp = DISAS_NORETURN;
break;
-#ifndef CONFIG_USER_ONLY
- case 0x002: /* URET */
- gen_exception_illegal(ctx);
- break;
- case 0x102: /* SRET */
- if (riscv_has_ext(env, RVS)) {
- gen_helper_sret(cpu_pc, cpu_env, cpu_pc);
- tcg_gen_exit_tb(NULL, 0); /* no chaining */
- ctx->base.is_jmp = DISAS_NORETURN;
- } else {
- gen_exception_illegal(ctx);
- }
- break;
- case 0x202: /* HRET */
- gen_exception_illegal(ctx);
- break;
- case 0x302: /* MRET */
- gen_helper_mret(cpu_pc, cpu_env, cpu_pc);
- tcg_gen_exit_tb(NULL, 0); /* no chaining */
- ctx->base.is_jmp = DISAS_NORETURN;
- break;
- case 0x7b2: /* DRET */
- gen_exception_illegal(ctx);
- break;
- case 0x105: /* WFI */
- tcg_gen_movi_tl(cpu_pc, ctx->pc_succ_insn);
- gen_helper_wfi(cpu_env);
- break;
- case 0x104: /* SFENCE.VM */
- gen_helper_tlb_flush(cpu_env);
- break;
-#endif
default:
gen_exception_illegal(ctx);
break;
}
break;
}
- tcg_temp_free(source1);
- tcg_temp_free(dest);
}
static void decode_RV32_64C0(DisasContext *ctx)
@@ -1056,6 +1008,7 @@ bool decode_insn32(DisasContext *ctx, uint32_t insn);
#include "insn_trans/trans_rva.inc.c"
#include "insn_trans/trans_rvf.inc.c"
#include "insn_trans/trans_rvd.inc.c"
+#include "insn_trans/trans_privileged.inc.c"
static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
{
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 16/35] target/riscv: Convert RV priv insns to decodetree
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- illegal instruction call replaced with return false
- dropped insn argument of trans_foo functions
target/riscv/insn32.decode | 15 +++
.../riscv/insn_trans/trans_privileged.inc.c | 108 ++++++++++++++++++
target/riscv/translate.c | 49 +-------
3 files changed, 124 insertions(+), 48 deletions(-)
create mode 100644 target/riscv/insn_trans/trans_privileged.inc.c
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index b933c349a4..379a5d5438 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -57,6 +57,21 @@
@r2_rm ....... ..... ..... ... ..... ....... %rs1 %rm %rd
@r2 ....... ..... ..... ... ..... ....... %rs1 %rd
+@sfence_vma ....... ..... ..... ... ..... ....... %rs2 %rs1
+@sfence_vm ....... ..... ..... ... ..... ....... %rs1
+
+
+# *** Privileged Instructions ***
+ecall 000000000000 00000 000 00000 1110011
+ebreak 000000000001 00000 000 00000 1110011
+uret 0000000 00010 00000 000 00000 1110011
+sret 0001000 00010 00000 000 00000 1110011
+hret 0010000 00010 00000 000 00000 1110011
+mret 0011000 00010 00000 000 00000 1110011
+wfi 0001000 00101 00000 000 00000 1110011
+sfence_vma 0001001 ..... ..... 000 00000 1110011 @sfence_vma
+sfence_vm 0001000 00100 ..... 000 00000 1110011 @sfence_vm
+
# *** RV32I Base Instruction Set ***
lui .................... ..... 0110111 @u
auipc .................... ..... 0010111 @u
diff --git a/target/riscv/insn_trans/trans_privileged.inc.c b/target/riscv/insn_trans/trans_privileged.inc.c
new file mode 100644
index 0000000000..7a4f0c7846
--- /dev/null
+++ b/target/riscv/insn_trans/trans_privileged.inc.c
@@ -0,0 +1,108 @@
+/*
+ * RISC-V translation routines for the RISC-V privileged instructions.
+ *
+ * Copyright (c) 2016-2017 Sagar Karandikar, sagark@eecs.berkeley.edu
+ * Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
+ * Bastian Koppelmann, kbastian@mail.uni-paderborn.de
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2 or later, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+static bool trans_ecall(DisasContext *ctx, arg_ecall *a)
+{
+ /* always generates U-level ECALL, fixed in do_interrupt handler */
+ generate_exception(ctx, RISCV_EXCP_U_ECALL);
+ tcg_gen_exit_tb(NULL, 0); /* no chaining */
+ ctx->base.is_jmp = DISAS_NORETURN;
+ return true;
+}
+
+static bool trans_ebreak(DisasContext *ctx, arg_ebreak *a)
+{
+ generate_exception(ctx, RISCV_EXCP_BREAKPOINT);
+ tcg_gen_exit_tb(NULL, 0); /* no chaining */
+ ctx->base.is_jmp = DISAS_NORETURN;
+ return true;
+}
+
+static bool trans_uret(DisasContext *ctx, arg_uret *a)
+{
+ return false;
+}
+
+static bool trans_sret(DisasContext *ctx, arg_sret *a)
+{
+#ifndef CONFIG_USER_ONLY
+ tcg_gen_movi_tl(cpu_pc, ctx->base.pc_next);
+ CPURISCVState *env = current_cpu->env_ptr;
+ if (riscv_has_ext(env, RVS)) {
+ gen_helper_sret(cpu_pc, cpu_env, cpu_pc);
+ tcg_gen_exit_tb(NULL, 0); /* no chaining */
+ ctx->base.is_jmp = DISAS_NORETURN;
+ } else {
+ return false;
+ }
+ return true;
+#else
+ return false;
+#endif
+}
+
+static bool trans_hret(DisasContext *ctx, arg_hret *a)
+{
+ return false;
+}
+
+static bool trans_mret(DisasContext *ctx, arg_mret *a)
+{
+#ifndef CONFIG_USER_ONLY
+ tcg_gen_movi_tl(cpu_pc, ctx->base.pc_next);
+ gen_helper_mret(cpu_pc, cpu_env, cpu_pc);
+ tcg_gen_exit_tb(NULL, 0); /* no chaining */
+ ctx->base.is_jmp = DISAS_NORETURN;
+ return true;
+#else
+ return false;
+#endif
+}
+
+static bool trans_wfi(DisasContext *ctx, arg_wfi *a)
+{
+#ifndef CONFIG_USER_ONLY
+ tcg_gen_movi_tl(cpu_pc, ctx->pc_succ_insn);
+ gen_helper_wfi(cpu_env);
+ return true;
+#else
+ return false;
+#endif
+}
+
+static bool trans_sfence_vma(DisasContext *ctx, arg_sfence_vma *a)
+{
+#ifndef CONFIG_USER_ONLY
+ gen_helper_tlb_flush(cpu_env);
+ return true;
+#else
+ return false;
+#endif
+}
+
+static bool trans_sfence_vm(DisasContext *ctx, arg_sfence_vm *a)
+{
+#ifndef CONFIG_USER_ONLY
+ gen_helper_tlb_flush(cpu_env);
+ return true;
+#else
+ return false;
+#endif
+}
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 8e91fa8284..885e48881c 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -690,22 +690,8 @@ static void gen_set_rm(DisasContext *ctx, int rm)
static void gen_system(CPURISCVState *env, DisasContext *ctx, uint32_t opc,
int rd, int rs1, int csr)
{
- TCGv source1, dest;
- source1 = tcg_temp_new();
- dest = tcg_temp_new();
- gen_get_gpr(source1, rs1);
tcg_gen_movi_tl(cpu_pc, ctx->base.pc_next);
-#ifndef CONFIG_USER_ONLY
- /* Extract funct7 value and check whether it matches SFENCE.VMA */
- if ((opc == OPC_RISC_ECALL) && ((csr >> 5) == 9)) {
- /* sfence.vma */
- /* TODO: handle ASID specific fences */
- gen_helper_tlb_flush(cpu_env);
- return;
- }
-#endif
-
switch (opc) {
case OPC_RISC_ECALL:
switch (csr) {
@@ -720,46 +706,12 @@ static void gen_system(CPURISCVState *env, DisasContext *ctx, uint32_t opc,
tcg_gen_exit_tb(NULL, 0); /* no chaining */
ctx->base.is_jmp = DISAS_NORETURN;
break;
-#ifndef CONFIG_USER_ONLY
- case 0x002: /* URET */
- gen_exception_illegal(ctx);
- break;
- case 0x102: /* SRET */
- if (riscv_has_ext(env, RVS)) {
- gen_helper_sret(cpu_pc, cpu_env, cpu_pc);
- tcg_gen_exit_tb(NULL, 0); /* no chaining */
- ctx->base.is_jmp = DISAS_NORETURN;
- } else {
- gen_exception_illegal(ctx);
- }
- break;
- case 0x202: /* HRET */
- gen_exception_illegal(ctx);
- break;
- case 0x302: /* MRET */
- gen_helper_mret(cpu_pc, cpu_env, cpu_pc);
- tcg_gen_exit_tb(NULL, 0); /* no chaining */
- ctx->base.is_jmp = DISAS_NORETURN;
- break;
- case 0x7b2: /* DRET */
- gen_exception_illegal(ctx);
- break;
- case 0x105: /* WFI */
- tcg_gen_movi_tl(cpu_pc, ctx->pc_succ_insn);
- gen_helper_wfi(cpu_env);
- break;
- case 0x104: /* SFENCE.VM */
- gen_helper_tlb_flush(cpu_env);
- break;
-#endif
default:
gen_exception_illegal(ctx);
break;
}
break;
}
- tcg_temp_free(source1);
- tcg_temp_free(dest);
}
static void decode_RV32_64C0(DisasContext *ctx)
@@ -1056,6 +1008,7 @@ bool decode_insn32(DisasContext *ctx, uint32_t insn);
#include "insn_trans/trans_rva.inc.c"
#include "insn_trans/trans_rvf.inc.c"
#include "insn_trans/trans_rvd.inc.c"
+#include "insn_trans/trans_privileged.inc.c"
static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
{
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 17/35] target/riscv: Convert quadrant 0 of RVXC insns to decodetree
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- trans_c_flw_ld and trans_c_fsw_sd now return false for now, as we
cannot use the insn argument anymore for manual decoding. We fix this
in a later patch when rvc is properly split up into insn16-32.decode
and insn16-64.decode.
- special case of trans_c_addi4spn() returns false in this patch
- dropped insn argument of trans_foo functions
target/riscv/Makefile.objs | 9 ++-
target/riscv/insn16.decode | 55 ++++++++++++++++++
target/riscv/insn_trans/trans_rvc.inc.c | 75 +++++++++++++++++++++++++
target/riscv/translate.c | 53 ++++++-----------
4 files changed, 154 insertions(+), 38 deletions(-)
create mode 100644 target/riscv/insn16.decode
create mode 100644 target/riscv/insn_trans/trans_rvc.inc.c
diff --git a/target/riscv/Makefile.objs b/target/riscv/Makefile.objs
index b9b8152cc2..ba661b69e3 100644
--- a/target/riscv/Makefile.objs
+++ b/target/riscv/Makefile.objs
@@ -10,4 +10,11 @@ target/riscv/decode_insn32.inc.c: $(decode32-y) $(DECODETREE)
$(PYTHON) $(DECODETREE) -o $@ --decode decode_insn32 $(decode32-y), \
"GEN", $(TARGET_DIR)$@)
-target/riscv/translate.o: target/riscv/decode_insn32.inc.c
+target/riscv/decode_insn16.inc.c: \
+ $(SRC_PATH)/target/riscv/insn16.decode $(DECODETREE)
+ $(call quiet-command, \
+ $(PYTHON) $(DECODETREE) -o $@ --decode decode_insn16 --insnwidth 16 $<, \
+ "GEN", $(TARGET_DIR)$@)
+
+target/riscv/translate.o: target/riscv/decode_insn32.inc.c \
+ target/riscv/decode_insn16.inc.c
diff --git a/target/riscv/insn16.decode b/target/riscv/insn16.decode
new file mode 100644
index 0000000000..558c0c41f0
--- /dev/null
+++ b/target/riscv/insn16.decode
@@ -0,0 +1,55 @@
+#
+# RISC-V translation routines for the RVXI Base Integer Instruction Set.
+#
+# Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
+# Bastian Koppelmann, kbastian@mail.uni-paderborn.de
+#
+# This program is free software; you can redistribute it and/or modify it
+# under the terms and conditions of the GNU General Public License,
+# version 2 or later, as published by the Free Software Foundation.
+#
+# This program is distributed in the hope it will be useful, but WITHOUT
+# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+# FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+# more details.
+#
+# You should have received a copy of the GNU General Public License along with
+# this program. If not, see <http://www.gnu.org/licenses/>.
+
+# Fields:
+%rd 7:5
+%rs1_3 7:3 !function=ex_rvc_register
+%rs2_3 2:3 !function=ex_rvc_register
+
+# Immediates:
+%nzuimm_ciw 7:4 11:2 5:1 6:1 !function=ex_shift_2
+%uimm_cl_d 5:2 10:3 !function=ex_shift_3
+%uimm_cl_w 5:1 10:3 6:1 !function=ex_shift_2
+
+
+# Argument sets:
+&cl rs1 rd
+&cl_dw uimm rs1 rd
+&ciw nzuimm rd
+&cs rs1 rs2
+&cs_dw uimm rs1 rs2
+
+
+# Formats 16:
+@ciw ... ........ ... .. &ciw nzuimm=%nzuimm_ciw rd=%rs2_3
+@cl_d ... ... ... .. ... .. &cl_dw uimm=%uimm_cl_d rs1=%rs1_3 rd=%rs2_3
+@cl_w ... ... ... .. ... .. &cl_dw uimm=%uimm_cl_w rs1=%rs1_3 rd=%rs2_3
+@cl ... ... ... .. ... .. &cl rs1=%rs1_3 rd=%rs2_3
+@cs ... ... ... .. ... .. &cs rs1=%rs1_3 rs2=%rs2_3
+@cs_d ... ... ... .. ... .. &cs_dw uimm=%uimm_cl_d rs1=%rs1_3 rs2=%rs2_3
+@cs_w ... ... ... .. ... .. &cs_dw uimm=%uimm_cl_w rs1=%rs1_3 rs2=%rs2_3
+
+
+# *** RV64C Standard Extension (Quadrant 0) ***
+c_addi4spn 000 ........ ... 00 @ciw
+c_fld 001 ... ... .. ... 00 @cl_d
+c_lw 010 ... ... .. ... 00 @cl_w
+c_flw_ld 011 --- ... -- ... 00 @cl #Note: Must parse uimm manually
+c_fsd 101 ... ... .. ... 00 @cs_d
+c_sw 110 ... ... .. ... 00 @cs_w
+c_fsw_sd 111 --- ... -- ... 00 @cs #Note: Must parse uimm manually
diff --git a/target/riscv/insn_trans/trans_rvc.inc.c b/target/riscv/insn_trans/trans_rvc.inc.c
new file mode 100644
index 0000000000..93ec8aa30b
--- /dev/null
+++ b/target/riscv/insn_trans/trans_rvc.inc.c
@@ -0,0 +1,75 @@
+/*
+ * RISC-V translation routines for the RVC Compressed Instruction Set.
+ *
+ * Copyright (c) 2016-2017 Sagar Karandikar, sagark@eecs.berkeley.edu
+ * Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
+ * Bastian Koppelmann, kbastian@mail.uni-paderborn.de
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2 or later, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+static bool trans_c_addi4spn(DisasContext *ctx, arg_c_addi4spn *a)
+{
+ if (a->nzuimm == 0) {
+ /* Reserved in ISA */
+ return false;
+ }
+ arg_addi arg = { .rd = a->rd, .rs1 = 2, .imm = a->nzuimm };
+ return trans_addi(ctx, &arg);
+}
+
+static bool trans_c_fld(DisasContext *ctx, arg_c_fld *a)
+{
+ arg_fld arg = { .rd = a->rd, .rs1 = a->rs1, .imm = a->uimm };
+ return trans_fld(ctx, &arg);
+}
+
+static bool trans_c_lw(DisasContext *ctx, arg_c_lw *a)
+{
+ arg_lw arg = { .rd = a->rd, .rs1 = a->rs1, .imm = a->uimm };
+ return trans_lw(ctx, &arg);
+}
+
+static bool trans_c_flw_ld(DisasContext *ctx, arg_c_flw_ld *a)
+{
+#ifdef TARGET_RISCV32
+ /* C.FLW ( RV32FC-only ) */
+ return false;
+#else
+ /* C.LD ( RV64C/RV128C-only ) */
+ return false;
+#endif
+}
+
+static bool trans_c_fsd(DisasContext *ctx, arg_c_fsd *a)
+{
+ arg_fsd arg = { .rs1 = a->rs1, .rs2 = a->rs2, .imm = a->uimm };
+ return trans_fsd(ctx, &arg);
+}
+
+static bool trans_c_sw(DisasContext *ctx, arg_c_sw *a)
+{
+ arg_sw arg = { .rs1 = a->rs1, .rs2 = a->rs2, .imm = a->uimm };
+ return trans_sw(ctx, &arg);
+}
+
+static bool trans_c_fsw_sd(DisasContext *ctx, arg_c_fsw_sd *a)
+{
+#ifdef TARGET_RISCV32
+ /* C.FSW ( RV32FC-only ) */
+ return false;
+#else
+ /* C.SD ( RV64C/RV128C-only ) */
+ return false;
+#endif
+}
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 885e48881c..8bddebb35e 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -721,27 +721,6 @@ static void decode_RV32_64C0(DisasContext *ctx)
uint8_t rs1s = GET_C_RS1S(ctx->opcode);
switch (funct3) {
- case 0:
- /* illegal */
- if (ctx->opcode == 0) {
- gen_exception_illegal(ctx);
- } else {
- /* C.ADDI4SPN -> addi rd', x2, zimm[9:2]*/
- gen_arith_imm(ctx, OPC_RISC_ADDI, rd_rs2, 2,
- GET_C_ADDI4SPN_IMM(ctx->opcode));
- }
- break;
- case 1:
- /* C.FLD -> fld rd', offset[7:3](rs1')*/
- gen_fp_load(ctx, OPC_RISC_FLD, rd_rs2, rs1s,
- GET_C_LD_IMM(ctx->opcode));
- /* C.LQ(RV128) */
- break;
- case 2:
- /* C.LW -> lw rd', offset[6:2](rs1') */
- gen_load(ctx, OPC_RISC_LW, rd_rs2, rs1s,
- GET_C_LW_IMM(ctx->opcode));
- break;
case 3:
#if defined(TARGET_RISCV64)
/* C.LD(RV64/128) -> ld rd', offset[7:3](rs1')*/
@@ -753,21 +732,6 @@ static void decode_RV32_64C0(DisasContext *ctx)
GET_C_LW_IMM(ctx->opcode));
#endif
break;
- case 4:
- /* reserved */
- gen_exception_illegal(ctx);
- break;
- case 5:
- /* C.FSD(RV32/64) -> fsd rs2', offset[7:3](rs1') */
- gen_fp_store(ctx, OPC_RISC_FSD, rs1s, rd_rs2,
- GET_C_LD_IMM(ctx->opcode));
- /* C.SQ (RV128) */
- break;
- case 6:
- /* C.SW -> sw rs2', offset[6:2](rs1')*/
- gen_store(ctx, OPC_RISC_SW, rs1s, rd_rs2,
- GET_C_LW_IMM(ctx->opcode));
- break;
case 7:
#if defined(TARGET_RISCV64)
/* C.SD (RV64/128) -> sd rs2', offset[7:3](rs1')*/
@@ -997,8 +961,15 @@ static void decode_RV32_64C(CPURISCVState *env, DisasContext *ctx)
return imm << amount; \
}
EX_SH(1)
+EX_SH(2)
+EX_SH(3)
EX_SH(12)
+static int ex_rvc_register(int reg)
+{
+ return 8 + reg;
+}
+
bool decode_insn32(DisasContext *ctx, uint32_t insn);
/* Include the auto-generated decoder for 32 bit insn */
#include "decode_insn32.inc.c"
@@ -1010,6 +981,11 @@ bool decode_insn32(DisasContext *ctx, uint32_t insn);
#include "insn_trans/trans_rvd.inc.c"
#include "insn_trans/trans_privileged.inc.c"
+bool decode_insn16(DisasContext *ctx, uint16_t insn);
+/* auto-generated decoder*/
+#include "decode_insn16.inc.c"
+#include "insn_trans/trans_rvc.inc.c"
+
static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
{
int rs1, rd;
@@ -1043,7 +1019,10 @@ static void decode_opc(DisasContext *ctx)
gen_exception_illegal(ctx);
} else {
ctx->pc_succ_insn = ctx->base.pc_next + 2;
- decode_RV32_64C(ctx->env, ctx);
+ if (!decode_insn16(ctx, ctx->opcode)) {
+ /* fall back to old decoder */
+ decode_RV32_64C(ctx->env, ctx);
+ }
}
} else {
ctx->pc_succ_insn = ctx->base.pc_next + 4;
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 17/35] target/riscv: Convert quadrant 0 of RVXC insns to decodetree
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- trans_c_flw_ld and trans_c_fsw_sd now return false for now, as we
cannot use the insn argument anymore for manual decoding. We fix this
in a later patch when rvc is properly split up into insn16-32.decode
and insn16-64.decode.
- special case of trans_c_addi4spn() returns false in this patch
- dropped insn argument of trans_foo functions
target/riscv/Makefile.objs | 9 ++-
target/riscv/insn16.decode | 55 ++++++++++++++++++
target/riscv/insn_trans/trans_rvc.inc.c | 75 +++++++++++++++++++++++++
target/riscv/translate.c | 53 ++++++-----------
4 files changed, 154 insertions(+), 38 deletions(-)
create mode 100644 target/riscv/insn16.decode
create mode 100644 target/riscv/insn_trans/trans_rvc.inc.c
diff --git a/target/riscv/Makefile.objs b/target/riscv/Makefile.objs
index b9b8152cc2..ba661b69e3 100644
--- a/target/riscv/Makefile.objs
+++ b/target/riscv/Makefile.objs
@@ -10,4 +10,11 @@ target/riscv/decode_insn32.inc.c: $(decode32-y) $(DECODETREE)
$(PYTHON) $(DECODETREE) -o $@ --decode decode_insn32 $(decode32-y), \
"GEN", $(TARGET_DIR)$@)
-target/riscv/translate.o: target/riscv/decode_insn32.inc.c
+target/riscv/decode_insn16.inc.c: \
+ $(SRC_PATH)/target/riscv/insn16.decode $(DECODETREE)
+ $(call quiet-command, \
+ $(PYTHON) $(DECODETREE) -o $@ --decode decode_insn16 --insnwidth 16 $<, \
+ "GEN", $(TARGET_DIR)$@)
+
+target/riscv/translate.o: target/riscv/decode_insn32.inc.c \
+ target/riscv/decode_insn16.inc.c
diff --git a/target/riscv/insn16.decode b/target/riscv/insn16.decode
new file mode 100644
index 0000000000..558c0c41f0
--- /dev/null
+++ b/target/riscv/insn16.decode
@@ -0,0 +1,55 @@
+#
+# RISC-V translation routines for the RVXI Base Integer Instruction Set.
+#
+# Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
+# Bastian Koppelmann, kbastian@mail.uni-paderborn.de
+#
+# This program is free software; you can redistribute it and/or modify it
+# under the terms and conditions of the GNU General Public License,
+# version 2 or later, as published by the Free Software Foundation.
+#
+# This program is distributed in the hope it will be useful, but WITHOUT
+# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+# FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+# more details.
+#
+# You should have received a copy of the GNU General Public License along with
+# this program. If not, see <http://www.gnu.org/licenses/>.
+
+# Fields:
+%rd 7:5
+%rs1_3 7:3 !function=ex_rvc_register
+%rs2_3 2:3 !function=ex_rvc_register
+
+# Immediates:
+%nzuimm_ciw 7:4 11:2 5:1 6:1 !function=ex_shift_2
+%uimm_cl_d 5:2 10:3 !function=ex_shift_3
+%uimm_cl_w 5:1 10:3 6:1 !function=ex_shift_2
+
+
+# Argument sets:
+&cl rs1 rd
+&cl_dw uimm rs1 rd
+&ciw nzuimm rd
+&cs rs1 rs2
+&cs_dw uimm rs1 rs2
+
+
+# Formats 16:
+@ciw ... ........ ... .. &ciw nzuimm=%nzuimm_ciw rd=%rs2_3
+@cl_d ... ... ... .. ... .. &cl_dw uimm=%uimm_cl_d rs1=%rs1_3 rd=%rs2_3
+@cl_w ... ... ... .. ... .. &cl_dw uimm=%uimm_cl_w rs1=%rs1_3 rd=%rs2_3
+@cl ... ... ... .. ... .. &cl rs1=%rs1_3 rd=%rs2_3
+@cs ... ... ... .. ... .. &cs rs1=%rs1_3 rs2=%rs2_3
+@cs_d ... ... ... .. ... .. &cs_dw uimm=%uimm_cl_d rs1=%rs1_3 rs2=%rs2_3
+@cs_w ... ... ... .. ... .. &cs_dw uimm=%uimm_cl_w rs1=%rs1_3 rs2=%rs2_3
+
+
+# *** RV64C Standard Extension (Quadrant 0) ***
+c_addi4spn 000 ........ ... 00 @ciw
+c_fld 001 ... ... .. ... 00 @cl_d
+c_lw 010 ... ... .. ... 00 @cl_w
+c_flw_ld 011 --- ... -- ... 00 @cl #Note: Must parse uimm manually
+c_fsd 101 ... ... .. ... 00 @cs_d
+c_sw 110 ... ... .. ... 00 @cs_w
+c_fsw_sd 111 --- ... -- ... 00 @cs #Note: Must parse uimm manually
diff --git a/target/riscv/insn_trans/trans_rvc.inc.c b/target/riscv/insn_trans/trans_rvc.inc.c
new file mode 100644
index 0000000000..93ec8aa30b
--- /dev/null
+++ b/target/riscv/insn_trans/trans_rvc.inc.c
@@ -0,0 +1,75 @@
+/*
+ * RISC-V translation routines for the RVC Compressed Instruction Set.
+ *
+ * Copyright (c) 2016-2017 Sagar Karandikar, sagark@eecs.berkeley.edu
+ * Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
+ * Bastian Koppelmann, kbastian@mail.uni-paderborn.de
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2 or later, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+static bool trans_c_addi4spn(DisasContext *ctx, arg_c_addi4spn *a)
+{
+ if (a->nzuimm == 0) {
+ /* Reserved in ISA */
+ return false;
+ }
+ arg_addi arg = { .rd = a->rd, .rs1 = 2, .imm = a->nzuimm };
+ return trans_addi(ctx, &arg);
+}
+
+static bool trans_c_fld(DisasContext *ctx, arg_c_fld *a)
+{
+ arg_fld arg = { .rd = a->rd, .rs1 = a->rs1, .imm = a->uimm };
+ return trans_fld(ctx, &arg);
+}
+
+static bool trans_c_lw(DisasContext *ctx, arg_c_lw *a)
+{
+ arg_lw arg = { .rd = a->rd, .rs1 = a->rs1, .imm = a->uimm };
+ return trans_lw(ctx, &arg);
+}
+
+static bool trans_c_flw_ld(DisasContext *ctx, arg_c_flw_ld *a)
+{
+#ifdef TARGET_RISCV32
+ /* C.FLW ( RV32FC-only ) */
+ return false;
+#else
+ /* C.LD ( RV64C/RV128C-only ) */
+ return false;
+#endif
+}
+
+static bool trans_c_fsd(DisasContext *ctx, arg_c_fsd *a)
+{
+ arg_fsd arg = { .rs1 = a->rs1, .rs2 = a->rs2, .imm = a->uimm };
+ return trans_fsd(ctx, &arg);
+}
+
+static bool trans_c_sw(DisasContext *ctx, arg_c_sw *a)
+{
+ arg_sw arg = { .rs1 = a->rs1, .rs2 = a->rs2, .imm = a->uimm };
+ return trans_sw(ctx, &arg);
+}
+
+static bool trans_c_fsw_sd(DisasContext *ctx, arg_c_fsw_sd *a)
+{
+#ifdef TARGET_RISCV32
+ /* C.FSW ( RV32FC-only ) */
+ return false;
+#else
+ /* C.SD ( RV64C/RV128C-only ) */
+ return false;
+#endif
+}
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 885e48881c..8bddebb35e 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -721,27 +721,6 @@ static void decode_RV32_64C0(DisasContext *ctx)
uint8_t rs1s = GET_C_RS1S(ctx->opcode);
switch (funct3) {
- case 0:
- /* illegal */
- if (ctx->opcode == 0) {
- gen_exception_illegal(ctx);
- } else {
- /* C.ADDI4SPN -> addi rd', x2, zimm[9:2]*/
- gen_arith_imm(ctx, OPC_RISC_ADDI, rd_rs2, 2,
- GET_C_ADDI4SPN_IMM(ctx->opcode));
- }
- break;
- case 1:
- /* C.FLD -> fld rd', offset[7:3](rs1')*/
- gen_fp_load(ctx, OPC_RISC_FLD, rd_rs2, rs1s,
- GET_C_LD_IMM(ctx->opcode));
- /* C.LQ(RV128) */
- break;
- case 2:
- /* C.LW -> lw rd', offset[6:2](rs1') */
- gen_load(ctx, OPC_RISC_LW, rd_rs2, rs1s,
- GET_C_LW_IMM(ctx->opcode));
- break;
case 3:
#if defined(TARGET_RISCV64)
/* C.LD(RV64/128) -> ld rd', offset[7:3](rs1')*/
@@ -753,21 +732,6 @@ static void decode_RV32_64C0(DisasContext *ctx)
GET_C_LW_IMM(ctx->opcode));
#endif
break;
- case 4:
- /* reserved */
- gen_exception_illegal(ctx);
- break;
- case 5:
- /* C.FSD(RV32/64) -> fsd rs2', offset[7:3](rs1') */
- gen_fp_store(ctx, OPC_RISC_FSD, rs1s, rd_rs2,
- GET_C_LD_IMM(ctx->opcode));
- /* C.SQ (RV128) */
- break;
- case 6:
- /* C.SW -> sw rs2', offset[6:2](rs1')*/
- gen_store(ctx, OPC_RISC_SW, rs1s, rd_rs2,
- GET_C_LW_IMM(ctx->opcode));
- break;
case 7:
#if defined(TARGET_RISCV64)
/* C.SD (RV64/128) -> sd rs2', offset[7:3](rs1')*/
@@ -997,8 +961,15 @@ static void decode_RV32_64C(CPURISCVState *env, DisasContext *ctx)
return imm << amount; \
}
EX_SH(1)
+EX_SH(2)
+EX_SH(3)
EX_SH(12)
+static int ex_rvc_register(int reg)
+{
+ return 8 + reg;
+}
+
bool decode_insn32(DisasContext *ctx, uint32_t insn);
/* Include the auto-generated decoder for 32 bit insn */
#include "decode_insn32.inc.c"
@@ -1010,6 +981,11 @@ bool decode_insn32(DisasContext *ctx, uint32_t insn);
#include "insn_trans/trans_rvd.inc.c"
#include "insn_trans/trans_privileged.inc.c"
+bool decode_insn16(DisasContext *ctx, uint16_t insn);
+/* auto-generated decoder*/
+#include "decode_insn16.inc.c"
+#include "insn_trans/trans_rvc.inc.c"
+
static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
{
int rs1, rd;
@@ -1043,7 +1019,10 @@ static void decode_opc(DisasContext *ctx)
gen_exception_illegal(ctx);
} else {
ctx->pc_succ_insn = ctx->base.pc_next + 2;
- decode_RV32_64C(ctx->env, ctx);
+ if (!decode_insn16(ctx, ctx->opcode)) {
+ /* fall back to old decoder */
+ decode_RV32_64C(ctx->env, ctx);
+ }
}
} else {
ctx->pc_succ_insn = ctx->base.pc_next + 4;
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 18/35] target/riscv: Convert quadrant 1 of RVXC insns to decodetree
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- simplified trans_c_srai by Richard's suggestion
- dropped insn argument of trans_foo functions
target/riscv/insn16.decode | 43 +++++++
target/riscv/insn_trans/trans_rvc.inc.c | 151 ++++++++++++++++++++++++
target/riscv/translate.c | 118 +-----------------
3 files changed, 195 insertions(+), 117 deletions(-)
diff --git a/target/riscv/insn16.decode b/target/riscv/insn16.decode
index 558c0c41f0..29dade0fa1 100644
--- a/target/riscv/insn16.decode
+++ b/target/riscv/insn16.decode
@@ -22,28 +22,53 @@
%rs2_3 2:3 !function=ex_rvc_register
# Immediates:
+%imm_ci 12:s1 2:5
%nzuimm_ciw 7:4 11:2 5:1 6:1 !function=ex_shift_2
%uimm_cl_d 5:2 10:3 !function=ex_shift_3
%uimm_cl_w 5:1 10:3 6:1 !function=ex_shift_2
+%imm_cb 12:s1 5:2 2:1 10:2 3:2 !function=ex_shift_1
+%imm_cj 12:s1 8:1 9:2 6:1 7:1 2:1 11:1 3:3 !function=ex_shift_1
+
+%nzuimm_6bit 12:1 2:5
+
+%imm_addi16sp 12:s1 3:2 5:1 2:1 6:1 !function=ex_shift_4
+%imm_lui 12:s1 2:5 !function=ex_shift_12
+
# Argument sets:
&cl rs1 rd
&cl_dw uimm rs1 rd
+&ci imm rd
&ciw nzuimm rd
&cs rs1 rs2
&cs_dw uimm rs1 rs2
+&cb imm rs1
+&cr rd rs2
+&c_j imm
+&c_shift shamt rd
+
+&c_addi16sp_lui imm_lui imm_addi16sp rd
# Formats 16:
+@ci ... . ..... ..... .. &ci imm=%imm_ci %rd
@ciw ... ........ ... .. &ciw nzuimm=%nzuimm_ciw rd=%rs2_3
@cl_d ... ... ... .. ... .. &cl_dw uimm=%uimm_cl_d rs1=%rs1_3 rd=%rs2_3
@cl_w ... ... ... .. ... .. &cl_dw uimm=%uimm_cl_w rs1=%rs1_3 rd=%rs2_3
@cl ... ... ... .. ... .. &cl rs1=%rs1_3 rd=%rs2_3
@cs ... ... ... .. ... .. &cs rs1=%rs1_3 rs2=%rs2_3
+@cs_2 ... ... ... .. ... .. &cr rd=%rs1_3 rs2=%rs2_3
@cs_d ... ... ... .. ... .. &cs_dw uimm=%uimm_cl_d rs1=%rs1_3 rs2=%rs2_3
@cs_w ... ... ... .. ... .. &cs_dw uimm=%uimm_cl_w rs1=%rs1_3 rs2=%rs2_3
+@cb ... ... ... .. ... .. &cb imm=%imm_cb rs1=%rs1_3
+@cj ... ........... .. &c_j imm=%imm_cj
+@c_addi16sp_lui ... . ..... ..... .. &c_addi16sp_lui %imm_lui %imm_addi16sp %rd
+
+@c_shift ... . .. ... ..... .. &c_shift rd=%rs1_3 shamt=%nzuimm_6bit
+
+@c_andi ... . .. ... ..... .. &ci imm=%imm_ci rd=%rs1_3
# *** RV64C Standard Extension (Quadrant 0) ***
c_addi4spn 000 ........ ... 00 @ciw
@@ -53,3 +78,21 @@ c_flw_ld 011 --- ... -- ... 00 @cl #Note: Must parse uimm manually
c_fsd 101 ... ... .. ... 00 @cs_d
c_sw 110 ... ... .. ... 00 @cs_w
c_fsw_sd 111 --- ... -- ... 00 @cs #Note: Must parse uimm manually
+
+# *** RV64C Standard Extension (Quadrant 1) ***
+c_addi 000 . ..... ..... 01 @ci
+c_jal_addiw 001 . ..... ..... 01 @ci #Note: parse rd and/or imm manually
+c_li 010 . ..... ..... 01 @ci
+c_addi16sp_lui 011 . ..... ..... 01 @c_addi16sp_lui # shares opc with C.LUI
+c_srli 100 . 00 ... ..... 01 @c_shift
+c_srai 100 . 01 ... ..... 01 @c_shift
+c_andi 100 . 10 ... ..... 01 @c_andi
+c_sub 100 0 11 ... 00 ... 01 @cs_2
+c_xor 100 0 11 ... 01 ... 01 @cs_2
+c_or 100 0 11 ... 10 ... 01 @cs_2
+c_and 100 0 11 ... 11 ... 01 @cs_2
+c_subw 100 1 11 ... 00 ... 01 @cs_2
+c_addw 100 1 11 ... 01 ... 01 @cs_2
+c_j 101 ........... 01 @cj
+c_beqz 110 ... ... ..... 01 @cb
+c_bnez 111 ... ... ..... 01 @cb
diff --git a/target/riscv/insn_trans/trans_rvc.inc.c b/target/riscv/insn_trans/trans_rvc.inc.c
index 93ec8aa30b..b06c435c98 100644
--- a/target/riscv/insn_trans/trans_rvc.inc.c
+++ b/target/riscv/insn_trans/trans_rvc.inc.c
@@ -73,3 +73,154 @@ static bool trans_c_fsw_sd(DisasContext *ctx, arg_c_fsw_sd *a)
return false;
#endif
}
+
+static bool trans_c_addi(DisasContext *ctx, arg_c_addi *a)
+{
+ if (a->imm == 0) {
+ /* Hint: insn is valid but does not affect state */
+ return true;
+ }
+ arg_addi arg = { .rd = a->rd, .rs1 = a->rd, .imm = a->imm };
+ return trans_addi(ctx, &arg);
+}
+
+static bool trans_c_jal_addiw(DisasContext *ctx, arg_c_jal_addiw *a)
+{
+#ifdef TARGET_RISCV32
+ /* C.JAL */
+ arg_jal arg = { .rd = 1, .imm = a->imm };
+ return trans_jal(ctx, &arg);
+#else
+ /* C.ADDIW */
+ arg_addiw arg = { .rd = a->rd, .rs1 = a->rd, .imm = a->imm };
+ return trans_addiw(ctx, &arg);
+#endif
+}
+
+static bool trans_c_li(DisasContext *ctx, arg_c_li *a)
+{
+ if (a->rd == 0) {
+ /* Hint: insn is valid but does not affect state */
+ return true;
+ }
+ arg_addi arg = { .rd = a->rd, .rs1 = 0, .imm = a->imm };
+ return trans_addi(ctx, &arg);
+}
+
+static bool trans_c_addi16sp_lui(DisasContext *ctx, arg_c_addi16sp_lui *a)
+{
+ if (a->rd == 2) {
+ /* C.ADDI16SP */
+ arg_addi arg = { .rd = 2, .rs1 = 2, .imm = a->imm_addi16sp };
+ return trans_addi(ctx, &arg);
+ } else if (a->imm_lui != 0) {
+ /* C.LUI */
+ if (a->rd == 0) {
+ /* Hint: insn is valid but does not affect state */
+ return true;
+ }
+ arg_lui arg = { .rd = a->rd, .imm = a->imm_lui };
+ return trans_lui(ctx, &arg);
+ }
+ return false;
+}
+
+static bool trans_c_srli(DisasContext *ctx, arg_c_srli *a)
+{
+ int shamt = a->shamt;
+ if (shamt == 0) {
+ /* For RV128 a shamt of 0 means a shift by 64 */
+ shamt = 64;
+ }
+ /* Ensure, that shamt[5] is zero for RV32 */
+ if (shamt >= TARGET_LONG_BITS) {
+ return false;
+ }
+
+ arg_srli arg = { .rd = a->rd, .rs1 = a->rd, .shamt = a->shamt };
+ return trans_srli(ctx, &arg);
+}
+
+static bool trans_c_srai(DisasContext *ctx, arg_c_srai *a)
+{
+ int shamt = a->shamt;
+ if (shamt == 0) {
+ /* For RV128 a shamt of 0 means a shift by 64 */
+ shamt = 64;
+ }
+ /* Ensure, that shamt[5] is zero for RV32 */
+ if (shamt >= TARGET_LONG_BITS) {
+ return false;
+ }
+
+ arg_srai arg = { .rd = a->rd, .rs1 = a->rd, .shamt = a->shamt };
+ return trans_srai(ctx, &arg);
+}
+
+static bool trans_c_andi(DisasContext *ctx, arg_c_andi *a)
+{
+ arg_andi arg = { .rd = a->rd, .rs1 = a->rd, .imm = a->imm };
+ return trans_andi(ctx, &arg);
+}
+
+static bool trans_c_sub(DisasContext *ctx, arg_c_sub *a)
+{
+ arg_sub arg = { .rd = a->rd, .rs1 = a->rd, .rs2 = a->rs2 };
+ return trans_sub(ctx, &arg);
+}
+
+static bool trans_c_xor(DisasContext *ctx, arg_c_xor *a)
+{
+ arg_xor arg = { .rd = a->rd, .rs1 = a->rd, .rs2 = a->rs2 };
+ return trans_xor(ctx, &arg);
+}
+
+static bool trans_c_or(DisasContext *ctx, arg_c_or *a)
+{
+ arg_or arg = { .rd = a->rd, .rs1 = a->rd, .rs2 = a->rs2 };
+ return trans_or(ctx, &arg);
+}
+
+static bool trans_c_and(DisasContext *ctx, arg_c_and *a)
+{
+ arg_and arg = { .rd = a->rd, .rs1 = a->rd, .rs2 = a->rs2 };
+ return trans_and(ctx, &arg);
+}
+
+static bool trans_c_subw(DisasContext *ctx, arg_c_subw *a)
+{
+#ifdef TARGET_RISCV64
+ arg_subw arg = { .rd = a->rd, .rs1 = a->rd, .rs2 = a->rs2 };
+ return trans_subw(ctx, &arg);
+#else
+ return false;
+#endif
+}
+
+static bool trans_c_addw(DisasContext *ctx, arg_c_addw *a)
+{
+#ifdef TARGET_RISCV64
+ arg_addw arg = { .rd = a->rd, .rs1 = a->rd, .rs2 = a->rs2 };
+ return trans_addw(ctx, &arg);
+#else
+ return false;
+#endif
+}
+
+static bool trans_c_j(DisasContext *ctx, arg_c_j *a)
+{
+ arg_jal arg = { .rd = 0, .imm = a->imm };
+ return trans_jal(ctx, &arg);
+}
+
+static bool trans_c_beqz(DisasContext *ctx, arg_c_beqz *a)
+{
+ arg_beq arg = { .rs1 = a->rs1, .rs2 = 0, .imm = a->imm };
+ return trans_beq(ctx, &arg);
+}
+
+static bool trans_c_bnez(DisasContext *ctx, arg_c_bnez *a)
+{
+ arg_bne arg = { .rs1 = a->rs1, .rs2 = 0, .imm = a->imm };
+ return trans_bne(ctx, &arg);
+}
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 8bddebb35e..dc4d2c42ba 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -746,120 +746,6 @@ static void decode_RV32_64C0(DisasContext *ctx)
}
}
-static void decode_RV32_64C1(CPURISCVState *env, DisasContext *ctx)
-{
- uint8_t funct3 = extract32(ctx->opcode, 13, 3);
- uint8_t rd_rs1 = GET_C_RS1(ctx->opcode);
- uint8_t rs1s, rs2s;
- uint8_t funct2;
-
- switch (funct3) {
- case 0:
- /* C.ADDI -> addi rd, rd, nzimm[5:0] */
- gen_arith_imm(ctx, OPC_RISC_ADDI, rd_rs1, rd_rs1,
- GET_C_IMM(ctx->opcode));
- break;
- case 1:
-#if defined(TARGET_RISCV64)
- /* C.ADDIW (RV64/128) -> addiw rd, rd, imm[5:0]*/
- gen_arith_imm(ctx, OPC_RISC_ADDIW, rd_rs1, rd_rs1,
- GET_C_IMM(ctx->opcode));
-#else
- /* C.JAL(RV32) -> jal x1, offset[11:1] */
- gen_jal(env, ctx, 1, GET_C_J_IMM(ctx->opcode));
-#endif
- break;
- case 2:
- /* C.LI -> addi rd, x0, imm[5:0]*/
- gen_arith_imm(ctx, OPC_RISC_ADDI, rd_rs1, 0, GET_C_IMM(ctx->opcode));
- break;
- case 3:
- if (rd_rs1 == 2) {
- /* C.ADDI16SP -> addi x2, x2, nzimm[9:4]*/
- gen_arith_imm(ctx, OPC_RISC_ADDI, 2, 2,
- GET_C_ADDI16SP_IMM(ctx->opcode));
- } else if (rd_rs1 != 0) {
- /* C.LUI (rs1/rd =/= {0,2}) -> lui rd, nzimm[17:12]*/
- tcg_gen_movi_tl(cpu_gpr[rd_rs1],
- GET_C_IMM(ctx->opcode) << 12);
- }
- break;
- case 4:
- funct2 = extract32(ctx->opcode, 10, 2);
- rs1s = GET_C_RS1S(ctx->opcode);
- switch (funct2) {
- case 0: /* C.SRLI(RV32) -> srli rd', rd', shamt[5:0] */
- gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_I, rs1s, rs1s,
- GET_C_ZIMM(ctx->opcode));
- /* C.SRLI64(RV128) */
- break;
- case 1:
- /* C.SRAI -> srai rd', rd', shamt[5:0]*/
- gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_I, rs1s, rs1s,
- GET_C_ZIMM(ctx->opcode) | 0x400);
- /* C.SRAI64(RV128) */
- break;
- case 2:
- /* C.ANDI -> andi rd', rd', imm[5:0]*/
- gen_arith_imm(ctx, OPC_RISC_ANDI, rs1s, rs1s,
- GET_C_IMM(ctx->opcode));
- break;
- case 3:
- funct2 = extract32(ctx->opcode, 5, 2);
- rs2s = GET_C_RS2S(ctx->opcode);
- switch (funct2) {
- case 0:
- /* C.SUB -> sub rd', rd', rs2' */
- if (extract32(ctx->opcode, 12, 1) == 0) {
- gen_arith(ctx, OPC_RISC_SUB, rs1s, rs1s, rs2s);
- }
-#if defined(TARGET_RISCV64)
- else {
- gen_arith(ctx, OPC_RISC_SUBW, rs1s, rs1s, rs2s);
- }
-#endif
- break;
- case 1:
- /* C.XOR -> xor rs1', rs1', rs2' */
- if (extract32(ctx->opcode, 12, 1) == 0) {
- gen_arith(ctx, OPC_RISC_XOR, rs1s, rs1s, rs2s);
- }
-#if defined(TARGET_RISCV64)
- else {
- /* C.ADDW (RV64/128) */
- gen_arith(ctx, OPC_RISC_ADDW, rs1s, rs1s, rs2s);
- }
-#endif
- break;
- case 2:
- /* C.OR -> or rs1', rs1', rs2' */
- gen_arith(ctx, OPC_RISC_OR, rs1s, rs1s, rs2s);
- break;
- case 3:
- /* C.AND -> and rs1', rs1', rs2' */
- gen_arith(ctx, OPC_RISC_AND, rs1s, rs1s, rs2s);
- break;
- }
- break;
- }
- break;
- case 5:
- /* C.J -> jal x0, offset[11:1]*/
- gen_jal(env, ctx, 0, GET_C_J_IMM(ctx->opcode));
- break;
- case 6:
- /* C.BEQZ -> beq rs1', x0, offset[8:1]*/
- rs1s = GET_C_RS1S(ctx->opcode);
- gen_branch(env, ctx, OPC_RISC_BEQ, rs1s, 0, GET_C_B_IMM(ctx->opcode));
- break;
- case 7:
- /* C.BNEZ -> bne rs1', x0, offset[8:1]*/
- rs1s = GET_C_RS1S(ctx->opcode);
- gen_branch(env, ctx, OPC_RISC_BNE, rs1s, 0, GET_C_B_IMM(ctx->opcode));
- break;
- }
-}
-
static void decode_RV32_64C2(CPURISCVState *env, DisasContext *ctx)
{
uint8_t rd, rs2;
@@ -946,9 +832,6 @@ static void decode_RV32_64C(CPURISCVState *env, DisasContext *ctx)
case 0:
decode_RV32_64C0(ctx);
break;
- case 1:
- decode_RV32_64C1(env, ctx);
- break;
case 2:
decode_RV32_64C2(env, ctx);
break;
@@ -963,6 +846,7 @@ static void decode_RV32_64C(CPURISCVState *env, DisasContext *ctx)
EX_SH(1)
EX_SH(2)
EX_SH(3)
+EX_SH(4)
EX_SH(12)
static int ex_rvc_register(int reg)
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 18/35] target/riscv: Convert quadrant 1 of RVXC insns to decodetree
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- simplified trans_c_srai by Richard's suggestion
- dropped insn argument of trans_foo functions
target/riscv/insn16.decode | 43 +++++++
target/riscv/insn_trans/trans_rvc.inc.c | 151 ++++++++++++++++++++++++
target/riscv/translate.c | 118 +-----------------
3 files changed, 195 insertions(+), 117 deletions(-)
diff --git a/target/riscv/insn16.decode b/target/riscv/insn16.decode
index 558c0c41f0..29dade0fa1 100644
--- a/target/riscv/insn16.decode
+++ b/target/riscv/insn16.decode
@@ -22,28 +22,53 @@
%rs2_3 2:3 !function=ex_rvc_register
# Immediates:
+%imm_ci 12:s1 2:5
%nzuimm_ciw 7:4 11:2 5:1 6:1 !function=ex_shift_2
%uimm_cl_d 5:2 10:3 !function=ex_shift_3
%uimm_cl_w 5:1 10:3 6:1 !function=ex_shift_2
+%imm_cb 12:s1 5:2 2:1 10:2 3:2 !function=ex_shift_1
+%imm_cj 12:s1 8:1 9:2 6:1 7:1 2:1 11:1 3:3 !function=ex_shift_1
+
+%nzuimm_6bit 12:1 2:5
+
+%imm_addi16sp 12:s1 3:2 5:1 2:1 6:1 !function=ex_shift_4
+%imm_lui 12:s1 2:5 !function=ex_shift_12
+
# Argument sets:
&cl rs1 rd
&cl_dw uimm rs1 rd
+&ci imm rd
&ciw nzuimm rd
&cs rs1 rs2
&cs_dw uimm rs1 rs2
+&cb imm rs1
+&cr rd rs2
+&c_j imm
+&c_shift shamt rd
+
+&c_addi16sp_lui imm_lui imm_addi16sp rd
# Formats 16:
+@ci ... . ..... ..... .. &ci imm=%imm_ci %rd
@ciw ... ........ ... .. &ciw nzuimm=%nzuimm_ciw rd=%rs2_3
@cl_d ... ... ... .. ... .. &cl_dw uimm=%uimm_cl_d rs1=%rs1_3 rd=%rs2_3
@cl_w ... ... ... .. ... .. &cl_dw uimm=%uimm_cl_w rs1=%rs1_3 rd=%rs2_3
@cl ... ... ... .. ... .. &cl rs1=%rs1_3 rd=%rs2_3
@cs ... ... ... .. ... .. &cs rs1=%rs1_3 rs2=%rs2_3
+@cs_2 ... ... ... .. ... .. &cr rd=%rs1_3 rs2=%rs2_3
@cs_d ... ... ... .. ... .. &cs_dw uimm=%uimm_cl_d rs1=%rs1_3 rs2=%rs2_3
@cs_w ... ... ... .. ... .. &cs_dw uimm=%uimm_cl_w rs1=%rs1_3 rs2=%rs2_3
+@cb ... ... ... .. ... .. &cb imm=%imm_cb rs1=%rs1_3
+@cj ... ........... .. &c_j imm=%imm_cj
+@c_addi16sp_lui ... . ..... ..... .. &c_addi16sp_lui %imm_lui %imm_addi16sp %rd
+
+@c_shift ... . .. ... ..... .. &c_shift rd=%rs1_3 shamt=%nzuimm_6bit
+
+@c_andi ... . .. ... ..... .. &ci imm=%imm_ci rd=%rs1_3
# *** RV64C Standard Extension (Quadrant 0) ***
c_addi4spn 000 ........ ... 00 @ciw
@@ -53,3 +78,21 @@ c_flw_ld 011 --- ... -- ... 00 @cl #Note: Must parse uimm manually
c_fsd 101 ... ... .. ... 00 @cs_d
c_sw 110 ... ... .. ... 00 @cs_w
c_fsw_sd 111 --- ... -- ... 00 @cs #Note: Must parse uimm manually
+
+# *** RV64C Standard Extension (Quadrant 1) ***
+c_addi 000 . ..... ..... 01 @ci
+c_jal_addiw 001 . ..... ..... 01 @ci #Note: parse rd and/or imm manually
+c_li 010 . ..... ..... 01 @ci
+c_addi16sp_lui 011 . ..... ..... 01 @c_addi16sp_lui # shares opc with C.LUI
+c_srli 100 . 00 ... ..... 01 @c_shift
+c_srai 100 . 01 ... ..... 01 @c_shift
+c_andi 100 . 10 ... ..... 01 @c_andi
+c_sub 100 0 11 ... 00 ... 01 @cs_2
+c_xor 100 0 11 ... 01 ... 01 @cs_2
+c_or 100 0 11 ... 10 ... 01 @cs_2
+c_and 100 0 11 ... 11 ... 01 @cs_2
+c_subw 100 1 11 ... 00 ... 01 @cs_2
+c_addw 100 1 11 ... 01 ... 01 @cs_2
+c_j 101 ........... 01 @cj
+c_beqz 110 ... ... ..... 01 @cb
+c_bnez 111 ... ... ..... 01 @cb
diff --git a/target/riscv/insn_trans/trans_rvc.inc.c b/target/riscv/insn_trans/trans_rvc.inc.c
index 93ec8aa30b..b06c435c98 100644
--- a/target/riscv/insn_trans/trans_rvc.inc.c
+++ b/target/riscv/insn_trans/trans_rvc.inc.c
@@ -73,3 +73,154 @@ static bool trans_c_fsw_sd(DisasContext *ctx, arg_c_fsw_sd *a)
return false;
#endif
}
+
+static bool trans_c_addi(DisasContext *ctx, arg_c_addi *a)
+{
+ if (a->imm == 0) {
+ /* Hint: insn is valid but does not affect state */
+ return true;
+ }
+ arg_addi arg = { .rd = a->rd, .rs1 = a->rd, .imm = a->imm };
+ return trans_addi(ctx, &arg);
+}
+
+static bool trans_c_jal_addiw(DisasContext *ctx, arg_c_jal_addiw *a)
+{
+#ifdef TARGET_RISCV32
+ /* C.JAL */
+ arg_jal arg = { .rd = 1, .imm = a->imm };
+ return trans_jal(ctx, &arg);
+#else
+ /* C.ADDIW */
+ arg_addiw arg = { .rd = a->rd, .rs1 = a->rd, .imm = a->imm };
+ return trans_addiw(ctx, &arg);
+#endif
+}
+
+static bool trans_c_li(DisasContext *ctx, arg_c_li *a)
+{
+ if (a->rd == 0) {
+ /* Hint: insn is valid but does not affect state */
+ return true;
+ }
+ arg_addi arg = { .rd = a->rd, .rs1 = 0, .imm = a->imm };
+ return trans_addi(ctx, &arg);
+}
+
+static bool trans_c_addi16sp_lui(DisasContext *ctx, arg_c_addi16sp_lui *a)
+{
+ if (a->rd == 2) {
+ /* C.ADDI16SP */
+ arg_addi arg = { .rd = 2, .rs1 = 2, .imm = a->imm_addi16sp };
+ return trans_addi(ctx, &arg);
+ } else if (a->imm_lui != 0) {
+ /* C.LUI */
+ if (a->rd == 0) {
+ /* Hint: insn is valid but does not affect state */
+ return true;
+ }
+ arg_lui arg = { .rd = a->rd, .imm = a->imm_lui };
+ return trans_lui(ctx, &arg);
+ }
+ return false;
+}
+
+static bool trans_c_srli(DisasContext *ctx, arg_c_srli *a)
+{
+ int shamt = a->shamt;
+ if (shamt == 0) {
+ /* For RV128 a shamt of 0 means a shift by 64 */
+ shamt = 64;
+ }
+ /* Ensure, that shamt[5] is zero for RV32 */
+ if (shamt >= TARGET_LONG_BITS) {
+ return false;
+ }
+
+ arg_srli arg = { .rd = a->rd, .rs1 = a->rd, .shamt = a->shamt };
+ return trans_srli(ctx, &arg);
+}
+
+static bool trans_c_srai(DisasContext *ctx, arg_c_srai *a)
+{
+ int shamt = a->shamt;
+ if (shamt == 0) {
+ /* For RV128 a shamt of 0 means a shift by 64 */
+ shamt = 64;
+ }
+ /* Ensure, that shamt[5] is zero for RV32 */
+ if (shamt >= TARGET_LONG_BITS) {
+ return false;
+ }
+
+ arg_srai arg = { .rd = a->rd, .rs1 = a->rd, .shamt = a->shamt };
+ return trans_srai(ctx, &arg);
+}
+
+static bool trans_c_andi(DisasContext *ctx, arg_c_andi *a)
+{
+ arg_andi arg = { .rd = a->rd, .rs1 = a->rd, .imm = a->imm };
+ return trans_andi(ctx, &arg);
+}
+
+static bool trans_c_sub(DisasContext *ctx, arg_c_sub *a)
+{
+ arg_sub arg = { .rd = a->rd, .rs1 = a->rd, .rs2 = a->rs2 };
+ return trans_sub(ctx, &arg);
+}
+
+static bool trans_c_xor(DisasContext *ctx, arg_c_xor *a)
+{
+ arg_xor arg = { .rd = a->rd, .rs1 = a->rd, .rs2 = a->rs2 };
+ return trans_xor(ctx, &arg);
+}
+
+static bool trans_c_or(DisasContext *ctx, arg_c_or *a)
+{
+ arg_or arg = { .rd = a->rd, .rs1 = a->rd, .rs2 = a->rs2 };
+ return trans_or(ctx, &arg);
+}
+
+static bool trans_c_and(DisasContext *ctx, arg_c_and *a)
+{
+ arg_and arg = { .rd = a->rd, .rs1 = a->rd, .rs2 = a->rs2 };
+ return trans_and(ctx, &arg);
+}
+
+static bool trans_c_subw(DisasContext *ctx, arg_c_subw *a)
+{
+#ifdef TARGET_RISCV64
+ arg_subw arg = { .rd = a->rd, .rs1 = a->rd, .rs2 = a->rs2 };
+ return trans_subw(ctx, &arg);
+#else
+ return false;
+#endif
+}
+
+static bool trans_c_addw(DisasContext *ctx, arg_c_addw *a)
+{
+#ifdef TARGET_RISCV64
+ arg_addw arg = { .rd = a->rd, .rs1 = a->rd, .rs2 = a->rs2 };
+ return trans_addw(ctx, &arg);
+#else
+ return false;
+#endif
+}
+
+static bool trans_c_j(DisasContext *ctx, arg_c_j *a)
+{
+ arg_jal arg = { .rd = 0, .imm = a->imm };
+ return trans_jal(ctx, &arg);
+}
+
+static bool trans_c_beqz(DisasContext *ctx, arg_c_beqz *a)
+{
+ arg_beq arg = { .rs1 = a->rs1, .rs2 = 0, .imm = a->imm };
+ return trans_beq(ctx, &arg);
+}
+
+static bool trans_c_bnez(DisasContext *ctx, arg_c_bnez *a)
+{
+ arg_bne arg = { .rs1 = a->rs1, .rs2 = 0, .imm = a->imm };
+ return trans_bne(ctx, &arg);
+}
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 8bddebb35e..dc4d2c42ba 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -746,120 +746,6 @@ static void decode_RV32_64C0(DisasContext *ctx)
}
}
-static void decode_RV32_64C1(CPURISCVState *env, DisasContext *ctx)
-{
- uint8_t funct3 = extract32(ctx->opcode, 13, 3);
- uint8_t rd_rs1 = GET_C_RS1(ctx->opcode);
- uint8_t rs1s, rs2s;
- uint8_t funct2;
-
- switch (funct3) {
- case 0:
- /* C.ADDI -> addi rd, rd, nzimm[5:0] */
- gen_arith_imm(ctx, OPC_RISC_ADDI, rd_rs1, rd_rs1,
- GET_C_IMM(ctx->opcode));
- break;
- case 1:
-#if defined(TARGET_RISCV64)
- /* C.ADDIW (RV64/128) -> addiw rd, rd, imm[5:0]*/
- gen_arith_imm(ctx, OPC_RISC_ADDIW, rd_rs1, rd_rs1,
- GET_C_IMM(ctx->opcode));
-#else
- /* C.JAL(RV32) -> jal x1, offset[11:1] */
- gen_jal(env, ctx, 1, GET_C_J_IMM(ctx->opcode));
-#endif
- break;
- case 2:
- /* C.LI -> addi rd, x0, imm[5:0]*/
- gen_arith_imm(ctx, OPC_RISC_ADDI, rd_rs1, 0, GET_C_IMM(ctx->opcode));
- break;
- case 3:
- if (rd_rs1 == 2) {
- /* C.ADDI16SP -> addi x2, x2, nzimm[9:4]*/
- gen_arith_imm(ctx, OPC_RISC_ADDI, 2, 2,
- GET_C_ADDI16SP_IMM(ctx->opcode));
- } else if (rd_rs1 != 0) {
- /* C.LUI (rs1/rd =/= {0,2}) -> lui rd, nzimm[17:12]*/
- tcg_gen_movi_tl(cpu_gpr[rd_rs1],
- GET_C_IMM(ctx->opcode) << 12);
- }
- break;
- case 4:
- funct2 = extract32(ctx->opcode, 10, 2);
- rs1s = GET_C_RS1S(ctx->opcode);
- switch (funct2) {
- case 0: /* C.SRLI(RV32) -> srli rd', rd', shamt[5:0] */
- gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_I, rs1s, rs1s,
- GET_C_ZIMM(ctx->opcode));
- /* C.SRLI64(RV128) */
- break;
- case 1:
- /* C.SRAI -> srai rd', rd', shamt[5:0]*/
- gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_I, rs1s, rs1s,
- GET_C_ZIMM(ctx->opcode) | 0x400);
- /* C.SRAI64(RV128) */
- break;
- case 2:
- /* C.ANDI -> andi rd', rd', imm[5:0]*/
- gen_arith_imm(ctx, OPC_RISC_ANDI, rs1s, rs1s,
- GET_C_IMM(ctx->opcode));
- break;
- case 3:
- funct2 = extract32(ctx->opcode, 5, 2);
- rs2s = GET_C_RS2S(ctx->opcode);
- switch (funct2) {
- case 0:
- /* C.SUB -> sub rd', rd', rs2' */
- if (extract32(ctx->opcode, 12, 1) == 0) {
- gen_arith(ctx, OPC_RISC_SUB, rs1s, rs1s, rs2s);
- }
-#if defined(TARGET_RISCV64)
- else {
- gen_arith(ctx, OPC_RISC_SUBW, rs1s, rs1s, rs2s);
- }
-#endif
- break;
- case 1:
- /* C.XOR -> xor rs1', rs1', rs2' */
- if (extract32(ctx->opcode, 12, 1) == 0) {
- gen_arith(ctx, OPC_RISC_XOR, rs1s, rs1s, rs2s);
- }
-#if defined(TARGET_RISCV64)
- else {
- /* C.ADDW (RV64/128) */
- gen_arith(ctx, OPC_RISC_ADDW, rs1s, rs1s, rs2s);
- }
-#endif
- break;
- case 2:
- /* C.OR -> or rs1', rs1', rs2' */
- gen_arith(ctx, OPC_RISC_OR, rs1s, rs1s, rs2s);
- break;
- case 3:
- /* C.AND -> and rs1', rs1', rs2' */
- gen_arith(ctx, OPC_RISC_AND, rs1s, rs1s, rs2s);
- break;
- }
- break;
- }
- break;
- case 5:
- /* C.J -> jal x0, offset[11:1]*/
- gen_jal(env, ctx, 0, GET_C_J_IMM(ctx->opcode));
- break;
- case 6:
- /* C.BEQZ -> beq rs1', x0, offset[8:1]*/
- rs1s = GET_C_RS1S(ctx->opcode);
- gen_branch(env, ctx, OPC_RISC_BEQ, rs1s, 0, GET_C_B_IMM(ctx->opcode));
- break;
- case 7:
- /* C.BNEZ -> bne rs1', x0, offset[8:1]*/
- rs1s = GET_C_RS1S(ctx->opcode);
- gen_branch(env, ctx, OPC_RISC_BNE, rs1s, 0, GET_C_B_IMM(ctx->opcode));
- break;
- }
-}
-
static void decode_RV32_64C2(CPURISCVState *env, DisasContext *ctx)
{
uint8_t rd, rs2;
@@ -946,9 +832,6 @@ static void decode_RV32_64C(CPURISCVState *env, DisasContext *ctx)
case 0:
decode_RV32_64C0(ctx);
break;
- case 1:
- decode_RV32_64C1(env, ctx);
- break;
case 2:
decode_RV32_64C2(env, ctx);
break;
@@ -963,6 +846,7 @@ static void decode_RV32_64C(CPURISCVState *env, DisasContext *ctx)
EX_SH(1)
EX_SH(2)
EX_SH(3)
+EX_SH(4)
EX_SH(12)
static int ex_rvc_register(int reg)
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 19/35] target/riscv: Convert quadrant 2 of RVXC insns to decodetree
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- dropped insn argument of trans_foo functions
target/riscv/insn16.decode | 31 ++++++++
target/riscv/insn_trans/trans_rvc.inc.c | 101 ++++++++++++++++++++++++
target/riscv/translate.c | 83 +------------------
3 files changed, 134 insertions(+), 81 deletions(-)
diff --git a/target/riscv/insn16.decode b/target/riscv/insn16.decode
index 29dade0fa1..0829e3bc59 100644
--- a/target/riscv/insn16.decode
+++ b/target/riscv/insn16.decode
@@ -20,6 +20,7 @@
%rd 7:5
%rs1_3 7:3 !function=ex_rvc_register
%rs2_3 2:3 !function=ex_rvc_register
+%rs2_5 2:5
# Immediates:
%imm_ci 12:s1 2:5
@@ -30,6 +31,10 @@
%imm_cj 12:s1 8:1 9:2 6:1 7:1 2:1 11:1 3:3 !function=ex_shift_1
%nzuimm_6bit 12:1 2:5
+%uimm_6bit_ld 2:3 12:1 5:2 !function=ex_shift_3
+%uimm_6bit_lw 2:2 12:1 4:3 !function=ex_shift_2
+%uimm_6bit_sd 7:3 10:3 !function=ex_shift_3
+%uimm_6bit_sw 7:2 9:4 !function=ex_shift_2
%imm_addi16sp 12:s1 3:2 5:1 2:1 6:1 !function=ex_shift_4
%imm_lui 12:s1 2:5 !function=ex_shift_12
@@ -48,10 +53,15 @@
&c_j imm
&c_shift shamt rd
+&c_ld uimm rd
+&c_sd uimm rs2
&c_addi16sp_lui imm_lui imm_addi16sp rd
+&c_flwsp_ldsp uimm_flwsp uimm_ldsp rd
+&c_fswsp_sdsp uimm_fswsp uimm_sdsp rs2
# Formats 16:
+@cr .... ..... ..... .. &cr rs2=%rs2_5 %rd
@ci ... . ..... ..... .. &ci imm=%imm_ci %rd
@ciw ... ........ ... .. &ciw nzuimm=%nzuimm_ciw rd=%rs2_3
@cl_d ... ... ... .. ... .. &cl_dw uimm=%uimm_cl_d rs1=%rs1_3 rd=%rs2_3
@@ -64,9 +74,19 @@
@cb ... ... ... .. ... .. &cb imm=%imm_cb rs1=%rs1_3
@cj ... ........... .. &c_j imm=%imm_cj
+@c_ld ... . ..... ..... .. &c_ld uimm=%uimm_6bit_ld %rd
+@c_lw ... . ..... ..... .. &c_ld uimm=%uimm_6bit_lw %rd
+@c_sd ... . ..... ..... .. &c_sd uimm=%uimm_6bit_sd rs2=%rs2_5
+@c_sw ... . ..... ..... .. &c_sd uimm=%uimm_6bit_sw rs2=%rs2_5
+
@c_addi16sp_lui ... . ..... ..... .. &c_addi16sp_lui %imm_lui %imm_addi16sp %rd
+@c_flwsp_ldsp ... . ..... ..... .. &c_flwsp_ldsp uimm_flwsp=%uimm_6bit_lw \
+ uimm_ldsp=%uimm_6bit_ld %rd
+@c_fswsp_sdsp ... . ..... ..... .. &c_fswsp_sdsp uimm_fswsp=%uimm_6bit_sw \
+ uimm_sdsp=%uimm_6bit_sd rs2=%rs2_5
@c_shift ... . .. ... ..... .. &c_shift rd=%rs1_3 shamt=%nzuimm_6bit
+@c_shift2 ... . .. ... ..... .. &c_shift rd=%rd shamt=%nzuimm_6bit
@c_andi ... . .. ... ..... .. &ci imm=%imm_ci rd=%rs1_3
@@ -96,3 +116,14 @@ c_addw 100 1 11 ... 01 ... 01 @cs_2
c_j 101 ........... 01 @cj
c_beqz 110 ... ... ..... 01 @cb
c_bnez 111 ... ... ..... 01 @cb
+
+# *** RV64C Standard Extension (Quadrant 2) ***
+c_slli 000 . ..... ..... 10 @c_shift2
+c_fldsp 001 . ..... ..... 10 @c_ld
+c_lwsp 010 . ..... ..... 10 @c_lw
+c_flwsp_ldsp 011 . ..... ..... 10 @c_flwsp_ldsp #C.LDSP:RV64;C.FLWSP:RV32
+c_jr_mv 100 0 ..... ..... 10 @cr
+c_ebreak_jalr_add 100 1 ..... ..... 10 @cr
+c_fsdsp 101 ...... ..... 10 @c_sd
+c_swsp 110 . ..... ..... 10 @c_sw
+c_fswsp_sdsp 111 . ..... ..... 10 @c_fswsp_sdsp #C.SDSP:RV64;C.FSWSP:RV32
diff --git a/target/riscv/insn_trans/trans_rvc.inc.c b/target/riscv/insn_trans/trans_rvc.inc.c
index b06c435c98..bcdf64d3b7 100644
--- a/target/riscv/insn_trans/trans_rvc.inc.c
+++ b/target/riscv/insn_trans/trans_rvc.inc.c
@@ -224,3 +224,104 @@ static bool trans_c_bnez(DisasContext *ctx, arg_c_bnez *a)
arg_bne arg = { .rs1 = a->rs1, .rs2 = 0, .imm = a->imm };
return trans_bne(ctx, &arg);
}
+
+static bool trans_c_slli(DisasContext *ctx, arg_c_slli *a)
+{
+ int shamt = a->shamt;
+ if (shamt == 0) {
+ /* For RV128 a shamt of 0 means a shift by 64 */
+ shamt = 64;
+ }
+ /* Ensure, that shamt[5] is zero for RV32 */
+ if (shamt >= TARGET_LONG_BITS) {
+ return false;
+ }
+
+ arg_slli arg = { .rd = a->rd, .rs1 = a->rd, .shamt = a->shamt };
+ return trans_slli(ctx, &arg);
+}
+
+static bool trans_c_fldsp(DisasContext *ctx, arg_c_fldsp *a)
+{
+ arg_fld arg = { .rd = a->rd, .rs1 = 2, .imm = a->uimm };
+ return trans_fld(ctx, &arg);
+}
+
+static bool trans_c_lwsp(DisasContext *ctx, arg_c_lwsp *a)
+{
+ arg_lw arg = { .rd = a->rd, .rs1 = 2, .imm = a->uimm };
+ return trans_lw(ctx, &arg);
+}
+
+static bool trans_c_flwsp_ldsp(DisasContext *ctx, arg_c_flwsp_ldsp *a)
+{
+#ifdef TARGET_RISCV32
+ /* C.FLWSP */
+ arg_flw arg_flw = { .rd = a->rd, .rs1 = 2, .imm = a->uimm_flwsp };
+ return trans_flw(ctx, &arg_flw);
+#else
+ /* C.LDSP */
+ arg_ld arg_ld = { .rd = a->rd, .rs1 = 2, .imm = a->uimm_ldsp };
+ return trans_ld(ctx, &arg_ld);
+#endif
+ return false;
+}
+
+static bool trans_c_jr_mv(DisasContext *ctx, arg_c_jr_mv *a)
+{
+ if (a->rd != 0 && a->rs2 == 0) {
+ /* C.JR */
+ arg_jalr arg = { .rd = 0, .rs1 = a->rd, .imm = 0 };
+ return trans_jalr(ctx, &arg);
+ } else if (a->rd != 0 && a->rs2 != 0) {
+ /* C.MV */
+ arg_add arg = { .rd = a->rd, .rs1 = 0, .rs2 = a->rs2 };
+ return trans_add(ctx, &arg);
+ }
+ return false;
+}
+
+static bool trans_c_ebreak_jalr_add(DisasContext *ctx, arg_c_ebreak_jalr_add *a)
+{
+ if (a->rd == 0 && a->rs2 == 0) {
+ /* C.EBREAK */
+ arg_ebreak arg = { };
+ return trans_ebreak(ctx, &arg);
+ } else if (a->rd != 0) {
+ if (a->rs2 == 0) {
+ /* C.JALR */
+ arg_jalr arg = { .rd = 1, .rs1 = a->rd, .imm = 0 };
+ return trans_jalr(ctx, &arg);
+ } else {
+ /* C.ADD */
+ arg_add arg = { .rd = a->rd, .rs1 = a->rd, .rs2 = a->rs2 };
+ return trans_add(ctx, &arg);
+ }
+ }
+ return false;
+}
+
+static bool trans_c_fsdsp(DisasContext *ctx, arg_c_fsdsp *a)
+{
+ arg_fsd arg = { .rs1 = 2, .rs2 = a->rs2, .imm = a->uimm };
+ return trans_fsd(ctx, &arg);
+}
+
+static bool trans_c_swsp(DisasContext *ctx, arg_c_swsp *a)
+{
+ arg_sw arg = { .rs1 = 2, .rs2 = a->rs2, .imm = a->uimm };
+ return trans_sw(ctx, &arg);
+}
+
+static bool trans_c_fswsp_sdsp(DisasContext *ctx, arg_c_fswsp_sdsp *a)
+{
+#ifdef TARGET_RISCV32
+ /* C.FSWSP */
+ arg_fsw a_fsw = { .rs1 = a->rs2, .rs2 = 2, .imm = a->uimm_fswsp };
+ return trans_fsw(ctx, &a_fsw);
+#else
+ /* C.SDSP */
+ arg_sd a_sd = { .rs1 = 2, .rs2 = a->rs2, .imm = a->uimm_sdsp };
+ return trans_sd(ctx, &a_sd);
+#endif
+}
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index dc4d2c42ba..41d66fae18 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -614,6 +614,7 @@ static void gen_store(DisasContext *ctx, uint32_t opc, int rs1, int rs2,
tcg_temp_free(dat);
}
+#ifdef TARGET_RISCV32
static void gen_fp_load(DisasContext *ctx, uint32_t opc, int rd,
int rs1, target_long imm)
{
@@ -672,6 +673,7 @@ static void gen_fp_store(DisasContext *ctx, uint32_t opc, int rs1,
tcg_temp_free(t0);
}
+#endif
static void gen_set_rm(DisasContext *ctx, int rm)
{
@@ -746,84 +748,6 @@ static void decode_RV32_64C0(DisasContext *ctx)
}
}
-static void decode_RV32_64C2(CPURISCVState *env, DisasContext *ctx)
-{
- uint8_t rd, rs2;
- uint8_t funct3 = extract32(ctx->opcode, 13, 3);
-
-
- rd = GET_RD(ctx->opcode);
-
- switch (funct3) {
- case 0: /* C.SLLI -> slli rd, rd, shamt[5:0]
- C.SLLI64 -> */
- gen_arith_imm(ctx, OPC_RISC_SLLI, rd, rd, GET_C_ZIMM(ctx->opcode));
- break;
- case 1: /* C.FLDSP(RV32/64DC) -> fld rd, offset[8:3](x2) */
- gen_fp_load(ctx, OPC_RISC_FLD, rd, 2, GET_C_LDSP_IMM(ctx->opcode));
- break;
- case 2: /* C.LWSP -> lw rd, offset[7:2](x2) */
- gen_load(ctx, OPC_RISC_LW, rd, 2, GET_C_LWSP_IMM(ctx->opcode));
- break;
- case 3:
-#if defined(TARGET_RISCV64)
- /* C.LDSP(RVC64) -> ld rd, offset[8:3](x2) */
- gen_load(ctx, OPC_RISC_LD, rd, 2, GET_C_LDSP_IMM(ctx->opcode));
-#else
- /* C.FLWSP(RV32FC) -> flw rd, offset[7:2](x2) */
- gen_fp_load(ctx, OPC_RISC_FLW, rd, 2, GET_C_LWSP_IMM(ctx->opcode));
-#endif
- break;
- case 4:
- rs2 = GET_C_RS2(ctx->opcode);
-
- if (extract32(ctx->opcode, 12, 1) == 0) {
- if (rs2 == 0) {
- /* C.JR -> jalr x0, rs1, 0*/
- gen_jalr(env, ctx, OPC_RISC_JALR, 0, rd, 0);
- } else {
- /* C.MV -> add rd, x0, rs2 */
- gen_arith(ctx, OPC_RISC_ADD, rd, 0, rs2);
- }
- } else {
- if (rd == 0) {
- /* C.EBREAK -> ebreak*/
- gen_system(env, ctx, OPC_RISC_ECALL, 0, 0, 0x1);
- } else {
- if (rs2 == 0) {
- /* C.JALR -> jalr x1, rs1, 0*/
- gen_jalr(env, ctx, OPC_RISC_JALR, 1, rd, 0);
- } else {
- /* C.ADD -> add rd, rd, rs2 */
- gen_arith(ctx, OPC_RISC_ADD, rd, rd, rs2);
- }
- }
- }
- break;
- case 5:
- /* C.FSDSP -> fsd rs2, offset[8:3](x2)*/
- gen_fp_store(ctx, OPC_RISC_FSD, 2, GET_C_RS2(ctx->opcode),
- GET_C_SDSP_IMM(ctx->opcode));
- /* C.SQSP */
- break;
- case 6: /* C.SWSP -> sw rs2, offset[7:2](x2)*/
- gen_store(ctx, OPC_RISC_SW, 2, GET_C_RS2(ctx->opcode),
- GET_C_SWSP_IMM(ctx->opcode));
- break;
- case 7:
-#if defined(TARGET_RISCV64)
- /* C.SDSP(Rv64/128) -> sd rs2, offset[8:3](x2)*/
- gen_store(ctx, OPC_RISC_SD, 2, GET_C_RS2(ctx->opcode),
- GET_C_SDSP_IMM(ctx->opcode));
-#else
- /* C.FSWSP(RV32) -> fsw rs2, offset[7:2](x2) */
- gen_fp_store(ctx, OPC_RISC_FSW, 2, GET_C_RS2(ctx->opcode),
- GET_C_SWSP_IMM(ctx->opcode));
-#endif
- break;
- }
-}
-
static void decode_RV32_64C(CPURISCVState *env, DisasContext *ctx)
{
uint8_t op = extract32(ctx->opcode, 0, 2);
@@ -832,9 +756,6 @@ static void decode_RV32_64C(CPURISCVState *env, DisasContext *ctx)
case 0:
decode_RV32_64C0(ctx);
break;
- case 2:
- decode_RV32_64C2(env, ctx);
- break;
}
}
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 19/35] target/riscv: Convert quadrant 2 of RVXC insns to decodetree
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- dropped insn argument of trans_foo functions
target/riscv/insn16.decode | 31 ++++++++
target/riscv/insn_trans/trans_rvc.inc.c | 101 ++++++++++++++++++++++++
target/riscv/translate.c | 83 +------------------
3 files changed, 134 insertions(+), 81 deletions(-)
diff --git a/target/riscv/insn16.decode b/target/riscv/insn16.decode
index 29dade0fa1..0829e3bc59 100644
--- a/target/riscv/insn16.decode
+++ b/target/riscv/insn16.decode
@@ -20,6 +20,7 @@
%rd 7:5
%rs1_3 7:3 !function=ex_rvc_register
%rs2_3 2:3 !function=ex_rvc_register
+%rs2_5 2:5
# Immediates:
%imm_ci 12:s1 2:5
@@ -30,6 +31,10 @@
%imm_cj 12:s1 8:1 9:2 6:1 7:1 2:1 11:1 3:3 !function=ex_shift_1
%nzuimm_6bit 12:1 2:5
+%uimm_6bit_ld 2:3 12:1 5:2 !function=ex_shift_3
+%uimm_6bit_lw 2:2 12:1 4:3 !function=ex_shift_2
+%uimm_6bit_sd 7:3 10:3 !function=ex_shift_3
+%uimm_6bit_sw 7:2 9:4 !function=ex_shift_2
%imm_addi16sp 12:s1 3:2 5:1 2:1 6:1 !function=ex_shift_4
%imm_lui 12:s1 2:5 !function=ex_shift_12
@@ -48,10 +53,15 @@
&c_j imm
&c_shift shamt rd
+&c_ld uimm rd
+&c_sd uimm rs2
&c_addi16sp_lui imm_lui imm_addi16sp rd
+&c_flwsp_ldsp uimm_flwsp uimm_ldsp rd
+&c_fswsp_sdsp uimm_fswsp uimm_sdsp rs2
# Formats 16:
+@cr .... ..... ..... .. &cr rs2=%rs2_5 %rd
@ci ... . ..... ..... .. &ci imm=%imm_ci %rd
@ciw ... ........ ... .. &ciw nzuimm=%nzuimm_ciw rd=%rs2_3
@cl_d ... ... ... .. ... .. &cl_dw uimm=%uimm_cl_d rs1=%rs1_3 rd=%rs2_3
@@ -64,9 +74,19 @@
@cb ... ... ... .. ... .. &cb imm=%imm_cb rs1=%rs1_3
@cj ... ........... .. &c_j imm=%imm_cj
+@c_ld ... . ..... ..... .. &c_ld uimm=%uimm_6bit_ld %rd
+@c_lw ... . ..... ..... .. &c_ld uimm=%uimm_6bit_lw %rd
+@c_sd ... . ..... ..... .. &c_sd uimm=%uimm_6bit_sd rs2=%rs2_5
+@c_sw ... . ..... ..... .. &c_sd uimm=%uimm_6bit_sw rs2=%rs2_5
+
@c_addi16sp_lui ... . ..... ..... .. &c_addi16sp_lui %imm_lui %imm_addi16sp %rd
+@c_flwsp_ldsp ... . ..... ..... .. &c_flwsp_ldsp uimm_flwsp=%uimm_6bit_lw \
+ uimm_ldsp=%uimm_6bit_ld %rd
+@c_fswsp_sdsp ... . ..... ..... .. &c_fswsp_sdsp uimm_fswsp=%uimm_6bit_sw \
+ uimm_sdsp=%uimm_6bit_sd rs2=%rs2_5
@c_shift ... . .. ... ..... .. &c_shift rd=%rs1_3 shamt=%nzuimm_6bit
+@c_shift2 ... . .. ... ..... .. &c_shift rd=%rd shamt=%nzuimm_6bit
@c_andi ... . .. ... ..... .. &ci imm=%imm_ci rd=%rs1_3
@@ -96,3 +116,14 @@ c_addw 100 1 11 ... 01 ... 01 @cs_2
c_j 101 ........... 01 @cj
c_beqz 110 ... ... ..... 01 @cb
c_bnez 111 ... ... ..... 01 @cb
+
+# *** RV64C Standard Extension (Quadrant 2) ***
+c_slli 000 . ..... ..... 10 @c_shift2
+c_fldsp 001 . ..... ..... 10 @c_ld
+c_lwsp 010 . ..... ..... 10 @c_lw
+c_flwsp_ldsp 011 . ..... ..... 10 @c_flwsp_ldsp #C.LDSP:RV64;C.FLWSP:RV32
+c_jr_mv 100 0 ..... ..... 10 @cr
+c_ebreak_jalr_add 100 1 ..... ..... 10 @cr
+c_fsdsp 101 ...... ..... 10 @c_sd
+c_swsp 110 . ..... ..... 10 @c_sw
+c_fswsp_sdsp 111 . ..... ..... 10 @c_fswsp_sdsp #C.SDSP:RV64;C.FSWSP:RV32
diff --git a/target/riscv/insn_trans/trans_rvc.inc.c b/target/riscv/insn_trans/trans_rvc.inc.c
index b06c435c98..bcdf64d3b7 100644
--- a/target/riscv/insn_trans/trans_rvc.inc.c
+++ b/target/riscv/insn_trans/trans_rvc.inc.c
@@ -224,3 +224,104 @@ static bool trans_c_bnez(DisasContext *ctx, arg_c_bnez *a)
arg_bne arg = { .rs1 = a->rs1, .rs2 = 0, .imm = a->imm };
return trans_bne(ctx, &arg);
}
+
+static bool trans_c_slli(DisasContext *ctx, arg_c_slli *a)
+{
+ int shamt = a->shamt;
+ if (shamt == 0) {
+ /* For RV128 a shamt of 0 means a shift by 64 */
+ shamt = 64;
+ }
+ /* Ensure, that shamt[5] is zero for RV32 */
+ if (shamt >= TARGET_LONG_BITS) {
+ return false;
+ }
+
+ arg_slli arg = { .rd = a->rd, .rs1 = a->rd, .shamt = a->shamt };
+ return trans_slli(ctx, &arg);
+}
+
+static bool trans_c_fldsp(DisasContext *ctx, arg_c_fldsp *a)
+{
+ arg_fld arg = { .rd = a->rd, .rs1 = 2, .imm = a->uimm };
+ return trans_fld(ctx, &arg);
+}
+
+static bool trans_c_lwsp(DisasContext *ctx, arg_c_lwsp *a)
+{
+ arg_lw arg = { .rd = a->rd, .rs1 = 2, .imm = a->uimm };
+ return trans_lw(ctx, &arg);
+}
+
+static bool trans_c_flwsp_ldsp(DisasContext *ctx, arg_c_flwsp_ldsp *a)
+{
+#ifdef TARGET_RISCV32
+ /* C.FLWSP */
+ arg_flw arg_flw = { .rd = a->rd, .rs1 = 2, .imm = a->uimm_flwsp };
+ return trans_flw(ctx, &arg_flw);
+#else
+ /* C.LDSP */
+ arg_ld arg_ld = { .rd = a->rd, .rs1 = 2, .imm = a->uimm_ldsp };
+ return trans_ld(ctx, &arg_ld);
+#endif
+ return false;
+}
+
+static bool trans_c_jr_mv(DisasContext *ctx, arg_c_jr_mv *a)
+{
+ if (a->rd != 0 && a->rs2 == 0) {
+ /* C.JR */
+ arg_jalr arg = { .rd = 0, .rs1 = a->rd, .imm = 0 };
+ return trans_jalr(ctx, &arg);
+ } else if (a->rd != 0 && a->rs2 != 0) {
+ /* C.MV */
+ arg_add arg = { .rd = a->rd, .rs1 = 0, .rs2 = a->rs2 };
+ return trans_add(ctx, &arg);
+ }
+ return false;
+}
+
+static bool trans_c_ebreak_jalr_add(DisasContext *ctx, arg_c_ebreak_jalr_add *a)
+{
+ if (a->rd == 0 && a->rs2 == 0) {
+ /* C.EBREAK */
+ arg_ebreak arg = { };
+ return trans_ebreak(ctx, &arg);
+ } else if (a->rd != 0) {
+ if (a->rs2 == 0) {
+ /* C.JALR */
+ arg_jalr arg = { .rd = 1, .rs1 = a->rd, .imm = 0 };
+ return trans_jalr(ctx, &arg);
+ } else {
+ /* C.ADD */
+ arg_add arg = { .rd = a->rd, .rs1 = a->rd, .rs2 = a->rs2 };
+ return trans_add(ctx, &arg);
+ }
+ }
+ return false;
+}
+
+static bool trans_c_fsdsp(DisasContext *ctx, arg_c_fsdsp *a)
+{
+ arg_fsd arg = { .rs1 = 2, .rs2 = a->rs2, .imm = a->uimm };
+ return trans_fsd(ctx, &arg);
+}
+
+static bool trans_c_swsp(DisasContext *ctx, arg_c_swsp *a)
+{
+ arg_sw arg = { .rs1 = 2, .rs2 = a->rs2, .imm = a->uimm };
+ return trans_sw(ctx, &arg);
+}
+
+static bool trans_c_fswsp_sdsp(DisasContext *ctx, arg_c_fswsp_sdsp *a)
+{
+#ifdef TARGET_RISCV32
+ /* C.FSWSP */
+ arg_fsw a_fsw = { .rs1 = a->rs2, .rs2 = 2, .imm = a->uimm_fswsp };
+ return trans_fsw(ctx, &a_fsw);
+#else
+ /* C.SDSP */
+ arg_sd a_sd = { .rs1 = 2, .rs2 = a->rs2, .imm = a->uimm_sdsp };
+ return trans_sd(ctx, &a_sd);
+#endif
+}
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index dc4d2c42ba..41d66fae18 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -614,6 +614,7 @@ static void gen_store(DisasContext *ctx, uint32_t opc, int rs1, int rs2,
tcg_temp_free(dat);
}
+#ifdef TARGET_RISCV32
static void gen_fp_load(DisasContext *ctx, uint32_t opc, int rd,
int rs1, target_long imm)
{
@@ -672,6 +673,7 @@ static void gen_fp_store(DisasContext *ctx, uint32_t opc, int rs1,
tcg_temp_free(t0);
}
+#endif
static void gen_set_rm(DisasContext *ctx, int rm)
{
@@ -746,84 +748,6 @@ static void decode_RV32_64C0(DisasContext *ctx)
}
}
-static void decode_RV32_64C2(CPURISCVState *env, DisasContext *ctx)
-{
- uint8_t rd, rs2;
- uint8_t funct3 = extract32(ctx->opcode, 13, 3);
-
-
- rd = GET_RD(ctx->opcode);
-
- switch (funct3) {
- case 0: /* C.SLLI -> slli rd, rd, shamt[5:0]
- C.SLLI64 -> */
- gen_arith_imm(ctx, OPC_RISC_SLLI, rd, rd, GET_C_ZIMM(ctx->opcode));
- break;
- case 1: /* C.FLDSP(RV32/64DC) -> fld rd, offset[8:3](x2) */
- gen_fp_load(ctx, OPC_RISC_FLD, rd, 2, GET_C_LDSP_IMM(ctx->opcode));
- break;
- case 2: /* C.LWSP -> lw rd, offset[7:2](x2) */
- gen_load(ctx, OPC_RISC_LW, rd, 2, GET_C_LWSP_IMM(ctx->opcode));
- break;
- case 3:
-#if defined(TARGET_RISCV64)
- /* C.LDSP(RVC64) -> ld rd, offset[8:3](x2) */
- gen_load(ctx, OPC_RISC_LD, rd, 2, GET_C_LDSP_IMM(ctx->opcode));
-#else
- /* C.FLWSP(RV32FC) -> flw rd, offset[7:2](x2) */
- gen_fp_load(ctx, OPC_RISC_FLW, rd, 2, GET_C_LWSP_IMM(ctx->opcode));
-#endif
- break;
- case 4:
- rs2 = GET_C_RS2(ctx->opcode);
-
- if (extract32(ctx->opcode, 12, 1) == 0) {
- if (rs2 == 0) {
- /* C.JR -> jalr x0, rs1, 0*/
- gen_jalr(env, ctx, OPC_RISC_JALR, 0, rd, 0);
- } else {
- /* C.MV -> add rd, x0, rs2 */
- gen_arith(ctx, OPC_RISC_ADD, rd, 0, rs2);
- }
- } else {
- if (rd == 0) {
- /* C.EBREAK -> ebreak*/
- gen_system(env, ctx, OPC_RISC_ECALL, 0, 0, 0x1);
- } else {
- if (rs2 == 0) {
- /* C.JALR -> jalr x1, rs1, 0*/
- gen_jalr(env, ctx, OPC_RISC_JALR, 1, rd, 0);
- } else {
- /* C.ADD -> add rd, rd, rs2 */
- gen_arith(ctx, OPC_RISC_ADD, rd, rd, rs2);
- }
- }
- }
- break;
- case 5:
- /* C.FSDSP -> fsd rs2, offset[8:3](x2)*/
- gen_fp_store(ctx, OPC_RISC_FSD, 2, GET_C_RS2(ctx->opcode),
- GET_C_SDSP_IMM(ctx->opcode));
- /* C.SQSP */
- break;
- case 6: /* C.SWSP -> sw rs2, offset[7:2](x2)*/
- gen_store(ctx, OPC_RISC_SW, 2, GET_C_RS2(ctx->opcode),
- GET_C_SWSP_IMM(ctx->opcode));
- break;
- case 7:
-#if defined(TARGET_RISCV64)
- /* C.SDSP(Rv64/128) -> sd rs2, offset[8:3](x2)*/
- gen_store(ctx, OPC_RISC_SD, 2, GET_C_RS2(ctx->opcode),
- GET_C_SDSP_IMM(ctx->opcode));
-#else
- /* C.FSWSP(RV32) -> fsw rs2, offset[7:2](x2) */
- gen_fp_store(ctx, OPC_RISC_FSW, 2, GET_C_RS2(ctx->opcode),
- GET_C_SWSP_IMM(ctx->opcode));
-#endif
- break;
- }
-}
-
static void decode_RV32_64C(CPURISCVState *env, DisasContext *ctx)
{
uint8_t op = extract32(ctx->opcode, 0, 2);
@@ -832,9 +756,6 @@ static void decode_RV32_64C(CPURISCVState *env, DisasContext *ctx)
case 0:
decode_RV32_64C0(ctx);
break;
- case 2:
- decode_RV32_64C2(env, ctx);
- break;
}
}
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 20/35] target/riscv: Remove gen_jalr()
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
trans_jalr() is the only caller, so move the code into trans_jalr().
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
target/riscv/insn_trans/trans_rvi.inc.c | 28 +++++++++++++++++-
target/riscv/translate.c | 38 -------------------------
2 files changed, 27 insertions(+), 39 deletions(-)
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
index 09e7a0052a..4d090d68e7 100644
--- a/target/riscv/insn_trans/trans_rvi.inc.c
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -42,7 +42,33 @@ static bool trans_jal(DisasContext *ctx, arg_jal *a)
static bool trans_jalr(DisasContext *ctx, arg_jalr *a)
{
- gen_jalr(ctx->env, ctx, OPC_RISC_JALR, a->rd, a->rs1, a->imm);
+ /* no chaining with JALR */
+ TCGLabel *misaligned = NULL;
+ TCGv t0 = tcg_temp_new();
+
+
+ gen_get_gpr(cpu_pc, a->rs1);
+ tcg_gen_addi_tl(cpu_pc, cpu_pc, a->imm);
+ tcg_gen_andi_tl(cpu_pc, cpu_pc, (target_ulong)-2);
+
+ if (!riscv_has_ext(ctx->env, RVC)) {
+ misaligned = gen_new_label();
+ tcg_gen_andi_tl(t0, cpu_pc, 0x2);
+ tcg_gen_brcondi_tl(TCG_COND_NE, t0, 0x0, misaligned);
+ }
+
+ if (a->rd != 0) {
+ tcg_gen_movi_tl(cpu_gpr[a->rd], ctx->pc_succ_insn);
+ }
+ tcg_gen_lookup_and_goto_ptr();
+
+ if (misaligned) {
+ gen_set_label(misaligned);
+ gen_exception_inst_addr_mis(ctx);
+ }
+ ctx->base.is_jmp = DISAS_NORETURN;
+
+ tcg_temp_free(t0);
return true;
}
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 41d66fae18..b78e423d94 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -489,44 +489,6 @@ static void gen_jal(CPURISCVState *env, DisasContext *ctx, int rd,
ctx->base.is_jmp = DISAS_NORETURN;
}
-static void gen_jalr(CPURISCVState *env, DisasContext *ctx, uint32_t opc,
- int rd, int rs1, target_long imm)
-{
- /* no chaining with JALR */
- TCGLabel *misaligned = NULL;
- TCGv t0 = tcg_temp_new();
-
- switch (opc) {
- case OPC_RISC_JALR:
- gen_get_gpr(cpu_pc, rs1);
- tcg_gen_addi_tl(cpu_pc, cpu_pc, imm);
- tcg_gen_andi_tl(cpu_pc, cpu_pc, (target_ulong)-2);
-
- if (!riscv_has_ext(env, RVC)) {
- misaligned = gen_new_label();
- tcg_gen_andi_tl(t0, cpu_pc, 0x2);
- tcg_gen_brcondi_tl(TCG_COND_NE, t0, 0x0, misaligned);
- }
-
- if (rd != 0) {
- tcg_gen_movi_tl(cpu_gpr[rd], ctx->pc_succ_insn);
- }
- tcg_gen_lookup_and_goto_ptr();
-
- if (misaligned) {
- gen_set_label(misaligned);
- gen_exception_inst_addr_mis(ctx);
- }
- ctx->base.is_jmp = DISAS_NORETURN;
- break;
-
- default:
- gen_exception_illegal(ctx);
- break;
- }
- tcg_temp_free(t0);
-}
-
static void gen_branch(CPURISCVState *env, DisasContext *ctx, uint32_t opc,
int rs1, int rs2, target_long bimm)
{
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 20/35] target/riscv: Remove gen_jalr()
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
trans_jalr() is the only caller, so move the code into trans_jalr().
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
target/riscv/insn_trans/trans_rvi.inc.c | 28 +++++++++++++++++-
target/riscv/translate.c | 38 -------------------------
2 files changed, 27 insertions(+), 39 deletions(-)
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
index 09e7a0052a..4d090d68e7 100644
--- a/target/riscv/insn_trans/trans_rvi.inc.c
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -42,7 +42,33 @@ static bool trans_jal(DisasContext *ctx, arg_jal *a)
static bool trans_jalr(DisasContext *ctx, arg_jalr *a)
{
- gen_jalr(ctx->env, ctx, OPC_RISC_JALR, a->rd, a->rs1, a->imm);
+ /* no chaining with JALR */
+ TCGLabel *misaligned = NULL;
+ TCGv t0 = tcg_temp_new();
+
+
+ gen_get_gpr(cpu_pc, a->rs1);
+ tcg_gen_addi_tl(cpu_pc, cpu_pc, a->imm);
+ tcg_gen_andi_tl(cpu_pc, cpu_pc, (target_ulong)-2);
+
+ if (!riscv_has_ext(ctx->env, RVC)) {
+ misaligned = gen_new_label();
+ tcg_gen_andi_tl(t0, cpu_pc, 0x2);
+ tcg_gen_brcondi_tl(TCG_COND_NE, t0, 0x0, misaligned);
+ }
+
+ if (a->rd != 0) {
+ tcg_gen_movi_tl(cpu_gpr[a->rd], ctx->pc_succ_insn);
+ }
+ tcg_gen_lookup_and_goto_ptr();
+
+ if (misaligned) {
+ gen_set_label(misaligned);
+ gen_exception_inst_addr_mis(ctx);
+ }
+ ctx->base.is_jmp = DISAS_NORETURN;
+
+ tcg_temp_free(t0);
return true;
}
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 41d66fae18..b78e423d94 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -489,44 +489,6 @@ static void gen_jal(CPURISCVState *env, DisasContext *ctx, int rd,
ctx->base.is_jmp = DISAS_NORETURN;
}
-static void gen_jalr(CPURISCVState *env, DisasContext *ctx, uint32_t opc,
- int rd, int rs1, target_long imm)
-{
- /* no chaining with JALR */
- TCGLabel *misaligned = NULL;
- TCGv t0 = tcg_temp_new();
-
- switch (opc) {
- case OPC_RISC_JALR:
- gen_get_gpr(cpu_pc, rs1);
- tcg_gen_addi_tl(cpu_pc, cpu_pc, imm);
- tcg_gen_andi_tl(cpu_pc, cpu_pc, (target_ulong)-2);
-
- if (!riscv_has_ext(env, RVC)) {
- misaligned = gen_new_label();
- tcg_gen_andi_tl(t0, cpu_pc, 0x2);
- tcg_gen_brcondi_tl(TCG_COND_NE, t0, 0x0, misaligned);
- }
-
- if (rd != 0) {
- tcg_gen_movi_tl(cpu_gpr[rd], ctx->pc_succ_insn);
- }
- tcg_gen_lookup_and_goto_ptr();
-
- if (misaligned) {
- gen_set_label(misaligned);
- gen_exception_inst_addr_mis(ctx);
- }
- ctx->base.is_jmp = DISAS_NORETURN;
- break;
-
- default:
- gen_exception_illegal(ctx);
- break;
- }
- tcg_temp_free(t0);
-}
-
static void gen_branch(CPURISCVState *env, DisasContext *ctx, uint32_t opc,
int rs1, int rs2, target_long bimm)
{
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 21/35] target/riscv: Remove manual decoding from gen_branch()
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
We now utilizes argument-sets of decodetree such that no manual
decoding is necessary.
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
target/riscv/insn_trans/trans_rvi.inc.c | 46 +++++++++++++++++-------
target/riscv/translate.c | 47 -------------------------
2 files changed, 33 insertions(+), 60 deletions(-)
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
index 4d090d68e7..4f63b904a3 100644
--- a/target/riscv/insn_trans/trans_rvi.inc.c
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -72,41 +72,61 @@ static bool trans_jalr(DisasContext *ctx, arg_jalr *a)
return true;
}
-static bool trans_beq(DisasContext *ctx, arg_beq *a)
+static bool gen_branch(DisasContext *ctx, arg_b *a, TCGCond cond)
{
- gen_branch(ctx->env, ctx, OPC_RISC_BEQ, a->rs1, a->rs2, a->imm);
+ TCGLabel *l = gen_new_label();
+ TCGv source1, source2;
+ source1 = tcg_temp_new();
+ source2 = tcg_temp_new();
+ gen_get_gpr(source1, a->rs1);
+ gen_get_gpr(source2, a->rs2);
+
+ tcg_gen_brcond_tl(cond, source1, source2, l);
+ gen_goto_tb(ctx, 1, ctx->pc_succ_insn);
+ gen_set_label(l); /* branch taken */
+
+ if (!riscv_has_ext(ctx->env, RVC) && ((ctx->base.pc_next + a->imm) & 0x3)) {
+ /* misaligned */
+ gen_exception_inst_addr_mis(ctx);
+ } else {
+ gen_goto_tb(ctx, 0, ctx->base.pc_next + a->imm);
+ }
+ ctx->base.is_jmp = DISAS_NORETURN;
+
+ tcg_temp_free(source1);
+ tcg_temp_free(source2);
+
return true;
}
+static bool trans_beq(DisasContext *ctx, arg_beq *a)
+{
+ return gen_branch(ctx, a, TCG_COND_EQ);
+}
+
static bool trans_bne(DisasContext *ctx, arg_bne *a)
{
- gen_branch(ctx->env, ctx, OPC_RISC_BNE, a->rs1, a->rs2, a->imm);
- return true;
+ return gen_branch(ctx, a, TCG_COND_NE);
}
static bool trans_blt(DisasContext *ctx, arg_blt *a)
{
- gen_branch(ctx->env, ctx, OPC_RISC_BLT, a->rs1, a->rs2, a->imm);
- return true;
+ return gen_branch(ctx, a, TCG_COND_LT);
}
static bool trans_bge(DisasContext *ctx, arg_bge *a)
{
- gen_branch(ctx->env, ctx, OPC_RISC_BGE, a->rs1, a->rs2, a->imm);
- return true;
+ return gen_branch(ctx, a, TCG_COND_GE);
}
static bool trans_bltu(DisasContext *ctx, arg_bltu *a)
{
- gen_branch(ctx->env, ctx, OPC_RISC_BLTU, a->rs1, a->rs2, a->imm);
- return true;
+ return gen_branch(ctx, a, TCG_COND_LTU);
}
static bool trans_bgeu(DisasContext *ctx, arg_bgeu *a)
{
-
- gen_branch(ctx->env, ctx, OPC_RISC_BGEU, a->rs1, a->rs2, a->imm);
- return true;
+ return gen_branch(ctx, a, TCG_COND_GEU);
}
static bool trans_lb(DisasContext *ctx, arg_lb *a)
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index b78e423d94..585378d1e1 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -489,53 +489,6 @@ static void gen_jal(CPURISCVState *env, DisasContext *ctx, int rd,
ctx->base.is_jmp = DISAS_NORETURN;
}
-static void gen_branch(CPURISCVState *env, DisasContext *ctx, uint32_t opc,
- int rs1, int rs2, target_long bimm)
-{
- TCGLabel *l = gen_new_label();
- TCGv source1, source2;
- source1 = tcg_temp_new();
- source2 = tcg_temp_new();
- gen_get_gpr(source1, rs1);
- gen_get_gpr(source2, rs2);
-
- switch (opc) {
- case OPC_RISC_BEQ:
- tcg_gen_brcond_tl(TCG_COND_EQ, source1, source2, l);
- break;
- case OPC_RISC_BNE:
- tcg_gen_brcond_tl(TCG_COND_NE, source1, source2, l);
- break;
- case OPC_RISC_BLT:
- tcg_gen_brcond_tl(TCG_COND_LT, source1, source2, l);
- break;
- case OPC_RISC_BGE:
- tcg_gen_brcond_tl(TCG_COND_GE, source1, source2, l);
- break;
- case OPC_RISC_BLTU:
- tcg_gen_brcond_tl(TCG_COND_LTU, source1, source2, l);
- break;
- case OPC_RISC_BGEU:
- tcg_gen_brcond_tl(TCG_COND_GEU, source1, source2, l);
- break;
- default:
- gen_exception_illegal(ctx);
- return;
- }
- tcg_temp_free(source1);
- tcg_temp_free(source2);
-
- gen_goto_tb(ctx, 1, ctx->pc_succ_insn);
- gen_set_label(l); /* branch taken */
- if (!riscv_has_ext(env, RVC) && ((ctx->base.pc_next + bimm) & 0x3)) {
- /* misaligned */
- gen_exception_inst_addr_mis(ctx);
- } else {
- gen_goto_tb(ctx, 0, ctx->base.pc_next + bimm);
- }
- ctx->base.is_jmp = DISAS_NORETURN;
-}
-
static void gen_load(DisasContext *ctx, uint32_t opc, int rd, int rs1,
target_long imm)
{
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 21/35] target/riscv: Remove manual decoding from gen_branch()
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
We now utilizes argument-sets of decodetree such that no manual
decoding is necessary.
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
target/riscv/insn_trans/trans_rvi.inc.c | 46 +++++++++++++++++-------
target/riscv/translate.c | 47 -------------------------
2 files changed, 33 insertions(+), 60 deletions(-)
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
index 4d090d68e7..4f63b904a3 100644
--- a/target/riscv/insn_trans/trans_rvi.inc.c
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -72,41 +72,61 @@ static bool trans_jalr(DisasContext *ctx, arg_jalr *a)
return true;
}
-static bool trans_beq(DisasContext *ctx, arg_beq *a)
+static bool gen_branch(DisasContext *ctx, arg_b *a, TCGCond cond)
{
- gen_branch(ctx->env, ctx, OPC_RISC_BEQ, a->rs1, a->rs2, a->imm);
+ TCGLabel *l = gen_new_label();
+ TCGv source1, source2;
+ source1 = tcg_temp_new();
+ source2 = tcg_temp_new();
+ gen_get_gpr(source1, a->rs1);
+ gen_get_gpr(source2, a->rs2);
+
+ tcg_gen_brcond_tl(cond, source1, source2, l);
+ gen_goto_tb(ctx, 1, ctx->pc_succ_insn);
+ gen_set_label(l); /* branch taken */
+
+ if (!riscv_has_ext(ctx->env, RVC) && ((ctx->base.pc_next + a->imm) & 0x3)) {
+ /* misaligned */
+ gen_exception_inst_addr_mis(ctx);
+ } else {
+ gen_goto_tb(ctx, 0, ctx->base.pc_next + a->imm);
+ }
+ ctx->base.is_jmp = DISAS_NORETURN;
+
+ tcg_temp_free(source1);
+ tcg_temp_free(source2);
+
return true;
}
+static bool trans_beq(DisasContext *ctx, arg_beq *a)
+{
+ return gen_branch(ctx, a, TCG_COND_EQ);
+}
+
static bool trans_bne(DisasContext *ctx, arg_bne *a)
{
- gen_branch(ctx->env, ctx, OPC_RISC_BNE, a->rs1, a->rs2, a->imm);
- return true;
+ return gen_branch(ctx, a, TCG_COND_NE);
}
static bool trans_blt(DisasContext *ctx, arg_blt *a)
{
- gen_branch(ctx->env, ctx, OPC_RISC_BLT, a->rs1, a->rs2, a->imm);
- return true;
+ return gen_branch(ctx, a, TCG_COND_LT);
}
static bool trans_bge(DisasContext *ctx, arg_bge *a)
{
- gen_branch(ctx->env, ctx, OPC_RISC_BGE, a->rs1, a->rs2, a->imm);
- return true;
+ return gen_branch(ctx, a, TCG_COND_GE);
}
static bool trans_bltu(DisasContext *ctx, arg_bltu *a)
{
- gen_branch(ctx->env, ctx, OPC_RISC_BLTU, a->rs1, a->rs2, a->imm);
- return true;
+ return gen_branch(ctx, a, TCG_COND_LTU);
}
static bool trans_bgeu(DisasContext *ctx, arg_bgeu *a)
{
-
- gen_branch(ctx->env, ctx, OPC_RISC_BGEU, a->rs1, a->rs2, a->imm);
- return true;
+ return gen_branch(ctx, a, TCG_COND_GEU);
}
static bool trans_lb(DisasContext *ctx, arg_lb *a)
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index b78e423d94..585378d1e1 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -489,53 +489,6 @@ static void gen_jal(CPURISCVState *env, DisasContext *ctx, int rd,
ctx->base.is_jmp = DISAS_NORETURN;
}
-static void gen_branch(CPURISCVState *env, DisasContext *ctx, uint32_t opc,
- int rs1, int rs2, target_long bimm)
-{
- TCGLabel *l = gen_new_label();
- TCGv source1, source2;
- source1 = tcg_temp_new();
- source2 = tcg_temp_new();
- gen_get_gpr(source1, rs1);
- gen_get_gpr(source2, rs2);
-
- switch (opc) {
- case OPC_RISC_BEQ:
- tcg_gen_brcond_tl(TCG_COND_EQ, source1, source2, l);
- break;
- case OPC_RISC_BNE:
- tcg_gen_brcond_tl(TCG_COND_NE, source1, source2, l);
- break;
- case OPC_RISC_BLT:
- tcg_gen_brcond_tl(TCG_COND_LT, source1, source2, l);
- break;
- case OPC_RISC_BGE:
- tcg_gen_brcond_tl(TCG_COND_GE, source1, source2, l);
- break;
- case OPC_RISC_BLTU:
- tcg_gen_brcond_tl(TCG_COND_LTU, source1, source2, l);
- break;
- case OPC_RISC_BGEU:
- tcg_gen_brcond_tl(TCG_COND_GEU, source1, source2, l);
- break;
- default:
- gen_exception_illegal(ctx);
- return;
- }
- tcg_temp_free(source1);
- tcg_temp_free(source2);
-
- gen_goto_tb(ctx, 1, ctx->pc_succ_insn);
- gen_set_label(l); /* branch taken */
- if (!riscv_has_ext(env, RVC) && ((ctx->base.pc_next + bimm) & 0x3)) {
- /* misaligned */
- gen_exception_inst_addr_mis(ctx);
- } else {
- gen_goto_tb(ctx, 0, ctx->base.pc_next + bimm);
- }
- ctx->base.is_jmp = DISAS_NORETURN;
-}
-
static void gen_load(DisasContext *ctx, uint32_t opc, int rd, int rs1,
target_long imm)
{
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 22/35] target/riscv: Remove manual decoding from gen_load()
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
With decodetree we don't need to convert RISC-V opcodes into to MemOps
as the old gen_load() did.
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de><Paste>
---
v2 -> v3:
- Since trans_c_flw_ld and trans_c_fsw_sd still rely on the old decoder
we need to keep gen_load(). Thus we renamed it to gen_load_c.
target/riscv/insn_trans/trans_rvi.inc.c | 35 +++++++++++++++----------
target/riscv/translate.c | 6 +++--
2 files changed, 25 insertions(+), 16 deletions(-)
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
index 4f63b904a3..2816c77375 100644
--- a/target/riscv/insn_trans/trans_rvi.inc.c
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -129,34 +129,43 @@ static bool trans_bgeu(DisasContext *ctx, arg_bgeu *a)
return gen_branch(ctx, a, TCG_COND_GEU);
}
-static bool trans_lb(DisasContext *ctx, arg_lb *a)
+static bool gen_load(DisasContext *ctx, arg_lb *a, int memop)
{
- gen_load(ctx, OPC_RISC_LB, a->rd, a->rs1, a->imm);
+ TCGv t0 = tcg_temp_new();
+ TCGv t1 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+ tcg_gen_addi_tl(t0, t0, a->imm);
+
+ tcg_gen_qemu_ld_tl(t1, t0, ctx->mem_idx, memop);
+ gen_set_gpr(a->rd, t1);
+ tcg_temp_free(t0);
+ tcg_temp_free(t1);
return true;
}
+static bool trans_lb(DisasContext *ctx, arg_lb *a)
+{
+ return gen_load(ctx, a, MO_SB);
+}
+
static bool trans_lh(DisasContext *ctx, arg_lh *a)
{
- gen_load(ctx, OPC_RISC_LH, a->rd, a->rs1, a->imm);
- return true;
+ return gen_load(ctx, a, MO_TESW);
}
static bool trans_lw(DisasContext *ctx, arg_lw *a)
{
- gen_load(ctx, OPC_RISC_LW, a->rd, a->rs1, a->imm);
- return true;
+ return gen_load(ctx, a, MO_TESL);
}
static bool trans_lbu(DisasContext *ctx, arg_lbu *a)
{
- gen_load(ctx, OPC_RISC_LBU, a->rd, a->rs1, a->imm);
- return true;
+ return gen_load(ctx, a, MO_UB);
}
static bool trans_lhu(DisasContext *ctx, arg_lhu *a)
{
- gen_load(ctx, OPC_RISC_LHU, a->rd, a->rs1, a->imm);
- return true;
+ return gen_load(ctx, a, MO_TEUW);
}
static bool trans_sb(DisasContext *ctx, arg_sb *a)
@@ -180,14 +189,12 @@ static bool trans_sw(DisasContext *ctx, arg_sw *a)
#ifdef TARGET_RISCV64
static bool trans_lwu(DisasContext *ctx, arg_lwu *a)
{
- gen_load(ctx, OPC_RISC_LWU, a->rd, a->rs1, a->imm);
- return true;
+ return gen_load(ctx, a, MO_TEUL);
}
static bool trans_ld(DisasContext *ctx, arg_ld *a)
{
- gen_load(ctx, OPC_RISC_LD, a->rd, a->rs1, a->imm);
- return true;
+ return gen_load(ctx, a, MO_TEQ);
}
static bool trans_sd(DisasContext *ctx, arg_sd *a)
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 585378d1e1..605742989d 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -489,7 +489,8 @@ static void gen_jal(CPURISCVState *env, DisasContext *ctx, int rd,
ctx->base.is_jmp = DISAS_NORETURN;
}
-static void gen_load(DisasContext *ctx, uint32_t opc, int rd, int rs1,
+#ifdef TARGET_RISCV64
+static void gen_load_c(DisasContext *ctx, uint32_t opc, int rd, int rs1,
target_long imm)
{
TCGv t0 = tcg_temp_new();
@@ -508,6 +509,7 @@ static void gen_load(DisasContext *ctx, uint32_t opc, int rd, int rs1,
tcg_temp_free(t0);
tcg_temp_free(t1);
}
+#endif
static void gen_store(DisasContext *ctx, uint32_t opc, int rs1, int rs2,
target_long imm)
@@ -641,7 +643,7 @@ static void decode_RV32_64C0(DisasContext *ctx)
case 3:
#if defined(TARGET_RISCV64)
/* C.LD(RV64/128) -> ld rd', offset[7:3](rs1')*/
- gen_load(ctx, OPC_RISC_LD, rd_rs2, rs1s,
+ gen_load_c(ctx, OPC_RISC_LD, rd_rs2, rs1s,
GET_C_LD_IMM(ctx->opcode));
#else
/* C.FLW (RV32) -> flw rd', offset[6:2](rs1')*/
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 22/35] target/riscv: Remove manual decoding from gen_load()
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
With decodetree we don't need to convert RISC-V opcodes into to MemOps
as the old gen_load() did.
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de><Paste>
---
v2 -> v3:
- Since trans_c_flw_ld and trans_c_fsw_sd still rely on the old decoder
we need to keep gen_load(). Thus we renamed it to gen_load_c.
target/riscv/insn_trans/trans_rvi.inc.c | 35 +++++++++++++++----------
target/riscv/translate.c | 6 +++--
2 files changed, 25 insertions(+), 16 deletions(-)
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
index 4f63b904a3..2816c77375 100644
--- a/target/riscv/insn_trans/trans_rvi.inc.c
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -129,34 +129,43 @@ static bool trans_bgeu(DisasContext *ctx, arg_bgeu *a)
return gen_branch(ctx, a, TCG_COND_GEU);
}
-static bool trans_lb(DisasContext *ctx, arg_lb *a)
+static bool gen_load(DisasContext *ctx, arg_lb *a, int memop)
{
- gen_load(ctx, OPC_RISC_LB, a->rd, a->rs1, a->imm);
+ TCGv t0 = tcg_temp_new();
+ TCGv t1 = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+ tcg_gen_addi_tl(t0, t0, a->imm);
+
+ tcg_gen_qemu_ld_tl(t1, t0, ctx->mem_idx, memop);
+ gen_set_gpr(a->rd, t1);
+ tcg_temp_free(t0);
+ tcg_temp_free(t1);
return true;
}
+static bool trans_lb(DisasContext *ctx, arg_lb *a)
+{
+ return gen_load(ctx, a, MO_SB);
+}
+
static bool trans_lh(DisasContext *ctx, arg_lh *a)
{
- gen_load(ctx, OPC_RISC_LH, a->rd, a->rs1, a->imm);
- return true;
+ return gen_load(ctx, a, MO_TESW);
}
static bool trans_lw(DisasContext *ctx, arg_lw *a)
{
- gen_load(ctx, OPC_RISC_LW, a->rd, a->rs1, a->imm);
- return true;
+ return gen_load(ctx, a, MO_TESL);
}
static bool trans_lbu(DisasContext *ctx, arg_lbu *a)
{
- gen_load(ctx, OPC_RISC_LBU, a->rd, a->rs1, a->imm);
- return true;
+ return gen_load(ctx, a, MO_UB);
}
static bool trans_lhu(DisasContext *ctx, arg_lhu *a)
{
- gen_load(ctx, OPC_RISC_LHU, a->rd, a->rs1, a->imm);
- return true;
+ return gen_load(ctx, a, MO_TEUW);
}
static bool trans_sb(DisasContext *ctx, arg_sb *a)
@@ -180,14 +189,12 @@ static bool trans_sw(DisasContext *ctx, arg_sw *a)
#ifdef TARGET_RISCV64
static bool trans_lwu(DisasContext *ctx, arg_lwu *a)
{
- gen_load(ctx, OPC_RISC_LWU, a->rd, a->rs1, a->imm);
- return true;
+ return gen_load(ctx, a, MO_TEUL);
}
static bool trans_ld(DisasContext *ctx, arg_ld *a)
{
- gen_load(ctx, OPC_RISC_LD, a->rd, a->rs1, a->imm);
- return true;
+ return gen_load(ctx, a, MO_TEQ);
}
static bool trans_sd(DisasContext *ctx, arg_sd *a)
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 585378d1e1..605742989d 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -489,7 +489,8 @@ static void gen_jal(CPURISCVState *env, DisasContext *ctx, int rd,
ctx->base.is_jmp = DISAS_NORETURN;
}
-static void gen_load(DisasContext *ctx, uint32_t opc, int rd, int rs1,
+#ifdef TARGET_RISCV64
+static void gen_load_c(DisasContext *ctx, uint32_t opc, int rd, int rs1,
target_long imm)
{
TCGv t0 = tcg_temp_new();
@@ -508,6 +509,7 @@ static void gen_load(DisasContext *ctx, uint32_t opc, int rd, int rs1,
tcg_temp_free(t0);
tcg_temp_free(t1);
}
+#endif
static void gen_store(DisasContext *ctx, uint32_t opc, int rs1, int rs2,
target_long imm)
@@ -641,7 +643,7 @@ static void decode_RV32_64C0(DisasContext *ctx)
case 3:
#if defined(TARGET_RISCV64)
/* C.LD(RV64/128) -> ld rd', offset[7:3](rs1')*/
- gen_load(ctx, OPC_RISC_LD, rd_rs2, rs1s,
+ gen_load_c(ctx, OPC_RISC_LD, rd_rs2, rs1s,
GET_C_LD_IMM(ctx->opcode));
#else
/* C.FLW (RV32) -> flw rd', offset[6:2](rs1')*/
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 23/35] target/riscv: Remove manual decoding from gen_store()
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
With decodetree we don't need to convert RISC-V opcodes into to MemOps
as the old gen_store() did.
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- Since trans_c_flw_ld and trans_c_fsw_sd still rely on the old decoder
we need to keep gen_store(). Thus we renamed it to gen_store_c.
- tcg_memop_lookup is now only used by TARGET_RISCV64 insn so we wrapped
it in a ifdef
target/riscv/insn_trans/trans_rvi.inc.c | 27 +++++++++++++++++--------
target/riscv/translate.c | 8 +++++---
2 files changed, 24 insertions(+), 11 deletions(-)
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
index 2816c77375..cf05176ba2 100644
--- a/target/riscv/insn_trans/trans_rvi.inc.c
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -168,22 +168,34 @@ static bool trans_lhu(DisasContext *ctx, arg_lhu *a)
return gen_load(ctx, a, MO_TEUW);
}
-static bool trans_sb(DisasContext *ctx, arg_sb *a)
+static bool gen_store(DisasContext *ctx, arg_sb *a, int memop)
{
- gen_store(ctx, OPC_RISC_SB, a->rs1, a->rs2, a->imm);
+ TCGv t0 = tcg_temp_new();
+ TCGv dat = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+ tcg_gen_addi_tl(t0, t0, a->imm);
+ gen_get_gpr(dat, a->rs2);
+
+ tcg_gen_qemu_st_tl(dat, t0, ctx->mem_idx, memop);
+ tcg_temp_free(t0);
+ tcg_temp_free(dat);
return true;
}
+
+static bool trans_sb(DisasContext *ctx, arg_sb *a)
+{
+ return gen_store(ctx, a, MO_SB);
+}
+
static bool trans_sh(DisasContext *ctx, arg_sh *a)
{
- gen_store(ctx, OPC_RISC_SH, a->rs1, a->rs2, a->imm);
- return true;
+ return gen_store(ctx, a, MO_TESW);
}
static bool trans_sw(DisasContext *ctx, arg_sw *a)
{
- gen_store(ctx, OPC_RISC_SW, a->rs1, a->rs2, a->imm);
- return true;
+ return gen_store(ctx, a, MO_TESL);
}
#ifdef TARGET_RISCV64
@@ -199,8 +211,7 @@ static bool trans_ld(DisasContext *ctx, arg_ld *a)
static bool trans_sd(DisasContext *ctx, arg_sd *a)
{
- gen_store(ctx, OPC_RISC_SD, a->rs1, a->rs2, a->imm);
- return true;
+ return gen_store(ctx, a, MO_TEQ);
}
#endif
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 605742989d..0741a543aa 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -55,6 +55,7 @@ typedef struct DisasContext {
CPURISCVState *env;
} DisasContext;
+#ifdef TARGET_RISCV64
/* convert riscv funct3 to qemu memop for load/store */
static const int tcg_memop_lookup[8] = {
[0 ... 7] = -1,
@@ -68,6 +69,7 @@ static const int tcg_memop_lookup[8] = {
[6] = MO_TEUL,
#endif
};
+#endif
#ifdef TARGET_RISCV64
#define CASE_OP_32_64(X) case X: case glue(X, W)
@@ -509,9 +511,8 @@ static void gen_load_c(DisasContext *ctx, uint32_t opc, int rd, int rs1,
tcg_temp_free(t0);
tcg_temp_free(t1);
}
-#endif
-static void gen_store(DisasContext *ctx, uint32_t opc, int rs1, int rs2,
+static void gen_store_c(DisasContext *ctx, uint32_t opc, int rs1, int rs2,
target_long imm)
{
TCGv t0 = tcg_temp_new();
@@ -530,6 +531,7 @@ static void gen_store(DisasContext *ctx, uint32_t opc, int rs1, int rs2,
tcg_temp_free(t0);
tcg_temp_free(dat);
}
+#endif
#ifdef TARGET_RISCV32
static void gen_fp_load(DisasContext *ctx, uint32_t opc, int rd,
@@ -654,7 +656,7 @@ static void decode_RV32_64C0(DisasContext *ctx)
case 7:
#if defined(TARGET_RISCV64)
/* C.SD (RV64/128) -> sd rs2', offset[7:3](rs1')*/
- gen_store(ctx, OPC_RISC_SD, rs1s, rd_rs2,
+ gen_store_c(ctx, OPC_RISC_SD, rs1s, rd_rs2,
GET_C_LD_IMM(ctx->opcode));
#else
/* C.FSW (RV32) -> fsw rs2', offset[6:2](rs1')*/
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 23/35] target/riscv: Remove manual decoding from gen_store()
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
With decodetree we don't need to convert RISC-V opcodes into to MemOps
as the old gen_store() did.
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- Since trans_c_flw_ld and trans_c_fsw_sd still rely on the old decoder
we need to keep gen_store(). Thus we renamed it to gen_store_c.
- tcg_memop_lookup is now only used by TARGET_RISCV64 insn so we wrapped
it in a ifdef
target/riscv/insn_trans/trans_rvi.inc.c | 27 +++++++++++++++++--------
target/riscv/translate.c | 8 +++++---
2 files changed, 24 insertions(+), 11 deletions(-)
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
index 2816c77375..cf05176ba2 100644
--- a/target/riscv/insn_trans/trans_rvi.inc.c
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -168,22 +168,34 @@ static bool trans_lhu(DisasContext *ctx, arg_lhu *a)
return gen_load(ctx, a, MO_TEUW);
}
-static bool trans_sb(DisasContext *ctx, arg_sb *a)
+static bool gen_store(DisasContext *ctx, arg_sb *a, int memop)
{
- gen_store(ctx, OPC_RISC_SB, a->rs1, a->rs2, a->imm);
+ TCGv t0 = tcg_temp_new();
+ TCGv dat = tcg_temp_new();
+ gen_get_gpr(t0, a->rs1);
+ tcg_gen_addi_tl(t0, t0, a->imm);
+ gen_get_gpr(dat, a->rs2);
+
+ tcg_gen_qemu_st_tl(dat, t0, ctx->mem_idx, memop);
+ tcg_temp_free(t0);
+ tcg_temp_free(dat);
return true;
}
+
+static bool trans_sb(DisasContext *ctx, arg_sb *a)
+{
+ return gen_store(ctx, a, MO_SB);
+}
+
static bool trans_sh(DisasContext *ctx, arg_sh *a)
{
- gen_store(ctx, OPC_RISC_SH, a->rs1, a->rs2, a->imm);
- return true;
+ return gen_store(ctx, a, MO_TESW);
}
static bool trans_sw(DisasContext *ctx, arg_sw *a)
{
- gen_store(ctx, OPC_RISC_SW, a->rs1, a->rs2, a->imm);
- return true;
+ return gen_store(ctx, a, MO_TESL);
}
#ifdef TARGET_RISCV64
@@ -199,8 +211,7 @@ static bool trans_ld(DisasContext *ctx, arg_ld *a)
static bool trans_sd(DisasContext *ctx, arg_sd *a)
{
- gen_store(ctx, OPC_RISC_SD, a->rs1, a->rs2, a->imm);
- return true;
+ return gen_store(ctx, a, MO_TEQ);
}
#endif
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 605742989d..0741a543aa 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -55,6 +55,7 @@ typedef struct DisasContext {
CPURISCVState *env;
} DisasContext;
+#ifdef TARGET_RISCV64
/* convert riscv funct3 to qemu memop for load/store */
static const int tcg_memop_lookup[8] = {
[0 ... 7] = -1,
@@ -68,6 +69,7 @@ static const int tcg_memop_lookup[8] = {
[6] = MO_TEUL,
#endif
};
+#endif
#ifdef TARGET_RISCV64
#define CASE_OP_32_64(X) case X: case glue(X, W)
@@ -509,9 +511,8 @@ static void gen_load_c(DisasContext *ctx, uint32_t opc, int rd, int rs1,
tcg_temp_free(t0);
tcg_temp_free(t1);
}
-#endif
-static void gen_store(DisasContext *ctx, uint32_t opc, int rs1, int rs2,
+static void gen_store_c(DisasContext *ctx, uint32_t opc, int rs1, int rs2,
target_long imm)
{
TCGv t0 = tcg_temp_new();
@@ -530,6 +531,7 @@ static void gen_store(DisasContext *ctx, uint32_t opc, int rs1, int rs2,
tcg_temp_free(t0);
tcg_temp_free(dat);
}
+#endif
#ifdef TARGET_RISCV32
static void gen_fp_load(DisasContext *ctx, uint32_t opc, int rd,
@@ -654,7 +656,7 @@ static void decode_RV32_64C0(DisasContext *ctx)
case 7:
#if defined(TARGET_RISCV64)
/* C.SD (RV64/128) -> sd rs2', offset[7:3](rs1')*/
- gen_store(ctx, OPC_RISC_SD, rs1s, rd_rs2,
+ gen_store_c(ctx, OPC_RISC_SD, rs1s, rd_rs2,
GET_C_LD_IMM(ctx->opcode));
#else
/* C.FSW (RV32) -> fsw rs2', offset[6:2](rs1')*/
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 24/35] target/riscv: Move gen_arith_imm() decoding into trans_* functions
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
gen_arith_imm() does a lot of decoding manually, which was hard to read
in
case of the shift instructions and is not necessary anymore with
decodetree.
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- trans_srli/srai now use tcg_gen_shri/srai_tl
- trans_addiw uses its own gen_addiw function which properly extends the
result
- &arith_imm -> &i
target/riscv/insn32.decode | 3 +-
target/riscv/insn_trans/trans_rvi.inc.c | 89 +++++++++++++++-----
target/riscv/translate.c | 107 ++++++------------------
3 files changed, 99 insertions(+), 100 deletions(-)
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index 379a5d5438..70e00412b6 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -35,12 +35,13 @@
# Argument sets:
&b imm rs2 rs1
+&i imm rs1 rd
&shift shamt rs1 rd
&atomic aq rl rs2 rs1 rd
# Formats 32:
@r ....... ..... ..... ... ..... ....... %rs2 %rs1 %rd
-@i ............ ..... ... ..... ....... imm=%imm_i %rs1 %rd
+@i ............ ..... ... ..... ....... &i imm=%imm_i %rs1 %rd
@b ....... ..... ..... ... ..... ....... &b imm=%imm_b %rs2 %rs1
@s ....... ..... ..... ... ..... ....... imm=%imm_s %rs2 %rs1
@u .................... ..... ....... imm=%imm_u %rd
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
index cf05176ba2..8bd2752f31 100644
--- a/target/riscv/insn_trans/trans_rvi.inc.c
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -217,52 +217,85 @@ static bool trans_sd(DisasContext *ctx, arg_sd *a)
static bool trans_addi(DisasContext *ctx, arg_addi *a)
{
- gen_arith_imm(ctx, OPC_RISC_ADDI, a->rd, a->rs1, a->imm);
- return true;
+ return gen_arith_imm(ctx, a, &tcg_gen_add_tl);
}
static bool trans_slti(DisasContext *ctx, arg_slti *a)
{
- gen_arith_imm(ctx, OPC_RISC_SLTI, a->rd, a->rs1, a->imm);
+ TCGv source1;
+ source1 = tcg_temp_new();
+ gen_get_gpr(source1, a->rs1);
+
+ tcg_gen_setcondi_tl(TCG_COND_LT, source1, source1, a->imm);
+
+ gen_set_gpr(a->rd, source1);
+ tcg_temp_free(source1);
return true;
}
static bool trans_sltiu(DisasContext *ctx, arg_sltiu *a)
{
- gen_arith_imm(ctx, OPC_RISC_SLTIU, a->rd, a->rs1, a->imm);
+ TCGv source1;
+ source1 = tcg_temp_new();
+ gen_get_gpr(source1, a->rs1);
+
+ tcg_gen_setcondi_tl(TCG_COND_LTU, source1, source1, a->imm);
+
+ gen_set_gpr(a->rd, source1);
+ tcg_temp_free(source1);
return true;
}
static bool trans_xori(DisasContext *ctx, arg_xori *a)
{
- gen_arith_imm(ctx, OPC_RISC_XORI, a->rd, a->rs1, a->imm);
- return true;
+ return gen_arith_imm(ctx, a, &tcg_gen_xor_tl);
}
static bool trans_ori(DisasContext *ctx, arg_ori *a)
{
- gen_arith_imm(ctx, OPC_RISC_ORI, a->rd, a->rs1, a->imm);
- return true;
+ return gen_arith_imm(ctx, a, &tcg_gen_or_tl);
}
static bool trans_andi(DisasContext *ctx, arg_andi *a)
{
- gen_arith_imm(ctx, OPC_RISC_ANDI, a->rd, a->rs1, a->imm);
- return true;
+ return gen_arith_imm(ctx, a, &tcg_gen_and_tl);
}
static bool trans_slli(DisasContext *ctx, arg_slli *a)
{
- gen_arith_imm(ctx, OPC_RISC_SLLI, a->rd, a->rs1, a->shamt);
+ if (a->rd != 0) {
+ TCGv t = tcg_temp_new();
+ gen_get_gpr(t, a->rs1);
+
+ if (a->shamt >= TARGET_LONG_BITS) {
+ return false;
+ }
+ tcg_gen_shli_tl(t, t, a->shamt);
+
+ gen_set_gpr(a->rd, t);
+ tcg_temp_free(t);
+ } /* NOP otherwise */
return true;
}
static bool trans_srli(DisasContext *ctx, arg_srli *a)
{
- gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_I, a->rd, a->rs1, a->shamt);
+ if (a->rd != 0) {
+ TCGv t = tcg_temp_new();
+ gen_get_gpr(t, a->rs1);
+ tcg_gen_shri_tl(t, t, a->shamt);
+ gen_set_gpr(a->rd, t);
+ tcg_temp_free(t);
+ } /* NOP otherwise */
return true;
}
static bool trans_srai(DisasContext *ctx, arg_srai *a)
{
- gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_I, a->rd, a->rs1, a->shamt | 0x400);
+ if (a->rd != 0) {
+ TCGv t = tcg_temp_new();
+ gen_get_gpr(t, a->rs1);
+ tcg_gen_sari_tl(t, t, a->shamt);
+ gen_set_gpr(a->rd, t);
+ tcg_temp_free(t);
+ } /* NOP otherwise */
return true;
}
@@ -329,26 +362,44 @@ static bool trans_and(DisasContext *ctx, arg_and *a)
#ifdef TARGET_RISCV64
static bool trans_addiw(DisasContext *ctx, arg_addiw *a)
{
- gen_arith_imm(ctx, OPC_RISC_ADDIW, a->rd, a->rs1, a->imm);
- return true;
+ return gen_arith_imm(ctx, a, &gen_addiw);
}
static bool trans_slliw(DisasContext *ctx, arg_slliw *a)
{
- gen_arith_imm(ctx, OPC_RISC_SLLIW, a->rd, a->rs1, a->shamt);
+ TCGv source1;
+ source1 = tcg_temp_new();
+ gen_get_gpr(source1, a->rs1);
+
+ tcg_gen_shli_tl(source1, source1, a->shamt);
+ tcg_gen_ext32s_tl(source1, source1);
+ gen_set_gpr(a->rd, source1);
+
+ tcg_temp_free(source1);
return true;
}
static bool trans_srliw(DisasContext *ctx, arg_srliw *a)
{
- gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_IW, a->rd, a->rs1, a->shamt);
+ TCGv t = tcg_temp_new();
+ gen_get_gpr(t, a->rs1);
+ tcg_gen_extract_tl(t, t, a->shamt, 32 - a->shamt);
+ /* sign-extend for W instructions */
+ tcg_gen_ext32s_tl(t, t);
+ gen_set_gpr(a->rd, t);
+ tcg_temp_free(t);
return true;
}
static bool trans_sraiw(DisasContext *ctx, arg_sraiw *a)
{
- gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_IW , a->rd, a->rs1,
- a->shamt | 0x400);
+ TCGv t = tcg_temp_new();
+ gen_get_gpr(t, a->rs1);
+ tcg_gen_sextract_tl(t, t, a->shamt, 32 - a->shamt);
+ /* sign-extend for W instructions */
+ tcg_gen_ext32s_tl(t, t);
+ gen_set_gpr(a->rd, t);
+ tcg_temp_free(t);
return true;
}
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 0741a543aa..de60686df4 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -390,86 +390,6 @@ static void gen_arith(DisasContext *ctx, uint32_t opc, int rd, int rs1,
tcg_temp_free(source2);
}
-static void gen_arith_imm(DisasContext *ctx, uint32_t opc, int rd,
- int rs1, target_long imm)
-{
- TCGv source1 = tcg_temp_new();
- int shift_len = TARGET_LONG_BITS;
- int shift_a;
-
- gen_get_gpr(source1, rs1);
-
- switch (opc) {
- case OPC_RISC_ADDI:
-#if defined(TARGET_RISCV64)
- case OPC_RISC_ADDIW:
-#endif
- tcg_gen_addi_tl(source1, source1, imm);
- break;
- case OPC_RISC_SLTI:
- tcg_gen_setcondi_tl(TCG_COND_LT, source1, source1, imm);
- break;
- case OPC_RISC_SLTIU:
- tcg_gen_setcondi_tl(TCG_COND_LTU, source1, source1, imm);
- break;
- case OPC_RISC_XORI:
- tcg_gen_xori_tl(source1, source1, imm);
- break;
- case OPC_RISC_ORI:
- tcg_gen_ori_tl(source1, source1, imm);
- break;
- case OPC_RISC_ANDI:
- tcg_gen_andi_tl(source1, source1, imm);
- break;
-#if defined(TARGET_RISCV64)
- case OPC_RISC_SLLIW:
- shift_len = 32;
- /* FALLTHRU */
-#endif
- case OPC_RISC_SLLI:
- if (imm >= shift_len) {
- goto do_illegal;
- }
- tcg_gen_shli_tl(source1, source1, imm);
- break;
-#if defined(TARGET_RISCV64)
- case OPC_RISC_SHIFT_RIGHT_IW:
- shift_len = 32;
- /* FALLTHRU */
-#endif
- case OPC_RISC_SHIFT_RIGHT_I:
- /* differentiate on IMM */
- shift_a = imm & 0x400;
- imm &= 0x3ff;
- if (imm >= shift_len) {
- goto do_illegal;
- }
- if (imm != 0) {
- if (shift_a) {
- /* SRAI[W] */
- tcg_gen_sextract_tl(source1, source1, imm, shift_len - imm);
- } else {
- /* SRLI[W] */
- tcg_gen_extract_tl(source1, source1, imm, shift_len - imm);
- }
- /* No further sign-extension needed for W instructions. */
- opc &= ~0x8;
- }
- break;
- default:
- do_illegal:
- gen_exception_illegal(ctx);
- return;
- }
-
- if (opc & 0x8) { /* sign-extend for W instructions */
- tcg_gen_ext32s_tl(source1, source1);
- }
-
- gen_set_gpr(rd, source1);
- tcg_temp_free(source1);
-}
-
static void gen_jal(CPURISCVState *env, DisasContext *ctx, int rd,
target_ulong imm)
{
@@ -697,6 +617,33 @@ static int ex_rvc_register(int reg)
bool decode_insn32(DisasContext *ctx, uint32_t insn);
/* Include the auto-generated decoder for 32 bit insn */
#include "decode_insn32.inc.c"
+
+static bool gen_arith_imm(DisasContext *ctx, arg_i *a,
+ void(*func)(TCGv, TCGv, TCGv))
+{
+ TCGv source1, source2;
+ source1 = tcg_temp_new();
+ source2 = tcg_temp_new();
+
+ gen_get_gpr(source1, a->rs1);
+ tcg_gen_movi_tl(source2, a->imm);
+
+ (*func)(source1, source1, source2);
+
+ gen_set_gpr(a->rd, source1);
+ tcg_temp_free(source1);
+ tcg_temp_free(source2);
+ return true;
+}
+
+#ifdef TARGET_RISCV64
+static void gen_addiw(TCGv ret, TCGv arg1, TCGv arg2)
+{
+ tcg_gen_add_tl(ret, arg1, arg2);
+ tcg_gen_ext32s_tl(ret, ret);
+}
+#endif
+
/* Include insn module translation function */
#include "insn_trans/trans_rvi.inc.c"
#include "insn_trans/trans_rvm.inc.c"
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 24/35] target/riscv: Move gen_arith_imm() decoding into trans_* functions
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
gen_arith_imm() does a lot of decoding manually, which was hard to read
in
case of the shift instructions and is not necessary anymore with
decodetree.
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- trans_srli/srai now use tcg_gen_shri/srai_tl
- trans_addiw uses its own gen_addiw function which properly extends the
result
- &arith_imm -> &i
target/riscv/insn32.decode | 3 +-
target/riscv/insn_trans/trans_rvi.inc.c | 89 +++++++++++++++-----
target/riscv/translate.c | 107 ++++++------------------
3 files changed, 99 insertions(+), 100 deletions(-)
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index 379a5d5438..70e00412b6 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -35,12 +35,13 @@
# Argument sets:
&b imm rs2 rs1
+&i imm rs1 rd
&shift shamt rs1 rd
&atomic aq rl rs2 rs1 rd
# Formats 32:
@r ....... ..... ..... ... ..... ....... %rs2 %rs1 %rd
-@i ............ ..... ... ..... ....... imm=%imm_i %rs1 %rd
+@i ............ ..... ... ..... ....... &i imm=%imm_i %rs1 %rd
@b ....... ..... ..... ... ..... ....... &b imm=%imm_b %rs2 %rs1
@s ....... ..... ..... ... ..... ....... imm=%imm_s %rs2 %rs1
@u .................... ..... ....... imm=%imm_u %rd
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
index cf05176ba2..8bd2752f31 100644
--- a/target/riscv/insn_trans/trans_rvi.inc.c
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -217,52 +217,85 @@ static bool trans_sd(DisasContext *ctx, arg_sd *a)
static bool trans_addi(DisasContext *ctx, arg_addi *a)
{
- gen_arith_imm(ctx, OPC_RISC_ADDI, a->rd, a->rs1, a->imm);
- return true;
+ return gen_arith_imm(ctx, a, &tcg_gen_add_tl);
}
static bool trans_slti(DisasContext *ctx, arg_slti *a)
{
- gen_arith_imm(ctx, OPC_RISC_SLTI, a->rd, a->rs1, a->imm);
+ TCGv source1;
+ source1 = tcg_temp_new();
+ gen_get_gpr(source1, a->rs1);
+
+ tcg_gen_setcondi_tl(TCG_COND_LT, source1, source1, a->imm);
+
+ gen_set_gpr(a->rd, source1);
+ tcg_temp_free(source1);
return true;
}
static bool trans_sltiu(DisasContext *ctx, arg_sltiu *a)
{
- gen_arith_imm(ctx, OPC_RISC_SLTIU, a->rd, a->rs1, a->imm);
+ TCGv source1;
+ source1 = tcg_temp_new();
+ gen_get_gpr(source1, a->rs1);
+
+ tcg_gen_setcondi_tl(TCG_COND_LTU, source1, source1, a->imm);
+
+ gen_set_gpr(a->rd, source1);
+ tcg_temp_free(source1);
return true;
}
static bool trans_xori(DisasContext *ctx, arg_xori *a)
{
- gen_arith_imm(ctx, OPC_RISC_XORI, a->rd, a->rs1, a->imm);
- return true;
+ return gen_arith_imm(ctx, a, &tcg_gen_xor_tl);
}
static bool trans_ori(DisasContext *ctx, arg_ori *a)
{
- gen_arith_imm(ctx, OPC_RISC_ORI, a->rd, a->rs1, a->imm);
- return true;
+ return gen_arith_imm(ctx, a, &tcg_gen_or_tl);
}
static bool trans_andi(DisasContext *ctx, arg_andi *a)
{
- gen_arith_imm(ctx, OPC_RISC_ANDI, a->rd, a->rs1, a->imm);
- return true;
+ return gen_arith_imm(ctx, a, &tcg_gen_and_tl);
}
static bool trans_slli(DisasContext *ctx, arg_slli *a)
{
- gen_arith_imm(ctx, OPC_RISC_SLLI, a->rd, a->rs1, a->shamt);
+ if (a->rd != 0) {
+ TCGv t = tcg_temp_new();
+ gen_get_gpr(t, a->rs1);
+
+ if (a->shamt >= TARGET_LONG_BITS) {
+ return false;
+ }
+ tcg_gen_shli_tl(t, t, a->shamt);
+
+ gen_set_gpr(a->rd, t);
+ tcg_temp_free(t);
+ } /* NOP otherwise */
return true;
}
static bool trans_srli(DisasContext *ctx, arg_srli *a)
{
- gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_I, a->rd, a->rs1, a->shamt);
+ if (a->rd != 0) {
+ TCGv t = tcg_temp_new();
+ gen_get_gpr(t, a->rs1);
+ tcg_gen_shri_tl(t, t, a->shamt);
+ gen_set_gpr(a->rd, t);
+ tcg_temp_free(t);
+ } /* NOP otherwise */
return true;
}
static bool trans_srai(DisasContext *ctx, arg_srai *a)
{
- gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_I, a->rd, a->rs1, a->shamt | 0x400);
+ if (a->rd != 0) {
+ TCGv t = tcg_temp_new();
+ gen_get_gpr(t, a->rs1);
+ tcg_gen_sari_tl(t, t, a->shamt);
+ gen_set_gpr(a->rd, t);
+ tcg_temp_free(t);
+ } /* NOP otherwise */
return true;
}
@@ -329,26 +362,44 @@ static bool trans_and(DisasContext *ctx, arg_and *a)
#ifdef TARGET_RISCV64
static bool trans_addiw(DisasContext *ctx, arg_addiw *a)
{
- gen_arith_imm(ctx, OPC_RISC_ADDIW, a->rd, a->rs1, a->imm);
- return true;
+ return gen_arith_imm(ctx, a, &gen_addiw);
}
static bool trans_slliw(DisasContext *ctx, arg_slliw *a)
{
- gen_arith_imm(ctx, OPC_RISC_SLLIW, a->rd, a->rs1, a->shamt);
+ TCGv source1;
+ source1 = tcg_temp_new();
+ gen_get_gpr(source1, a->rs1);
+
+ tcg_gen_shli_tl(source1, source1, a->shamt);
+ tcg_gen_ext32s_tl(source1, source1);
+ gen_set_gpr(a->rd, source1);
+
+ tcg_temp_free(source1);
return true;
}
static bool trans_srliw(DisasContext *ctx, arg_srliw *a)
{
- gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_IW, a->rd, a->rs1, a->shamt);
+ TCGv t = tcg_temp_new();
+ gen_get_gpr(t, a->rs1);
+ tcg_gen_extract_tl(t, t, a->shamt, 32 - a->shamt);
+ /* sign-extend for W instructions */
+ tcg_gen_ext32s_tl(t, t);
+ gen_set_gpr(a->rd, t);
+ tcg_temp_free(t);
return true;
}
static bool trans_sraiw(DisasContext *ctx, arg_sraiw *a)
{
- gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_IW , a->rd, a->rs1,
- a->shamt | 0x400);
+ TCGv t = tcg_temp_new();
+ gen_get_gpr(t, a->rs1);
+ tcg_gen_sextract_tl(t, t, a->shamt, 32 - a->shamt);
+ /* sign-extend for W instructions */
+ tcg_gen_ext32s_tl(t, t);
+ gen_set_gpr(a->rd, t);
+ tcg_temp_free(t);
return true;
}
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 0741a543aa..de60686df4 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -390,86 +390,6 @@ static void gen_arith(DisasContext *ctx, uint32_t opc, int rd, int rs1,
tcg_temp_free(source2);
}
-static void gen_arith_imm(DisasContext *ctx, uint32_t opc, int rd,
- int rs1, target_long imm)
-{
- TCGv source1 = tcg_temp_new();
- int shift_len = TARGET_LONG_BITS;
- int shift_a;
-
- gen_get_gpr(source1, rs1);
-
- switch (opc) {
- case OPC_RISC_ADDI:
-#if defined(TARGET_RISCV64)
- case OPC_RISC_ADDIW:
-#endif
- tcg_gen_addi_tl(source1, source1, imm);
- break;
- case OPC_RISC_SLTI:
- tcg_gen_setcondi_tl(TCG_COND_LT, source1, source1, imm);
- break;
- case OPC_RISC_SLTIU:
- tcg_gen_setcondi_tl(TCG_COND_LTU, source1, source1, imm);
- break;
- case OPC_RISC_XORI:
- tcg_gen_xori_tl(source1, source1, imm);
- break;
- case OPC_RISC_ORI:
- tcg_gen_ori_tl(source1, source1, imm);
- break;
- case OPC_RISC_ANDI:
- tcg_gen_andi_tl(source1, source1, imm);
- break;
-#if defined(TARGET_RISCV64)
- case OPC_RISC_SLLIW:
- shift_len = 32;
- /* FALLTHRU */
-#endif
- case OPC_RISC_SLLI:
- if (imm >= shift_len) {
- goto do_illegal;
- }
- tcg_gen_shli_tl(source1, source1, imm);
- break;
-#if defined(TARGET_RISCV64)
- case OPC_RISC_SHIFT_RIGHT_IW:
- shift_len = 32;
- /* FALLTHRU */
-#endif
- case OPC_RISC_SHIFT_RIGHT_I:
- /* differentiate on IMM */
- shift_a = imm & 0x400;
- imm &= 0x3ff;
- if (imm >= shift_len) {
- goto do_illegal;
- }
- if (imm != 0) {
- if (shift_a) {
- /* SRAI[W] */
- tcg_gen_sextract_tl(source1, source1, imm, shift_len - imm);
- } else {
- /* SRLI[W] */
- tcg_gen_extract_tl(source1, source1, imm, shift_len - imm);
- }
- /* No further sign-extension needed for W instructions. */
- opc &= ~0x8;
- }
- break;
- default:
- do_illegal:
- gen_exception_illegal(ctx);
- return;
- }
-
- if (opc & 0x8) { /* sign-extend for W instructions */
- tcg_gen_ext32s_tl(source1, source1);
- }
-
- gen_set_gpr(rd, source1);
- tcg_temp_free(source1);
-}
-
static void gen_jal(CPURISCVState *env, DisasContext *ctx, int rd,
target_ulong imm)
{
@@ -697,6 +617,33 @@ static int ex_rvc_register(int reg)
bool decode_insn32(DisasContext *ctx, uint32_t insn);
/* Include the auto-generated decoder for 32 bit insn */
#include "decode_insn32.inc.c"
+
+static bool gen_arith_imm(DisasContext *ctx, arg_i *a,
+ void(*func)(TCGv, TCGv, TCGv))
+{
+ TCGv source1, source2;
+ source1 = tcg_temp_new();
+ source2 = tcg_temp_new();
+
+ gen_get_gpr(source1, a->rs1);
+ tcg_gen_movi_tl(source2, a->imm);
+
+ (*func)(source1, source1, source2);
+
+ gen_set_gpr(a->rd, source1);
+ tcg_temp_free(source1);
+ tcg_temp_free(source2);
+ return true;
+}
+
+#ifdef TARGET_RISCV64
+static void gen_addiw(TCGv ret, TCGv arg1, TCGv arg2)
+{
+ tcg_gen_add_tl(ret, arg1, arg2);
+ tcg_gen_ext32s_tl(ret, ret);
+}
+#endif
+
/* Include insn module translation function */
#include "insn_trans/trans_rvi.inc.c"
#include "insn_trans/trans_rvm.inc.c"
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 25/35] target/riscv: make ADD/SUB/OR/XOR/AND insn use arg lists
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
manual decoding in gen_arith() is not necessary with decodetree. For now
the function is called trans_arith as the original gen_arith still
exists. The former will be renamed to gen_arith as soon as the old
gen_arith can be removed.
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- &arith -> &r
target/riscv/insn32.decode | 3 ++-
target/riscv/insn_trans/trans_rvi.inc.c | 21 ++++++----------
target/riscv/translate.c | 33 ++++++++++++++-----------
3 files changed, 27 insertions(+), 30 deletions(-)
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index 70e00412b6..520954e9c6 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -36,11 +36,12 @@
# Argument sets:
&b imm rs2 rs1
&i imm rs1 rd
+&r rd rs1 rs2
&shift shamt rs1 rd
&atomic aq rl rs2 rs1 rd
# Formats 32:
-@r ....... ..... ..... ... ..... ....... %rs2 %rs1 %rd
+@r ....... ..... ..... ... ..... ....... &r %rs2 %rs1 %rd
@i ............ ..... ... ..... ....... &i imm=%imm_i %rs1 %rd
@b ....... ..... ..... ... ..... ....... &b imm=%imm_b %rs2 %rs1
@s ....... ..... ..... ... ..... ....... imm=%imm_s %rs2 %rs1
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
index 8bd2752f31..5b5999954a 100644
--- a/target/riscv/insn_trans/trans_rvi.inc.c
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -301,14 +301,12 @@ static bool trans_srai(DisasContext *ctx, arg_srai *a)
static bool trans_add(DisasContext *ctx, arg_add *a)
{
- gen_arith(ctx, OPC_RISC_ADD, a->rd, a->rs1, a->rs2);
- return true;
+ return trans_arith(ctx, a, &tcg_gen_add_tl);
}
static bool trans_sub(DisasContext *ctx, arg_sub *a)
{
- gen_arith(ctx, OPC_RISC_SUB, a->rd, a->rs1, a->rs2);
- return true;
+ return trans_arith(ctx, a, &tcg_gen_sub_tl);
}
static bool trans_sll(DisasContext *ctx, arg_sll *a)
@@ -331,8 +329,7 @@ static bool trans_sltu(DisasContext *ctx, arg_sltu *a)
static bool trans_xor(DisasContext *ctx, arg_xor *a)
{
- gen_arith(ctx, OPC_RISC_XOR, a->rd, a->rs1, a->rs2);
- return true;
+ return trans_arith(ctx, a, &tcg_gen_xor_tl);
}
static bool trans_srl(DisasContext *ctx, arg_srl *a)
@@ -349,14 +346,12 @@ static bool trans_sra(DisasContext *ctx, arg_sra *a)
static bool trans_or(DisasContext *ctx, arg_or *a)
{
- gen_arith(ctx, OPC_RISC_OR, a->rd, a->rs1, a->rs2);
- return true;
+ return trans_arith(ctx, a, &tcg_gen_or_tl);
}
static bool trans_and(DisasContext *ctx, arg_and *a)
{
- gen_arith(ctx, OPC_RISC_AND, a->rd, a->rs1, a->rs2);
- return true;
+ return trans_arith(ctx, a, &tcg_gen_and_tl);
}
#ifdef TARGET_RISCV64
@@ -405,14 +400,12 @@ static bool trans_sraiw(DisasContext *ctx, arg_sraiw *a)
static bool trans_addw(DisasContext *ctx, arg_addw *a)
{
- gen_arith(ctx, OPC_RISC_ADDW, a->rd, a->rs1, a->rs2);
- return true;
+ return trans_arith(ctx, a, &tcg_gen_add_tl);
}
static bool trans_subw(DisasContext *ctx, arg_subw *a)
{
- gen_arith(ctx, OPC_RISC_SUBW, a->rd, a->rs1, a->rs2);
- return true;
+ return trans_arith(ctx, a, &tcg_gen_sub_tl);
}
static bool trans_sllw(DisasContext *ctx, arg_sllw *a)
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index de60686df4..a781aba08c 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -192,12 +192,6 @@ static void gen_arith(DisasContext *ctx, uint32_t opc, int rd, int rs1,
gen_get_gpr(source2, rs2);
switch (opc) {
- CASE_OP_32_64(OPC_RISC_ADD):
- tcg_gen_add_tl(source1, source1, source2);
- break;
- CASE_OP_32_64(OPC_RISC_SUB):
- tcg_gen_sub_tl(source1, source1, source2);
- break;
#if defined(TARGET_RISCV64)
case OPC_RISC_SLLW:
tcg_gen_andi_tl(source2, source2, 0x1F);
@@ -214,9 +208,6 @@ static void gen_arith(DisasContext *ctx, uint32_t opc, int rd, int rs1,
case OPC_RISC_SLTU:
tcg_gen_setcond_tl(TCG_COND_LTU, source1, source1, source2);
break;
- case OPC_RISC_XOR:
- tcg_gen_xor_tl(source1, source1, source2);
- break;
#if defined(TARGET_RISCV64)
case OPC_RISC_SRLW:
/* clear upper 32 */
@@ -242,12 +233,6 @@ static void gen_arith(DisasContext *ctx, uint32_t opc, int rd, int rs1,
tcg_gen_andi_tl(source2, source2, TARGET_LONG_BITS - 1);
tcg_gen_sar_tl(source1, source1, source2);
break;
- case OPC_RISC_OR:
- tcg_gen_or_tl(source1, source1, source2);
- break;
- case OPC_RISC_AND:
- tcg_gen_and_tl(source1, source1, source2);
- break;
CASE_OP_32_64(OPC_RISC_MUL):
tcg_gen_mul_tl(source1, source1, source2);
break;
@@ -644,6 +629,24 @@ static void gen_addiw(TCGv ret, TCGv arg1, TCGv arg2)
}
#endif
+static bool trans_arith(DisasContext *ctx, arg_r *a,
+ void(*func)(TCGv, TCGv, TCGv))
+{
+ TCGv source1, source2;
+ source1 = tcg_temp_new();
+ source2 = tcg_temp_new();
+
+ gen_get_gpr(source1, a->rs1);
+ gen_get_gpr(source2, a->rs2);
+
+ (*func)(source1, source1, source2);
+
+ gen_set_gpr(a->rd, source1);
+ tcg_temp_free(source1);
+ tcg_temp_free(source2);
+ return true;
+}
+
/* Include insn module translation function */
#include "insn_trans/trans_rvi.inc.c"
#include "insn_trans/trans_rvm.inc.c"
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 25/35] target/riscv: make ADD/SUB/OR/XOR/AND insn use arg lists
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
manual decoding in gen_arith() is not necessary with decodetree. For now
the function is called trans_arith as the original gen_arith still
exists. The former will be renamed to gen_arith as soon as the old
gen_arith can be removed.
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- &arith -> &r
target/riscv/insn32.decode | 3 ++-
target/riscv/insn_trans/trans_rvi.inc.c | 21 ++++++----------
target/riscv/translate.c | 33 ++++++++++++++-----------
3 files changed, 27 insertions(+), 30 deletions(-)
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index 70e00412b6..520954e9c6 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -36,11 +36,12 @@
# Argument sets:
&b imm rs2 rs1
&i imm rs1 rd
+&r rd rs1 rs2
&shift shamt rs1 rd
&atomic aq rl rs2 rs1 rd
# Formats 32:
-@r ....... ..... ..... ... ..... ....... %rs2 %rs1 %rd
+@r ....... ..... ..... ... ..... ....... &r %rs2 %rs1 %rd
@i ............ ..... ... ..... ....... &i imm=%imm_i %rs1 %rd
@b ....... ..... ..... ... ..... ....... &b imm=%imm_b %rs2 %rs1
@s ....... ..... ..... ... ..... ....... imm=%imm_s %rs2 %rs1
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
index 8bd2752f31..5b5999954a 100644
--- a/target/riscv/insn_trans/trans_rvi.inc.c
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -301,14 +301,12 @@ static bool trans_srai(DisasContext *ctx, arg_srai *a)
static bool trans_add(DisasContext *ctx, arg_add *a)
{
- gen_arith(ctx, OPC_RISC_ADD, a->rd, a->rs1, a->rs2);
- return true;
+ return trans_arith(ctx, a, &tcg_gen_add_tl);
}
static bool trans_sub(DisasContext *ctx, arg_sub *a)
{
- gen_arith(ctx, OPC_RISC_SUB, a->rd, a->rs1, a->rs2);
- return true;
+ return trans_arith(ctx, a, &tcg_gen_sub_tl);
}
static bool trans_sll(DisasContext *ctx, arg_sll *a)
@@ -331,8 +329,7 @@ static bool trans_sltu(DisasContext *ctx, arg_sltu *a)
static bool trans_xor(DisasContext *ctx, arg_xor *a)
{
- gen_arith(ctx, OPC_RISC_XOR, a->rd, a->rs1, a->rs2);
- return true;
+ return trans_arith(ctx, a, &tcg_gen_xor_tl);
}
static bool trans_srl(DisasContext *ctx, arg_srl *a)
@@ -349,14 +346,12 @@ static bool trans_sra(DisasContext *ctx, arg_sra *a)
static bool trans_or(DisasContext *ctx, arg_or *a)
{
- gen_arith(ctx, OPC_RISC_OR, a->rd, a->rs1, a->rs2);
- return true;
+ return trans_arith(ctx, a, &tcg_gen_or_tl);
}
static bool trans_and(DisasContext *ctx, arg_and *a)
{
- gen_arith(ctx, OPC_RISC_AND, a->rd, a->rs1, a->rs2);
- return true;
+ return trans_arith(ctx, a, &tcg_gen_and_tl);
}
#ifdef TARGET_RISCV64
@@ -405,14 +400,12 @@ static bool trans_sraiw(DisasContext *ctx, arg_sraiw *a)
static bool trans_addw(DisasContext *ctx, arg_addw *a)
{
- gen_arith(ctx, OPC_RISC_ADDW, a->rd, a->rs1, a->rs2);
- return true;
+ return trans_arith(ctx, a, &tcg_gen_add_tl);
}
static bool trans_subw(DisasContext *ctx, arg_subw *a)
{
- gen_arith(ctx, OPC_RISC_SUBW, a->rd, a->rs1, a->rs2);
- return true;
+ return trans_arith(ctx, a, &tcg_gen_sub_tl);
}
static bool trans_sllw(DisasContext *ctx, arg_sllw *a)
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index de60686df4..a781aba08c 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -192,12 +192,6 @@ static void gen_arith(DisasContext *ctx, uint32_t opc, int rd, int rs1,
gen_get_gpr(source2, rs2);
switch (opc) {
- CASE_OP_32_64(OPC_RISC_ADD):
- tcg_gen_add_tl(source1, source1, source2);
- break;
- CASE_OP_32_64(OPC_RISC_SUB):
- tcg_gen_sub_tl(source1, source1, source2);
- break;
#if defined(TARGET_RISCV64)
case OPC_RISC_SLLW:
tcg_gen_andi_tl(source2, source2, 0x1F);
@@ -214,9 +208,6 @@ static void gen_arith(DisasContext *ctx, uint32_t opc, int rd, int rs1,
case OPC_RISC_SLTU:
tcg_gen_setcond_tl(TCG_COND_LTU, source1, source1, source2);
break;
- case OPC_RISC_XOR:
- tcg_gen_xor_tl(source1, source1, source2);
- break;
#if defined(TARGET_RISCV64)
case OPC_RISC_SRLW:
/* clear upper 32 */
@@ -242,12 +233,6 @@ static void gen_arith(DisasContext *ctx, uint32_t opc, int rd, int rs1,
tcg_gen_andi_tl(source2, source2, TARGET_LONG_BITS - 1);
tcg_gen_sar_tl(source1, source1, source2);
break;
- case OPC_RISC_OR:
- tcg_gen_or_tl(source1, source1, source2);
- break;
- case OPC_RISC_AND:
- tcg_gen_and_tl(source1, source1, source2);
- break;
CASE_OP_32_64(OPC_RISC_MUL):
tcg_gen_mul_tl(source1, source1, source2);
break;
@@ -644,6 +629,24 @@ static void gen_addiw(TCGv ret, TCGv arg1, TCGv arg2)
}
#endif
+static bool trans_arith(DisasContext *ctx, arg_r *a,
+ void(*func)(TCGv, TCGv, TCGv))
+{
+ TCGv source1, source2;
+ source1 = tcg_temp_new();
+ source2 = tcg_temp_new();
+
+ gen_get_gpr(source1, a->rs1);
+ gen_get_gpr(source2, a->rs2);
+
+ (*func)(source1, source1, source2);
+
+ gen_set_gpr(a->rd, source1);
+ tcg_temp_free(source1);
+ tcg_temp_free(source2);
+ return true;
+}
+
/* Include insn module translation function */
#include "insn_trans/trans_rvi.inc.c"
#include "insn_trans/trans_rvm.inc.c"
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 26/35] target/riscv: Remove shift and slt insn manual decoding
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
target/riscv/insn_trans/trans_rvi.inc.c | 78 +++++++++++++++++++++----
target/riscv/translate.c | 59 ++++++-------------
2 files changed, 85 insertions(+), 52 deletions(-)
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
index 5b5999954a..65548d1281 100644
--- a/target/riscv/insn_trans/trans_rvi.inc.c
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -311,19 +311,38 @@ static bool trans_sub(DisasContext *ctx, arg_sub *a)
static bool trans_sll(DisasContext *ctx, arg_sll *a)
{
- gen_arith(ctx, OPC_RISC_SLL, a->rd, a->rs1, a->rs2);
- return true;
+ return gen_shift(ctx, a, &tcg_gen_shl_tl);
}
static bool trans_slt(DisasContext *ctx, arg_slt *a)
{
- gen_arith(ctx, OPC_RISC_SLT, a->rd, a->rs1, a->rs2);
+ TCGv source1 = tcg_temp_new();
+ TCGv source2 = tcg_temp_new();
+
+ gen_get_gpr(source1, a->rs1);
+ gen_get_gpr(source2, a->rs2);
+
+ tcg_gen_setcond_tl(TCG_COND_LT, source1, source1, source2);
+
+ gen_set_gpr(a->rd, source1);
+ tcg_temp_free(source1);
+ tcg_temp_free(source2);
return true;
}
static bool trans_sltu(DisasContext *ctx, arg_sltu *a)
{
- gen_arith(ctx, OPC_RISC_SLTU, a->rd, a->rs1, a->rs2);
+ TCGv source1 = tcg_temp_new();
+ TCGv source2 = tcg_temp_new();
+
+ gen_get_gpr(source1, a->rs1);
+ gen_get_gpr(source2, a->rs2);
+
+ tcg_gen_setcond_tl(TCG_COND_LTU, source1, source1, source2);
+
+ gen_set_gpr(a->rd, source1);
+ tcg_temp_free(source1);
+ tcg_temp_free(source2);
return true;
}
@@ -334,14 +353,12 @@ static bool trans_xor(DisasContext *ctx, arg_xor *a)
static bool trans_srl(DisasContext *ctx, arg_srl *a)
{
- gen_arith(ctx, OPC_RISC_SRL, a->rd, a->rs1, a->rs2);
- return true;
+ return gen_shift(ctx, a, &tcg_gen_shr_tl);
}
static bool trans_sra(DisasContext *ctx, arg_sra *a)
{
- gen_arith(ctx, OPC_RISC_SRA, a->rd, a->rs1, a->rs2);
- return true;
+ return gen_shift(ctx, a, &tcg_gen_sar_tl);
}
static bool trans_or(DisasContext *ctx, arg_or *a)
@@ -410,19 +427,58 @@ static bool trans_subw(DisasContext *ctx, arg_subw *a)
static bool trans_sllw(DisasContext *ctx, arg_sllw *a)
{
- gen_arith(ctx, OPC_RISC_SLLW, a->rd, a->rs1, a->rs2);
+ TCGv source1 = tcg_temp_new();
+ TCGv source2 = tcg_temp_new();
+
+ gen_get_gpr(source1, a->rs1);
+ gen_get_gpr(source2, a->rs2);
+
+ tcg_gen_andi_tl(source2, source2, 0x1F);
+ tcg_gen_shl_tl(source1, source1, source2);
+
+ gen_set_gpr(a->rd, source1);
+ tcg_temp_free(source1);
+ tcg_temp_free(source2);
return true;
}
static bool trans_srlw(DisasContext *ctx, arg_srlw *a)
{
- gen_arith(ctx, OPC_RISC_SRLW, a->rd, a->rs1, a->rs2);
+ TCGv source1 = tcg_temp_new();
+ TCGv source2 = tcg_temp_new();
+
+ gen_get_gpr(source1, a->rs1);
+ gen_get_gpr(source2, a->rs2);
+
+ /* clear upper 32 */
+ tcg_gen_ext32u_tl(source1, source1);
+ tcg_gen_andi_tl(source2, source2, 0x1F);
+ tcg_gen_shr_tl(source1, source1, source2);
+
+ gen_set_gpr(a->rd, source1);
+ tcg_temp_free(source1);
+ tcg_temp_free(source2);
return true;
}
static bool trans_sraw(DisasContext *ctx, arg_sraw *a)
{
- gen_arith(ctx, OPC_RISC_SRAW, a->rd, a->rs1, a->rs2);
+ TCGv source1 = tcg_temp_new();
+ TCGv source2 = tcg_temp_new();
+
+ gen_get_gpr(source1, a->rs1);
+ gen_get_gpr(source2, a->rs2);
+
+ /* first, trick to get it to act like working on 32 bits (get rid of
+ upper 32, sign extend to fill space) */
+ tcg_gen_ext32s_tl(source1, source1);
+ tcg_gen_andi_tl(source2, source2, 0x1F);
+ tcg_gen_sar_tl(source1, source1, source2);
+
+ gen_set_gpr(a->rd, source1);
+ tcg_temp_free(source1);
+ tcg_temp_free(source2);
+
return true;
}
#endif
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index a781aba08c..5b4d091561 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -192,47 +192,6 @@ static void gen_arith(DisasContext *ctx, uint32_t opc, int rd, int rs1,
gen_get_gpr(source2, rs2);
switch (opc) {
-#if defined(TARGET_RISCV64)
- case OPC_RISC_SLLW:
- tcg_gen_andi_tl(source2, source2, 0x1F);
- tcg_gen_shl_tl(source1, source1, source2);
- break;
-#endif
- case OPC_RISC_SLL:
- tcg_gen_andi_tl(source2, source2, TARGET_LONG_BITS - 1);
- tcg_gen_shl_tl(source1, source1, source2);
- break;
- case OPC_RISC_SLT:
- tcg_gen_setcond_tl(TCG_COND_LT, source1, source1, source2);
- break;
- case OPC_RISC_SLTU:
- tcg_gen_setcond_tl(TCG_COND_LTU, source1, source1, source2);
- break;
-#if defined(TARGET_RISCV64)
- case OPC_RISC_SRLW:
- /* clear upper 32 */
- tcg_gen_ext32u_tl(source1, source1);
- tcg_gen_andi_tl(source2, source2, 0x1F);
- tcg_gen_shr_tl(source1, source1, source2);
- break;
-#endif
- case OPC_RISC_SRL:
- tcg_gen_andi_tl(source2, source2, TARGET_LONG_BITS - 1);
- tcg_gen_shr_tl(source1, source1, source2);
- break;
-#if defined(TARGET_RISCV64)
- case OPC_RISC_SRAW:
- /* first, trick to get it to act like working on 32 bits (get rid of
- upper 32, sign extend to fill space) */
- tcg_gen_ext32s_tl(source1, source1);
- tcg_gen_andi_tl(source2, source2, 0x1F);
- tcg_gen_sar_tl(source1, source1, source2);
- break;
-#endif
- case OPC_RISC_SRA:
- tcg_gen_andi_tl(source2, source2, TARGET_LONG_BITS - 1);
- tcg_gen_sar_tl(source1, source1, source2);
- break;
CASE_OP_32_64(OPC_RISC_MUL):
tcg_gen_mul_tl(source1, source1, source2);
break;
@@ -647,6 +606,24 @@ static bool trans_arith(DisasContext *ctx, arg_r *a,
return true;
}
+static bool gen_shift(DisasContext *ctx, arg_r *a,
+ void(*func)(TCGv, TCGv, TCGv))
+{
+ TCGv source1 = tcg_temp_new();
+ TCGv source2 = tcg_temp_new();
+
+ gen_get_gpr(source1, a->rs1);
+ gen_get_gpr(source2, a->rs2);
+
+ tcg_gen_andi_tl(source2, source2, TARGET_LONG_BITS - 1);
+ (*func)(source1, source1, source2);
+
+ gen_set_gpr(a->rd, source1);
+ tcg_temp_free(source1);
+ tcg_temp_free(source2);
+ return true;
+}
+
/* Include insn module translation function */
#include "insn_trans/trans_rvi.inc.c"
#include "insn_trans/trans_rvm.inc.c"
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 26/35] target/riscv: Remove shift and slt insn manual decoding
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
target/riscv/insn_trans/trans_rvi.inc.c | 78 +++++++++++++++++++++----
target/riscv/translate.c | 59 ++++++-------------
2 files changed, 85 insertions(+), 52 deletions(-)
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
index 5b5999954a..65548d1281 100644
--- a/target/riscv/insn_trans/trans_rvi.inc.c
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -311,19 +311,38 @@ static bool trans_sub(DisasContext *ctx, arg_sub *a)
static bool trans_sll(DisasContext *ctx, arg_sll *a)
{
- gen_arith(ctx, OPC_RISC_SLL, a->rd, a->rs1, a->rs2);
- return true;
+ return gen_shift(ctx, a, &tcg_gen_shl_tl);
}
static bool trans_slt(DisasContext *ctx, arg_slt *a)
{
- gen_arith(ctx, OPC_RISC_SLT, a->rd, a->rs1, a->rs2);
+ TCGv source1 = tcg_temp_new();
+ TCGv source2 = tcg_temp_new();
+
+ gen_get_gpr(source1, a->rs1);
+ gen_get_gpr(source2, a->rs2);
+
+ tcg_gen_setcond_tl(TCG_COND_LT, source1, source1, source2);
+
+ gen_set_gpr(a->rd, source1);
+ tcg_temp_free(source1);
+ tcg_temp_free(source2);
return true;
}
static bool trans_sltu(DisasContext *ctx, arg_sltu *a)
{
- gen_arith(ctx, OPC_RISC_SLTU, a->rd, a->rs1, a->rs2);
+ TCGv source1 = tcg_temp_new();
+ TCGv source2 = tcg_temp_new();
+
+ gen_get_gpr(source1, a->rs1);
+ gen_get_gpr(source2, a->rs2);
+
+ tcg_gen_setcond_tl(TCG_COND_LTU, source1, source1, source2);
+
+ gen_set_gpr(a->rd, source1);
+ tcg_temp_free(source1);
+ tcg_temp_free(source2);
return true;
}
@@ -334,14 +353,12 @@ static bool trans_xor(DisasContext *ctx, arg_xor *a)
static bool trans_srl(DisasContext *ctx, arg_srl *a)
{
- gen_arith(ctx, OPC_RISC_SRL, a->rd, a->rs1, a->rs2);
- return true;
+ return gen_shift(ctx, a, &tcg_gen_shr_tl);
}
static bool trans_sra(DisasContext *ctx, arg_sra *a)
{
- gen_arith(ctx, OPC_RISC_SRA, a->rd, a->rs1, a->rs2);
- return true;
+ return gen_shift(ctx, a, &tcg_gen_sar_tl);
}
static bool trans_or(DisasContext *ctx, arg_or *a)
@@ -410,19 +427,58 @@ static bool trans_subw(DisasContext *ctx, arg_subw *a)
static bool trans_sllw(DisasContext *ctx, arg_sllw *a)
{
- gen_arith(ctx, OPC_RISC_SLLW, a->rd, a->rs1, a->rs2);
+ TCGv source1 = tcg_temp_new();
+ TCGv source2 = tcg_temp_new();
+
+ gen_get_gpr(source1, a->rs1);
+ gen_get_gpr(source2, a->rs2);
+
+ tcg_gen_andi_tl(source2, source2, 0x1F);
+ tcg_gen_shl_tl(source1, source1, source2);
+
+ gen_set_gpr(a->rd, source1);
+ tcg_temp_free(source1);
+ tcg_temp_free(source2);
return true;
}
static bool trans_srlw(DisasContext *ctx, arg_srlw *a)
{
- gen_arith(ctx, OPC_RISC_SRLW, a->rd, a->rs1, a->rs2);
+ TCGv source1 = tcg_temp_new();
+ TCGv source2 = tcg_temp_new();
+
+ gen_get_gpr(source1, a->rs1);
+ gen_get_gpr(source2, a->rs2);
+
+ /* clear upper 32 */
+ tcg_gen_ext32u_tl(source1, source1);
+ tcg_gen_andi_tl(source2, source2, 0x1F);
+ tcg_gen_shr_tl(source1, source1, source2);
+
+ gen_set_gpr(a->rd, source1);
+ tcg_temp_free(source1);
+ tcg_temp_free(source2);
return true;
}
static bool trans_sraw(DisasContext *ctx, arg_sraw *a)
{
- gen_arith(ctx, OPC_RISC_SRAW, a->rd, a->rs1, a->rs2);
+ TCGv source1 = tcg_temp_new();
+ TCGv source2 = tcg_temp_new();
+
+ gen_get_gpr(source1, a->rs1);
+ gen_get_gpr(source2, a->rs2);
+
+ /* first, trick to get it to act like working on 32 bits (get rid of
+ upper 32, sign extend to fill space) */
+ tcg_gen_ext32s_tl(source1, source1);
+ tcg_gen_andi_tl(source2, source2, 0x1F);
+ tcg_gen_sar_tl(source1, source1, source2);
+
+ gen_set_gpr(a->rd, source1);
+ tcg_temp_free(source1);
+ tcg_temp_free(source2);
+
return true;
}
#endif
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index a781aba08c..5b4d091561 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -192,47 +192,6 @@ static void gen_arith(DisasContext *ctx, uint32_t opc, int rd, int rs1,
gen_get_gpr(source2, rs2);
switch (opc) {
-#if defined(TARGET_RISCV64)
- case OPC_RISC_SLLW:
- tcg_gen_andi_tl(source2, source2, 0x1F);
- tcg_gen_shl_tl(source1, source1, source2);
- break;
-#endif
- case OPC_RISC_SLL:
- tcg_gen_andi_tl(source2, source2, TARGET_LONG_BITS - 1);
- tcg_gen_shl_tl(source1, source1, source2);
- break;
- case OPC_RISC_SLT:
- tcg_gen_setcond_tl(TCG_COND_LT, source1, source1, source2);
- break;
- case OPC_RISC_SLTU:
- tcg_gen_setcond_tl(TCG_COND_LTU, source1, source1, source2);
- break;
-#if defined(TARGET_RISCV64)
- case OPC_RISC_SRLW:
- /* clear upper 32 */
- tcg_gen_ext32u_tl(source1, source1);
- tcg_gen_andi_tl(source2, source2, 0x1F);
- tcg_gen_shr_tl(source1, source1, source2);
- break;
-#endif
- case OPC_RISC_SRL:
- tcg_gen_andi_tl(source2, source2, TARGET_LONG_BITS - 1);
- tcg_gen_shr_tl(source1, source1, source2);
- break;
-#if defined(TARGET_RISCV64)
- case OPC_RISC_SRAW:
- /* first, trick to get it to act like working on 32 bits (get rid of
- upper 32, sign extend to fill space) */
- tcg_gen_ext32s_tl(source1, source1);
- tcg_gen_andi_tl(source2, source2, 0x1F);
- tcg_gen_sar_tl(source1, source1, source2);
- break;
-#endif
- case OPC_RISC_SRA:
- tcg_gen_andi_tl(source2, source2, TARGET_LONG_BITS - 1);
- tcg_gen_sar_tl(source1, source1, source2);
- break;
CASE_OP_32_64(OPC_RISC_MUL):
tcg_gen_mul_tl(source1, source1, source2);
break;
@@ -647,6 +606,24 @@ static bool trans_arith(DisasContext *ctx, arg_r *a,
return true;
}
+static bool gen_shift(DisasContext *ctx, arg_r *a,
+ void(*func)(TCGv, TCGv, TCGv))
+{
+ TCGv source1 = tcg_temp_new();
+ TCGv source2 = tcg_temp_new();
+
+ gen_get_gpr(source1, a->rs1);
+ gen_get_gpr(source2, a->rs2);
+
+ tcg_gen_andi_tl(source2, source2, TARGET_LONG_BITS - 1);
+ (*func)(source1, source1, source2);
+
+ gen_set_gpr(a->rd, source1);
+ tcg_temp_free(source1);
+ tcg_temp_free(source2);
+ return true;
+}
+
/* Include insn module translation function */
#include "insn_trans/trans_rvi.inc.c"
#include "insn_trans/trans_rvm.inc.c"
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 27/35] target/riscv: Remove manual decoding of RV32/64M insn
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- trans_mulw now uses gen_mulw which properly sign extends the result
- gen_arith_w -> gen_arith_div_w
- gen_arith_div_w properly sign extends the resul
target/riscv/insn_trans/trans_rvm.inc.c | 55 +++--
target/riscv/translate.c | 281 +++++++++++-------------
2 files changed, 162 insertions(+), 174 deletions(-)
diff --git a/target/riscv/insn_trans/trans_rvm.inc.c b/target/riscv/insn_trans/trans_rvm.inc.c
index ec3197ede8..949f59ddb2 100644
--- a/target/riscv/insn_trans/trans_rvm.inc.c
+++ b/target/riscv/insn_trans/trans_rvm.inc.c
@@ -21,80 +21,87 @@
static bool trans_mul(DisasContext *ctx, arg_mul *a)
{
- gen_arith(ctx, OPC_RISC_MUL, a->rd, a->rs1, a->rs2);
- return true;
+ return trans_arith(ctx, a, &tcg_gen_mul_tl);
}
static bool trans_mulh(DisasContext *ctx, arg_mulh *a)
{
- gen_arith(ctx, OPC_RISC_MULH, a->rd, a->rs1, a->rs2);
+ TCGv source1 = tcg_temp_new();
+ TCGv source2 = tcg_temp_new();
+ gen_get_gpr(source1, a->rs1);
+ gen_get_gpr(source2, a->rs2);
+
+ tcg_gen_muls2_tl(source2, source1, source1, source2);
+
+ gen_set_gpr(a->rd, source1);
+ tcg_temp_free(source1);
+ tcg_temp_free(source2);
return true;
}
static bool trans_mulhsu(DisasContext *ctx, arg_mulhsu *a)
{
- gen_arith(ctx, OPC_RISC_MULHSU, a->rd, a->rs1, a->rs2);
- return true;
+ return trans_arith(ctx, a, &gen_mulhsu);
}
static bool trans_mulhu(DisasContext *ctx, arg_mulhu *a)
{
- gen_arith(ctx, OPC_RISC_MULHU, a->rd, a->rs1, a->rs2);
+ TCGv source1 = tcg_temp_new();
+ TCGv source2 = tcg_temp_new();
+ gen_get_gpr(source1, a->rs1);
+ gen_get_gpr(source2, a->rs2);
+
+ tcg_gen_mulu2_tl(source2, source1, source1, source2);
+
+ gen_set_gpr(a->rd, source1);
+ tcg_temp_free(source1);
+ tcg_temp_free(source2);
return true;
}
static bool trans_div(DisasContext *ctx, arg_div *a)
{
- gen_arith(ctx, OPC_RISC_DIV, a->rd, a->rs1, a->rs2);
- return true;
+ return trans_arith(ctx, a, &gen_div);
}
static bool trans_divu(DisasContext *ctx, arg_divu *a)
{
- gen_arith(ctx, OPC_RISC_DIVU, a->rd, a->rs1, a->rs2);
- return true;
+ return trans_arith(ctx, a, &gen_divu);
}
static bool trans_rem(DisasContext *ctx, arg_rem *a)
{
- gen_arith(ctx, OPC_RISC_REM, a->rd, a->rs1, a->rs2);
- return true;
+ return trans_arith(ctx, a, &gen_rem);
}
static bool trans_remu(DisasContext *ctx, arg_remu *a)
{
- gen_arith(ctx, OPC_RISC_REMU, a->rd, a->rs1, a->rs2);
- return true;
+ return trans_arith(ctx, a, &gen_remu);
}
#ifdef TARGET_RISCV64
static bool trans_mulw(DisasContext *ctx, arg_mulw *a)
{
- gen_arith(ctx, OPC_RISC_MULW, a->rd, a->rs1, a->rs2);
- return true;
+ return trans_arith(ctx, a, &gen_mulw);
}
static bool trans_divw(DisasContext *ctx, arg_divw *a)
{
- gen_arith(ctx, OPC_RISC_DIVW, a->rd, a->rs1, a->rs2);
- return true;
+ return gen_arith_div_w(ctx, a, &gen_div);
}
static bool trans_divuw(DisasContext *ctx, arg_divuw *a)
{
- gen_arith(ctx, OPC_RISC_DIVUW, a->rd, a->rs1, a->rs2);
- return true;
+ return gen_arith_div_w(ctx, a, &gen_divu);
}
static bool trans_remw(DisasContext *ctx, arg_remw *a)
{
- gen_arith(ctx, OPC_RISC_REMW, a->rd, a->rs1, a->rs2);
- return true;
+ return gen_arith_div_w(ctx, a, &gen_rem);
}
static bool trans_remuw(DisasContext *ctx, arg_remuw *a)
{
- gen_arith(ctx, OPC_RISC_REMUW, a->rd, a->rs1, a->rs2);
- return true;
+ return gen_arith_div_w(ctx, a, &gen_remu);
}
#endif
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 5b4d091561..fd4b63b4c4 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -182,156 +182,110 @@ static void gen_mulhsu(TCGv ret, TCGv arg1, TCGv arg2)
tcg_temp_free(rh);
}
-static void gen_arith(DisasContext *ctx, uint32_t opc, int rd, int rs1,
- int rs2)
-{
- TCGv source1, source2, cond1, cond2, zeroreg, resultopt1;
- source1 = tcg_temp_new();
- source2 = tcg_temp_new();
- gen_get_gpr(source1, rs1);
- gen_get_gpr(source2, rs2);
-
- switch (opc) {
- CASE_OP_32_64(OPC_RISC_MUL):
- tcg_gen_mul_tl(source1, source1, source2);
- break;
- case OPC_RISC_MULH:
- tcg_gen_muls2_tl(source2, source1, source1, source2);
- break;
- case OPC_RISC_MULHSU:
- gen_mulhsu(source1, source1, source2);
- break;
- case OPC_RISC_MULHU:
- tcg_gen_mulu2_tl(source2, source1, source1, source2);
- break;
-#if defined(TARGET_RISCV64)
- case OPC_RISC_DIVW:
- tcg_gen_ext32s_tl(source1, source1);
- tcg_gen_ext32s_tl(source2, source2);
- /* fall through to DIV */
-#endif
- case OPC_RISC_DIV:
- /* Handle by altering args to tcg_gen_div to produce req'd results:
- * For overflow: want source1 in source1 and 1 in source2
- * For div by zero: want -1 in source1 and 1 in source2 -> -1 result */
- cond1 = tcg_temp_new();
- cond2 = tcg_temp_new();
- zeroreg = tcg_const_tl(0);
- resultopt1 = tcg_temp_new();
-
- tcg_gen_movi_tl(resultopt1, (target_ulong)-1);
- tcg_gen_setcondi_tl(TCG_COND_EQ, cond2, source2, (target_ulong)(~0L));
- tcg_gen_setcondi_tl(TCG_COND_EQ, cond1, source1,
- ((target_ulong)1) << (TARGET_LONG_BITS - 1));
- tcg_gen_and_tl(cond1, cond1, cond2); /* cond1 = overflow */
- tcg_gen_setcondi_tl(TCG_COND_EQ, cond2, source2, 0); /* cond2 = div 0 */
- /* if div by zero, set source1 to -1, otherwise don't change */
- tcg_gen_movcond_tl(TCG_COND_EQ, source1, cond2, zeroreg, source1,
- resultopt1);
- /* if overflow or div by zero, set source2 to 1, else don't change */
- tcg_gen_or_tl(cond1, cond1, cond2);
- tcg_gen_movi_tl(resultopt1, (target_ulong)1);
- tcg_gen_movcond_tl(TCG_COND_EQ, source2, cond1, zeroreg, source2,
- resultopt1);
- tcg_gen_div_tl(source1, source1, source2);
-
- tcg_temp_free(cond1);
- tcg_temp_free(cond2);
- tcg_temp_free(zeroreg);
- tcg_temp_free(resultopt1);
- break;
-#if defined(TARGET_RISCV64)
- case OPC_RISC_DIVUW:
- tcg_gen_ext32u_tl(source1, source1);
- tcg_gen_ext32u_tl(source2, source2);
- /* fall through to DIVU */
-#endif
- case OPC_RISC_DIVU:
- cond1 = tcg_temp_new();
- zeroreg = tcg_const_tl(0);
- resultopt1 = tcg_temp_new();
-
- tcg_gen_setcondi_tl(TCG_COND_EQ, cond1, source2, 0);
- tcg_gen_movi_tl(resultopt1, (target_ulong)-1);
- tcg_gen_movcond_tl(TCG_COND_EQ, source1, cond1, zeroreg, source1,
- resultopt1);
- tcg_gen_movi_tl(resultopt1, (target_ulong)1);
- tcg_gen_movcond_tl(TCG_COND_EQ, source2, cond1, zeroreg, source2,
- resultopt1);
- tcg_gen_divu_tl(source1, source1, source2);
-
- tcg_temp_free(cond1);
- tcg_temp_free(zeroreg);
- tcg_temp_free(resultopt1);
- break;
-#if defined(TARGET_RISCV64)
- case OPC_RISC_REMW:
- tcg_gen_ext32s_tl(source1, source1);
- tcg_gen_ext32s_tl(source2, source2);
- /* fall through to REM */
-#endif
- case OPC_RISC_REM:
- cond1 = tcg_temp_new();
- cond2 = tcg_temp_new();
- zeroreg = tcg_const_tl(0);
- resultopt1 = tcg_temp_new();
-
- tcg_gen_movi_tl(resultopt1, 1L);
- tcg_gen_setcondi_tl(TCG_COND_EQ, cond2, source2, (target_ulong)-1);
- tcg_gen_setcondi_tl(TCG_COND_EQ, cond1, source1,
- (target_ulong)1 << (TARGET_LONG_BITS - 1));
- tcg_gen_and_tl(cond2, cond1, cond2); /* cond1 = overflow */
- tcg_gen_setcondi_tl(TCG_COND_EQ, cond1, source2, 0); /* cond2 = div 0 */
- /* if overflow or div by zero, set source2 to 1, else don't change */
- tcg_gen_or_tl(cond2, cond1, cond2);
- tcg_gen_movcond_tl(TCG_COND_EQ, source2, cond2, zeroreg, source2,
- resultopt1);
- tcg_gen_rem_tl(resultopt1, source1, source2);
- /* if div by zero, just return the original dividend */
- tcg_gen_movcond_tl(TCG_COND_EQ, source1, cond1, zeroreg, resultopt1,
- source1);
-
- tcg_temp_free(cond1);
- tcg_temp_free(cond2);
- tcg_temp_free(zeroreg);
- tcg_temp_free(resultopt1);
- break;
-#if defined(TARGET_RISCV64)
- case OPC_RISC_REMUW:
- tcg_gen_ext32u_tl(source1, source1);
- tcg_gen_ext32u_tl(source2, source2);
- /* fall through to REMU */
-#endif
- case OPC_RISC_REMU:
- cond1 = tcg_temp_new();
- zeroreg = tcg_const_tl(0);
- resultopt1 = tcg_temp_new();
-
- tcg_gen_movi_tl(resultopt1, (target_ulong)1);
- tcg_gen_setcondi_tl(TCG_COND_EQ, cond1, source2, 0);
- tcg_gen_movcond_tl(TCG_COND_EQ, source2, cond1, zeroreg, source2,
- resultopt1);
- tcg_gen_remu_tl(resultopt1, source1, source2);
- /* if div by zero, just return the original dividend */
- tcg_gen_movcond_tl(TCG_COND_EQ, source1, cond1, zeroreg, resultopt1,
- source1);
-
- tcg_temp_free(cond1);
- tcg_temp_free(zeroreg);
- tcg_temp_free(resultopt1);
- break;
- default:
- gen_exception_illegal(ctx);
- return;
- }
-
- if (opc & 0x8) { /* sign extend for W instructions */
- tcg_gen_ext32s_tl(source1, source1);
- }
-
- gen_set_gpr(rd, source1);
- tcg_temp_free(source1);
- tcg_temp_free(source2);
+static void gen_div(TCGv ret, TCGv source1, TCGv source2)
+{
+ TCGv cond1, cond2, zeroreg, resultopt1;
+ /* Handle by altering args to tcg_gen_div to produce req'd results:
+ * For overflow: want source1 in source1 and 1 in source2
+ * For div by zero: want -1 in source1 and 1 in source2 -> -1 result */
+ cond1 = tcg_temp_new();
+ cond2 = tcg_temp_new();
+ zeroreg = tcg_const_tl(0);
+ resultopt1 = tcg_temp_new();
+
+ tcg_gen_movi_tl(resultopt1, (target_ulong)-1);
+ tcg_gen_setcondi_tl(TCG_COND_EQ, cond2, source2, (target_ulong)(~0L));
+ tcg_gen_setcondi_tl(TCG_COND_EQ, cond1, source1,
+ ((target_ulong)1) << (TARGET_LONG_BITS - 1));
+ tcg_gen_and_tl(cond1, cond1, cond2); /* cond1 = overflow */
+ tcg_gen_setcondi_tl(TCG_COND_EQ, cond2, source2, 0); /* cond2 = div 0 */
+ /* if div by zero, set source1 to -1, otherwise don't change */
+ tcg_gen_movcond_tl(TCG_COND_EQ, source1, cond2, zeroreg, source1,
+ resultopt1);
+ /* if overflow or div by zero, set source2 to 1, else don't change */
+ tcg_gen_or_tl(cond1, cond1, cond2);
+ tcg_gen_movi_tl(resultopt1, (target_ulong)1);
+ tcg_gen_movcond_tl(TCG_COND_EQ, source2, cond1, zeroreg, source2,
+ resultopt1);
+ tcg_gen_div_tl(ret, source1, source2);
+
+ tcg_temp_free(cond1);
+ tcg_temp_free(cond2);
+ tcg_temp_free(zeroreg);
+ tcg_temp_free(resultopt1);
+}
+
+static void gen_divu(TCGv ret, TCGv source1, TCGv source2)
+{
+ TCGv cond1, zeroreg, resultopt1;
+ cond1 = tcg_temp_new();
+
+ zeroreg = tcg_const_tl(0);
+ resultopt1 = tcg_temp_new();
+
+ tcg_gen_setcondi_tl(TCG_COND_EQ, cond1, source2, 0);
+ tcg_gen_movi_tl(resultopt1, (target_ulong)-1);
+ tcg_gen_movcond_tl(TCG_COND_EQ, source1, cond1, zeroreg, source1,
+ resultopt1);
+ tcg_gen_movi_tl(resultopt1, (target_ulong)1);
+ tcg_gen_movcond_tl(TCG_COND_EQ, source2, cond1, zeroreg, source2,
+ resultopt1);
+ tcg_gen_divu_tl(ret, source1, source2);
+
+ tcg_temp_free(cond1);
+ tcg_temp_free(zeroreg);
+ tcg_temp_free(resultopt1);
+}
+
+static void gen_rem(TCGv ret, TCGv source1, TCGv source2)
+{
+ TCGv cond1, cond2, zeroreg, resultopt1;
+
+ cond1 = tcg_temp_new();
+ cond2 = tcg_temp_new();
+ zeroreg = tcg_const_tl(0);
+ resultopt1 = tcg_temp_new();
+
+ tcg_gen_movi_tl(resultopt1, 1L);
+ tcg_gen_setcondi_tl(TCG_COND_EQ, cond2, source2, (target_ulong)-1);
+ tcg_gen_setcondi_tl(TCG_COND_EQ, cond1, source1,
+ (target_ulong)1 << (TARGET_LONG_BITS - 1));
+ tcg_gen_and_tl(cond2, cond1, cond2); /* cond1 = overflow */
+ tcg_gen_setcondi_tl(TCG_COND_EQ, cond1, source2, 0); /* cond2 = div 0 */
+ /* if overflow or div by zero, set source2 to 1, else don't change */
+ tcg_gen_or_tl(cond2, cond1, cond2);
+ tcg_gen_movcond_tl(TCG_COND_EQ, source2, cond2, zeroreg, source2,
+ resultopt1);
+ tcg_gen_rem_tl(resultopt1, source1, source2);
+ /* if div by zero, just return the original dividend */
+ tcg_gen_movcond_tl(TCG_COND_EQ, ret, cond1, zeroreg, resultopt1,
+ source1);
+
+ tcg_temp_free(cond1);
+ tcg_temp_free(cond2);
+ tcg_temp_free(zeroreg);
+ tcg_temp_free(resultopt1);
+}
+
+static void gen_remu(TCGv ret, TCGv source1, TCGv source2)
+{
+ TCGv cond1, zeroreg, resultopt1;
+ cond1 = tcg_temp_new();
+ zeroreg = tcg_const_tl(0);
+ resultopt1 = tcg_temp_new();
+
+ tcg_gen_movi_tl(resultopt1, (target_ulong)1);
+ tcg_gen_setcondi_tl(TCG_COND_EQ, cond1, source2, 0);
+ tcg_gen_movcond_tl(TCG_COND_EQ, source2, cond1, zeroreg, source2,
+ resultopt1);
+ tcg_gen_remu_tl(resultopt1, source1, source2);
+ /* if div by zero, just return the original dividend */
+ tcg_gen_movcond_tl(TCG_COND_EQ, ret, cond1, zeroreg, resultopt1,
+ source1);
+
+ tcg_temp_free(cond1);
+ tcg_temp_free(zeroreg);
+ tcg_temp_free(resultopt1);
}
static void gen_jal(CPURISCVState *env, DisasContext *ctx, int rd,
@@ -586,6 +540,33 @@ static void gen_addiw(TCGv ret, TCGv arg1, TCGv arg2)
tcg_gen_add_tl(ret, arg1, arg2);
tcg_gen_ext32s_tl(ret, ret);
}
+
+static void gen_mulw(TCGv ret, TCGv arg1, TCGv arg2)
+{
+ tcg_gen_mul_tl(ret, arg1, arg2);
+ tcg_gen_ext32s_tl(ret, ret);
+}
+
+static bool gen_arith_div_w(DisasContext *ctx, arg_r *a,
+ void(*func)(TCGv, TCGv, TCGv))
+{
+ TCGv source1, source2;
+ source1 = tcg_temp_new();
+ source2 = tcg_temp_new();
+
+ gen_get_gpr(source1, a->rs1);
+ gen_get_gpr(source2, a->rs2);
+ tcg_gen_ext32s_tl(source1, source1);
+ tcg_gen_ext32s_tl(source2, source2);
+
+ (*func)(source1, source1, source2);
+
+ tcg_gen_ext32s_tl(source1, source1);
+ gen_set_gpr(a->rd, source1);
+ tcg_temp_free(source1);
+ tcg_temp_free(source2);
+ return true;
+}
#endif
static bool trans_arith(DisasContext *ctx, arg_r *a,
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 27/35] target/riscv: Remove manual decoding of RV32/64M insn
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
v2 -> v3:
- trans_mulw now uses gen_mulw which properly sign extends the result
- gen_arith_w -> gen_arith_div_w
- gen_arith_div_w properly sign extends the resul
target/riscv/insn_trans/trans_rvm.inc.c | 55 +++--
target/riscv/translate.c | 281 +++++++++++-------------
2 files changed, 162 insertions(+), 174 deletions(-)
diff --git a/target/riscv/insn_trans/trans_rvm.inc.c b/target/riscv/insn_trans/trans_rvm.inc.c
index ec3197ede8..949f59ddb2 100644
--- a/target/riscv/insn_trans/trans_rvm.inc.c
+++ b/target/riscv/insn_trans/trans_rvm.inc.c
@@ -21,80 +21,87 @@
static bool trans_mul(DisasContext *ctx, arg_mul *a)
{
- gen_arith(ctx, OPC_RISC_MUL, a->rd, a->rs1, a->rs2);
- return true;
+ return trans_arith(ctx, a, &tcg_gen_mul_tl);
}
static bool trans_mulh(DisasContext *ctx, arg_mulh *a)
{
- gen_arith(ctx, OPC_RISC_MULH, a->rd, a->rs1, a->rs2);
+ TCGv source1 = tcg_temp_new();
+ TCGv source2 = tcg_temp_new();
+ gen_get_gpr(source1, a->rs1);
+ gen_get_gpr(source2, a->rs2);
+
+ tcg_gen_muls2_tl(source2, source1, source1, source2);
+
+ gen_set_gpr(a->rd, source1);
+ tcg_temp_free(source1);
+ tcg_temp_free(source2);
return true;
}
static bool trans_mulhsu(DisasContext *ctx, arg_mulhsu *a)
{
- gen_arith(ctx, OPC_RISC_MULHSU, a->rd, a->rs1, a->rs2);
- return true;
+ return trans_arith(ctx, a, &gen_mulhsu);
}
static bool trans_mulhu(DisasContext *ctx, arg_mulhu *a)
{
- gen_arith(ctx, OPC_RISC_MULHU, a->rd, a->rs1, a->rs2);
+ TCGv source1 = tcg_temp_new();
+ TCGv source2 = tcg_temp_new();
+ gen_get_gpr(source1, a->rs1);
+ gen_get_gpr(source2, a->rs2);
+
+ tcg_gen_mulu2_tl(source2, source1, source1, source2);
+
+ gen_set_gpr(a->rd, source1);
+ tcg_temp_free(source1);
+ tcg_temp_free(source2);
return true;
}
static bool trans_div(DisasContext *ctx, arg_div *a)
{
- gen_arith(ctx, OPC_RISC_DIV, a->rd, a->rs1, a->rs2);
- return true;
+ return trans_arith(ctx, a, &gen_div);
}
static bool trans_divu(DisasContext *ctx, arg_divu *a)
{
- gen_arith(ctx, OPC_RISC_DIVU, a->rd, a->rs1, a->rs2);
- return true;
+ return trans_arith(ctx, a, &gen_divu);
}
static bool trans_rem(DisasContext *ctx, arg_rem *a)
{
- gen_arith(ctx, OPC_RISC_REM, a->rd, a->rs1, a->rs2);
- return true;
+ return trans_arith(ctx, a, &gen_rem);
}
static bool trans_remu(DisasContext *ctx, arg_remu *a)
{
- gen_arith(ctx, OPC_RISC_REMU, a->rd, a->rs1, a->rs2);
- return true;
+ return trans_arith(ctx, a, &gen_remu);
}
#ifdef TARGET_RISCV64
static bool trans_mulw(DisasContext *ctx, arg_mulw *a)
{
- gen_arith(ctx, OPC_RISC_MULW, a->rd, a->rs1, a->rs2);
- return true;
+ return trans_arith(ctx, a, &gen_mulw);
}
static bool trans_divw(DisasContext *ctx, arg_divw *a)
{
- gen_arith(ctx, OPC_RISC_DIVW, a->rd, a->rs1, a->rs2);
- return true;
+ return gen_arith_div_w(ctx, a, &gen_div);
}
static bool trans_divuw(DisasContext *ctx, arg_divuw *a)
{
- gen_arith(ctx, OPC_RISC_DIVUW, a->rd, a->rs1, a->rs2);
- return true;
+ return gen_arith_div_w(ctx, a, &gen_divu);
}
static bool trans_remw(DisasContext *ctx, arg_remw *a)
{
- gen_arith(ctx, OPC_RISC_REMW, a->rd, a->rs1, a->rs2);
- return true;
+ return gen_arith_div_w(ctx, a, &gen_rem);
}
static bool trans_remuw(DisasContext *ctx, arg_remuw *a)
{
- gen_arith(ctx, OPC_RISC_REMUW, a->rd, a->rs1, a->rs2);
- return true;
+ return gen_arith_div_w(ctx, a, &gen_remu);
}
#endif
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 5b4d091561..fd4b63b4c4 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -182,156 +182,110 @@ static void gen_mulhsu(TCGv ret, TCGv arg1, TCGv arg2)
tcg_temp_free(rh);
}
-static void gen_arith(DisasContext *ctx, uint32_t opc, int rd, int rs1,
- int rs2)
-{
- TCGv source1, source2, cond1, cond2, zeroreg, resultopt1;
- source1 = tcg_temp_new();
- source2 = tcg_temp_new();
- gen_get_gpr(source1, rs1);
- gen_get_gpr(source2, rs2);
-
- switch (opc) {
- CASE_OP_32_64(OPC_RISC_MUL):
- tcg_gen_mul_tl(source1, source1, source2);
- break;
- case OPC_RISC_MULH:
- tcg_gen_muls2_tl(source2, source1, source1, source2);
- break;
- case OPC_RISC_MULHSU:
- gen_mulhsu(source1, source1, source2);
- break;
- case OPC_RISC_MULHU:
- tcg_gen_mulu2_tl(source2, source1, source1, source2);
- break;
-#if defined(TARGET_RISCV64)
- case OPC_RISC_DIVW:
- tcg_gen_ext32s_tl(source1, source1);
- tcg_gen_ext32s_tl(source2, source2);
- /* fall through to DIV */
-#endif
- case OPC_RISC_DIV:
- /* Handle by altering args to tcg_gen_div to produce req'd results:
- * For overflow: want source1 in source1 and 1 in source2
- * For div by zero: want -1 in source1 and 1 in source2 -> -1 result */
- cond1 = tcg_temp_new();
- cond2 = tcg_temp_new();
- zeroreg = tcg_const_tl(0);
- resultopt1 = tcg_temp_new();
-
- tcg_gen_movi_tl(resultopt1, (target_ulong)-1);
- tcg_gen_setcondi_tl(TCG_COND_EQ, cond2, source2, (target_ulong)(~0L));
- tcg_gen_setcondi_tl(TCG_COND_EQ, cond1, source1,
- ((target_ulong)1) << (TARGET_LONG_BITS - 1));
- tcg_gen_and_tl(cond1, cond1, cond2); /* cond1 = overflow */
- tcg_gen_setcondi_tl(TCG_COND_EQ, cond2, source2, 0); /* cond2 = div 0 */
- /* if div by zero, set source1 to -1, otherwise don't change */
- tcg_gen_movcond_tl(TCG_COND_EQ, source1, cond2, zeroreg, source1,
- resultopt1);
- /* if overflow or div by zero, set source2 to 1, else don't change */
- tcg_gen_or_tl(cond1, cond1, cond2);
- tcg_gen_movi_tl(resultopt1, (target_ulong)1);
- tcg_gen_movcond_tl(TCG_COND_EQ, source2, cond1, zeroreg, source2,
- resultopt1);
- tcg_gen_div_tl(source1, source1, source2);
-
- tcg_temp_free(cond1);
- tcg_temp_free(cond2);
- tcg_temp_free(zeroreg);
- tcg_temp_free(resultopt1);
- break;
-#if defined(TARGET_RISCV64)
- case OPC_RISC_DIVUW:
- tcg_gen_ext32u_tl(source1, source1);
- tcg_gen_ext32u_tl(source2, source2);
- /* fall through to DIVU */
-#endif
- case OPC_RISC_DIVU:
- cond1 = tcg_temp_new();
- zeroreg = tcg_const_tl(0);
- resultopt1 = tcg_temp_new();
-
- tcg_gen_setcondi_tl(TCG_COND_EQ, cond1, source2, 0);
- tcg_gen_movi_tl(resultopt1, (target_ulong)-1);
- tcg_gen_movcond_tl(TCG_COND_EQ, source1, cond1, zeroreg, source1,
- resultopt1);
- tcg_gen_movi_tl(resultopt1, (target_ulong)1);
- tcg_gen_movcond_tl(TCG_COND_EQ, source2, cond1, zeroreg, source2,
- resultopt1);
- tcg_gen_divu_tl(source1, source1, source2);
-
- tcg_temp_free(cond1);
- tcg_temp_free(zeroreg);
- tcg_temp_free(resultopt1);
- break;
-#if defined(TARGET_RISCV64)
- case OPC_RISC_REMW:
- tcg_gen_ext32s_tl(source1, source1);
- tcg_gen_ext32s_tl(source2, source2);
- /* fall through to REM */
-#endif
- case OPC_RISC_REM:
- cond1 = tcg_temp_new();
- cond2 = tcg_temp_new();
- zeroreg = tcg_const_tl(0);
- resultopt1 = tcg_temp_new();
-
- tcg_gen_movi_tl(resultopt1, 1L);
- tcg_gen_setcondi_tl(TCG_COND_EQ, cond2, source2, (target_ulong)-1);
- tcg_gen_setcondi_tl(TCG_COND_EQ, cond1, source1,
- (target_ulong)1 << (TARGET_LONG_BITS - 1));
- tcg_gen_and_tl(cond2, cond1, cond2); /* cond1 = overflow */
- tcg_gen_setcondi_tl(TCG_COND_EQ, cond1, source2, 0); /* cond2 = div 0 */
- /* if overflow or div by zero, set source2 to 1, else don't change */
- tcg_gen_or_tl(cond2, cond1, cond2);
- tcg_gen_movcond_tl(TCG_COND_EQ, source2, cond2, zeroreg, source2,
- resultopt1);
- tcg_gen_rem_tl(resultopt1, source1, source2);
- /* if div by zero, just return the original dividend */
- tcg_gen_movcond_tl(TCG_COND_EQ, source1, cond1, zeroreg, resultopt1,
- source1);
-
- tcg_temp_free(cond1);
- tcg_temp_free(cond2);
- tcg_temp_free(zeroreg);
- tcg_temp_free(resultopt1);
- break;
-#if defined(TARGET_RISCV64)
- case OPC_RISC_REMUW:
- tcg_gen_ext32u_tl(source1, source1);
- tcg_gen_ext32u_tl(source2, source2);
- /* fall through to REMU */
-#endif
- case OPC_RISC_REMU:
- cond1 = tcg_temp_new();
- zeroreg = tcg_const_tl(0);
- resultopt1 = tcg_temp_new();
-
- tcg_gen_movi_tl(resultopt1, (target_ulong)1);
- tcg_gen_setcondi_tl(TCG_COND_EQ, cond1, source2, 0);
- tcg_gen_movcond_tl(TCG_COND_EQ, source2, cond1, zeroreg, source2,
- resultopt1);
- tcg_gen_remu_tl(resultopt1, source1, source2);
- /* if div by zero, just return the original dividend */
- tcg_gen_movcond_tl(TCG_COND_EQ, source1, cond1, zeroreg, resultopt1,
- source1);
-
- tcg_temp_free(cond1);
- tcg_temp_free(zeroreg);
- tcg_temp_free(resultopt1);
- break;
- default:
- gen_exception_illegal(ctx);
- return;
- }
-
- if (opc & 0x8) { /* sign extend for W instructions */
- tcg_gen_ext32s_tl(source1, source1);
- }
-
- gen_set_gpr(rd, source1);
- tcg_temp_free(source1);
- tcg_temp_free(source2);
+static void gen_div(TCGv ret, TCGv source1, TCGv source2)
+{
+ TCGv cond1, cond2, zeroreg, resultopt1;
+ /* Handle by altering args to tcg_gen_div to produce req'd results:
+ * For overflow: want source1 in source1 and 1 in source2
+ * For div by zero: want -1 in source1 and 1 in source2 -> -1 result */
+ cond1 = tcg_temp_new();
+ cond2 = tcg_temp_new();
+ zeroreg = tcg_const_tl(0);
+ resultopt1 = tcg_temp_new();
+
+ tcg_gen_movi_tl(resultopt1, (target_ulong)-1);
+ tcg_gen_setcondi_tl(TCG_COND_EQ, cond2, source2, (target_ulong)(~0L));
+ tcg_gen_setcondi_tl(TCG_COND_EQ, cond1, source1,
+ ((target_ulong)1) << (TARGET_LONG_BITS - 1));
+ tcg_gen_and_tl(cond1, cond1, cond2); /* cond1 = overflow */
+ tcg_gen_setcondi_tl(TCG_COND_EQ, cond2, source2, 0); /* cond2 = div 0 */
+ /* if div by zero, set source1 to -1, otherwise don't change */
+ tcg_gen_movcond_tl(TCG_COND_EQ, source1, cond2, zeroreg, source1,
+ resultopt1);
+ /* if overflow or div by zero, set source2 to 1, else don't change */
+ tcg_gen_or_tl(cond1, cond1, cond2);
+ tcg_gen_movi_tl(resultopt1, (target_ulong)1);
+ tcg_gen_movcond_tl(TCG_COND_EQ, source2, cond1, zeroreg, source2,
+ resultopt1);
+ tcg_gen_div_tl(ret, source1, source2);
+
+ tcg_temp_free(cond1);
+ tcg_temp_free(cond2);
+ tcg_temp_free(zeroreg);
+ tcg_temp_free(resultopt1);
+}
+
+static void gen_divu(TCGv ret, TCGv source1, TCGv source2)
+{
+ TCGv cond1, zeroreg, resultopt1;
+ cond1 = tcg_temp_new();
+
+ zeroreg = tcg_const_tl(0);
+ resultopt1 = tcg_temp_new();
+
+ tcg_gen_setcondi_tl(TCG_COND_EQ, cond1, source2, 0);
+ tcg_gen_movi_tl(resultopt1, (target_ulong)-1);
+ tcg_gen_movcond_tl(TCG_COND_EQ, source1, cond1, zeroreg, source1,
+ resultopt1);
+ tcg_gen_movi_tl(resultopt1, (target_ulong)1);
+ tcg_gen_movcond_tl(TCG_COND_EQ, source2, cond1, zeroreg, source2,
+ resultopt1);
+ tcg_gen_divu_tl(ret, source1, source2);
+
+ tcg_temp_free(cond1);
+ tcg_temp_free(zeroreg);
+ tcg_temp_free(resultopt1);
+}
+
+static void gen_rem(TCGv ret, TCGv source1, TCGv source2)
+{
+ TCGv cond1, cond2, zeroreg, resultopt1;
+
+ cond1 = tcg_temp_new();
+ cond2 = tcg_temp_new();
+ zeroreg = tcg_const_tl(0);
+ resultopt1 = tcg_temp_new();
+
+ tcg_gen_movi_tl(resultopt1, 1L);
+ tcg_gen_setcondi_tl(TCG_COND_EQ, cond2, source2, (target_ulong)-1);
+ tcg_gen_setcondi_tl(TCG_COND_EQ, cond1, source1,
+ (target_ulong)1 << (TARGET_LONG_BITS - 1));
+ tcg_gen_and_tl(cond2, cond1, cond2); /* cond1 = overflow */
+ tcg_gen_setcondi_tl(TCG_COND_EQ, cond1, source2, 0); /* cond2 = div 0 */
+ /* if overflow or div by zero, set source2 to 1, else don't change */
+ tcg_gen_or_tl(cond2, cond1, cond2);
+ tcg_gen_movcond_tl(TCG_COND_EQ, source2, cond2, zeroreg, source2,
+ resultopt1);
+ tcg_gen_rem_tl(resultopt1, source1, source2);
+ /* if div by zero, just return the original dividend */
+ tcg_gen_movcond_tl(TCG_COND_EQ, ret, cond1, zeroreg, resultopt1,
+ source1);
+
+ tcg_temp_free(cond1);
+ tcg_temp_free(cond2);
+ tcg_temp_free(zeroreg);
+ tcg_temp_free(resultopt1);
+}
+
+static void gen_remu(TCGv ret, TCGv source1, TCGv source2)
+{
+ TCGv cond1, zeroreg, resultopt1;
+ cond1 = tcg_temp_new();
+ zeroreg = tcg_const_tl(0);
+ resultopt1 = tcg_temp_new();
+
+ tcg_gen_movi_tl(resultopt1, (target_ulong)1);
+ tcg_gen_setcondi_tl(TCG_COND_EQ, cond1, source2, 0);
+ tcg_gen_movcond_tl(TCG_COND_EQ, source2, cond1, zeroreg, source2,
+ resultopt1);
+ tcg_gen_remu_tl(resultopt1, source1, source2);
+ /* if div by zero, just return the original dividend */
+ tcg_gen_movcond_tl(TCG_COND_EQ, ret, cond1, zeroreg, resultopt1,
+ source1);
+
+ tcg_temp_free(cond1);
+ tcg_temp_free(zeroreg);
+ tcg_temp_free(resultopt1);
}
static void gen_jal(CPURISCVState *env, DisasContext *ctx, int rd,
@@ -586,6 +540,33 @@ static void gen_addiw(TCGv ret, TCGv arg1, TCGv arg2)
tcg_gen_add_tl(ret, arg1, arg2);
tcg_gen_ext32s_tl(ret, ret);
}
+
+static void gen_mulw(TCGv ret, TCGv arg1, TCGv arg2)
+{
+ tcg_gen_mul_tl(ret, arg1, arg2);
+ tcg_gen_ext32s_tl(ret, ret);
+}
+
+static bool gen_arith_div_w(DisasContext *ctx, arg_r *a,
+ void(*func)(TCGv, TCGv, TCGv))
+{
+ TCGv source1, source2;
+ source1 = tcg_temp_new();
+ source2 = tcg_temp_new();
+
+ gen_get_gpr(source1, a->rs1);
+ gen_get_gpr(source2, a->rs2);
+ tcg_gen_ext32s_tl(source1, source1);
+ tcg_gen_ext32s_tl(source2, source2);
+
+ (*func)(source1, source1, source2);
+
+ tcg_gen_ext32s_tl(source1, source1);
+ gen_set_gpr(a->rd, source1);
+ tcg_temp_free(source1);
+ tcg_temp_free(source2);
+ return true;
+}
#endif
static bool trans_arith(DisasContext *ctx, arg_r *a,
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 28/35] target/riscv: Rename trans_arith to gen_arith
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
---
target/riscv/insn_trans/trans_rvi.inc.c | 14 +++++++-------
target/riscv/insn_trans/trans_rvm.inc.c | 14 +++++++-------
target/riscv/translate.c | 4 ++--
3 files changed, 16 insertions(+), 16 deletions(-)
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
index 65548d1281..227742232c 100644
--- a/target/riscv/insn_trans/trans_rvi.inc.c
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -301,12 +301,12 @@ static bool trans_srai(DisasContext *ctx, arg_srai *a)
static bool trans_add(DisasContext *ctx, arg_add *a)
{
- return trans_arith(ctx, a, &tcg_gen_add_tl);
+ return gen_arith(ctx, a, &tcg_gen_add_tl);
}
static bool trans_sub(DisasContext *ctx, arg_sub *a)
{
- return trans_arith(ctx, a, &tcg_gen_sub_tl);
+ return gen_arith(ctx, a, &tcg_gen_sub_tl);
}
static bool trans_sll(DisasContext *ctx, arg_sll *a)
@@ -348,7 +348,7 @@ static bool trans_sltu(DisasContext *ctx, arg_sltu *a)
static bool trans_xor(DisasContext *ctx, arg_xor *a)
{
- return trans_arith(ctx, a, &tcg_gen_xor_tl);
+ return gen_arith(ctx, a, &tcg_gen_xor_tl);
}
static bool trans_srl(DisasContext *ctx, arg_srl *a)
@@ -363,12 +363,12 @@ static bool trans_sra(DisasContext *ctx, arg_sra *a)
static bool trans_or(DisasContext *ctx, arg_or *a)
{
- return trans_arith(ctx, a, &tcg_gen_or_tl);
+ return gen_arith(ctx, a, &tcg_gen_or_tl);
}
static bool trans_and(DisasContext *ctx, arg_and *a)
{
- return trans_arith(ctx, a, &tcg_gen_and_tl);
+ return gen_arith(ctx, a, &tcg_gen_and_tl);
}
#ifdef TARGET_RISCV64
@@ -417,12 +417,12 @@ static bool trans_sraiw(DisasContext *ctx, arg_sraiw *a)
static bool trans_addw(DisasContext *ctx, arg_addw *a)
{
- return trans_arith(ctx, a, &tcg_gen_add_tl);
+ return gen_arith(ctx, a, &tcg_gen_add_tl);
}
static bool trans_subw(DisasContext *ctx, arg_subw *a)
{
- return trans_arith(ctx, a, &tcg_gen_sub_tl);
+ return gen_arith(ctx, a, &tcg_gen_sub_tl);
}
static bool trans_sllw(DisasContext *ctx, arg_sllw *a)
diff --git a/target/riscv/insn_trans/trans_rvm.inc.c b/target/riscv/insn_trans/trans_rvm.inc.c
index 949f59ddb2..5844d6f5be 100644
--- a/target/riscv/insn_trans/trans_rvm.inc.c
+++ b/target/riscv/insn_trans/trans_rvm.inc.c
@@ -21,7 +21,7 @@
static bool trans_mul(DisasContext *ctx, arg_mul *a)
{
- return trans_arith(ctx, a, &tcg_gen_mul_tl);
+ return gen_arith(ctx, a, &tcg_gen_mul_tl);
}
static bool trans_mulh(DisasContext *ctx, arg_mulh *a)
@@ -41,7 +41,7 @@ static bool trans_mulh(DisasContext *ctx, arg_mulh *a)
static bool trans_mulhsu(DisasContext *ctx, arg_mulhsu *a)
{
- return trans_arith(ctx, a, &gen_mulhsu);
+ return gen_arith(ctx, a, &gen_mulhsu);
}
static bool trans_mulhu(DisasContext *ctx, arg_mulhu *a)
@@ -61,28 +61,28 @@ static bool trans_mulhu(DisasContext *ctx, arg_mulhu *a)
static bool trans_div(DisasContext *ctx, arg_div *a)
{
- return trans_arith(ctx, a, &gen_div);
+ return gen_arith(ctx, a, &gen_div);
}
static bool trans_divu(DisasContext *ctx, arg_divu *a)
{
- return trans_arith(ctx, a, &gen_divu);
+ return gen_arith(ctx, a, &gen_divu);
}
static bool trans_rem(DisasContext *ctx, arg_rem *a)
{
- return trans_arith(ctx, a, &gen_rem);
+ return gen_arith(ctx, a, &gen_rem);
}
static bool trans_remu(DisasContext *ctx, arg_remu *a)
{
- return trans_arith(ctx, a, &gen_remu);
+ return gen_arith(ctx, a, &gen_remu);
}
#ifdef TARGET_RISCV64
static bool trans_mulw(DisasContext *ctx, arg_mulw *a)
{
- return trans_arith(ctx, a, &gen_mulw);
+ return gen_arith(ctx, a, &gen_mulw);
}
static bool trans_divw(DisasContext *ctx, arg_divw *a)
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index fd4b63b4c4..68223925ba 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -569,8 +569,8 @@ static bool gen_arith_div_w(DisasContext *ctx, arg_r *a,
}
#endif
-static bool trans_arith(DisasContext *ctx, arg_r *a,
- void(*func)(TCGv, TCGv, TCGv))
+static bool gen_arith(DisasContext *ctx, arg_r *a,
+ void(*func)(TCGv, TCGv, TCGv))
{
TCGv source1, source2;
source1 = tcg_temp_new();
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 28/35] target/riscv: Rename trans_arith to gen_arith
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
---
target/riscv/insn_trans/trans_rvi.inc.c | 14 +++++++-------
target/riscv/insn_trans/trans_rvm.inc.c | 14 +++++++-------
target/riscv/translate.c | 4 ++--
3 files changed, 16 insertions(+), 16 deletions(-)
diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
index 65548d1281..227742232c 100644
--- a/target/riscv/insn_trans/trans_rvi.inc.c
+++ b/target/riscv/insn_trans/trans_rvi.inc.c
@@ -301,12 +301,12 @@ static bool trans_srai(DisasContext *ctx, arg_srai *a)
static bool trans_add(DisasContext *ctx, arg_add *a)
{
- return trans_arith(ctx, a, &tcg_gen_add_tl);
+ return gen_arith(ctx, a, &tcg_gen_add_tl);
}
static bool trans_sub(DisasContext *ctx, arg_sub *a)
{
- return trans_arith(ctx, a, &tcg_gen_sub_tl);
+ return gen_arith(ctx, a, &tcg_gen_sub_tl);
}
static bool trans_sll(DisasContext *ctx, arg_sll *a)
@@ -348,7 +348,7 @@ static bool trans_sltu(DisasContext *ctx, arg_sltu *a)
static bool trans_xor(DisasContext *ctx, arg_xor *a)
{
- return trans_arith(ctx, a, &tcg_gen_xor_tl);
+ return gen_arith(ctx, a, &tcg_gen_xor_tl);
}
static bool trans_srl(DisasContext *ctx, arg_srl *a)
@@ -363,12 +363,12 @@ static bool trans_sra(DisasContext *ctx, arg_sra *a)
static bool trans_or(DisasContext *ctx, arg_or *a)
{
- return trans_arith(ctx, a, &tcg_gen_or_tl);
+ return gen_arith(ctx, a, &tcg_gen_or_tl);
}
static bool trans_and(DisasContext *ctx, arg_and *a)
{
- return trans_arith(ctx, a, &tcg_gen_and_tl);
+ return gen_arith(ctx, a, &tcg_gen_and_tl);
}
#ifdef TARGET_RISCV64
@@ -417,12 +417,12 @@ static bool trans_sraiw(DisasContext *ctx, arg_sraiw *a)
static bool trans_addw(DisasContext *ctx, arg_addw *a)
{
- return trans_arith(ctx, a, &tcg_gen_add_tl);
+ return gen_arith(ctx, a, &tcg_gen_add_tl);
}
static bool trans_subw(DisasContext *ctx, arg_subw *a)
{
- return trans_arith(ctx, a, &tcg_gen_sub_tl);
+ return gen_arith(ctx, a, &tcg_gen_sub_tl);
}
static bool trans_sllw(DisasContext *ctx, arg_sllw *a)
diff --git a/target/riscv/insn_trans/trans_rvm.inc.c b/target/riscv/insn_trans/trans_rvm.inc.c
index 949f59ddb2..5844d6f5be 100644
--- a/target/riscv/insn_trans/trans_rvm.inc.c
+++ b/target/riscv/insn_trans/trans_rvm.inc.c
@@ -21,7 +21,7 @@
static bool trans_mul(DisasContext *ctx, arg_mul *a)
{
- return trans_arith(ctx, a, &tcg_gen_mul_tl);
+ return gen_arith(ctx, a, &tcg_gen_mul_tl);
}
static bool trans_mulh(DisasContext *ctx, arg_mulh *a)
@@ -41,7 +41,7 @@ static bool trans_mulh(DisasContext *ctx, arg_mulh *a)
static bool trans_mulhsu(DisasContext *ctx, arg_mulhsu *a)
{
- return trans_arith(ctx, a, &gen_mulhsu);
+ return gen_arith(ctx, a, &gen_mulhsu);
}
static bool trans_mulhu(DisasContext *ctx, arg_mulhu *a)
@@ -61,28 +61,28 @@ static bool trans_mulhu(DisasContext *ctx, arg_mulhu *a)
static bool trans_div(DisasContext *ctx, arg_div *a)
{
- return trans_arith(ctx, a, &gen_div);
+ return gen_arith(ctx, a, &gen_div);
}
static bool trans_divu(DisasContext *ctx, arg_divu *a)
{
- return trans_arith(ctx, a, &gen_divu);
+ return gen_arith(ctx, a, &gen_divu);
}
static bool trans_rem(DisasContext *ctx, arg_rem *a)
{
- return trans_arith(ctx, a, &gen_rem);
+ return gen_arith(ctx, a, &gen_rem);
}
static bool trans_remu(DisasContext *ctx, arg_remu *a)
{
- return trans_arith(ctx, a, &gen_remu);
+ return gen_arith(ctx, a, &gen_remu);
}
#ifdef TARGET_RISCV64
static bool trans_mulw(DisasContext *ctx, arg_mulw *a)
{
- return trans_arith(ctx, a, &gen_mulw);
+ return gen_arith(ctx, a, &gen_mulw);
}
static bool trans_divw(DisasContext *ctx, arg_divw *a)
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index fd4b63b4c4..68223925ba 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -569,8 +569,8 @@ static bool gen_arith_div_w(DisasContext *ctx, arg_r *a,
}
#endif
-static bool trans_arith(DisasContext *ctx, arg_r *a,
- void(*func)(TCGv, TCGv, TCGv))
+static bool gen_arith(DisasContext *ctx, arg_r *a,
+ void(*func)(TCGv, TCGv, TCGv))
{
TCGv source1, source2;
source1 = tcg_temp_new();
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 29/35] target/riscv: Remove gen_system()
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
with all 16 bit insns moved to decodetree no path is falling back to
gen_system(), so we can remove it.
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
target/riscv/translate.c | 35 -----------------------------------
1 file changed, 35 deletions(-)
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 68223925ba..0ff3da641e 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -425,34 +425,6 @@ static void gen_set_rm(DisasContext *ctx, int rm)
tcg_temp_free_i32(t0);
}
-
-static void gen_system(CPURISCVState *env, DisasContext *ctx, uint32_t opc,
- int rd, int rs1, int csr)
-{
- tcg_gen_movi_tl(cpu_pc, ctx->base.pc_next);
-
- switch (opc) {
- case OPC_RISC_ECALL:
- switch (csr) {
- case 0x0: /* ECALL */
- /* always generates U-level ECALL, fixed in do_interrupt handler */
- generate_exception(ctx, RISCV_EXCP_U_ECALL);
- tcg_gen_exit_tb(NULL, 0); /* no chaining */
- ctx->base.is_jmp = DISAS_NORETURN;
- break;
- case 0x1: /* EBREAK */
- generate_exception(ctx, RISCV_EXCP_BREAKPOINT);
- tcg_gen_exit_tb(NULL, 0); /* no chaining */
- ctx->base.is_jmp = DISAS_NORETURN;
- break;
- default:
- gen_exception_illegal(ctx);
- break;
- }
- break;
- }
-}
-
static void decode_RV32_64C0(DisasContext *ctx)
{
uint8_t funct3 = extract32(ctx->opcode, 13, 3);
@@ -620,7 +592,6 @@ bool decode_insn16(DisasContext *ctx, uint16_t insn);
static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
{
- int rs1, rd;
uint32_t op;
/* We do not do misaligned address check here: the address should never be
@@ -629,14 +600,8 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
* perform the misaligned instruction fetch */
op = MASK_OP_MAJOR(ctx->opcode);
- rs1 = GET_RS1(ctx->opcode);
- rd = GET_RD(ctx->opcode);
switch (op) {
- case OPC_RISC_SYSTEM:
- gen_system(env, ctx, MASK_OP_SYSTEM(ctx->opcode), rd, rs1,
- (ctx->opcode & 0xFFF00000) >> 20);
- break;
default:
gen_exception_illegal(ctx);
break;
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 29/35] target/riscv: Remove gen_system()
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
with all 16 bit insns moved to decodetree no path is falling back to
gen_system(), so we can remove it.
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
target/riscv/translate.c | 35 -----------------------------------
1 file changed, 35 deletions(-)
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 68223925ba..0ff3da641e 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -425,34 +425,6 @@ static void gen_set_rm(DisasContext *ctx, int rm)
tcg_temp_free_i32(t0);
}
-
-static void gen_system(CPURISCVState *env, DisasContext *ctx, uint32_t opc,
- int rd, int rs1, int csr)
-{
- tcg_gen_movi_tl(cpu_pc, ctx->base.pc_next);
-
- switch (opc) {
- case OPC_RISC_ECALL:
- switch (csr) {
- case 0x0: /* ECALL */
- /* always generates U-level ECALL, fixed in do_interrupt handler */
- generate_exception(ctx, RISCV_EXCP_U_ECALL);
- tcg_gen_exit_tb(NULL, 0); /* no chaining */
- ctx->base.is_jmp = DISAS_NORETURN;
- break;
- case 0x1: /* EBREAK */
- generate_exception(ctx, RISCV_EXCP_BREAKPOINT);
- tcg_gen_exit_tb(NULL, 0); /* no chaining */
- ctx->base.is_jmp = DISAS_NORETURN;
- break;
- default:
- gen_exception_illegal(ctx);
- break;
- }
- break;
- }
-}
-
static void decode_RV32_64C0(DisasContext *ctx)
{
uint8_t funct3 = extract32(ctx->opcode, 13, 3);
@@ -620,7 +592,6 @@ bool decode_insn16(DisasContext *ctx, uint16_t insn);
static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
{
- int rs1, rd;
uint32_t op;
/* We do not do misaligned address check here: the address should never be
@@ -629,14 +600,8 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
* perform the misaligned instruction fetch */
op = MASK_OP_MAJOR(ctx->opcode);
- rs1 = GET_RS1(ctx->opcode);
- rd = GET_RD(ctx->opcode);
switch (op) {
- case OPC_RISC_SYSTEM:
- gen_system(env, ctx, MASK_OP_SYSTEM(ctx->opcode), rd, rs1,
- (ctx->opcode & 0xFFF00000) >> 20);
- break;
default:
gen_exception_illegal(ctx);
break;
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 30/35] target/riscv: Remove decode_RV32_64G()
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
decodetree handles all instructions now so the fallback is not necessary
anymore.
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
target/riscv/translate.c | 21 +--------------------
1 file changed, 1 insertion(+), 20 deletions(-)
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 0ff3da641e..38e0ff32bb 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -590,24 +590,6 @@ bool decode_insn16(DisasContext *ctx, uint16_t insn);
#include "decode_insn16.inc.c"
#include "insn_trans/trans_rvc.inc.c"
-static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
-{
- uint32_t op;
-
- /* We do not do misaligned address check here: the address should never be
- * misaligned at this point. Instructions that set PC must do the check,
- * since epc must be the address of the instruction that caused us to
- * perform the misaligned instruction fetch */
-
- op = MASK_OP_MAJOR(ctx->opcode);
-
- switch (op) {
- default:
- gen_exception_illegal(ctx);
- break;
- }
-}
-
static void decode_opc(DisasContext *ctx)
{
/* check for compressed insn */
@@ -624,8 +606,7 @@ static void decode_opc(DisasContext *ctx)
} else {
ctx->pc_succ_insn = ctx->base.pc_next + 4;
if (!decode_insn32(ctx, ctx->opcode)) {
- /* fallback to old decoder */
- decode_RV32_64G(ctx->env, ctx);
+ gen_exception_illegal(ctx);
}
}
}
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 30/35] target/riscv: Remove decode_RV32_64G()
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
decodetree handles all instructions now so the fallback is not necessary
anymore.
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
---
target/riscv/translate.c | 21 +--------------------
1 file changed, 1 insertion(+), 20 deletions(-)
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 0ff3da641e..38e0ff32bb 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -590,24 +590,6 @@ bool decode_insn16(DisasContext *ctx, uint16_t insn);
#include "decode_insn16.inc.c"
#include "insn_trans/trans_rvc.inc.c"
-static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
-{
- uint32_t op;
-
- /* We do not do misaligned address check here: the address should never be
- * misaligned at this point. Instructions that set PC must do the check,
- * since epc must be the address of the instruction that caused us to
- * perform the misaligned instruction fetch */
-
- op = MASK_OP_MAJOR(ctx->opcode);
-
- switch (op) {
- default:
- gen_exception_illegal(ctx);
- break;
- }
-}
-
static void decode_opc(DisasContext *ctx)
{
/* check for compressed insn */
@@ -624,8 +606,7 @@ static void decode_opc(DisasContext *ctx)
} else {
ctx->pc_succ_insn = ctx->base.pc_next + 4;
if (!decode_insn32(ctx, ctx->opcode)) {
- /* fallback to old decoder */
- decode_RV32_64G(ctx->env, ctx);
+ gen_exception_illegal(ctx);
}
}
}
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 31/35] target/riscv: Convert @cs_2 insns to share translation functions
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
These all expand simply to R format instructions.
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
---
target/riscv/Makefile.objs | 10 +++---
target/riscv/insn16-64.decode | 24 ++++++++++++++
target/riscv/insn16.decode | 15 +++++----
target/riscv/insn_trans/trans_rvc.inc.c | 44 -------------------------
target/riscv/translate.c | 20 ++++++++---
5 files changed, 54 insertions(+), 59 deletions(-)
create mode 100644 target/riscv/insn16-64.decode
diff --git a/target/riscv/Makefile.objs b/target/riscv/Makefile.objs
index ba661b69e3..515c12b4f8 100644
--- a/target/riscv/Makefile.objs
+++ b/target/riscv/Makefile.objs
@@ -5,16 +5,18 @@ DECODETREE = $(SRC_PATH)/scripts/decodetree.py
decode32-y = $(SRC_PATH)/target/riscv/insn32.decode
decode32-$(TARGET_RISCV64) += $(SRC_PATH)/target/riscv/insn64.decode
+decode16-y = $(SRC_PATH)/target/riscv/insn16.decode
+decode16-$(TARGET_RISCV64) += $(SRC_PATH)/target/riscv/insn16-64.decode
+
target/riscv/decode_insn32.inc.c: $(decode32-y) $(DECODETREE)
$(call quiet-command, \
$(PYTHON) $(DECODETREE) -o $@ --decode decode_insn32 $(decode32-y), \
"GEN", $(TARGET_DIR)$@)
-target/riscv/decode_insn16.inc.c: \
- $(SRC_PATH)/target/riscv/insn16.decode $(DECODETREE)
+target/riscv/decode_insn16.inc.c: $(decode16-y) $(DECODETREE)
$(call quiet-command, \
- $(PYTHON) $(DECODETREE) -o $@ --decode decode_insn16 --insnwidth 16 $<, \
- "GEN", $(TARGET_DIR)$@)
+ $(PYTHON) $(DECODETREE) -o $@ --decode decode_insn16 --insnwidth 16 \
+ $(decode16-y), "GEN", $(TARGET_DIR)$@)
target/riscv/translate.o: target/riscv/decode_insn32.inc.c \
target/riscv/decode_insn16.inc.c
diff --git a/target/riscv/insn16-64.decode b/target/riscv/insn16-64.decode
new file mode 100644
index 0000000000..5af2e2b072
--- /dev/null
+++ b/target/riscv/insn16-64.decode
@@ -0,0 +1,24 @@
+#
+# RISC-V translation routines for the RVC Instruction Set.
+#
+# Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
+# Bastian Koppelmann, kbastian@mail.uni-paderborn.de
+#
+# This program is free software; you can redistribute it and/or modify it
+# under the terms and conditions of the GNU General Public License,
+# version 2 or later, as published by the Free Software Foundation.
+#
+# This program is distributed in the hope it will be useful, but WITHOUT
+# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+# FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+# more details.
+#
+# You should have received a copy of the GNU General Public License along with
+# this program. If not, see <http://www.gnu.org/licenses/>.
+
+# This is concatenated with insn16.decode for risc64 targets.
+# All of the fields and formats are there.
+
+# *** RV64C Standard Extension (Quadrant 1) ***
+subw 100 1 11 ... 00 ... 01 @cs_2
+addw 100 1 11 ... 01 ... 01 @cs_2
diff --git a/target/riscv/insn16.decode b/target/riscv/insn16.decode
index 0829e3bc59..c7a58d80e5 100644
--- a/target/riscv/insn16.decode
+++ b/target/riscv/insn16.decode
@@ -41,6 +41,9 @@
+# Argument sets imported from insn32.decode:
+&r rd rs1 rs2 !extern
+
# Argument sets:
&cl rs1 rd
&cl_dw uimm rs1 rd
@@ -68,7 +71,7 @@
@cl_w ... ... ... .. ... .. &cl_dw uimm=%uimm_cl_w rs1=%rs1_3 rd=%rs2_3
@cl ... ... ... .. ... .. &cl rs1=%rs1_3 rd=%rs2_3
@cs ... ... ... .. ... .. &cs rs1=%rs1_3 rs2=%rs2_3
-@cs_2 ... ... ... .. ... .. &cr rd=%rs1_3 rs2=%rs2_3
+@cs_2 ... ... ... .. ... .. &r rd=%rs1_3 rs1=%rs1_3 rs2=%rs2_3
@cs_d ... ... ... .. ... .. &cs_dw uimm=%uimm_cl_d rs1=%rs1_3 rs2=%rs2_3
@cs_w ... ... ... .. ... .. &cs_dw uimm=%uimm_cl_w rs1=%rs1_3 rs2=%rs2_3
@cb ... ... ... .. ... .. &cb imm=%imm_cb rs1=%rs1_3
@@ -107,12 +110,10 @@ c_addi16sp_lui 011 . ..... ..... 01 @c_addi16sp_lui # shares opc with C.LUI
c_srli 100 . 00 ... ..... 01 @c_shift
c_srai 100 . 01 ... ..... 01 @c_shift
c_andi 100 . 10 ... ..... 01 @c_andi
-c_sub 100 0 11 ... 00 ... 01 @cs_2
-c_xor 100 0 11 ... 01 ... 01 @cs_2
-c_or 100 0 11 ... 10 ... 01 @cs_2
-c_and 100 0 11 ... 11 ... 01 @cs_2
-c_subw 100 1 11 ... 00 ... 01 @cs_2
-c_addw 100 1 11 ... 01 ... 01 @cs_2
+sub 100 0 11 ... 00 ... 01 @cs_2
+xor 100 0 11 ... 01 ... 01 @cs_2
+or 100 0 11 ... 10 ... 01 @cs_2
+and 100 0 11 ... 11 ... 01 @cs_2
c_j 101 ........... 01 @cj
c_beqz 110 ... ... ..... 01 @cb
c_bnez 111 ... ... ..... 01 @cb
diff --git a/target/riscv/insn_trans/trans_rvc.inc.c b/target/riscv/insn_trans/trans_rvc.inc.c
index bcdf64d3b7..639c381edf 100644
--- a/target/riscv/insn_trans/trans_rvc.inc.c
+++ b/target/riscv/insn_trans/trans_rvc.inc.c
@@ -163,50 +163,6 @@ static bool trans_c_andi(DisasContext *ctx, arg_c_andi *a)
return trans_andi(ctx, &arg);
}
-static bool trans_c_sub(DisasContext *ctx, arg_c_sub *a)
-{
- arg_sub arg = { .rd = a->rd, .rs1 = a->rd, .rs2 = a->rs2 };
- return trans_sub(ctx, &arg);
-}
-
-static bool trans_c_xor(DisasContext *ctx, arg_c_xor *a)
-{
- arg_xor arg = { .rd = a->rd, .rs1 = a->rd, .rs2 = a->rs2 };
- return trans_xor(ctx, &arg);
-}
-
-static bool trans_c_or(DisasContext *ctx, arg_c_or *a)
-{
- arg_or arg = { .rd = a->rd, .rs1 = a->rd, .rs2 = a->rs2 };
- return trans_or(ctx, &arg);
-}
-
-static bool trans_c_and(DisasContext *ctx, arg_c_and *a)
-{
- arg_and arg = { .rd = a->rd, .rs1 = a->rd, .rs2 = a->rs2 };
- return trans_and(ctx, &arg);
-}
-
-static bool trans_c_subw(DisasContext *ctx, arg_c_subw *a)
-{
-#ifdef TARGET_RISCV64
- arg_subw arg = { .rd = a->rd, .rs1 = a->rd, .rs2 = a->rs2 };
- return trans_subw(ctx, &arg);
-#else
- return false;
-#endif
-}
-
-static bool trans_c_addw(DisasContext *ctx, arg_c_addw *a)
-{
-#ifdef TARGET_RISCV64
- arg_addw arg = { .rd = a->rd, .rs1 = a->rd, .rs2 = a->rs2 };
- return trans_addw(ctx, &arg);
-#else
- return false;
-#endif
-}
-
static bool trans_c_j(DisasContext *ctx, arg_c_j *a)
{
arg_jal arg = { .rd = 0, .imm = a->imm };
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 38e0ff32bb..56b650a9d7 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -484,10 +484,25 @@ static int ex_rvc_register(int reg)
return 8 + reg;
}
+/*
+ * Include the auto-generated decoders.
+ * Note that the 16-bit decoder reuses some of the trans_* functions
+ * from the 32-bit decoder, which results in duplicate declarations
+ * of the relevant helpers. Suppress the warning.
+ */
bool decode_insn32(DisasContext *ctx, uint32_t insn);
-/* Include the auto-generated decoder for 32 bit insn */
+bool decode_insn16(DisasContext *ctx, uint16_t insn);
+
#include "decode_insn32.inc.c"
+#pragma GCC diagnostic push
+#pragma GCC diagnostic ignored "-Wredundant-decls"
+
+#include "decode_insn16.inc.c"
+
+#pragma GCC diagnostic pop
+
+
static bool gen_arith_imm(DisasContext *ctx, arg_i *a,
void(*func)(TCGv, TCGv, TCGv))
{
@@ -585,9 +600,6 @@ static bool gen_shift(DisasContext *ctx, arg_r *a,
#include "insn_trans/trans_rvd.inc.c"
#include "insn_trans/trans_privileged.inc.c"
-bool decode_insn16(DisasContext *ctx, uint16_t insn);
-/* auto-generated decoder*/
-#include "decode_insn16.inc.c"
#include "insn_trans/trans_rvc.inc.c"
static void decode_opc(DisasContext *ctx)
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 31/35] target/riscv: Convert @cs_2 insns to share translation functions
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
These all expand simply to R format instructions.
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
---
target/riscv/Makefile.objs | 10 +++---
target/riscv/insn16-64.decode | 24 ++++++++++++++
target/riscv/insn16.decode | 15 +++++----
target/riscv/insn_trans/trans_rvc.inc.c | 44 -------------------------
target/riscv/translate.c | 20 ++++++++---
5 files changed, 54 insertions(+), 59 deletions(-)
create mode 100644 target/riscv/insn16-64.decode
diff --git a/target/riscv/Makefile.objs b/target/riscv/Makefile.objs
index ba661b69e3..515c12b4f8 100644
--- a/target/riscv/Makefile.objs
+++ b/target/riscv/Makefile.objs
@@ -5,16 +5,18 @@ DECODETREE = $(SRC_PATH)/scripts/decodetree.py
decode32-y = $(SRC_PATH)/target/riscv/insn32.decode
decode32-$(TARGET_RISCV64) += $(SRC_PATH)/target/riscv/insn64.decode
+decode16-y = $(SRC_PATH)/target/riscv/insn16.decode
+decode16-$(TARGET_RISCV64) += $(SRC_PATH)/target/riscv/insn16-64.decode
+
target/riscv/decode_insn32.inc.c: $(decode32-y) $(DECODETREE)
$(call quiet-command, \
$(PYTHON) $(DECODETREE) -o $@ --decode decode_insn32 $(decode32-y), \
"GEN", $(TARGET_DIR)$@)
-target/riscv/decode_insn16.inc.c: \
- $(SRC_PATH)/target/riscv/insn16.decode $(DECODETREE)
+target/riscv/decode_insn16.inc.c: $(decode16-y) $(DECODETREE)
$(call quiet-command, \
- $(PYTHON) $(DECODETREE) -o $@ --decode decode_insn16 --insnwidth 16 $<, \
- "GEN", $(TARGET_DIR)$@)
+ $(PYTHON) $(DECODETREE) -o $@ --decode decode_insn16 --insnwidth 16 \
+ $(decode16-y), "GEN", $(TARGET_DIR)$@)
target/riscv/translate.o: target/riscv/decode_insn32.inc.c \
target/riscv/decode_insn16.inc.c
diff --git a/target/riscv/insn16-64.decode b/target/riscv/insn16-64.decode
new file mode 100644
index 0000000000..5af2e2b072
--- /dev/null
+++ b/target/riscv/insn16-64.decode
@@ -0,0 +1,24 @@
+#
+# RISC-V translation routines for the RVC Instruction Set.
+#
+# Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
+# Bastian Koppelmann, kbastian@mail.uni-paderborn.de
+#
+# This program is free software; you can redistribute it and/or modify it
+# under the terms and conditions of the GNU General Public License,
+# version 2 or later, as published by the Free Software Foundation.
+#
+# This program is distributed in the hope it will be useful, but WITHOUT
+# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+# FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+# more details.
+#
+# You should have received a copy of the GNU General Public License along with
+# this program. If not, see <http://www.gnu.org/licenses/>.
+
+# This is concatenated with insn16.decode for risc64 targets.
+# All of the fields and formats are there.
+
+# *** RV64C Standard Extension (Quadrant 1) ***
+subw 100 1 11 ... 00 ... 01 @cs_2
+addw 100 1 11 ... 01 ... 01 @cs_2
diff --git a/target/riscv/insn16.decode b/target/riscv/insn16.decode
index 0829e3bc59..c7a58d80e5 100644
--- a/target/riscv/insn16.decode
+++ b/target/riscv/insn16.decode
@@ -41,6 +41,9 @@
+# Argument sets imported from insn32.decode:
+&r rd rs1 rs2 !extern
+
# Argument sets:
&cl rs1 rd
&cl_dw uimm rs1 rd
@@ -68,7 +71,7 @@
@cl_w ... ... ... .. ... .. &cl_dw uimm=%uimm_cl_w rs1=%rs1_3 rd=%rs2_3
@cl ... ... ... .. ... .. &cl rs1=%rs1_3 rd=%rs2_3
@cs ... ... ... .. ... .. &cs rs1=%rs1_3 rs2=%rs2_3
-@cs_2 ... ... ... .. ... .. &cr rd=%rs1_3 rs2=%rs2_3
+@cs_2 ... ... ... .. ... .. &r rd=%rs1_3 rs1=%rs1_3 rs2=%rs2_3
@cs_d ... ... ... .. ... .. &cs_dw uimm=%uimm_cl_d rs1=%rs1_3 rs2=%rs2_3
@cs_w ... ... ... .. ... .. &cs_dw uimm=%uimm_cl_w rs1=%rs1_3 rs2=%rs2_3
@cb ... ... ... .. ... .. &cb imm=%imm_cb rs1=%rs1_3
@@ -107,12 +110,10 @@ c_addi16sp_lui 011 . ..... ..... 01 @c_addi16sp_lui # shares opc with C.LUI
c_srli 100 . 00 ... ..... 01 @c_shift
c_srai 100 . 01 ... ..... 01 @c_shift
c_andi 100 . 10 ... ..... 01 @c_andi
-c_sub 100 0 11 ... 00 ... 01 @cs_2
-c_xor 100 0 11 ... 01 ... 01 @cs_2
-c_or 100 0 11 ... 10 ... 01 @cs_2
-c_and 100 0 11 ... 11 ... 01 @cs_2
-c_subw 100 1 11 ... 00 ... 01 @cs_2
-c_addw 100 1 11 ... 01 ... 01 @cs_2
+sub 100 0 11 ... 00 ... 01 @cs_2
+xor 100 0 11 ... 01 ... 01 @cs_2
+or 100 0 11 ... 10 ... 01 @cs_2
+and 100 0 11 ... 11 ... 01 @cs_2
c_j 101 ........... 01 @cj
c_beqz 110 ... ... ..... 01 @cb
c_bnez 111 ... ... ..... 01 @cb
diff --git a/target/riscv/insn_trans/trans_rvc.inc.c b/target/riscv/insn_trans/trans_rvc.inc.c
index bcdf64d3b7..639c381edf 100644
--- a/target/riscv/insn_trans/trans_rvc.inc.c
+++ b/target/riscv/insn_trans/trans_rvc.inc.c
@@ -163,50 +163,6 @@ static bool trans_c_andi(DisasContext *ctx, arg_c_andi *a)
return trans_andi(ctx, &arg);
}
-static bool trans_c_sub(DisasContext *ctx, arg_c_sub *a)
-{
- arg_sub arg = { .rd = a->rd, .rs1 = a->rd, .rs2 = a->rs2 };
- return trans_sub(ctx, &arg);
-}
-
-static bool trans_c_xor(DisasContext *ctx, arg_c_xor *a)
-{
- arg_xor arg = { .rd = a->rd, .rs1 = a->rd, .rs2 = a->rs2 };
- return trans_xor(ctx, &arg);
-}
-
-static bool trans_c_or(DisasContext *ctx, arg_c_or *a)
-{
- arg_or arg = { .rd = a->rd, .rs1 = a->rd, .rs2 = a->rs2 };
- return trans_or(ctx, &arg);
-}
-
-static bool trans_c_and(DisasContext *ctx, arg_c_and *a)
-{
- arg_and arg = { .rd = a->rd, .rs1 = a->rd, .rs2 = a->rs2 };
- return trans_and(ctx, &arg);
-}
-
-static bool trans_c_subw(DisasContext *ctx, arg_c_subw *a)
-{
-#ifdef TARGET_RISCV64
- arg_subw arg = { .rd = a->rd, .rs1 = a->rd, .rs2 = a->rs2 };
- return trans_subw(ctx, &arg);
-#else
- return false;
-#endif
-}
-
-static bool trans_c_addw(DisasContext *ctx, arg_c_addw *a)
-{
-#ifdef TARGET_RISCV64
- arg_addw arg = { .rd = a->rd, .rs1 = a->rd, .rs2 = a->rs2 };
- return trans_addw(ctx, &arg);
-#else
- return false;
-#endif
-}
-
static bool trans_c_j(DisasContext *ctx, arg_c_j *a)
{
arg_jal arg = { .rd = 0, .imm = a->imm };
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 38e0ff32bb..56b650a9d7 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -484,10 +484,25 @@ static int ex_rvc_register(int reg)
return 8 + reg;
}
+/*
+ * Include the auto-generated decoders.
+ * Note that the 16-bit decoder reuses some of the trans_* functions
+ * from the 32-bit decoder, which results in duplicate declarations
+ * of the relevant helpers. Suppress the warning.
+ */
bool decode_insn32(DisasContext *ctx, uint32_t insn);
-/* Include the auto-generated decoder for 32 bit insn */
+bool decode_insn16(DisasContext *ctx, uint16_t insn);
+
#include "decode_insn32.inc.c"
+#pragma GCC diagnostic push
+#pragma GCC diagnostic ignored "-Wredundant-decls"
+
+#include "decode_insn16.inc.c"
+
+#pragma GCC diagnostic pop
+
+
static bool gen_arith_imm(DisasContext *ctx, arg_i *a,
void(*func)(TCGv, TCGv, TCGv))
{
@@ -585,9 +600,6 @@ static bool gen_shift(DisasContext *ctx, arg_r *a,
#include "insn_trans/trans_rvd.inc.c"
#include "insn_trans/trans_privileged.inc.c"
-bool decode_insn16(DisasContext *ctx, uint16_t insn);
-/* auto-generated decoder*/
-#include "decode_insn16.inc.c"
#include "insn_trans/trans_rvc.inc.c"
static void decode_opc(DisasContext *ctx)
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 32/35] target/riscv: Convert @cl_d, @cl_w, @cs_d, @cs_w insns
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
---
target/riscv/insn16.decode | 20 ++++++++++----------
target/riscv/insn32.decode | 3 ++-
target/riscv/insn_trans/trans_rvc.inc.c | 24 ------------------------
3 files changed, 12 insertions(+), 35 deletions(-)
diff --git a/target/riscv/insn16.decode b/target/riscv/insn16.decode
index c7a58d80e5..c215867ff9 100644
--- a/target/riscv/insn16.decode
+++ b/target/riscv/insn16.decode
@@ -43,14 +43,14 @@
# Argument sets imported from insn32.decode:
&r rd rs1 rs2 !extern
+&i imm rs1 rd !extern
+&s imm rs1 rs2 !extern
# Argument sets:
&cl rs1 rd
-&cl_dw uimm rs1 rd
&ci imm rd
&ciw nzuimm rd
&cs rs1 rs2
-&cs_dw uimm rs1 rs2
&cb imm rs1
&cr rd rs2
&c_j imm
@@ -67,13 +67,13 @@
@cr .... ..... ..... .. &cr rs2=%rs2_5 %rd
@ci ... . ..... ..... .. &ci imm=%imm_ci %rd
@ciw ... ........ ... .. &ciw nzuimm=%nzuimm_ciw rd=%rs2_3
-@cl_d ... ... ... .. ... .. &cl_dw uimm=%uimm_cl_d rs1=%rs1_3 rd=%rs2_3
-@cl_w ... ... ... .. ... .. &cl_dw uimm=%uimm_cl_w rs1=%rs1_3 rd=%rs2_3
+@cl_d ... ... ... .. ... .. &i imm=%uimm_cl_d rs1=%rs1_3 rd=%rs2_3
+@cl_w ... ... ... .. ... .. &i imm=%uimm_cl_w rs1=%rs1_3 rd=%rs2_3
@cl ... ... ... .. ... .. &cl rs1=%rs1_3 rd=%rs2_3
@cs ... ... ... .. ... .. &cs rs1=%rs1_3 rs2=%rs2_3
@cs_2 ... ... ... .. ... .. &r rd=%rs1_3 rs1=%rs1_3 rs2=%rs2_3
-@cs_d ... ... ... .. ... .. &cs_dw uimm=%uimm_cl_d rs1=%rs1_3 rs2=%rs2_3
-@cs_w ... ... ... .. ... .. &cs_dw uimm=%uimm_cl_w rs1=%rs1_3 rs2=%rs2_3
+@cs_d ... ... ... .. ... .. &s imm=%uimm_cl_d rs1=%rs1_3 rs2=%rs2_3
+@cs_w ... ... ... .. ... .. &s imm=%uimm_cl_w rs1=%rs1_3 rs2=%rs2_3
@cb ... ... ... .. ... .. &cb imm=%imm_cb rs1=%rs1_3
@cj ... ........... .. &c_j imm=%imm_cj
@@ -95,11 +95,11 @@
# *** RV64C Standard Extension (Quadrant 0) ***
c_addi4spn 000 ........ ... 00 @ciw
-c_fld 001 ... ... .. ... 00 @cl_d
-c_lw 010 ... ... .. ... 00 @cl_w
+fld 001 ... ... .. ... 00 @cl_d
+lw 010 ... ... .. ... 00 @cl_w
c_flw_ld 011 --- ... -- ... 00 @cl #Note: Must parse uimm manually
-c_fsd 101 ... ... .. ... 00 @cs_d
-c_sw 110 ... ... .. ... 00 @cs_w
+fsd 101 ... ... .. ... 00 @cs_d
+sw 110 ... ... .. ... 00 @cs_w
c_fsw_sd 111 --- ... -- ... 00 @cs #Note: Must parse uimm manually
# *** RV64C Standard Extension (Quadrant 1) ***
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index 520954e9c6..9f13fbcab9 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -37,6 +37,7 @@
&b imm rs2 rs1
&i imm rs1 rd
&r rd rs1 rs2
+&s imm rs2 rs1
&shift shamt rs1 rd
&atomic aq rl rs2 rs1 rd
@@ -44,7 +45,7 @@
@r ....... ..... ..... ... ..... ....... &r %rs2 %rs1 %rd
@i ............ ..... ... ..... ....... &i imm=%imm_i %rs1 %rd
@b ....... ..... ..... ... ..... ....... &b imm=%imm_b %rs2 %rs1
-@s ....... ..... ..... ... ..... ....... imm=%imm_s %rs2 %rs1
+@s ....... ..... ..... ... ..... ....... &s imm=%imm_s %rs2 %rs1
@u .................... ..... ....... imm=%imm_u %rd
@j .................... ..... ....... imm=%imm_j %rd
diff --git a/target/riscv/insn_trans/trans_rvc.inc.c b/target/riscv/insn_trans/trans_rvc.inc.c
index 639c381edf..d932bfd3e0 100644
--- a/target/riscv/insn_trans/trans_rvc.inc.c
+++ b/target/riscv/insn_trans/trans_rvc.inc.c
@@ -28,18 +28,6 @@ static bool trans_c_addi4spn(DisasContext *ctx, arg_c_addi4spn *a)
return trans_addi(ctx, &arg);
}
-static bool trans_c_fld(DisasContext *ctx, arg_c_fld *a)
-{
- arg_fld arg = { .rd = a->rd, .rs1 = a->rs1, .imm = a->uimm };
- return trans_fld(ctx, &arg);
-}
-
-static bool trans_c_lw(DisasContext *ctx, arg_c_lw *a)
-{
- arg_lw arg = { .rd = a->rd, .rs1 = a->rs1, .imm = a->uimm };
- return trans_lw(ctx, &arg);
-}
-
static bool trans_c_flw_ld(DisasContext *ctx, arg_c_flw_ld *a)
{
#ifdef TARGET_RISCV32
@@ -51,18 +39,6 @@ static bool trans_c_flw_ld(DisasContext *ctx, arg_c_flw_ld *a)
#endif
}
-static bool trans_c_fsd(DisasContext *ctx, arg_c_fsd *a)
-{
- arg_fsd arg = { .rs1 = a->rs1, .rs2 = a->rs2, .imm = a->uimm };
- return trans_fsd(ctx, &arg);
-}
-
-static bool trans_c_sw(DisasContext *ctx, arg_c_sw *a)
-{
- arg_sw arg = { .rs1 = a->rs1, .rs2 = a->rs2, .imm = a->uimm };
- return trans_sw(ctx, &arg);
-}
-
static bool trans_c_fsw_sd(DisasContext *ctx, arg_c_fsw_sd *a)
{
#ifdef TARGET_RISCV32
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 32/35] target/riscv: Convert @cl_d, @cl_w, @cs_d, @cs_w insns
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
---
target/riscv/insn16.decode | 20 ++++++++++----------
target/riscv/insn32.decode | 3 ++-
target/riscv/insn_trans/trans_rvc.inc.c | 24 ------------------------
3 files changed, 12 insertions(+), 35 deletions(-)
diff --git a/target/riscv/insn16.decode b/target/riscv/insn16.decode
index c7a58d80e5..c215867ff9 100644
--- a/target/riscv/insn16.decode
+++ b/target/riscv/insn16.decode
@@ -43,14 +43,14 @@
# Argument sets imported from insn32.decode:
&r rd rs1 rs2 !extern
+&i imm rs1 rd !extern
+&s imm rs1 rs2 !extern
# Argument sets:
&cl rs1 rd
-&cl_dw uimm rs1 rd
&ci imm rd
&ciw nzuimm rd
&cs rs1 rs2
-&cs_dw uimm rs1 rs2
&cb imm rs1
&cr rd rs2
&c_j imm
@@ -67,13 +67,13 @@
@cr .... ..... ..... .. &cr rs2=%rs2_5 %rd
@ci ... . ..... ..... .. &ci imm=%imm_ci %rd
@ciw ... ........ ... .. &ciw nzuimm=%nzuimm_ciw rd=%rs2_3
-@cl_d ... ... ... .. ... .. &cl_dw uimm=%uimm_cl_d rs1=%rs1_3 rd=%rs2_3
-@cl_w ... ... ... .. ... .. &cl_dw uimm=%uimm_cl_w rs1=%rs1_3 rd=%rs2_3
+@cl_d ... ... ... .. ... .. &i imm=%uimm_cl_d rs1=%rs1_3 rd=%rs2_3
+@cl_w ... ... ... .. ... .. &i imm=%uimm_cl_w rs1=%rs1_3 rd=%rs2_3
@cl ... ... ... .. ... .. &cl rs1=%rs1_3 rd=%rs2_3
@cs ... ... ... .. ... .. &cs rs1=%rs1_3 rs2=%rs2_3
@cs_2 ... ... ... .. ... .. &r rd=%rs1_3 rs1=%rs1_3 rs2=%rs2_3
-@cs_d ... ... ... .. ... .. &cs_dw uimm=%uimm_cl_d rs1=%rs1_3 rs2=%rs2_3
-@cs_w ... ... ... .. ... .. &cs_dw uimm=%uimm_cl_w rs1=%rs1_3 rs2=%rs2_3
+@cs_d ... ... ... .. ... .. &s imm=%uimm_cl_d rs1=%rs1_3 rs2=%rs2_3
+@cs_w ... ... ... .. ... .. &s imm=%uimm_cl_w rs1=%rs1_3 rs2=%rs2_3
@cb ... ... ... .. ... .. &cb imm=%imm_cb rs1=%rs1_3
@cj ... ........... .. &c_j imm=%imm_cj
@@ -95,11 +95,11 @@
# *** RV64C Standard Extension (Quadrant 0) ***
c_addi4spn 000 ........ ... 00 @ciw
-c_fld 001 ... ... .. ... 00 @cl_d
-c_lw 010 ... ... .. ... 00 @cl_w
+fld 001 ... ... .. ... 00 @cl_d
+lw 010 ... ... .. ... 00 @cl_w
c_flw_ld 011 --- ... -- ... 00 @cl #Note: Must parse uimm manually
-c_fsd 101 ... ... .. ... 00 @cs_d
-c_sw 110 ... ... .. ... 00 @cs_w
+fsd 101 ... ... .. ... 00 @cs_d
+sw 110 ... ... .. ... 00 @cs_w
c_fsw_sd 111 --- ... -- ... 00 @cs #Note: Must parse uimm manually
# *** RV64C Standard Extension (Quadrant 1) ***
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index 520954e9c6..9f13fbcab9 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -37,6 +37,7 @@
&b imm rs2 rs1
&i imm rs1 rd
&r rd rs1 rs2
+&s imm rs2 rs1
&shift shamt rs1 rd
&atomic aq rl rs2 rs1 rd
@@ -44,7 +45,7 @@
@r ....... ..... ..... ... ..... ....... &r %rs2 %rs1 %rd
@i ............ ..... ... ..... ....... &i imm=%imm_i %rs1 %rd
@b ....... ..... ..... ... ..... ....... &b imm=%imm_b %rs2 %rs1
-@s ....... ..... ..... ... ..... ....... imm=%imm_s %rs2 %rs1
+@s ....... ..... ..... ... ..... ....... &s imm=%imm_s %rs2 %rs1
@u .................... ..... ....... imm=%imm_u %rd
@j .................... ..... ....... imm=%imm_j %rd
diff --git a/target/riscv/insn_trans/trans_rvc.inc.c b/target/riscv/insn_trans/trans_rvc.inc.c
index 639c381edf..d932bfd3e0 100644
--- a/target/riscv/insn_trans/trans_rvc.inc.c
+++ b/target/riscv/insn_trans/trans_rvc.inc.c
@@ -28,18 +28,6 @@ static bool trans_c_addi4spn(DisasContext *ctx, arg_c_addi4spn *a)
return trans_addi(ctx, &arg);
}
-static bool trans_c_fld(DisasContext *ctx, arg_c_fld *a)
-{
- arg_fld arg = { .rd = a->rd, .rs1 = a->rs1, .imm = a->uimm };
- return trans_fld(ctx, &arg);
-}
-
-static bool trans_c_lw(DisasContext *ctx, arg_c_lw *a)
-{
- arg_lw arg = { .rd = a->rd, .rs1 = a->rs1, .imm = a->uimm };
- return trans_lw(ctx, &arg);
-}
-
static bool trans_c_flw_ld(DisasContext *ctx, arg_c_flw_ld *a)
{
#ifdef TARGET_RISCV32
@@ -51,18 +39,6 @@ static bool trans_c_flw_ld(DisasContext *ctx, arg_c_flw_ld *a)
#endif
}
-static bool trans_c_fsd(DisasContext *ctx, arg_c_fsd *a)
-{
- arg_fsd arg = { .rs1 = a->rs1, .rs2 = a->rs2, .imm = a->uimm };
- return trans_fsd(ctx, &arg);
-}
-
-static bool trans_c_sw(DisasContext *ctx, arg_c_sw *a)
-{
- arg_sw arg = { .rs1 = a->rs1, .rs2 = a->rs2, .imm = a->uimm };
- return trans_sw(ctx, &arg);
-}
-
static bool trans_c_fsw_sd(DisasContext *ctx, arg_c_fsw_sd *a)
{
#ifdef TARGET_RISCV32
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 33/35] target/riscv: Splice fsw_sd and flw_ld for riscv32 vs riscv64
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
this finally removes the old decoder functions that we carried along
with it.
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
---
target/riscv/Makefile.objs | 1 +
target/riscv/insn16-32.decode | 24 ++++
target/riscv/insn16-64.decode | 4 +
target/riscv/insn16.decode | 7 +-
target/riscv/insn_trans/trans_rvc.inc.c | 22 ----
target/riscv/translate.c | 165 +-----------------------
6 files changed, 31 insertions(+), 192 deletions(-)
create mode 100644 target/riscv/insn16-32.decode
diff --git a/target/riscv/Makefile.objs b/target/riscv/Makefile.objs
index 515c12b4f8..3086ef0c29 100644
--- a/target/riscv/Makefile.objs
+++ b/target/riscv/Makefile.objs
@@ -6,6 +6,7 @@ decode32-y = $(SRC_PATH)/target/riscv/insn32.decode
decode32-$(TARGET_RISCV64) += $(SRC_PATH)/target/riscv/insn64.decode
decode16-y = $(SRC_PATH)/target/riscv/insn16.decode
+decode16-$(TARGET_RISCV32) += $(SRC_PATH)/target/riscv/insn16-32.decode
decode16-$(TARGET_RISCV64) += $(SRC_PATH)/target/riscv/insn16-64.decode
target/riscv/decode_insn32.inc.c: $(decode32-y) $(DECODETREE)
diff --git a/target/riscv/insn16-32.decode b/target/riscv/insn16-32.decode
new file mode 100644
index 0000000000..e21a701056
--- /dev/null
+++ b/target/riscv/insn16-32.decode
@@ -0,0 +1,24 @@
+#
+# RISC-V translation routines for the RVC Instruction Set.
+#
+# Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
+# Bastian Koppelmann, kbastian@mail.uni-paderborn.de
+#
+# This program is free software; you can redistribute it and/or modify it
+# under the terms and conditions of the GNU General Public License,
+# version 2 or later, as published by the Free Software Foundation.
+#
+# This program is distributed in the hope it will be useful, but WITHOUT
+# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+# FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+# more details.
+#
+# You should have received a copy of the GNU General Public License along with
+# this program. If not, see <http://www.gnu.org/licenses/>.
+
+# This is concatenated with insn16.decode for risc32 targets.
+# All of the fields and formats are there.
+
+# *** RV32C Standard Extension (Quadrant 0) ***
+flw 011 ... ... .. ... 00 @cl_w
+fsw 111 ... ... .. ... 00 @cs_w
diff --git a/target/riscv/insn16-64.decode b/target/riscv/insn16-64.decode
index 5af2e2b072..de97a45acf 100644
--- a/target/riscv/insn16-64.decode
+++ b/target/riscv/insn16-64.decode
@@ -19,6 +19,10 @@
# This is concatenated with insn16.decode for risc64 targets.
# All of the fields and formats are there.
+# *** RV64C Standard Extension (Quadrant 0) ***
+ld 011 ... ... .. ... 00 @cl_d
+sd 111 ... ... .. ... 00 @cs_d
+
# *** RV64C Standard Extension (Quadrant 1) ***
subw 100 1 11 ... 00 ... 01 @cs_2
addw 100 1 11 ... 01 ... 01 @cs_2
diff --git a/target/riscv/insn16.decode b/target/riscv/insn16.decode
index c215867ff9..b075336062 100644
--- a/target/riscv/insn16.decode
+++ b/target/riscv/insn16.decode
@@ -47,10 +47,9 @@
&s imm rs1 rs2 !extern
# Argument sets:
-&cl rs1 rd
&ci imm rd
&ciw nzuimm rd
-&cs rs1 rs2
+&cs_dw uimm rs1 rs2
&cb imm rs1
&cr rd rs2
&c_j imm
@@ -69,8 +68,6 @@
@ciw ... ........ ... .. &ciw nzuimm=%nzuimm_ciw rd=%rs2_3
@cl_d ... ... ... .. ... .. &i imm=%uimm_cl_d rs1=%rs1_3 rd=%rs2_3
@cl_w ... ... ... .. ... .. &i imm=%uimm_cl_w rs1=%rs1_3 rd=%rs2_3
-@cl ... ... ... .. ... .. &cl rs1=%rs1_3 rd=%rs2_3
-@cs ... ... ... .. ... .. &cs rs1=%rs1_3 rs2=%rs2_3
@cs_2 ... ... ... .. ... .. &r rd=%rs1_3 rs1=%rs1_3 rs2=%rs2_3
@cs_d ... ... ... .. ... .. &s imm=%uimm_cl_d rs1=%rs1_3 rs2=%rs2_3
@cs_w ... ... ... .. ... .. &s imm=%uimm_cl_w rs1=%rs1_3 rs2=%rs2_3
@@ -97,10 +94,8 @@
c_addi4spn 000 ........ ... 00 @ciw
fld 001 ... ... .. ... 00 @cl_d
lw 010 ... ... .. ... 00 @cl_w
-c_flw_ld 011 --- ... -- ... 00 @cl #Note: Must parse uimm manually
fsd 101 ... ... .. ... 00 @cs_d
sw 110 ... ... .. ... 00 @cs_w
-c_fsw_sd 111 --- ... -- ... 00 @cs #Note: Must parse uimm manually
# *** RV64C Standard Extension (Quadrant 1) ***
c_addi 000 . ..... ..... 01 @ci
diff --git a/target/riscv/insn_trans/trans_rvc.inc.c b/target/riscv/insn_trans/trans_rvc.inc.c
index d932bfd3e0..f521daf32e 100644
--- a/target/riscv/insn_trans/trans_rvc.inc.c
+++ b/target/riscv/insn_trans/trans_rvc.inc.c
@@ -28,28 +28,6 @@ static bool trans_c_addi4spn(DisasContext *ctx, arg_c_addi4spn *a)
return trans_addi(ctx, &arg);
}
-static bool trans_c_flw_ld(DisasContext *ctx, arg_c_flw_ld *a)
-{
-#ifdef TARGET_RISCV32
- /* C.FLW ( RV32FC-only ) */
- return false;
-#else
- /* C.LD ( RV64C/RV128C-only ) */
- return false;
-#endif
-}
-
-static bool trans_c_fsw_sd(DisasContext *ctx, arg_c_fsw_sd *a)
-{
-#ifdef TARGET_RISCV32
- /* C.FSW ( RV32FC-only ) */
- return false;
-#else
- /* C.SD ( RV64C/RV128C-only ) */
- return false;
-#endif
-}
-
static bool trans_c_addi(DisasContext *ctx, arg_c_addi *a)
{
if (a->imm == 0) {
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 56b650a9d7..1d4f848e9a 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -55,22 +55,6 @@ typedef struct DisasContext {
CPURISCVState *env;
} DisasContext;
-#ifdef TARGET_RISCV64
-/* convert riscv funct3 to qemu memop for load/store */
-static const int tcg_memop_lookup[8] = {
- [0 ... 7] = -1,
- [0] = MO_SB,
- [1] = MO_TESW,
- [2] = MO_TESL,
- [4] = MO_UB,
- [5] = MO_TEUW,
-#ifdef TARGET_RISCV64
- [3] = MO_TEQ,
- [6] = MO_TEUL,
-#endif
-};
-#endif
-
#ifdef TARGET_RISCV64
#define CASE_OP_32_64(X) case X: case glue(X, W)
#else
@@ -309,109 +293,6 @@ static void gen_jal(CPURISCVState *env, DisasContext *ctx, int rd,
ctx->base.is_jmp = DISAS_NORETURN;
}
-#ifdef TARGET_RISCV64
-static void gen_load_c(DisasContext *ctx, uint32_t opc, int rd, int rs1,
- target_long imm)
-{
- TCGv t0 = tcg_temp_new();
- TCGv t1 = tcg_temp_new();
- gen_get_gpr(t0, rs1);
- tcg_gen_addi_tl(t0, t0, imm);
- int memop = tcg_memop_lookup[(opc >> 12) & 0x7];
-
- if (memop < 0) {
- gen_exception_illegal(ctx);
- return;
- }
-
- tcg_gen_qemu_ld_tl(t1, t0, ctx->mem_idx, memop);
- gen_set_gpr(rd, t1);
- tcg_temp_free(t0);
- tcg_temp_free(t1);
-}
-
-static void gen_store_c(DisasContext *ctx, uint32_t opc, int rs1, int rs2,
- target_long imm)
-{
- TCGv t0 = tcg_temp_new();
- TCGv dat = tcg_temp_new();
- gen_get_gpr(t0, rs1);
- tcg_gen_addi_tl(t0, t0, imm);
- gen_get_gpr(dat, rs2);
- int memop = tcg_memop_lookup[(opc >> 12) & 0x7];
-
- if (memop < 0) {
- gen_exception_illegal(ctx);
- return;
- }
-
- tcg_gen_qemu_st_tl(dat, t0, ctx->mem_idx, memop);
- tcg_temp_free(t0);
- tcg_temp_free(dat);
-}
-#endif
-
-#ifdef TARGET_RISCV32
-static void gen_fp_load(DisasContext *ctx, uint32_t opc, int rd,
- int rs1, target_long imm)
-{
- TCGv t0;
-
- if (!(ctx->flags & TB_FLAGS_FP_ENABLE)) {
- gen_exception_illegal(ctx);
- return;
- }
-
- t0 = tcg_temp_new();
- gen_get_gpr(t0, rs1);
- tcg_gen_addi_tl(t0, t0, imm);
-
- switch (opc) {
- case OPC_RISC_FLW:
- tcg_gen_qemu_ld_i64(cpu_fpr[rd], t0, ctx->mem_idx, MO_TEUL);
- /* RISC-V requires NaN-boxing of narrower width floating point values */
- tcg_gen_ori_i64(cpu_fpr[rd], cpu_fpr[rd], 0xffffffff00000000ULL);
- break;
- case OPC_RISC_FLD:
- tcg_gen_qemu_ld_i64(cpu_fpr[rd], t0, ctx->mem_idx, MO_TEQ);
- break;
- default:
- gen_exception_illegal(ctx);
- break;
- }
- tcg_temp_free(t0);
-}
-
-static void gen_fp_store(DisasContext *ctx, uint32_t opc, int rs1,
- int rs2, target_long imm)
-{
- TCGv t0;
-
- if (!(ctx->flags & TB_FLAGS_FP_ENABLE)) {
- gen_exception_illegal(ctx);
- return;
- }
-
- t0 = tcg_temp_new();
- gen_get_gpr(t0, rs1);
- tcg_gen_addi_tl(t0, t0, imm);
-
- switch (opc) {
- case OPC_RISC_FSW:
- tcg_gen_qemu_st_i64(cpu_fpr[rs2], t0, ctx->mem_idx, MO_TEUL);
- break;
- case OPC_RISC_FSD:
- tcg_gen_qemu_st_i64(cpu_fpr[rs2], t0, ctx->mem_idx, MO_TEQ);
- break;
- default:
- gen_exception_illegal(ctx);
- break;
- }
-
- tcg_temp_free(t0);
-}
-#endif
-
static void gen_set_rm(DisasContext *ctx, int rm)
{
TCGv_i32 t0;
@@ -425,49 +306,6 @@ static void gen_set_rm(DisasContext *ctx, int rm)
tcg_temp_free_i32(t0);
}
-static void decode_RV32_64C0(DisasContext *ctx)
-{
- uint8_t funct3 = extract32(ctx->opcode, 13, 3);
- uint8_t rd_rs2 = GET_C_RS2S(ctx->opcode);
- uint8_t rs1s = GET_C_RS1S(ctx->opcode);
-
- switch (funct3) {
- case 3:
-#if defined(TARGET_RISCV64)
- /* C.LD(RV64/128) -> ld rd', offset[7:3](rs1')*/
- gen_load_c(ctx, OPC_RISC_LD, rd_rs2, rs1s,
- GET_C_LD_IMM(ctx->opcode));
-#else
- /* C.FLW (RV32) -> flw rd', offset[6:2](rs1')*/
- gen_fp_load(ctx, OPC_RISC_FLW, rd_rs2, rs1s,
- GET_C_LW_IMM(ctx->opcode));
-#endif
- break;
- case 7:
-#if defined(TARGET_RISCV64)
- /* C.SD (RV64/128) -> sd rs2', offset[7:3](rs1')*/
- gen_store_c(ctx, OPC_RISC_SD, rs1s, rd_rs2,
- GET_C_LD_IMM(ctx->opcode));
-#else
- /* C.FSW (RV32) -> fsw rs2', offset[6:2](rs1')*/
- gen_fp_store(ctx, OPC_RISC_FSW, rs1s, rd_rs2,
- GET_C_LW_IMM(ctx->opcode));
-#endif
- break;
- }
-}
-
-static void decode_RV32_64C(CPURISCVState *env, DisasContext *ctx)
-{
- uint8_t op = extract32(ctx->opcode, 0, 2);
-
- switch (op) {
- case 0:
- decode_RV32_64C0(ctx);
- break;
- }
-}
-
#define EX_SH(amount) \
static int ex_shift_##amount(int imm) \
{ \
@@ -611,8 +449,7 @@ static void decode_opc(DisasContext *ctx)
} else {
ctx->pc_succ_insn = ctx->base.pc_next + 2;
if (!decode_insn16(ctx, ctx->opcode)) {
- /* fall back to old decoder */
- decode_RV32_64C(ctx->env, ctx);
+ gen_exception_illegal(ctx);
}
}
} else {
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 33/35] target/riscv: Splice fsw_sd and flw_ld for riscv32 vs riscv64
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
this finally removes the old decoder functions that we carried along
with it.
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
---
target/riscv/Makefile.objs | 1 +
target/riscv/insn16-32.decode | 24 ++++
target/riscv/insn16-64.decode | 4 +
target/riscv/insn16.decode | 7 +-
target/riscv/insn_trans/trans_rvc.inc.c | 22 ----
target/riscv/translate.c | 165 +-----------------------
6 files changed, 31 insertions(+), 192 deletions(-)
create mode 100644 target/riscv/insn16-32.decode
diff --git a/target/riscv/Makefile.objs b/target/riscv/Makefile.objs
index 515c12b4f8..3086ef0c29 100644
--- a/target/riscv/Makefile.objs
+++ b/target/riscv/Makefile.objs
@@ -6,6 +6,7 @@ decode32-y = $(SRC_PATH)/target/riscv/insn32.decode
decode32-$(TARGET_RISCV64) += $(SRC_PATH)/target/riscv/insn64.decode
decode16-y = $(SRC_PATH)/target/riscv/insn16.decode
+decode16-$(TARGET_RISCV32) += $(SRC_PATH)/target/riscv/insn16-32.decode
decode16-$(TARGET_RISCV64) += $(SRC_PATH)/target/riscv/insn16-64.decode
target/riscv/decode_insn32.inc.c: $(decode32-y) $(DECODETREE)
diff --git a/target/riscv/insn16-32.decode b/target/riscv/insn16-32.decode
new file mode 100644
index 0000000000..e21a701056
--- /dev/null
+++ b/target/riscv/insn16-32.decode
@@ -0,0 +1,24 @@
+#
+# RISC-V translation routines for the RVC Instruction Set.
+#
+# Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
+# Bastian Koppelmann, kbastian@mail.uni-paderborn.de
+#
+# This program is free software; you can redistribute it and/or modify it
+# under the terms and conditions of the GNU General Public License,
+# version 2 or later, as published by the Free Software Foundation.
+#
+# This program is distributed in the hope it will be useful, but WITHOUT
+# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+# FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+# more details.
+#
+# You should have received a copy of the GNU General Public License along with
+# this program. If not, see <http://www.gnu.org/licenses/>.
+
+# This is concatenated with insn16.decode for risc32 targets.
+# All of the fields and formats are there.
+
+# *** RV32C Standard Extension (Quadrant 0) ***
+flw 011 ... ... .. ... 00 @cl_w
+fsw 111 ... ... .. ... 00 @cs_w
diff --git a/target/riscv/insn16-64.decode b/target/riscv/insn16-64.decode
index 5af2e2b072..de97a45acf 100644
--- a/target/riscv/insn16-64.decode
+++ b/target/riscv/insn16-64.decode
@@ -19,6 +19,10 @@
# This is concatenated with insn16.decode for risc64 targets.
# All of the fields and formats are there.
+# *** RV64C Standard Extension (Quadrant 0) ***
+ld 011 ... ... .. ... 00 @cl_d
+sd 111 ... ... .. ... 00 @cs_d
+
# *** RV64C Standard Extension (Quadrant 1) ***
subw 100 1 11 ... 00 ... 01 @cs_2
addw 100 1 11 ... 01 ... 01 @cs_2
diff --git a/target/riscv/insn16.decode b/target/riscv/insn16.decode
index c215867ff9..b075336062 100644
--- a/target/riscv/insn16.decode
+++ b/target/riscv/insn16.decode
@@ -47,10 +47,9 @@
&s imm rs1 rs2 !extern
# Argument sets:
-&cl rs1 rd
&ci imm rd
&ciw nzuimm rd
-&cs rs1 rs2
+&cs_dw uimm rs1 rs2
&cb imm rs1
&cr rd rs2
&c_j imm
@@ -69,8 +68,6 @@
@ciw ... ........ ... .. &ciw nzuimm=%nzuimm_ciw rd=%rs2_3
@cl_d ... ... ... .. ... .. &i imm=%uimm_cl_d rs1=%rs1_3 rd=%rs2_3
@cl_w ... ... ... .. ... .. &i imm=%uimm_cl_w rs1=%rs1_3 rd=%rs2_3
-@cl ... ... ... .. ... .. &cl rs1=%rs1_3 rd=%rs2_3
-@cs ... ... ... .. ... .. &cs rs1=%rs1_3 rs2=%rs2_3
@cs_2 ... ... ... .. ... .. &r rd=%rs1_3 rs1=%rs1_3 rs2=%rs2_3
@cs_d ... ... ... .. ... .. &s imm=%uimm_cl_d rs1=%rs1_3 rs2=%rs2_3
@cs_w ... ... ... .. ... .. &s imm=%uimm_cl_w rs1=%rs1_3 rs2=%rs2_3
@@ -97,10 +94,8 @@
c_addi4spn 000 ........ ... 00 @ciw
fld 001 ... ... .. ... 00 @cl_d
lw 010 ... ... .. ... 00 @cl_w
-c_flw_ld 011 --- ... -- ... 00 @cl #Note: Must parse uimm manually
fsd 101 ... ... .. ... 00 @cs_d
sw 110 ... ... .. ... 00 @cs_w
-c_fsw_sd 111 --- ... -- ... 00 @cs #Note: Must parse uimm manually
# *** RV64C Standard Extension (Quadrant 1) ***
c_addi 000 . ..... ..... 01 @ci
diff --git a/target/riscv/insn_trans/trans_rvc.inc.c b/target/riscv/insn_trans/trans_rvc.inc.c
index d932bfd3e0..f521daf32e 100644
--- a/target/riscv/insn_trans/trans_rvc.inc.c
+++ b/target/riscv/insn_trans/trans_rvc.inc.c
@@ -28,28 +28,6 @@ static bool trans_c_addi4spn(DisasContext *ctx, arg_c_addi4spn *a)
return trans_addi(ctx, &arg);
}
-static bool trans_c_flw_ld(DisasContext *ctx, arg_c_flw_ld *a)
-{
-#ifdef TARGET_RISCV32
- /* C.FLW ( RV32FC-only ) */
- return false;
-#else
- /* C.LD ( RV64C/RV128C-only ) */
- return false;
-#endif
-}
-
-static bool trans_c_fsw_sd(DisasContext *ctx, arg_c_fsw_sd *a)
-{
-#ifdef TARGET_RISCV32
- /* C.FSW ( RV32FC-only ) */
- return false;
-#else
- /* C.SD ( RV64C/RV128C-only ) */
- return false;
-#endif
-}
-
static bool trans_c_addi(DisasContext *ctx, arg_c_addi *a)
{
if (a->imm == 0) {
diff --git a/target/riscv/translate.c b/target/riscv/translate.c
index 56b650a9d7..1d4f848e9a 100644
--- a/target/riscv/translate.c
+++ b/target/riscv/translate.c
@@ -55,22 +55,6 @@ typedef struct DisasContext {
CPURISCVState *env;
} DisasContext;
-#ifdef TARGET_RISCV64
-/* convert riscv funct3 to qemu memop for load/store */
-static const int tcg_memop_lookup[8] = {
- [0 ... 7] = -1,
- [0] = MO_SB,
- [1] = MO_TESW,
- [2] = MO_TESL,
- [4] = MO_UB,
- [5] = MO_TEUW,
-#ifdef TARGET_RISCV64
- [3] = MO_TEQ,
- [6] = MO_TEUL,
-#endif
-};
-#endif
-
#ifdef TARGET_RISCV64
#define CASE_OP_32_64(X) case X: case glue(X, W)
#else
@@ -309,109 +293,6 @@ static void gen_jal(CPURISCVState *env, DisasContext *ctx, int rd,
ctx->base.is_jmp = DISAS_NORETURN;
}
-#ifdef TARGET_RISCV64
-static void gen_load_c(DisasContext *ctx, uint32_t opc, int rd, int rs1,
- target_long imm)
-{
- TCGv t0 = tcg_temp_new();
- TCGv t1 = tcg_temp_new();
- gen_get_gpr(t0, rs1);
- tcg_gen_addi_tl(t0, t0, imm);
- int memop = tcg_memop_lookup[(opc >> 12) & 0x7];
-
- if (memop < 0) {
- gen_exception_illegal(ctx);
- return;
- }
-
- tcg_gen_qemu_ld_tl(t1, t0, ctx->mem_idx, memop);
- gen_set_gpr(rd, t1);
- tcg_temp_free(t0);
- tcg_temp_free(t1);
-}
-
-static void gen_store_c(DisasContext *ctx, uint32_t opc, int rs1, int rs2,
- target_long imm)
-{
- TCGv t0 = tcg_temp_new();
- TCGv dat = tcg_temp_new();
- gen_get_gpr(t0, rs1);
- tcg_gen_addi_tl(t0, t0, imm);
- gen_get_gpr(dat, rs2);
- int memop = tcg_memop_lookup[(opc >> 12) & 0x7];
-
- if (memop < 0) {
- gen_exception_illegal(ctx);
- return;
- }
-
- tcg_gen_qemu_st_tl(dat, t0, ctx->mem_idx, memop);
- tcg_temp_free(t0);
- tcg_temp_free(dat);
-}
-#endif
-
-#ifdef TARGET_RISCV32
-static void gen_fp_load(DisasContext *ctx, uint32_t opc, int rd,
- int rs1, target_long imm)
-{
- TCGv t0;
-
- if (!(ctx->flags & TB_FLAGS_FP_ENABLE)) {
- gen_exception_illegal(ctx);
- return;
- }
-
- t0 = tcg_temp_new();
- gen_get_gpr(t0, rs1);
- tcg_gen_addi_tl(t0, t0, imm);
-
- switch (opc) {
- case OPC_RISC_FLW:
- tcg_gen_qemu_ld_i64(cpu_fpr[rd], t0, ctx->mem_idx, MO_TEUL);
- /* RISC-V requires NaN-boxing of narrower width floating point values */
- tcg_gen_ori_i64(cpu_fpr[rd], cpu_fpr[rd], 0xffffffff00000000ULL);
- break;
- case OPC_RISC_FLD:
- tcg_gen_qemu_ld_i64(cpu_fpr[rd], t0, ctx->mem_idx, MO_TEQ);
- break;
- default:
- gen_exception_illegal(ctx);
- break;
- }
- tcg_temp_free(t0);
-}
-
-static void gen_fp_store(DisasContext *ctx, uint32_t opc, int rs1,
- int rs2, target_long imm)
-{
- TCGv t0;
-
- if (!(ctx->flags & TB_FLAGS_FP_ENABLE)) {
- gen_exception_illegal(ctx);
- return;
- }
-
- t0 = tcg_temp_new();
- gen_get_gpr(t0, rs1);
- tcg_gen_addi_tl(t0, t0, imm);
-
- switch (opc) {
- case OPC_RISC_FSW:
- tcg_gen_qemu_st_i64(cpu_fpr[rs2], t0, ctx->mem_idx, MO_TEUL);
- break;
- case OPC_RISC_FSD:
- tcg_gen_qemu_st_i64(cpu_fpr[rs2], t0, ctx->mem_idx, MO_TEQ);
- break;
- default:
- gen_exception_illegal(ctx);
- break;
- }
-
- tcg_temp_free(t0);
-}
-#endif
-
static void gen_set_rm(DisasContext *ctx, int rm)
{
TCGv_i32 t0;
@@ -425,49 +306,6 @@ static void gen_set_rm(DisasContext *ctx, int rm)
tcg_temp_free_i32(t0);
}
-static void decode_RV32_64C0(DisasContext *ctx)
-{
- uint8_t funct3 = extract32(ctx->opcode, 13, 3);
- uint8_t rd_rs2 = GET_C_RS2S(ctx->opcode);
- uint8_t rs1s = GET_C_RS1S(ctx->opcode);
-
- switch (funct3) {
- case 3:
-#if defined(TARGET_RISCV64)
- /* C.LD(RV64/128) -> ld rd', offset[7:3](rs1')*/
- gen_load_c(ctx, OPC_RISC_LD, rd_rs2, rs1s,
- GET_C_LD_IMM(ctx->opcode));
-#else
- /* C.FLW (RV32) -> flw rd', offset[6:2](rs1')*/
- gen_fp_load(ctx, OPC_RISC_FLW, rd_rs2, rs1s,
- GET_C_LW_IMM(ctx->opcode));
-#endif
- break;
- case 7:
-#if defined(TARGET_RISCV64)
- /* C.SD (RV64/128) -> sd rs2', offset[7:3](rs1')*/
- gen_store_c(ctx, OPC_RISC_SD, rs1s, rd_rs2,
- GET_C_LD_IMM(ctx->opcode));
-#else
- /* C.FSW (RV32) -> fsw rs2', offset[6:2](rs1')*/
- gen_fp_store(ctx, OPC_RISC_FSW, rs1s, rd_rs2,
- GET_C_LW_IMM(ctx->opcode));
-#endif
- break;
- }
-}
-
-static void decode_RV32_64C(CPURISCVState *env, DisasContext *ctx)
-{
- uint8_t op = extract32(ctx->opcode, 0, 2);
-
- switch (op) {
- case 0:
- decode_RV32_64C0(ctx);
- break;
- }
-}
-
#define EX_SH(amount) \
static int ex_shift_##amount(int imm) \
{ \
@@ -611,8 +449,7 @@ static void decode_opc(DisasContext *ctx)
} else {
ctx->pc_succ_insn = ctx->base.pc_next + 2;
if (!decode_insn16(ctx, ctx->opcode)) {
- /* fall back to old decoder */
- decode_RV32_64C(ctx->env, ctx);
+ gen_exception_illegal(ctx);
}
}
} else {
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 34/35] target/riscv: Splice remaining compressed insn pairs for riscv32 vs riscv64
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
it splices flwsp_ldsp, fswsp_sdsp, and jal_addiw and makes each of them
reuse the code generator used for the non compressed insns.
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
---
target/riscv/insn16-32.decode | 7 +++++
target/riscv/insn16-64.decode | 5 ++++
target/riscv/insn16.decode | 12 ++------
target/riscv/insn32.decode | 3 +-
target/riscv/insn_trans/trans_rvc.inc.c | 40 -------------------------
5 files changed, 16 insertions(+), 51 deletions(-)
diff --git a/target/riscv/insn16-32.decode b/target/riscv/insn16-32.decode
index e21a701056..978b8d5834 100644
--- a/target/riscv/insn16-32.decode
+++ b/target/riscv/insn16-32.decode
@@ -22,3 +22,10 @@
# *** RV32C Standard Extension (Quadrant 0) ***
flw 011 ... ... .. ... 00 @cl_w
fsw 111 ... ... .. ... 00 @cs_w
+
+# *** RV32C Standard Extension (Quadrant 1) ***
+jal 001 ...... ..... 01 &j imm=%imm_cj rd=1
+
+# *** RV32C Standard Extension (Quadrant 2) ***
+flw 011 . ..... ..... 10 &i imm=%uimm_6bit_lw %rd rs1=2
+fsw 111 ...... ..... 10 &s imm=%uimm_6bit_sw rs2=2 rs1=%rs2_5
diff --git a/target/riscv/insn16-64.decode b/target/riscv/insn16-64.decode
index de97a45acf..d43055837a 100644
--- a/target/riscv/insn16-64.decode
+++ b/target/riscv/insn16-64.decode
@@ -24,5 +24,10 @@ ld 011 ... ... .. ... 00 @cl_d
sd 111 ... ... .. ... 00 @cs_d
# *** RV64C Standard Extension (Quadrant 1) ***
+addiw 001 . ..... ..... 01 @ci
subw 100 1 11 ... 00 ... 01 @cs_2
addw 100 1 11 ... 01 ... 01 @cs_2
+
+# *** RV64C Standard Extension (Quadrant 2) ***
+ld 011 . ..... ..... 10 &i imm=%uimm_6bit_ld %rd rs1=2
+sd 111 ...... ..... 10 &s imm=%uimm_6bit_sd rs2=%rs2_5 rs1=2
diff --git a/target/riscv/insn16.decode b/target/riscv/insn16.decode
index b075336062..98dd672c7f 100644
--- a/target/riscv/insn16.decode
+++ b/target/riscv/insn16.decode
@@ -45,6 +45,7 @@
&r rd rs1 rs2 !extern
&i imm rs1 rd !extern
&s imm rs1 rs2 !extern
+&j imm rd !extern
# Argument sets:
&ci imm rd
@@ -59,12 +60,10 @@
&c_sd uimm rs2
&c_addi16sp_lui imm_lui imm_addi16sp rd
-&c_flwsp_ldsp uimm_flwsp uimm_ldsp rd
-&c_fswsp_sdsp uimm_fswsp uimm_sdsp rs2
# Formats 16:
@cr .... ..... ..... .. &cr rs2=%rs2_5 %rd
-@ci ... . ..... ..... .. &ci imm=%imm_ci %rd
+@ci ... . ..... ..... .. &i imm=%imm_ci %rd rs1=%rd
@ciw ... ........ ... .. &ciw nzuimm=%nzuimm_ciw rd=%rs2_3
@cl_d ... ... ... .. ... .. &i imm=%uimm_cl_d rs1=%rs1_3 rd=%rs2_3
@cl_w ... ... ... .. ... .. &i imm=%uimm_cl_w rs1=%rs1_3 rd=%rs2_3
@@ -80,10 +79,6 @@
@c_sw ... . ..... ..... .. &c_sd uimm=%uimm_6bit_sw rs2=%rs2_5
@c_addi16sp_lui ... . ..... ..... .. &c_addi16sp_lui %imm_lui %imm_addi16sp %rd
-@c_flwsp_ldsp ... . ..... ..... .. &c_flwsp_ldsp uimm_flwsp=%uimm_6bit_lw \
- uimm_ldsp=%uimm_6bit_ld %rd
-@c_fswsp_sdsp ... . ..... ..... .. &c_fswsp_sdsp uimm_fswsp=%uimm_6bit_sw \
- uimm_sdsp=%uimm_6bit_sd rs2=%rs2_5
@c_shift ... . .. ... ..... .. &c_shift rd=%rs1_3 shamt=%nzuimm_6bit
@c_shift2 ... . .. ... ..... .. &c_shift rd=%rd shamt=%nzuimm_6bit
@@ -99,7 +94,6 @@ sw 110 ... ... .. ... 00 @cs_w
# *** RV64C Standard Extension (Quadrant 1) ***
c_addi 000 . ..... ..... 01 @ci
-c_jal_addiw 001 . ..... ..... 01 @ci #Note: parse rd and/or imm manually
c_li 010 . ..... ..... 01 @ci
c_addi16sp_lui 011 . ..... ..... 01 @c_addi16sp_lui # shares opc with C.LUI
c_srli 100 . 00 ... ..... 01 @c_shift
@@ -117,9 +111,7 @@ c_bnez 111 ... ... ..... 01 @cb
c_slli 000 . ..... ..... 10 @c_shift2
c_fldsp 001 . ..... ..... 10 @c_ld
c_lwsp 010 . ..... ..... 10 @c_lw
-c_flwsp_ldsp 011 . ..... ..... 10 @c_flwsp_ldsp #C.LDSP:RV64;C.FLWSP:RV32
c_jr_mv 100 0 ..... ..... 10 @cr
c_ebreak_jalr_add 100 1 ..... ..... 10 @cr
c_fsdsp 101 ...... ..... 10 @c_sd
c_swsp 110 . ..... ..... 10 @c_sw
-c_fswsp_sdsp 111 . ..... ..... 10 @c_fswsp_sdsp #C.SDSP:RV64;C.FSWSP:RV32
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index 9f13fbcab9..1cb7e2cde0 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -38,6 +38,7 @@
&i imm rs1 rd
&r rd rs1 rs2
&s imm rs2 rs1
+&j imm rd
&shift shamt rs1 rd
&atomic aq rl rs2 rs1 rd
@@ -47,7 +48,7 @@
@b ....... ..... ..... ... ..... ....... &b imm=%imm_b %rs2 %rs1
@s ....... ..... ..... ... ..... ....... &s imm=%imm_s %rs2 %rs1
@u .................... ..... ....... imm=%imm_u %rd
-@j .................... ..... ....... imm=%imm_j %rd
+@j .................... ..... ....... &j imm=%imm_j %rd
@sh6 ...... ...... ..... ... ..... ....... &shift shamt=%sh6 %rs1 %rd
@csr ............ ..... ... ..... ....... %csr %rs1 %rd
diff --git a/target/riscv/insn_trans/trans_rvc.inc.c b/target/riscv/insn_trans/trans_rvc.inc.c
index f521daf32e..db9119ec9b 100644
--- a/target/riscv/insn_trans/trans_rvc.inc.c
+++ b/target/riscv/insn_trans/trans_rvc.inc.c
@@ -38,19 +38,6 @@ static bool trans_c_addi(DisasContext *ctx, arg_c_addi *a)
return trans_addi(ctx, &arg);
}
-static bool trans_c_jal_addiw(DisasContext *ctx, arg_c_jal_addiw *a)
-{
-#ifdef TARGET_RISCV32
- /* C.JAL */
- arg_jal arg = { .rd = 1, .imm = a->imm };
- return trans_jal(ctx, &arg);
-#else
- /* C.ADDIW */
- arg_addiw arg = { .rd = a->rd, .rs1 = a->rd, .imm = a->imm };
- return trans_addiw(ctx, &arg);
-#endif
-}
-
static bool trans_c_li(DisasContext *ctx, arg_c_li *a)
{
if (a->rd == 0) {
@@ -163,20 +150,6 @@ static bool trans_c_lwsp(DisasContext *ctx, arg_c_lwsp *a)
return trans_lw(ctx, &arg);
}
-static bool trans_c_flwsp_ldsp(DisasContext *ctx, arg_c_flwsp_ldsp *a)
-{
-#ifdef TARGET_RISCV32
- /* C.FLWSP */
- arg_flw arg_flw = { .rd = a->rd, .rs1 = 2, .imm = a->uimm_flwsp };
- return trans_flw(ctx, &arg_flw);
-#else
- /* C.LDSP */
- arg_ld arg_ld = { .rd = a->rd, .rs1 = 2, .imm = a->uimm_ldsp };
- return trans_ld(ctx, &arg_ld);
-#endif
- return false;
-}
-
static bool trans_c_jr_mv(DisasContext *ctx, arg_c_jr_mv *a)
{
if (a->rd != 0 && a->rs2 == 0) {
@@ -222,16 +195,3 @@ static bool trans_c_swsp(DisasContext *ctx, arg_c_swsp *a)
arg_sw arg = { .rs1 = 2, .rs2 = a->rs2, .imm = a->uimm };
return trans_sw(ctx, &arg);
}
-
-static bool trans_c_fswsp_sdsp(DisasContext *ctx, arg_c_fswsp_sdsp *a)
-{
-#ifdef TARGET_RISCV32
- /* C.FSWSP */
- arg_fsw a_fsw = { .rs1 = a->rs2, .rs2 = 2, .imm = a->uimm_fswsp };
- return trans_fsw(ctx, &a_fsw);
-#else
- /* C.SDSP */
- arg_sd a_sd = { .rs1 = 2, .rs2 = a->rs2, .imm = a->uimm_sdsp };
- return trans_sd(ctx, &a_sd);
-#endif
-}
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 34/35] target/riscv: Splice remaining compressed insn pairs for riscv32 vs riscv64
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
it splices flwsp_ldsp, fswsp_sdsp, and jal_addiw and makes each of them
reuse the code generator used for the non compressed insns.
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
---
target/riscv/insn16-32.decode | 7 +++++
target/riscv/insn16-64.decode | 5 ++++
target/riscv/insn16.decode | 12 ++------
target/riscv/insn32.decode | 3 +-
target/riscv/insn_trans/trans_rvc.inc.c | 40 -------------------------
5 files changed, 16 insertions(+), 51 deletions(-)
diff --git a/target/riscv/insn16-32.decode b/target/riscv/insn16-32.decode
index e21a701056..978b8d5834 100644
--- a/target/riscv/insn16-32.decode
+++ b/target/riscv/insn16-32.decode
@@ -22,3 +22,10 @@
# *** RV32C Standard Extension (Quadrant 0) ***
flw 011 ... ... .. ... 00 @cl_w
fsw 111 ... ... .. ... 00 @cs_w
+
+# *** RV32C Standard Extension (Quadrant 1) ***
+jal 001 ...... ..... 01 &j imm=%imm_cj rd=1
+
+# *** RV32C Standard Extension (Quadrant 2) ***
+flw 011 . ..... ..... 10 &i imm=%uimm_6bit_lw %rd rs1=2
+fsw 111 ...... ..... 10 &s imm=%uimm_6bit_sw rs2=2 rs1=%rs2_5
diff --git a/target/riscv/insn16-64.decode b/target/riscv/insn16-64.decode
index de97a45acf..d43055837a 100644
--- a/target/riscv/insn16-64.decode
+++ b/target/riscv/insn16-64.decode
@@ -24,5 +24,10 @@ ld 011 ... ... .. ... 00 @cl_d
sd 111 ... ... .. ... 00 @cs_d
# *** RV64C Standard Extension (Quadrant 1) ***
+addiw 001 . ..... ..... 01 @ci
subw 100 1 11 ... 00 ... 01 @cs_2
addw 100 1 11 ... 01 ... 01 @cs_2
+
+# *** RV64C Standard Extension (Quadrant 2) ***
+ld 011 . ..... ..... 10 &i imm=%uimm_6bit_ld %rd rs1=2
+sd 111 ...... ..... 10 &s imm=%uimm_6bit_sd rs2=%rs2_5 rs1=2
diff --git a/target/riscv/insn16.decode b/target/riscv/insn16.decode
index b075336062..98dd672c7f 100644
--- a/target/riscv/insn16.decode
+++ b/target/riscv/insn16.decode
@@ -45,6 +45,7 @@
&r rd rs1 rs2 !extern
&i imm rs1 rd !extern
&s imm rs1 rs2 !extern
+&j imm rd !extern
# Argument sets:
&ci imm rd
@@ -59,12 +60,10 @@
&c_sd uimm rs2
&c_addi16sp_lui imm_lui imm_addi16sp rd
-&c_flwsp_ldsp uimm_flwsp uimm_ldsp rd
-&c_fswsp_sdsp uimm_fswsp uimm_sdsp rs2
# Formats 16:
@cr .... ..... ..... .. &cr rs2=%rs2_5 %rd
-@ci ... . ..... ..... .. &ci imm=%imm_ci %rd
+@ci ... . ..... ..... .. &i imm=%imm_ci %rd rs1=%rd
@ciw ... ........ ... .. &ciw nzuimm=%nzuimm_ciw rd=%rs2_3
@cl_d ... ... ... .. ... .. &i imm=%uimm_cl_d rs1=%rs1_3 rd=%rs2_3
@cl_w ... ... ... .. ... .. &i imm=%uimm_cl_w rs1=%rs1_3 rd=%rs2_3
@@ -80,10 +79,6 @@
@c_sw ... . ..... ..... .. &c_sd uimm=%uimm_6bit_sw rs2=%rs2_5
@c_addi16sp_lui ... . ..... ..... .. &c_addi16sp_lui %imm_lui %imm_addi16sp %rd
-@c_flwsp_ldsp ... . ..... ..... .. &c_flwsp_ldsp uimm_flwsp=%uimm_6bit_lw \
- uimm_ldsp=%uimm_6bit_ld %rd
-@c_fswsp_sdsp ... . ..... ..... .. &c_fswsp_sdsp uimm_fswsp=%uimm_6bit_sw \
- uimm_sdsp=%uimm_6bit_sd rs2=%rs2_5
@c_shift ... . .. ... ..... .. &c_shift rd=%rs1_3 shamt=%nzuimm_6bit
@c_shift2 ... . .. ... ..... .. &c_shift rd=%rd shamt=%nzuimm_6bit
@@ -99,7 +94,6 @@ sw 110 ... ... .. ... 00 @cs_w
# *** RV64C Standard Extension (Quadrant 1) ***
c_addi 000 . ..... ..... 01 @ci
-c_jal_addiw 001 . ..... ..... 01 @ci #Note: parse rd and/or imm manually
c_li 010 . ..... ..... 01 @ci
c_addi16sp_lui 011 . ..... ..... 01 @c_addi16sp_lui # shares opc with C.LUI
c_srli 100 . 00 ... ..... 01 @c_shift
@@ -117,9 +111,7 @@ c_bnez 111 ... ... ..... 01 @cb
c_slli 000 . ..... ..... 10 @c_shift2
c_fldsp 001 . ..... ..... 10 @c_ld
c_lwsp 010 . ..... ..... 10 @c_lw
-c_flwsp_ldsp 011 . ..... ..... 10 @c_flwsp_ldsp #C.LDSP:RV64;C.FLWSP:RV32
c_jr_mv 100 0 ..... ..... 10 @cr
c_ebreak_jalr_add 100 1 ..... ..... 10 @cr
c_fsdsp 101 ...... ..... 10 @c_sd
c_swsp 110 . ..... ..... 10 @c_sw
-c_fswsp_sdsp 111 . ..... ..... 10 @c_fswsp_sdsp #C.SDSP:RV64;C.FSWSP:RV32
diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
index 9f13fbcab9..1cb7e2cde0 100644
--- a/target/riscv/insn32.decode
+++ b/target/riscv/insn32.decode
@@ -38,6 +38,7 @@
&i imm rs1 rd
&r rd rs1 rs2
&s imm rs2 rs1
+&j imm rd
&shift shamt rs1 rd
&atomic aq rl rs2 rs1 rd
@@ -47,7 +48,7 @@
@b ....... ..... ..... ... ..... ....... &b imm=%imm_b %rs2 %rs1
@s ....... ..... ..... ... ..... ....... &s imm=%imm_s %rs2 %rs1
@u .................... ..... ....... imm=%imm_u %rd
-@j .................... ..... ....... imm=%imm_j %rd
+@j .................... ..... ....... &j imm=%imm_j %rd
@sh6 ...... ...... ..... ... ..... ....... &shift shamt=%sh6 %rs1 %rd
@csr ............ ..... ... ..... ....... %csr %rs1 %rd
diff --git a/target/riscv/insn_trans/trans_rvc.inc.c b/target/riscv/insn_trans/trans_rvc.inc.c
index f521daf32e..db9119ec9b 100644
--- a/target/riscv/insn_trans/trans_rvc.inc.c
+++ b/target/riscv/insn_trans/trans_rvc.inc.c
@@ -38,19 +38,6 @@ static bool trans_c_addi(DisasContext *ctx, arg_c_addi *a)
return trans_addi(ctx, &arg);
}
-static bool trans_c_jal_addiw(DisasContext *ctx, arg_c_jal_addiw *a)
-{
-#ifdef TARGET_RISCV32
- /* C.JAL */
- arg_jal arg = { .rd = 1, .imm = a->imm };
- return trans_jal(ctx, &arg);
-#else
- /* C.ADDIW */
- arg_addiw arg = { .rd = a->rd, .rs1 = a->rd, .imm = a->imm };
- return trans_addiw(ctx, &arg);
-#endif
-}
-
static bool trans_c_li(DisasContext *ctx, arg_c_li *a)
{
if (a->rd == 0) {
@@ -163,20 +150,6 @@ static bool trans_c_lwsp(DisasContext *ctx, arg_c_lwsp *a)
return trans_lw(ctx, &arg);
}
-static bool trans_c_flwsp_ldsp(DisasContext *ctx, arg_c_flwsp_ldsp *a)
-{
-#ifdef TARGET_RISCV32
- /* C.FLWSP */
- arg_flw arg_flw = { .rd = a->rd, .rs1 = 2, .imm = a->uimm_flwsp };
- return trans_flw(ctx, &arg_flw);
-#else
- /* C.LDSP */
- arg_ld arg_ld = { .rd = a->rd, .rs1 = 2, .imm = a->uimm_ldsp };
- return trans_ld(ctx, &arg_ld);
-#endif
- return false;
-}
-
static bool trans_c_jr_mv(DisasContext *ctx, arg_c_jr_mv *a)
{
if (a->rd != 0 && a->rs2 == 0) {
@@ -222,16 +195,3 @@ static bool trans_c_swsp(DisasContext *ctx, arg_c_swsp *a)
arg_sw arg = { .rs1 = 2, .rs2 = a->rs2, .imm = a->uimm };
return trans_sw(ctx, &arg);
}
-
-static bool trans_c_fswsp_sdsp(DisasContext *ctx, arg_c_fswsp_sdsp *a)
-{
-#ifdef TARGET_RISCV32
- /* C.FSWSP */
- arg_fsw a_fsw = { .rs1 = a->rs2, .rs2 = 2, .imm = a->uimm_fswsp };
- return trans_fsw(ctx, &a_fsw);
-#else
- /* C.SDSP */
- arg_sd a_sd = { .rs1 = 2, .rs2 = a->rs2, .imm = a->uimm_sdsp };
- return trans_sd(ctx, &a_sd);
-#endif
-}
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-devel] [PATCH v3 35/35] target/riscv: Remaining rvc insn reuse 32 bit translators
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 13:20 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
only a translate functions of rvc need to handle special cases. For
the other rvc instruction we can remove the extra layer of indirection.
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
---
target/riscv/insn16.decode | 37 +++++++++----------
target/riscv/insn_trans/trans_rvc.inc.c | 48 -------------------------
2 files changed, 17 insertions(+), 68 deletions(-)
diff --git a/target/riscv/insn16.decode b/target/riscv/insn16.decode
index 98dd672c7f..d88a0c78ab 100644
--- a/target/riscv/insn16.decode
+++ b/target/riscv/insn16.decode
@@ -46,19 +46,15 @@
&i imm rs1 rd !extern
&s imm rs1 rs2 !extern
&j imm rd !extern
+&b imm rs2 rs1 !extern
# Argument sets:
&ci imm rd
&ciw nzuimm rd
&cs_dw uimm rs1 rs2
-&cb imm rs1
&cr rd rs2
-&c_j imm
&c_shift shamt rd
-&c_ld uimm rd
-&c_sd uimm rs2
-
&c_addi16sp_lui imm_lui imm_addi16sp rd
# Formats 16:
@@ -70,20 +66,21 @@
@cs_2 ... ... ... .. ... .. &r rd=%rs1_3 rs1=%rs1_3 rs2=%rs2_3
@cs_d ... ... ... .. ... .. &s imm=%uimm_cl_d rs1=%rs1_3 rs2=%rs2_3
@cs_w ... ... ... .. ... .. &s imm=%uimm_cl_w rs1=%rs1_3 rs2=%rs2_3
-@cb ... ... ... .. ... .. &cb imm=%imm_cb rs1=%rs1_3
-@cj ... ........... .. &c_j imm=%imm_cj
+@cb ... ... ... .. ... .. &b imm=%imm_cb rs1=%rs1_3 rs2=0
+@cj ... ........... .. &j imm=%imm_cj rd=0
-@c_ld ... . ..... ..... .. &c_ld uimm=%uimm_6bit_ld %rd
-@c_lw ... . ..... ..... .. &c_ld uimm=%uimm_6bit_lw %rd
-@c_sd ... . ..... ..... .. &c_sd uimm=%uimm_6bit_sd rs2=%rs2_5
-@c_sw ... . ..... ..... .. &c_sd uimm=%uimm_6bit_sw rs2=%rs2_5
+@c_ld ... . ..... ..... .. &i imm=%uimm_6bit_ld %rd rs1=2
+@c_lw ... . ..... ..... .. &i imm=%uimm_6bit_lw %rd rs1=2
+@c_sd ... . ..... ..... .. &s imm=%uimm_6bit_sd rs1=2 rs2=%rs2_5
+@c_sw ... . ..... ..... .. &s imm=%uimm_6bit_sw rs1=2 rs2=%rs2_5
@c_addi16sp_lui ... . ..... ..... .. &c_addi16sp_lui %imm_lui %imm_addi16sp %rd
@c_shift ... . .. ... ..... .. &c_shift rd=%rs1_3 shamt=%nzuimm_6bit
@c_shift2 ... . .. ... ..... .. &c_shift rd=%rd shamt=%nzuimm_6bit
-@c_andi ... . .. ... ..... .. &ci imm=%imm_ci rd=%rs1_3
+@c_andi ... . .. ... ..... .. &i imm=%imm_ci rd=%rs1_3 rs1=%rs1_3
+
# *** RV64C Standard Extension (Quadrant 0) ***
c_addi4spn 000 ........ ... 00 @ciw
@@ -98,20 +95,20 @@ c_li 010 . ..... ..... 01 @ci
c_addi16sp_lui 011 . ..... ..... 01 @c_addi16sp_lui # shares opc with C.LUI
c_srli 100 . 00 ... ..... 01 @c_shift
c_srai 100 . 01 ... ..... 01 @c_shift
-c_andi 100 . 10 ... ..... 01 @c_andi
+andi 100 . 10 ... ..... 01 @c_andi
sub 100 0 11 ... 00 ... 01 @cs_2
xor 100 0 11 ... 01 ... 01 @cs_2
or 100 0 11 ... 10 ... 01 @cs_2
and 100 0 11 ... 11 ... 01 @cs_2
-c_j 101 ........... 01 @cj
-c_beqz 110 ... ... ..... 01 @cb
-c_bnez 111 ... ... ..... 01 @cb
+jal 101 ........... 01 @cj # c_j
+beq 110 ... ... ..... 01 @cb # c_beqz
+bne 111 ... ... ..... 01 @cb # c_bnez
# *** RV64C Standard Extension (Quadrant 2) ***
c_slli 000 . ..... ..... 10 @c_shift2
-c_fldsp 001 . ..... ..... 10 @c_ld
-c_lwsp 010 . ..... ..... 10 @c_lw
+fld 001 . ..... ..... 10 @c_ld # fldsp
+lw 010 . ..... ..... 10 @c_lw # lwsp
c_jr_mv 100 0 ..... ..... 10 @cr
c_ebreak_jalr_add 100 1 ..... ..... 10 @cr
-c_fsdsp 101 ...... ..... 10 @c_sd
-c_swsp 110 . ..... ..... 10 @c_sw
+fsd 101 ...... ..... 10 @c_sd # fsdsp
+sw 110 . ..... ..... 10 @c_sw # swsp
diff --git a/target/riscv/insn_trans/trans_rvc.inc.c b/target/riscv/insn_trans/trans_rvc.inc.c
index db9119ec9b..631e72c8b5 100644
--- a/target/riscv/insn_trans/trans_rvc.inc.c
+++ b/target/riscv/insn_trans/trans_rvc.inc.c
@@ -98,30 +98,6 @@ static bool trans_c_srai(DisasContext *ctx, arg_c_srai *a)
return trans_srai(ctx, &arg);
}
-static bool trans_c_andi(DisasContext *ctx, arg_c_andi *a)
-{
- arg_andi arg = { .rd = a->rd, .rs1 = a->rd, .imm = a->imm };
- return trans_andi(ctx, &arg);
-}
-
-static bool trans_c_j(DisasContext *ctx, arg_c_j *a)
-{
- arg_jal arg = { .rd = 0, .imm = a->imm };
- return trans_jal(ctx, &arg);
-}
-
-static bool trans_c_beqz(DisasContext *ctx, arg_c_beqz *a)
-{
- arg_beq arg = { .rs1 = a->rs1, .rs2 = 0, .imm = a->imm };
- return trans_beq(ctx, &arg);
-}
-
-static bool trans_c_bnez(DisasContext *ctx, arg_c_bnez *a)
-{
- arg_bne arg = { .rs1 = a->rs1, .rs2 = 0, .imm = a->imm };
- return trans_bne(ctx, &arg);
-}
-
static bool trans_c_slli(DisasContext *ctx, arg_c_slli *a)
{
int shamt = a->shamt;
@@ -138,18 +114,6 @@ static bool trans_c_slli(DisasContext *ctx, arg_c_slli *a)
return trans_slli(ctx, &arg);
}
-static bool trans_c_fldsp(DisasContext *ctx, arg_c_fldsp *a)
-{
- arg_fld arg = { .rd = a->rd, .rs1 = 2, .imm = a->uimm };
- return trans_fld(ctx, &arg);
-}
-
-static bool trans_c_lwsp(DisasContext *ctx, arg_c_lwsp *a)
-{
- arg_lw arg = { .rd = a->rd, .rs1 = 2, .imm = a->uimm };
- return trans_lw(ctx, &arg);
-}
-
static bool trans_c_jr_mv(DisasContext *ctx, arg_c_jr_mv *a)
{
if (a->rd != 0 && a->rs2 == 0) {
@@ -183,15 +147,3 @@ static bool trans_c_ebreak_jalr_add(DisasContext *ctx, arg_c_ebreak_jalr_add *a)
}
return false;
}
-
-static bool trans_c_fsdsp(DisasContext *ctx, arg_c_fsdsp *a)
-{
- arg_fsd arg = { .rs1 = 2, .rs2 = a->rs2, .imm = a->uimm };
- return trans_fsd(ctx, &arg);
-}
-
-static bool trans_c_swsp(DisasContext *ctx, arg_c_swsp *a)
-{
- arg_sw arg = { .rs1 = 2, .rs2 = a->rs2, .imm = a->uimm };
- return trans_sw(ctx, &arg);
-}
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* [Qemu-riscv] [PATCH v3 35/35] target/riscv: Remaining rvc insn reuse 32 bit translators
@ 2018-10-31 13:20 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-10-31 13:20 UTC (permalink / raw)
To: mjc, sagark, palmer, kbastian, Alistair.Francis
Cc: richard.henderson, peer.adelt, qemu-devel, qemu-riscv
only a translate functions of rvc need to handle special cases. For
the other rvc instruction we can remove the extra layer of indirection.
Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
---
target/riscv/insn16.decode | 37 +++++++++----------
target/riscv/insn_trans/trans_rvc.inc.c | 48 -------------------------
2 files changed, 17 insertions(+), 68 deletions(-)
diff --git a/target/riscv/insn16.decode b/target/riscv/insn16.decode
index 98dd672c7f..d88a0c78ab 100644
--- a/target/riscv/insn16.decode
+++ b/target/riscv/insn16.decode
@@ -46,19 +46,15 @@
&i imm rs1 rd !extern
&s imm rs1 rs2 !extern
&j imm rd !extern
+&b imm rs2 rs1 !extern
# Argument sets:
&ci imm rd
&ciw nzuimm rd
&cs_dw uimm rs1 rs2
-&cb imm rs1
&cr rd rs2
-&c_j imm
&c_shift shamt rd
-&c_ld uimm rd
-&c_sd uimm rs2
-
&c_addi16sp_lui imm_lui imm_addi16sp rd
# Formats 16:
@@ -70,20 +66,21 @@
@cs_2 ... ... ... .. ... .. &r rd=%rs1_3 rs1=%rs1_3 rs2=%rs2_3
@cs_d ... ... ... .. ... .. &s imm=%uimm_cl_d rs1=%rs1_3 rs2=%rs2_3
@cs_w ... ... ... .. ... .. &s imm=%uimm_cl_w rs1=%rs1_3 rs2=%rs2_3
-@cb ... ... ... .. ... .. &cb imm=%imm_cb rs1=%rs1_3
-@cj ... ........... .. &c_j imm=%imm_cj
+@cb ... ... ... .. ... .. &b imm=%imm_cb rs1=%rs1_3 rs2=0
+@cj ... ........... .. &j imm=%imm_cj rd=0
-@c_ld ... . ..... ..... .. &c_ld uimm=%uimm_6bit_ld %rd
-@c_lw ... . ..... ..... .. &c_ld uimm=%uimm_6bit_lw %rd
-@c_sd ... . ..... ..... .. &c_sd uimm=%uimm_6bit_sd rs2=%rs2_5
-@c_sw ... . ..... ..... .. &c_sd uimm=%uimm_6bit_sw rs2=%rs2_5
+@c_ld ... . ..... ..... .. &i imm=%uimm_6bit_ld %rd rs1=2
+@c_lw ... . ..... ..... .. &i imm=%uimm_6bit_lw %rd rs1=2
+@c_sd ... . ..... ..... .. &s imm=%uimm_6bit_sd rs1=2 rs2=%rs2_5
+@c_sw ... . ..... ..... .. &s imm=%uimm_6bit_sw rs1=2 rs2=%rs2_5
@c_addi16sp_lui ... . ..... ..... .. &c_addi16sp_lui %imm_lui %imm_addi16sp %rd
@c_shift ... . .. ... ..... .. &c_shift rd=%rs1_3 shamt=%nzuimm_6bit
@c_shift2 ... . .. ... ..... .. &c_shift rd=%rd shamt=%nzuimm_6bit
-@c_andi ... . .. ... ..... .. &ci imm=%imm_ci rd=%rs1_3
+@c_andi ... . .. ... ..... .. &i imm=%imm_ci rd=%rs1_3 rs1=%rs1_3
+
# *** RV64C Standard Extension (Quadrant 0) ***
c_addi4spn 000 ........ ... 00 @ciw
@@ -98,20 +95,20 @@ c_li 010 . ..... ..... 01 @ci
c_addi16sp_lui 011 . ..... ..... 01 @c_addi16sp_lui # shares opc with C.LUI
c_srli 100 . 00 ... ..... 01 @c_shift
c_srai 100 . 01 ... ..... 01 @c_shift
-c_andi 100 . 10 ... ..... 01 @c_andi
+andi 100 . 10 ... ..... 01 @c_andi
sub 100 0 11 ... 00 ... 01 @cs_2
xor 100 0 11 ... 01 ... 01 @cs_2
or 100 0 11 ... 10 ... 01 @cs_2
and 100 0 11 ... 11 ... 01 @cs_2
-c_j 101 ........... 01 @cj
-c_beqz 110 ... ... ..... 01 @cb
-c_bnez 111 ... ... ..... 01 @cb
+jal 101 ........... 01 @cj # c_j
+beq 110 ... ... ..... 01 @cb # c_beqz
+bne 111 ... ... ..... 01 @cb # c_bnez
# *** RV64C Standard Extension (Quadrant 2) ***
c_slli 000 . ..... ..... 10 @c_shift2
-c_fldsp 001 . ..... ..... 10 @c_ld
-c_lwsp 010 . ..... ..... 10 @c_lw
+fld 001 . ..... ..... 10 @c_ld # fldsp
+lw 010 . ..... ..... 10 @c_lw # lwsp
c_jr_mv 100 0 ..... ..... 10 @cr
c_ebreak_jalr_add 100 1 ..... ..... 10 @cr
-c_fsdsp 101 ...... ..... 10 @c_sd
-c_swsp 110 . ..... ..... 10 @c_sw
+fsd 101 ...... ..... 10 @c_sd # fsdsp
+sw 110 . ..... ..... 10 @c_sw # swsp
diff --git a/target/riscv/insn_trans/trans_rvc.inc.c b/target/riscv/insn_trans/trans_rvc.inc.c
index db9119ec9b..631e72c8b5 100644
--- a/target/riscv/insn_trans/trans_rvc.inc.c
+++ b/target/riscv/insn_trans/trans_rvc.inc.c
@@ -98,30 +98,6 @@ static bool trans_c_srai(DisasContext *ctx, arg_c_srai *a)
return trans_srai(ctx, &arg);
}
-static bool trans_c_andi(DisasContext *ctx, arg_c_andi *a)
-{
- arg_andi arg = { .rd = a->rd, .rs1 = a->rd, .imm = a->imm };
- return trans_andi(ctx, &arg);
-}
-
-static bool trans_c_j(DisasContext *ctx, arg_c_j *a)
-{
- arg_jal arg = { .rd = 0, .imm = a->imm };
- return trans_jal(ctx, &arg);
-}
-
-static bool trans_c_beqz(DisasContext *ctx, arg_c_beqz *a)
-{
- arg_beq arg = { .rs1 = a->rs1, .rs2 = 0, .imm = a->imm };
- return trans_beq(ctx, &arg);
-}
-
-static bool trans_c_bnez(DisasContext *ctx, arg_c_bnez *a)
-{
- arg_bne arg = { .rs1 = a->rs1, .rs2 = 0, .imm = a->imm };
- return trans_bne(ctx, &arg);
-}
-
static bool trans_c_slli(DisasContext *ctx, arg_c_slli *a)
{
int shamt = a->shamt;
@@ -138,18 +114,6 @@ static bool trans_c_slli(DisasContext *ctx, arg_c_slli *a)
return trans_slli(ctx, &arg);
}
-static bool trans_c_fldsp(DisasContext *ctx, arg_c_fldsp *a)
-{
- arg_fld arg = { .rd = a->rd, .rs1 = 2, .imm = a->uimm };
- return trans_fld(ctx, &arg);
-}
-
-static bool trans_c_lwsp(DisasContext *ctx, arg_c_lwsp *a)
-{
- arg_lw arg = { .rd = a->rd, .rs1 = 2, .imm = a->uimm };
- return trans_lw(ctx, &arg);
-}
-
static bool trans_c_jr_mv(DisasContext *ctx, arg_c_jr_mv *a)
{
if (a->rd != 0 && a->rs2 == 0) {
@@ -183,15 +147,3 @@ static bool trans_c_ebreak_jalr_add(DisasContext *ctx, arg_c_ebreak_jalr_add *a)
}
return false;
}
-
-static bool trans_c_fsdsp(DisasContext *ctx, arg_c_fsdsp *a)
-{
- arg_fsd arg = { .rs1 = 2, .rs2 = a->rs2, .imm = a->uimm };
- return trans_fsd(ctx, &arg);
-}
-
-static bool trans_c_swsp(DisasContext *ctx, arg_c_swsp *a)
-{
- arg_sw arg = { .rs1 = 2, .rs2 = a->rs2, .imm = a->uimm };
- return trans_sw(ctx, &arg);
-}
--
2.19.1
^ permalink raw reply related [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 01/35] target/riscv: Move CPURISCVState pointer to DisasContext
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 17:07 ` Richard Henderson
-1 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 17:07 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:19 PM, Bastian Koppelmann wrote:
> CPURISCVState is rarely used, so there is no need to pass it to every
> translate function. This paves the way for decodetree which only passes
> DisasContext to translate functions.
>
> Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> ---
> target/riscv/translate.c | 15 ++++++++-------
> 1 file changed, 8 insertions(+), 7 deletions(-)
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [PATCH v3 01/35] target/riscv: Move CPURISCVState pointer to DisasContext
@ 2018-10-31 17:07 ` Richard Henderson
0 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 17:07 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:19 PM, Bastian Koppelmann wrote:
> CPURISCVState is rarely used, so there is no need to pass it to every
> translate function. This paves the way for decodetree which only passes
> DisasContext to translate functions.
>
> Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> ---
> target/riscv/translate.c | 15 ++++++++-------
> 1 file changed, 8 insertions(+), 7 deletions(-)
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 03/35] target/riscv: Convert RVXI branch insns to decodetree
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 17:11 ` Richard Henderson
-1 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 17:11 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:19 PM, Bastian Koppelmann wrote:
> +static bool trans_jal(DisasContext *ctx, arg_jal *a, uint32_t insn)
> +{
> + gen_jal(ctx->env, ctx, a->rd, a->imm);
> + return true;
> +}
This patch didn't get updated properly. It looks like the removal of the insn
parameter got squashed into patch 4 instead.
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [PATCH v3 03/35] target/riscv: Convert RVXI branch insns to decodetree
@ 2018-10-31 17:11 ` Richard Henderson
0 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 17:11 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:19 PM, Bastian Koppelmann wrote:
> +static bool trans_jal(DisasContext *ctx, arg_jal *a, uint32_t insn)
> +{
> + gen_jal(ctx->env, ctx, a->rd, a->imm);
> + return true;
> +}
This patch didn't get updated properly. It looks like the removal of the insn
parameter got squashed into patch 4 instead.
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 05/35] target/riscv: Convert RV64I load/store insns to decodetree
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 17:14 ` Richard Henderson
-1 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 17:14 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:19 PM, Bastian Koppelmann wrote:
> this splits the 64-bit only instructions into its own decode file such
> that we generate the decoder for these instructions only for the RISC-V
> 64 bit target.
>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
> ---
> target/riscv/Makefile.objs | 8 +++++---
> target/riscv/insn64.decode | 25 +++++++++++++++++++++++++
> target/riscv/insn_trans/trans_rvi.inc.c | 20 ++++++++++++++++++++
> target/riscv/translate.c | 7 -------
> 4 files changed, 50 insertions(+), 10 deletions(-)
> create mode 100644 target/riscv/insn64.decode
I did suggest using insn32-64.decode, so that insn64.decode is available for an
actual 64-bit instruction word, which is mentioned in the "Extensions" section
of the ISA manual.
However,
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [PATCH v3 05/35] target/riscv: Convert RV64I load/store insns to decodetree
@ 2018-10-31 17:14 ` Richard Henderson
0 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 17:14 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:19 PM, Bastian Koppelmann wrote:
> this splits the 64-bit only instructions into its own decode file such
> that we generate the decoder for these instructions only for the RISC-V
> 64 bit target.
>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
> ---
> target/riscv/Makefile.objs | 8 +++++---
> target/riscv/insn64.decode | 25 +++++++++++++++++++++++++
> target/riscv/insn_trans/trans_rvi.inc.c | 20 ++++++++++++++++++++
> target/riscv/translate.c | 7 -------
> 4 files changed, 50 insertions(+), 10 deletions(-)
> create mode 100644 target/riscv/insn64.decode
I did suggest using insn32-64.decode, so that insn64.decode is available for an
actual 64-bit instruction word, which is mentioned in the "Extensions" section
of the ISA manual.
However,
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 06/35] target/riscv: Convert RVXI arithmetic insns to decodetree
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 17:15 ` Richard Henderson
-1 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 17:15 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
> we cannot remove the call to gen_arith() in decode_RV32_64G() since it
> is used to translate multiply instructions.
>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
> ---
> v2 -> v3:
> - moved 64-bit only insn to insn64.decode
> - dropped insn argument of trans_foo functions
>
> target/riscv/insn32.decode | 25 ++++
> target/riscv/insn64.decode | 13 ++
> target/riscv/insn_trans/trans_rvi.inc.c | 168 ++++++++++++++++++++++++
> target/riscv/translate.c | 9 --
> 4 files changed, 206 insertions(+), 9 deletions(-)
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [PATCH v3 06/35] target/riscv: Convert RVXI arithmetic insns to decodetree
@ 2018-10-31 17:15 ` Richard Henderson
0 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 17:15 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
> we cannot remove the call to gen_arith() in decode_RV32_64G() since it
> is used to translate multiply instructions.
>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
> ---
> v2 -> v3:
> - moved 64-bit only insn to insn64.decode
> - dropped insn argument of trans_foo functions
>
> target/riscv/insn32.decode | 25 ++++
> target/riscv/insn64.decode | 13 ++
> target/riscv/insn_trans/trans_rvi.inc.c | 168 ++++++++++++++++++++++++
> target/riscv/translate.c | 9 --
> 4 files changed, 206 insertions(+), 9 deletions(-)
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 01/35] target/riscv: Move CPURISCVState pointer to DisasContext
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 20:14 ` Alistair Francis
-1 siblings, 0 replies; 140+ messages in thread
From: Alistair Francis @ 2018-10-31 20:14 UTC (permalink / raw)
To: Bastian Koppelmann
Cc: Michael Clark, Sagar Karandikar, Palmer Dabbelt,
Alistair Francis, qemu-riscv, peer.adelt, Richard Henderson,
qemu-devel@nongnu.org Developers
On Wed, Oct 31, 2018 at 6:27 AM Bastian Koppelmann
<kbastian@mail.uni-paderborn.de> wrote:
>
> CPURISCVState is rarely used, so there is no need to pass it to every
> translate function. This paves the way for decodetree which only passes
> DisasContext to translate functions.
>
> Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> target/riscv/translate.c | 15 ++++++++-------
> 1 file changed, 8 insertions(+), 7 deletions(-)
>
> diff --git a/target/riscv/translate.c b/target/riscv/translate.c
> index 18d7b6d147..e81b9f097e 100644
> --- a/target/riscv/translate.c
> +++ b/target/riscv/translate.c
> @@ -52,6 +52,7 @@ typedef struct DisasContext {
> to any system register, which includes CSR_FRM, so we do not have
> to reset this known value. */
> int frm;
> + CPURISCVState *env;
> } DisasContext;
>
> /* convert riscv funct3 to qemu memop for load/store */
> @@ -1789,19 +1790,19 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
> }
> }
>
> -static void decode_opc(CPURISCVState *env, DisasContext *ctx)
> +static void decode_opc(DisasContext *ctx)
> {
> /* check for compressed insn */
> if (extract32(ctx->opcode, 0, 2) != 3) {
> - if (!riscv_has_ext(env, RVC)) {
> + if (!riscv_has_ext(ctx->env, RVC)) {
> gen_exception_illegal(ctx);
> } else {
> ctx->pc_succ_insn = ctx->base.pc_next + 2;
> - decode_RV32_64C(env, ctx);
> + decode_RV32_64C(ctx->env, ctx);
> }
> } else {
> ctx->pc_succ_insn = ctx->base.pc_next + 4;
> - decode_RV32_64G(env, ctx);
> + decode_RV32_64G(ctx->env, ctx);
> }
> }
>
> @@ -1846,10 +1847,10 @@ static bool riscv_tr_breakpoint_check(DisasContextBase *dcbase, CPUState *cpu,
> static void riscv_tr_translate_insn(DisasContextBase *dcbase, CPUState *cpu)
> {
> DisasContext *ctx = container_of(dcbase, DisasContext, base);
> - CPURISCVState *env = cpu->env_ptr;
> + ctx->env = cpu->env_ptr;
>
> - ctx->opcode = cpu_ldl_code(env, ctx->base.pc_next);
> - decode_opc(env, ctx);
> + ctx->opcode = cpu_ldl_code(ctx->env, ctx->base.pc_next);
> + decode_opc(ctx);
> ctx->base.pc_next = ctx->pc_succ_insn;
>
> if (ctx->base.is_jmp == DISAS_NEXT) {
> --
> 2.19.1
>
>
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [Qemu-devel] [PATCH v3 01/35] target/riscv: Move CPURISCVState pointer to DisasContext
@ 2018-10-31 20:14 ` Alistair Francis
0 siblings, 0 replies; 140+ messages in thread
From: Alistair Francis @ 2018-10-31 20:14 UTC (permalink / raw)
To: Bastian Koppelmann
Cc: Michael Clark, Sagar Karandikar, Palmer Dabbelt,
Alistair Francis, qemu-riscv, peer.adelt, Richard Henderson,
qemu-devel@nongnu.org Developers
On Wed, Oct 31, 2018 at 6:27 AM Bastian Koppelmann
<kbastian@mail.uni-paderborn.de> wrote:
>
> CPURISCVState is rarely used, so there is no need to pass it to every
> translate function. This paves the way for decodetree which only passes
> DisasContext to translate functions.
>
> Reviewed-by: Palmer Dabbelt <palmer@sifive.com>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> target/riscv/translate.c | 15 ++++++++-------
> 1 file changed, 8 insertions(+), 7 deletions(-)
>
> diff --git a/target/riscv/translate.c b/target/riscv/translate.c
> index 18d7b6d147..e81b9f097e 100644
> --- a/target/riscv/translate.c
> +++ b/target/riscv/translate.c
> @@ -52,6 +52,7 @@ typedef struct DisasContext {
> to any system register, which includes CSR_FRM, so we do not have
> to reset this known value. */
> int frm;
> + CPURISCVState *env;
> } DisasContext;
>
> /* convert riscv funct3 to qemu memop for load/store */
> @@ -1789,19 +1790,19 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
> }
> }
>
> -static void decode_opc(CPURISCVState *env, DisasContext *ctx)
> +static void decode_opc(DisasContext *ctx)
> {
> /* check for compressed insn */
> if (extract32(ctx->opcode, 0, 2) != 3) {
> - if (!riscv_has_ext(env, RVC)) {
> + if (!riscv_has_ext(ctx->env, RVC)) {
> gen_exception_illegal(ctx);
> } else {
> ctx->pc_succ_insn = ctx->base.pc_next + 2;
> - decode_RV32_64C(env, ctx);
> + decode_RV32_64C(ctx->env, ctx);
> }
> } else {
> ctx->pc_succ_insn = ctx->base.pc_next + 4;
> - decode_RV32_64G(env, ctx);
> + decode_RV32_64G(ctx->env, ctx);
> }
> }
>
> @@ -1846,10 +1847,10 @@ static bool riscv_tr_breakpoint_check(DisasContextBase *dcbase, CPUState *cpu,
> static void riscv_tr_translate_insn(DisasContextBase *dcbase, CPUState *cpu)
> {
> DisasContext *ctx = container_of(dcbase, DisasContext, base);
> - CPURISCVState *env = cpu->env_ptr;
> + ctx->env = cpu->env_ptr;
>
> - ctx->opcode = cpu_ldl_code(env, ctx->base.pc_next);
> - decode_opc(env, ctx);
> + ctx->opcode = cpu_ldl_code(ctx->env, ctx->base.pc_next);
> + decode_opc(ctx);
> ctx->base.pc_next = ctx->pc_succ_insn;
>
> if (ctx->base.is_jmp == DISAS_NEXT) {
> --
> 2.19.1
>
>
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 02/35] target/riscv: Activate decodetree and implemnt LUI & AUIPC
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 20:20 ` Alistair
-1 siblings, 0 replies; 140+ messages in thread
From: Alistair @ 2018-10-31 20:20 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: qemu-riscv, peer.adelt, richard.henderson, qemu-devel
On 10/31/18 6:19 AM, Bastian Koppelmann wrote:
> for now only LUI & AUIPC are decoded and translated. If decodetree fails, we
> fall back to the old decoder.
>
> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> v2 -> v3:
> - ex_shift_amount returns int
> - dropped insn argument of trans_foo functions
>
> target/riscv/Makefile.objs | 10 +++++++
> target/riscv/insn32.decode | 30 +++++++++++++++++++++
> target/riscv/insn_trans/trans_rvi.inc.c | 35 +++++++++++++++++++++++++
> target/riscv/translate.c | 31 ++++++++++++----------
> 4 files changed, 92 insertions(+), 14 deletions(-)
> create mode 100644 target/riscv/insn32.decode
> create mode 100644 target/riscv/insn_trans/trans_rvi.inc.c
>
> diff --git a/target/riscv/Makefile.objs b/target/riscv/Makefile.objs
> index fcc5d34c1f..ee995b3fc7 100644
> --- a/target/riscv/Makefile.objs
> +++ b/target/riscv/Makefile.objs
> @@ -1 +1,11 @@
> obj-y += translate.o op_helper.o cpu_helper.o cpu.o fpu_helper.o gdbstub.o pmp.o
> +
> +DECODETREE = $(SRC_PATH)/scripts/decodetree.py
> +
> +target/riscv/decode_insn32.inc.c: \
> + $(SRC_PATH)/target/riscv/insn32.decode $(DECODETREE)
> + $(call quiet-command, \
> + $(PYTHON) $(DECODETREE) -o $@ --decode decode_insn32 $<, \
> + "GEN", $(TARGET_DIR)$@)
> +
> +target/riscv/translate.o: target/riscv/decode_insn32.inc.c
> diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
> new file mode 100644
> index 0000000000..44d4e922b6
> --- /dev/null
> +++ b/target/riscv/insn32.decode
> @@ -0,0 +1,30 @@
> +#
> +# RISC-V translation routines for the RVXI Base Integer Instruction Set.
> +#
> +# Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
> +# Bastian Koppelmann, kbastian@mail.uni-paderborn.de
> +#
> +# This program is free software; you can redistribute it and/or modify it
> +# under the terms and conditions of the GNU General Public License,
> +# version 2 or later, as published by the Free Software Foundation.
> +#
> +# This program is distributed in the hope it will be useful, but WITHOUT
> +# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
> +# FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
> +# more details.
> +#
> +# You should have received a copy of the GNU General Public License along with
> +# this program. If not, see <http://www.gnu.org/licenses/>.
> +
> +# Fields:
> +%rd 7:5
> +
> +# immediates:
> +%imm_u 12:s20 !function=ex_shift_12
> +
> +# Formats 32:
> +@u .................... ..... ....... imm=%imm_u %rd
> +
> +# *** RV32I Base Instruction Set ***
> +lui .................... ..... 0110111 @u
> +auipc .................... ..... 0010111 @u
> diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
> new file mode 100644
> index 0000000000..9885a8d275
> --- /dev/null
> +++ b/target/riscv/insn_trans/trans_rvi.inc.c
> @@ -0,0 +1,35 @@
> +/*
> + * RISC-V translation routines for the RVXI Base Integer Instruction Set.
> + *
> + * Copyright (c) 2016-2017 Sagar Karandikar, sagark@eecs.berkeley.edu
> + * Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
> + * Bastian Koppelmann, kbastian@mail.uni-paderborn.de
> + *
> + * This program is free software; you can redistribute it and/or modify it
> + * under the terms and conditions of the GNU General Public License,
> + * version 2 or later, as published by the Free Software Foundation.
> + *
> + * This program is distributed in the hope it will be useful, but WITHOUT
> + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
> + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
> + * more details.
> + *
> + * You should have received a copy of the GNU General Public License along with
> + * this program. If not, see <http://www.gnu.org/licenses/>.
> + */
> +
> +static bool trans_lui(DisasContext *ctx, arg_lui *a)
> +{
> + if (a->rd != 0) {
> + tcg_gen_movi_tl(cpu_gpr[a->rd], a->imm);
> + }
> + return true;
> +}
> +
> +static bool trans_auipc(DisasContext *ctx, arg_auipc *a)
> +{
> + if (a->rd != 0) {
> + tcg_gen_movi_tl(cpu_gpr[a->rd], a->imm + ctx->base.pc_next);
> + }
> + return true;
> +}
> diff --git a/target/riscv/translate.c b/target/riscv/translate.c
> index e81b9f097e..ba57cb29f5 100644
> --- a/target/riscv/translate.c
> +++ b/target/riscv/translate.c
> @@ -1667,6 +1667,19 @@ static void decode_RV32_64C(CPURISCVState *env, DisasContext *ctx)
> }
> }
>
> +#define EX_SH(amount) \
> + static int ex_shift_##amount(int imm) \
> + { \
> + return imm << amount; \
> + }
> +EX_SH(12)
> +
> +bool decode_insn32(DisasContext *ctx, uint32_t insn);
> +/* Include the auto-generated decoder for 32 bit insn */
> +#include "decode_insn32.inc.c"
> +/* Include insn module translation function */
> +#include "insn_trans/trans_rvi.inc.c"
> +
> static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
> {
> int rs1;
> @@ -1687,19 +1700,6 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
> imm = GET_IMM(ctx->opcode);
>
> switch (op) {
> - case OPC_RISC_LUI:
> - if (rd == 0) {
> - break; /* NOP */
> - }
> - tcg_gen_movi_tl(cpu_gpr[rd], sextract64(ctx->opcode, 12, 20) << 12);
> - break;
> - case OPC_RISC_AUIPC:
> - if (rd == 0) {
> - break; /* NOP */
> - }
> - tcg_gen_movi_tl(cpu_gpr[rd], (sextract64(ctx->opcode, 12, 20) << 12) +
> - ctx->base.pc_next);
> - break;
> case OPC_RISC_JAL:
> imm = GET_JAL_IMM(ctx->opcode);
> gen_jal(env, ctx, rd, imm);
> @@ -1802,7 +1802,10 @@ static void decode_opc(DisasContext *ctx)
> }
> } else {
> ctx->pc_succ_insn = ctx->base.pc_next + 4;
> - decode_RV32_64G(ctx->env, ctx);
> + if (!decode_insn32(ctx, ctx->opcode)) {
> + /* fallback to old decoder */
> + decode_RV32_64G(ctx->env, ctx);
> + }
> }
> }
>
>
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [Qemu-devel] [PATCH v3 02/35] target/riscv: Activate decodetree and implemnt LUI & AUIPC
@ 2018-10-31 20:20 ` Alistair
0 siblings, 0 replies; 140+ messages in thread
From: Alistair @ 2018-10-31 20:20 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: qemu-riscv, peer.adelt, richard.henderson, qemu-devel
On 10/31/18 6:19 AM, Bastian Koppelmann wrote:
> for now only LUI & AUIPC are decoded and translated. If decodetree fails, we
> fall back to the old decoder.
>
> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> v2 -> v3:
> - ex_shift_amount returns int
> - dropped insn argument of trans_foo functions
>
> target/riscv/Makefile.objs | 10 +++++++
> target/riscv/insn32.decode | 30 +++++++++++++++++++++
> target/riscv/insn_trans/trans_rvi.inc.c | 35 +++++++++++++++++++++++++
> target/riscv/translate.c | 31 ++++++++++++----------
> 4 files changed, 92 insertions(+), 14 deletions(-)
> create mode 100644 target/riscv/insn32.decode
> create mode 100644 target/riscv/insn_trans/trans_rvi.inc.c
>
> diff --git a/target/riscv/Makefile.objs b/target/riscv/Makefile.objs
> index fcc5d34c1f..ee995b3fc7 100644
> --- a/target/riscv/Makefile.objs
> +++ b/target/riscv/Makefile.objs
> @@ -1 +1,11 @@
> obj-y += translate.o op_helper.o cpu_helper.o cpu.o fpu_helper.o gdbstub.o pmp.o
> +
> +DECODETREE = $(SRC_PATH)/scripts/decodetree.py
> +
> +target/riscv/decode_insn32.inc.c: \
> + $(SRC_PATH)/target/riscv/insn32.decode $(DECODETREE)
> + $(call quiet-command, \
> + $(PYTHON) $(DECODETREE) -o $@ --decode decode_insn32 $<, \
> + "GEN", $(TARGET_DIR)$@)
> +
> +target/riscv/translate.o: target/riscv/decode_insn32.inc.c
> diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
> new file mode 100644
> index 0000000000..44d4e922b6
> --- /dev/null
> +++ b/target/riscv/insn32.decode
> @@ -0,0 +1,30 @@
> +#
> +# RISC-V translation routines for the RVXI Base Integer Instruction Set.
> +#
> +# Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
> +# Bastian Koppelmann, kbastian@mail.uni-paderborn.de
> +#
> +# This program is free software; you can redistribute it and/or modify it
> +# under the terms and conditions of the GNU General Public License,
> +# version 2 or later, as published by the Free Software Foundation.
> +#
> +# This program is distributed in the hope it will be useful, but WITHOUT
> +# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
> +# FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
> +# more details.
> +#
> +# You should have received a copy of the GNU General Public License along with
> +# this program. If not, see <http://www.gnu.org/licenses/>.
> +
> +# Fields:
> +%rd 7:5
> +
> +# immediates:
> +%imm_u 12:s20 !function=ex_shift_12
> +
> +# Formats 32:
> +@u .................... ..... ....... imm=%imm_u %rd
> +
> +# *** RV32I Base Instruction Set ***
> +lui .................... ..... 0110111 @u
> +auipc .................... ..... 0010111 @u
> diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
> new file mode 100644
> index 0000000000..9885a8d275
> --- /dev/null
> +++ b/target/riscv/insn_trans/trans_rvi.inc.c
> @@ -0,0 +1,35 @@
> +/*
> + * RISC-V translation routines for the RVXI Base Integer Instruction Set.
> + *
> + * Copyright (c) 2016-2017 Sagar Karandikar, sagark@eecs.berkeley.edu
> + * Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
> + * Bastian Koppelmann, kbastian@mail.uni-paderborn.de
> + *
> + * This program is free software; you can redistribute it and/or modify it
> + * under the terms and conditions of the GNU General Public License,
> + * version 2 or later, as published by the Free Software Foundation.
> + *
> + * This program is distributed in the hope it will be useful, but WITHOUT
> + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
> + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
> + * more details.
> + *
> + * You should have received a copy of the GNU General Public License along with
> + * this program. If not, see <http://www.gnu.org/licenses/>.
> + */
> +
> +static bool trans_lui(DisasContext *ctx, arg_lui *a)
> +{
> + if (a->rd != 0) {
> + tcg_gen_movi_tl(cpu_gpr[a->rd], a->imm);
> + }
> + return true;
> +}
> +
> +static bool trans_auipc(DisasContext *ctx, arg_auipc *a)
> +{
> + if (a->rd != 0) {
> + tcg_gen_movi_tl(cpu_gpr[a->rd], a->imm + ctx->base.pc_next);
> + }
> + return true;
> +}
> diff --git a/target/riscv/translate.c b/target/riscv/translate.c
> index e81b9f097e..ba57cb29f5 100644
> --- a/target/riscv/translate.c
> +++ b/target/riscv/translate.c
> @@ -1667,6 +1667,19 @@ static void decode_RV32_64C(CPURISCVState *env, DisasContext *ctx)
> }
> }
>
> +#define EX_SH(amount) \
> + static int ex_shift_##amount(int imm) \
> + { \
> + return imm << amount; \
> + }
> +EX_SH(12)
> +
> +bool decode_insn32(DisasContext *ctx, uint32_t insn);
> +/* Include the auto-generated decoder for 32 bit insn */
> +#include "decode_insn32.inc.c"
> +/* Include insn module translation function */
> +#include "insn_trans/trans_rvi.inc.c"
> +
> static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
> {
> int rs1;
> @@ -1687,19 +1700,6 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
> imm = GET_IMM(ctx->opcode);
>
> switch (op) {
> - case OPC_RISC_LUI:
> - if (rd == 0) {
> - break; /* NOP */
> - }
> - tcg_gen_movi_tl(cpu_gpr[rd], sextract64(ctx->opcode, 12, 20) << 12);
> - break;
> - case OPC_RISC_AUIPC:
> - if (rd == 0) {
> - break; /* NOP */
> - }
> - tcg_gen_movi_tl(cpu_gpr[rd], (sextract64(ctx->opcode, 12, 20) << 12) +
> - ctx->base.pc_next);
> - break;
> case OPC_RISC_JAL:
> imm = GET_JAL_IMM(ctx->opcode);
> gen_jal(env, ctx, rd, imm);
> @@ -1802,7 +1802,10 @@ static void decode_opc(DisasContext *ctx)
> }
> } else {
> ctx->pc_succ_insn = ctx->base.pc_next + 4;
> - decode_RV32_64G(ctx->env, ctx);
> + if (!decode_insn32(ctx, ctx->opcode)) {
> + /* fallback to old decoder */
> + decode_RV32_64G(ctx->env, ctx);
> + }
> }
> }
>
>
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 05/35] target/riscv: Convert RV64I load/store insns to decodetree
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 20:26 ` Alistair
-1 siblings, 0 replies; 140+ messages in thread
From: Alistair @ 2018-10-31 20:26 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: qemu-riscv, peer.adelt, richard.henderson, qemu-devel
On 10/31/18 6:19 AM, Bastian Koppelmann wrote:
> this splits the 64-bit only instructions into its own decode file such
> that we generate the decoder for these instructions only for the RISC-V
> 64 bit target.
>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> target/riscv/Makefile.objs | 8 +++++---
> target/riscv/insn64.decode | 25 +++++++++++++++++++++++++
> target/riscv/insn_trans/trans_rvi.inc.c | 20 ++++++++++++++++++++
> target/riscv/translate.c | 7 -------
> 4 files changed, 50 insertions(+), 10 deletions(-)
> create mode 100644 target/riscv/insn64.decode
>
> diff --git a/target/riscv/Makefile.objs b/target/riscv/Makefile.objs
> index ee995b3fc7..b9b8152cc2 100644
> --- a/target/riscv/Makefile.objs
> +++ b/target/riscv/Makefile.objs
> @@ -2,10 +2,12 @@ obj-y += translate.o op_helper.o cpu_helper.o cpu.o fpu_helper.o gdbstub.o pmp.o
>
> DECODETREE = $(SRC_PATH)/scripts/decodetree.py
>
> -target/riscv/decode_insn32.inc.c: \
> - $(SRC_PATH)/target/riscv/insn32.decode $(DECODETREE)
> +decode32-y = $(SRC_PATH)/target/riscv/insn32.decode
> +decode32-$(TARGET_RISCV64) += $(SRC_PATH)/target/riscv/insn64.decode
> +
> +target/riscv/decode_insn32.inc.c: $(decode32-y) $(DECODETREE)
> $(call quiet-command, \
> - $(PYTHON) $(DECODETREE) -o $@ --decode decode_insn32 $<, \
> + $(PYTHON) $(DECODETREE) -o $@ --decode decode_insn32 $(decode32-y), \
> "GEN", $(TARGET_DIR)$@)
>
> target/riscv/translate.o: target/riscv/decode_insn32.inc.c
> diff --git a/target/riscv/insn64.decode b/target/riscv/insn64.decode
> new file mode 100644
> index 0000000000..439d4e2c58
> --- /dev/null
> +++ b/target/riscv/insn64.decode
> @@ -0,0 +1,25 @@
> +#
> +# RISC-V translation routines for the RV Instruction Set.
> +#
> +# Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
> +# Bastian Koppelmann, kbastian@mail.uni-paderborn.de
> +#
> +# This program is free software; you can redistribute it and/or modify it
> +# under the terms and conditions of the GNU General Public License,
> +# version 2 or later, as published by the Free Software Foundation.
> +#
> +# This program is distributed in the hope it will be useful, but WITHOUT
> +# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
> +# FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
> +# more details.
> +#
> +# You should have received a copy of the GNU General Public License along with
> +# this program. If not, see <http://www.gnu.org/licenses/>.
> +
> +# This is concatenated with insn32.decode for risc64 targets.
> +# Most of the fields and formats are there.
> +
> +# *** RV64I Base Instruction Set (in addition to RV32I) ***
> +lwu ............ ..... 110 ..... 0000011 @i
> +ld ............ ..... 011 ..... 0000011 @i
> +sd ....... ..... ..... 011 ..... 0100011 @s
> diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
> index f3b88ebb69..39a20a70e8 100644
> --- a/target/riscv/insn_trans/trans_rvi.inc.c
> +++ b/target/riscv/insn_trans/trans_rvi.inc.c
> @@ -130,3 +130,23 @@ static bool trans_sw(DisasContext *ctx, arg_sw *a)
> gen_store(ctx, OPC_RISC_SW, a->rs1, a->rs2, a->imm);
> return true;
> }
> +
> +#ifdef TARGET_RISCV64
> +static bool trans_lwu(DisasContext *ctx, arg_lwu *a)
> +{
> + gen_load(ctx, OPC_RISC_LWU, a->rd, a->rs1, a->imm);
> + return true;
> +}
> +
> +static bool trans_ld(DisasContext *ctx, arg_ld *a)
> +{
> + gen_load(ctx, OPC_RISC_LD, a->rd, a->rs1, a->imm);
> + return true;
> +}
> +
> +static bool trans_sd(DisasContext *ctx, arg_sd *a)
> +{
> + gen_store(ctx, OPC_RISC_SD, a->rs1, a->rs2, a->imm);
> + return true;
> +}
> +#endif
> diff --git a/target/riscv/translate.c b/target/riscv/translate.c
> index 909f7cd013..244855c82d 100644
> --- a/target/riscv/translate.c
> +++ b/target/riscv/translate.c
> @@ -1701,13 +1701,6 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
> imm = GET_IMM(ctx->opcode);
>
> switch (op) {
> - case OPC_RISC_LOAD:
> - gen_load(ctx, MASK_OP_LOAD(ctx->opcode), rd, rs1, imm);
> - break;
> - case OPC_RISC_STORE:
> - gen_store(ctx, MASK_OP_STORE(ctx->opcode), rs1, rs2,
> - GET_STORE_IMM(ctx->opcode));
> - break;
> case OPC_RISC_ARITH_IMM:
> #if defined(TARGET_RISCV64)
> case OPC_RISC_ARITH_IMM_W:
>
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [Qemu-devel] [PATCH v3 05/35] target/riscv: Convert RV64I load/store insns to decodetree
@ 2018-10-31 20:26 ` Alistair
0 siblings, 0 replies; 140+ messages in thread
From: Alistair @ 2018-10-31 20:26 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: qemu-riscv, peer.adelt, richard.henderson, qemu-devel
On 10/31/18 6:19 AM, Bastian Koppelmann wrote:
> this splits the 64-bit only instructions into its own decode file such
> that we generate the decoder for these instructions only for the RISC-V
> 64 bit target.
>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> target/riscv/Makefile.objs | 8 +++++---
> target/riscv/insn64.decode | 25 +++++++++++++++++++++++++
> target/riscv/insn_trans/trans_rvi.inc.c | 20 ++++++++++++++++++++
> target/riscv/translate.c | 7 -------
> 4 files changed, 50 insertions(+), 10 deletions(-)
> create mode 100644 target/riscv/insn64.decode
>
> diff --git a/target/riscv/Makefile.objs b/target/riscv/Makefile.objs
> index ee995b3fc7..b9b8152cc2 100644
> --- a/target/riscv/Makefile.objs
> +++ b/target/riscv/Makefile.objs
> @@ -2,10 +2,12 @@ obj-y += translate.o op_helper.o cpu_helper.o cpu.o fpu_helper.o gdbstub.o pmp.o
>
> DECODETREE = $(SRC_PATH)/scripts/decodetree.py
>
> -target/riscv/decode_insn32.inc.c: \
> - $(SRC_PATH)/target/riscv/insn32.decode $(DECODETREE)
> +decode32-y = $(SRC_PATH)/target/riscv/insn32.decode
> +decode32-$(TARGET_RISCV64) += $(SRC_PATH)/target/riscv/insn64.decode
> +
> +target/riscv/decode_insn32.inc.c: $(decode32-y) $(DECODETREE)
> $(call quiet-command, \
> - $(PYTHON) $(DECODETREE) -o $@ --decode decode_insn32 $<, \
> + $(PYTHON) $(DECODETREE) -o $@ --decode decode_insn32 $(decode32-y), \
> "GEN", $(TARGET_DIR)$@)
>
> target/riscv/translate.o: target/riscv/decode_insn32.inc.c
> diff --git a/target/riscv/insn64.decode b/target/riscv/insn64.decode
> new file mode 100644
> index 0000000000..439d4e2c58
> --- /dev/null
> +++ b/target/riscv/insn64.decode
> @@ -0,0 +1,25 @@
> +#
> +# RISC-V translation routines for the RV Instruction Set.
> +#
> +# Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
> +# Bastian Koppelmann, kbastian@mail.uni-paderborn.de
> +#
> +# This program is free software; you can redistribute it and/or modify it
> +# under the terms and conditions of the GNU General Public License,
> +# version 2 or later, as published by the Free Software Foundation.
> +#
> +# This program is distributed in the hope it will be useful, but WITHOUT
> +# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
> +# FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
> +# more details.
> +#
> +# You should have received a copy of the GNU General Public License along with
> +# this program. If not, see <http://www.gnu.org/licenses/>.
> +
> +# This is concatenated with insn32.decode for risc64 targets.
> +# Most of the fields and formats are there.
> +
> +# *** RV64I Base Instruction Set (in addition to RV32I) ***
> +lwu ............ ..... 110 ..... 0000011 @i
> +ld ............ ..... 011 ..... 0000011 @i
> +sd ....... ..... ..... 011 ..... 0100011 @s
> diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
> index f3b88ebb69..39a20a70e8 100644
> --- a/target/riscv/insn_trans/trans_rvi.inc.c
> +++ b/target/riscv/insn_trans/trans_rvi.inc.c
> @@ -130,3 +130,23 @@ static bool trans_sw(DisasContext *ctx, arg_sw *a)
> gen_store(ctx, OPC_RISC_SW, a->rs1, a->rs2, a->imm);
> return true;
> }
> +
> +#ifdef TARGET_RISCV64
> +static bool trans_lwu(DisasContext *ctx, arg_lwu *a)
> +{
> + gen_load(ctx, OPC_RISC_LWU, a->rd, a->rs1, a->imm);
> + return true;
> +}
> +
> +static bool trans_ld(DisasContext *ctx, arg_ld *a)
> +{
> + gen_load(ctx, OPC_RISC_LD, a->rd, a->rs1, a->imm);
> + return true;
> +}
> +
> +static bool trans_sd(DisasContext *ctx, arg_sd *a)
> +{
> + gen_store(ctx, OPC_RISC_SD, a->rs1, a->rs2, a->imm);
> + return true;
> +}
> +#endif
> diff --git a/target/riscv/translate.c b/target/riscv/translate.c
> index 909f7cd013..244855c82d 100644
> --- a/target/riscv/translate.c
> +++ b/target/riscv/translate.c
> @@ -1701,13 +1701,6 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
> imm = GET_IMM(ctx->opcode);
>
> switch (op) {
> - case OPC_RISC_LOAD:
> - gen_load(ctx, MASK_OP_LOAD(ctx->opcode), rd, rs1, imm);
> - break;
> - case OPC_RISC_STORE:
> - gen_store(ctx, MASK_OP_STORE(ctx->opcode), rs1, rs2,
> - GET_STORE_IMM(ctx->opcode));
> - break;
> case OPC_RISC_ARITH_IMM:
> #if defined(TARGET_RISCV64)
> case OPC_RISC_ARITH_IMM_W:
>
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 06/35] target/riscv: Convert RVXI arithmetic insns to decodetree
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 20:29 ` Alistair
-1 siblings, 0 replies; 140+ messages in thread
From: Alistair @ 2018-10-31 20:29 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: qemu-riscv, peer.adelt, richard.henderson, qemu-devel
On 10/31/18 6:20 AM, Bastian Koppelmann wrote:
> we cannot remove the call to gen_arith() in decode_RV32_64G() since it
> is used to translate multiply instructions.
>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> v2 -> v3:
> - moved 64-bit only insn to insn64.decode
> - dropped insn argument of trans_foo functions
>
> target/riscv/insn32.decode | 25 ++++
> target/riscv/insn64.decode | 13 ++
> target/riscv/insn_trans/trans_rvi.inc.c | 168 ++++++++++++++++++++++++
> target/riscv/translate.c | 9 --
> 4 files changed, 206 insertions(+), 9 deletions(-)
>
> diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
> index 076de873c4..4fd88f48d2 100644
> --- a/target/riscv/insn32.decode
> +++ b/target/riscv/insn32.decode
> @@ -21,6 +21,8 @@
> %rs1 15:5
> %rd 7:5
>
> +%sh6 20:6
> +
> # immediates:
> %imm_i 20:s12
> %imm_s 25:s7 7:5
> @@ -30,14 +32,18 @@
>
> # Argument sets:
> &b imm rs2 rs1
> +&shift shamt rs1 rd
>
> # Formats 32:
> +@r ....... ..... ..... ... ..... ....... %rs2 %rs1 %rd
> @i ............ ..... ... ..... ....... imm=%imm_i %rs1 %rd
> @b ....... ..... ..... ... ..... ....... &b imm=%imm_b %rs2 %rs1
> @s ....... ..... ..... ... ..... ....... imm=%imm_s %rs2 %rs1
> @u .................... ..... ....... imm=%imm_u %rd
> @j .................... ..... ....... imm=%imm_j %rd
>
> +@sh6 ...... ...... ..... ... ..... ....... &shift shamt=%sh6 %rs1 %rd
> +
> # *** RV32I Base Instruction Set ***
> lui .................... ..... 0110111 @u
> auipc .................... ..... 0010111 @u
> @@ -57,3 +63,22 @@ lhu ............ ..... 101 ..... 0000011 @i
> sb ....... ..... ..... 000 ..... 0100011 @s
> sh ....... ..... ..... 001 ..... 0100011 @s
> sw ....... ..... ..... 010 ..... 0100011 @s
> +addi ............ ..... 000 ..... 0010011 @i
> +slti ............ ..... 010 ..... 0010011 @i
> +sltiu ............ ..... 011 ..... 0010011 @i
> +xori ............ ..... 100 ..... 0010011 @i
> +ori ............ ..... 110 ..... 0010011 @i
> +andi ............ ..... 111 ..... 0010011 @i
> +slli 000000 ...... ..... 001 ..... 0010011 @sh6
> +srli 000000 ...... ..... 101 ..... 0010011 @sh6
> +srai 010000 ...... ..... 101 ..... 0010011 @sh6
> +add 0000000 ..... ..... 000 ..... 0110011 @r
> +sub 0100000 ..... ..... 000 ..... 0110011 @r
> +sll 0000000 ..... ..... 001 ..... 0110011 @r
> +slt 0000000 ..... ..... 010 ..... 0110011 @r
> +sltu 0000000 ..... ..... 011 ..... 0110011 @r
> +xor 0000000 ..... ..... 100 ..... 0110011 @r
> +srl 0000000 ..... ..... 101 ..... 0110011 @r
> +sra 0100000 ..... ..... 101 ..... 0110011 @r
> +or 0000000 ..... ..... 110 ..... 0110011 @r
> +and 0000000 ..... ..... 111 ..... 0110011 @r
> diff --git a/target/riscv/insn64.decode b/target/riscv/insn64.decode
> index 439d4e2c58..9a35f2aa19 100644
> --- a/target/riscv/insn64.decode
> +++ b/target/riscv/insn64.decode
> @@ -19,7 +19,20 @@
> # This is concatenated with insn32.decode for risc64 targets.
> # Most of the fields and formats are there.
>
> +%sh5 20:5
> +
> +@sh5 ....... ..... ..... ... ..... ....... &shift shamt=%sh5 %rs1 %rd
> +
> # *** RV64I Base Instruction Set (in addition to RV32I) ***
> lwu ............ ..... 110 ..... 0000011 @i
> ld ............ ..... 011 ..... 0000011 @i
> sd ....... ..... ..... 011 ..... 0100011 @s
> +addiw ............ ..... 000 ..... 0011011 @i
> +slliw 0000000 ..... ..... 001 ..... 0011011 @sh5
> +srliw 0000000 ..... ..... 101 ..... 0011011 @sh5
> +sraiw 0100000 ..... ..... 101 ..... 0011011 @sh5
> +addw 0000000 ..... ..... 000 ..... 0111011 @r
> +subw 0100000 ..... ..... 000 ..... 0111011 @r
> +sllw 0000000 ..... ..... 001 ..... 0111011 @r
> +srlw 0000000 ..... ..... 101 ..... 0111011 @r
> +sraw 0100000 ..... ..... 101 ..... 0111011 @r
> diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
> index 39a20a70e8..01f751650a 100644
> --- a/target/riscv/insn_trans/trans_rvi.inc.c
> +++ b/target/riscv/insn_trans/trans_rvi.inc.c
> @@ -150,3 +150,171 @@ static bool trans_sd(DisasContext *ctx, arg_sd *a)
> return true;
> }
> #endif
> +
> +static bool trans_addi(DisasContext *ctx, arg_addi *a)
> +{
> + gen_arith_imm(ctx, OPC_RISC_ADDI, a->rd, a->rs1, a->imm);
> + return true;
> +}
> +
> +static bool trans_slti(DisasContext *ctx, arg_slti *a)
> +{
> + gen_arith_imm(ctx, OPC_RISC_SLTI, a->rd, a->rs1, a->imm);
> + return true;
> +}
> +
> +static bool trans_sltiu(DisasContext *ctx, arg_sltiu *a)
> +{
> + gen_arith_imm(ctx, OPC_RISC_SLTIU, a->rd, a->rs1, a->imm);
> + return true;
> +}
> +
> +static bool trans_xori(DisasContext *ctx, arg_xori *a)
> +{
> + gen_arith_imm(ctx, OPC_RISC_XORI, a->rd, a->rs1, a->imm);
> + return true;
> +}
> +static bool trans_ori(DisasContext *ctx, arg_ori *a)
> +{
> + gen_arith_imm(ctx, OPC_RISC_ORI, a->rd, a->rs1, a->imm);
> + return true;
> +}
> +static bool trans_andi(DisasContext *ctx, arg_andi *a)
> +{
> + gen_arith_imm(ctx, OPC_RISC_ANDI, a->rd, a->rs1, a->imm);
> + return true;
> +}
> +static bool trans_slli(DisasContext *ctx, arg_slli *a)
> +{
> + gen_arith_imm(ctx, OPC_RISC_SLLI, a->rd, a->rs1, a->shamt);
> + return true;
> +}
> +
> +static bool trans_srli(DisasContext *ctx, arg_srli *a)
> +{
> + gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_I, a->rd, a->rs1, a->shamt);
> + return true;
> +}
> +
> +static bool trans_srai(DisasContext *ctx, arg_srai *a)
> +{
> + gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_I, a->rd, a->rs1, a->shamt | 0x400);
> + return true;
> +}
> +
> +static bool trans_add(DisasContext *ctx, arg_add *a)
> +{
> + gen_arith(ctx, OPC_RISC_ADD, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_sub(DisasContext *ctx, arg_sub *a)
> +{
> + gen_arith(ctx, OPC_RISC_SUB, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_sll(DisasContext *ctx, arg_sll *a)
> +{
> + gen_arith(ctx, OPC_RISC_SLL, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_slt(DisasContext *ctx, arg_slt *a)
> +{
> + gen_arith(ctx, OPC_RISC_SLT, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_sltu(DisasContext *ctx, arg_sltu *a)
> +{
> + gen_arith(ctx, OPC_RISC_SLTU, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_xor(DisasContext *ctx, arg_xor *a)
> +{
> + gen_arith(ctx, OPC_RISC_XOR, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_srl(DisasContext *ctx, arg_srl *a)
> +{
> + gen_arith(ctx, OPC_RISC_SRL, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_sra(DisasContext *ctx, arg_sra *a)
> +{
> + gen_arith(ctx, OPC_RISC_SRA, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_or(DisasContext *ctx, arg_or *a)
> +{
> + gen_arith(ctx, OPC_RISC_OR, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_and(DisasContext *ctx, arg_and *a)
> +{
> + gen_arith(ctx, OPC_RISC_AND, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +#ifdef TARGET_RISCV64
> +static bool trans_addiw(DisasContext *ctx, arg_addiw *a)
> +{
> + gen_arith_imm(ctx, OPC_RISC_ADDIW, a->rd, a->rs1, a->imm);
> + return true;
> +}
> +
> +static bool trans_slliw(DisasContext *ctx, arg_slliw *a)
> +{
> + gen_arith_imm(ctx, OPC_RISC_SLLIW, a->rd, a->rs1, a->shamt);
> + return true;
> +}
> +
> +static bool trans_srliw(DisasContext *ctx, arg_srliw *a)
> +{
> + gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_IW, a->rd, a->rs1, a->shamt);
> + return true;
> +}
> +
> +static bool trans_sraiw(DisasContext *ctx, arg_sraiw *a)
> +{
> + gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_IW , a->rd, a->rs1,
> + a->shamt | 0x400);
> + return true;
> +}
> +
> +static bool trans_addw(DisasContext *ctx, arg_addw *a)
> +{
> + gen_arith(ctx, OPC_RISC_ADDW, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_subw(DisasContext *ctx, arg_subw *a)
> +{
> + gen_arith(ctx, OPC_RISC_SUBW, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_sllw(DisasContext *ctx, arg_sllw *a)
> +{
> + gen_arith(ctx, OPC_RISC_SLLW, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_srlw(DisasContext *ctx, arg_srlw *a)
> +{
> + gen_arith(ctx, OPC_RISC_SRLW, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_sraw(DisasContext *ctx, arg_sraw *a)
> +{
> + gen_arith(ctx, OPC_RISC_SRAW, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +#endif
> diff --git a/target/riscv/translate.c b/target/riscv/translate.c
> index 244855c82d..855c241e97 100644
> --- a/target/riscv/translate.c
> +++ b/target/riscv/translate.c
> @@ -1701,15 +1701,6 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
> imm = GET_IMM(ctx->opcode);
>
> switch (op) {
> - case OPC_RISC_ARITH_IMM:
> -#if defined(TARGET_RISCV64)
> - case OPC_RISC_ARITH_IMM_W:
> -#endif
> - if (rd == 0) {
> - break; /* NOP */
> - }
> - gen_arith_imm(ctx, MASK_OP_ARITH_IMM(ctx->opcode), rd, rs1, imm);
> - break;
> case OPC_RISC_ARITH:
> #if defined(TARGET_RISCV64)
> case OPC_RISC_ARITH_W:
>
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [Qemu-devel] [PATCH v3 06/35] target/riscv: Convert RVXI arithmetic insns to decodetree
@ 2018-10-31 20:29 ` Alistair
0 siblings, 0 replies; 140+ messages in thread
From: Alistair @ 2018-10-31 20:29 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: qemu-riscv, peer.adelt, richard.henderson, qemu-devel
On 10/31/18 6:20 AM, Bastian Koppelmann wrote:
> we cannot remove the call to gen_arith() in decode_RV32_64G() since it
> is used to translate multiply instructions.
>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> v2 -> v3:
> - moved 64-bit only insn to insn64.decode
> - dropped insn argument of trans_foo functions
>
> target/riscv/insn32.decode | 25 ++++
> target/riscv/insn64.decode | 13 ++
> target/riscv/insn_trans/trans_rvi.inc.c | 168 ++++++++++++++++++++++++
> target/riscv/translate.c | 9 --
> 4 files changed, 206 insertions(+), 9 deletions(-)
>
> diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
> index 076de873c4..4fd88f48d2 100644
> --- a/target/riscv/insn32.decode
> +++ b/target/riscv/insn32.decode
> @@ -21,6 +21,8 @@
> %rs1 15:5
> %rd 7:5
>
> +%sh6 20:6
> +
> # immediates:
> %imm_i 20:s12
> %imm_s 25:s7 7:5
> @@ -30,14 +32,18 @@
>
> # Argument sets:
> &b imm rs2 rs1
> +&shift shamt rs1 rd
>
> # Formats 32:
> +@r ....... ..... ..... ... ..... ....... %rs2 %rs1 %rd
> @i ............ ..... ... ..... ....... imm=%imm_i %rs1 %rd
> @b ....... ..... ..... ... ..... ....... &b imm=%imm_b %rs2 %rs1
> @s ....... ..... ..... ... ..... ....... imm=%imm_s %rs2 %rs1
> @u .................... ..... ....... imm=%imm_u %rd
> @j .................... ..... ....... imm=%imm_j %rd
>
> +@sh6 ...... ...... ..... ... ..... ....... &shift shamt=%sh6 %rs1 %rd
> +
> # *** RV32I Base Instruction Set ***
> lui .................... ..... 0110111 @u
> auipc .................... ..... 0010111 @u
> @@ -57,3 +63,22 @@ lhu ............ ..... 101 ..... 0000011 @i
> sb ....... ..... ..... 000 ..... 0100011 @s
> sh ....... ..... ..... 001 ..... 0100011 @s
> sw ....... ..... ..... 010 ..... 0100011 @s
> +addi ............ ..... 000 ..... 0010011 @i
> +slti ............ ..... 010 ..... 0010011 @i
> +sltiu ............ ..... 011 ..... 0010011 @i
> +xori ............ ..... 100 ..... 0010011 @i
> +ori ............ ..... 110 ..... 0010011 @i
> +andi ............ ..... 111 ..... 0010011 @i
> +slli 000000 ...... ..... 001 ..... 0010011 @sh6
> +srli 000000 ...... ..... 101 ..... 0010011 @sh6
> +srai 010000 ...... ..... 101 ..... 0010011 @sh6
> +add 0000000 ..... ..... 000 ..... 0110011 @r
> +sub 0100000 ..... ..... 000 ..... 0110011 @r
> +sll 0000000 ..... ..... 001 ..... 0110011 @r
> +slt 0000000 ..... ..... 010 ..... 0110011 @r
> +sltu 0000000 ..... ..... 011 ..... 0110011 @r
> +xor 0000000 ..... ..... 100 ..... 0110011 @r
> +srl 0000000 ..... ..... 101 ..... 0110011 @r
> +sra 0100000 ..... ..... 101 ..... 0110011 @r
> +or 0000000 ..... ..... 110 ..... 0110011 @r
> +and 0000000 ..... ..... 111 ..... 0110011 @r
> diff --git a/target/riscv/insn64.decode b/target/riscv/insn64.decode
> index 439d4e2c58..9a35f2aa19 100644
> --- a/target/riscv/insn64.decode
> +++ b/target/riscv/insn64.decode
> @@ -19,7 +19,20 @@
> # This is concatenated with insn32.decode for risc64 targets.
> # Most of the fields and formats are there.
>
> +%sh5 20:5
> +
> +@sh5 ....... ..... ..... ... ..... ....... &shift shamt=%sh5 %rs1 %rd
> +
> # *** RV64I Base Instruction Set (in addition to RV32I) ***
> lwu ............ ..... 110 ..... 0000011 @i
> ld ............ ..... 011 ..... 0000011 @i
> sd ....... ..... ..... 011 ..... 0100011 @s
> +addiw ............ ..... 000 ..... 0011011 @i
> +slliw 0000000 ..... ..... 001 ..... 0011011 @sh5
> +srliw 0000000 ..... ..... 101 ..... 0011011 @sh5
> +sraiw 0100000 ..... ..... 101 ..... 0011011 @sh5
> +addw 0000000 ..... ..... 000 ..... 0111011 @r
> +subw 0100000 ..... ..... 000 ..... 0111011 @r
> +sllw 0000000 ..... ..... 001 ..... 0111011 @r
> +srlw 0000000 ..... ..... 101 ..... 0111011 @r
> +sraw 0100000 ..... ..... 101 ..... 0111011 @r
> diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
> index 39a20a70e8..01f751650a 100644
> --- a/target/riscv/insn_trans/trans_rvi.inc.c
> +++ b/target/riscv/insn_trans/trans_rvi.inc.c
> @@ -150,3 +150,171 @@ static bool trans_sd(DisasContext *ctx, arg_sd *a)
> return true;
> }
> #endif
> +
> +static bool trans_addi(DisasContext *ctx, arg_addi *a)
> +{
> + gen_arith_imm(ctx, OPC_RISC_ADDI, a->rd, a->rs1, a->imm);
> + return true;
> +}
> +
> +static bool trans_slti(DisasContext *ctx, arg_slti *a)
> +{
> + gen_arith_imm(ctx, OPC_RISC_SLTI, a->rd, a->rs1, a->imm);
> + return true;
> +}
> +
> +static bool trans_sltiu(DisasContext *ctx, arg_sltiu *a)
> +{
> + gen_arith_imm(ctx, OPC_RISC_SLTIU, a->rd, a->rs1, a->imm);
> + return true;
> +}
> +
> +static bool trans_xori(DisasContext *ctx, arg_xori *a)
> +{
> + gen_arith_imm(ctx, OPC_RISC_XORI, a->rd, a->rs1, a->imm);
> + return true;
> +}
> +static bool trans_ori(DisasContext *ctx, arg_ori *a)
> +{
> + gen_arith_imm(ctx, OPC_RISC_ORI, a->rd, a->rs1, a->imm);
> + return true;
> +}
> +static bool trans_andi(DisasContext *ctx, arg_andi *a)
> +{
> + gen_arith_imm(ctx, OPC_RISC_ANDI, a->rd, a->rs1, a->imm);
> + return true;
> +}
> +static bool trans_slli(DisasContext *ctx, arg_slli *a)
> +{
> + gen_arith_imm(ctx, OPC_RISC_SLLI, a->rd, a->rs1, a->shamt);
> + return true;
> +}
> +
> +static bool trans_srli(DisasContext *ctx, arg_srli *a)
> +{
> + gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_I, a->rd, a->rs1, a->shamt);
> + return true;
> +}
> +
> +static bool trans_srai(DisasContext *ctx, arg_srai *a)
> +{
> + gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_I, a->rd, a->rs1, a->shamt | 0x400);
> + return true;
> +}
> +
> +static bool trans_add(DisasContext *ctx, arg_add *a)
> +{
> + gen_arith(ctx, OPC_RISC_ADD, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_sub(DisasContext *ctx, arg_sub *a)
> +{
> + gen_arith(ctx, OPC_RISC_SUB, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_sll(DisasContext *ctx, arg_sll *a)
> +{
> + gen_arith(ctx, OPC_RISC_SLL, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_slt(DisasContext *ctx, arg_slt *a)
> +{
> + gen_arith(ctx, OPC_RISC_SLT, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_sltu(DisasContext *ctx, arg_sltu *a)
> +{
> + gen_arith(ctx, OPC_RISC_SLTU, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_xor(DisasContext *ctx, arg_xor *a)
> +{
> + gen_arith(ctx, OPC_RISC_XOR, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_srl(DisasContext *ctx, arg_srl *a)
> +{
> + gen_arith(ctx, OPC_RISC_SRL, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_sra(DisasContext *ctx, arg_sra *a)
> +{
> + gen_arith(ctx, OPC_RISC_SRA, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_or(DisasContext *ctx, arg_or *a)
> +{
> + gen_arith(ctx, OPC_RISC_OR, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_and(DisasContext *ctx, arg_and *a)
> +{
> + gen_arith(ctx, OPC_RISC_AND, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +#ifdef TARGET_RISCV64
> +static bool trans_addiw(DisasContext *ctx, arg_addiw *a)
> +{
> + gen_arith_imm(ctx, OPC_RISC_ADDIW, a->rd, a->rs1, a->imm);
> + return true;
> +}
> +
> +static bool trans_slliw(DisasContext *ctx, arg_slliw *a)
> +{
> + gen_arith_imm(ctx, OPC_RISC_SLLIW, a->rd, a->rs1, a->shamt);
> + return true;
> +}
> +
> +static bool trans_srliw(DisasContext *ctx, arg_srliw *a)
> +{
> + gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_IW, a->rd, a->rs1, a->shamt);
> + return true;
> +}
> +
> +static bool trans_sraiw(DisasContext *ctx, arg_sraiw *a)
> +{
> + gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_IW , a->rd, a->rs1,
> + a->shamt | 0x400);
> + return true;
> +}
> +
> +static bool trans_addw(DisasContext *ctx, arg_addw *a)
> +{
> + gen_arith(ctx, OPC_RISC_ADDW, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_subw(DisasContext *ctx, arg_subw *a)
> +{
> + gen_arith(ctx, OPC_RISC_SUBW, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_sllw(DisasContext *ctx, arg_sllw *a)
> +{
> + gen_arith(ctx, OPC_RISC_SLLW, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_srlw(DisasContext *ctx, arg_srlw *a)
> +{
> + gen_arith(ctx, OPC_RISC_SRLW, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_sraw(DisasContext *ctx, arg_sraw *a)
> +{
> + gen_arith(ctx, OPC_RISC_SRAW, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +#endif
> diff --git a/target/riscv/translate.c b/target/riscv/translate.c
> index 244855c82d..855c241e97 100644
> --- a/target/riscv/translate.c
> +++ b/target/riscv/translate.c
> @@ -1701,15 +1701,6 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
> imm = GET_IMM(ctx->opcode);
>
> switch (op) {
> - case OPC_RISC_ARITH_IMM:
> -#if defined(TARGET_RISCV64)
> - case OPC_RISC_ARITH_IMM_W:
> -#endif
> - if (rd == 0) {
> - break; /* NOP */
> - }
> - gen_arith_imm(ctx, MASK_OP_ARITH_IMM(ctx->opcode), rd, rs1, imm);
> - break;
> case OPC_RISC_ARITH:
> #if defined(TARGET_RISCV64)
> case OPC_RISC_ARITH_W:
>
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 07/35] target/riscv: Convert RVXI fence insns to decodetree
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 20:30 ` Alistair
-1 siblings, 0 replies; 140+ messages in thread
From: Alistair @ 2018-10-31 20:30 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: qemu-riscv, peer.adelt, richard.henderson, qemu-devel
On 10/31/18 6:20 AM, Bastian Koppelmann wrote:
> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> v2 -> v3:
> - removed %pred/%succ
> - dropped insn argument of trans_foo functions
>
> target/riscv/insn32.decode | 2 ++
> target/riscv/insn_trans/trans_rvi.inc.c | 21 +++++++++++++++++++++
> target/riscv/translate.c | 14 --------------
> 3 files changed, 23 insertions(+), 14 deletions(-)
>
> diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
> index 4fd88f48d2..6d750b4c5a 100644
> --- a/target/riscv/insn32.decode
> +++ b/target/riscv/insn32.decode
> @@ -82,3 +82,5 @@ srl 0000000 ..... ..... 101 ..... 0110011 @r
> sra 0100000 ..... ..... 101 ..... 0110011 @r
> or 0000000 ..... ..... 110 ..... 0110011 @r
> and 0000000 ..... ..... 111 ..... 0110011 @r
> +fence ---- pred:4 succ:4 ----- 000 ----- 0001111
> +fence_i ---- ---- ---- ----- 001 ----- 0001111
> diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
> index 01f751650a..a149e913b1 100644
> --- a/target/riscv/insn_trans/trans_rvi.inc.c
> +++ b/target/riscv/insn_trans/trans_rvi.inc.c
> @@ -318,3 +318,24 @@ static bool trans_sraw(DisasContext *ctx, arg_sraw *a)
> return true;
> }
> #endif
> +
> +static bool trans_fence(DisasContext *ctx, arg_fence *a)
> +{
> +#ifndef CONFIG_USER_ONLY
> + /* FENCE is a full memory barrier. */
> + tcg_gen_mb(TCG_MO_ALL | TCG_BAR_SC);
> +#endif
> + return true;
> +}
> +
> +static bool trans_fence_i(DisasContext *ctx, arg_fence_i *a)
> +{
> +#ifndef CONFIG_USER_ONLY
> + /* FENCE_I is a no-op in QEMU,
> + * however we need to end the translation block */
> + tcg_gen_movi_tl(cpu_pc, ctx->pc_succ_insn);
> + tcg_gen_exit_tb(NULL, 0);
> + ctx->base.is_jmp = DISAS_NORETURN;
> +#endif
> + return true;
> +}
> diff --git a/target/riscv/translate.c b/target/riscv/translate.c
> index 855c241e97..80f18fb6aa 100644
> --- a/target/riscv/translate.c
> +++ b/target/riscv/translate.c
> @@ -1740,20 +1740,6 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
> gen_fp_arith(ctx, MASK_OP_FP_ARITH(ctx->opcode), rd, rs1, rs2,
> GET_RM(ctx->opcode));
> break;
> - case OPC_RISC_FENCE:
> -#ifndef CONFIG_USER_ONLY
> - if (ctx->opcode & 0x1000) {
> - /* FENCE_I is a no-op in QEMU,
> - * however we need to end the translation block */
> - tcg_gen_movi_tl(cpu_pc, ctx->pc_succ_insn);
> - tcg_gen_exit_tb(NULL, 0);
> - ctx->base.is_jmp = DISAS_NORETURN;
> - } else {
> - /* FENCE is a full memory barrier. */
> - tcg_gen_mb(TCG_MO_ALL | TCG_BAR_SC);
> - }
> -#endif
> - break;
> case OPC_RISC_SYSTEM:
> gen_system(env, ctx, MASK_OP_SYSTEM(ctx->opcode), rd, rs1,
> (ctx->opcode & 0xFFF00000) >> 20);
>
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [Qemu-devel] [PATCH v3 07/35] target/riscv: Convert RVXI fence insns to decodetree
@ 2018-10-31 20:30 ` Alistair
0 siblings, 0 replies; 140+ messages in thread
From: Alistair @ 2018-10-31 20:30 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: qemu-riscv, peer.adelt, richard.henderson, qemu-devel
On 10/31/18 6:20 AM, Bastian Koppelmann wrote:
> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> v2 -> v3:
> - removed %pred/%succ
> - dropped insn argument of trans_foo functions
>
> target/riscv/insn32.decode | 2 ++
> target/riscv/insn_trans/trans_rvi.inc.c | 21 +++++++++++++++++++++
> target/riscv/translate.c | 14 --------------
> 3 files changed, 23 insertions(+), 14 deletions(-)
>
> diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
> index 4fd88f48d2..6d750b4c5a 100644
> --- a/target/riscv/insn32.decode
> +++ b/target/riscv/insn32.decode
> @@ -82,3 +82,5 @@ srl 0000000 ..... ..... 101 ..... 0110011 @r
> sra 0100000 ..... ..... 101 ..... 0110011 @r
> or 0000000 ..... ..... 110 ..... 0110011 @r
> and 0000000 ..... ..... 111 ..... 0110011 @r
> +fence ---- pred:4 succ:4 ----- 000 ----- 0001111
> +fence_i ---- ---- ---- ----- 001 ----- 0001111
> diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
> index 01f751650a..a149e913b1 100644
> --- a/target/riscv/insn_trans/trans_rvi.inc.c
> +++ b/target/riscv/insn_trans/trans_rvi.inc.c
> @@ -318,3 +318,24 @@ static bool trans_sraw(DisasContext *ctx, arg_sraw *a)
> return true;
> }
> #endif
> +
> +static bool trans_fence(DisasContext *ctx, arg_fence *a)
> +{
> +#ifndef CONFIG_USER_ONLY
> + /* FENCE is a full memory barrier. */
> + tcg_gen_mb(TCG_MO_ALL | TCG_BAR_SC);
> +#endif
> + return true;
> +}
> +
> +static bool trans_fence_i(DisasContext *ctx, arg_fence_i *a)
> +{
> +#ifndef CONFIG_USER_ONLY
> + /* FENCE_I is a no-op in QEMU,
> + * however we need to end the translation block */
> + tcg_gen_movi_tl(cpu_pc, ctx->pc_succ_insn);
> + tcg_gen_exit_tb(NULL, 0);
> + ctx->base.is_jmp = DISAS_NORETURN;
> +#endif
> + return true;
> +}
> diff --git a/target/riscv/translate.c b/target/riscv/translate.c
> index 855c241e97..80f18fb6aa 100644
> --- a/target/riscv/translate.c
> +++ b/target/riscv/translate.c
> @@ -1740,20 +1740,6 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
> gen_fp_arith(ctx, MASK_OP_FP_ARITH(ctx->opcode), rd, rs1, rs2,
> GET_RM(ctx->opcode));
> break;
> - case OPC_RISC_FENCE:
> -#ifndef CONFIG_USER_ONLY
> - if (ctx->opcode & 0x1000) {
> - /* FENCE_I is a no-op in QEMU,
> - * however we need to end the translation block */
> - tcg_gen_movi_tl(cpu_pc, ctx->pc_succ_insn);
> - tcg_gen_exit_tb(NULL, 0);
> - ctx->base.is_jmp = DISAS_NORETURN;
> - } else {
> - /* FENCE is a full memory barrier. */
> - tcg_gen_mb(TCG_MO_ALL | TCG_BAR_SC);
> - }
> -#endif
> - break;
> case OPC_RISC_SYSTEM:
> gen_system(env, ctx, MASK_OP_SYSTEM(ctx->opcode), rd, rs1,
> (ctx->opcode & 0xFFF00000) >> 20);
>
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 09/35] target/riscv: Convert RVXM insns to decodetree
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 20:38 ` Alistair
-1 siblings, 0 replies; 140+ messages in thread
From: Alistair @ 2018-10-31 20:38 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: qemu-riscv, peer.adelt, richard.henderson, qemu-devel
On 10/31/18 6:20 AM, Bastian Koppelmann wrote:
> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> v2 -> v3:
> - moved 64-bit only insn to insn64.decode
> - dropped insn argument of trans_foo functions
>
> target/riscv/insn32.decode | 10 +++
> target/riscv/insn64.decode | 7 ++
> target/riscv/insn_trans/trans_rvm.inc.c | 100 ++++++++++++++++++++++++
> target/riscv/translate.c | 10 +--
> 4 files changed, 118 insertions(+), 9 deletions(-)
> create mode 100644 target/riscv/insn_trans/trans_rvm.inc.c
>
> diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
> index b67d3a9437..b37fece633 100644
> --- a/target/riscv/insn32.decode
> +++ b/target/riscv/insn32.decode
> @@ -92,3 +92,13 @@ csrrc ............ ..... 011 ..... 1110011 @csr
> csrrwi ............ ..... 101 ..... 1110011 @csr
> csrrsi ............ ..... 110 ..... 1110011 @csr
> csrrci ............ ..... 111 ..... 1110011 @csr
> +
> +# *** RV32M Standard Extension ***
> +mul 0000001 ..... ..... 000 ..... 0110011 @r
> +mulh 0000001 ..... ..... 001 ..... 0110011 @r
> +mulhsu 0000001 ..... ..... 010 ..... 0110011 @r
> +mulhu 0000001 ..... ..... 011 ..... 0110011 @r
> +div 0000001 ..... ..... 100 ..... 0110011 @r
> +divu 0000001 ..... ..... 101 ..... 0110011 @r
> +rem 0000001 ..... ..... 110 ..... 0110011 @r
> +remu 0000001 ..... ..... 111 ..... 0110011 @r
> diff --git a/target/riscv/insn64.decode b/target/riscv/insn64.decode
> index 9a35f2aa19..008f100546 100644
> --- a/target/riscv/insn64.decode
> +++ b/target/riscv/insn64.decode
> @@ -36,3 +36,10 @@ subw 0100000 ..... ..... 000 ..... 0111011 @r
> sllw 0000000 ..... ..... 001 ..... 0111011 @r
> srlw 0000000 ..... ..... 101 ..... 0111011 @r
> sraw 0100000 ..... ..... 101 ..... 0111011 @r
> +
> +# *** RV64M Standard Extension (in addition to RV32M) ***
> +mulw 0000001 ..... ..... 000 ..... 0111011 @r
> +divw 0000001 ..... ..... 100 ..... 0111011 @r
> +divuw 0000001 ..... ..... 101 ..... 0111011 @r
> +remw 0000001 ..... ..... 110 ..... 0111011 @r
> +remuw 0000001 ..... ..... 111 ..... 0111011 @r
> diff --git a/target/riscv/insn_trans/trans_rvm.inc.c b/target/riscv/insn_trans/trans_rvm.inc.c
> new file mode 100644
> index 0000000000..ec3197ede8
> --- /dev/null
> +++ b/target/riscv/insn_trans/trans_rvm.inc.c
> @@ -0,0 +1,100 @@
> +/*
> + * RISC-V translation routines for the RV64M Standard Extension.
> + *
> + * Copyright (c) 2016-2017 Sagar Karandikar, sagark@eecs.berkeley.edu
> + * Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
> + * Bastian Koppelmann, kbastian@mail.uni-paderborn.de
> + *
> + * This program is free software; you can redistribute it and/or modify it
> + * under the terms and conditions of the GNU General Public License,
> + * version 2 or later, as published by the Free Software Foundation.
> + *
> + * This program is distributed in the hope it will be useful, but WITHOUT
> + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
> + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
> + * more details.
> + *
> + * You should have received a copy of the GNU General Public License along with
> + * this program. If not, see <http://www.gnu.org/licenses/>.
> + */
> +
> +
> +static bool trans_mul(DisasContext *ctx, arg_mul *a)
> +{
> + gen_arith(ctx, OPC_RISC_MUL, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_mulh(DisasContext *ctx, arg_mulh *a)
> +{
> + gen_arith(ctx, OPC_RISC_MULH, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_mulhsu(DisasContext *ctx, arg_mulhsu *a)
> +{
> + gen_arith(ctx, OPC_RISC_MULHSU, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_mulhu(DisasContext *ctx, arg_mulhu *a)
> +{
> + gen_arith(ctx, OPC_RISC_MULHU, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_div(DisasContext *ctx, arg_div *a)
> +{
> + gen_arith(ctx, OPC_RISC_DIV, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_divu(DisasContext *ctx, arg_divu *a)
> +{
> + gen_arith(ctx, OPC_RISC_DIVU, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_rem(DisasContext *ctx, arg_rem *a)
> +{
> + gen_arith(ctx, OPC_RISC_REM, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_remu(DisasContext *ctx, arg_remu *a)
> +{
> + gen_arith(ctx, OPC_RISC_REMU, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +#ifdef TARGET_RISCV64
> +static bool trans_mulw(DisasContext *ctx, arg_mulw *a)
> +{
> + gen_arith(ctx, OPC_RISC_MULW, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_divw(DisasContext *ctx, arg_divw *a)
> +{
> + gen_arith(ctx, OPC_RISC_DIVW, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_divuw(DisasContext *ctx, arg_divuw *a)
> +{
> + gen_arith(ctx, OPC_RISC_DIVUW, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_remw(DisasContext *ctx, arg_remw *a)
> +{
> + gen_arith(ctx, OPC_RISC_REMW, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_remuw(DisasContext *ctx, arg_remuw *a)
> +{
> + gen_arith(ctx, OPC_RISC_REMUW, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +#endif
> diff --git a/target/riscv/translate.c b/target/riscv/translate.c
> index be03acd066..8c876187b1 100644
> --- a/target/riscv/translate.c
> +++ b/target/riscv/translate.c
> @@ -1639,6 +1639,7 @@ bool decode_insn32(DisasContext *ctx, uint32_t insn);
> #include "decode_insn32.inc.c"
> /* Include insn module translation function */
> #include "insn_trans/trans_rvi.inc.c"
> +#include "insn_trans/trans_rvm.inc.c"
>
> static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
> {
> @@ -1660,15 +1661,6 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
> imm = GET_IMM(ctx->opcode);
>
> switch (op) {
> - case OPC_RISC_ARITH:
> -#if defined(TARGET_RISCV64)
> - case OPC_RISC_ARITH_W:
> -#endif
> - if (rd == 0) {
> - break; /* NOP */
> - }
> - gen_arith(ctx, MASK_OP_ARITH(ctx->opcode), rd, rs1, rs2);
> - break;
> case OPC_RISC_FP_LOAD:
> gen_fp_load(ctx, MASK_OP_FP_LOAD(ctx->opcode), rd, rs1, imm);
> break;
>
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [Qemu-devel] [PATCH v3 09/35] target/riscv: Convert RVXM insns to decodetree
@ 2018-10-31 20:38 ` Alistair
0 siblings, 0 replies; 140+ messages in thread
From: Alistair @ 2018-10-31 20:38 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: qemu-riscv, peer.adelt, richard.henderson, qemu-devel
On 10/31/18 6:20 AM, Bastian Koppelmann wrote:
> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> v2 -> v3:
> - moved 64-bit only insn to insn64.decode
> - dropped insn argument of trans_foo functions
>
> target/riscv/insn32.decode | 10 +++
> target/riscv/insn64.decode | 7 ++
> target/riscv/insn_trans/trans_rvm.inc.c | 100 ++++++++++++++++++++++++
> target/riscv/translate.c | 10 +--
> 4 files changed, 118 insertions(+), 9 deletions(-)
> create mode 100644 target/riscv/insn_trans/trans_rvm.inc.c
>
> diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
> index b67d3a9437..b37fece633 100644
> --- a/target/riscv/insn32.decode
> +++ b/target/riscv/insn32.decode
> @@ -92,3 +92,13 @@ csrrc ............ ..... 011 ..... 1110011 @csr
> csrrwi ............ ..... 101 ..... 1110011 @csr
> csrrsi ............ ..... 110 ..... 1110011 @csr
> csrrci ............ ..... 111 ..... 1110011 @csr
> +
> +# *** RV32M Standard Extension ***
> +mul 0000001 ..... ..... 000 ..... 0110011 @r
> +mulh 0000001 ..... ..... 001 ..... 0110011 @r
> +mulhsu 0000001 ..... ..... 010 ..... 0110011 @r
> +mulhu 0000001 ..... ..... 011 ..... 0110011 @r
> +div 0000001 ..... ..... 100 ..... 0110011 @r
> +divu 0000001 ..... ..... 101 ..... 0110011 @r
> +rem 0000001 ..... ..... 110 ..... 0110011 @r
> +remu 0000001 ..... ..... 111 ..... 0110011 @r
> diff --git a/target/riscv/insn64.decode b/target/riscv/insn64.decode
> index 9a35f2aa19..008f100546 100644
> --- a/target/riscv/insn64.decode
> +++ b/target/riscv/insn64.decode
> @@ -36,3 +36,10 @@ subw 0100000 ..... ..... 000 ..... 0111011 @r
> sllw 0000000 ..... ..... 001 ..... 0111011 @r
> srlw 0000000 ..... ..... 101 ..... 0111011 @r
> sraw 0100000 ..... ..... 101 ..... 0111011 @r
> +
> +# *** RV64M Standard Extension (in addition to RV32M) ***
> +mulw 0000001 ..... ..... 000 ..... 0111011 @r
> +divw 0000001 ..... ..... 100 ..... 0111011 @r
> +divuw 0000001 ..... ..... 101 ..... 0111011 @r
> +remw 0000001 ..... ..... 110 ..... 0111011 @r
> +remuw 0000001 ..... ..... 111 ..... 0111011 @r
> diff --git a/target/riscv/insn_trans/trans_rvm.inc.c b/target/riscv/insn_trans/trans_rvm.inc.c
> new file mode 100644
> index 0000000000..ec3197ede8
> --- /dev/null
> +++ b/target/riscv/insn_trans/trans_rvm.inc.c
> @@ -0,0 +1,100 @@
> +/*
> + * RISC-V translation routines for the RV64M Standard Extension.
> + *
> + * Copyright (c) 2016-2017 Sagar Karandikar, sagark@eecs.berkeley.edu
> + * Copyright (c) 2018 Peer Adelt, peer.adelt@hni.uni-paderborn.de
> + * Bastian Koppelmann, kbastian@mail.uni-paderborn.de
> + *
> + * This program is free software; you can redistribute it and/or modify it
> + * under the terms and conditions of the GNU General Public License,
> + * version 2 or later, as published by the Free Software Foundation.
> + *
> + * This program is distributed in the hope it will be useful, but WITHOUT
> + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
> + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
> + * more details.
> + *
> + * You should have received a copy of the GNU General Public License along with
> + * this program. If not, see <http://www.gnu.org/licenses/>.
> + */
> +
> +
> +static bool trans_mul(DisasContext *ctx, arg_mul *a)
> +{
> + gen_arith(ctx, OPC_RISC_MUL, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_mulh(DisasContext *ctx, arg_mulh *a)
> +{
> + gen_arith(ctx, OPC_RISC_MULH, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_mulhsu(DisasContext *ctx, arg_mulhsu *a)
> +{
> + gen_arith(ctx, OPC_RISC_MULHSU, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_mulhu(DisasContext *ctx, arg_mulhu *a)
> +{
> + gen_arith(ctx, OPC_RISC_MULHU, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_div(DisasContext *ctx, arg_div *a)
> +{
> + gen_arith(ctx, OPC_RISC_DIV, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_divu(DisasContext *ctx, arg_divu *a)
> +{
> + gen_arith(ctx, OPC_RISC_DIVU, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_rem(DisasContext *ctx, arg_rem *a)
> +{
> + gen_arith(ctx, OPC_RISC_REM, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_remu(DisasContext *ctx, arg_remu *a)
> +{
> + gen_arith(ctx, OPC_RISC_REMU, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +#ifdef TARGET_RISCV64
> +static bool trans_mulw(DisasContext *ctx, arg_mulw *a)
> +{
> + gen_arith(ctx, OPC_RISC_MULW, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_divw(DisasContext *ctx, arg_divw *a)
> +{
> + gen_arith(ctx, OPC_RISC_DIVW, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_divuw(DisasContext *ctx, arg_divuw *a)
> +{
> + gen_arith(ctx, OPC_RISC_DIVUW, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_remw(DisasContext *ctx, arg_remw *a)
> +{
> + gen_arith(ctx, OPC_RISC_REMW, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +
> +static bool trans_remuw(DisasContext *ctx, arg_remuw *a)
> +{
> + gen_arith(ctx, OPC_RISC_REMUW, a->rd, a->rs1, a->rs2);
> + return true;
> +}
> +#endif
> diff --git a/target/riscv/translate.c b/target/riscv/translate.c
> index be03acd066..8c876187b1 100644
> --- a/target/riscv/translate.c
> +++ b/target/riscv/translate.c
> @@ -1639,6 +1639,7 @@ bool decode_insn32(DisasContext *ctx, uint32_t insn);
> #include "decode_insn32.inc.c"
> /* Include insn module translation function */
> #include "insn_trans/trans_rvi.inc.c"
> +#include "insn_trans/trans_rvm.inc.c"
>
> static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
> {
> @@ -1660,15 +1661,6 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
> imm = GET_IMM(ctx->opcode);
>
> switch (op) {
> - case OPC_RISC_ARITH:
> -#if defined(TARGET_RISCV64)
> - case OPC_RISC_ARITH_W:
> -#endif
> - if (rd == 0) {
> - break; /* NOP */
> - }
> - gen_arith(ctx, MASK_OP_ARITH(ctx->opcode), rd, rs1, rs2);
> - break;
> case OPC_RISC_FP_LOAD:
> gen_fp_load(ctx, MASK_OP_FP_LOAD(ctx->opcode), rd, rs1, imm);
> break;
>
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 25/35] target/riscv: make ADD/SUB/OR/XOR/AND insn use arg lists
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 20:44 ` Alistair Francis
-1 siblings, 0 replies; 140+ messages in thread
From: Alistair Francis @ 2018-10-31 20:44 UTC (permalink / raw)
To: Bastian Koppelmann
Cc: Michael Clark, Sagar Karandikar, Palmer Dabbelt,
Alistair Francis, qemu-riscv, peer.adelt, Richard Henderson,
qemu-devel@nongnu.org Developers
On Wed, Oct 31, 2018 at 6:29 AM Bastian Koppelmann
<kbastian@mail.uni-paderborn.de> wrote:
>
> manual decoding in gen_arith() is not necessary with decodetree. For now
> the function is called trans_arith as the original gen_arith still
> exists. The former will be renamed to gen_arith as soon as the old
> gen_arith can be removed.
This series causes a lot of kernel oops during boot (see below for an
example) and I bisected it to this patch.
[ 3.292000] Unable to handle kernel paging request at virtual
address ffffffa078a9401d
[ 3.292000] Oops [#2]
[ 3.292000] Modules linked in:
[ 3.292000] CPU: 0 PID: 64 Comm: systemd-journal Tainted: G D
4.18.0-riscv #1
[ 3.292000] sepc: ffffffe00057e266 ra : ffffffe000136094 sp :
ffffffe078aa9d10
[ 3.292000] gp : ffffffe00078eb98 tp : ffffffe078a81180 t0 :
00000000000001f8
[ 3.292000] t1 : 0000000000000022 t2 : 0000000000000000 s0 :
ffffffe078aa9d20
[ 3.292000] s1 : ffffffe078a81180 a0 : ffffffa078a9401d a1 :
0000004000000136
[ 3.292000] a2 : 0000000000000001 a3 : 0000000000000000 a4 :
ffffffc00000001d
[ 3.292000] a5 : ffffffe078a94000 a6 : ffffffe00000b508 a7 :
0000003fff89ce87
[ 3.292000] s2 : 0000000000000153 s3 : ffffffe078aa9e98 s4 :
0000002aaaac44e0
[ 3.292000] s5 : 0000000000000000 s6 : 0000000000000000 s7 :
0000000000000153
[ 3.292000] s8 : 0000000000000153 s9 : 0000003fff89ce87 s10:
0000003fff89cea5
[ 3.292000] s11: ffffffe078a18d00 t3 : 000000003e264cd7 t4 :
0000000000000078
[ 3.292000] t5 : 000000000000006d t6 : ffffffe078a94153
[ 3.292000] sstatus: 0000000000000120 sbadaddr: ffffffa078a9401d
scause: 000000000000000d
[ 3.296000] ---[ end trace 3cae92dc85ff86eb ]---
[FAILED] Failed to start Journal Service.
Alistair
>
> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
> ---
> v2 -> v3:
> - &arith -> &r
>
> target/riscv/insn32.decode | 3 ++-
> target/riscv/insn_trans/trans_rvi.inc.c | 21 ++++++----------
> target/riscv/translate.c | 33 ++++++++++++++-----------
> 3 files changed, 27 insertions(+), 30 deletions(-)
>
> diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
> index 70e00412b6..520954e9c6 100644
> --- a/target/riscv/insn32.decode
> +++ b/target/riscv/insn32.decode
> @@ -36,11 +36,12 @@
> # Argument sets:
> &b imm rs2 rs1
> &i imm rs1 rd
> +&r rd rs1 rs2
> &shift shamt rs1 rd
> &atomic aq rl rs2 rs1 rd
>
> # Formats 32:
> -@r ....... ..... ..... ... ..... ....... %rs2 %rs1 %rd
> +@r ....... ..... ..... ... ..... ....... &r %rs2 %rs1 %rd
> @i ............ ..... ... ..... ....... &i imm=%imm_i %rs1 %rd
> @b ....... ..... ..... ... ..... ....... &b imm=%imm_b %rs2 %rs1
> @s ....... ..... ..... ... ..... ....... imm=%imm_s %rs2 %rs1
> diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
> index 8bd2752f31..5b5999954a 100644
> --- a/target/riscv/insn_trans/trans_rvi.inc.c
> +++ b/target/riscv/insn_trans/trans_rvi.inc.c
> @@ -301,14 +301,12 @@ static bool trans_srai(DisasContext *ctx, arg_srai *a)
>
> static bool trans_add(DisasContext *ctx, arg_add *a)
> {
> - gen_arith(ctx, OPC_RISC_ADD, a->rd, a->rs1, a->rs2);
> - return true;
> + return trans_arith(ctx, a, &tcg_gen_add_tl);
> }
>
> static bool trans_sub(DisasContext *ctx, arg_sub *a)
> {
> - gen_arith(ctx, OPC_RISC_SUB, a->rd, a->rs1, a->rs2);
> - return true;
> + return trans_arith(ctx, a, &tcg_gen_sub_tl);
> }
>
> static bool trans_sll(DisasContext *ctx, arg_sll *a)
> @@ -331,8 +329,7 @@ static bool trans_sltu(DisasContext *ctx, arg_sltu *a)
>
> static bool trans_xor(DisasContext *ctx, arg_xor *a)
> {
> - gen_arith(ctx, OPC_RISC_XOR, a->rd, a->rs1, a->rs2);
> - return true;
> + return trans_arith(ctx, a, &tcg_gen_xor_tl);
> }
>
> static bool trans_srl(DisasContext *ctx, arg_srl *a)
> @@ -349,14 +346,12 @@ static bool trans_sra(DisasContext *ctx, arg_sra *a)
>
> static bool trans_or(DisasContext *ctx, arg_or *a)
> {
> - gen_arith(ctx, OPC_RISC_OR, a->rd, a->rs1, a->rs2);
> - return true;
> + return trans_arith(ctx, a, &tcg_gen_or_tl);
> }
>
> static bool trans_and(DisasContext *ctx, arg_and *a)
> {
> - gen_arith(ctx, OPC_RISC_AND, a->rd, a->rs1, a->rs2);
> - return true;
> + return trans_arith(ctx, a, &tcg_gen_and_tl);
> }
>
> #ifdef TARGET_RISCV64
> @@ -405,14 +400,12 @@ static bool trans_sraiw(DisasContext *ctx, arg_sraiw *a)
>
> static bool trans_addw(DisasContext *ctx, arg_addw *a)
> {
> - gen_arith(ctx, OPC_RISC_ADDW, a->rd, a->rs1, a->rs2);
> - return true;
> + return trans_arith(ctx, a, &tcg_gen_add_tl);
> }
>
> static bool trans_subw(DisasContext *ctx, arg_subw *a)
> {
> - gen_arith(ctx, OPC_RISC_SUBW, a->rd, a->rs1, a->rs2);
> - return true;
> + return trans_arith(ctx, a, &tcg_gen_sub_tl);
> }
>
> static bool trans_sllw(DisasContext *ctx, arg_sllw *a)
> diff --git a/target/riscv/translate.c b/target/riscv/translate.c
> index de60686df4..a781aba08c 100644
> --- a/target/riscv/translate.c
> +++ b/target/riscv/translate.c
> @@ -192,12 +192,6 @@ static void gen_arith(DisasContext *ctx, uint32_t opc, int rd, int rs1,
> gen_get_gpr(source2, rs2);
>
> switch (opc) {
> - CASE_OP_32_64(OPC_RISC_ADD):
> - tcg_gen_add_tl(source1, source1, source2);
> - break;
> - CASE_OP_32_64(OPC_RISC_SUB):
> - tcg_gen_sub_tl(source1, source1, source2);
> - break;
> #if defined(TARGET_RISCV64)
> case OPC_RISC_SLLW:
> tcg_gen_andi_tl(source2, source2, 0x1F);
> @@ -214,9 +208,6 @@ static void gen_arith(DisasContext *ctx, uint32_t opc, int rd, int rs1,
> case OPC_RISC_SLTU:
> tcg_gen_setcond_tl(TCG_COND_LTU, source1, source1, source2);
> break;
> - case OPC_RISC_XOR:
> - tcg_gen_xor_tl(source1, source1, source2);
> - break;
> #if defined(TARGET_RISCV64)
> case OPC_RISC_SRLW:
> /* clear upper 32 */
> @@ -242,12 +233,6 @@ static void gen_arith(DisasContext *ctx, uint32_t opc, int rd, int rs1,
> tcg_gen_andi_tl(source2, source2, TARGET_LONG_BITS - 1);
> tcg_gen_sar_tl(source1, source1, source2);
> break;
> - case OPC_RISC_OR:
> - tcg_gen_or_tl(source1, source1, source2);
> - break;
> - case OPC_RISC_AND:
> - tcg_gen_and_tl(source1, source1, source2);
> - break;
> CASE_OP_32_64(OPC_RISC_MUL):
> tcg_gen_mul_tl(source1, source1, source2);
> break;
> @@ -644,6 +629,24 @@ static void gen_addiw(TCGv ret, TCGv arg1, TCGv arg2)
> }
> #endif
>
> +static bool trans_arith(DisasContext *ctx, arg_r *a,
> + void(*func)(TCGv, TCGv, TCGv))
> +{
> + TCGv source1, source2;
> + source1 = tcg_temp_new();
> + source2 = tcg_temp_new();
> +
> + gen_get_gpr(source1, a->rs1);
> + gen_get_gpr(source2, a->rs2);
> +
> + (*func)(source1, source1, source2);
> +
> + gen_set_gpr(a->rd, source1);
> + tcg_temp_free(source1);
> + tcg_temp_free(source2);
> + return true;
> +}
> +
> /* Include insn module translation function */
> #include "insn_trans/trans_rvi.inc.c"
> #include "insn_trans/trans_rvm.inc.c"
> --
> 2.19.1
>
>
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [Qemu-devel] [PATCH v3 25/35] target/riscv: make ADD/SUB/OR/XOR/AND insn use arg lists
@ 2018-10-31 20:44 ` Alistair Francis
0 siblings, 0 replies; 140+ messages in thread
From: Alistair Francis @ 2018-10-31 20:44 UTC (permalink / raw)
To: Bastian Koppelmann
Cc: Michael Clark, Sagar Karandikar, Palmer Dabbelt,
Alistair Francis, qemu-riscv, peer.adelt, Richard Henderson,
qemu-devel@nongnu.org Developers
On Wed, Oct 31, 2018 at 6:29 AM Bastian Koppelmann
<kbastian@mail.uni-paderborn.de> wrote:
>
> manual decoding in gen_arith() is not necessary with decodetree. For now
> the function is called trans_arith as the original gen_arith still
> exists. The former will be renamed to gen_arith as soon as the old
> gen_arith can be removed.
This series causes a lot of kernel oops during boot (see below for an
example) and I bisected it to this patch.
[ 3.292000] Unable to handle kernel paging request at virtual
address ffffffa078a9401d
[ 3.292000] Oops [#2]
[ 3.292000] Modules linked in:
[ 3.292000] CPU: 0 PID: 64 Comm: systemd-journal Tainted: G D
4.18.0-riscv #1
[ 3.292000] sepc: ffffffe00057e266 ra : ffffffe000136094 sp :
ffffffe078aa9d10
[ 3.292000] gp : ffffffe00078eb98 tp : ffffffe078a81180 t0 :
00000000000001f8
[ 3.292000] t1 : 0000000000000022 t2 : 0000000000000000 s0 :
ffffffe078aa9d20
[ 3.292000] s1 : ffffffe078a81180 a0 : ffffffa078a9401d a1 :
0000004000000136
[ 3.292000] a2 : 0000000000000001 a3 : 0000000000000000 a4 :
ffffffc00000001d
[ 3.292000] a5 : ffffffe078a94000 a6 : ffffffe00000b508 a7 :
0000003fff89ce87
[ 3.292000] s2 : 0000000000000153 s3 : ffffffe078aa9e98 s4 :
0000002aaaac44e0
[ 3.292000] s5 : 0000000000000000 s6 : 0000000000000000 s7 :
0000000000000153
[ 3.292000] s8 : 0000000000000153 s9 : 0000003fff89ce87 s10:
0000003fff89cea5
[ 3.292000] s11: ffffffe078a18d00 t3 : 000000003e264cd7 t4 :
0000000000000078
[ 3.292000] t5 : 000000000000006d t6 : ffffffe078a94153
[ 3.292000] sstatus: 0000000000000120 sbadaddr: ffffffa078a9401d
scause: 000000000000000d
[ 3.296000] ---[ end trace 3cae92dc85ff86eb ]---
[FAILED] Failed to start Journal Service.
Alistair
>
> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
> ---
> v2 -> v3:
> - &arith -> &r
>
> target/riscv/insn32.decode | 3 ++-
> target/riscv/insn_trans/trans_rvi.inc.c | 21 ++++++----------
> target/riscv/translate.c | 33 ++++++++++++++-----------
> 3 files changed, 27 insertions(+), 30 deletions(-)
>
> diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
> index 70e00412b6..520954e9c6 100644
> --- a/target/riscv/insn32.decode
> +++ b/target/riscv/insn32.decode
> @@ -36,11 +36,12 @@
> # Argument sets:
> &b imm rs2 rs1
> &i imm rs1 rd
> +&r rd rs1 rs2
> &shift shamt rs1 rd
> &atomic aq rl rs2 rs1 rd
>
> # Formats 32:
> -@r ....... ..... ..... ... ..... ....... %rs2 %rs1 %rd
> +@r ....... ..... ..... ... ..... ....... &r %rs2 %rs1 %rd
> @i ............ ..... ... ..... ....... &i imm=%imm_i %rs1 %rd
> @b ....... ..... ..... ... ..... ....... &b imm=%imm_b %rs2 %rs1
> @s ....... ..... ..... ... ..... ....... imm=%imm_s %rs2 %rs1
> diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
> index 8bd2752f31..5b5999954a 100644
> --- a/target/riscv/insn_trans/trans_rvi.inc.c
> +++ b/target/riscv/insn_trans/trans_rvi.inc.c
> @@ -301,14 +301,12 @@ static bool trans_srai(DisasContext *ctx, arg_srai *a)
>
> static bool trans_add(DisasContext *ctx, arg_add *a)
> {
> - gen_arith(ctx, OPC_RISC_ADD, a->rd, a->rs1, a->rs2);
> - return true;
> + return trans_arith(ctx, a, &tcg_gen_add_tl);
> }
>
> static bool trans_sub(DisasContext *ctx, arg_sub *a)
> {
> - gen_arith(ctx, OPC_RISC_SUB, a->rd, a->rs1, a->rs2);
> - return true;
> + return trans_arith(ctx, a, &tcg_gen_sub_tl);
> }
>
> static bool trans_sll(DisasContext *ctx, arg_sll *a)
> @@ -331,8 +329,7 @@ static bool trans_sltu(DisasContext *ctx, arg_sltu *a)
>
> static bool trans_xor(DisasContext *ctx, arg_xor *a)
> {
> - gen_arith(ctx, OPC_RISC_XOR, a->rd, a->rs1, a->rs2);
> - return true;
> + return trans_arith(ctx, a, &tcg_gen_xor_tl);
> }
>
> static bool trans_srl(DisasContext *ctx, arg_srl *a)
> @@ -349,14 +346,12 @@ static bool trans_sra(DisasContext *ctx, arg_sra *a)
>
> static bool trans_or(DisasContext *ctx, arg_or *a)
> {
> - gen_arith(ctx, OPC_RISC_OR, a->rd, a->rs1, a->rs2);
> - return true;
> + return trans_arith(ctx, a, &tcg_gen_or_tl);
> }
>
> static bool trans_and(DisasContext *ctx, arg_and *a)
> {
> - gen_arith(ctx, OPC_RISC_AND, a->rd, a->rs1, a->rs2);
> - return true;
> + return trans_arith(ctx, a, &tcg_gen_and_tl);
> }
>
> #ifdef TARGET_RISCV64
> @@ -405,14 +400,12 @@ static bool trans_sraiw(DisasContext *ctx, arg_sraiw *a)
>
> static bool trans_addw(DisasContext *ctx, arg_addw *a)
> {
> - gen_arith(ctx, OPC_RISC_ADDW, a->rd, a->rs1, a->rs2);
> - return true;
> + return trans_arith(ctx, a, &tcg_gen_add_tl);
> }
>
> static bool trans_subw(DisasContext *ctx, arg_subw *a)
> {
> - gen_arith(ctx, OPC_RISC_SUBW, a->rd, a->rs1, a->rs2);
> - return true;
> + return trans_arith(ctx, a, &tcg_gen_sub_tl);
> }
>
> static bool trans_sllw(DisasContext *ctx, arg_sllw *a)
> diff --git a/target/riscv/translate.c b/target/riscv/translate.c
> index de60686df4..a781aba08c 100644
> --- a/target/riscv/translate.c
> +++ b/target/riscv/translate.c
> @@ -192,12 +192,6 @@ static void gen_arith(DisasContext *ctx, uint32_t opc, int rd, int rs1,
> gen_get_gpr(source2, rs2);
>
> switch (opc) {
> - CASE_OP_32_64(OPC_RISC_ADD):
> - tcg_gen_add_tl(source1, source1, source2);
> - break;
> - CASE_OP_32_64(OPC_RISC_SUB):
> - tcg_gen_sub_tl(source1, source1, source2);
> - break;
> #if defined(TARGET_RISCV64)
> case OPC_RISC_SLLW:
> tcg_gen_andi_tl(source2, source2, 0x1F);
> @@ -214,9 +208,6 @@ static void gen_arith(DisasContext *ctx, uint32_t opc, int rd, int rs1,
> case OPC_RISC_SLTU:
> tcg_gen_setcond_tl(TCG_COND_LTU, source1, source1, source2);
> break;
> - case OPC_RISC_XOR:
> - tcg_gen_xor_tl(source1, source1, source2);
> - break;
> #if defined(TARGET_RISCV64)
> case OPC_RISC_SRLW:
> /* clear upper 32 */
> @@ -242,12 +233,6 @@ static void gen_arith(DisasContext *ctx, uint32_t opc, int rd, int rs1,
> tcg_gen_andi_tl(source2, source2, TARGET_LONG_BITS - 1);
> tcg_gen_sar_tl(source1, source1, source2);
> break;
> - case OPC_RISC_OR:
> - tcg_gen_or_tl(source1, source1, source2);
> - break;
> - case OPC_RISC_AND:
> - tcg_gen_and_tl(source1, source1, source2);
> - break;
> CASE_OP_32_64(OPC_RISC_MUL):
> tcg_gen_mul_tl(source1, source1, source2);
> break;
> @@ -644,6 +629,24 @@ static void gen_addiw(TCGv ret, TCGv arg1, TCGv arg2)
> }
> #endif
>
> +static bool trans_arith(DisasContext *ctx, arg_r *a,
> + void(*func)(TCGv, TCGv, TCGv))
> +{
> + TCGv source1, source2;
> + source1 = tcg_temp_new();
> + source2 = tcg_temp_new();
> +
> + gen_get_gpr(source1, a->rs1);
> + gen_get_gpr(source2, a->rs2);
> +
> + (*func)(source1, source1, source2);
> +
> + gen_set_gpr(a->rd, source1);
> + tcg_temp_free(source1);
> + tcg_temp_free(source2);
> + return true;
> +}
> +
> /* Include insn module translation function */
> #include "insn_trans/trans_rvi.inc.c"
> #include "insn_trans/trans_rvm.inc.c"
> --
> 2.19.1
>
>
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 08/35] target/riscv: Convert RVXI csr insns to decodetree
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 20:46 ` Alistair
-1 siblings, 0 replies; 140+ messages in thread
From: Alistair @ 2018-10-31 20:46 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: qemu-riscv, peer.adelt, richard.henderson, qemu-devel
On 10/31/18 6:20 AM, Bastian Koppelmann wrote:
> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> v2 -> v3:
> - dropped insn argument of trans_foo functions
>
> target/riscv/insn32.decode | 8 +++
> target/riscv/insn_trans/trans_rvi.inc.c | 79 +++++++++++++++++++++++++
> target/riscv/translate.c | 43 +-------------
> 3 files changed, 88 insertions(+), 42 deletions(-)
>
> diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
> index 6d750b4c5a..b67d3a9437 100644
> --- a/target/riscv/insn32.decode
> +++ b/target/riscv/insn32.decode
> @@ -22,6 +22,7 @@
> %rd 7:5
>
> %sh6 20:6
> +%csr 20:12
>
> # immediates:
> %imm_i 20:s12
> @@ -43,6 +44,7 @@
> @j .................... ..... ....... imm=%imm_j %rd
>
> @sh6 ...... ...... ..... ... ..... ....... &shift shamt=%sh6 %rs1 %rd
> +@csr ............ ..... ... ..... ....... %csr %rs1 %rd
>
> # *** RV32I Base Instruction Set ***
> lui .................... ..... 0110111 @u
> @@ -84,3 +86,9 @@ or 0000000 ..... ..... 110 ..... 0110011 @r
> and 0000000 ..... ..... 111 ..... 0110011 @r
> fence ---- pred:4 succ:4 ----- 000 ----- 0001111
> fence_i ---- ---- ---- ----- 001 ----- 0001111
> +csrrw ............ ..... 001 ..... 1110011 @csr
> +csrrs ............ ..... 010 ..... 1110011 @csr
> +csrrc ............ ..... 011 ..... 1110011 @csr
> +csrrwi ............ ..... 101 ..... 1110011 @csr
> +csrrsi ............ ..... 110 ..... 1110011 @csr
> +csrrci ............ ..... 111 ..... 1110011 @csr
> diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
> index a149e913b1..09e7a0052a 100644
> --- a/target/riscv/insn_trans/trans_rvi.inc.c
> +++ b/target/riscv/insn_trans/trans_rvi.inc.c
> @@ -339,3 +339,82 @@ static bool trans_fence_i(DisasContext *ctx, arg_fence_i *a)
> #endif
> return true;
> }
> +
> +#define RISCV_OP_CSR_PRE do {\
> + source1 = tcg_temp_new(); \
> + csr_store = tcg_temp_new(); \
> + dest = tcg_temp_new(); \
> + rs1_pass = tcg_temp_new(); \
> + gen_get_gpr(source1, a->rs1); \
> + tcg_gen_movi_tl(cpu_pc, ctx->base.pc_next); \
> + tcg_gen_movi_tl(rs1_pass, a->rs1); \
> + tcg_gen_movi_tl(csr_store, a->csr); \
> + gen_io_start();\
> +} while (0)
> +
> +#define RISCV_OP_CSR_POST do {\
> + gen_io_end(); \
> + gen_set_gpr(a->rd, dest); \
> + tcg_gen_movi_tl(cpu_pc, ctx->pc_succ_insn); \
> + tcg_gen_exit_tb(NULL, 0); \
> + ctx->base.is_jmp = DISAS_NORETURN; \
> + tcg_temp_free(source1); \
> + tcg_temp_free(csr_store); \
> + tcg_temp_free(dest); \
> + tcg_temp_free(rs1_pass); \
> +} while (0)
> +
> +
> +static bool trans_csrrw(DisasContext *ctx, arg_csrrw *a)
> +{
> + TCGv source1, csr_store, dest, rs1_pass;
> + RISCV_OP_CSR_PRE;
> + gen_helper_csrrw(dest, cpu_env, source1, csr_store);
> + RISCV_OP_CSR_POST;
> + return true;
> +}
> +
> +static bool trans_csrrs(DisasContext *ctx, arg_csrrs *a)
> +{
> + TCGv source1, csr_store, dest, rs1_pass;
> + RISCV_OP_CSR_PRE;
> + gen_helper_csrrs(dest, cpu_env, source1, csr_store, rs1_pass);
> + RISCV_OP_CSR_POST;
> + return true;
> +}
> +
> +static bool trans_csrrc(DisasContext *ctx, arg_csrrc *a)
> +{
> + TCGv source1, csr_store, dest, rs1_pass;
> + RISCV_OP_CSR_PRE;
> + gen_helper_csrrc(dest, cpu_env, source1, csr_store, rs1_pass);
> + RISCV_OP_CSR_POST;
> + return true;
> +}
> +
> +static bool trans_csrrwi(DisasContext *ctx, arg_csrrwi *a)
> +{
> + TCGv source1, csr_store, dest, rs1_pass;
> + RISCV_OP_CSR_PRE;
> + gen_helper_csrrw(dest, cpu_env, rs1_pass, csr_store);
> + RISCV_OP_CSR_POST;
> + return true;
> +}
> +
> +static bool trans_csrrsi(DisasContext *ctx, arg_csrrsi *a)
> +{
> + TCGv source1, csr_store, dest, rs1_pass;
> + RISCV_OP_CSR_PRE;
> + gen_helper_csrrs(dest, cpu_env, rs1_pass, csr_store, rs1_pass);
> + RISCV_OP_CSR_POST;
> + return true;
> +}
> +
> +static bool trans_csrrci(DisasContext *ctx, arg_csrrci *a)
> +{
> + TCGv source1, csr_store, dest, rs1_pass;
> + RISCV_OP_CSR_PRE;
> + gen_helper_csrrc(dest, cpu_env, rs1_pass, csr_store, rs1_pass);
> + RISCV_OP_CSR_POST;
> + return true;
> +}
> diff --git a/target/riscv/translate.c b/target/riscv/translate.c
> index 80f18fb6aa..be03acd066 100644
> --- a/target/riscv/translate.c
> +++ b/target/riscv/translate.c
> @@ -1277,16 +1277,11 @@ static void gen_fp_arith(DisasContext *ctx, uint32_t opc, int rd,
> static void gen_system(CPURISCVState *env, DisasContext *ctx, uint32_t opc,
> int rd, int rs1, int csr)
> {
> - TCGv source1, csr_store, dest, rs1_pass, imm_rs1;
> + TCGv source1, dest;
> source1 = tcg_temp_new();
> - csr_store = tcg_temp_new();
> dest = tcg_temp_new();
> - rs1_pass = tcg_temp_new();
> - imm_rs1 = tcg_temp_new();
> gen_get_gpr(source1, rs1);
> tcg_gen_movi_tl(cpu_pc, ctx->base.pc_next);
> - tcg_gen_movi_tl(rs1_pass, rs1);
> - tcg_gen_movi_tl(csr_store, csr); /* copy into temp reg to feed to helper */
>
> #ifndef CONFIG_USER_ONLY
> /* Extract funct7 value and check whether it matches SFENCE.VMA */
> @@ -1349,45 +1344,9 @@ static void gen_system(CPURISCVState *env, DisasContext *ctx, uint32_t opc,
> break;
> }
> break;
> - default:
> - tcg_gen_movi_tl(imm_rs1, rs1);
> - gen_io_start();
> - switch (opc) {
> - case OPC_RISC_CSRRW:
> - gen_helper_csrrw(dest, cpu_env, source1, csr_store);
> - break;
> - case OPC_RISC_CSRRS:
> - gen_helper_csrrs(dest, cpu_env, source1, csr_store, rs1_pass);
> - break;
> - case OPC_RISC_CSRRC:
> - gen_helper_csrrc(dest, cpu_env, source1, csr_store, rs1_pass);
> - break;
> - case OPC_RISC_CSRRWI:
> - gen_helper_csrrw(dest, cpu_env, imm_rs1, csr_store);
> - break;
> - case OPC_RISC_CSRRSI:
> - gen_helper_csrrs(dest, cpu_env, imm_rs1, csr_store, rs1_pass);
> - break;
> - case OPC_RISC_CSRRCI:
> - gen_helper_csrrc(dest, cpu_env, imm_rs1, csr_store, rs1_pass);
> - break;
> - default:
> - gen_exception_illegal(ctx);
> - return;
> - }
> - gen_io_end();
> - gen_set_gpr(rd, dest);
> - /* end tb since we may be changing priv modes, to get mmu_index right */
> - tcg_gen_movi_tl(cpu_pc, ctx->pc_succ_insn);
> - tcg_gen_exit_tb(NULL, 0); /* no chaining */
> - ctx->base.is_jmp = DISAS_NORETURN;
> - break;
> }
> tcg_temp_free(source1);
> - tcg_temp_free(csr_store);
> tcg_temp_free(dest);
> - tcg_temp_free(rs1_pass);
> - tcg_temp_free(imm_rs1);
> }
>
> static void decode_RV32_64C0(DisasContext *ctx)
>
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [Qemu-devel] [PATCH v3 08/35] target/riscv: Convert RVXI csr insns to decodetree
@ 2018-10-31 20:46 ` Alistair
0 siblings, 0 replies; 140+ messages in thread
From: Alistair @ 2018-10-31 20:46 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: qemu-riscv, peer.adelt, richard.henderson, qemu-devel
On 10/31/18 6:20 AM, Bastian Koppelmann wrote:
> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> v2 -> v3:
> - dropped insn argument of trans_foo functions
>
> target/riscv/insn32.decode | 8 +++
> target/riscv/insn_trans/trans_rvi.inc.c | 79 +++++++++++++++++++++++++
> target/riscv/translate.c | 43 +-------------
> 3 files changed, 88 insertions(+), 42 deletions(-)
>
> diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode
> index 6d750b4c5a..b67d3a9437 100644
> --- a/target/riscv/insn32.decode
> +++ b/target/riscv/insn32.decode
> @@ -22,6 +22,7 @@
> %rd 7:5
>
> %sh6 20:6
> +%csr 20:12
>
> # immediates:
> %imm_i 20:s12
> @@ -43,6 +44,7 @@
> @j .................... ..... ....... imm=%imm_j %rd
>
> @sh6 ...... ...... ..... ... ..... ....... &shift shamt=%sh6 %rs1 %rd
> +@csr ............ ..... ... ..... ....... %csr %rs1 %rd
>
> # *** RV32I Base Instruction Set ***
> lui .................... ..... 0110111 @u
> @@ -84,3 +86,9 @@ or 0000000 ..... ..... 110 ..... 0110011 @r
> and 0000000 ..... ..... 111 ..... 0110011 @r
> fence ---- pred:4 succ:4 ----- 000 ----- 0001111
> fence_i ---- ---- ---- ----- 001 ----- 0001111
> +csrrw ............ ..... 001 ..... 1110011 @csr
> +csrrs ............ ..... 010 ..... 1110011 @csr
> +csrrc ............ ..... 011 ..... 1110011 @csr
> +csrrwi ............ ..... 101 ..... 1110011 @csr
> +csrrsi ............ ..... 110 ..... 1110011 @csr
> +csrrci ............ ..... 111 ..... 1110011 @csr
> diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
> index a149e913b1..09e7a0052a 100644
> --- a/target/riscv/insn_trans/trans_rvi.inc.c
> +++ b/target/riscv/insn_trans/trans_rvi.inc.c
> @@ -339,3 +339,82 @@ static bool trans_fence_i(DisasContext *ctx, arg_fence_i *a)
> #endif
> return true;
> }
> +
> +#define RISCV_OP_CSR_PRE do {\
> + source1 = tcg_temp_new(); \
> + csr_store = tcg_temp_new(); \
> + dest = tcg_temp_new(); \
> + rs1_pass = tcg_temp_new(); \
> + gen_get_gpr(source1, a->rs1); \
> + tcg_gen_movi_tl(cpu_pc, ctx->base.pc_next); \
> + tcg_gen_movi_tl(rs1_pass, a->rs1); \
> + tcg_gen_movi_tl(csr_store, a->csr); \
> + gen_io_start();\
> +} while (0)
> +
> +#define RISCV_OP_CSR_POST do {\
> + gen_io_end(); \
> + gen_set_gpr(a->rd, dest); \
> + tcg_gen_movi_tl(cpu_pc, ctx->pc_succ_insn); \
> + tcg_gen_exit_tb(NULL, 0); \
> + ctx->base.is_jmp = DISAS_NORETURN; \
> + tcg_temp_free(source1); \
> + tcg_temp_free(csr_store); \
> + tcg_temp_free(dest); \
> + tcg_temp_free(rs1_pass); \
> +} while (0)
> +
> +
> +static bool trans_csrrw(DisasContext *ctx, arg_csrrw *a)
> +{
> + TCGv source1, csr_store, dest, rs1_pass;
> + RISCV_OP_CSR_PRE;
> + gen_helper_csrrw(dest, cpu_env, source1, csr_store);
> + RISCV_OP_CSR_POST;
> + return true;
> +}
> +
> +static bool trans_csrrs(DisasContext *ctx, arg_csrrs *a)
> +{
> + TCGv source1, csr_store, dest, rs1_pass;
> + RISCV_OP_CSR_PRE;
> + gen_helper_csrrs(dest, cpu_env, source1, csr_store, rs1_pass);
> + RISCV_OP_CSR_POST;
> + return true;
> +}
> +
> +static bool trans_csrrc(DisasContext *ctx, arg_csrrc *a)
> +{
> + TCGv source1, csr_store, dest, rs1_pass;
> + RISCV_OP_CSR_PRE;
> + gen_helper_csrrc(dest, cpu_env, source1, csr_store, rs1_pass);
> + RISCV_OP_CSR_POST;
> + return true;
> +}
> +
> +static bool trans_csrrwi(DisasContext *ctx, arg_csrrwi *a)
> +{
> + TCGv source1, csr_store, dest, rs1_pass;
> + RISCV_OP_CSR_PRE;
> + gen_helper_csrrw(dest, cpu_env, rs1_pass, csr_store);
> + RISCV_OP_CSR_POST;
> + return true;
> +}
> +
> +static bool trans_csrrsi(DisasContext *ctx, arg_csrrsi *a)
> +{
> + TCGv source1, csr_store, dest, rs1_pass;
> + RISCV_OP_CSR_PRE;
> + gen_helper_csrrs(dest, cpu_env, rs1_pass, csr_store, rs1_pass);
> + RISCV_OP_CSR_POST;
> + return true;
> +}
> +
> +static bool trans_csrrci(DisasContext *ctx, arg_csrrci *a)
> +{
> + TCGv source1, csr_store, dest, rs1_pass;
> + RISCV_OP_CSR_PRE;
> + gen_helper_csrrc(dest, cpu_env, rs1_pass, csr_store, rs1_pass);
> + RISCV_OP_CSR_POST;
> + return true;
> +}
> diff --git a/target/riscv/translate.c b/target/riscv/translate.c
> index 80f18fb6aa..be03acd066 100644
> --- a/target/riscv/translate.c
> +++ b/target/riscv/translate.c
> @@ -1277,16 +1277,11 @@ static void gen_fp_arith(DisasContext *ctx, uint32_t opc, int rd,
> static void gen_system(CPURISCVState *env, DisasContext *ctx, uint32_t opc,
> int rd, int rs1, int csr)
> {
> - TCGv source1, csr_store, dest, rs1_pass, imm_rs1;
> + TCGv source1, dest;
> source1 = tcg_temp_new();
> - csr_store = tcg_temp_new();
> dest = tcg_temp_new();
> - rs1_pass = tcg_temp_new();
> - imm_rs1 = tcg_temp_new();
> gen_get_gpr(source1, rs1);
> tcg_gen_movi_tl(cpu_pc, ctx->base.pc_next);
> - tcg_gen_movi_tl(rs1_pass, rs1);
> - tcg_gen_movi_tl(csr_store, csr); /* copy into temp reg to feed to helper */
>
> #ifndef CONFIG_USER_ONLY
> /* Extract funct7 value and check whether it matches SFENCE.VMA */
> @@ -1349,45 +1344,9 @@ static void gen_system(CPURISCVState *env, DisasContext *ctx, uint32_t opc,
> break;
> }
> break;
> - default:
> - tcg_gen_movi_tl(imm_rs1, rs1);
> - gen_io_start();
> - switch (opc) {
> - case OPC_RISC_CSRRW:
> - gen_helper_csrrw(dest, cpu_env, source1, csr_store);
> - break;
> - case OPC_RISC_CSRRS:
> - gen_helper_csrrs(dest, cpu_env, source1, csr_store, rs1_pass);
> - break;
> - case OPC_RISC_CSRRC:
> - gen_helper_csrrc(dest, cpu_env, source1, csr_store, rs1_pass);
> - break;
> - case OPC_RISC_CSRRWI:
> - gen_helper_csrrw(dest, cpu_env, imm_rs1, csr_store);
> - break;
> - case OPC_RISC_CSRRSI:
> - gen_helper_csrrs(dest, cpu_env, imm_rs1, csr_store, rs1_pass);
> - break;
> - case OPC_RISC_CSRRCI:
> - gen_helper_csrrc(dest, cpu_env, imm_rs1, csr_store, rs1_pass);
> - break;
> - default:
> - gen_exception_illegal(ctx);
> - return;
> - }
> - gen_io_end();
> - gen_set_gpr(rd, dest);
> - /* end tb since we may be changing priv modes, to get mmu_index right */
> - tcg_gen_movi_tl(cpu_pc, ctx->pc_succ_insn);
> - tcg_gen_exit_tb(NULL, 0); /* no chaining */
> - ctx->base.is_jmp = DISAS_NORETURN;
> - break;
> }
> tcg_temp_free(source1);
> - tcg_temp_free(csr_store);
> tcg_temp_free(dest);
> - tcg_temp_free(rs1_pass);
> - tcg_temp_free(imm_rs1);
> }
>
> static void decode_RV32_64C0(DisasContext *ctx)
>
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 11/35] target/riscv: Convert RV64A insns to decodetree
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 20:49 ` Alistair
-1 siblings, 0 replies; 140+ messages in thread
From: Alistair @ 2018-10-31 20:49 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: qemu-riscv, peer.adelt, richard.henderson, qemu-devel
On 10/31/18 6:20 AM, Bastian Koppelmann wrote:
> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> v2 -> v3:
> - moved 64-bit only insn to insn64.decode
> - dropped insn argument of trans_foo functions
>
> target/riscv/insn64.decode | 13 +++
> target/riscv/insn_trans/trans_rva.inc.c | 58 ++++++++++
> target/riscv/translate.c | 140 ------------------------
> 3 files changed, 71 insertions(+), 140 deletions(-)
>
> diff --git a/target/riscv/insn64.decode b/target/riscv/insn64.decode
> index 008f100546..0bee95c984 100644
> --- a/target/riscv/insn64.decode
> +++ b/target/riscv/insn64.decode
> @@ -43,3 +43,16 @@ divw 0000001 ..... ..... 100 ..... 0111011 @r
> divuw 0000001 ..... ..... 101 ..... 0111011 @r
> remw 0000001 ..... ..... 110 ..... 0111011 @r
> remuw 0000001 ..... ..... 111 ..... 0111011 @r
> +
> +# *** RV64A Standard Extension (in addition to RV32A) ***
> +lr_d 00010 . . 00000 ..... 011 ..... 0101111 @atom_ld
> +sc_d 00011 . . ..... ..... 011 ..... 0101111 @atom_st
> +amoswap_d 00001 . . ..... ..... 011 ..... 0101111 @atom_st
> +amoadd_d 00000 . . ..... ..... 011 ..... 0101111 @atom_st
> +amoxor_d 00100 . . ..... ..... 011 ..... 0101111 @atom_st
> +amoand_d 01100 . . ..... ..... 011 ..... 0101111 @atom_st
> +amoor_d 01000 . . ..... ..... 011 ..... 0101111 @atom_st
> +amomin_d 10000 . . ..... ..... 011 ..... 0101111 @atom_st
> +amomax_d 10100 . . ..... ..... 011 ..... 0101111 @atom_st
> +amominu_d 11000 . . ..... ..... 011 ..... 0101111 @atom_st
> +amomaxu_d 11100 . . ..... ..... 011 ..... 0101111 @atom_st
> diff --git a/target/riscv/insn_trans/trans_rva.inc.c b/target/riscv/insn_trans/trans_rva.inc.c
> index d42d36ebf7..fbfca7f835 100644
> --- a/target/riscv/insn_trans/trans_rva.inc.c
> +++ b/target/riscv/insn_trans/trans_rva.inc.c
> @@ -143,3 +143,61 @@ static bool trans_amomaxu_w(DisasContext *ctx, arg_amomaxu_w *a)
> {
> return gen_amo(ctx, a, &tcg_gen_atomic_fetch_umax_tl, (MO_ALIGN | MO_TESL));
> }
> +
> +#ifdef TARGET_RISCV64
> +
> +static bool trans_lr_d(DisasContext *ctx, arg_lr_d *a)
> +{
> + return gen_lr(ctx, a, MO_ALIGN | MO_TEQ);
> +}
> +
> +static bool trans_sc_d(DisasContext *ctx, arg_sc_d *a)
> +{
> + return gen_sc(ctx, a, (MO_ALIGN | MO_TEQ));
> +}
> +
> +static bool trans_amoswap_d(DisasContext *ctx, arg_amoswap_d *a)
> +{
> + return gen_amo(ctx, a, &tcg_gen_atomic_xchg_tl, (MO_ALIGN | MO_TEQ));
> +}
> +
> +static bool trans_amoadd_d(DisasContext *ctx, arg_amoadd_d *a)
> +{
> + return gen_amo(ctx, a, &tcg_gen_atomic_fetch_add_tl, (MO_ALIGN | MO_TEQ));
> +}
> +
> +static bool trans_amoxor_d(DisasContext *ctx, arg_amoxor_d *a)
> +{
> + return gen_amo(ctx, a, &tcg_gen_atomic_fetch_xor_tl, (MO_ALIGN | MO_TEQ));
> +}
> +
> +static bool trans_amoand_d(DisasContext *ctx, arg_amoand_d *a)
> +{
> + return gen_amo(ctx, a, &tcg_gen_atomic_fetch_and_tl, (MO_ALIGN | MO_TEQ));
> +}
> +
> +static bool trans_amoor_d(DisasContext *ctx, arg_amoor_d *a)
> +{
> + return gen_amo(ctx, a, &tcg_gen_atomic_fetch_or_tl, (MO_ALIGN | MO_TEQ));
> +}
> +
> +static bool trans_amomin_d(DisasContext *ctx, arg_amomin_d *a)
> +{
> + return gen_amo(ctx, a, &tcg_gen_atomic_fetch_smin_tl, (MO_ALIGN | MO_TEQ));
> +}
> +
> +static bool trans_amomax_d(DisasContext *ctx, arg_amomax_d *a)
> +{
> + return gen_amo(ctx, a, &tcg_gen_atomic_fetch_smax_tl, (MO_ALIGN | MO_TEQ));
> +}
> +
> +static bool trans_amominu_d(DisasContext *ctx, arg_amominu_d *a)
> +{
> + return gen_amo(ctx, a, &tcg_gen_atomic_fetch_umin_tl, (MO_ALIGN | MO_TEQ));
> +}
> +
> +static bool trans_amomaxu_d(DisasContext *ctx, arg_amomaxu_d *a)
> +{
> + return gen_amo(ctx, a, &tcg_gen_atomic_fetch_umax_tl, (MO_ALIGN | MO_TEQ));
> +}
> +#endif
> diff --git a/target/riscv/translate.c b/target/riscv/translate.c
> index d8fe319f62..a1b949ad0a 100644
> --- a/target/riscv/translate.c
> +++ b/target/riscv/translate.c
> @@ -711,143 +711,6 @@ static void gen_fp_store(DisasContext *ctx, uint32_t opc, int rs1,
> tcg_temp_free(t0);
> }
>
> -static void gen_atomic(DisasContext *ctx, uint32_t opc,
> - int rd, int rs1, int rs2)
> -{
> - TCGv src1, src2, dat;
> - TCGLabel *l1, *l2;
> - TCGMemOp mop;
> - bool aq, rl;
> -
> - /* Extract the size of the atomic operation. */
> - switch (extract32(opc, 12, 3)) {
> - case 2: /* 32-bit */
> - mop = MO_ALIGN | MO_TESL;
> - break;
> -#if defined(TARGET_RISCV64)
> - case 3: /* 64-bit */
> - mop = MO_ALIGN | MO_TEQ;
> - break;
> -#endif
> - default:
> - gen_exception_illegal(ctx);
> - return;
> - }
> - rl = extract32(opc, 25, 1);
> - aq = extract32(opc, 26, 1);
> -
> - src1 = tcg_temp_new();
> - src2 = tcg_temp_new();
> -
> - switch (MASK_OP_ATOMIC_NO_AQ_RL_SZ(opc)) {
> - case OPC_RISC_LR:
> - /* Put addr in load_res, data in load_val. */
> - gen_get_gpr(src1, rs1);
> - if (rl) {
> - tcg_gen_mb(TCG_MO_ALL | TCG_BAR_STRL);
> - }
> - tcg_gen_qemu_ld_tl(load_val, src1, ctx->mem_idx, mop);
> - if (aq) {
> - tcg_gen_mb(TCG_MO_ALL | TCG_BAR_LDAQ);
> - }
> - tcg_gen_mov_tl(load_res, src1);
> - gen_set_gpr(rd, load_val);
> - break;
> -
> - case OPC_RISC_SC:
> - l1 = gen_new_label();
> - l2 = gen_new_label();
> - dat = tcg_temp_new();
> -
> - gen_get_gpr(src1, rs1);
> - tcg_gen_brcond_tl(TCG_COND_NE, load_res, src1, l1);
> -
> - gen_get_gpr(src2, rs2);
> - /* Note that the TCG atomic primitives are SC,
> - so we can ignore AQ/RL along this path. */
> - tcg_gen_atomic_cmpxchg_tl(src1, load_res, load_val, src2,
> - ctx->mem_idx, mop);
> - tcg_gen_setcond_tl(TCG_COND_NE, dat, src1, load_val);
> - gen_set_gpr(rd, dat);
> - tcg_gen_br(l2);
> -
> - gen_set_label(l1);
> - /* Address comparion failure. However, we still need to
> - provide the memory barrier implied by AQ/RL. */
> - tcg_gen_mb(TCG_MO_ALL + aq * TCG_BAR_LDAQ + rl * TCG_BAR_STRL);
> - tcg_gen_movi_tl(dat, 1);
> - gen_set_gpr(rd, dat);
> -
> - gen_set_label(l2);
> - tcg_temp_free(dat);
> - break;
> -
> - case OPC_RISC_AMOSWAP:
> - /* Note that the TCG atomic primitives are SC,
> - so we can ignore AQ/RL along this path. */
> - gen_get_gpr(src1, rs1);
> - gen_get_gpr(src2, rs2);
> - tcg_gen_atomic_xchg_tl(src2, src1, src2, ctx->mem_idx, mop);
> - gen_set_gpr(rd, src2);
> - break;
> - case OPC_RISC_AMOADD:
> - gen_get_gpr(src1, rs1);
> - gen_get_gpr(src2, rs2);
> - tcg_gen_atomic_fetch_add_tl(src2, src1, src2, ctx->mem_idx, mop);
> - gen_set_gpr(rd, src2);
> - break;
> - case OPC_RISC_AMOXOR:
> - gen_get_gpr(src1, rs1);
> - gen_get_gpr(src2, rs2);
> - tcg_gen_atomic_fetch_xor_tl(src2, src1, src2, ctx->mem_idx, mop);
> - gen_set_gpr(rd, src2);
> - break;
> - case OPC_RISC_AMOAND:
> - gen_get_gpr(src1, rs1);
> - gen_get_gpr(src2, rs2);
> - tcg_gen_atomic_fetch_and_tl(src2, src1, src2, ctx->mem_idx, mop);
> - gen_set_gpr(rd, src2);
> - break;
> - case OPC_RISC_AMOOR:
> - gen_get_gpr(src1, rs1);
> - gen_get_gpr(src2, rs2);
> - tcg_gen_atomic_fetch_or_tl(src2, src1, src2, ctx->mem_idx, mop);
> - gen_set_gpr(rd, src2);
> - break;
> - case OPC_RISC_AMOMIN:
> - gen_get_gpr(src1, rs1);
> - gen_get_gpr(src2, rs2);
> - tcg_gen_atomic_fetch_smin_tl(src2, src1, src2, ctx->mem_idx, mop);
> - gen_set_gpr(rd, src2);
> - break;
> - case OPC_RISC_AMOMAX:
> - gen_get_gpr(src1, rs1);
> - gen_get_gpr(src2, rs2);
> - tcg_gen_atomic_fetch_smax_tl(src2, src1, src2, ctx->mem_idx, mop);
> - gen_set_gpr(rd, src2);
> - break;
> - case OPC_RISC_AMOMINU:
> - gen_get_gpr(src1, rs1);
> - gen_get_gpr(src2, rs2);
> - tcg_gen_atomic_fetch_umin_tl(src2, src1, src2, ctx->mem_idx, mop);
> - gen_set_gpr(rd, src2);
> - break;
> - case OPC_RISC_AMOMAXU:
> - gen_get_gpr(src1, rs1);
> - gen_get_gpr(src2, rs2);
> - tcg_gen_atomic_fetch_umax_tl(src2, src1, src2, ctx->mem_idx, mop);
> - gen_set_gpr(rd, src2);
> - break;
> -
> - default:
> - gen_exception_illegal(ctx);
> - break;
> - }
> -
> - tcg_temp_free(src1);
> - tcg_temp_free(src2);
> -}
> -
> static void gen_set_rm(DisasContext *ctx, int rm)
> {
> TCGv_i32 t0;
> @@ -1669,9 +1532,6 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
> gen_fp_store(ctx, MASK_OP_FP_STORE(ctx->opcode), rs1, rs2,
> GET_STORE_IMM(ctx->opcode));
> break;
> - case OPC_RISC_ATOMIC:
> - gen_atomic(ctx, MASK_OP_ATOMIC(ctx->opcode), rd, rs1, rs2);
> - break;
> case OPC_RISC_FMADD:
> gen_fp_fmadd(ctx, MASK_OP_FP_FMADD(ctx->opcode), rd, rs1, rs2,
> GET_RS3(ctx->opcode), GET_RM(ctx->opcode));
>
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [Qemu-devel] [PATCH v3 11/35] target/riscv: Convert RV64A insns to decodetree
@ 2018-10-31 20:49 ` Alistair
0 siblings, 0 replies; 140+ messages in thread
From: Alistair @ 2018-10-31 20:49 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: qemu-riscv, peer.adelt, richard.henderson, qemu-devel
On 10/31/18 6:20 AM, Bastian Koppelmann wrote:
> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> v2 -> v3:
> - moved 64-bit only insn to insn64.decode
> - dropped insn argument of trans_foo functions
>
> target/riscv/insn64.decode | 13 +++
> target/riscv/insn_trans/trans_rva.inc.c | 58 ++++++++++
> target/riscv/translate.c | 140 ------------------------
> 3 files changed, 71 insertions(+), 140 deletions(-)
>
> diff --git a/target/riscv/insn64.decode b/target/riscv/insn64.decode
> index 008f100546..0bee95c984 100644
> --- a/target/riscv/insn64.decode
> +++ b/target/riscv/insn64.decode
> @@ -43,3 +43,16 @@ divw 0000001 ..... ..... 100 ..... 0111011 @r
> divuw 0000001 ..... ..... 101 ..... 0111011 @r
> remw 0000001 ..... ..... 110 ..... 0111011 @r
> remuw 0000001 ..... ..... 111 ..... 0111011 @r
> +
> +# *** RV64A Standard Extension (in addition to RV32A) ***
> +lr_d 00010 . . 00000 ..... 011 ..... 0101111 @atom_ld
> +sc_d 00011 . . ..... ..... 011 ..... 0101111 @atom_st
> +amoswap_d 00001 . . ..... ..... 011 ..... 0101111 @atom_st
> +amoadd_d 00000 . . ..... ..... 011 ..... 0101111 @atom_st
> +amoxor_d 00100 . . ..... ..... 011 ..... 0101111 @atom_st
> +amoand_d 01100 . . ..... ..... 011 ..... 0101111 @atom_st
> +amoor_d 01000 . . ..... ..... 011 ..... 0101111 @atom_st
> +amomin_d 10000 . . ..... ..... 011 ..... 0101111 @atom_st
> +amomax_d 10100 . . ..... ..... 011 ..... 0101111 @atom_st
> +amominu_d 11000 . . ..... ..... 011 ..... 0101111 @atom_st
> +amomaxu_d 11100 . . ..... ..... 011 ..... 0101111 @atom_st
> diff --git a/target/riscv/insn_trans/trans_rva.inc.c b/target/riscv/insn_trans/trans_rva.inc.c
> index d42d36ebf7..fbfca7f835 100644
> --- a/target/riscv/insn_trans/trans_rva.inc.c
> +++ b/target/riscv/insn_trans/trans_rva.inc.c
> @@ -143,3 +143,61 @@ static bool trans_amomaxu_w(DisasContext *ctx, arg_amomaxu_w *a)
> {
> return gen_amo(ctx, a, &tcg_gen_atomic_fetch_umax_tl, (MO_ALIGN | MO_TESL));
> }
> +
> +#ifdef TARGET_RISCV64
> +
> +static bool trans_lr_d(DisasContext *ctx, arg_lr_d *a)
> +{
> + return gen_lr(ctx, a, MO_ALIGN | MO_TEQ);
> +}
> +
> +static bool trans_sc_d(DisasContext *ctx, arg_sc_d *a)
> +{
> + return gen_sc(ctx, a, (MO_ALIGN | MO_TEQ));
> +}
> +
> +static bool trans_amoswap_d(DisasContext *ctx, arg_amoswap_d *a)
> +{
> + return gen_amo(ctx, a, &tcg_gen_atomic_xchg_tl, (MO_ALIGN | MO_TEQ));
> +}
> +
> +static bool trans_amoadd_d(DisasContext *ctx, arg_amoadd_d *a)
> +{
> + return gen_amo(ctx, a, &tcg_gen_atomic_fetch_add_tl, (MO_ALIGN | MO_TEQ));
> +}
> +
> +static bool trans_amoxor_d(DisasContext *ctx, arg_amoxor_d *a)
> +{
> + return gen_amo(ctx, a, &tcg_gen_atomic_fetch_xor_tl, (MO_ALIGN | MO_TEQ));
> +}
> +
> +static bool trans_amoand_d(DisasContext *ctx, arg_amoand_d *a)
> +{
> + return gen_amo(ctx, a, &tcg_gen_atomic_fetch_and_tl, (MO_ALIGN | MO_TEQ));
> +}
> +
> +static bool trans_amoor_d(DisasContext *ctx, arg_amoor_d *a)
> +{
> + return gen_amo(ctx, a, &tcg_gen_atomic_fetch_or_tl, (MO_ALIGN | MO_TEQ));
> +}
> +
> +static bool trans_amomin_d(DisasContext *ctx, arg_amomin_d *a)
> +{
> + return gen_amo(ctx, a, &tcg_gen_atomic_fetch_smin_tl, (MO_ALIGN | MO_TEQ));
> +}
> +
> +static bool trans_amomax_d(DisasContext *ctx, arg_amomax_d *a)
> +{
> + return gen_amo(ctx, a, &tcg_gen_atomic_fetch_smax_tl, (MO_ALIGN | MO_TEQ));
> +}
> +
> +static bool trans_amominu_d(DisasContext *ctx, arg_amominu_d *a)
> +{
> + return gen_amo(ctx, a, &tcg_gen_atomic_fetch_umin_tl, (MO_ALIGN | MO_TEQ));
> +}
> +
> +static bool trans_amomaxu_d(DisasContext *ctx, arg_amomaxu_d *a)
> +{
> + return gen_amo(ctx, a, &tcg_gen_atomic_fetch_umax_tl, (MO_ALIGN | MO_TEQ));
> +}
> +#endif
> diff --git a/target/riscv/translate.c b/target/riscv/translate.c
> index d8fe319f62..a1b949ad0a 100644
> --- a/target/riscv/translate.c
> +++ b/target/riscv/translate.c
> @@ -711,143 +711,6 @@ static void gen_fp_store(DisasContext *ctx, uint32_t opc, int rs1,
> tcg_temp_free(t0);
> }
>
> -static void gen_atomic(DisasContext *ctx, uint32_t opc,
> - int rd, int rs1, int rs2)
> -{
> - TCGv src1, src2, dat;
> - TCGLabel *l1, *l2;
> - TCGMemOp mop;
> - bool aq, rl;
> -
> - /* Extract the size of the atomic operation. */
> - switch (extract32(opc, 12, 3)) {
> - case 2: /* 32-bit */
> - mop = MO_ALIGN | MO_TESL;
> - break;
> -#if defined(TARGET_RISCV64)
> - case 3: /* 64-bit */
> - mop = MO_ALIGN | MO_TEQ;
> - break;
> -#endif
> - default:
> - gen_exception_illegal(ctx);
> - return;
> - }
> - rl = extract32(opc, 25, 1);
> - aq = extract32(opc, 26, 1);
> -
> - src1 = tcg_temp_new();
> - src2 = tcg_temp_new();
> -
> - switch (MASK_OP_ATOMIC_NO_AQ_RL_SZ(opc)) {
> - case OPC_RISC_LR:
> - /* Put addr in load_res, data in load_val. */
> - gen_get_gpr(src1, rs1);
> - if (rl) {
> - tcg_gen_mb(TCG_MO_ALL | TCG_BAR_STRL);
> - }
> - tcg_gen_qemu_ld_tl(load_val, src1, ctx->mem_idx, mop);
> - if (aq) {
> - tcg_gen_mb(TCG_MO_ALL | TCG_BAR_LDAQ);
> - }
> - tcg_gen_mov_tl(load_res, src1);
> - gen_set_gpr(rd, load_val);
> - break;
> -
> - case OPC_RISC_SC:
> - l1 = gen_new_label();
> - l2 = gen_new_label();
> - dat = tcg_temp_new();
> -
> - gen_get_gpr(src1, rs1);
> - tcg_gen_brcond_tl(TCG_COND_NE, load_res, src1, l1);
> -
> - gen_get_gpr(src2, rs2);
> - /* Note that the TCG atomic primitives are SC,
> - so we can ignore AQ/RL along this path. */
> - tcg_gen_atomic_cmpxchg_tl(src1, load_res, load_val, src2,
> - ctx->mem_idx, mop);
> - tcg_gen_setcond_tl(TCG_COND_NE, dat, src1, load_val);
> - gen_set_gpr(rd, dat);
> - tcg_gen_br(l2);
> -
> - gen_set_label(l1);
> - /* Address comparion failure. However, we still need to
> - provide the memory barrier implied by AQ/RL. */
> - tcg_gen_mb(TCG_MO_ALL + aq * TCG_BAR_LDAQ + rl * TCG_BAR_STRL);
> - tcg_gen_movi_tl(dat, 1);
> - gen_set_gpr(rd, dat);
> -
> - gen_set_label(l2);
> - tcg_temp_free(dat);
> - break;
> -
> - case OPC_RISC_AMOSWAP:
> - /* Note that the TCG atomic primitives are SC,
> - so we can ignore AQ/RL along this path. */
> - gen_get_gpr(src1, rs1);
> - gen_get_gpr(src2, rs2);
> - tcg_gen_atomic_xchg_tl(src2, src1, src2, ctx->mem_idx, mop);
> - gen_set_gpr(rd, src2);
> - break;
> - case OPC_RISC_AMOADD:
> - gen_get_gpr(src1, rs1);
> - gen_get_gpr(src2, rs2);
> - tcg_gen_atomic_fetch_add_tl(src2, src1, src2, ctx->mem_idx, mop);
> - gen_set_gpr(rd, src2);
> - break;
> - case OPC_RISC_AMOXOR:
> - gen_get_gpr(src1, rs1);
> - gen_get_gpr(src2, rs2);
> - tcg_gen_atomic_fetch_xor_tl(src2, src1, src2, ctx->mem_idx, mop);
> - gen_set_gpr(rd, src2);
> - break;
> - case OPC_RISC_AMOAND:
> - gen_get_gpr(src1, rs1);
> - gen_get_gpr(src2, rs2);
> - tcg_gen_atomic_fetch_and_tl(src2, src1, src2, ctx->mem_idx, mop);
> - gen_set_gpr(rd, src2);
> - break;
> - case OPC_RISC_AMOOR:
> - gen_get_gpr(src1, rs1);
> - gen_get_gpr(src2, rs2);
> - tcg_gen_atomic_fetch_or_tl(src2, src1, src2, ctx->mem_idx, mop);
> - gen_set_gpr(rd, src2);
> - break;
> - case OPC_RISC_AMOMIN:
> - gen_get_gpr(src1, rs1);
> - gen_get_gpr(src2, rs2);
> - tcg_gen_atomic_fetch_smin_tl(src2, src1, src2, ctx->mem_idx, mop);
> - gen_set_gpr(rd, src2);
> - break;
> - case OPC_RISC_AMOMAX:
> - gen_get_gpr(src1, rs1);
> - gen_get_gpr(src2, rs2);
> - tcg_gen_atomic_fetch_smax_tl(src2, src1, src2, ctx->mem_idx, mop);
> - gen_set_gpr(rd, src2);
> - break;
> - case OPC_RISC_AMOMINU:
> - gen_get_gpr(src1, rs1);
> - gen_get_gpr(src2, rs2);
> - tcg_gen_atomic_fetch_umin_tl(src2, src1, src2, ctx->mem_idx, mop);
> - gen_set_gpr(rd, src2);
> - break;
> - case OPC_RISC_AMOMAXU:
> - gen_get_gpr(src1, rs1);
> - gen_get_gpr(src2, rs2);
> - tcg_gen_atomic_fetch_umax_tl(src2, src1, src2, ctx->mem_idx, mop);
> - gen_set_gpr(rd, src2);
> - break;
> -
> - default:
> - gen_exception_illegal(ctx);
> - break;
> - }
> -
> - tcg_temp_free(src1);
> - tcg_temp_free(src2);
> -}
> -
> static void gen_set_rm(DisasContext *ctx, int rm)
> {
> TCGv_i32 t0;
> @@ -1669,9 +1532,6 @@ static void decode_RV32_64G(CPURISCVState *env, DisasContext *ctx)
> gen_fp_store(ctx, MASK_OP_FP_STORE(ctx->opcode), rs1, rs2,
> GET_STORE_IMM(ctx->opcode));
> break;
> - case OPC_RISC_ATOMIC:
> - gen_atomic(ctx, MASK_OP_ATOMIC(ctx->opcode), rd, rs1, rs2);
> - break;
> case OPC_RISC_FMADD:
> gen_fp_fmadd(ctx, MASK_OP_FP_FMADD(ctx->opcode), rd, rs1, rs2,
> GET_RS3(ctx->opcode), GET_RM(ctx->opcode));
>
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 20/35] target/riscv: Remove gen_jalr()
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 20:50 ` Alistair
-1 siblings, 0 replies; 140+ messages in thread
From: Alistair @ 2018-10-31 20:50 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: qemu-riscv, peer.adelt, richard.henderson, qemu-devel
On 10/31/18 6:20 AM, Bastian Koppelmann wrote:
> trans_jalr() is the only caller, so move the code into trans_jalr().
>
> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> target/riscv/insn_trans/trans_rvi.inc.c | 28 +++++++++++++++++-
> target/riscv/translate.c | 38 -------------------------
> 2 files changed, 27 insertions(+), 39 deletions(-)
>
> diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
> index 09e7a0052a..4d090d68e7 100644
> --- a/target/riscv/insn_trans/trans_rvi.inc.c
> +++ b/target/riscv/insn_trans/trans_rvi.inc.c
> @@ -42,7 +42,33 @@ static bool trans_jal(DisasContext *ctx, arg_jal *a)
>
> static bool trans_jalr(DisasContext *ctx, arg_jalr *a)
> {
> - gen_jalr(ctx->env, ctx, OPC_RISC_JALR, a->rd, a->rs1, a->imm);
> + /* no chaining with JALR */
> + TCGLabel *misaligned = NULL;
> + TCGv t0 = tcg_temp_new();
> +
> +
> + gen_get_gpr(cpu_pc, a->rs1);
> + tcg_gen_addi_tl(cpu_pc, cpu_pc, a->imm);
> + tcg_gen_andi_tl(cpu_pc, cpu_pc, (target_ulong)-2);
> +
> + if (!riscv_has_ext(ctx->env, RVC)) {
> + misaligned = gen_new_label();
> + tcg_gen_andi_tl(t0, cpu_pc, 0x2);
> + tcg_gen_brcondi_tl(TCG_COND_NE, t0, 0x0, misaligned);
> + }
> +
> + if (a->rd != 0) {
> + tcg_gen_movi_tl(cpu_gpr[a->rd], ctx->pc_succ_insn);
> + }
> + tcg_gen_lookup_and_goto_ptr();
> +
> + if (misaligned) {
> + gen_set_label(misaligned);
> + gen_exception_inst_addr_mis(ctx);
> + }
> + ctx->base.is_jmp = DISAS_NORETURN;
> +
> + tcg_temp_free(t0);
> return true;
> }
>
> diff --git a/target/riscv/translate.c b/target/riscv/translate.c
> index 41d66fae18..b78e423d94 100644
> --- a/target/riscv/translate.c
> +++ b/target/riscv/translate.c
> @@ -489,44 +489,6 @@ static void gen_jal(CPURISCVState *env, DisasContext *ctx, int rd,
> ctx->base.is_jmp = DISAS_NORETURN;
> }
>
> -static void gen_jalr(CPURISCVState *env, DisasContext *ctx, uint32_t opc,
> - int rd, int rs1, target_long imm)
> -{
> - /* no chaining with JALR */
> - TCGLabel *misaligned = NULL;
> - TCGv t0 = tcg_temp_new();
> -
> - switch (opc) {
> - case OPC_RISC_JALR:
> - gen_get_gpr(cpu_pc, rs1);
> - tcg_gen_addi_tl(cpu_pc, cpu_pc, imm);
> - tcg_gen_andi_tl(cpu_pc, cpu_pc, (target_ulong)-2);
> -
> - if (!riscv_has_ext(env, RVC)) {
> - misaligned = gen_new_label();
> - tcg_gen_andi_tl(t0, cpu_pc, 0x2);
> - tcg_gen_brcondi_tl(TCG_COND_NE, t0, 0x0, misaligned);
> - }
> -
> - if (rd != 0) {
> - tcg_gen_movi_tl(cpu_gpr[rd], ctx->pc_succ_insn);
> - }
> - tcg_gen_lookup_and_goto_ptr();
> -
> - if (misaligned) {
> - gen_set_label(misaligned);
> - gen_exception_inst_addr_mis(ctx);
> - }
> - ctx->base.is_jmp = DISAS_NORETURN;
> - break;
> -
> - default:
> - gen_exception_illegal(ctx);
> - break;
> - }
> - tcg_temp_free(t0);
> -}
> -
> static void gen_branch(CPURISCVState *env, DisasContext *ctx, uint32_t opc,
> int rs1, int rs2, target_long bimm)
> {
>
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [Qemu-devel] [PATCH v3 20/35] target/riscv: Remove gen_jalr()
@ 2018-10-31 20:50 ` Alistair
0 siblings, 0 replies; 140+ messages in thread
From: Alistair @ 2018-10-31 20:50 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: qemu-riscv, peer.adelt, richard.henderson, qemu-devel
On 10/31/18 6:20 AM, Bastian Koppelmann wrote:
> trans_jalr() is the only caller, so move the code into trans_jalr().
>
> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Alistair
> ---
> target/riscv/insn_trans/trans_rvi.inc.c | 28 +++++++++++++++++-
> target/riscv/translate.c | 38 -------------------------
> 2 files changed, 27 insertions(+), 39 deletions(-)
>
> diff --git a/target/riscv/insn_trans/trans_rvi.inc.c b/target/riscv/insn_trans/trans_rvi.inc.c
> index 09e7a0052a..4d090d68e7 100644
> --- a/target/riscv/insn_trans/trans_rvi.inc.c
> +++ b/target/riscv/insn_trans/trans_rvi.inc.c
> @@ -42,7 +42,33 @@ static bool trans_jal(DisasContext *ctx, arg_jal *a)
>
> static bool trans_jalr(DisasContext *ctx, arg_jalr *a)
> {
> - gen_jalr(ctx->env, ctx, OPC_RISC_JALR, a->rd, a->rs1, a->imm);
> + /* no chaining with JALR */
> + TCGLabel *misaligned = NULL;
> + TCGv t0 = tcg_temp_new();
> +
> +
> + gen_get_gpr(cpu_pc, a->rs1);
> + tcg_gen_addi_tl(cpu_pc, cpu_pc, a->imm);
> + tcg_gen_andi_tl(cpu_pc, cpu_pc, (target_ulong)-2);
> +
> + if (!riscv_has_ext(ctx->env, RVC)) {
> + misaligned = gen_new_label();
> + tcg_gen_andi_tl(t0, cpu_pc, 0x2);
> + tcg_gen_brcondi_tl(TCG_COND_NE, t0, 0x0, misaligned);
> + }
> +
> + if (a->rd != 0) {
> + tcg_gen_movi_tl(cpu_gpr[a->rd], ctx->pc_succ_insn);
> + }
> + tcg_gen_lookup_and_goto_ptr();
> +
> + if (misaligned) {
> + gen_set_label(misaligned);
> + gen_exception_inst_addr_mis(ctx);
> + }
> + ctx->base.is_jmp = DISAS_NORETURN;
> +
> + tcg_temp_free(t0);
> return true;
> }
>
> diff --git a/target/riscv/translate.c b/target/riscv/translate.c
> index 41d66fae18..b78e423d94 100644
> --- a/target/riscv/translate.c
> +++ b/target/riscv/translate.c
> @@ -489,44 +489,6 @@ static void gen_jal(CPURISCVState *env, DisasContext *ctx, int rd,
> ctx->base.is_jmp = DISAS_NORETURN;
> }
>
> -static void gen_jalr(CPURISCVState *env, DisasContext *ctx, uint32_t opc,
> - int rd, int rs1, target_long imm)
> -{
> - /* no chaining with JALR */
> - TCGLabel *misaligned = NULL;
> - TCGv t0 = tcg_temp_new();
> -
> - switch (opc) {
> - case OPC_RISC_JALR:
> - gen_get_gpr(cpu_pc, rs1);
> - tcg_gen_addi_tl(cpu_pc, cpu_pc, imm);
> - tcg_gen_andi_tl(cpu_pc, cpu_pc, (target_ulong)-2);
> -
> - if (!riscv_has_ext(env, RVC)) {
> - misaligned = gen_new_label();
> - tcg_gen_andi_tl(t0, cpu_pc, 0x2);
> - tcg_gen_brcondi_tl(TCG_COND_NE, t0, 0x0, misaligned);
> - }
> -
> - if (rd != 0) {
> - tcg_gen_movi_tl(cpu_gpr[rd], ctx->pc_succ_insn);
> - }
> - tcg_gen_lookup_and_goto_ptr();
> -
> - if (misaligned) {
> - gen_set_label(misaligned);
> - gen_exception_inst_addr_mis(ctx);
> - }
> - ctx->base.is_jmp = DISAS_NORETURN;
> - break;
> -
> - default:
> - gen_exception_illegal(ctx);
> - break;
> - }
> - tcg_temp_free(t0);
> -}
> -
> static void gen_branch(CPURISCVState *env, DisasContext *ctx, uint32_t opc,
> int rs1, int rs2, target_long bimm)
> {
>
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 22/35] target/riscv: Remove manual decoding from gen_load()
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 22:09 ` Richard Henderson
-1 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 22:09 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
> With decodetree we don't need to convert RISC-V opcodes into to MemOps
> as the old gen_load() did.
>
> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de><Paste>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
> +static bool gen_load(DisasContext *ctx, arg_lb *a, int memop)
Better with TCGMemOp instead of int.
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [PATCH v3 22/35] target/riscv: Remove manual decoding from gen_load()
@ 2018-10-31 22:09 ` Richard Henderson
0 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 22:09 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
> With decodetree we don't need to convert RISC-V opcodes into to MemOps
> as the old gen_load() did.
>
> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de><Paste>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
> +static bool gen_load(DisasContext *ctx, arg_lb *a, int memop)
Better with TCGMemOp instead of int.
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 23/35] target/riscv: Remove manual decoding from gen_store()
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 22:09 ` Richard Henderson
-1 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 22:09 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
> +static bool gen_store(DisasContext *ctx, arg_sb *a, int memop)
Likewise with TCGMemOp.
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [PATCH v3 23/35] target/riscv: Remove manual decoding from gen_store()
@ 2018-10-31 22:09 ` Richard Henderson
0 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 22:09 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
> +static bool gen_store(DisasContext *ctx, arg_sb *a, int memop)
Likewise with TCGMemOp.
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 24/35] target/riscv: Move gen_arith_imm() decoding into trans_* functions
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 22:18 ` Richard Henderson
-1 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 22:18 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
> static bool trans_slli(DisasContext *ctx, arg_slli *a)
> {
> - gen_arith_imm(ctx, OPC_RISC_SLLI, a->rd, a->rs1, a->shamt);
> + if (a->rd != 0) {
> + TCGv t = tcg_temp_new();
> + gen_get_gpr(t, a->rs1);
> +
> + if (a->shamt >= TARGET_LONG_BITS) {
> + return false;
> + }
> + tcg_gen_shli_tl(t, t, a->shamt);
> +
> + gen_set_gpr(a->rd, t);
> + tcg_temp_free(t);
> + } /* NOP otherwise */
> return true;
> }
>
> static bool trans_srli(DisasContext *ctx, arg_srli *a)
> {
> - gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_I, a->rd, a->rs1, a->shamt);
> + if (a->rd != 0) {
> + TCGv t = tcg_temp_new();
> + gen_get_gpr(t, a->rs1);
> + tcg_gen_shri_tl(t, t, a->shamt);
> + gen_set_gpr(a->rd, t);
> + tcg_temp_free(t);
> + } /* NOP otherwise */
> return true;
> }
>
> static bool trans_srai(DisasContext *ctx, arg_srai *a)
> {
> - gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_I, a->rd, a->rs1, a->shamt | 0x400);
> + if (a->rd != 0) {
> + TCGv t = tcg_temp_new();
> + gen_get_gpr(t, a->rs1);
> + tcg_gen_sari_tl(t, t, a->shamt);
> + gen_set_gpr(a->rd, t);
> + tcg_temp_free(t);
> + } /* NOP otherwise */
> return true;
> }
Surely the shri and sari functions need the same shamt >= TARGET_LONG_BITS
check as slli. Otherwise RV32 shri should definitely produce an assert in
tcg_gen_shri_tl.
I did wonder about changing the decode of the shift functions such that only
the top two bits of the imm are reserved for secondary parsing of the shift
type, and the other 10 bits are passed down into trans_foo. At which point the
TARGET_LONG_BITS check takes care of other illegalities.
Which means that the parsing for slli and slliw are identical, and also that
for the far future when RV128 is a thing, we don't have to change the parsing.
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [PATCH v3 24/35] target/riscv: Move gen_arith_imm() decoding into trans_* functions
@ 2018-10-31 22:18 ` Richard Henderson
0 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 22:18 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
> static bool trans_slli(DisasContext *ctx, arg_slli *a)
> {
> - gen_arith_imm(ctx, OPC_RISC_SLLI, a->rd, a->rs1, a->shamt);
> + if (a->rd != 0) {
> + TCGv t = tcg_temp_new();
> + gen_get_gpr(t, a->rs1);
> +
> + if (a->shamt >= TARGET_LONG_BITS) {
> + return false;
> + }
> + tcg_gen_shli_tl(t, t, a->shamt);
> +
> + gen_set_gpr(a->rd, t);
> + tcg_temp_free(t);
> + } /* NOP otherwise */
> return true;
> }
>
> static bool trans_srli(DisasContext *ctx, arg_srli *a)
> {
> - gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_I, a->rd, a->rs1, a->shamt);
> + if (a->rd != 0) {
> + TCGv t = tcg_temp_new();
> + gen_get_gpr(t, a->rs1);
> + tcg_gen_shri_tl(t, t, a->shamt);
> + gen_set_gpr(a->rd, t);
> + tcg_temp_free(t);
> + } /* NOP otherwise */
> return true;
> }
>
> static bool trans_srai(DisasContext *ctx, arg_srai *a)
> {
> - gen_arith_imm(ctx, OPC_RISC_SHIFT_RIGHT_I, a->rd, a->rs1, a->shamt | 0x400);
> + if (a->rd != 0) {
> + TCGv t = tcg_temp_new();
> + gen_get_gpr(t, a->rs1);
> + tcg_gen_sari_tl(t, t, a->shamt);
> + gen_set_gpr(a->rd, t);
> + tcg_temp_free(t);
> + } /* NOP otherwise */
> return true;
> }
Surely the shri and sari functions need the same shamt >= TARGET_LONG_BITS
check as slli. Otherwise RV32 shri should definitely produce an assert in
tcg_gen_shri_tl.
I did wonder about changing the decode of the shift functions such that only
the top two bits of the imm are reserved for secondary parsing of the shift
type, and the other 10 bits are passed down into trans_foo. At which point the
TARGET_LONG_BITS check takes care of other illegalities.
Which means that the parsing for slli and slliw are identical, and also that
for the far future when RV128 is a thing, we don't have to change the parsing.
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 24/35] target/riscv: Move gen_arith_imm() decoding into trans_* functions
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 22:26 ` Richard Henderson
-1 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 22:26 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
> +
> +#ifdef TARGET_RISCV64
> +static void gen_addiw(TCGv ret, TCGv arg1, TCGv arg2)
> +{
> + tcg_gen_add_tl(ret, arg1, arg2);
> + tcg_gen_ext32s_tl(ret, ret);
> +}
> +#endif
This should have been called gen_addw. It takes no immediate.
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [PATCH v3 24/35] target/riscv: Move gen_arith_imm() decoding into trans_* functions
@ 2018-10-31 22:26 ` Richard Henderson
0 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 22:26 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
> +
> +#ifdef TARGET_RISCV64
> +static void gen_addiw(TCGv ret, TCGv arg1, TCGv arg2)
> +{
> + tcg_gen_add_tl(ret, arg1, arg2);
> + tcg_gen_ext32s_tl(ret, ret);
> +}
> +#endif
This should have been called gen_addw. It takes no immediate.
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 25/35] target/riscv: make ADD/SUB/OR/XOR/AND insn use arg lists
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 22:26 ` Richard Henderson
-1 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 22:26 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
> static bool trans_addw(DisasContext *ctx, arg_addw *a)
> {
> - gen_arith(ctx, OPC_RISC_ADDW, a->rd, a->rs1, a->rs2);
> - return true;
> + return trans_arith(ctx, a, &tcg_gen_add_tl);
> }
This should be using gen_addw.
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [PATCH v3 25/35] target/riscv: make ADD/SUB/OR/XOR/AND insn use arg lists
@ 2018-10-31 22:26 ` Richard Henderson
0 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 22:26 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
> static bool trans_addw(DisasContext *ctx, arg_addw *a)
> {
> - gen_arith(ctx, OPC_RISC_ADDW, a->rd, a->rs1, a->rs2);
> - return true;
> + return trans_arith(ctx, a, &tcg_gen_add_tl);
> }
This should be using gen_addw.
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 25/35] target/riscv: make ADD/SUB/OR/XOR/AND insn use arg lists
2018-10-31 20:44 ` [Qemu-riscv] " Alistair Francis
@ 2018-10-31 22:27 ` Richard Henderson
-1 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 22:27 UTC (permalink / raw)
To: Alistair Francis, Bastian Koppelmann
Cc: Michael Clark, Sagar Karandikar, Palmer Dabbelt,
Alistair Francis, qemu-riscv, peer.adelt,
qemu-devel@nongnu.org Developers
On 10/31/18 8:44 PM, Alistair Francis wrote:
> This series causes a lot of kernel oops during boot...
The only bug I can see is in trans_addw. But that's a biggie and could easily
explain, since non-extended values could have caused a compare to misfire.
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [Qemu-devel] [PATCH v3 25/35] target/riscv: make ADD/SUB/OR/XOR/AND insn use arg lists
@ 2018-10-31 22:27 ` Richard Henderson
0 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 22:27 UTC (permalink / raw)
To: Alistair Francis, Bastian Koppelmann
Cc: Michael Clark, Sagar Karandikar, Palmer Dabbelt,
Alistair Francis, qemu-riscv, peer.adelt,
qemu-devel@nongnu.org Developers
On 10/31/18 8:44 PM, Alistair Francis wrote:
> This series causes a lot of kernel oops during boot...
The only bug I can see is in trans_addw. But that's a biggie and could easily
explain, since non-extended values could have caused a compare to misfire.
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 26/35] target/riscv: Remove shift and slt insn manual decoding
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 22:38 ` Richard Henderson
-1 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 22:38 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
> static bool trans_slt(DisasContext *ctx, arg_slt *a)
> {
> - gen_arith(ctx, OPC_RISC_SLT, a->rd, a->rs1, a->rs2);
> + TCGv source1 = tcg_temp_new();
> + TCGv source2 = tcg_temp_new();
> +
> + gen_get_gpr(source1, a->rs1);
> + gen_get_gpr(source2, a->rs2);
> +
> + tcg_gen_setcond_tl(TCG_COND_LT, source1, source1, source2);
I do wonder about extracting this one line to gen_slt so that you can re-use
gen_arith and gen_arithi.
> + tcg_gen_setcond_tl(TCG_COND_LTU, source1, source1, source2);
Similarly.
> static bool trans_sllw(DisasContext *ctx, arg_sllw *a)
> {
> - gen_arith(ctx, OPC_RISC_SLLW, a->rd, a->rs1, a->rs2);
> + TCGv source1 = tcg_temp_new();
> + TCGv source2 = tcg_temp_new();
> +
> + gen_get_gpr(source1, a->rs1);
> + gen_get_gpr(source2, a->rs2);
> +
> + tcg_gen_andi_tl(source2, source2, 0x1F);
> + tcg_gen_shl_tl(source1, source1, source2);
> +
> + gen_set_gpr(a->rd, source1);
Missing the ext32s after the shift.
> static bool trans_srlw(DisasContext *ctx, arg_srlw *a)
> {
> - gen_arith(ctx, OPC_RISC_SRLW, a->rd, a->rs1, a->rs2);
> + TCGv source1 = tcg_temp_new();
> + TCGv source2 = tcg_temp_new();
> +
> + gen_get_gpr(source1, a->rs1);
> + gen_get_gpr(source2, a->rs2);
> +
> + /* clear upper 32 */
> + tcg_gen_ext32u_tl(source1, source1);
> + tcg_gen_andi_tl(source2, source2, 0x1F);
> + tcg_gen_shr_tl(source1, source1, source2);
Likewise. (Consider source2 == 0.)
> - case OPC_RISC_SRL:
> - tcg_gen_andi_tl(source2, source2, TARGET_LONG_BITS - 1);
> - tcg_gen_shr_tl(source1, source1, source2);
> - break;
...
> - case OPC_RISC_SRA:
> - tcg_gen_andi_tl(source2, source2, TARGET_LONG_BITS - 1);
> - tcg_gen_sar_tl(source1, source1, source2);
> - break;
I see that the bugs are in the original though, so fixing them in a separate
patch is certainly ok.
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [PATCH v3 26/35] target/riscv: Remove shift and slt insn manual decoding
@ 2018-10-31 22:38 ` Richard Henderson
0 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 22:38 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
> static bool trans_slt(DisasContext *ctx, arg_slt *a)
> {
> - gen_arith(ctx, OPC_RISC_SLT, a->rd, a->rs1, a->rs2);
> + TCGv source1 = tcg_temp_new();
> + TCGv source2 = tcg_temp_new();
> +
> + gen_get_gpr(source1, a->rs1);
> + gen_get_gpr(source2, a->rs2);
> +
> + tcg_gen_setcond_tl(TCG_COND_LT, source1, source1, source2);
I do wonder about extracting this one line to gen_slt so that you can re-use
gen_arith and gen_arithi.
> + tcg_gen_setcond_tl(TCG_COND_LTU, source1, source1, source2);
Similarly.
> static bool trans_sllw(DisasContext *ctx, arg_sllw *a)
> {
> - gen_arith(ctx, OPC_RISC_SLLW, a->rd, a->rs1, a->rs2);
> + TCGv source1 = tcg_temp_new();
> + TCGv source2 = tcg_temp_new();
> +
> + gen_get_gpr(source1, a->rs1);
> + gen_get_gpr(source2, a->rs2);
> +
> + tcg_gen_andi_tl(source2, source2, 0x1F);
> + tcg_gen_shl_tl(source1, source1, source2);
> +
> + gen_set_gpr(a->rd, source1);
Missing the ext32s after the shift.
> static bool trans_srlw(DisasContext *ctx, arg_srlw *a)
> {
> - gen_arith(ctx, OPC_RISC_SRLW, a->rd, a->rs1, a->rs2);
> + TCGv source1 = tcg_temp_new();
> + TCGv source2 = tcg_temp_new();
> +
> + gen_get_gpr(source1, a->rs1);
> + gen_get_gpr(source2, a->rs2);
> +
> + /* clear upper 32 */
> + tcg_gen_ext32u_tl(source1, source1);
> + tcg_gen_andi_tl(source2, source2, 0x1F);
> + tcg_gen_shr_tl(source1, source1, source2);
Likewise. (Consider source2 == 0.)
> - case OPC_RISC_SRL:
> - tcg_gen_andi_tl(source2, source2, TARGET_LONG_BITS - 1);
> - tcg_gen_shr_tl(source1, source1, source2);
> - break;
...
> - case OPC_RISC_SRA:
> - tcg_gen_andi_tl(source2, source2, TARGET_LONG_BITS - 1);
> - tcg_gen_sar_tl(source1, source1, source2);
> - break;
I see that the bugs are in the original though, so fixing them in a separate
patch is certainly ok.
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 27/35] target/riscv: Remove manual decoding of RV32/64M insn
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 22:39 ` Richard Henderson
-1 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 22:39 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
> ---
> v2 -> v3:
> - trans_mulw now uses gen_mulw which properly sign extends the result
> - gen_arith_w -> gen_arith_div_w
> - gen_arith_div_w properly sign extends the resul
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [PATCH v3 27/35] target/riscv: Remove manual decoding of RV32/64M insn
@ 2018-10-31 22:39 ` Richard Henderson
0 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 22:39 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> Signed-off-by: Peer Adelt <peer.adelt@hni.uni-paderborn.de>
> ---
> v2 -> v3:
> - trans_mulw now uses gen_mulw which properly sign extends the result
> - gen_arith_w -> gen_arith_div_w
> - gen_arith_div_w properly sign extends the resul
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 31/35] target/riscv: Convert @cs_2 insns to share translation functions
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 22:42 ` Richard Henderson
-1 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 22:42 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
> These all expand simply to R format instructions.
>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> ---
> target/riscv/Makefile.objs | 10 +++---
> target/riscv/insn16-64.decode | 24 ++++++++++++++
> target/riscv/insn16.decode | 15 +++++----
> target/riscv/insn_trans/trans_rvc.inc.c | 44 -------------------------
> target/riscv/translate.c | 20 ++++++++---
> 5 files changed, 54 insertions(+), 59 deletions(-)
> create mode 100644 target/riscv/insn16-64.decode
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [PATCH v3 31/35] target/riscv: Convert @cs_2 insns to share translation functions
@ 2018-10-31 22:42 ` Richard Henderson
0 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 22:42 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
> These all expand simply to R format instructions.
>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> ---
> target/riscv/Makefile.objs | 10 +++---
> target/riscv/insn16-64.decode | 24 ++++++++++++++
> target/riscv/insn16.decode | 15 +++++----
> target/riscv/insn_trans/trans_rvc.inc.c | 44 -------------------------
> target/riscv/translate.c | 20 ++++++++---
> 5 files changed, 54 insertions(+), 59 deletions(-)
> create mode 100644 target/riscv/insn16-64.decode
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 32/35] target/riscv: Convert @cl_d, @cl_w, @cs_d, @cs_w insns
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 22:43 ` Richard Henderson
-1 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 22:43 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> ---
> target/riscv/insn16.decode | 20 ++++++++++----------
> target/riscv/insn32.decode | 3 ++-
> target/riscv/insn_trans/trans_rvc.inc.c | 24 ------------------------
> 3 files changed, 12 insertions(+), 35 deletions(-)
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [PATCH v3 32/35] target/riscv: Convert @cl_d, @cl_w, @cs_d, @cs_w insns
@ 2018-10-31 22:43 ` Richard Henderson
0 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 22:43 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> ---
> target/riscv/insn16.decode | 20 ++++++++++----------
> target/riscv/insn32.decode | 3 ++-
> target/riscv/insn_trans/trans_rvc.inc.c | 24 ------------------------
> 3 files changed, 12 insertions(+), 35 deletions(-)
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 33/35] target/riscv: Splice fsw_sd and flw_ld for riscv32 vs riscv64
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 22:45 ` Richard Henderson
-1 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 22:45 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
> this finally removes the old decoder functions that we carried along
> with it.
>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> ---
> target/riscv/Makefile.objs | 1 +
> target/riscv/insn16-32.decode | 24 ++++
> target/riscv/insn16-64.decode | 4 +
> target/riscv/insn16.decode | 7 +-
> target/riscv/insn_trans/trans_rvc.inc.c | 22 ----
> target/riscv/translate.c | 165 +-----------------------
> 6 files changed, 31 insertions(+), 192 deletions(-)
> create mode 100644 target/riscv/insn16-32.decode
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [PATCH v3 33/35] target/riscv: Splice fsw_sd and flw_ld for riscv32 vs riscv64
@ 2018-10-31 22:45 ` Richard Henderson
0 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 22:45 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
> this finally removes the old decoder functions that we carried along
> with it.
>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> ---
> target/riscv/Makefile.objs | 1 +
> target/riscv/insn16-32.decode | 24 ++++
> target/riscv/insn16-64.decode | 4 +
> target/riscv/insn16.decode | 7 +-
> target/riscv/insn_trans/trans_rvc.inc.c | 22 ----
> target/riscv/translate.c | 165 +-----------------------
> 6 files changed, 31 insertions(+), 192 deletions(-)
> create mode 100644 target/riscv/insn16-32.decode
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 34/35] target/riscv: Splice remaining compressed insn pairs for riscv32 vs riscv64
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 22:47 ` Richard Henderson
-1 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 22:47 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
> it splices flwsp_ldsp, fswsp_sdsp, and jal_addiw and makes each of them
> reuse the code generator used for the non compressed insns.
>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> ---
> target/riscv/insn16-32.decode | 7 +++++
> target/riscv/insn16-64.decode | 5 ++++
> target/riscv/insn16.decode | 12 ++------
> target/riscv/insn32.decode | 3 +-
> target/riscv/insn_trans/trans_rvc.inc.c | 40 -------------------------
> 5 files changed, 16 insertions(+), 51 deletions(-)
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [PATCH v3 34/35] target/riscv: Splice remaining compressed insn pairs for riscv32 vs riscv64
@ 2018-10-31 22:47 ` Richard Henderson
0 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 22:47 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
> it splices flwsp_ldsp, fswsp_sdsp, and jal_addiw and makes each of them
> reuse the code generator used for the non compressed insns.
>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> ---
> target/riscv/insn16-32.decode | 7 +++++
> target/riscv/insn16-64.decode | 5 ++++
> target/riscv/insn16.decode | 12 ++------
> target/riscv/insn32.decode | 3 +-
> target/riscv/insn_trans/trans_rvc.inc.c | 40 -------------------------
> 5 files changed, 16 insertions(+), 51 deletions(-)
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 35/35] target/riscv: Remaining rvc insn reuse 32 bit translators
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-10-31 22:49 ` Richard Henderson
-1 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 22:49 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
> only a translate functions of rvc need to handle special cases. For
> the other rvc instruction we can remove the extra layer of indirection.
>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> ---
> target/riscv/insn16.decode | 37 +++++++++----------
> target/riscv/insn_trans/trans_rvc.inc.c | 48 -------------------------
> 2 files changed, 17 insertions(+), 68 deletions(-)
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [PATCH v3 35/35] target/riscv: Remaining rvc insn reuse 32 bit translators
@ 2018-10-31 22:49 ` Richard Henderson
0 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2018-10-31 22:49 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-devel, qemu-riscv
On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
> only a translate functions of rvc need to handle special cases. For
> the other rvc instruction we can remove the extra layer of indirection.
>
> Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
> ---
> target/riscv/insn16.decode | 37 +++++++++----------
> target/riscv/insn_trans/trans_rvc.inc.c | 48 -------------------------
> 2 files changed, 17 insertions(+), 68 deletions(-)
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 26/35] target/riscv: Remove shift and slt insn manual decoding
2018-10-31 22:38 ` [Qemu-riscv] " Richard Henderson
@ 2018-11-01 15:59 ` Palmer Dabbelt
-1 siblings, 0 replies; 140+ messages in thread
From: Palmer Dabbelt @ 2018-11-01 15:59 UTC (permalink / raw)
To: richard.henderson
Cc: Bastian Koppelmann, Michael Clark, sagark, Alistair Francis,
peer.adelt, qemu-devel, qemu-riscv
On Wed, 31 Oct 2018 15:38:08 PDT (-0700), richard.henderson@linaro.org wrote:
> On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
>> static bool trans_slt(DisasContext *ctx, arg_slt *a)
>> {
>> - gen_arith(ctx, OPC_RISC_SLT, a->rd, a->rs1, a->rs2);
>> + TCGv source1 = tcg_temp_new();
>> + TCGv source2 = tcg_temp_new();
>> +
>> + gen_get_gpr(source1, a->rs1);
>> + gen_get_gpr(source2, a->rs2);
>> +
>> + tcg_gen_setcond_tl(TCG_COND_LT, source1, source1, source2);
>
> I do wonder about extracting this one line to gen_slt so that you can re-use
> gen_arith and gen_arithi.
>
>> + tcg_gen_setcond_tl(TCG_COND_LTU, source1, source1, source2);
>
> Similarly.
>
>> static bool trans_sllw(DisasContext *ctx, arg_sllw *a)
>> {
>> - gen_arith(ctx, OPC_RISC_SLLW, a->rd, a->rs1, a->rs2);
>> + TCGv source1 = tcg_temp_new();
>> + TCGv source2 = tcg_temp_new();
>> +
>> + gen_get_gpr(source1, a->rs1);
>> + gen_get_gpr(source2, a->rs2);
>> +
>> + tcg_gen_andi_tl(source2, source2, 0x1F);
>> + tcg_gen_shl_tl(source1, source1, source2);
>> +
>> + gen_set_gpr(a->rd, source1);
>
> Missing the ext32s after the shift.
>
>> static bool trans_srlw(DisasContext *ctx, arg_srlw *a)
>> {
>> - gen_arith(ctx, OPC_RISC_SRLW, a->rd, a->rs1, a->rs2);
>> + TCGv source1 = tcg_temp_new();
>> + TCGv source2 = tcg_temp_new();
>> +
>> + gen_get_gpr(source1, a->rs1);
>> + gen_get_gpr(source2, a->rs2);
>> +
>> + /* clear upper 32 */
>> + tcg_gen_ext32u_tl(source1, source1);
>> + tcg_gen_andi_tl(source2, source2, 0x1F);
>> + tcg_gen_shr_tl(source1, source1, source2);
>
> Likewise. (Consider source2 == 0.)
>
>> - case OPC_RISC_SRL:
>> - tcg_gen_andi_tl(source2, source2, TARGET_LONG_BITS - 1);
>> - tcg_gen_shr_tl(source1, source1, source2);
>> - break;
> ...
>> - case OPC_RISC_SRA:
>> - tcg_gen_andi_tl(source2, source2, TARGET_LONG_BITS - 1);
>> - tcg_gen_sar_tl(source1, source1, source2);
>> - break;
>
> I see that the bugs are in the original though, so fixing them in a separate
> patch is certainly ok.
Since we're in the soft freeze now, I think it would be great to get the bug
fixes broken out as separate patches at the start of this series that we can
pick up for this release. It would be great if the decodetree conversion was a
non-functional change, as that will make it easier to review.
Sorry for adding a bunch of work for you, Bastian :)
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [PATCH v3 26/35] target/riscv: Remove shift and slt insn manual decoding
@ 2018-11-01 15:59 ` Palmer Dabbelt
0 siblings, 0 replies; 140+ messages in thread
From: Palmer Dabbelt @ 2018-11-01 15:59 UTC (permalink / raw)
To: richard.henderson
Cc: Bastian Koppelmann, Michael Clark, sagark, Alistair Francis,
peer.adelt, qemu-devel, qemu-riscv
On Wed, 31 Oct 2018 15:38:08 PDT (-0700), richard.henderson@linaro.org wrote:
> On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
>> static bool trans_slt(DisasContext *ctx, arg_slt *a)
>> {
>> - gen_arith(ctx, OPC_RISC_SLT, a->rd, a->rs1, a->rs2);
>> + TCGv source1 = tcg_temp_new();
>> + TCGv source2 = tcg_temp_new();
>> +
>> + gen_get_gpr(source1, a->rs1);
>> + gen_get_gpr(source2, a->rs2);
>> +
>> + tcg_gen_setcond_tl(TCG_COND_LT, source1, source1, source2);
>
> I do wonder about extracting this one line to gen_slt so that you can re-use
> gen_arith and gen_arithi.
>
>> + tcg_gen_setcond_tl(TCG_COND_LTU, source1, source1, source2);
>
> Similarly.
>
>> static bool trans_sllw(DisasContext *ctx, arg_sllw *a)
>> {
>> - gen_arith(ctx, OPC_RISC_SLLW, a->rd, a->rs1, a->rs2);
>> + TCGv source1 = tcg_temp_new();
>> + TCGv source2 = tcg_temp_new();
>> +
>> + gen_get_gpr(source1, a->rs1);
>> + gen_get_gpr(source2, a->rs2);
>> +
>> + tcg_gen_andi_tl(source2, source2, 0x1F);
>> + tcg_gen_shl_tl(source1, source1, source2);
>> +
>> + gen_set_gpr(a->rd, source1);
>
> Missing the ext32s after the shift.
>
>> static bool trans_srlw(DisasContext *ctx, arg_srlw *a)
>> {
>> - gen_arith(ctx, OPC_RISC_SRLW, a->rd, a->rs1, a->rs2);
>> + TCGv source1 = tcg_temp_new();
>> + TCGv source2 = tcg_temp_new();
>> +
>> + gen_get_gpr(source1, a->rs1);
>> + gen_get_gpr(source2, a->rs2);
>> +
>> + /* clear upper 32 */
>> + tcg_gen_ext32u_tl(source1, source1);
>> + tcg_gen_andi_tl(source2, source2, 0x1F);
>> + tcg_gen_shr_tl(source1, source1, source2);
>
> Likewise. (Consider source2 == 0.)
>
>> - case OPC_RISC_SRL:
>> - tcg_gen_andi_tl(source2, source2, TARGET_LONG_BITS - 1);
>> - tcg_gen_shr_tl(source1, source1, source2);
>> - break;
> ...
>> - case OPC_RISC_SRA:
>> - tcg_gen_andi_tl(source2, source2, TARGET_LONG_BITS - 1);
>> - tcg_gen_sar_tl(source1, source1, source2);
>> - break;
>
> I see that the bugs are in the original though, so fixing them in a separate
> patch is certainly ok.
Since we're in the soft freeze now, I think it would be great to get the bug
fixes broken out as separate patches at the start of this series that we can
pick up for this release. It would be great if the decodetree conversion was a
non-functional change, as that will make it easier to review.
Sorry for adding a bunch of work for you, Bastian :)
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 00/35] target/riscv: Convert to decodetree
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-11-02 8:48 ` no-reply
-1 siblings, 0 replies; 140+ messages in thread
From: no-reply @ 2018-11-02 8:48 UTC (permalink / raw)
To: kbastian
Cc: famz, mjc, sagark, palmer, Alistair.Francis, qemu-riscv,
peer.adelt, richard.henderson, qemu-devel
Hi,
This series seems to have some coding style problems. See output below for
more information:
Type: series
Message-id: 20181031132029.4887-1-kbastian@mail.uni-paderborn.de
Subject: [Qemu-devel] [PATCH v3 00/35] target/riscv: Convert to decodetree
=== TEST SCRIPT BEGIN ===
#!/bin/bash
BASE=base
n=1
total=$(git log --oneline $BASE.. | wc -l)
failed=0
git config --local diff.renamelimit 0
git config --local diff.renames True
git config --local diff.algorithm histogram
commits="$(git log --format=%H --reverse $BASE..)"
for c in $commits; do
echo "Checking PATCH $n/$total: $(git log -n 1 --format=%s $c)..."
if ! git show $c --format=email | ./scripts/checkpatch.pl --mailback -; then
failed=1
echo
fi
n=$((n+1))
done
exit $failed
=== TEST SCRIPT END ===
Updating 3c8cf5a9c21ff8782164d1def7f44bd888713384
Switched to a new branch 'test'
0dc9aaa65c target/riscv: Remaining rvc insn reuse 32 bit translators
8d508e3f89 target/riscv: Splice remaining compressed insn pairs for riscv32 vs riscv64
44e80c8498 target/riscv: Splice fsw_sd and flw_ld for riscv32 vs riscv64
07f644d3a2 target/riscv: Convert @cl_d, @cl_w, @cs_d, @cs_w insns
40b6158d54 target/riscv: Convert @cs_2 insns to share translation functions
ca95ca92a5 target/riscv: Remove decode_RV32_64G()
666c42149b target/riscv: Remove gen_system()
e1892386ec target/riscv: Rename trans_arith to gen_arith
1b33908223 target/riscv: Remove manual decoding of RV32/64M insn
1a5fef8c4e target/riscv: Remove shift and slt insn manual decoding
2731f425de target/riscv: make ADD/SUB/OR/XOR/AND insn use arg lists
9c34cbd280 target/riscv: Move gen_arith_imm() decoding into trans_* functions
2eaf68ef78 target/riscv: Remove manual decoding from gen_store()
31f0d4c241 target/riscv: Remove manual decoding from gen_load()
9d725fe480 target/riscv: Remove manual decoding from gen_branch()
9d8e1f844c target/riscv: Remove gen_jalr()
dc3753ad33 target/riscv: Convert quadrant 2 of RVXC insns to decodetree
1a083b100e target/riscv: Convert quadrant 1 of RVXC insns to decodetree
b831b8c4ec target/riscv: Convert quadrant 0 of RVXC insns to decodetree
c931914bd2 target/riscv: Convert RV priv insns to decodetree
767d172f2d target/riscv: Convert RV64D insns to decodetree
18d3406ee6 target/riscv: Convert RV32D insns to decodetree
515c07059a target/riscv: Convert RV64F insns to decodetree
3506dbc10d target/riscv: Convert RV32F insns to decodetree
2197c6b6b1 target/riscv: Convert RV64A insns to decodetree
c07ae6248a target/riscv: Convert RV32A insns to decodetree
508bf988f6 target/riscv: Convert RVXM insns to decodetree
2773626354 target/riscv: Convert RVXI csr insns to decodetree
55033facbb target/riscv: Convert RVXI fence insns to decodetree
e4d5253c9f target/riscv: Convert RVXI arithmetic insns to decodetree
7616bcfe80 target/riscv: Convert RV64I load/store insns to decodetree
7a348ac675 target/riscv: Convert RV32I load/store insns to decodetree
68f5d736d1 target/riscv: Convert RVXI branch insns to decodetree
2b6fed94f8 target/riscv: Activate decodetree and implemnt LUI & AUIPC
5fa58ba15a target/riscv: Move CPURISCVState pointer to DisasContext
=== OUTPUT BEGIN ===
Checking PATCH 1/35: target/riscv: Move CPURISCVState pointer to DisasContext...
Checking PATCH 2/35: target/riscv: Activate decodetree and implemnt LUI & AUIPC...
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating?
#32:
new file mode 100644
ERROR: externs should be avoided in .c files
#123: FILE: target/riscv/translate.c:1677:
+bool decode_insn32(DisasContext *ctx, uint32_t insn);
total: 1 errors, 1 warnings, 125 lines checked
Your patch has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
Checking PATCH 3/35: target/riscv: Convert RVXI branch insns to decodetree...
Checking PATCH 4/35: target/riscv: Convert RV32I load/store insns to decodetree...
Checking PATCH 5/35: target/riscv: Convert RV64I load/store insns to decodetree...
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating?
#37:
new file mode 100644
total: 0 errors, 1 warnings, 76 lines checked
Your patch has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
Checking PATCH 6/35: target/riscv: Convert RVXI arithmetic insns to decodetree...
Checking PATCH 7/35: target/riscv: Convert RVXI fence insns to decodetree...
Checking PATCH 8/35: target/riscv: Convert RVXI csr insns to decodetree...
Checking PATCH 9/35: target/riscv: Convert RVXM insns to decodetree...
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating?
#46:
new file mode 100644
total: 0 errors, 1 warnings, 145 lines checked
Your patch has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
Checking PATCH 10/35: target/riscv: Convert RV32A insns to decodetree...
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating?
#51:
new file mode 100644
total: 0 errors, 1 warnings, 184 lines checked
Your patch has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
Checking PATCH 11/35: target/riscv: Convert RV64A insns to decodetree...
Checking PATCH 12/35: target/riscv: Convert RV32F insns to decodetree...
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating?
#75:
new file mode 100644
total: 0 errors, 1 warnings, 397 lines checked
Your patch has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
Checking PATCH 13/35: target/riscv: Convert RV64F insns to decodetree...
Checking PATCH 14/35: target/riscv: Convert RV32D insns to decodetree...
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating?
#48:
new file mode 100644
total: 0 errors, 1 warnings, 353 lines checked
Your patch has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
Checking PATCH 15/35: target/riscv: Convert RV64D insns to decodetree...
Checking PATCH 16/35: target/riscv: Convert RV priv insns to decodetree...
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating?
#38:
new file mode 100644
total: 0 errors, 1 warnings, 204 lines checked
Your patch has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
Checking PATCH 17/35: target/riscv: Convert quadrant 0 of RVXC insns to decodetree...
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating?
#29:
new file mode 100644
ERROR: externs should be avoided in .c files
#244: FILE: target/riscv/translate.c:984:
+bool decode_insn16(DisasContext *ctx, uint16_t insn);
total: 1 errors, 1 warnings, 227 lines checked
Your patch has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
Checking PATCH 18/35: target/riscv: Convert quadrant 1 of RVXC insns to decodetree...
Checking PATCH 19/35: target/riscv: Convert quadrant 2 of RVXC insns to decodetree...
Checking PATCH 20/35: target/riscv: Remove gen_jalr()...
Checking PATCH 21/35: target/riscv: Remove manual decoding from gen_branch()...
Checking PATCH 22/35: target/riscv: Remove manual decoding from gen_load()...
Checking PATCH 23/35: target/riscv: Remove manual decoding from gen_store()...
Checking PATCH 24/35: target/riscv: Move gen_arith_imm() decoding into trans_* functions...
Checking PATCH 25/35: target/riscv: make ADD/SUB/OR/XOR/AND insn use arg lists...
Checking PATCH 26/35: target/riscv: Remove shift and slt insn manual decoding...
Checking PATCH 27/35: target/riscv: Remove manual decoding of RV32/64M insn...
Checking PATCH 28/35: target/riscv: Rename trans_arith to gen_arith...
Checking PATCH 29/35: target/riscv: Remove gen_system()...
Checking PATCH 30/35: target/riscv: Remove decode_RV32_64G()...
Checking PATCH 31/35: target/riscv: Convert @cs_2 insns to share translation functions...
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating?
#40:
new file mode 100644
ERROR: externs should be avoided in .c files
#180: FILE: target/riscv/translate.c:494:
+bool decode_insn16(DisasContext *ctx, uint16_t insn);
total: 1 errors, 1 warnings, 164 lines checked
Your patch has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
Checking PATCH 32/35: target/riscv: Convert @cl_d, @cl_w, @cs_d, @cs_w insns...
Checking PATCH 33/35: target/riscv: Splice fsw_sd and flw_ld for riscv32 vs riscv64...
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating?
#25:
new file mode 100644
total: 0 errors, 1 warnings, 287 lines checked
Your patch has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
Checking PATCH 34/35: target/riscv: Splice remaining compressed insn pairs for riscv32 vs riscv64...
Checking PATCH 35/35: target/riscv: Remaining rvc insn reuse 32 bit translators...
=== OUTPUT END ===
Test command exited with code: 1
---
Email generated automatically by Patchew [http://patchew.org/].
Please send your feedback to patchew-devel@redhat.com
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [Qemu-devel] [PATCH v3 00/35] target/riscv: Convert to decodetree
@ 2018-11-02 8:48 ` no-reply
0 siblings, 0 replies; 140+ messages in thread
From: no-reply @ 2018-11-02 8:48 UTC (permalink / raw)
To: kbastian
Cc: famz, mjc, sagark, palmer, kbastian, Alistair.Francis,
qemu-riscv, peer.adelt, richard.henderson, qemu-devel
Hi,
This series seems to have some coding style problems. See output below for
more information:
Type: series
Message-id: 20181031132029.4887-1-kbastian@mail.uni-paderborn.de
Subject: [Qemu-devel] [PATCH v3 00/35] target/riscv: Convert to decodetree
=== TEST SCRIPT BEGIN ===
#!/bin/bash
BASE=base
n=1
total=$(git log --oneline $BASE.. | wc -l)
failed=0
git config --local diff.renamelimit 0
git config --local diff.renames True
git config --local diff.algorithm histogram
commits="$(git log --format=%H --reverse $BASE..)"
for c in $commits; do
echo "Checking PATCH $n/$total: $(git log -n 1 --format=%s $c)..."
if ! git show $c --format=email | ./scripts/checkpatch.pl --mailback -; then
failed=1
echo
fi
n=$((n+1))
done
exit $failed
=== TEST SCRIPT END ===
Updating 3c8cf5a9c21ff8782164d1def7f44bd888713384
Switched to a new branch 'test'
0dc9aaa65c target/riscv: Remaining rvc insn reuse 32 bit translators
8d508e3f89 target/riscv: Splice remaining compressed insn pairs for riscv32 vs riscv64
44e80c8498 target/riscv: Splice fsw_sd and flw_ld for riscv32 vs riscv64
07f644d3a2 target/riscv: Convert @cl_d, @cl_w, @cs_d, @cs_w insns
40b6158d54 target/riscv: Convert @cs_2 insns to share translation functions
ca95ca92a5 target/riscv: Remove decode_RV32_64G()
666c42149b target/riscv: Remove gen_system()
e1892386ec target/riscv: Rename trans_arith to gen_arith
1b33908223 target/riscv: Remove manual decoding of RV32/64M insn
1a5fef8c4e target/riscv: Remove shift and slt insn manual decoding
2731f425de target/riscv: make ADD/SUB/OR/XOR/AND insn use arg lists
9c34cbd280 target/riscv: Move gen_arith_imm() decoding into trans_* functions
2eaf68ef78 target/riscv: Remove manual decoding from gen_store()
31f0d4c241 target/riscv: Remove manual decoding from gen_load()
9d725fe480 target/riscv: Remove manual decoding from gen_branch()
9d8e1f844c target/riscv: Remove gen_jalr()
dc3753ad33 target/riscv: Convert quadrant 2 of RVXC insns to decodetree
1a083b100e target/riscv: Convert quadrant 1 of RVXC insns to decodetree
b831b8c4ec target/riscv: Convert quadrant 0 of RVXC insns to decodetree
c931914bd2 target/riscv: Convert RV priv insns to decodetree
767d172f2d target/riscv: Convert RV64D insns to decodetree
18d3406ee6 target/riscv: Convert RV32D insns to decodetree
515c07059a target/riscv: Convert RV64F insns to decodetree
3506dbc10d target/riscv: Convert RV32F insns to decodetree
2197c6b6b1 target/riscv: Convert RV64A insns to decodetree
c07ae6248a target/riscv: Convert RV32A insns to decodetree
508bf988f6 target/riscv: Convert RVXM insns to decodetree
2773626354 target/riscv: Convert RVXI csr insns to decodetree
55033facbb target/riscv: Convert RVXI fence insns to decodetree
e4d5253c9f target/riscv: Convert RVXI arithmetic insns to decodetree
7616bcfe80 target/riscv: Convert RV64I load/store insns to decodetree
7a348ac675 target/riscv: Convert RV32I load/store insns to decodetree
68f5d736d1 target/riscv: Convert RVXI branch insns to decodetree
2b6fed94f8 target/riscv: Activate decodetree and implemnt LUI & AUIPC
5fa58ba15a target/riscv: Move CPURISCVState pointer to DisasContext
=== OUTPUT BEGIN ===
Checking PATCH 1/35: target/riscv: Move CPURISCVState pointer to DisasContext...
Checking PATCH 2/35: target/riscv: Activate decodetree and implemnt LUI & AUIPC...
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating?
#32:
new file mode 100644
ERROR: externs should be avoided in .c files
#123: FILE: target/riscv/translate.c:1677:
+bool decode_insn32(DisasContext *ctx, uint32_t insn);
total: 1 errors, 1 warnings, 125 lines checked
Your patch has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
Checking PATCH 3/35: target/riscv: Convert RVXI branch insns to decodetree...
Checking PATCH 4/35: target/riscv: Convert RV32I load/store insns to decodetree...
Checking PATCH 5/35: target/riscv: Convert RV64I load/store insns to decodetree...
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating?
#37:
new file mode 100644
total: 0 errors, 1 warnings, 76 lines checked
Your patch has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
Checking PATCH 6/35: target/riscv: Convert RVXI arithmetic insns to decodetree...
Checking PATCH 7/35: target/riscv: Convert RVXI fence insns to decodetree...
Checking PATCH 8/35: target/riscv: Convert RVXI csr insns to decodetree...
Checking PATCH 9/35: target/riscv: Convert RVXM insns to decodetree...
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating?
#46:
new file mode 100644
total: 0 errors, 1 warnings, 145 lines checked
Your patch has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
Checking PATCH 10/35: target/riscv: Convert RV32A insns to decodetree...
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating?
#51:
new file mode 100644
total: 0 errors, 1 warnings, 184 lines checked
Your patch has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
Checking PATCH 11/35: target/riscv: Convert RV64A insns to decodetree...
Checking PATCH 12/35: target/riscv: Convert RV32F insns to decodetree...
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating?
#75:
new file mode 100644
total: 0 errors, 1 warnings, 397 lines checked
Your patch has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
Checking PATCH 13/35: target/riscv: Convert RV64F insns to decodetree...
Checking PATCH 14/35: target/riscv: Convert RV32D insns to decodetree...
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating?
#48:
new file mode 100644
total: 0 errors, 1 warnings, 353 lines checked
Your patch has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
Checking PATCH 15/35: target/riscv: Convert RV64D insns to decodetree...
Checking PATCH 16/35: target/riscv: Convert RV priv insns to decodetree...
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating?
#38:
new file mode 100644
total: 0 errors, 1 warnings, 204 lines checked
Your patch has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
Checking PATCH 17/35: target/riscv: Convert quadrant 0 of RVXC insns to decodetree...
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating?
#29:
new file mode 100644
ERROR: externs should be avoided in .c files
#244: FILE: target/riscv/translate.c:984:
+bool decode_insn16(DisasContext *ctx, uint16_t insn);
total: 1 errors, 1 warnings, 227 lines checked
Your patch has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
Checking PATCH 18/35: target/riscv: Convert quadrant 1 of RVXC insns to decodetree...
Checking PATCH 19/35: target/riscv: Convert quadrant 2 of RVXC insns to decodetree...
Checking PATCH 20/35: target/riscv: Remove gen_jalr()...
Checking PATCH 21/35: target/riscv: Remove manual decoding from gen_branch()...
Checking PATCH 22/35: target/riscv: Remove manual decoding from gen_load()...
Checking PATCH 23/35: target/riscv: Remove manual decoding from gen_store()...
Checking PATCH 24/35: target/riscv: Move gen_arith_imm() decoding into trans_* functions...
Checking PATCH 25/35: target/riscv: make ADD/SUB/OR/XOR/AND insn use arg lists...
Checking PATCH 26/35: target/riscv: Remove shift and slt insn manual decoding...
Checking PATCH 27/35: target/riscv: Remove manual decoding of RV32/64M insn...
Checking PATCH 28/35: target/riscv: Rename trans_arith to gen_arith...
Checking PATCH 29/35: target/riscv: Remove gen_system()...
Checking PATCH 30/35: target/riscv: Remove decode_RV32_64G()...
Checking PATCH 31/35: target/riscv: Convert @cs_2 insns to share translation functions...
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating?
#40:
new file mode 100644
ERROR: externs should be avoided in .c files
#180: FILE: target/riscv/translate.c:494:
+bool decode_insn16(DisasContext *ctx, uint16_t insn);
total: 1 errors, 1 warnings, 164 lines checked
Your patch has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
Checking PATCH 32/35: target/riscv: Convert @cl_d, @cl_w, @cs_d, @cs_w insns...
Checking PATCH 33/35: target/riscv: Splice fsw_sd and flw_ld for riscv32 vs riscv64...
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating?
#25:
new file mode 100644
total: 0 errors, 1 warnings, 287 lines checked
Your patch has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
Checking PATCH 34/35: target/riscv: Splice remaining compressed insn pairs for riscv32 vs riscv64...
Checking PATCH 35/35: target/riscv: Remaining rvc insn reuse 32 bit translators...
=== OUTPUT END ===
Test command exited with code: 1
---
Email generated automatically by Patchew [http://patchew.org/].
Please send your feedback to patchew-devel@redhat.com
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 26/35] target/riscv: Remove shift and slt insn manual decoding
2018-11-01 15:59 ` [Qemu-riscv] " Palmer Dabbelt
@ 2018-11-05 17:00 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-11-05 17:00 UTC (permalink / raw)
To: Palmer Dabbelt, richard.henderson
Cc: qemu-riscv, sagark, qemu-devel, peer.adelt, Alistair Francis,
Michael Clark
On 11/1/18 4:59 PM, Palmer Dabbelt wrote:
> On Wed, 31 Oct 2018 15:38:08 PDT (-0700), richard.henderson@linaro.org
> wrote:
>> On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
>>> static bool trans_slt(DisasContext *ctx, arg_slt *a)
>>> {
>>> - gen_arith(ctx, OPC_RISC_SLT, a->rd, a->rs1, a->rs2);
>>> + TCGv source1 = tcg_temp_new();
>>> + TCGv source2 = tcg_temp_new();
>>> +
>>> + gen_get_gpr(source1, a->rs1);
>>> + gen_get_gpr(source2, a->rs2);
>>> +
>>> + tcg_gen_setcond_tl(TCG_COND_LT, source1, source1, source2);
>>
>> I do wonder about extracting this one line to gen_slt so that you can
>> re-use
>> gen_arith and gen_arithi.
>>
>>> + tcg_gen_setcond_tl(TCG_COND_LTU, source1, source1, source2);
>>
>> Similarly.
>>
>>> static bool trans_sllw(DisasContext *ctx, arg_sllw *a)
>>> {
>>> - gen_arith(ctx, OPC_RISC_SLLW, a->rd, a->rs1, a->rs2);
>>> + TCGv source1 = tcg_temp_new();
>>> + TCGv source2 = tcg_temp_new();
>>> +
>>> + gen_get_gpr(source1, a->rs1);
>>> + gen_get_gpr(source2, a->rs2);
>>> +
>>> + tcg_gen_andi_tl(source2, source2, 0x1F);
>>> + tcg_gen_shl_tl(source1, source1, source2);
>>> +
>>> + gen_set_gpr(a->rd, source1);
>>
>> Missing the ext32s after the shift.
>>
>>> static bool trans_srlw(DisasContext *ctx, arg_srlw *a)
>>> {
>>> - gen_arith(ctx, OPC_RISC_SRLW, a->rd, a->rs1, a->rs2);
>>> + TCGv source1 = tcg_temp_new();
>>> + TCGv source2 = tcg_temp_new();
>>> +
>>> + gen_get_gpr(source1, a->rs1);
>>> + gen_get_gpr(source2, a->rs2);
>>> +
>>> + /* clear upper 32 */
>>> + tcg_gen_ext32u_tl(source1, source1);
>>> + tcg_gen_andi_tl(source2, source2, 0x1F);
>>> + tcg_gen_shr_tl(source1, source1, source2);
>>
>> Likewise. (Consider source2 == 0.)
>>
>>> - case OPC_RISC_SRL:
>>> - tcg_gen_andi_tl(source2, source2, TARGET_LONG_BITS - 1);
>>> - tcg_gen_shr_tl(source1, source1, source2);
>>> - break;
>> ...
>>> - case OPC_RISC_SRA:
>>> - tcg_gen_andi_tl(source2, source2, TARGET_LONG_BITS - 1);
>>> - tcg_gen_sar_tl(source1, source1, source2);
>>> - break;
>>
>> I see that the bugs are in the original though, so fixing them in a
>> separate
>> patch is certainly ok.
>
> Since we're in the soft freeze now, I think it would be great to get
> the bug fixes broken out as separate patches at the start of this
> series that we can pick up for this release. It would be great if the
> decodetree conversion was a non-functional change, as that will make
> it easier to review.
Sounds good. Unfortunately, I don't have much time right now. I try at
least to extract the bugfixes, such that they make it for this release.
Cheers,
Bastian
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [Qemu-devel] [PATCH v3 26/35] target/riscv: Remove shift and slt insn manual decoding
@ 2018-11-05 17:00 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2018-11-05 17:00 UTC (permalink / raw)
To: Palmer Dabbelt, richard.henderson
Cc: qemu-riscv, sagark, qemu-devel, peer.adelt, Alistair Francis,
Michael Clark
On 11/1/18 4:59 PM, Palmer Dabbelt wrote:
> On Wed, 31 Oct 2018 15:38:08 PDT (-0700), richard.henderson@linaro.org
> wrote:
>> On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
>>> static bool trans_slt(DisasContext *ctx, arg_slt *a)
>>> {
>>> - gen_arith(ctx, OPC_RISC_SLT, a->rd, a->rs1, a->rs2);
>>> + TCGv source1 = tcg_temp_new();
>>> + TCGv source2 = tcg_temp_new();
>>> +
>>> + gen_get_gpr(source1, a->rs1);
>>> + gen_get_gpr(source2, a->rs2);
>>> +
>>> + tcg_gen_setcond_tl(TCG_COND_LT, source1, source1, source2);
>>
>> I do wonder about extracting this one line to gen_slt so that you can
>> re-use
>> gen_arith and gen_arithi.
>>
>>> + tcg_gen_setcond_tl(TCG_COND_LTU, source1, source1, source2);
>>
>> Similarly.
>>
>>> static bool trans_sllw(DisasContext *ctx, arg_sllw *a)
>>> {
>>> - gen_arith(ctx, OPC_RISC_SLLW, a->rd, a->rs1, a->rs2);
>>> + TCGv source1 = tcg_temp_new();
>>> + TCGv source2 = tcg_temp_new();
>>> +
>>> + gen_get_gpr(source1, a->rs1);
>>> + gen_get_gpr(source2, a->rs2);
>>> +
>>> + tcg_gen_andi_tl(source2, source2, 0x1F);
>>> + tcg_gen_shl_tl(source1, source1, source2);
>>> +
>>> + gen_set_gpr(a->rd, source1);
>>
>> Missing the ext32s after the shift.
>>
>>> static bool trans_srlw(DisasContext *ctx, arg_srlw *a)
>>> {
>>> - gen_arith(ctx, OPC_RISC_SRLW, a->rd, a->rs1, a->rs2);
>>> + TCGv source1 = tcg_temp_new();
>>> + TCGv source2 = tcg_temp_new();
>>> +
>>> + gen_get_gpr(source1, a->rs1);
>>> + gen_get_gpr(source2, a->rs2);
>>> +
>>> + /* clear upper 32 */
>>> + tcg_gen_ext32u_tl(source1, source1);
>>> + tcg_gen_andi_tl(source2, source2, 0x1F);
>>> + tcg_gen_shr_tl(source1, source1, source2);
>>
>> Likewise. (Consider source2 == 0.)
>>
>>> - case OPC_RISC_SRL:
>>> - tcg_gen_andi_tl(source2, source2, TARGET_LONG_BITS - 1);
>>> - tcg_gen_shr_tl(source1, source1, source2);
>>> - break;
>> ...
>>> - case OPC_RISC_SRA:
>>> - tcg_gen_andi_tl(source2, source2, TARGET_LONG_BITS - 1);
>>> - tcg_gen_sar_tl(source1, source1, source2);
>>> - break;
>>
>> I see that the bugs are in the original though, so fixing them in a
>> separate
>> patch is certainly ok.
>
> Since we're in the soft freeze now, I think it would be great to get
> the bug fixes broken out as separate patches at the start of this
> series that we can pick up for this release. It would be great if the
> decodetree conversion was a non-functional change, as that will make
> it easier to review.
Sounds good. Unfortunately, I don't have much time right now. I try at
least to extract the bugfixes, such that they make it for this release.
Cheers,
Bastian
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 26/35] target/riscv: Remove shift and slt insn manual decoding
2018-11-05 17:00 ` [Qemu-riscv] " Bastian Koppelmann
@ 2018-11-07 0:56 ` Palmer Dabbelt
-1 siblings, 0 replies; 140+ messages in thread
From: Palmer Dabbelt @ 2018-11-07 0:56 UTC (permalink / raw)
To: Bastian Koppelmann
Cc: richard.henderson, qemu-riscv, sagark, qemu-devel, peer.adelt,
Alistair Francis, Michael Clark
On Mon, 05 Nov 2018 09:00:10 PST (-0800), Bastian Koppelmann wrote:
>
> On 11/1/18 4:59 PM, Palmer Dabbelt wrote:
>> On Wed, 31 Oct 2018 15:38:08 PDT (-0700), richard.henderson@linaro.org
>> wrote:
>>> On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
>>>> static bool trans_slt(DisasContext *ctx, arg_slt *a)
>>>> {
>>>> - gen_arith(ctx, OPC_RISC_SLT, a->rd, a->rs1, a->rs2);
>>>> + TCGv source1 = tcg_temp_new();
>>>> + TCGv source2 = tcg_temp_new();
>>>> +
>>>> + gen_get_gpr(source1, a->rs1);
>>>> + gen_get_gpr(source2, a->rs2);
>>>> +
>>>> + tcg_gen_setcond_tl(TCG_COND_LT, source1, source1, source2);
>>>
>>> I do wonder about extracting this one line to gen_slt so that you can
>>> re-use
>>> gen_arith and gen_arithi.
>>>
>>>> + tcg_gen_setcond_tl(TCG_COND_LTU, source1, source1, source2);
>>>
>>> Similarly.
>>>
>>>> static bool trans_sllw(DisasContext *ctx, arg_sllw *a)
>>>> {
>>>> - gen_arith(ctx, OPC_RISC_SLLW, a->rd, a->rs1, a->rs2);
>>>> + TCGv source1 = tcg_temp_new();
>>>> + TCGv source2 = tcg_temp_new();
>>>> +
>>>> + gen_get_gpr(source1, a->rs1);
>>>> + gen_get_gpr(source2, a->rs2);
>>>> +
>>>> + tcg_gen_andi_tl(source2, source2, 0x1F);
>>>> + tcg_gen_shl_tl(source1, source1, source2);
>>>> +
>>>> + gen_set_gpr(a->rd, source1);
>>>
>>> Missing the ext32s after the shift.
>>>
>>>> static bool trans_srlw(DisasContext *ctx, arg_srlw *a)
>>>> {
>>>> - gen_arith(ctx, OPC_RISC_SRLW, a->rd, a->rs1, a->rs2);
>>>> + TCGv source1 = tcg_temp_new();
>>>> + TCGv source2 = tcg_temp_new();
>>>> +
>>>> + gen_get_gpr(source1, a->rs1);
>>>> + gen_get_gpr(source2, a->rs2);
>>>> +
>>>> + /* clear upper 32 */
>>>> + tcg_gen_ext32u_tl(source1, source1);
>>>> + tcg_gen_andi_tl(source2, source2, 0x1F);
>>>> + tcg_gen_shr_tl(source1, source1, source2);
>>>
>>> Likewise. (Consider source2 == 0.)
>>>
>>>> - case OPC_RISC_SRL:
>>>> - tcg_gen_andi_tl(source2, source2, TARGET_LONG_BITS - 1);
>>>> - tcg_gen_shr_tl(source1, source1, source2);
>>>> - break;
>>> ...
>>>> - case OPC_RISC_SRA:
>>>> - tcg_gen_andi_tl(source2, source2, TARGET_LONG_BITS - 1);
>>>> - tcg_gen_sar_tl(source1, source1, source2);
>>>> - break;
>>>
>>> I see that the bugs are in the original though, so fixing them in a
>>> separate
>>> patch is certainly ok.
>>
>> Since we're in the soft freeze now, I think it would be great to get
>> the bug fixes broken out as separate patches at the start of this
>> series that we can pick up for this release. It would be great if the
>> decodetree conversion was a non-functional change, as that will make
>> it easier to review.
>
>
> Sounds good. Unfortunately, I don't have much time right now. I try at
> least to extract the bugfixes, such that they make it for this release.
Thanks. You're welcome to just enumerate the bug fixes so I can dig them out
if that's easier.
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [Qemu-devel] [PATCH v3 26/35] target/riscv: Remove shift and slt insn manual decoding
@ 2018-11-07 0:56 ` Palmer Dabbelt
0 siblings, 0 replies; 140+ messages in thread
From: Palmer Dabbelt @ 2018-11-07 0:56 UTC (permalink / raw)
To: Bastian Koppelmann
Cc: richard.henderson, qemu-riscv, sagark, qemu-devel, peer.adelt,
Alistair Francis, Michael Clark
On Mon, 05 Nov 2018 09:00:10 PST (-0800), Bastian Koppelmann wrote:
>
> On 11/1/18 4:59 PM, Palmer Dabbelt wrote:
>> On Wed, 31 Oct 2018 15:38:08 PDT (-0700), richard.henderson@linaro.org
>> wrote:
>>> On 10/31/18 1:20 PM, Bastian Koppelmann wrote:
>>>> static bool trans_slt(DisasContext *ctx, arg_slt *a)
>>>> {
>>>> - gen_arith(ctx, OPC_RISC_SLT, a->rd, a->rs1, a->rs2);
>>>> + TCGv source1 = tcg_temp_new();
>>>> + TCGv source2 = tcg_temp_new();
>>>> +
>>>> + gen_get_gpr(source1, a->rs1);
>>>> + gen_get_gpr(source2, a->rs2);
>>>> +
>>>> + tcg_gen_setcond_tl(TCG_COND_LT, source1, source1, source2);
>>>
>>> I do wonder about extracting this one line to gen_slt so that you can
>>> re-use
>>> gen_arith and gen_arithi.
>>>
>>>> + tcg_gen_setcond_tl(TCG_COND_LTU, source1, source1, source2);
>>>
>>> Similarly.
>>>
>>>> static bool trans_sllw(DisasContext *ctx, arg_sllw *a)
>>>> {
>>>> - gen_arith(ctx, OPC_RISC_SLLW, a->rd, a->rs1, a->rs2);
>>>> + TCGv source1 = tcg_temp_new();
>>>> + TCGv source2 = tcg_temp_new();
>>>> +
>>>> + gen_get_gpr(source1, a->rs1);
>>>> + gen_get_gpr(source2, a->rs2);
>>>> +
>>>> + tcg_gen_andi_tl(source2, source2, 0x1F);
>>>> + tcg_gen_shl_tl(source1, source1, source2);
>>>> +
>>>> + gen_set_gpr(a->rd, source1);
>>>
>>> Missing the ext32s after the shift.
>>>
>>>> static bool trans_srlw(DisasContext *ctx, arg_srlw *a)
>>>> {
>>>> - gen_arith(ctx, OPC_RISC_SRLW, a->rd, a->rs1, a->rs2);
>>>> + TCGv source1 = tcg_temp_new();
>>>> + TCGv source2 = tcg_temp_new();
>>>> +
>>>> + gen_get_gpr(source1, a->rs1);
>>>> + gen_get_gpr(source2, a->rs2);
>>>> +
>>>> + /* clear upper 32 */
>>>> + tcg_gen_ext32u_tl(source1, source1);
>>>> + tcg_gen_andi_tl(source2, source2, 0x1F);
>>>> + tcg_gen_shr_tl(source1, source1, source2);
>>>
>>> Likewise. (Consider source2 == 0.)
>>>
>>>> - case OPC_RISC_SRL:
>>>> - tcg_gen_andi_tl(source2, source2, TARGET_LONG_BITS - 1);
>>>> - tcg_gen_shr_tl(source1, source1, source2);
>>>> - break;
>>> ...
>>>> - case OPC_RISC_SRA:
>>>> - tcg_gen_andi_tl(source2, source2, TARGET_LONG_BITS - 1);
>>>> - tcg_gen_sar_tl(source1, source1, source2);
>>>> - break;
>>>
>>> I see that the bugs are in the original though, so fixing them in a
>>> separate
>>> patch is certainly ok.
>>
>> Since we're in the soft freeze now, I think it would be great to get
>> the bug fixes broken out as separate patches at the start of this
>> series that we can pick up for this release. It would be great if the
>> decodetree conversion was a non-functional change, as that will make
>> it easier to review.
>
>
> Sounds good. Unfortunately, I don't have much time right now. I try at
> least to extract the bugfixes, such that they make it for this release.
Thanks. You're welcome to just enumerate the bug fixes so I can dig them out
if that's easier.
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 24/35] target/riscv: Move gen_arith_imm() decoding into trans_* functions
2018-10-31 22:18 ` [Qemu-riscv] " Richard Henderson
@ 2019-01-11 13:10 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2019-01-11 13:10 UTC (permalink / raw)
To: Richard Henderson, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-riscv, qemu-devel
On 10/31/18 11:18 PM, Richard Henderson wrote:
>
> Surely the shri and sari functions need the same shamt >= TARGET_LONG_BITS
> check as slli. Otherwise RV32 shri should definitely produce an assert in
> tcg_gen_shri_tl.
>
> I did wonder about changing the decode of the shift functions such that only
> the top two bits of the imm are reserved for secondary parsing of the shift
> type, and the other 10 bits are passed down into trans_foo. At which point the
> TARGET_LONG_BITS check takes care of other illegalities.
>
> Which means that the parsing for slli and slliw are identical, and also that
> for the far future when RV128 is a thing, we don't have to change the parsing.
I don't quite understand this. Do you want to have one entry in the
decode file for slli and slliw?
How is the parsing of slli and slliw identical with this change? As far
as I see it, they are different at least in the opcode.
Cheers,
Bastian
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [Qemu-devel] [PATCH v3 24/35] target/riscv: Move gen_arith_imm() decoding into trans_* functions
@ 2019-01-11 13:10 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2019-01-11 13:10 UTC (permalink / raw)
To: Richard Henderson, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-riscv, qemu-devel
On 10/31/18 11:18 PM, Richard Henderson wrote:
>
> Surely the shri and sari functions need the same shamt >= TARGET_LONG_BITS
> check as slli. Otherwise RV32 shri should definitely produce an assert in
> tcg_gen_shri_tl.
>
> I did wonder about changing the decode of the shift functions such that only
> the top two bits of the imm are reserved for secondary parsing of the shift
> type, and the other 10 bits are passed down into trans_foo. At which point the
> TARGET_LONG_BITS check takes care of other illegalities.
>
> Which means that the parsing for slli and slliw are identical, and also that
> for the far future when RV128 is a thing, we don't have to change the parsing.
I don't quite understand this. Do you want to have one entry in the
decode file for slli and slliw?
How is the parsing of slli and slliw identical with this change? As far
as I see it, they are different at least in the opcode.
Cheers,
Bastian
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 24/35] target/riscv: Move gen_arith_imm() decoding into trans_* functions
2019-01-11 13:10 ` [Qemu-riscv] " Bastian Koppelmann
@ 2019-01-11 21:00 ` Richard Henderson
-1 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2019-01-11 21:00 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-riscv, qemu-devel
On 1/12/19 12:10 AM, Bastian Koppelmann wrote:
>
> On 10/31/18 11:18 PM, Richard Henderson wrote:
>>
>> Surely the shri and sari functions need the same shamt >= TARGET_LONG_BITS
>> check as slli. Otherwise RV32 shri should definitely produce an assert in
>> tcg_gen_shri_tl.
>>
>> I did wonder about changing the decode of the shift functions such that only
>> the top two bits of the imm are reserved for secondary parsing of the shift
>> type, and the other 10 bits are passed down into trans_foo. At which point the
>> TARGET_LONG_BITS check takes care of other illegalities.
>>
>> Which means that the parsing for slli and slliw are identical, and also that
>> for the far future when RV128 is a thing, we don't have to change the parsing.
>
>
> I don't quite understand this. Do you want to have one entry in the decode file
> for slli and slliw?
>
> How is the parsing of slli and slliw identical with this change? As far as I
> see it, they are different at least in the opcode.
I meant in the extraction and validation of operands, I think.
I'm not really sure where else I was going with this. It has
been two months and I don't have the decode in front of me.
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [Qemu-devel] [PATCH v3 24/35] target/riscv: Move gen_arith_imm() decoding into trans_* functions
@ 2019-01-11 21:00 ` Richard Henderson
0 siblings, 0 replies; 140+ messages in thread
From: Richard Henderson @ 2019-01-11 21:00 UTC (permalink / raw)
To: Bastian Koppelmann, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-riscv, qemu-devel
On 1/12/19 12:10 AM, Bastian Koppelmann wrote:
>
> On 10/31/18 11:18 PM, Richard Henderson wrote:
>>
>> Surely the shri and sari functions need the same shamt >= TARGET_LONG_BITS
>> check as slli. Otherwise RV32 shri should definitely produce an assert in
>> tcg_gen_shri_tl.
>>
>> I did wonder about changing the decode of the shift functions such that only
>> the top two bits of the imm are reserved for secondary parsing of the shift
>> type, and the other 10 bits are passed down into trans_foo. At which point the
>> TARGET_LONG_BITS check takes care of other illegalities.
>>
>> Which means that the parsing for slli and slliw are identical, and also that
>> for the far future when RV128 is a thing, we don't have to change the parsing.
>
>
> I don't quite understand this. Do you want to have one entry in the decode file
> for slli and slliw?
>
> How is the parsing of slli and slliw identical with this change? As far as I
> see it, they are different at least in the opcode.
I meant in the extraction and validation of operands, I think.
I'm not really sure where else I was going with this. It has
been two months and I don't have the decode in front of me.
r~
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-devel] [PATCH v3 24/35] target/riscv: Move gen_arith_imm() decoding into trans_* functions
2019-01-11 21:00 ` [Qemu-riscv] " Richard Henderson
@ 2019-01-18 12:00 ` Bastian Koppelmann
-1 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2019-01-18 12:00 UTC (permalink / raw)
To: Richard Henderson, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-riscv, qemu-devel
On 1/11/19 10:00 PM, Richard Henderson wrote:
> On 1/12/19 12:10 AM, Bastian Koppelmann wrote:
>> On 10/31/18 11:18 PM, Richard Henderson wrote:
>>> Surely the shri and sari functions need the same shamt >= TARGET_LONG_BITS
>>> check as slli. Otherwise RV32 shri should definitely produce an assert in
>>> tcg_gen_shri_tl.
>>>
>>> I did wonder about changing the decode of the shift functions such that only
>>> the top two bits of the imm are reserved for secondary parsing of the shift
>>> type, and the other 10 bits are passed down into trans_foo. At which point the
>>> TARGET_LONG_BITS check takes care of other illegalities.
>>>
>>> Which means that the parsing for slli and slliw are identical, and also that
>>> for the far future when RV128 is a thing, we don't have to change the parsing.
>>
>> I don't quite understand this. Do you want to have one entry in the decode file
>> for slli and slliw?
>>
>> How is the parsing of slli and slliw identical with this change? As far as I
>> see it, they are different at least in the opcode.
> I meant in the extraction and validation of operands, I think.
> I'm not really sure where else I was going with this. It has
> been two months and I don't have the decode in front of me.
Okay, I extended the immediate to 10 bits for the next patch round
(which I send out soon) which should not break anything. I suggest we
leave this optimization for another patch as this set is already long
and complex.
Cheers,
Bastian
^ permalink raw reply [flat|nested] 140+ messages in thread
* Re: [Qemu-riscv] [Qemu-devel] [PATCH v3 24/35] target/riscv: Move gen_arith_imm() decoding into trans_* functions
@ 2019-01-18 12:00 ` Bastian Koppelmann
0 siblings, 0 replies; 140+ messages in thread
From: Bastian Koppelmann @ 2019-01-18 12:00 UTC (permalink / raw)
To: Richard Henderson, mjc, sagark, palmer, Alistair.Francis
Cc: peer.adelt, qemu-riscv, qemu-devel
On 1/11/19 10:00 PM, Richard Henderson wrote:
> On 1/12/19 12:10 AM, Bastian Koppelmann wrote:
>> On 10/31/18 11:18 PM, Richard Henderson wrote:
>>> Surely the shri and sari functions need the same shamt >= TARGET_LONG_BITS
>>> check as slli. Otherwise RV32 shri should definitely produce an assert in
>>> tcg_gen_shri_tl.
>>>
>>> I did wonder about changing the decode of the shift functions such that only
>>> the top two bits of the imm are reserved for secondary parsing of the shift
>>> type, and the other 10 bits are passed down into trans_foo. At which point the
>>> TARGET_LONG_BITS check takes care of other illegalities.
>>>
>>> Which means that the parsing for slli and slliw are identical, and also that
>>> for the far future when RV128 is a thing, we don't have to change the parsing.
>>
>> I don't quite understand this. Do you want to have one entry in the decode file
>> for slli and slliw?
>>
>> How is the parsing of slli and slliw identical with this change? As far as I
>> see it, they are different at least in the opcode.
> I meant in the extraction and validation of operands, I think.
> I'm not really sure where else I was going with this. It has
> been two months and I don't have the decode in front of me.
Okay, I extended the immediate to 10 bits for the next patch round
(which I send out soon) which should not break anything. I suggest we
leave this optimization for another patch as this set is already long
and complex.
Cheers,
Bastian
^ permalink raw reply [flat|nested] 140+ messages in thread
end of thread, other threads:[~2019-01-18 12:00 UTC | newest]
Thread overview: 140+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-10-31 13:19 [Qemu-devel] [PATCH v3 00/35] target/riscv: Convert to decodetree Bastian Koppelmann
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 13:19 ` [Qemu-devel] [PATCH v3 01/35] target/riscv: Move CPURISCVState pointer to DisasContext Bastian Koppelmann
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 17:07 ` [Qemu-devel] " Richard Henderson
2018-10-31 17:07 ` [Qemu-riscv] " Richard Henderson
2018-10-31 20:14 ` [Qemu-devel] " Alistair Francis
2018-10-31 20:14 ` [Qemu-riscv] " Alistair Francis
2018-10-31 13:19 ` [Qemu-devel] [PATCH v3 02/35] target/riscv: Activate decodetree and implemnt LUI & AUIPC Bastian Koppelmann
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 20:20 ` [Qemu-devel] " Alistair
2018-10-31 20:20 ` [Qemu-riscv] " Alistair
2018-10-31 13:19 ` [Qemu-devel] [PATCH v3 03/35] target/riscv: Convert RVXI branch insns to decodetree Bastian Koppelmann
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 17:11 ` [Qemu-devel] " Richard Henderson
2018-10-31 17:11 ` [Qemu-riscv] " Richard Henderson
2018-10-31 13:19 ` [Qemu-devel] [PATCH v3 04/35] target/riscv: Convert RV32I load/store " Bastian Koppelmann
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 13:19 ` [Qemu-devel] [PATCH v3 05/35] target/riscv: Convert RV64I " Bastian Koppelmann
2018-10-31 13:19 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 17:14 ` [Qemu-devel] " Richard Henderson
2018-10-31 17:14 ` [Qemu-riscv] " Richard Henderson
2018-10-31 20:26 ` [Qemu-devel] " Alistair
2018-10-31 20:26 ` [Qemu-riscv] " Alistair
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 06/35] target/riscv: Convert RVXI arithmetic " Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 17:15 ` [Qemu-devel] " Richard Henderson
2018-10-31 17:15 ` [Qemu-riscv] " Richard Henderson
2018-10-31 20:29 ` [Qemu-devel] " Alistair
2018-10-31 20:29 ` [Qemu-riscv] " Alistair
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 07/35] target/riscv: Convert RVXI fence " Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 20:30 ` [Qemu-devel] " Alistair
2018-10-31 20:30 ` [Qemu-riscv] " Alistair
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 08/35] target/riscv: Convert RVXI csr " Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 20:46 ` [Qemu-devel] " Alistair
2018-10-31 20:46 ` [Qemu-riscv] " Alistair
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 09/35] target/riscv: Convert RVXM " Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 20:38 ` [Qemu-devel] " Alistair
2018-10-31 20:38 ` [Qemu-riscv] " Alistair
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 10/35] target/riscv: Convert RV32A " Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 11/35] target/riscv: Convert RV64A " Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 20:49 ` [Qemu-devel] " Alistair
2018-10-31 20:49 ` [Qemu-riscv] " Alistair
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 12/35] target/riscv: Convert RV32F " Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 13/35] target/riscv: Convert RV64F " Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 14/35] target/riscv: Convert RV32D " Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 15/35] target/riscv: Convert RV64D " Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 16/35] target/riscv: Convert RV priv " Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 17/35] target/riscv: Convert quadrant 0 of RVXC " Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 18/35] target/riscv: Convert quadrant 1 " Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 19/35] target/riscv: Convert quadrant 2 " Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 20/35] target/riscv: Remove gen_jalr() Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 20:50 ` [Qemu-devel] " Alistair
2018-10-31 20:50 ` [Qemu-riscv] " Alistair
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 21/35] target/riscv: Remove manual decoding from gen_branch() Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 22/35] target/riscv: Remove manual decoding from gen_load() Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 22:09 ` [Qemu-devel] " Richard Henderson
2018-10-31 22:09 ` [Qemu-riscv] " Richard Henderson
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 23/35] target/riscv: Remove manual decoding from gen_store() Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 22:09 ` [Qemu-devel] " Richard Henderson
2018-10-31 22:09 ` [Qemu-riscv] " Richard Henderson
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 24/35] target/riscv: Move gen_arith_imm() decoding into trans_* functions Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 22:18 ` [Qemu-devel] " Richard Henderson
2018-10-31 22:18 ` [Qemu-riscv] " Richard Henderson
2019-01-11 13:10 ` [Qemu-devel] " Bastian Koppelmann
2019-01-11 13:10 ` [Qemu-riscv] " Bastian Koppelmann
2019-01-11 21:00 ` Richard Henderson
2019-01-11 21:00 ` [Qemu-riscv] " Richard Henderson
2019-01-18 12:00 ` Bastian Koppelmann
2019-01-18 12:00 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 22:26 ` Richard Henderson
2018-10-31 22:26 ` [Qemu-riscv] " Richard Henderson
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 25/35] target/riscv: make ADD/SUB/OR/XOR/AND insn use arg lists Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 20:44 ` [Qemu-devel] " Alistair Francis
2018-10-31 20:44 ` [Qemu-riscv] " Alistair Francis
2018-10-31 22:27 ` Richard Henderson
2018-10-31 22:27 ` [Qemu-riscv] " Richard Henderson
2018-10-31 22:26 ` Richard Henderson
2018-10-31 22:26 ` [Qemu-riscv] " Richard Henderson
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 26/35] target/riscv: Remove shift and slt insn manual decoding Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 22:38 ` [Qemu-devel] " Richard Henderson
2018-10-31 22:38 ` [Qemu-riscv] " Richard Henderson
2018-11-01 15:59 ` [Qemu-devel] " Palmer Dabbelt
2018-11-01 15:59 ` [Qemu-riscv] " Palmer Dabbelt
2018-11-05 17:00 ` [Qemu-devel] " Bastian Koppelmann
2018-11-05 17:00 ` [Qemu-riscv] " Bastian Koppelmann
2018-11-07 0:56 ` Palmer Dabbelt
2018-11-07 0:56 ` [Qemu-riscv] " Palmer Dabbelt
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 27/35] target/riscv: Remove manual decoding of RV32/64M insn Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 22:39 ` [Qemu-devel] " Richard Henderson
2018-10-31 22:39 ` [Qemu-riscv] " Richard Henderson
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 28/35] target/riscv: Rename trans_arith to gen_arith Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 29/35] target/riscv: Remove gen_system() Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 30/35] target/riscv: Remove decode_RV32_64G() Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 31/35] target/riscv: Convert @cs_2 insns to share translation functions Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 22:42 ` [Qemu-devel] " Richard Henderson
2018-10-31 22:42 ` [Qemu-riscv] " Richard Henderson
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 32/35] target/riscv: Convert @cl_d, @cl_w, @cs_d, @cs_w insns Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 22:43 ` [Qemu-devel] " Richard Henderson
2018-10-31 22:43 ` [Qemu-riscv] " Richard Henderson
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 33/35] target/riscv: Splice fsw_sd and flw_ld for riscv32 vs riscv64 Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 22:45 ` [Qemu-devel] " Richard Henderson
2018-10-31 22:45 ` [Qemu-riscv] " Richard Henderson
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 34/35] target/riscv: Splice remaining compressed insn pairs " Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 22:47 ` [Qemu-devel] " Richard Henderson
2018-10-31 22:47 ` [Qemu-riscv] " Richard Henderson
2018-10-31 13:20 ` [Qemu-devel] [PATCH v3 35/35] target/riscv: Remaining rvc insn reuse 32 bit translators Bastian Koppelmann
2018-10-31 13:20 ` [Qemu-riscv] " Bastian Koppelmann
2018-10-31 22:49 ` [Qemu-devel] " Richard Henderson
2018-10-31 22:49 ` [Qemu-riscv] " Richard Henderson
2018-11-02 8:48 ` [Qemu-devel] [PATCH v3 00/35] target/riscv: Convert to decodetree no-reply
2018-11-02 8:48 ` [Qemu-riscv] " no-reply
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.