From: Peter Maydell <peter.maydell@linaro.org>
To: qemu-arm@nongnu.org, qemu-devel@nongnu.org
Cc: Richard Henderson <richard.henderson@linaro.org>
Subject: [PATCH v3 09/44] tcg: Make gen_dup_i32/i64() public as tcg_gen_dup_i32/i64
Date: Thu, 17 Jun 2021 13:15:53 +0100 [thread overview]
Message-ID: <20210617121628.20116-10-peter.maydell@linaro.org> (raw)
In-Reply-To: <20210617121628.20116-1-peter.maydell@linaro.org>
The Arm MVE VDUP implementation would like to be able to emit code to
duplicate a byte or halfword value into an i32. We have code to do
this already in tcg-op-gvec.c, so all we need to do is make the
functions global.
For consistency with other functions made available to the frontends:
* we rename to tcg_gen_dup_*
* we expose both the _i32 and _i64 forms
* we provide the #define for a _tl form
Suggested-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
---
Changes v2->v3: renamed and used different header file, as suggested
by Richard. I also added the _tl #define.
---
include/tcg/tcg-op.h | 8 ++++++++
include/tcg/tcg.h | 1 -
tcg/tcg-op-gvec.c | 20 ++++++++++----------
3 files changed, 18 insertions(+), 11 deletions(-)
diff --git a/include/tcg/tcg-op.h b/include/tcg/tcg-op.h
index ef8a008ea74..1a2ae937583 100644
--- a/include/tcg/tcg-op.h
+++ b/include/tcg/tcg-op.h
@@ -338,6 +338,9 @@ void tcg_gen_umin_i32(TCGv_i32, TCGv_i32 arg1, TCGv_i32 arg2);
void tcg_gen_umax_i32(TCGv_i32, TCGv_i32 arg1, TCGv_i32 arg2);
void tcg_gen_abs_i32(TCGv_i32, TCGv_i32);
+/* Replicate a value of size @vece from @in to all the lanes in @out */
+void tcg_gen_dup_i32(unsigned vece, TCGv_i32 out, TCGv_i32 in);
+
static inline void tcg_gen_discard_i32(TCGv_i32 arg)
{
tcg_gen_op1_i32(INDEX_op_discard, arg);
@@ -534,6 +537,9 @@ void tcg_gen_umin_i64(TCGv_i64, TCGv_i64 arg1, TCGv_i64 arg2);
void tcg_gen_umax_i64(TCGv_i64, TCGv_i64 arg1, TCGv_i64 arg2);
void tcg_gen_abs_i64(TCGv_i64, TCGv_i64);
+/* Replicate a value of size @vece from @in to all the lanes in @out */
+void tcg_gen_dup_i64(unsigned vece, TCGv_i64 out, TCGv_i64 in);
+
#if TCG_TARGET_REG_BITS == 64
static inline void tcg_gen_discard_i64(TCGv_i64 arg)
{
@@ -1127,6 +1133,7 @@ void tcg_gen_stl_vec(TCGv_vec r, TCGv_ptr base, TCGArg offset, TCGType t);
#define tcg_gen_atomic_smax_fetch_tl tcg_gen_atomic_smax_fetch_i64
#define tcg_gen_atomic_umax_fetch_tl tcg_gen_atomic_umax_fetch_i64
#define tcg_gen_dup_tl_vec tcg_gen_dup_i64_vec
+#define tcg_gen_dup_tl tcg_gen_dup_i64
#else
#define tcg_gen_movi_tl tcg_gen_movi_i32
#define tcg_gen_mov_tl tcg_gen_mov_i32
@@ -1241,6 +1248,7 @@ void tcg_gen_stl_vec(TCGv_vec r, TCGv_ptr base, TCGArg offset, TCGType t);
#define tcg_gen_atomic_smax_fetch_tl tcg_gen_atomic_smax_fetch_i32
#define tcg_gen_atomic_umax_fetch_tl tcg_gen_atomic_umax_fetch_i32
#define tcg_gen_dup_tl_vec tcg_gen_dup_i32_vec
+#define tcg_gen_dup_tl tcg_gen_dup_i32
#endif
#if UINTPTR_MAX == UINT32_MAX
diff --git a/include/tcg/tcg.h b/include/tcg/tcg.h
index 064dab383bc..483e1e1f24e 100644
--- a/include/tcg/tcg.h
+++ b/include/tcg/tcg.h
@@ -1331,7 +1331,6 @@ uint64_t dup_const(unsigned vece, uint64_t c);
: (qemu_build_not_reached_always(), 0)) \
: dup_const(VECE, C))
-
/*
* Memory helpers that will be used by TCG generated code.
*/
diff --git a/tcg/tcg-op-gvec.c b/tcg/tcg-op-gvec.c
index 498a959839f..515db120cc6 100644
--- a/tcg/tcg-op-gvec.c
+++ b/tcg/tcg-op-gvec.c
@@ -386,7 +386,7 @@ uint64_t (dup_const)(unsigned vece, uint64_t c)
}
/* Duplicate IN into OUT as per VECE. */
-static void gen_dup_i32(unsigned vece, TCGv_i32 out, TCGv_i32 in)
+void tcg_gen_dup_i32(unsigned vece, TCGv_i32 out, TCGv_i32 in)
{
switch (vece) {
case MO_8:
@@ -404,7 +404,7 @@ static void gen_dup_i32(unsigned vece, TCGv_i32 out, TCGv_i32 in)
}
}
-static void gen_dup_i64(unsigned vece, TCGv_i64 out, TCGv_i64 in)
+void tcg_gen_dup_i64(unsigned vece, TCGv_i64 out, TCGv_i64 in)
{
switch (vece) {
case MO_8:
@@ -578,15 +578,15 @@ static void do_dup(unsigned vece, uint32_t dofs, uint32_t oprsz,
&& (vece != MO_32 || !check_size_impl(oprsz, 4))) {
t_64 = tcg_temp_new_i64();
tcg_gen_extu_i32_i64(t_64, in_32);
- gen_dup_i64(vece, t_64, t_64);
+ tcg_gen_dup_i64(vece, t_64, t_64);
} else {
t_32 = tcg_temp_new_i32();
- gen_dup_i32(vece, t_32, in_32);
+ tcg_gen_dup_i32(vece, t_32, in_32);
}
} else if (in_64) {
/* We are given a 64-bit variable input. */
t_64 = tcg_temp_new_i64();
- gen_dup_i64(vece, t_64, in_64);
+ tcg_gen_dup_i64(vece, t_64, in_64);
} else {
/* We are given a constant input. */
/* For 64-bit hosts, use 64-bit constants for "simple" constants
@@ -1311,14 +1311,14 @@ void tcg_gen_gvec_2s(uint32_t dofs, uint32_t aofs, uint32_t oprsz,
} else if (g->fni8 && check_size_impl(oprsz, 8)) {
TCGv_i64 t64 = tcg_temp_new_i64();
- gen_dup_i64(g->vece, t64, c);
+ tcg_gen_dup_i64(g->vece, t64, c);
expand_2s_i64(dofs, aofs, oprsz, t64, g->scalar_first, g->fni8);
tcg_temp_free_i64(t64);
} else if (g->fni4 && check_size_impl(oprsz, 4)) {
TCGv_i32 t32 = tcg_temp_new_i32();
tcg_gen_extrl_i64_i32(t32, c);
- gen_dup_i32(g->vece, t32, t32);
+ tcg_gen_dup_i32(g->vece, t32, t32);
expand_2s_i32(dofs, aofs, oprsz, t32, g->scalar_first, g->fni4);
tcg_temp_free_i32(t32);
} else {
@@ -2538,7 +2538,7 @@ void tcg_gen_gvec_ands(unsigned vece, uint32_t dofs, uint32_t aofs,
TCGv_i64 c, uint32_t oprsz, uint32_t maxsz)
{
TCGv_i64 tmp = tcg_temp_new_i64();
- gen_dup_i64(vece, tmp, c);
+ tcg_gen_dup_i64(vece, tmp, c);
tcg_gen_gvec_2s(dofs, aofs, oprsz, maxsz, tmp, &gop_ands);
tcg_temp_free_i64(tmp);
}
@@ -2562,7 +2562,7 @@ void tcg_gen_gvec_xors(unsigned vece, uint32_t dofs, uint32_t aofs,
TCGv_i64 c, uint32_t oprsz, uint32_t maxsz)
{
TCGv_i64 tmp = tcg_temp_new_i64();
- gen_dup_i64(vece, tmp, c);
+ tcg_gen_dup_i64(vece, tmp, c);
tcg_gen_gvec_2s(dofs, aofs, oprsz, maxsz, tmp, &gop_xors);
tcg_temp_free_i64(tmp);
}
@@ -2586,7 +2586,7 @@ void tcg_gen_gvec_ors(unsigned vece, uint32_t dofs, uint32_t aofs,
TCGv_i64 c, uint32_t oprsz, uint32_t maxsz)
{
TCGv_i64 tmp = tcg_temp_new_i64();
- gen_dup_i64(vece, tmp, c);
+ tcg_gen_dup_i64(vece, tmp, c);
tcg_gen_gvec_2s(dofs, aofs, oprsz, maxsz, tmp, &gop_ors);
tcg_temp_free_i64(tmp);
}
--
2.20.1
next prev parent reply other threads:[~2021-06-17 12:19 UTC|newest]
Thread overview: 54+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-06-17 12:15 [PATCH v3 00/44] target/arm: First slice of MVE implementation Peter Maydell
2021-06-17 12:15 ` [PATCH v3 01/44] target/arm: Implement MVE VLDR/VSTR (non-widening forms) Peter Maydell
2021-06-17 13:30 ` Claudio Fontana
2021-06-17 13:47 ` Peter Maydell
2021-06-18 14:44 ` Richard Henderson
2021-06-17 12:15 ` [PATCH v3 02/44] target/arm: Implement widening/narrowing MVE VLDR/VSTR insns Peter Maydell
2021-06-18 14:47 ` Richard Henderson
2021-06-17 12:15 ` [PATCH v3 03/44] target/arm: Implement MVE VCLZ Peter Maydell
2021-06-21 13:28 ` Peter Maydell
2021-06-21 16:12 ` Peter Maydell
2021-06-17 12:15 ` [PATCH v3 04/44] target/arm: Implement MVE VCLS Peter Maydell
2021-06-17 12:15 ` [PATCH v3 05/44] target/arm: Implement MVE VREV16, VREV32, VREV64 Peter Maydell
2021-06-17 12:15 ` [PATCH v3 06/44] target/arm: Implement MVE VMVN (register) Peter Maydell
2021-06-17 12:15 ` [PATCH v3 07/44] target/arm: Implement MVE VABS Peter Maydell
2021-06-17 12:15 ` [PATCH v3 08/44] target/arm: Implement MVE VNEG Peter Maydell
2021-06-17 12:15 ` Peter Maydell [this message]
2021-06-17 12:15 ` [PATCH v3 10/44] target/arm: Implement MVE VDUP Peter Maydell
2021-06-17 12:15 ` [PATCH v3 11/44] target/arm: Implement MVE VAND, VBIC, VORR, VORN, VEOR Peter Maydell
2021-06-17 12:15 ` [PATCH v3 12/44] target/arm: Implement MVE VADD, VSUB, VMUL Peter Maydell
2021-06-17 12:15 ` [PATCH v3 13/44] target/arm: Implement MVE VMULH Peter Maydell
2021-06-17 12:15 ` [PATCH v3 14/44] target/arm: Implement MVE VRMULH Peter Maydell
2021-06-17 12:15 ` [PATCH v3 15/44] target/arm: Implement MVE VMAX, VMIN Peter Maydell
2021-06-17 12:16 ` [PATCH v3 16/44] target/arm: Implement MVE VABD Peter Maydell
2021-06-17 12:16 ` [PATCH v3 17/44] target/arm: Implement MVE VHADD, VHSUB Peter Maydell
2021-06-17 12:16 ` [PATCH v3 18/44] target/arm: Implement MVE VMULL Peter Maydell
2021-06-17 12:16 ` [PATCH v3 19/44] target/arm: Implement MVE VMLALDAV Peter Maydell
2021-06-17 12:16 ` [PATCH v3 20/44] target/arm: Implement MVE VMLSLDAV Peter Maydell
2021-06-17 12:16 ` [PATCH v3 21/44] target/arm: Implement MVE VRMLALDAVH, VRMLSLDAVH Peter Maydell
2021-06-17 12:16 ` [PATCH v3 22/44] target/arm: Implement MVE VADD (scalar) Peter Maydell
2021-06-17 12:16 ` [PATCH v3 23/44] target/arm: Implement MVE VSUB, VMUL (scalar) Peter Maydell
2021-06-17 12:16 ` [PATCH v3 24/44] target/arm: Implement MVE VHADD, VHSUB (scalar) Peter Maydell
2021-06-17 12:16 ` [PATCH v3 25/44] target/arm: Implement MVE VBRSR Peter Maydell
2021-06-17 12:16 ` [PATCH v3 26/44] target/arm: Implement MVE VPST Peter Maydell
2021-06-17 12:16 ` [PATCH v3 27/44] target/arm: Implement MVE VQADD and VQSUB Peter Maydell
2021-06-17 12:16 ` [PATCH v3 28/44] target/arm: Implement MVE VQDMULH and VQRDMULH (scalar) Peter Maydell
2021-06-17 12:16 ` [PATCH v3 29/44] target/arm: Implement MVE VQDMULL scalar Peter Maydell
2021-06-17 12:16 ` [PATCH v3 30/44] target/arm: Implement MVE VQDMULH, VQRDMULH (vector) Peter Maydell
2021-06-17 12:16 ` [PATCH v3 31/44] target/arm: Implement MVE VQADD, VQSUB (vector) Peter Maydell
2021-06-17 12:16 ` [PATCH v3 32/44] target/arm: Implement MVE VQSHL (vector) Peter Maydell
2021-06-17 12:16 ` [PATCH v3 33/44] target/arm: Implement MVE VQRSHL Peter Maydell
2021-06-17 12:16 ` [PATCH v3 34/44] target/arm: Implement MVE VSHL insn Peter Maydell
2021-06-17 12:16 ` [PATCH v3 35/44] target/arm: Implmement MVE VRSHL Peter Maydell
2021-06-17 13:20 ` Claudio Fontana
2021-06-17 13:23 ` Peter Maydell
2021-06-17 12:16 ` [PATCH v3 36/44] target/arm: Implement MVE VQDMLADH and VQRDMLADH Peter Maydell
2021-06-17 12:16 ` [PATCH v3 37/44] target/arm: Implement MVE VQDMLSDH and VQRDMLSDH Peter Maydell
2021-06-17 12:16 ` [PATCH v3 38/44] target/arm: Implement MVE VQDMULL (vector) Peter Maydell
2021-06-17 12:16 ` [PATCH v3 39/44] target/arm: Implement MVE VRHADD Peter Maydell
2021-06-17 12:16 ` [PATCH v3 40/44] target/arm: Implement MVE VADC, VSBC Peter Maydell
2021-06-17 12:16 ` [PATCH v3 41/44] target/arm: Implement MVE VCADD Peter Maydell
2021-06-17 12:16 ` [PATCH v3 42/44] target/arm: Implement MVE VHCADD Peter Maydell
2021-06-17 12:16 ` [PATCH v3 43/44] target/arm: Implement MVE VADDV Peter Maydell
2021-06-17 12:16 ` [PATCH v3 44/44] target/arm: Make VMOV scalar <-> gpreg beatwise for MVE Peter Maydell
2021-06-17 13:10 ` [PATCH v3 00/44] target/arm: First slice of MVE implementation no-reply
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20210617121628.20116-10-peter.maydell@linaro.org \
--to=peter.maydell@linaro.org \
--cc=qemu-arm@nongnu.org \
--cc=qemu-devel@nongnu.org \
--cc=richard.henderson@linaro.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).