[PULL v2 00/18] MIPS + misc queue for June 15th, 2020

[PULL v2 00/18] MIPS + misc queue for June 15th, 2020

[Aleksandar Markovic]

The following changes since commit 7d3660e79830a069f1848bb4fa1cdf8f666424fb:

  Merge remote-tracking branch 'remotes/bonzini/tags/for-upstream' into staging (2020-06-12 23:06:22 +0100)

are available in the git repository at:

  https://github.com/AMarkovic/qemu tags/mips-queue-jun-15-2020

for you to fetch changes up to 250bc43a406f7d46e319abe87c19548d4f027828:

  translations: Add Swedish language (2020-06-15 20:51:10 +0200)

----------------------------------------------------------------

MIPS + misc queue for June 15th, 2020

  Highlights:

    This pull request, just exceptionally, contains two non-MIPS patches:

      - adjust sh4 maintainership
      - add Swedish translations

    The rest are MIPS patches:

      - refactor emulation of a number of MSA instructions
      - activate Loongson-related insn_flags

    Notes:

      - one checkpatch warning is benign
      - some of make check iotest-qcow2 tests fail on my system, both before
        and after applying the patches from this pull request

----------------------------------------------------------------


Aleksandar Markovic (15):
  target/mips: msa: Split helpers for MADDV.<B|H|W|D>
  target/mips: msa: Split helpers for MSUBV.<B|H|W|D>
  target/mips: msa: Split helpers for DPADD_S.<H|W|D>
  target/mips: msa: Split helpers for DPADD_U.<H|W|D>
  target/mips: msa: Split helpers for DPSUB_S.<H|W|D>
  target/mips: msa: Split helpers for DPSUB_U.<H|W|D>
  target/mips: msa: Split helpers for DOTP_S.<H|W|D>
  target/mips: msa: Split helpers for DOTP_U.<H|W|D>
  target/mips: msa: Split helpers for SUBS_S.<B|H|W|D>
  target/mips: msa: Split helpers for SUBS_U.<B|H|W|D>
  target/mips: msa: Split helpers for SUBSUS_U.<B|H|W|D>
  target/mips: msa: Split helpers for SUBSUU_S.<B|H|W|D>
  target/mips: msa: Split helpers for SUBV.<B|H|W|D>
  target/mips: msa: Split helpers for MULV.<B|H|W|D>
  MAINTAINERS: Adjust sh4 maintainership

Jiaxun Yang (2):
  target/mips: Legalize Loongson insn flags
  target/mips: Add comments for vendor-specific ASEs

Sebastian Rasmussen (1):
  translations: Add Swedish language

 target/mips/helper.h     |   73 ++-
 target/mips/mips-defs.h  |    8 +-
 target/mips/msa_helper.c | 1296 ++++++++++++++++++++++++++++++++++++----------
 target/mips/translate.c  |  213 +++++++-
 MAINTAINERS              |    8 +-
 po/sv.po                 |   75 +++
 6 files changed, 1364 insertions(+), 309 deletions(-)
 create mode 100644 po/sv.po

-- 
2.7.4

[PULL 01/18] target/mips: Legalize Loongson insn flags

[Aleksandar Markovic]

From: Jiaxun Yang <jiaxun.yang@flygoat.com>

To match the actual status of Loongson insn, we split flags
for LMMI and LEXT from INSN_LOONGSON2F.

As Loongson-2F only implemented interger part of LEXT, we'll
not enable LEXT for the processor, but instead we're still using
INSN_LOONGSON2F as switch flag of these instructions.

All multimedia instructions have been moved to LMMI flag. Loongson-2F
and Loongson-3A are sharing these instructions.

Signed-off-by: Jiaxun Yang <jiaxun.yang@flygoat.com>
Reviewed-by: Aleksandar Markovic <aleksandar.qemu.devel@gmail.com>
Signed-off-by: Aleksandar Markovic <aleksandar.qemu.devel@gmail.com>
Message-Id: <20200614080049.31134-2-jiaxun.yang@flygoat.com>
---
 target/mips/mips-defs.h |  4 ++--
 target/mips/translate.c | 13 +++++++------
 2 files changed, 9 insertions(+), 8 deletions(-)

diff --git a/target/mips/mips-defs.h b/target/mips/mips-defs.h
index 0c12910..f1b833f 100644
--- a/target/mips/mips-defs.h
+++ b/target/mips/mips-defs.h
@@ -70,7 +70,7 @@
 #define CPU_VR54XX      (CPU_MIPS4 | INSN_VR54XX)
 #define CPU_R5900       (CPU_MIPS3 | INSN_R5900)
 #define CPU_LOONGSON2E  (CPU_MIPS3 | INSN_LOONGSON2E)
-#define CPU_LOONGSON2F  (CPU_MIPS3 | INSN_LOONGSON2F)
+#define CPU_LOONGSON2F  (CPU_MIPS3 | INSN_LOONGSON2F | ASE_LMMI)
 
 #define CPU_MIPS5       (CPU_MIPS4 | ISA_MIPS5)
 
@@ -97,7 +97,7 @@
 /* Wave Computing: "nanoMIPS" */
 #define CPU_NANOMIPS32  (CPU_MIPS32R6 | ISA_NANOMIPS32)
 
-#define CPU_LOONGSON3A  (CPU_MIPS64R2 | INSN_LOONGSON3A)
+#define CPU_LOONGSON3A  (CPU_MIPS64R2 | INSN_LOONGSON3A | ASE_LMMI | ASE_LEXT)
 
 /*
  * Strictly follow the architecture standard:
diff --git a/target/mips/translate.c b/target/mips/translate.c
index 2caf4cb..e49f32f 100644
--- a/target/mips/translate.c
+++ b/target/mips/translate.c
@@ -1046,7 +1046,7 @@ enum {
     OPC_BC2NEZ  = (0x0D << 21) | OPC_CP2,
 };
 
-#define MASK_LMI(op)    (MASK_OP_MAJOR(op) | (op & (0x1F << 21)) | (op & 0x1F))
+#define MASK_LMMI(op)    (MASK_OP_MAJOR(op) | (op & (0x1F << 21)) | (op & 0x1F))
 
 enum {
     OPC_PADDSH      = (24 << 21) | (0x00) | OPC_CP2,
@@ -3421,7 +3421,8 @@ static void gen_ld(DisasContext *ctx, uint32_t opc,
     TCGv t0, t1, t2;
     int mem_idx = ctx->mem_idx;
 
-    if (rt == 0 && ctx->insn_flags & (INSN_LOONGSON2E | INSN_LOONGSON2F)) {
+    if (rt == 0 && ctx->insn_flags & (INSN_LOONGSON2E | INSN_LOONGSON2F |
+                                      INSN_LOONGSON3A)) {
         /*
          * Loongson CPU uses a load to zero register for prefetch.
          * We emulate it as a NOP. On other CPU we must perform the
@@ -5531,7 +5532,7 @@ static void gen_loongson_multimedia(DisasContext *ctx, int rd, int rs, int rt)
     TCGv_i64 t0, t1;
     TCGCond cond;
 
-    opc = MASK_LMI(ctx->opcode);
+    opc = MASK_LMMI(ctx->opcode);
     switch (opc) {
     case OPC_ADD_CP2:
     case OPC_SUB_CP2:
@@ -27161,7 +27162,7 @@ static void decode_opc_special2_legacy(CPUMIPSState *env, DisasContext *ctx)
     case OPC_MULTU_G_2F:
     case OPC_MOD_G_2F:
     case OPC_MODU_G_2F:
-        check_insn(ctx, INSN_LOONGSON2F);
+        check_insn(ctx, INSN_LOONGSON2F | ASE_LEXT);
         gen_loongson_integer(ctx, op1, rd, rs, rt);
         break;
     case OPC_CLO:
@@ -27194,7 +27195,7 @@ static void decode_opc_special2_legacy(CPUMIPSState *env, DisasContext *ctx)
     case OPC_DDIVU_G_2F:
     case OPC_DMOD_G_2F:
     case OPC_DMODU_G_2F:
-        check_insn(ctx, INSN_LOONGSON2F);
+        check_insn(ctx, INSN_LOONGSON2F | ASE_LEXT);
         gen_loongson_integer(ctx, op1, rd, rs, rt);
         break;
 #endif
@@ -30641,7 +30642,7 @@ static void decode_opc(CPUMIPSState *env, DisasContext *ctx)
         }
         break;
     case OPC_CP2:
-        check_insn(ctx, INSN_LOONGSON2F);
+        check_insn(ctx, ASE_LMMI);
         /* Note that these instructions use different fields.  */
         gen_loongson_multimedia(ctx, sa, rd, rt);
         break;
-- 
2.7.4

[PULL 02/18] target/mips: Add comments for vendor-specific ASEs

[Aleksandar Markovic]

From: Jiaxun Yang <jiaxun.yang@flygoat.com>

Abbreviations of vendor-specific ASEs looks very similiar.
Add comments to explain the full name and vendors of these flags.

Signed-off-by: Jiaxun Yang <jiaxun.yang@flygoat.com>
Reviewed-by: Aleksandar Markovic <aleksandar.qemu.devel@gmail.com>
Signed-off-by: Aleksandar Markovic <aleksandar.qemu.devel@gmail.com>
Message-Id: <20200614080049.31134-3-jiaxun.yang@flygoat.com>
---
 target/mips/mips-defs.h | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/target/mips/mips-defs.h b/target/mips/mips-defs.h
index f1b833f..ed6a7a9 100644
--- a/target/mips/mips-defs.h
+++ b/target/mips/mips-defs.h
@@ -57,9 +57,13 @@
 /*
  *   bits 52-63: vendor-specific ASEs
  */
+/* MultiMedia Instructions defined by R5900 */
 #define ASE_MMI           0x0010000000000000ULL
+/* MIPS eXtension/enhanced Unit defined by Ingenic */
 #define ASE_MXU           0x0020000000000000ULL
+/* Loongson MultiMedia Instructions */
 #define ASE_LMMI          0x0040000000000000ULL
+/* Loongson EXTensions */
 #define ASE_LEXT          0x0080000000000000ULL
 
 /* MIPS CPU defines. */
-- 
2.7.4

[PULL 03/18] target/mips: msa: Split helpers for MADDV.<B|H|W|D>

[Aleksandar Markovic]

Achieves clearer code and slightly better performance.

Reviewed-by: Aleksandar Rikalo <aleksandar.rikalo@syrmia.com>
Signed-off-by: Aleksandar Markovic <aleksandar.qemu.devel@gmail.com>
Message-Id: <20200613152133.8964-2-aleksandar.qemu.devel@gmail.com>
---
 target/mips/helper.h     |  6 +++-
 target/mips/msa_helper.c | 79 +++++++++++++++++++++++++++++++++++++++++++-----
 target/mips/translate.c  | 19 ++++++++++--
 3 files changed, 92 insertions(+), 12 deletions(-)

diff --git a/target/mips/helper.h b/target/mips/helper.h
index 84fdd9f..e479a22 100644
--- a/target/mips/helper.h
+++ b/target/mips/helper.h
@@ -950,6 +950,11 @@ DEF_HELPER_4(msa_mod_s_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_mod_s_w, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_mod_s_d, void, env, i32, i32, i32)
 
+DEF_HELPER_4(msa_maddv_b, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_maddv_h, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_maddv_w, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_maddv_d, void, env, i32, i32, i32)
+
 DEF_HELPER_4(msa_asub_s_b, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_asub_s_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_asub_s_w, void, env, i32, i32, i32)
@@ -1069,7 +1074,6 @@ DEF_HELPER_5(msa_subs_u_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_subsus_u_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_subsuu_s_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_mulv_df, void, env, i32, i32, i32, i32)
-DEF_HELPER_5(msa_maddv_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_msubv_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_dotp_s_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_dotp_u_df, void, env, i32, i32, i32, i32)
diff --git a/target/mips/msa_helper.c b/target/mips/msa_helper.c
index c3b2719..3b75bdc 100644
--- a/target/mips/msa_helper.c
+++ b/target/mips/msa_helper.c
@@ -2883,7 +2883,77 @@ void helper_msa_mod_u_d(CPUMIPSState *env,
  * +---------------+----------------------------------------------------------+
  */
 
-/* TODO: insert Int Multiply group helpers here */
+static inline int64_t msa_maddv_df(uint32_t df, int64_t dest, int64_t arg1,
+                                   int64_t arg2)
+{
+    return dest + arg1 * arg2;
+}
+
+void helper_msa_maddv_b(CPUMIPSState *env,
+                        uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->b[0]  = msa_maddv_df(DF_BYTE, pwt->b[0],  pws->b[0],  pwt->b[0]);
+    pwd->b[1]  = msa_maddv_df(DF_BYTE, pwt->b[1],  pws->b[1],  pwt->b[1]);
+    pwd->b[2]  = msa_maddv_df(DF_BYTE, pwt->b[2],  pws->b[2],  pwt->b[2]);
+    pwd->b[3]  = msa_maddv_df(DF_BYTE, pwt->b[3],  pws->b[3],  pwt->b[3]);
+    pwd->b[4]  = msa_maddv_df(DF_BYTE, pwt->b[4],  pws->b[4],  pwt->b[4]);
+    pwd->b[5]  = msa_maddv_df(DF_BYTE, pwt->b[5],  pws->b[5],  pwt->b[5]);
+    pwd->b[6]  = msa_maddv_df(DF_BYTE, pwt->b[6],  pws->b[6],  pwt->b[6]);
+    pwd->b[7]  = msa_maddv_df(DF_BYTE, pwt->b[7],  pws->b[7],  pwt->b[7]);
+    pwd->b[8]  = msa_maddv_df(DF_BYTE, pwt->b[8],  pws->b[8],  pwt->b[8]);
+    pwd->b[9]  = msa_maddv_df(DF_BYTE, pwt->b[9],  pws->b[9],  pwt->b[9]);
+    pwd->b[10] = msa_maddv_df(DF_BYTE, pwt->b[10], pws->b[10], pwt->b[10]);
+    pwd->b[11] = msa_maddv_df(DF_BYTE, pwt->b[11], pws->b[11], pwt->b[11]);
+    pwd->b[12] = msa_maddv_df(DF_BYTE, pwt->b[12], pws->b[12], pwt->b[12]);
+    pwd->b[13] = msa_maddv_df(DF_BYTE, pwt->b[13], pws->b[13], pwt->b[13]);
+    pwd->b[14] = msa_maddv_df(DF_BYTE, pwt->b[14], pws->b[14], pwt->b[14]);
+    pwd->b[15] = msa_maddv_df(DF_BYTE, pwt->b[15], pws->b[15], pwt->b[15]);
+}
+
+void helper_msa_maddv_h(CPUMIPSState *env,
+                        uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->h[0]  = msa_maddv_df(DF_HALF, pwd->h[0],  pws->h[0],  pwt->h[0]);
+    pwd->h[1]  = msa_maddv_df(DF_HALF, pwd->h[1],  pws->h[1],  pwt->h[1]);
+    pwd->h[2]  = msa_maddv_df(DF_HALF, pwd->h[2],  pws->h[2],  pwt->h[2]);
+    pwd->h[3]  = msa_maddv_df(DF_HALF, pwd->h[3],  pws->h[3],  pwt->h[3]);
+    pwd->h[4]  = msa_maddv_df(DF_HALF, pwd->h[4],  pws->h[4],  pwt->h[4]);
+    pwd->h[5]  = msa_maddv_df(DF_HALF, pwd->h[5],  pws->h[5],  pwt->h[5]);
+    pwd->h[6]  = msa_maddv_df(DF_HALF, pwd->h[6],  pws->h[6],  pwt->h[6]);
+    pwd->h[7]  = msa_maddv_df(DF_HALF, pwd->h[7],  pws->h[7],  pwt->h[7]);
+}
+
+void helper_msa_maddv_w(CPUMIPSState *env,
+                        uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->w[0]  = msa_maddv_df(DF_WORD, pwd->w[0],  pws->w[0],  pwt->w[0]);
+    pwd->w[1]  = msa_maddv_df(DF_WORD, pwd->w[1],  pws->w[1],  pwt->w[1]);
+    pwd->w[2]  = msa_maddv_df(DF_WORD, pwd->w[2],  pws->w[2],  pwt->w[2]);
+    pwd->w[3]  = msa_maddv_df(DF_WORD, pwd->w[3],  pws->w[3],  pwt->w[3]);
+}
+
+void helper_msa_maddv_d(CPUMIPSState *env,
+                        uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->d[0]  = msa_maddv_df(DF_DOUBLE, pwd->d[0],  pws->d[0],  pwt->d[0]);
+    pwd->d[1]  = msa_maddv_df(DF_DOUBLE, pwd->d[1],  pws->d[1],  pwt->d[1]);
+}
 
 
 /*
@@ -4816,12 +4886,6 @@ void helper_msa_sld_df(CPUMIPSState *env, uint32_t df, uint32_t wd,
     msa_sld_df(df, pwd, pws, env->active_tc.gpr[rt]);
 }
 
-static inline int64_t msa_maddv_df(uint32_t df, int64_t dest, int64_t arg1,
-                                   int64_t arg2)
-{
-    return dest + arg1 * arg2;
-}
-
 static inline int64_t msa_msubv_df(uint32_t df, int64_t dest, int64_t arg1,
                                    int64_t arg2)
 {
@@ -5002,7 +5066,6 @@ void helper_msa_ ## func ## _df(CPUMIPSState *env, uint32_t df, uint32_t wd,  \
     }                                                                         \
 }
 
-MSA_TEROP_DF(maddv)
 MSA_TEROP_DF(msubv)
 MSA_TEROP_DF(dpadd_s)
 MSA_TEROP_DF(dpadd_u)
diff --git a/target/mips/translate.c b/target/mips/translate.c
index e49f32f..4975e57 100644
--- a/target/mips/translate.c
+++ b/target/mips/translate.c
@@ -29058,6 +29058,22 @@ static void gen_msa_3r(CPUMIPSState *env, DisasContext *ctx)
             break;
         }
         break;
+    case OPC_MADDV_df:
+        switch (df) {
+        case DF_BYTE:
+            gen_helper_msa_maddv_b(cpu_env, twd, tws, twt);
+            break;
+        case DF_HALF:
+            gen_helper_msa_maddv_h(cpu_env, twd, tws, twt);
+            break;
+        case DF_WORD:
+            gen_helper_msa_maddv_w(cpu_env, twd, tws, twt);
+            break;
+        case DF_DOUBLE:
+            gen_helper_msa_maddv_d(cpu_env, twd, tws, twt);
+            break;
+        }
+        break;
     case OPC_ASUB_S_df:
         switch (df) {
         case DF_BYTE:
@@ -29284,9 +29300,6 @@ static void gen_msa_3r(CPUMIPSState *env, DisasContext *ctx)
     case OPC_SUBS_U_df:
         gen_helper_msa_subs_u_df(cpu_env, tdf, twd, tws, twt);
         break;
-    case OPC_MADDV_df:
-        gen_helper_msa_maddv_df(cpu_env, tdf, twd, tws, twt);
-        break;
     case OPC_SPLAT_df:
         gen_helper_msa_splat_df(cpu_env, tdf, twd, tws, twt);
         break;
-- 
2.7.4

[PULL 04/18] target/mips: msa: Split helpers for MSUBV.<B|H|W|D>

[Aleksandar Markovic]

Achieves clearer code and slightly better performance.

Reviewed-by: Aleksandar Rikalo <aleksandar.rikalo@syrmia.com>
Signed-off-by: Aleksandar Markovic <aleksandar.qemu.devel@gmail.com>
Message-Id: <20200613152133.8964-3-aleksandar.qemu.devel@gmail.com>
---
 target/mips/helper.h     |  6 +++-
 target/mips/msa_helper.c | 79 +++++++++++++++++++++++++++++++++++++++++++-----
 target/mips/translate.c  | 19 ++++++++++--
 3 files changed, 93 insertions(+), 11 deletions(-)

diff --git a/target/mips/helper.h b/target/mips/helper.h
index e479a22..7ca0036 100644
--- a/target/mips/helper.h
+++ b/target/mips/helper.h
@@ -955,6 +955,11 @@ DEF_HELPER_4(msa_maddv_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_maddv_w, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_maddv_d, void, env, i32, i32, i32)
 
+DEF_HELPER_4(msa_msubv_b, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_msubv_h, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_msubv_w, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_msubv_d, void, env, i32, i32, i32)
+
 DEF_HELPER_4(msa_asub_s_b, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_asub_s_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_asub_s_w, void, env, i32, i32, i32)
@@ -1074,7 +1079,6 @@ DEF_HELPER_5(msa_subs_u_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_subsus_u_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_subsuu_s_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_mulv_df, void, env, i32, i32, i32, i32)
-DEF_HELPER_5(msa_msubv_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_dotp_s_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_dotp_u_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_dpadd_s_df, void, env, i32, i32, i32, i32)
diff --git a/target/mips/msa_helper.c b/target/mips/msa_helper.c
index 3b75bdc..2b54de0 100644
--- a/target/mips/msa_helper.c
+++ b/target/mips/msa_helper.c
@@ -2955,6 +2955,78 @@ void helper_msa_maddv_d(CPUMIPSState *env,
     pwd->d[1]  = msa_maddv_df(DF_DOUBLE, pwd->d[1],  pws->d[1],  pwt->d[1]);
 }
 
+static inline int64_t msa_msubv_df(uint32_t df, int64_t dest, int64_t arg1,
+                                   int64_t arg2)
+{
+    return dest - arg1 * arg2;
+}
+
+void helper_msa_msubv_b(CPUMIPSState *env,
+                        uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->b[0]  = msa_msubv_df(DF_BYTE, pwt->b[0],  pws->b[0],  pwt->b[0]);
+    pwd->b[1]  = msa_msubv_df(DF_BYTE, pwt->b[1],  pws->b[1],  pwt->b[1]);
+    pwd->b[2]  = msa_msubv_df(DF_BYTE, pwt->b[2],  pws->b[2],  pwt->b[2]);
+    pwd->b[3]  = msa_msubv_df(DF_BYTE, pwt->b[3],  pws->b[3],  pwt->b[3]);
+    pwd->b[4]  = msa_msubv_df(DF_BYTE, pwt->b[4],  pws->b[4],  pwt->b[4]);
+    pwd->b[5]  = msa_msubv_df(DF_BYTE, pwt->b[5],  pws->b[5],  pwt->b[5]);
+    pwd->b[6]  = msa_msubv_df(DF_BYTE, pwt->b[6],  pws->b[6],  pwt->b[6]);
+    pwd->b[7]  = msa_msubv_df(DF_BYTE, pwt->b[7],  pws->b[7],  pwt->b[7]);
+    pwd->b[8]  = msa_msubv_df(DF_BYTE, pwt->b[8],  pws->b[8],  pwt->b[8]);
+    pwd->b[9]  = msa_msubv_df(DF_BYTE, pwt->b[9],  pws->b[9],  pwt->b[9]);
+    pwd->b[10] = msa_msubv_df(DF_BYTE, pwt->b[10], pws->b[10], pwt->b[10]);
+    pwd->b[11] = msa_msubv_df(DF_BYTE, pwt->b[11], pws->b[11], pwt->b[11]);
+    pwd->b[12] = msa_msubv_df(DF_BYTE, pwt->b[12], pws->b[12], pwt->b[12]);
+    pwd->b[13] = msa_msubv_df(DF_BYTE, pwt->b[13], pws->b[13], pwt->b[13]);
+    pwd->b[14] = msa_msubv_df(DF_BYTE, pwt->b[14], pws->b[14], pwt->b[14]);
+    pwd->b[15] = msa_msubv_df(DF_BYTE, pwt->b[15], pws->b[15], pwt->b[15]);
+}
+
+void helper_msa_msubv_h(CPUMIPSState *env,
+                        uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->h[0]  = msa_msubv_df(DF_HALF, pwd->h[0],  pws->h[0],  pwt->h[0]);
+    pwd->h[1]  = msa_msubv_df(DF_HALF, pwd->h[1],  pws->h[1],  pwt->h[1]);
+    pwd->h[2]  = msa_msubv_df(DF_HALF, pwd->h[2],  pws->h[2],  pwt->h[2]);
+    pwd->h[3]  = msa_msubv_df(DF_HALF, pwd->h[3],  pws->h[3],  pwt->h[3]);
+    pwd->h[4]  = msa_msubv_df(DF_HALF, pwd->h[4],  pws->h[4],  pwt->h[4]);
+    pwd->h[5]  = msa_msubv_df(DF_HALF, pwd->h[5],  pws->h[5],  pwt->h[5]);
+    pwd->h[6]  = msa_msubv_df(DF_HALF, pwd->h[6],  pws->h[6],  pwt->h[6]);
+    pwd->h[7]  = msa_msubv_df(DF_HALF, pwd->h[7],  pws->h[7],  pwt->h[7]);
+}
+
+void helper_msa_msubv_w(CPUMIPSState *env,
+                        uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->w[0]  = msa_msubv_df(DF_WORD, pwd->w[0],  pws->w[0],  pwt->w[0]);
+    pwd->w[1]  = msa_msubv_df(DF_WORD, pwd->w[1],  pws->w[1],  pwt->w[1]);
+    pwd->w[2]  = msa_msubv_df(DF_WORD, pwd->w[2],  pws->w[2],  pwt->w[2]);
+    pwd->w[3]  = msa_msubv_df(DF_WORD, pwd->w[3],  pws->w[3],  pwt->w[3]);
+}
+
+void helper_msa_msubv_d(CPUMIPSState *env,
+                        uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->d[0]  = msa_msubv_df(DF_DOUBLE, pwd->d[0],  pws->d[0],  pwt->d[0]);
+    pwd->d[1]  = msa_msubv_df(DF_DOUBLE, pwd->d[1],  pws->d[1],  pwt->d[1]);
+}
+
 
 /*
  * Int Subtract
@@ -4886,12 +4958,6 @@ void helper_msa_sld_df(CPUMIPSState *env, uint32_t df, uint32_t wd,
     msa_sld_df(df, pwd, pws, env->active_tc.gpr[rt]);
 }
 
-static inline int64_t msa_msubv_df(uint32_t df, int64_t dest, int64_t arg1,
-                                   int64_t arg2)
-{
-    return dest - arg1 * arg2;
-}
-
 static inline int64_t msa_dpadd_s_df(uint32_t df, int64_t dest, int64_t arg1,
                                      int64_t arg2)
 {
@@ -5066,7 +5132,6 @@ void helper_msa_ ## func ## _df(CPUMIPSState *env, uint32_t df, uint32_t wd,  \
     }                                                                         \
 }
 
-MSA_TEROP_DF(msubv)
 MSA_TEROP_DF(dpadd_s)
 MSA_TEROP_DF(dpadd_u)
 MSA_TEROP_DF(dpsub_s)
diff --git a/target/mips/translate.c b/target/mips/translate.c
index 4975e57..7388a16 100644
--- a/target/mips/translate.c
+++ b/target/mips/translate.c
@@ -29074,6 +29074,22 @@ static void gen_msa_3r(CPUMIPSState *env, DisasContext *ctx)
             break;
         }
         break;
+    case OPC_MSUBV_df:
+        switch (df) {
+        case DF_BYTE:
+            gen_helper_msa_msubv_b(cpu_env, twd, tws, twt);
+            break;
+        case DF_HALF:
+            gen_helper_msa_msubv_h(cpu_env, twd, tws, twt);
+            break;
+        case DF_WORD:
+            gen_helper_msa_msubv_w(cpu_env, twd, tws, twt);
+            break;
+        case DF_DOUBLE:
+            gen_helper_msa_msubv_d(cpu_env, twd, tws, twt);
+            break;
+        }
+        break;
     case OPC_ASUB_S_df:
         switch (df) {
         case DF_BYTE:
@@ -29306,9 +29322,6 @@ static void gen_msa_3r(CPUMIPSState *env, DisasContext *ctx)
     case OPC_SUBSUS_U_df:
         gen_helper_msa_subsus_u_df(cpu_env, tdf, twd, tws, twt);
         break;
-    case OPC_MSUBV_df:
-        gen_helper_msa_msubv_df(cpu_env, tdf, twd, tws, twt);
-        break;
     case OPC_SUBSUU_S_df:
         gen_helper_msa_subsuu_s_df(cpu_env, tdf, twd, tws, twt);
         break;
-- 
2.7.4

[PULL 05/18] target/mips: msa: Split helpers for DPADD_S.<H|W|D>

[Aleksandar Markovic]

Achieves clearer code and slightly better performance.

Reviewed-by: Aleksandar Rikalo <aleksandar.rikalo@syrmia.com>
Signed-off-by: Aleksandar Markovic <aleksandar.qemu.devel@gmail.com>
Message-Id: <20200613152133.8964-4-aleksandar.qemu.devel@gmail.com>
---
 target/mips/helper.h     |  4 ++-
 target/mips/msa_helper.c | 90 ++++++++++++++++++++++++++++++++++--------------
 target/mips/translate.c  | 12 ++++++-
 3 files changed, 78 insertions(+), 28 deletions(-)

diff --git a/target/mips/helper.h b/target/mips/helper.h
index 7ca0036..16f2d53 100644
--- a/target/mips/helper.h
+++ b/target/mips/helper.h
@@ -1081,7 +1081,9 @@ DEF_HELPER_5(msa_subsuu_s_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_mulv_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_dotp_s_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_dotp_u_df, void, env, i32, i32, i32, i32)
-DEF_HELPER_5(msa_dpadd_s_df, void, env, i32, i32, i32, i32)
+DEF_HELPER_4(msa_dpadd_s_h, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_dpadd_s_w, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_dpadd_s_d, void, env, i32, i32, i32)
 DEF_HELPER_5(msa_dpadd_u_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_dpsub_s_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_dpsub_u_df, void, env, i32, i32, i32, i32)
diff --git a/target/mips/msa_helper.c b/target/mips/msa_helper.c
index 2b54de0..086b56f 100644
--- a/target/mips/msa_helper.c
+++ b/target/mips/msa_helper.c
@@ -2224,7 +2224,70 @@ void helper_msa_div_u_d(CPUMIPSState *env,
  * +---------------+----------------------------------------------------------+
  */
 
-/* TODO: insert Int Dot Product group helpers here */
+#define SIGNED_EXTRACT(e, o, a, df)     \
+    do {                                \
+        e = SIGNED_EVEN(a, df);         \
+        o = SIGNED_ODD(a, df);          \
+    } while (0)
+
+#define UNSIGNED_EXTRACT(e, o, a, df)   \
+    do {                                \
+        e = UNSIGNED_EVEN(a, df);       \
+        o = UNSIGNED_ODD(a, df);        \
+    } while (0)
+
+static inline int64_t msa_dpadd_s_df(uint32_t df, int64_t dest, int64_t arg1,
+                                     int64_t arg2)
+{
+    int64_t even_arg1;
+    int64_t even_arg2;
+    int64_t odd_arg1;
+    int64_t odd_arg2;
+    SIGNED_EXTRACT(even_arg1, odd_arg1, arg1, df);
+    SIGNED_EXTRACT(even_arg2, odd_arg2, arg2, df);
+    return dest + (even_arg1 * even_arg2) + (odd_arg1 * odd_arg2);
+}
+
+void helper_msa_dpadd_s_h(CPUMIPSState *env,
+                          uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->h[0]  = msa_dpadd_s_df(DF_HALF, pwd->h[0],  pws->h[0],  pwt->h[0]);
+    pwd->h[1]  = msa_dpadd_s_df(DF_HALF, pwd->h[1],  pws->h[1],  pwt->h[1]);
+    pwd->h[2]  = msa_dpadd_s_df(DF_HALF, pwd->h[2],  pws->h[2],  pwt->h[2]);
+    pwd->h[3]  = msa_dpadd_s_df(DF_HALF, pwd->h[3],  pws->h[3],  pwt->h[3]);
+    pwd->h[4]  = msa_dpadd_s_df(DF_HALF, pwd->h[4],  pws->h[4],  pwt->h[4]);
+    pwd->h[5]  = msa_dpadd_s_df(DF_HALF, pwd->h[5],  pws->h[5],  pwt->h[5]);
+    pwd->h[6]  = msa_dpadd_s_df(DF_HALF, pwd->h[6],  pws->h[6],  pwt->h[6]);
+    pwd->h[7]  = msa_dpadd_s_df(DF_HALF, pwd->h[7],  pws->h[7],  pwt->h[7]);
+}
+
+void helper_msa_dpadd_s_w(CPUMIPSState *env,
+                          uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->w[0]  = msa_dpadd_s_df(DF_WORD, pwd->w[0],  pws->w[0],  pwt->w[0]);
+    pwd->w[1]  = msa_dpadd_s_df(DF_WORD, pwd->w[1],  pws->w[1],  pwt->w[1]);
+    pwd->w[2]  = msa_dpadd_s_df(DF_WORD, pwd->w[2],  pws->w[2],  pwt->w[2]);
+    pwd->w[3]  = msa_dpadd_s_df(DF_WORD, pwd->w[3],  pws->w[3],  pwt->w[3]);
+}
+
+void helper_msa_dpadd_s_d(CPUMIPSState *env,
+                          uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->d[0]  = msa_dpadd_s_df(DF_DOUBLE, pwd->d[0],  pws->d[0],  pwt->d[0]);
+    pwd->d[1]  = msa_dpadd_s_df(DF_DOUBLE, pwd->d[1],  pws->d[1],  pwt->d[1]);
+}
 
 
 /*
@@ -4785,18 +4848,6 @@ static inline int64_t msa_mulv_df(uint32_t df, int64_t arg1, int64_t arg2)
     return arg1 * arg2;
 }
 
-#define SIGNED_EXTRACT(e, o, a, df)     \
-    do {                                \
-        e = SIGNED_EVEN(a, df);         \
-        o = SIGNED_ODD(a, df);          \
-    } while (0)
-
-#define UNSIGNED_EXTRACT(e, o, a, df)   \
-    do {                                \
-        e = UNSIGNED_EVEN(a, df);       \
-        o = UNSIGNED_ODD(a, df);        \
-    } while (0)
-
 static inline int64_t msa_dotp_s_df(uint32_t df, int64_t arg1, int64_t arg2)
 {
     int64_t even_arg1;
@@ -4958,18 +5009,6 @@ void helper_msa_sld_df(CPUMIPSState *env, uint32_t df, uint32_t wd,
     msa_sld_df(df, pwd, pws, env->active_tc.gpr[rt]);
 }
 
-static inline int64_t msa_dpadd_s_df(uint32_t df, int64_t dest, int64_t arg1,
-                                     int64_t arg2)
-{
-    int64_t even_arg1;
-    int64_t even_arg2;
-    int64_t odd_arg1;
-    int64_t odd_arg2;
-    SIGNED_EXTRACT(even_arg1, odd_arg1, arg1, df);
-    SIGNED_EXTRACT(even_arg2, odd_arg2, arg2, df);
-    return dest + (even_arg1 * even_arg2) + (odd_arg1 * odd_arg2);
-}
-
 static inline int64_t msa_dpadd_u_df(uint32_t df, int64_t dest, int64_t arg1,
                                      int64_t arg2)
 {
@@ -5132,7 +5171,6 @@ void helper_msa_ ## func ## _df(CPUMIPSState *env, uint32_t df, uint32_t wd,  \
     }                                                                         \
 }
 
-MSA_TEROP_DF(dpadd_s)
 MSA_TEROP_DF(dpadd_u)
 MSA_TEROP_DF(dpsub_s)
 MSA_TEROP_DF(dpsub_u)
diff --git a/target/mips/translate.c b/target/mips/translate.c
index 7388a16..47ac4ea 100644
--- a/target/mips/translate.c
+++ b/target/mips/translate.c
@@ -29400,7 +29400,17 @@ static void gen_msa_3r(CPUMIPSState *env, DisasContext *ctx)
             gen_helper_msa_dotp_u_df(cpu_env, tdf, twd, tws, twt);
             break;
         case OPC_DPADD_S_df:
-            gen_helper_msa_dpadd_s_df(cpu_env, tdf, twd, tws, twt);
+            switch (df) {
+            case DF_HALF:
+                gen_helper_msa_dpadd_s_h(cpu_env, twd, tws, twt);
+                break;
+            case DF_WORD:
+                gen_helper_msa_dpadd_s_w(cpu_env, twd, tws, twt);
+                break;
+            case DF_DOUBLE:
+                gen_helper_msa_dpadd_s_d(cpu_env, twd, tws, twt);
+                break;
+            }
             break;
         case OPC_DPADD_U_df:
             gen_helper_msa_dpadd_u_df(cpu_env, tdf, twd, tws, twt);
-- 
2.7.4

[PULL 06/18] target/mips: msa: Split helpers for DPADD_U.<H|W|D>

[Aleksandar Markovic]

Achieves clearer code and slightly better performance.

Reviewed-by: Aleksandar Rikalo <aleksandar.rikalo@syrmia.com>
Signed-off-by: Aleksandar Markovic <aleksandar.qemu.devel@gmail.com>
Message-Id: <20200613152133.8964-5-aleksandar.qemu.devel@gmail.com>
---
 target/mips/helper.h     |  4 ++-
 target/mips/msa_helper.c | 67 ++++++++++++++++++++++++++++++++++++++----------
 target/mips/translate.c  | 12 ++++++++-
 3 files changed, 68 insertions(+), 15 deletions(-)

diff --git a/target/mips/helper.h b/target/mips/helper.h
index 16f2d53..155b6bb 100644
--- a/target/mips/helper.h
+++ b/target/mips/helper.h
@@ -1084,7 +1084,9 @@ DEF_HELPER_5(msa_dotp_u_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_4(msa_dpadd_s_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_dpadd_s_w, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_dpadd_s_d, void, env, i32, i32, i32)
-DEF_HELPER_5(msa_dpadd_u_df, void, env, i32, i32, i32, i32)
+DEF_HELPER_4(msa_dpadd_u_h, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_dpadd_u_w, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_dpadd_u_d, void, env, i32, i32, i32)
 DEF_HELPER_5(msa_dpsub_s_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_dpsub_u_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_sld_df, void, env, i32, i32, i32, i32)
diff --git a/target/mips/msa_helper.c b/target/mips/msa_helper.c
index 086b56f..9741c94 100644
--- a/target/mips/msa_helper.c
+++ b/target/mips/msa_helper.c
@@ -2290,6 +2290,60 @@ void helper_msa_dpadd_s_d(CPUMIPSState *env,
 }
 
 
+static inline int64_t msa_dpadd_u_df(uint32_t df, int64_t dest, int64_t arg1,
+                                     int64_t arg2)
+{
+    int64_t even_arg1;
+    int64_t even_arg2;
+    int64_t odd_arg1;
+    int64_t odd_arg2;
+    UNSIGNED_EXTRACT(even_arg1, odd_arg1, arg1, df);
+    UNSIGNED_EXTRACT(even_arg2, odd_arg2, arg2, df);
+    return dest + (even_arg1 * even_arg2) + (odd_arg1 * odd_arg2);
+}
+
+void helper_msa_dpadd_u_h(CPUMIPSState *env,
+                          uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->h[0]  = msa_dpadd_u_df(DF_HALF, pwd->h[0],  pws->h[0],  pwt->h[0]);
+    pwd->h[1]  = msa_dpadd_u_df(DF_HALF, pwd->h[1],  pws->h[1],  pwt->h[1]);
+    pwd->h[2]  = msa_dpadd_u_df(DF_HALF, pwd->h[2],  pws->h[2],  pwt->h[2]);
+    pwd->h[3]  = msa_dpadd_u_df(DF_HALF, pwd->h[3],  pws->h[3],  pwt->h[3]);
+    pwd->h[4]  = msa_dpadd_u_df(DF_HALF, pwd->h[4],  pws->h[4],  pwt->h[4]);
+    pwd->h[5]  = msa_dpadd_u_df(DF_HALF, pwd->h[5],  pws->h[5],  pwt->h[5]);
+    pwd->h[6]  = msa_dpadd_u_df(DF_HALF, pwd->h[6],  pws->h[6],  pwt->h[6]);
+    pwd->h[7]  = msa_dpadd_u_df(DF_HALF, pwd->h[7],  pws->h[7],  pwt->h[7]);
+}
+
+void helper_msa_dpadd_u_w(CPUMIPSState *env,
+                          uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->w[0]  = msa_dpadd_u_df(DF_WORD, pwd->w[0],  pws->w[0],  pwt->w[0]);
+    pwd->w[1]  = msa_dpadd_u_df(DF_WORD, pwd->w[1],  pws->w[1],  pwt->w[1]);
+    pwd->w[2]  = msa_dpadd_u_df(DF_WORD, pwd->w[2],  pws->w[2],  pwt->w[2]);
+    pwd->w[3]  = msa_dpadd_u_df(DF_WORD, pwd->w[3],  pws->w[3],  pwt->w[3]);
+}
+
+void helper_msa_dpadd_u_d(CPUMIPSState *env,
+                          uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->d[0]  = msa_dpadd_u_df(DF_DOUBLE, pwd->d[0],  pws->d[0],  pwt->d[0]);
+    pwd->d[1]  = msa_dpadd_u_df(DF_DOUBLE, pwd->d[1],  pws->d[1],  pwt->d[1]);
+}
+
+
 /*
  * Int Max Min
  * -----------
@@ -5009,18 +5063,6 @@ void helper_msa_sld_df(CPUMIPSState *env, uint32_t df, uint32_t wd,
     msa_sld_df(df, pwd, pws, env->active_tc.gpr[rt]);
 }
 
-static inline int64_t msa_dpadd_u_df(uint32_t df, int64_t dest, int64_t arg1,
-                                     int64_t arg2)
-{
-    int64_t even_arg1;
-    int64_t even_arg2;
-    int64_t odd_arg1;
-    int64_t odd_arg2;
-    UNSIGNED_EXTRACT(even_arg1, odd_arg1, arg1, df);
-    UNSIGNED_EXTRACT(even_arg2, odd_arg2, arg2, df);
-    return dest + (even_arg1 * even_arg2) + (odd_arg1 * odd_arg2);
-}
-
 static inline int64_t msa_dpsub_s_df(uint32_t df, int64_t dest, int64_t arg1,
                                      int64_t arg2)
 {
@@ -5171,7 +5213,6 @@ void helper_msa_ ## func ## _df(CPUMIPSState *env, uint32_t df, uint32_t wd,  \
     }                                                                         \
 }
 
-MSA_TEROP_DF(dpadd_u)
 MSA_TEROP_DF(dpsub_s)
 MSA_TEROP_DF(dpsub_u)
 MSA_TEROP_DF(binsl)
diff --git a/target/mips/translate.c b/target/mips/translate.c
index 47ac4ea..0f99768 100644
--- a/target/mips/translate.c
+++ b/target/mips/translate.c
@@ -29413,7 +29413,17 @@ static void gen_msa_3r(CPUMIPSState *env, DisasContext *ctx)
             }
             break;
         case OPC_DPADD_U_df:
-            gen_helper_msa_dpadd_u_df(cpu_env, tdf, twd, tws, twt);
+            switch (df) {
+            case DF_HALF:
+                gen_helper_msa_dpadd_u_h(cpu_env, twd, tws, twt);
+                break;
+            case DF_WORD:
+                gen_helper_msa_dpadd_u_w(cpu_env, twd, tws, twt);
+                break;
+            case DF_DOUBLE:
+                gen_helper_msa_dpadd_u_d(cpu_env, twd, tws, twt);
+                break;
+            }
             break;
         case OPC_DPSUB_S_df:
             gen_helper_msa_dpsub_s_df(cpu_env, tdf, twd, tws, twt);
-- 
2.7.4

[PULL 07/18] target/mips: msa: Split helpers for DPSUB_S.<H|W|D>

[Aleksandar Markovic]

Achieves clearer code and slightly better performance.

Reviewed-by: Aleksandar Rikalo <aleksandar.rikalo@syrmia.com>
Signed-off-by: Aleksandar Markovic <aleksandar.qemu.devel@gmail.com>
Message-Id: <20200613152133.8964-6-aleksandar.qemu.devel@gmail.com>
---
 target/mips/helper.h     |  4 ++-
 target/mips/msa_helper.c | 67 ++++++++++++++++++++++++++++++++++++++----------
 target/mips/translate.c  | 12 ++++++++-
 3 files changed, 68 insertions(+), 15 deletions(-)

diff --git a/target/mips/helper.h b/target/mips/helper.h
index 155b6bb..2de1454 100644
--- a/target/mips/helper.h
+++ b/target/mips/helper.h
@@ -1087,7 +1087,9 @@ DEF_HELPER_4(msa_dpadd_s_d, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_dpadd_u_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_dpadd_u_w, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_dpadd_u_d, void, env, i32, i32, i32)
-DEF_HELPER_5(msa_dpsub_s_df, void, env, i32, i32, i32, i32)
+DEF_HELPER_4(msa_dpsub_s_h, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_dpsub_s_w, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_dpsub_s_d, void, env, i32, i32, i32)
 DEF_HELPER_5(msa_dpsub_u_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_sld_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_splat_df, void, env, i32, i32, i32, i32)
diff --git a/target/mips/msa_helper.c b/target/mips/msa_helper.c
index 9741c94..934f705 100644
--- a/target/mips/msa_helper.c
+++ b/target/mips/msa_helper.c
@@ -2344,6 +2344,60 @@ void helper_msa_dpadd_u_d(CPUMIPSState *env,
 }
 
 
+static inline int64_t msa_dpsub_s_df(uint32_t df, int64_t dest, int64_t arg1,
+                                     int64_t arg2)
+{
+    int64_t even_arg1;
+    int64_t even_arg2;
+    int64_t odd_arg1;
+    int64_t odd_arg2;
+    SIGNED_EXTRACT(even_arg1, odd_arg1, arg1, df);
+    SIGNED_EXTRACT(even_arg2, odd_arg2, arg2, df);
+    return dest - ((even_arg1 * even_arg2) + (odd_arg1 * odd_arg2));
+}
+
+void helper_msa_dpsub_s_h(CPUMIPSState *env,
+                          uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->h[0]  = msa_dpsub_s_df(DF_HALF, pwd->h[0],  pws->h[0],  pwt->h[0]);
+    pwd->h[1]  = msa_dpsub_s_df(DF_HALF, pwd->h[1],  pws->h[1],  pwt->h[1]);
+    pwd->h[2]  = msa_dpsub_s_df(DF_HALF, pwd->h[2],  pws->h[2],  pwt->h[2]);
+    pwd->h[3]  = msa_dpsub_s_df(DF_HALF, pwd->h[3],  pws->h[3],  pwt->h[3]);
+    pwd->h[4]  = msa_dpsub_s_df(DF_HALF, pwd->h[4],  pws->h[4],  pwt->h[4]);
+    pwd->h[5]  = msa_dpsub_s_df(DF_HALF, pwd->h[5],  pws->h[5],  pwt->h[5]);
+    pwd->h[6]  = msa_dpsub_s_df(DF_HALF, pwd->h[6],  pws->h[6],  pwt->h[6]);
+    pwd->h[7]  = msa_dpsub_s_df(DF_HALF, pwd->h[7],  pws->h[7],  pwt->h[7]);
+}
+
+void helper_msa_dpsub_s_w(CPUMIPSState *env,
+                          uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->w[0]  = msa_dpsub_s_df(DF_WORD, pwd->w[0],  pws->w[0],  pwt->w[0]);
+    pwd->w[1]  = msa_dpsub_s_df(DF_WORD, pwd->w[1],  pws->w[1],  pwt->w[1]);
+    pwd->w[2]  = msa_dpsub_s_df(DF_WORD, pwd->w[2],  pws->w[2],  pwt->w[2]);
+    pwd->w[3]  = msa_dpsub_s_df(DF_WORD, pwd->w[3],  pws->w[3],  pwt->w[3]);
+}
+
+void helper_msa_dpsub_s_d(CPUMIPSState *env,
+                          uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->d[0]  = msa_dpsub_s_df(DF_DOUBLE, pwd->d[0],  pws->d[0],  pwt->d[0]);
+    pwd->d[1]  = msa_dpsub_s_df(DF_DOUBLE, pwd->d[1],  pws->d[1],  pwt->d[1]);
+}
+
+
 /*
  * Int Max Min
  * -----------
@@ -5063,18 +5117,6 @@ void helper_msa_sld_df(CPUMIPSState *env, uint32_t df, uint32_t wd,
     msa_sld_df(df, pwd, pws, env->active_tc.gpr[rt]);
 }
 
-static inline int64_t msa_dpsub_s_df(uint32_t df, int64_t dest, int64_t arg1,
-                                     int64_t arg2)
-{
-    int64_t even_arg1;
-    int64_t even_arg2;
-    int64_t odd_arg1;
-    int64_t odd_arg2;
-    SIGNED_EXTRACT(even_arg1, odd_arg1, arg1, df);
-    SIGNED_EXTRACT(even_arg2, odd_arg2, arg2, df);
-    return dest - ((even_arg1 * even_arg2) + (odd_arg1 * odd_arg2));
-}
-
 static inline int64_t msa_dpsub_u_df(uint32_t df, int64_t dest, int64_t arg1,
                                      int64_t arg2)
 {
@@ -5213,7 +5255,6 @@ void helper_msa_ ## func ## _df(CPUMIPSState *env, uint32_t df, uint32_t wd,  \
     }                                                                         \
 }
 
-MSA_TEROP_DF(dpsub_s)
 MSA_TEROP_DF(dpsub_u)
 MSA_TEROP_DF(binsl)
 MSA_TEROP_DF(binsr)
diff --git a/target/mips/translate.c b/target/mips/translate.c
index 0f99768..caf3adf 100644
--- a/target/mips/translate.c
+++ b/target/mips/translate.c
@@ -29426,7 +29426,17 @@ static void gen_msa_3r(CPUMIPSState *env, DisasContext *ctx)
             }
             break;
         case OPC_DPSUB_S_df:
-            gen_helper_msa_dpsub_s_df(cpu_env, tdf, twd, tws, twt);
+            switch (df) {
+            case DF_HALF:
+                gen_helper_msa_dpsub_s_h(cpu_env, twd, tws, twt);
+                break;
+            case DF_WORD:
+                gen_helper_msa_dpsub_s_w(cpu_env, twd, tws, twt);
+                break;
+            case DF_DOUBLE:
+                gen_helper_msa_dpsub_s_d(cpu_env, twd, tws, twt);
+                break;
+            }
             break;
         case OPC_DPSUB_U_df:
             gen_helper_msa_dpsub_u_df(cpu_env, tdf, twd, tws, twt);
-- 
2.7.4

[PULL 08/18] target/mips: msa: Split helpers for DPSUB_U.<H|W|D>

[Aleksandar Markovic]

Achieves clearer code and slightly better performance.

Reviewed-by: Aleksandar Rikalo <aleksandar.rikalo@syrmia.com>
Signed-off-by: Aleksandar Markovic <aleksandar.qemu.devel@gmail.com>
Message-Id: <20200613152133.8964-7-aleksandar.qemu.devel@gmail.com>
---
 target/mips/helper.h     |  4 ++-
 target/mips/msa_helper.c | 67 ++++++++++++++++++++++++++++++++++++++----------
 target/mips/translate.c  | 12 ++++++++-
 3 files changed, 68 insertions(+), 15 deletions(-)

diff --git a/target/mips/helper.h b/target/mips/helper.h
index 2de1454..575f4a5 100644
--- a/target/mips/helper.h
+++ b/target/mips/helper.h
@@ -1090,7 +1090,9 @@ DEF_HELPER_4(msa_dpadd_u_d, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_dpsub_s_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_dpsub_s_w, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_dpsub_s_d, void, env, i32, i32, i32)
-DEF_HELPER_5(msa_dpsub_u_df, void, env, i32, i32, i32, i32)
+DEF_HELPER_4(msa_dpsub_u_h, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_dpsub_u_w, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_dpsub_u_d, void, env, i32, i32, i32)
 DEF_HELPER_5(msa_sld_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_splat_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_vshf_df, void, env, i32, i32, i32, i32)
diff --git a/target/mips/msa_helper.c b/target/mips/msa_helper.c
index 934f705..33d5251 100644
--- a/target/mips/msa_helper.c
+++ b/target/mips/msa_helper.c
@@ -2398,6 +2398,60 @@ void helper_msa_dpsub_s_d(CPUMIPSState *env,
 }
 
 
+static inline int64_t msa_dpsub_u_df(uint32_t df, int64_t dest, int64_t arg1,
+                                     int64_t arg2)
+{
+    int64_t even_arg1;
+    int64_t even_arg2;
+    int64_t odd_arg1;
+    int64_t odd_arg2;
+    UNSIGNED_EXTRACT(even_arg1, odd_arg1, arg1, df);
+    UNSIGNED_EXTRACT(even_arg2, odd_arg2, arg2, df);
+    return dest - ((even_arg1 * even_arg2) + (odd_arg1 * odd_arg2));
+}
+
+void helper_msa_dpsub_u_h(CPUMIPSState *env,
+                          uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->h[0]  = msa_dpsub_u_df(DF_HALF, pwd->h[0],  pws->h[0],  pwt->h[0]);
+    pwd->h[1]  = msa_dpsub_u_df(DF_HALF, pwd->h[1],  pws->h[1],  pwt->h[1]);
+    pwd->h[2]  = msa_dpsub_u_df(DF_HALF, pwd->h[2],  pws->h[2],  pwt->h[2]);
+    pwd->h[3]  = msa_dpsub_u_df(DF_HALF, pwd->h[3],  pws->h[3],  pwt->h[3]);
+    pwd->h[4]  = msa_dpsub_u_df(DF_HALF, pwd->h[4],  pws->h[4],  pwt->h[4]);
+    pwd->h[5]  = msa_dpsub_u_df(DF_HALF, pwd->h[5],  pws->h[5],  pwt->h[5]);
+    pwd->h[6]  = msa_dpsub_u_df(DF_HALF, pwd->h[6],  pws->h[6],  pwt->h[6]);
+    pwd->h[7]  = msa_dpsub_u_df(DF_HALF, pwd->h[7],  pws->h[7],  pwt->h[7]);
+}
+
+void helper_msa_dpsub_u_w(CPUMIPSState *env,
+                          uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->w[0]  = msa_dpsub_u_df(DF_WORD, pwd->w[0],  pws->w[0],  pwt->w[0]);
+    pwd->w[1]  = msa_dpsub_u_df(DF_WORD, pwd->w[1],  pws->w[1],  pwt->w[1]);
+    pwd->w[2]  = msa_dpsub_u_df(DF_WORD, pwd->w[2],  pws->w[2],  pwt->w[2]);
+    pwd->w[3]  = msa_dpsub_u_df(DF_WORD, pwd->w[3],  pws->w[3],  pwt->w[3]);
+}
+
+void helper_msa_dpsub_u_d(CPUMIPSState *env,
+                          uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->d[0]  = msa_dpsub_u_df(DF_DOUBLE, pwd->d[0],  pws->d[0],  pwt->d[0]);
+    pwd->d[1]  = msa_dpsub_u_df(DF_DOUBLE, pwd->d[1],  pws->d[1],  pwt->d[1]);
+}
+
+
 /*
  * Int Max Min
  * -----------
@@ -5117,18 +5171,6 @@ void helper_msa_sld_df(CPUMIPSState *env, uint32_t df, uint32_t wd,
     msa_sld_df(df, pwd, pws, env->active_tc.gpr[rt]);
 }
 
-static inline int64_t msa_dpsub_u_df(uint32_t df, int64_t dest, int64_t arg1,
-                                     int64_t arg2)
-{
-    int64_t even_arg1;
-    int64_t even_arg2;
-    int64_t odd_arg1;
-    int64_t odd_arg2;
-    UNSIGNED_EXTRACT(even_arg1, odd_arg1, arg1, df);
-    UNSIGNED_EXTRACT(even_arg2, odd_arg2, arg2, df);
-    return dest - ((even_arg1 * even_arg2) + (odd_arg1 * odd_arg2));
-}
-
 static inline int64_t msa_madd_q_df(uint32_t df, int64_t dest, int64_t arg1,
                                     int64_t arg2)
 {
@@ -5255,7 +5297,6 @@ void helper_msa_ ## func ## _df(CPUMIPSState *env, uint32_t df, uint32_t wd,  \
     }                                                                         \
 }
 
-MSA_TEROP_DF(dpsub_u)
 MSA_TEROP_DF(binsl)
 MSA_TEROP_DF(binsr)
 MSA_TEROP_DF(madd_q)
diff --git a/target/mips/translate.c b/target/mips/translate.c
index caf3adf..bd6febc 100644
--- a/target/mips/translate.c
+++ b/target/mips/translate.c
@@ -29439,7 +29439,17 @@ static void gen_msa_3r(CPUMIPSState *env, DisasContext *ctx)
             }
             break;
         case OPC_DPSUB_U_df:
-            gen_helper_msa_dpsub_u_df(cpu_env, tdf, twd, tws, twt);
+            switch (df) {
+            case DF_HALF:
+                gen_helper_msa_dpsub_u_h(cpu_env, twd, tws, twt);
+                break;
+            case DF_WORD:
+                gen_helper_msa_dpsub_u_w(cpu_env, twd, tws, twt);
+                break;
+            case DF_DOUBLE:
+                gen_helper_msa_dpsub_u_d(cpu_env, twd, tws, twt);
+                break;
+            }
             break;
         }
         break;
-- 
2.7.4

[PULL 09/18] target/mips: msa: Split helpers for DOTP_S.<H|W|D>

[Aleksandar Markovic]

Achieves clearer code and slightly better performance.

Reviewed-by: Aleksandar Rikalo <aleksandar.rikalo@syrmia.com>
Signed-off-by: Aleksandar Markovic <aleksandar.qemu.devel@gmail.com>
Message-Id: <20200613152133.8964-8-aleksandar.qemu.devel@gmail.com>
---
 target/mips/helper.h     |  5 +++-
 target/mips/msa_helper.c | 66 +++++++++++++++++++++++++++++++++++++++---------
 target/mips/translate.c  | 12 ++++++++-
 3 files changed, 69 insertions(+), 14 deletions(-)

diff --git a/target/mips/helper.h b/target/mips/helper.h
index 575f4a5..06df3de 100644
--- a/target/mips/helper.h
+++ b/target/mips/helper.h
@@ -1079,7 +1079,10 @@ DEF_HELPER_5(msa_subs_u_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_subsus_u_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_subsuu_s_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_mulv_df, void, env, i32, i32, i32, i32)
-DEF_HELPER_5(msa_dotp_s_df, void, env, i32, i32, i32, i32)
+
+DEF_HELPER_4(msa_dotp_s_h, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_dotp_s_w, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_dotp_s_d, void, env, i32, i32, i32)
 DEF_HELPER_5(msa_dotp_u_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_4(msa_dpadd_s_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_dpadd_s_w, void, env, i32, i32, i32)
diff --git a/target/mips/msa_helper.c b/target/mips/msa_helper.c
index 33d5251..201283f 100644
--- a/target/mips/msa_helper.c
+++ b/target/mips/msa_helper.c
@@ -2236,6 +2236,60 @@ void helper_msa_div_u_d(CPUMIPSState *env,
         o = UNSIGNED_ODD(a, df);        \
     } while (0)
 
+
+static inline int64_t msa_dotp_s_df(uint32_t df, int64_t arg1, int64_t arg2)
+{
+    int64_t even_arg1;
+    int64_t even_arg2;
+    int64_t odd_arg1;
+    int64_t odd_arg2;
+    SIGNED_EXTRACT(even_arg1, odd_arg1, arg1, df);
+    SIGNED_EXTRACT(even_arg2, odd_arg2, arg2, df);
+    return (even_arg1 * even_arg2) + (odd_arg1 * odd_arg2);
+}
+
+void helper_msa_dotp_s_h(CPUMIPSState *env,
+                         uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->h[0]  = msa_dotp_s_df(DF_HALF, pws->h[0],  pwt->h[0]);
+    pwd->h[1]  = msa_dotp_s_df(DF_HALF, pws->h[1],  pwt->h[1]);
+    pwd->h[2]  = msa_dotp_s_df(DF_HALF, pws->h[2],  pwt->h[2]);
+    pwd->h[3]  = msa_dotp_s_df(DF_HALF, pws->h[3],  pwt->h[3]);
+    pwd->h[4]  = msa_dotp_s_df(DF_HALF, pws->h[4],  pwt->h[4]);
+    pwd->h[5]  = msa_dotp_s_df(DF_HALF, pws->h[5],  pwt->h[5]);
+    pwd->h[6]  = msa_dotp_s_df(DF_HALF, pws->h[6],  pwt->h[6]);
+    pwd->h[7]  = msa_dotp_s_df(DF_HALF, pws->h[7],  pwt->h[7]);
+}
+
+void helper_msa_dotp_s_w(CPUMIPSState *env,
+                         uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->w[0]  = msa_dotp_s_df(DF_WORD, pws->w[0],  pwt->w[0]);
+    pwd->w[1]  = msa_dotp_s_df(DF_WORD, pws->w[1],  pwt->w[1]);
+    pwd->w[2]  = msa_dotp_s_df(DF_WORD, pws->w[2],  pwt->w[2]);
+    pwd->w[3]  = msa_dotp_s_df(DF_WORD, pws->w[3],  pwt->w[3]);
+}
+
+void helper_msa_dotp_s_d(CPUMIPSState *env,
+                         uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->d[0]  = msa_dotp_s_df(DF_DOUBLE, pws->d[0],  pwt->d[0]);
+    pwd->d[1]  = msa_dotp_s_df(DF_DOUBLE, pws->d[1],  pwt->d[1]);
+}
+
+
 static inline int64_t msa_dpadd_s_df(uint32_t df, int64_t dest, int64_t arg1,
                                      int64_t arg2)
 {
@@ -5010,17 +5064,6 @@ static inline int64_t msa_mulv_df(uint32_t df, int64_t arg1, int64_t arg2)
     return arg1 * arg2;
 }
 
-static inline int64_t msa_dotp_s_df(uint32_t df, int64_t arg1, int64_t arg2)
-{
-    int64_t even_arg1;
-    int64_t even_arg2;
-    int64_t odd_arg1;
-    int64_t odd_arg2;
-    SIGNED_EXTRACT(even_arg1, odd_arg1, arg1, df);
-    SIGNED_EXTRACT(even_arg2, odd_arg2, arg2, df);
-    return (even_arg1 * even_arg2) + (odd_arg1 * odd_arg2);
-}
-
 static inline int64_t msa_dotp_u_df(uint32_t df, int64_t arg1, int64_t arg2)
 {
     int64_t even_arg1;
@@ -5155,7 +5198,6 @@ MSA_BINOP_DF(subs_u)
 MSA_BINOP_DF(subsus_u)
 MSA_BINOP_DF(subsuu_s)
 MSA_BINOP_DF(mulv)
-MSA_BINOP_DF(dotp_s)
 MSA_BINOP_DF(dotp_u)
 
 MSA_BINOP_DF(mul_q)
diff --git a/target/mips/translate.c b/target/mips/translate.c
index bd6febc..e150454 100644
--- a/target/mips/translate.c
+++ b/target/mips/translate.c
@@ -29394,7 +29394,17 @@ static void gen_msa_3r(CPUMIPSState *env, DisasContext *ctx)
             }
             break;
         case OPC_DOTP_S_df:
-            gen_helper_msa_dotp_s_df(cpu_env, tdf, twd, tws, twt);
+            switch (df) {
+            case DF_HALF:
+                gen_helper_msa_dotp_s_h(cpu_env, twd, tws, twt);
+                break;
+            case DF_WORD:
+                gen_helper_msa_dotp_s_w(cpu_env, twd, tws, twt);
+                break;
+            case DF_DOUBLE:
+                gen_helper_msa_dotp_s_d(cpu_env, twd, tws, twt);
+                break;
+            }
             break;
         case OPC_DOTP_U_df:
             gen_helper_msa_dotp_u_df(cpu_env, tdf, twd, tws, twt);
-- 
2.7.4

[PULL 10/18] target/mips: msa: Split helpers for DOTP_U.<H|W|D>

[Aleksandar Markovic]

Achieves clearer code and slightly better performance.

Reviewed-by: Aleksandar Rikalo <aleksandar.rikalo@syrmia.com>
Signed-off-by: Aleksandar Markovic <aleksandar.qemu.devel@gmail.com>
Message-Id: <20200613152133.8964-9-aleksandar.qemu.devel@gmail.com>
---
 target/mips/helper.h     |  4 ++-
 target/mips/msa_helper.c | 65 +++++++++++++++++++++++++++++++++++++++---------
 target/mips/translate.c  | 12 ++++++++-
 3 files changed, 67 insertions(+), 14 deletions(-)

diff --git a/target/mips/helper.h b/target/mips/helper.h
index 06df3de..05d5533 100644
--- a/target/mips/helper.h
+++ b/target/mips/helper.h
@@ -1083,7 +1083,9 @@ DEF_HELPER_5(msa_mulv_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_4(msa_dotp_s_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_dotp_s_w, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_dotp_s_d, void, env, i32, i32, i32)
-DEF_HELPER_5(msa_dotp_u_df, void, env, i32, i32, i32, i32)
+DEF_HELPER_4(msa_dotp_u_h, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_dotp_u_w, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_dotp_u_d, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_dpadd_s_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_dpadd_s_w, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_dpadd_s_d, void, env, i32, i32, i32)
diff --git a/target/mips/msa_helper.c b/target/mips/msa_helper.c
index 201283f..84d0073 100644
--- a/target/mips/msa_helper.c
+++ b/target/mips/msa_helper.c
@@ -2290,6 +2290,59 @@ void helper_msa_dotp_s_d(CPUMIPSState *env,
 }
 
 
+static inline int64_t msa_dotp_u_df(uint32_t df, int64_t arg1, int64_t arg2)
+{
+    int64_t even_arg1;
+    int64_t even_arg2;
+    int64_t odd_arg1;
+    int64_t odd_arg2;
+    UNSIGNED_EXTRACT(even_arg1, odd_arg1, arg1, df);
+    UNSIGNED_EXTRACT(even_arg2, odd_arg2, arg2, df);
+    return (even_arg1 * even_arg2) + (odd_arg1 * odd_arg2);
+}
+
+void helper_msa_dotp_u_h(CPUMIPSState *env,
+                         uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->h[0]  = msa_dotp_u_df(DF_HALF, pws->h[0],  pwt->h[0]);
+    pwd->h[1]  = msa_dotp_u_df(DF_HALF, pws->h[1],  pwt->h[1]);
+    pwd->h[2]  = msa_dotp_u_df(DF_HALF, pws->h[2],  pwt->h[2]);
+    pwd->h[3]  = msa_dotp_u_df(DF_HALF, pws->h[3],  pwt->h[3]);
+    pwd->h[4]  = msa_dotp_u_df(DF_HALF, pws->h[4],  pwt->h[4]);
+    pwd->h[5]  = msa_dotp_u_df(DF_HALF, pws->h[5],  pwt->h[5]);
+    pwd->h[6]  = msa_dotp_u_df(DF_HALF, pws->h[6],  pwt->h[6]);
+    pwd->h[7]  = msa_dotp_u_df(DF_HALF, pws->h[7],  pwt->h[7]);
+}
+
+void helper_msa_dotp_u_w(CPUMIPSState *env,
+                         uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->w[0]  = msa_dotp_u_df(DF_WORD, pws->w[0],  pwt->w[0]);
+    pwd->w[1]  = msa_dotp_u_df(DF_WORD, pws->w[1],  pwt->w[1]);
+    pwd->w[2]  = msa_dotp_u_df(DF_WORD, pws->w[2],  pwt->w[2]);
+    pwd->w[3]  = msa_dotp_u_df(DF_WORD, pws->w[3],  pwt->w[3]);
+}
+
+void helper_msa_dotp_u_d(CPUMIPSState *env,
+                         uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->d[0]  = msa_dotp_u_df(DF_DOUBLE, pws->d[0],  pwt->d[0]);
+    pwd->d[1]  = msa_dotp_u_df(DF_DOUBLE, pws->d[1],  pwt->d[1]);
+}
+
+
 static inline int64_t msa_dpadd_s_df(uint32_t df, int64_t dest, int64_t arg1,
                                      int64_t arg2)
 {
@@ -5064,17 +5117,6 @@ static inline int64_t msa_mulv_df(uint32_t df, int64_t arg1, int64_t arg2)
     return arg1 * arg2;
 }
 
-static inline int64_t msa_dotp_u_df(uint32_t df, int64_t arg1, int64_t arg2)
-{
-    int64_t even_arg1;
-    int64_t even_arg2;
-    int64_t odd_arg1;
-    int64_t odd_arg2;
-    UNSIGNED_EXTRACT(even_arg1, odd_arg1, arg1, df);
-    UNSIGNED_EXTRACT(even_arg2, odd_arg2, arg2, df);
-    return (even_arg1 * even_arg2) + (odd_arg1 * odd_arg2);
-}
-
 #define CONCATENATE_AND_SLIDE(s, k)             \
     do {                                        \
         for (i = 0; i < s; i++) {               \
@@ -5198,7 +5240,6 @@ MSA_BINOP_DF(subs_u)
 MSA_BINOP_DF(subsus_u)
 MSA_BINOP_DF(subsuu_s)
 MSA_BINOP_DF(mulv)
-MSA_BINOP_DF(dotp_u)
 
 MSA_BINOP_DF(mul_q)
 MSA_BINOP_DF(mulr_q)
diff --git a/target/mips/translate.c b/target/mips/translate.c
index e150454..77e2d95 100644
--- a/target/mips/translate.c
+++ b/target/mips/translate.c
@@ -29407,7 +29407,17 @@ static void gen_msa_3r(CPUMIPSState *env, DisasContext *ctx)
             }
             break;
         case OPC_DOTP_U_df:
-            gen_helper_msa_dotp_u_df(cpu_env, tdf, twd, tws, twt);
+            switch (df) {
+            case DF_HALF:
+                gen_helper_msa_dotp_u_h(cpu_env, twd, tws, twt);
+                break;
+            case DF_WORD:
+                gen_helper_msa_dotp_u_w(cpu_env, twd, tws, twt);
+                break;
+            case DF_DOUBLE:
+                gen_helper_msa_dotp_u_d(cpu_env, twd, tws, twt);
+                break;
+            }
             break;
         case OPC_DPADD_S_df:
             switch (df) {
-- 
2.7.4

[PULL 11/18] target/mips: msa: Split helpers for SUBS_S.<B|H|W|D>

[Aleksandar Markovic]

Achieves clearer code and slightly better performance.

Reviewed-by: Aleksandar Rikalo <aleksandar.rikalo@syrmia.com>
Signed-off-by: Aleksandar Markovic <aleksandar.qemu.devel@gmail.com>
Message-Id: <20200613152133.8964-10-aleksandar.qemu.devel@gmail.com>
---
 target/mips/helper.h     |  6 +++-
 target/mips/msa_helper.c | 90 +++++++++++++++++++++++++++++++++++++++++-------
 target/mips/translate.c  | 15 +++++++-
 3 files changed, 97 insertions(+), 14 deletions(-)

diff --git a/target/mips/helper.h b/target/mips/helper.h
index 05d5533..a93402a 100644
--- a/target/mips/helper.h
+++ b/target/mips/helper.h
@@ -978,6 +978,11 @@ DEF_HELPER_4(msa_hsub_u_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_hsub_u_w, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_hsub_u_d, void, env, i32, i32, i32)
 
+DEF_HELPER_4(msa_subs_s_b, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_subs_s_h, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_subs_s_w, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_subs_s_d, void, env, i32, i32, i32)
+
 DEF_HELPER_4(msa_ilvev_b, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_ilvev_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_ilvev_w, void, env, i32, i32, i32)
@@ -1074,7 +1079,6 @@ DEF_HELPER_5(msa_srlri_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_binsl_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_binsr_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_subv_df, void, env, i32, i32, i32, i32)
-DEF_HELPER_5(msa_subs_s_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_subs_u_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_subsus_u_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_subsuu_s_df, void, env, i32, i32, i32, i32)
diff --git a/target/mips/msa_helper.c b/target/mips/msa_helper.c
index 84d0073..f08beba 100644
--- a/target/mips/msa_helper.c
+++ b/target/mips/msa_helper.c
@@ -3650,6 +3650,84 @@ void helper_msa_hsub_u_d(CPUMIPSState *env,
 }
 
 
+static inline int64_t msa_subs_s_df(uint32_t df, int64_t arg1, int64_t arg2)
+{
+    int64_t max_int = DF_MAX_INT(df);
+    int64_t min_int = DF_MIN_INT(df);
+    if (arg2 > 0) {
+        return (min_int + arg2 < arg1) ? arg1 - arg2 : min_int;
+    } else {
+        return (arg1 < max_int + arg2) ? arg1 - arg2 : max_int;
+    }
+}
+
+void helper_msa_subs_s_b(CPUMIPSState *env,
+                         uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->b[0]  = msa_subs_s_df(DF_BYTE, pws->b[0],  pwt->b[0]);
+    pwd->b[1]  = msa_subs_s_df(DF_BYTE, pws->b[1],  pwt->b[1]);
+    pwd->b[2]  = msa_subs_s_df(DF_BYTE, pws->b[2],  pwt->b[2]);
+    pwd->b[3]  = msa_subs_s_df(DF_BYTE, pws->b[3],  pwt->b[3]);
+    pwd->b[4]  = msa_subs_s_df(DF_BYTE, pws->b[4],  pwt->b[4]);
+    pwd->b[5]  = msa_subs_s_df(DF_BYTE, pws->b[5],  pwt->b[5]);
+    pwd->b[6]  = msa_subs_s_df(DF_BYTE, pws->b[6],  pwt->b[6]);
+    pwd->b[7]  = msa_subs_s_df(DF_BYTE, pws->b[7],  pwt->b[7]);
+    pwd->b[8]  = msa_subs_s_df(DF_BYTE, pws->b[8],  pwt->b[8]);
+    pwd->b[9]  = msa_subs_s_df(DF_BYTE, pws->b[9],  pwt->b[9]);
+    pwd->b[10] = msa_subs_s_df(DF_BYTE, pws->b[10], pwt->b[10]);
+    pwd->b[11] = msa_subs_s_df(DF_BYTE, pws->b[11], pwt->b[11]);
+    pwd->b[12] = msa_subs_s_df(DF_BYTE, pws->b[12], pwt->b[12]);
+    pwd->b[13] = msa_subs_s_df(DF_BYTE, pws->b[13], pwt->b[13]);
+    pwd->b[14] = msa_subs_s_df(DF_BYTE, pws->b[14], pwt->b[14]);
+    pwd->b[15] = msa_subs_s_df(DF_BYTE, pws->b[15], pwt->b[15]);
+}
+
+void helper_msa_subs_s_h(CPUMIPSState *env,
+                         uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->h[0]  = msa_subs_s_df(DF_HALF, pws->h[0],  pwt->h[0]);
+    pwd->h[1]  = msa_subs_s_df(DF_HALF, pws->h[1],  pwt->h[1]);
+    pwd->h[2]  = msa_subs_s_df(DF_HALF, pws->h[2],  pwt->h[2]);
+    pwd->h[3]  = msa_subs_s_df(DF_HALF, pws->h[3],  pwt->h[3]);
+    pwd->h[4]  = msa_subs_s_df(DF_HALF, pws->h[4],  pwt->h[4]);
+    pwd->h[5]  = msa_subs_s_df(DF_HALF, pws->h[5],  pwt->h[5]);
+    pwd->h[6]  = msa_subs_s_df(DF_HALF, pws->h[6],  pwt->h[6]);
+    pwd->h[7]  = msa_subs_s_df(DF_HALF, pws->h[7],  pwt->h[7]);
+}
+
+void helper_msa_subs_s_w(CPUMIPSState *env,
+                         uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->w[0]  = msa_subs_s_df(DF_WORD, pws->w[0],  pwt->w[0]);
+    pwd->w[1]  = msa_subs_s_df(DF_WORD, pws->w[1],  pwt->w[1]);
+    pwd->w[2]  = msa_subs_s_df(DF_WORD, pws->w[2],  pwt->w[2]);
+    pwd->w[3]  = msa_subs_s_df(DF_WORD, pws->w[3],  pwt->w[3]);
+}
+
+void helper_msa_subs_s_d(CPUMIPSState *env,
+                         uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->d[0]  = msa_subs_s_df(DF_DOUBLE, pws->d[0],  pwt->d[0]);
+    pwd->d[1]  = msa_subs_s_df(DF_DOUBLE, pws->d[1],  pwt->d[1]);
+}
+
+
 /*
  * Interleave
  * ----------
@@ -5060,17 +5138,6 @@ MSA_TEROP_IMMU_DF(binsli, binsl)
 MSA_TEROP_IMMU_DF(binsri, binsr)
 #undef MSA_TEROP_IMMU_DF
 
-static inline int64_t msa_subs_s_df(uint32_t df, int64_t arg1, int64_t arg2)
-{
-    int64_t max_int = DF_MAX_INT(df);
-    int64_t min_int = DF_MIN_INT(df);
-    if (arg2 > 0) {
-        return (min_int + arg2 < arg1) ? arg1 - arg2 : min_int;
-    } else {
-        return (arg1 < max_int + arg2) ? arg1 - arg2 : max_int;
-    }
-}
-
 static inline int64_t msa_subs_u_df(uint32_t df, int64_t arg1, int64_t arg2)
 {
     uint64_t u_arg1 = UNSIGNED(arg1, df);
@@ -5235,7 +5302,6 @@ void helper_msa_ ## func ## _df(CPUMIPSState *env, uint32_t df,         \
 }
 
 MSA_BINOP_DF(subv)
-MSA_BINOP_DF(subs_s)
 MSA_BINOP_DF(subs_u)
 MSA_BINOP_DF(subsus_u)
 MSA_BINOP_DF(subsuu_s)
diff --git a/target/mips/translate.c b/target/mips/translate.c
index 77e2d95..f33121a 100644
--- a/target/mips/translate.c
+++ b/target/mips/translate.c
@@ -29299,7 +29299,20 @@ static void gen_msa_3r(CPUMIPSState *env, DisasContext *ctx)
         }
         break;
     case OPC_SUBS_S_df:
-        gen_helper_msa_subs_s_df(cpu_env, tdf, twd, tws, twt);
+        switch (df) {
+        case DF_BYTE:
+            gen_helper_msa_subs_s_b(cpu_env, twd, tws, twt);
+            break;
+        case DF_HALF:
+            gen_helper_msa_subs_s_h(cpu_env, twd, tws, twt);
+            break;
+        case DF_WORD:
+            gen_helper_msa_subs_s_w(cpu_env, twd, tws, twt);
+            break;
+        case DF_DOUBLE:
+            gen_helper_msa_subs_s_d(cpu_env, twd, tws, twt);
+            break;
+        }
         break;
     case OPC_MULV_df:
         gen_helper_msa_mulv_df(cpu_env, tdf, twd, tws, twt);
-- 
2.7.4

[PULL 12/18] target/mips: msa: Split helpers for SUBS_U.<B|H|W|D>

[Aleksandar Markovic]

Achieves clearer code and slightly better performance.

Reviewed-by: Aleksandar Rikalo <aleksandar.rikalo@syrmia.com>
Signed-off-by: Aleksandar Markovic <aleksandar.qemu.devel@gmail.com>
Message-Id: <20200613152133.8964-11-aleksandar.qemu.devel@gmail.com>
---
 target/mips/helper.h     |  6 +++-
 target/mips/msa_helper.c | 82 +++++++++++++++++++++++++++++++++++++++++++-----
 target/mips/translate.c  | 15 ++++++++-
 3 files changed, 93 insertions(+), 10 deletions(-)

diff --git a/target/mips/helper.h b/target/mips/helper.h
index a93402a..61dc1ed 100644
--- a/target/mips/helper.h
+++ b/target/mips/helper.h
@@ -983,6 +983,11 @@ DEF_HELPER_4(msa_subs_s_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_subs_s_w, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_subs_s_d, void, env, i32, i32, i32)
 
+DEF_HELPER_4(msa_subs_u_b, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_subs_u_h, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_subs_u_w, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_subs_u_d, void, env, i32, i32, i32)
+
 DEF_HELPER_4(msa_ilvev_b, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_ilvev_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_ilvev_w, void, env, i32, i32, i32)
@@ -1079,7 +1084,6 @@ DEF_HELPER_5(msa_srlri_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_binsl_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_binsr_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_subv_df, void, env, i32, i32, i32, i32)
-DEF_HELPER_5(msa_subs_u_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_subsus_u_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_subsuu_s_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_mulv_df, void, env, i32, i32, i32, i32)
diff --git a/target/mips/msa_helper.c b/target/mips/msa_helper.c
index f08beba..bce32ab 100644
--- a/target/mips/msa_helper.c
+++ b/target/mips/msa_helper.c
@@ -3728,6 +3728,80 @@ void helper_msa_subs_s_d(CPUMIPSState *env,
 }
 
 
+static inline int64_t msa_subs_u_df(uint32_t df, int64_t arg1, int64_t arg2)
+{
+    uint64_t u_arg1 = UNSIGNED(arg1, df);
+    uint64_t u_arg2 = UNSIGNED(arg2, df);
+    return (u_arg1 > u_arg2) ? u_arg1 - u_arg2 : 0;
+}
+
+void helper_msa_subs_u_b(CPUMIPSState *env,
+                         uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->b[0]  = msa_subs_u_df(DF_BYTE, pws->b[0],  pwt->b[0]);
+    pwd->b[1]  = msa_subs_u_df(DF_BYTE, pws->b[1],  pwt->b[1]);
+    pwd->b[2]  = msa_subs_u_df(DF_BYTE, pws->b[2],  pwt->b[2]);
+    pwd->b[3]  = msa_subs_u_df(DF_BYTE, pws->b[3],  pwt->b[3]);
+    pwd->b[4]  = msa_subs_u_df(DF_BYTE, pws->b[4],  pwt->b[4]);
+    pwd->b[5]  = msa_subs_u_df(DF_BYTE, pws->b[5],  pwt->b[5]);
+    pwd->b[6]  = msa_subs_u_df(DF_BYTE, pws->b[6],  pwt->b[6]);
+    pwd->b[7]  = msa_subs_u_df(DF_BYTE, pws->b[7],  pwt->b[7]);
+    pwd->b[8]  = msa_subs_u_df(DF_BYTE, pws->b[8],  pwt->b[8]);
+    pwd->b[9]  = msa_subs_u_df(DF_BYTE, pws->b[9],  pwt->b[9]);
+    pwd->b[10] = msa_subs_u_df(DF_BYTE, pws->b[10], pwt->b[10]);
+    pwd->b[11] = msa_subs_u_df(DF_BYTE, pws->b[11], pwt->b[11]);
+    pwd->b[12] = msa_subs_u_df(DF_BYTE, pws->b[12], pwt->b[12]);
+    pwd->b[13] = msa_subs_u_df(DF_BYTE, pws->b[13], pwt->b[13]);
+    pwd->b[14] = msa_subs_u_df(DF_BYTE, pws->b[14], pwt->b[14]);
+    pwd->b[15] = msa_subs_u_df(DF_BYTE, pws->b[15], pwt->b[15]);
+}
+
+void helper_msa_subs_u_h(CPUMIPSState *env,
+                         uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->h[0]  = msa_subs_u_df(DF_HALF, pws->h[0],  pwt->h[0]);
+    pwd->h[1]  = msa_subs_u_df(DF_HALF, pws->h[1],  pwt->h[1]);
+    pwd->h[2]  = msa_subs_u_df(DF_HALF, pws->h[2],  pwt->h[2]);
+    pwd->h[3]  = msa_subs_u_df(DF_HALF, pws->h[3],  pwt->h[3]);
+    pwd->h[4]  = msa_subs_u_df(DF_HALF, pws->h[4],  pwt->h[4]);
+    pwd->h[5]  = msa_subs_u_df(DF_HALF, pws->h[5],  pwt->h[5]);
+    pwd->h[6]  = msa_subs_u_df(DF_HALF, pws->h[6],  pwt->h[6]);
+    pwd->h[7]  = msa_subs_u_df(DF_HALF, pws->h[7],  pwt->h[7]);
+}
+
+void helper_msa_subs_u_w(CPUMIPSState *env,
+                         uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->w[0]  = msa_subs_u_df(DF_WORD, pws->w[0],  pwt->w[0]);
+    pwd->w[1]  = msa_subs_u_df(DF_WORD, pws->w[1],  pwt->w[1]);
+    pwd->w[2]  = msa_subs_u_df(DF_WORD, pws->w[2],  pwt->w[2]);
+    pwd->w[3]  = msa_subs_u_df(DF_WORD, pws->w[3],  pwt->w[3]);
+}
+
+void helper_msa_subs_u_d(CPUMIPSState *env,
+                         uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->d[0]  = msa_subs_u_df(DF_DOUBLE, pws->d[0],  pwt->d[0]);
+    pwd->d[1]  = msa_subs_u_df(DF_DOUBLE, pws->d[1],  pwt->d[1]);
+}
+
+
 /*
  * Interleave
  * ----------
@@ -5138,13 +5212,6 @@ MSA_TEROP_IMMU_DF(binsli, binsl)
 MSA_TEROP_IMMU_DF(binsri, binsr)
 #undef MSA_TEROP_IMMU_DF
 
-static inline int64_t msa_subs_u_df(uint32_t df, int64_t arg1, int64_t arg2)
-{
-    uint64_t u_arg1 = UNSIGNED(arg1, df);
-    uint64_t u_arg2 = UNSIGNED(arg2, df);
-    return (u_arg1 > u_arg2) ? u_arg1 - u_arg2 : 0;
-}
-
 static inline int64_t msa_subsus_u_df(uint32_t df, int64_t arg1, int64_t arg2)
 {
     uint64_t u_arg1 = UNSIGNED(arg1, df);
@@ -5302,7 +5369,6 @@ void helper_msa_ ## func ## _df(CPUMIPSState *env, uint32_t df,         \
 }
 
 MSA_BINOP_DF(subv)
-MSA_BINOP_DF(subs_u)
 MSA_BINOP_DF(subsus_u)
 MSA_BINOP_DF(subsuu_s)
 MSA_BINOP_DF(mulv)
diff --git a/target/mips/translate.c b/target/mips/translate.c
index f33121a..7671a49 100644
--- a/target/mips/translate.c
+++ b/target/mips/translate.c
@@ -29327,7 +29327,20 @@ static void gen_msa_3r(CPUMIPSState *env, DisasContext *ctx)
         gen_helper_msa_subv_df(cpu_env, tdf, twd, tws, twt);
         break;
     case OPC_SUBS_U_df:
-        gen_helper_msa_subs_u_df(cpu_env, tdf, twd, tws, twt);
+        switch (df) {
+        case DF_BYTE:
+            gen_helper_msa_subs_u_b(cpu_env, twd, tws, twt);
+            break;
+        case DF_HALF:
+            gen_helper_msa_subs_u_h(cpu_env, twd, tws, twt);
+            break;
+        case DF_WORD:
+            gen_helper_msa_subs_u_w(cpu_env, twd, tws, twt);
+            break;
+        case DF_DOUBLE:
+            gen_helper_msa_subs_u_d(cpu_env, twd, tws, twt);
+            break;
+        }
         break;
     case OPC_SPLAT_df:
         gen_helper_msa_splat_df(cpu_env, tdf, twd, tws, twt);
-- 
2.7.4

[PULL 13/18] target/mips: msa: Split helpers for SUBSUS_U.<B|H|W|D>

[Aleksandar Markovic]

Achieves clearer code and slightly better performance.

Reviewed-by: Aleksandar Rikalo <aleksandar.rikalo@syrmia.com>
Signed-off-by: Aleksandar Markovic <aleksandar.qemu.devel@gmail.com>
Message-Id: <20200613152133.8964-12-aleksandar.qemu.devel@gmail.com>
---
 target/mips/helper.h     |   6 ++-
 target/mips/msa_helper.c | 102 ++++++++++++++++++++++++++++++++++++++---------
 target/mips/translate.c  |  15 ++++++-
 3 files changed, 103 insertions(+), 20 deletions(-)

diff --git a/target/mips/helper.h b/target/mips/helper.h
index 61dc1ed..227ff76 100644
--- a/target/mips/helper.h
+++ b/target/mips/helper.h
@@ -988,6 +988,11 @@ DEF_HELPER_4(msa_subs_u_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_subs_u_w, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_subs_u_d, void, env, i32, i32, i32)
 
+DEF_HELPER_4(msa_subsus_u_b, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_subsus_u_h, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_subsus_u_w, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_subsus_u_d, void, env, i32, i32, i32)
+
 DEF_HELPER_4(msa_ilvev_b, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_ilvev_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_ilvev_w, void, env, i32, i32, i32)
@@ -1084,7 +1089,6 @@ DEF_HELPER_5(msa_srlri_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_binsl_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_binsr_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_subv_df, void, env, i32, i32, i32, i32)
-DEF_HELPER_5(msa_subsus_u_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_subsuu_s_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_mulv_df, void, env, i32, i32, i32, i32)
 
diff --git a/target/mips/msa_helper.c b/target/mips/msa_helper.c
index bce32ab..f7e5c01 100644
--- a/target/mips/msa_helper.c
+++ b/target/mips/msa_helper.c
@@ -3802,6 +3802,90 @@ void helper_msa_subs_u_d(CPUMIPSState *env,
 }
 
 
+static inline int64_t msa_subsus_u_df(uint32_t df, int64_t arg1, int64_t arg2)
+{
+    uint64_t u_arg1 = UNSIGNED(arg1, df);
+    uint64_t max_uint = DF_MAX_UINT(df);
+    if (arg2 >= 0) {
+        uint64_t u_arg2 = (uint64_t)arg2;
+        return (u_arg1 > u_arg2) ?
+            (int64_t)(u_arg1 - u_arg2) :
+            0;
+    } else {
+        uint64_t u_arg2 = (uint64_t)(-arg2);
+        return (u_arg1 < max_uint - u_arg2) ?
+            (int64_t)(u_arg1 + u_arg2) :
+            (int64_t)max_uint;
+    }
+}
+
+void helper_msa_subsus_u_b(CPUMIPSState *env,
+                           uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->b[0]  = msa_subsus_u_df(DF_BYTE, pws->b[0],  pwt->b[0]);
+    pwd->b[1]  = msa_subsus_u_df(DF_BYTE, pws->b[1],  pwt->b[1]);
+    pwd->b[2]  = msa_subsus_u_df(DF_BYTE, pws->b[2],  pwt->b[2]);
+    pwd->b[3]  = msa_subsus_u_df(DF_BYTE, pws->b[3],  pwt->b[3]);
+    pwd->b[4]  = msa_subsus_u_df(DF_BYTE, pws->b[4],  pwt->b[4]);
+    pwd->b[5]  = msa_subsus_u_df(DF_BYTE, pws->b[5],  pwt->b[5]);
+    pwd->b[6]  = msa_subsus_u_df(DF_BYTE, pws->b[6],  pwt->b[6]);
+    pwd->b[7]  = msa_subsus_u_df(DF_BYTE, pws->b[7],  pwt->b[7]);
+    pwd->b[8]  = msa_subsus_u_df(DF_BYTE, pws->b[8],  pwt->b[8]);
+    pwd->b[9]  = msa_subsus_u_df(DF_BYTE, pws->b[9],  pwt->b[9]);
+    pwd->b[10] = msa_subsus_u_df(DF_BYTE, pws->b[10], pwt->b[10]);
+    pwd->b[11] = msa_subsus_u_df(DF_BYTE, pws->b[11], pwt->b[11]);
+    pwd->b[12] = msa_subsus_u_df(DF_BYTE, pws->b[12], pwt->b[12]);
+    pwd->b[13] = msa_subsus_u_df(DF_BYTE, pws->b[13], pwt->b[13]);
+    pwd->b[14] = msa_subsus_u_df(DF_BYTE, pws->b[14], pwt->b[14]);
+    pwd->b[15] = msa_subsus_u_df(DF_BYTE, pws->b[15], pwt->b[15]);
+}
+
+void helper_msa_subsus_u_h(CPUMIPSState *env,
+                           uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->h[0]  = msa_subsus_u_df(DF_HALF, pws->h[0],  pwt->h[0]);
+    pwd->h[1]  = msa_subsus_u_df(DF_HALF, pws->h[1],  pwt->h[1]);
+    pwd->h[2]  = msa_subsus_u_df(DF_HALF, pws->h[2],  pwt->h[2]);
+    pwd->h[3]  = msa_subsus_u_df(DF_HALF, pws->h[3],  pwt->h[3]);
+    pwd->h[4]  = msa_subsus_u_df(DF_HALF, pws->h[4],  pwt->h[4]);
+    pwd->h[5]  = msa_subsus_u_df(DF_HALF, pws->h[5],  pwt->h[5]);
+    pwd->h[6]  = msa_subsus_u_df(DF_HALF, pws->h[6],  pwt->h[6]);
+    pwd->h[7]  = msa_subsus_u_df(DF_HALF, pws->h[7],  pwt->h[7]);
+}
+
+void helper_msa_subsus_u_w(CPUMIPSState *env,
+                           uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->w[0]  = msa_subsus_u_df(DF_WORD, pws->w[0],  pwt->w[0]);
+    pwd->w[1]  = msa_subsus_u_df(DF_WORD, pws->w[1],  pwt->w[1]);
+    pwd->w[2]  = msa_subsus_u_df(DF_WORD, pws->w[2],  pwt->w[2]);
+    pwd->w[3]  = msa_subsus_u_df(DF_WORD, pws->w[3],  pwt->w[3]);
+}
+
+void helper_msa_subsus_u_d(CPUMIPSState *env,
+                           uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->d[0]  = msa_subsus_u_df(DF_DOUBLE, pws->d[0],  pwt->d[0]);
+    pwd->d[1]  = msa_subsus_u_df(DF_DOUBLE, pws->d[1],  pwt->d[1]);
+}
+
+
 /*
  * Interleave
  * ----------
@@ -5212,23 +5296,6 @@ MSA_TEROP_IMMU_DF(binsli, binsl)
 MSA_TEROP_IMMU_DF(binsri, binsr)
 #undef MSA_TEROP_IMMU_DF
 
-static inline int64_t msa_subsus_u_df(uint32_t df, int64_t arg1, int64_t arg2)
-{
-    uint64_t u_arg1 = UNSIGNED(arg1, df);
-    uint64_t max_uint = DF_MAX_UINT(df);
-    if (arg2 >= 0) {
-        uint64_t u_arg2 = (uint64_t)arg2;
-        return (u_arg1 > u_arg2) ?
-            (int64_t)(u_arg1 - u_arg2) :
-            0;
-    } else {
-        uint64_t u_arg2 = (uint64_t)(-arg2);
-        return (u_arg1 < max_uint - u_arg2) ?
-            (int64_t)(u_arg1 + u_arg2) :
-            (int64_t)max_uint;
-    }
-}
-
 static inline int64_t msa_subsuu_s_df(uint32_t df, int64_t arg1, int64_t arg2)
 {
     uint64_t u_arg1 = UNSIGNED(arg1, df);
@@ -5369,7 +5436,6 @@ void helper_msa_ ## func ## _df(CPUMIPSState *env, uint32_t df,         \
 }
 
 MSA_BINOP_DF(subv)
-MSA_BINOP_DF(subsus_u)
 MSA_BINOP_DF(subsuu_s)
 MSA_BINOP_DF(mulv)
 
diff --git a/target/mips/translate.c b/target/mips/translate.c
index 7671a49..b273f4d 100644
--- a/target/mips/translate.c
+++ b/target/mips/translate.c
@@ -29346,7 +29346,20 @@ static void gen_msa_3r(CPUMIPSState *env, DisasContext *ctx)
         gen_helper_msa_splat_df(cpu_env, tdf, twd, tws, twt);
         break;
     case OPC_SUBSUS_U_df:
-        gen_helper_msa_subsus_u_df(cpu_env, tdf, twd, tws, twt);
+        switch (df) {
+        case DF_BYTE:
+            gen_helper_msa_subsus_u_b(cpu_env, twd, tws, twt);
+            break;
+        case DF_HALF:
+            gen_helper_msa_subsus_u_h(cpu_env, twd, tws, twt);
+            break;
+        case DF_WORD:
+            gen_helper_msa_subsus_u_w(cpu_env, twd, tws, twt);
+            break;
+        case DF_DOUBLE:
+            gen_helper_msa_subsus_u_d(cpu_env, twd, tws, twt);
+            break;
+        }
         break;
     case OPC_SUBSUU_S_df:
         gen_helper_msa_subsuu_s_df(cpu_env, tdf, twd, tws, twt);
-- 
2.7.4

[PULL 14/18] target/mips: msa: Split helpers for SUBSUU_S.<B|H|W|D>

[Aleksandar Markovic]

Achieves clearer code and slightly better performance.

Reviewed-by: Aleksandar Rikalo <aleksandar.rikalo@syrmia.com>
Signed-off-by: Aleksandar Markovic <aleksandar.qemu.devel@gmail.com>
Message-Id: <20200613152133.8964-13-aleksandar.qemu.devel@gmail.com>
---
 target/mips/helper.h     |   6 ++-
 target/mips/msa_helper.c | 102 ++++++++++++++++++++++++++++++++++++++---------
 target/mips/translate.c  |  15 ++++++-
 3 files changed, 103 insertions(+), 20 deletions(-)

diff --git a/target/mips/helper.h b/target/mips/helper.h
index 227ff76..4795c97 100644
--- a/target/mips/helper.h
+++ b/target/mips/helper.h
@@ -993,6 +993,11 @@ DEF_HELPER_4(msa_subsus_u_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_subsus_u_w, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_subsus_u_d, void, env, i32, i32, i32)
 
+DEF_HELPER_4(msa_subsuu_s_b, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_subsuu_s_h, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_subsuu_s_w, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_subsuu_s_d, void, env, i32, i32, i32)
+
 DEF_HELPER_4(msa_ilvev_b, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_ilvev_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_ilvev_w, void, env, i32, i32, i32)
@@ -1089,7 +1094,6 @@ DEF_HELPER_5(msa_srlri_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_binsl_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_binsr_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_subv_df, void, env, i32, i32, i32, i32)
-DEF_HELPER_5(msa_subsuu_s_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_mulv_df, void, env, i32, i32, i32, i32)
 
 DEF_HELPER_4(msa_dotp_s_h, void, env, i32, i32, i32)
diff --git a/target/mips/msa_helper.c b/target/mips/msa_helper.c
index f7e5c01..27a9c36 100644
--- a/target/mips/msa_helper.c
+++ b/target/mips/msa_helper.c
@@ -3886,6 +3886,90 @@ void helper_msa_subsus_u_d(CPUMIPSState *env,
 }
 
 
+static inline int64_t msa_subsuu_s_df(uint32_t df, int64_t arg1, int64_t arg2)
+{
+    uint64_t u_arg1 = UNSIGNED(arg1, df);
+    uint64_t u_arg2 = UNSIGNED(arg2, df);
+    int64_t max_int = DF_MAX_INT(df);
+    int64_t min_int = DF_MIN_INT(df);
+    if (u_arg1 > u_arg2) {
+        return u_arg1 - u_arg2 < (uint64_t)max_int ?
+            (int64_t)(u_arg1 - u_arg2) :
+            max_int;
+    } else {
+        return u_arg2 - u_arg1 < (uint64_t)(-min_int) ?
+            (int64_t)(u_arg1 - u_arg2) :
+            min_int;
+    }
+}
+
+void helper_msa_subsuu_s_b(CPUMIPSState *env,
+                           uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->b[0]  = msa_subsuu_s_df(DF_BYTE, pws->b[0],  pwt->b[0]);
+    pwd->b[1]  = msa_subsuu_s_df(DF_BYTE, pws->b[1],  pwt->b[1]);
+    pwd->b[2]  = msa_subsuu_s_df(DF_BYTE, pws->b[2],  pwt->b[2]);
+    pwd->b[3]  = msa_subsuu_s_df(DF_BYTE, pws->b[3],  pwt->b[3]);
+    pwd->b[4]  = msa_subsuu_s_df(DF_BYTE, pws->b[4],  pwt->b[4]);
+    pwd->b[5]  = msa_subsuu_s_df(DF_BYTE, pws->b[5],  pwt->b[5]);
+    pwd->b[6]  = msa_subsuu_s_df(DF_BYTE, pws->b[6],  pwt->b[6]);
+    pwd->b[7]  = msa_subsuu_s_df(DF_BYTE, pws->b[7],  pwt->b[7]);
+    pwd->b[8]  = msa_subsuu_s_df(DF_BYTE, pws->b[8],  pwt->b[8]);
+    pwd->b[9]  = msa_subsuu_s_df(DF_BYTE, pws->b[9],  pwt->b[9]);
+    pwd->b[10] = msa_subsuu_s_df(DF_BYTE, pws->b[10], pwt->b[10]);
+    pwd->b[11] = msa_subsuu_s_df(DF_BYTE, pws->b[11], pwt->b[11]);
+    pwd->b[12] = msa_subsuu_s_df(DF_BYTE, pws->b[12], pwt->b[12]);
+    pwd->b[13] = msa_subsuu_s_df(DF_BYTE, pws->b[13], pwt->b[13]);
+    pwd->b[14] = msa_subsuu_s_df(DF_BYTE, pws->b[14], pwt->b[14]);
+    pwd->b[15] = msa_subsuu_s_df(DF_BYTE, pws->b[15], pwt->b[15]);
+}
+
+void helper_msa_subsuu_s_h(CPUMIPSState *env,
+                           uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->h[0]  = msa_subsuu_s_df(DF_HALF, pws->h[0],  pwt->h[0]);
+    pwd->h[1]  = msa_subsuu_s_df(DF_HALF, pws->h[1],  pwt->h[1]);
+    pwd->h[2]  = msa_subsuu_s_df(DF_HALF, pws->h[2],  pwt->h[2]);
+    pwd->h[3]  = msa_subsuu_s_df(DF_HALF, pws->h[3],  pwt->h[3]);
+    pwd->h[4]  = msa_subsuu_s_df(DF_HALF, pws->h[4],  pwt->h[4]);
+    pwd->h[5]  = msa_subsuu_s_df(DF_HALF, pws->h[5],  pwt->h[5]);
+    pwd->h[6]  = msa_subsuu_s_df(DF_HALF, pws->h[6],  pwt->h[6]);
+    pwd->h[7]  = msa_subsuu_s_df(DF_HALF, pws->h[7],  pwt->h[7]);
+}
+
+void helper_msa_subsuu_s_w(CPUMIPSState *env,
+                           uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->w[0]  = msa_subsuu_s_df(DF_WORD, pws->w[0],  pwt->w[0]);
+    pwd->w[1]  = msa_subsuu_s_df(DF_WORD, pws->w[1],  pwt->w[1]);
+    pwd->w[2]  = msa_subsuu_s_df(DF_WORD, pws->w[2],  pwt->w[2]);
+    pwd->w[3]  = msa_subsuu_s_df(DF_WORD, pws->w[3],  pwt->w[3]);
+}
+
+void helper_msa_subsuu_s_d(CPUMIPSState *env,
+                           uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->d[0]  = msa_subsuu_s_df(DF_DOUBLE, pws->d[0],  pwt->d[0]);
+    pwd->d[1]  = msa_subsuu_s_df(DF_DOUBLE, pws->d[1],  pwt->d[1]);
+}
+
+
 /*
  * Interleave
  * ----------
@@ -5296,23 +5380,6 @@ MSA_TEROP_IMMU_DF(binsli, binsl)
 MSA_TEROP_IMMU_DF(binsri, binsr)
 #undef MSA_TEROP_IMMU_DF
 
-static inline int64_t msa_subsuu_s_df(uint32_t df, int64_t arg1, int64_t arg2)
-{
-    uint64_t u_arg1 = UNSIGNED(arg1, df);
-    uint64_t u_arg2 = UNSIGNED(arg2, df);
-    int64_t max_int = DF_MAX_INT(df);
-    int64_t min_int = DF_MIN_INT(df);
-    if (u_arg1 > u_arg2) {
-        return u_arg1 - u_arg2 < (uint64_t)max_int ?
-            (int64_t)(u_arg1 - u_arg2) :
-            max_int;
-    } else {
-        return u_arg2 - u_arg1 < (uint64_t)(-min_int) ?
-            (int64_t)(u_arg1 - u_arg2) :
-            min_int;
-    }
-}
-
 static inline int64_t msa_mulv_df(uint32_t df, int64_t arg1, int64_t arg2)
 {
     return arg1 * arg2;
@@ -5436,7 +5503,6 @@ void helper_msa_ ## func ## _df(CPUMIPSState *env, uint32_t df,         \
 }
 
 MSA_BINOP_DF(subv)
-MSA_BINOP_DF(subsuu_s)
 MSA_BINOP_DF(mulv)
 
 MSA_BINOP_DF(mul_q)
diff --git a/target/mips/translate.c b/target/mips/translate.c
index b273f4d..6d9ae1d 100644
--- a/target/mips/translate.c
+++ b/target/mips/translate.c
@@ -29362,7 +29362,20 @@ static void gen_msa_3r(CPUMIPSState *env, DisasContext *ctx)
         }
         break;
     case OPC_SUBSUU_S_df:
-        gen_helper_msa_subsuu_s_df(cpu_env, tdf, twd, tws, twt);
+        switch (df) {
+        case DF_BYTE:
+            gen_helper_msa_subsuu_s_b(cpu_env, twd, tws, twt);
+            break;
+        case DF_HALF:
+            gen_helper_msa_subsuu_s_h(cpu_env, twd, tws, twt);
+            break;
+        case DF_WORD:
+            gen_helper_msa_subsuu_s_w(cpu_env, twd, tws, twt);
+            break;
+        case DF_DOUBLE:
+            gen_helper_msa_subsuu_s_d(cpu_env, twd, tws, twt);
+            break;
+        }
         break;
 
     case OPC_DOTP_S_df:
-- 
2.7.4

[PULL 15/18] target/mips: msa: Split helpers for SUBV.<B|H|W|D>

[Aleksandar Markovic]

Achieves clearer code and slightly better performance.

Reviewed-by: Aleksandar Rikalo <aleksandar.rikalo@syrmia.com>
Signed-off-by: Aleksandar Markovic <aleksandar.qemu.devel@gmail.com>
Message-Id: <20200613152133.8964-14-aleksandar.qemu.devel@gmail.com>
---
 target/mips/helper.h     |  6 +++-
 target/mips/msa_helper.c | 81 ++++++++++++++++++++++++++++++++++++++++++------
 target/mips/translate.c  | 15 ++++++++-
 3 files changed, 91 insertions(+), 11 deletions(-)

diff --git a/target/mips/helper.h b/target/mips/helper.h
index 4795c97..5d7ba6a 100644
--- a/target/mips/helper.h
+++ b/target/mips/helper.h
@@ -998,6 +998,11 @@ DEF_HELPER_4(msa_subsuu_s_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_subsuu_s_w, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_subsuu_s_d, void, env, i32, i32, i32)
 
+DEF_HELPER_4(msa_subv_b, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_subv_h, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_subv_w, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_subv_d, void, env, i32, i32, i32)
+
 DEF_HELPER_4(msa_ilvev_b, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_ilvev_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_ilvev_w, void, env, i32, i32, i32)
@@ -1093,7 +1098,6 @@ DEF_HELPER_5(msa_srlri_df, void, env, i32, i32, i32, i32)
 
 DEF_HELPER_5(msa_binsl_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_binsr_df, void, env, i32, i32, i32, i32)
-DEF_HELPER_5(msa_subv_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_mulv_df, void, env, i32, i32, i32, i32)
 
 DEF_HELPER_4(msa_dotp_s_h, void, env, i32, i32, i32)
diff --git a/target/mips/msa_helper.c b/target/mips/msa_helper.c
index 27a9c36..d099e00 100644
--- a/target/mips/msa_helper.c
+++ b/target/mips/msa_helper.c
@@ -3553,9 +3553,6 @@ void helper_msa_asub_u_d(CPUMIPSState *env,
 }
 
 
-/* TODO: insert the rest of Int Subtract group helpers here */
-
-
 static inline int64_t msa_hsub_s_df(uint32_t df, int64_t arg1, int64_t arg2)
 {
     return SIGNED_ODD(arg1, df) - SIGNED_EVEN(arg2, df);
@@ -3970,6 +3967,78 @@ void helper_msa_subsuu_s_d(CPUMIPSState *env,
 }
 
 
+static inline int64_t msa_subv_df(uint32_t df, int64_t arg1, int64_t arg2)
+{
+    return arg1 - arg2;
+}
+
+void helper_msa_subv_b(CPUMIPSState *env,
+                       uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->b[0]  = msa_subv_df(DF_BYTE, pws->b[0],  pwt->b[0]);
+    pwd->b[1]  = msa_subv_df(DF_BYTE, pws->b[1],  pwt->b[1]);
+    pwd->b[2]  = msa_subv_df(DF_BYTE, pws->b[2],  pwt->b[2]);
+    pwd->b[3]  = msa_subv_df(DF_BYTE, pws->b[3],  pwt->b[3]);
+    pwd->b[4]  = msa_subv_df(DF_BYTE, pws->b[4],  pwt->b[4]);
+    pwd->b[5]  = msa_subv_df(DF_BYTE, pws->b[5],  pwt->b[5]);
+    pwd->b[6]  = msa_subv_df(DF_BYTE, pws->b[6],  pwt->b[6]);
+    pwd->b[7]  = msa_subv_df(DF_BYTE, pws->b[7],  pwt->b[7]);
+    pwd->b[8]  = msa_subv_df(DF_BYTE, pws->b[8],  pwt->b[8]);
+    pwd->b[9]  = msa_subv_df(DF_BYTE, pws->b[9],  pwt->b[9]);
+    pwd->b[10] = msa_subv_df(DF_BYTE, pws->b[10], pwt->b[10]);
+    pwd->b[11] = msa_subv_df(DF_BYTE, pws->b[11], pwt->b[11]);
+    pwd->b[12] = msa_subv_df(DF_BYTE, pws->b[12], pwt->b[12]);
+    pwd->b[13] = msa_subv_df(DF_BYTE, pws->b[13], pwt->b[13]);
+    pwd->b[14] = msa_subv_df(DF_BYTE, pws->b[14], pwt->b[14]);
+    pwd->b[15] = msa_subv_df(DF_BYTE, pws->b[15], pwt->b[15]);
+}
+
+void helper_msa_subv_h(CPUMIPSState *env,
+                       uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->h[0]  = msa_subv_df(DF_HALF, pws->h[0],  pwt->h[0]);
+    pwd->h[1]  = msa_subv_df(DF_HALF, pws->h[1],  pwt->h[1]);
+    pwd->h[2]  = msa_subv_df(DF_HALF, pws->h[2],  pwt->h[2]);
+    pwd->h[3]  = msa_subv_df(DF_HALF, pws->h[3],  pwt->h[3]);
+    pwd->h[4]  = msa_subv_df(DF_HALF, pws->h[4],  pwt->h[4]);
+    pwd->h[5]  = msa_subv_df(DF_HALF, pws->h[5],  pwt->h[5]);
+    pwd->h[6]  = msa_subv_df(DF_HALF, pws->h[6],  pwt->h[6]);
+    pwd->h[7]  = msa_subv_df(DF_HALF, pws->h[7],  pwt->h[7]);
+}
+
+void helper_msa_subv_w(CPUMIPSState *env,
+                       uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->w[0]  = msa_subv_df(DF_WORD, pws->w[0],  pwt->w[0]);
+    pwd->w[1]  = msa_subv_df(DF_WORD, pws->w[1],  pwt->w[1]);
+    pwd->w[2]  = msa_subv_df(DF_WORD, pws->w[2],  pwt->w[2]);
+    pwd->w[3]  = msa_subv_df(DF_WORD, pws->w[3],  pwt->w[3]);
+}
+
+void helper_msa_subv_d(CPUMIPSState *env,
+                       uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->d[0]  = msa_subv_df(DF_DOUBLE, pws->d[0],  pwt->d[0]);
+    pwd->d[1]  = msa_subv_df(DF_DOUBLE, pws->d[1],  pwt->d[1]);
+}
+
+
 /*
  * Interleave
  * ----------
@@ -5194,11 +5263,6 @@ void helper_msa_shf_df(CPUMIPSState *env, uint32_t df, uint32_t wd,
     msa_move_v(pwd, pwx);
 }
 
-static inline int64_t msa_subv_df(uint32_t df, int64_t arg1, int64_t arg2)
-{
-    return arg1 - arg2;
-}
-
 #define MSA_BINOP_IMM_DF(helper, func)                                  \
 void helper_msa_ ## helper ## _df(CPUMIPSState *env, uint32_t df,       \
                         uint32_t wd, uint32_t ws, int32_t u5)           \
@@ -5502,7 +5566,6 @@ void helper_msa_ ## func ## _df(CPUMIPSState *env, uint32_t df,         \
     }                                                                   \
 }
 
-MSA_BINOP_DF(subv)
 MSA_BINOP_DF(mulv)
 
 MSA_BINOP_DF(mul_q)
diff --git a/target/mips/translate.c b/target/mips/translate.c
index 6d9ae1d..9ca17ed 100644
--- a/target/mips/translate.c
+++ b/target/mips/translate.c
@@ -29324,7 +29324,20 @@ static void gen_msa_3r(CPUMIPSState *env, DisasContext *ctx)
         gen_helper_msa_vshf_df(cpu_env, tdf, twd, tws, twt);
         break;
     case OPC_SUBV_df:
-        gen_helper_msa_subv_df(cpu_env, tdf, twd, tws, twt);
+        switch (df) {
+        case DF_BYTE:
+            gen_helper_msa_subv_b(cpu_env, twd, tws, twt);
+            break;
+        case DF_HALF:
+            gen_helper_msa_subv_h(cpu_env, twd, tws, twt);
+            break;
+        case DF_WORD:
+            gen_helper_msa_subv_w(cpu_env, twd, tws, twt);
+            break;
+        case DF_DOUBLE:
+            gen_helper_msa_subv_d(cpu_env, twd, tws, twt);
+            break;
+        }
         break;
     case OPC_SUBS_U_df:
         switch (df) {
-- 
2.7.4

[PULL 16/18] target/mips: msa: Split helpers for MULV.<B|H|W|D>

[Aleksandar Markovic]

Achieves clearer code and slightly better performance.

Reviewed-by: Aleksandar Rikalo <aleksandar.rikalo@syrmia.com>
Signed-off-by: Aleksandar Markovic <aleksandar.qemu.devel@gmail.com>
Message-Id: <20200613152133.8964-15-aleksandar.qemu.devel@gmail.com>
---
 target/mips/helper.h     |  6 +++-
 target/mips/msa_helper.c | 79 +++++++++++++++++++++++++++++++++++++++++++-----
 target/mips/translate.c  | 15 ++++++++-
 3 files changed, 91 insertions(+), 9 deletions(-)

diff --git a/target/mips/helper.h b/target/mips/helper.h
index 5d7ba6a..e97655d 100644
--- a/target/mips/helper.h
+++ b/target/mips/helper.h
@@ -960,6 +960,11 @@ DEF_HELPER_4(msa_msubv_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_msubv_w, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_msubv_d, void, env, i32, i32, i32)
 
+DEF_HELPER_4(msa_mulv_b, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_mulv_h, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_mulv_w, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_mulv_d, void, env, i32, i32, i32)
+
 DEF_HELPER_4(msa_asub_s_b, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_asub_s_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_asub_s_w, void, env, i32, i32, i32)
@@ -1098,7 +1103,6 @@ DEF_HELPER_5(msa_srlri_df, void, env, i32, i32, i32, i32)
 
 DEF_HELPER_5(msa_binsl_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_binsr_df, void, env, i32, i32, i32, i32)
-DEF_HELPER_5(msa_mulv_df, void, env, i32, i32, i32, i32)
 
 DEF_HELPER_4(msa_dotp_s_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_dotp_s_w, void, env, i32, i32, i32)
diff --git a/target/mips/msa_helper.c b/target/mips/msa_helper.c
index d099e00..6865add 100644
--- a/target/mips/msa_helper.c
+++ b/target/mips/msa_helper.c
@@ -3360,6 +3360,78 @@ void helper_msa_msubv_d(CPUMIPSState *env,
 }
 
 
+static inline int64_t msa_mulv_df(uint32_t df, int64_t arg1, int64_t arg2)
+{
+    return arg1 * arg2;
+}
+
+void helper_msa_mulv_b(CPUMIPSState *env,
+                       uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->b[0]  = msa_mulv_df(DF_BYTE, pws->b[0],  pwt->b[0]);
+    pwd->b[1]  = msa_mulv_df(DF_BYTE, pws->b[1],  pwt->b[1]);
+    pwd->b[2]  = msa_mulv_df(DF_BYTE, pws->b[2],  pwt->b[2]);
+    pwd->b[3]  = msa_mulv_df(DF_BYTE, pws->b[3],  pwt->b[3]);
+    pwd->b[4]  = msa_mulv_df(DF_BYTE, pws->b[4],  pwt->b[4]);
+    pwd->b[5]  = msa_mulv_df(DF_BYTE, pws->b[5],  pwt->b[5]);
+    pwd->b[6]  = msa_mulv_df(DF_BYTE, pws->b[6],  pwt->b[6]);
+    pwd->b[7]  = msa_mulv_df(DF_BYTE, pws->b[7],  pwt->b[7]);
+    pwd->b[8]  = msa_mulv_df(DF_BYTE, pws->b[8],  pwt->b[8]);
+    pwd->b[9]  = msa_mulv_df(DF_BYTE, pws->b[9],  pwt->b[9]);
+    pwd->b[10] = msa_mulv_df(DF_BYTE, pws->b[10], pwt->b[10]);
+    pwd->b[11] = msa_mulv_df(DF_BYTE, pws->b[11], pwt->b[11]);
+    pwd->b[12] = msa_mulv_df(DF_BYTE, pws->b[12], pwt->b[12]);
+    pwd->b[13] = msa_mulv_df(DF_BYTE, pws->b[13], pwt->b[13]);
+    pwd->b[14] = msa_mulv_df(DF_BYTE, pws->b[14], pwt->b[14]);
+    pwd->b[15] = msa_mulv_df(DF_BYTE, pws->b[15], pwt->b[15]);
+}
+
+void helper_msa_mulv_h(CPUMIPSState *env,
+                       uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->h[0]  = msa_mulv_df(DF_HALF, pws->h[0],  pwt->h[0]);
+    pwd->h[1]  = msa_mulv_df(DF_HALF, pws->h[1],  pwt->h[1]);
+    pwd->h[2]  = msa_mulv_df(DF_HALF, pws->h[2],  pwt->h[2]);
+    pwd->h[3]  = msa_mulv_df(DF_HALF, pws->h[3],  pwt->h[3]);
+    pwd->h[4]  = msa_mulv_df(DF_HALF, pws->h[4],  pwt->h[4]);
+    pwd->h[5]  = msa_mulv_df(DF_HALF, pws->h[5],  pwt->h[5]);
+    pwd->h[6]  = msa_mulv_df(DF_HALF, pws->h[6],  pwt->h[6]);
+    pwd->h[7]  = msa_mulv_df(DF_HALF, pws->h[7],  pwt->h[7]);
+}
+
+void helper_msa_mulv_w(CPUMIPSState *env,
+                       uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->w[0]  = msa_mulv_df(DF_WORD, pws->w[0],  pwt->w[0]);
+    pwd->w[1]  = msa_mulv_df(DF_WORD, pws->w[1],  pwt->w[1]);
+    pwd->w[2]  = msa_mulv_df(DF_WORD, pws->w[2],  pwt->w[2]);
+    pwd->w[3]  = msa_mulv_df(DF_WORD, pws->w[3],  pwt->w[3]);
+}
+
+void helper_msa_mulv_d(CPUMIPSState *env,
+                       uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->d[0]  = msa_mulv_df(DF_DOUBLE, pws->d[0],  pwt->d[0]);
+    pwd->d[1]  = msa_mulv_df(DF_DOUBLE, pws->d[1],  pwt->d[1]);
+}
+
+
 /*
  * Int Subtract
  * ------------
@@ -5444,11 +5516,6 @@ MSA_TEROP_IMMU_DF(binsli, binsl)
 MSA_TEROP_IMMU_DF(binsri, binsr)
 #undef MSA_TEROP_IMMU_DF
 
-static inline int64_t msa_mulv_df(uint32_t df, int64_t arg1, int64_t arg2)
-{
-    return arg1 * arg2;
-}
-
 #define CONCATENATE_AND_SLIDE(s, k)             \
     do {                                        \
         for (i = 0; i < s; i++) {               \
@@ -5566,8 +5633,6 @@ void helper_msa_ ## func ## _df(CPUMIPSState *env, uint32_t df,         \
     }                                                                   \
 }
 
-MSA_BINOP_DF(mulv)
-
 MSA_BINOP_DF(mul_q)
 MSA_BINOP_DF(mulr_q)
 #undef MSA_BINOP_DF
diff --git a/target/mips/translate.c b/target/mips/translate.c
index 9ca17ed..9fad58e 100644
--- a/target/mips/translate.c
+++ b/target/mips/translate.c
@@ -29315,7 +29315,20 @@ static void gen_msa_3r(CPUMIPSState *env, DisasContext *ctx)
         }
         break;
     case OPC_MULV_df:
-        gen_helper_msa_mulv_df(cpu_env, tdf, twd, tws, twt);
+        switch (df) {
+        case DF_BYTE:
+            gen_helper_msa_mulv_b(cpu_env, twd, tws, twt);
+            break;
+        case DF_HALF:
+            gen_helper_msa_mulv_h(cpu_env, twd, tws, twt);
+            break;
+        case DF_WORD:
+            gen_helper_msa_mulv_w(cpu_env, twd, tws, twt);
+            break;
+        case DF_DOUBLE:
+            gen_helper_msa_mulv_d(cpu_env, twd, tws, twt);
+            break;
+        }
         break;
     case OPC_SLD_df:
         gen_helper_msa_sld_df(cpu_env, tdf, twd, tws, twt);
-- 
2.7.4

[PULL 17/18] MAINTAINERS: Adjust sh4 maintainership

[Aleksandar Markovic]

This patch transfers sh4 sections to Yoshinori Sato, who is
best positioned in the community to assume sh4 maintainership.
He is the maintainer of the related target rx as well, which
means that some synergy between the two targets can be expected
in future.

Further adjustments, reorganizations, and improvements of sh4
sections are left to the future maintainer to be devised and
executed, as he deems suitable.

Aurelien and Magnus are deleted as maintainers in some sections
of the MAINTAINERS file with this patch. However, they will not
be deleted from QEMU Hall of Fame, where their names will always
remained carved in stone as QEMU pioneers and granddaddies.

Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Acked-by: Aurelien Jarno <aurelien@aurel32.net>
Acked-by: Magnus Damm <magnus.damm@gmail.com>
Acked-by: Yoshinori Sato <ysato@users.sourceforge.jp>
Signed-off-by: Aleksandar Markovic <aleksandar.qemu.devel@gmail.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20200611095316.10133-2-aleksandar.qemu.devel@gmail.com>
---
 MAINTAINERS | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/MAINTAINERS b/MAINTAINERS
index a922775..d59b605 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -296,7 +296,7 @@ F: tests/tcg/s390x/
 L: qemu-s390x@nongnu.org
 
 SH4 TCG CPUs
-M: Aurelien Jarno <aurelien@aurel32.net>
+M: Yoshinori Sato <ysato@users.sourceforge.jp>
 S: Odd Fixes
 F: target/sh4/
 F: hw/sh4/
@@ -1253,14 +1253,16 @@ F: include/hw/riscv/opentitan.h
 SH4 Machines
 ------------
 R2D
-M: Magnus Damm <magnus.damm@gmail.com>
+M: Yoshinori Sato <ysato@users.sourceforge.jp>
+R: Magnus Damm <magnus.damm@gmail.com>
 S: Maintained
 F: hw/sh4/r2d.c
 F: hw/intc/sh_intc.c
 F: hw/timer/sh_timer.c
 
 Shix
-M: Magnus Damm <magnus.damm@gmail.com>
+M: Yoshinori Sato <ysato@users.sourceforge.jp>
+R: Magnus Damm <magnus.damm@gmail.com>
 S: Odd Fixes
 F: hw/sh4/shix.c
 
-- 
2.7.4

[PULL 18/18] translations: Add Swedish language

[Aleksandar Markovic]

From: Sebastian Rasmussen <sebras@gmail.com>

This patch adds translation of QEMU to Swedish.

Signed-off-by: Sebastian Rasmussen <sebras@gmail.com>
Signed-off-by: Aleksandar Markovic <aleksandar.qemu.devel@gmail.com>
Reviewed-by: Aleksandar Markovic <aleksandar.qemu.devel@gmail.com>
Message-Id: <20200611114523.15584-2-aleksandar.qemu.devel@gmail.com>
---
 po/sv.po | 75 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 75 insertions(+)
 create mode 100644 po/sv.po

diff --git a/po/sv.po b/po/sv.po
new file mode 100644
index 0000000..1e430ed
--- /dev/null
+++ b/po/sv.po
@@ -0,0 +1,75 @@
+# Swedish translation of qemu po-file.
+# This file is put in the public domain.
+# Sebastian Rasmussen <sebras@gmail.com>, 2019.
+#
+msgid ""
+msgstr ""
+"Project-Id-Version: QEMU 2.12.91\n"
+"Report-Msgid-Bugs-To: qemu-devel@nongnu.org\n"
+"POT-Creation-Date: 2018-07-18 07:56+0200\n"
+"PO-Revision-Date: 2019-08-16 21:19+0200\n"
+"Last-Translator: Sebastian Rasmussen <sebras@gmail.com>\n"
+"Language-Team: Swedish <tp-sv@listor.tp-sv.se>\n"
+"Language: sv\n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=UTF-8\n"
+"Content-Transfer-Encoding: 8bit\n"
+"Plural-Forms: nplurals=2; plural=(n != 1);\n"
+"X-Generator: Poedit 2.2.3\n"
+
+msgid " - Press Ctrl+Alt+G to release grab"
+msgstr " - Tryck Ctrl+Alt+G för att sluta fånga"
+
+msgid " [Paused]"
+msgstr " [Pausad]"
+
+msgid "_Pause"
+msgstr "_Paus"
+
+msgid "_Reset"
+msgstr "_Starta om"
+
+msgid "Power _Down"
+msgstr "Stäng _ner"
+
+msgid "_Quit"
+msgstr "_Avsluta"
+
+msgid "_Fullscreen"
+msgstr "_Helskärm"
+
+msgid "_Copy"
+msgstr "_Kopiera"
+
+msgid "Zoom _In"
+msgstr "Zooma _in"
+
+msgid "Zoom _Out"
+msgstr "Zooma _ut"
+
+msgid "Best _Fit"
+msgstr "Anpassad _storlek"
+
+msgid "Zoom To _Fit"
+msgstr "Zooma ti_ll anpassad storlek"
+
+msgid "Grab On _Hover"
+msgstr "Fånga vi_d hovring"
+
+msgid "_Grab Input"
+msgstr "Fån_ga inmatning"
+
+msgid "Show _Tabs"
+msgstr "Visa _flika"
+
+msgid "Detach Tab"
+msgstr "Frigör flik"
+
+msgid "Show Menubar"
+msgstr "Visa menyrad"
+
+msgid "_Machine"
+msgstr "_Maskin"
+
+msgid "_View"
+msgstr "_Visa"
-- 
2.7.4

Re: [PULL v2 00/18] MIPS + misc queue for June 15th, 2020

[Peter Maydell]

On Mon, 15 Jun 2020 at 20:29, Aleksandar Markovic
<aleksandar.qemu.devel@gmail.com> wrote:
>
> The following changes since commit 7d3660e79830a069f1848bb4fa1cdf8f666424fb:
>
>   Merge remote-tracking branch 'remotes/bonzini/tags/for-upstream' into staging (2020-06-12 23:06:22 +0100)
>
> are available in the git repository at:
>
>   https://github.com/AMarkovic/qemu tags/mips-queue-jun-15-2020
>
> for you to fetch changes up to 250bc43a406f7d46e319abe87c19548d4f027828:
>
>   translations: Add Swedish language (2020-06-15 20:51:10 +0200)
>
> ----------------------------------------------------------------
>
> MIPS + misc queue for June 15th, 2020
>
>   Highlights:
>
>     This pull request, just exceptionally, contains two non-MIPS patches:
>
>       - adjust sh4 maintainership
>       - add Swedish translations
>
>     The rest are MIPS patches:
>
>       - refactor emulation of a number of MSA instructions
>       - activate Loongson-related insn_flags
>
>     Notes:
>
>       - one checkpatch warning is benign
>       - some of make check iotest-qcow2 tests fail on my system, both before
>         and after applying the patches from this pull request


Applied, thanks.

Please update the changelog at https://wiki.qemu.org/ChangeLog/5.1
for any user-visible changes.

-- PMM

[PULL 14/18] target/mips: msa: Split helpers for SUBSUU_S.<B|H|W|D>

[Aleksandar Markovic]

Achieves clearer code and slightly better performance.

Reviewed-by: Aleksandar Rikalo <aleksandar.rikalo@syrmia.com>
Signed-off-by: Aleksandar Markovic <aleksandar.qemu.devel@gmail.com>
Message-Id: <20200613152133.8964-13-aleksandar.qemu.devel@gmail.com>
---
 target/mips/helper.h     |   6 ++-
 target/mips/msa_helper.c | 102 ++++++++++++++++++++++++++++++++++++++---------
 target/mips/translate.c  |  15 ++++++-
 3 files changed, 103 insertions(+), 20 deletions(-)

diff --git a/target/mips/helper.h b/target/mips/helper.h
index 227ff76..4795c97 100644
--- a/target/mips/helper.h
+++ b/target/mips/helper.h
@@ -993,6 +993,11 @@ DEF_HELPER_4(msa_subsus_u_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_subsus_u_w, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_subsus_u_d, void, env, i32, i32, i32)
 
+DEF_HELPER_4(msa_subsuu_s_b, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_subsuu_s_h, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_subsuu_s_w, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_subsuu_s_d, void, env, i32, i32, i32)
+
 DEF_HELPER_4(msa_ilvev_b, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_ilvev_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_ilvev_w, void, env, i32, i32, i32)
@@ -1089,7 +1094,6 @@ DEF_HELPER_5(msa_srlri_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_binsl_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_binsr_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_subv_df, void, env, i32, i32, i32, i32)
-DEF_HELPER_5(msa_subsuu_s_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_mulv_df, void, env, i32, i32, i32, i32)
 
 DEF_HELPER_4(msa_dotp_s_h, void, env, i32, i32, i32)
diff --git a/target/mips/msa_helper.c b/target/mips/msa_helper.c
index f7e5c01..27a9c36 100644
--- a/target/mips/msa_helper.c
+++ b/target/mips/msa_helper.c
@@ -3886,6 +3886,90 @@ void helper_msa_subsus_u_d(CPUMIPSState *env,
 }
 
 
+static inline int64_t msa_subsuu_s_df(uint32_t df, int64_t arg1, int64_t arg2)
+{
+    uint64_t u_arg1 = UNSIGNED(arg1, df);
+    uint64_t u_arg2 = UNSIGNED(arg2, df);
+    int64_t max_int = DF_MAX_INT(df);
+    int64_t min_int = DF_MIN_INT(df);
+    if (u_arg1 > u_arg2) {
+        return u_arg1 - u_arg2 < (uint64_t)max_int ?
+            (int64_t)(u_arg1 - u_arg2) :
+            max_int;
+    } else {
+        return u_arg2 - u_arg1 < (uint64_t)(-min_int) ?
+            (int64_t)(u_arg1 - u_arg2) :
+            min_int;
+    }
+}
+
+void helper_msa_subsuu_s_b(CPUMIPSState *env,
+                           uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->b[0]  = msa_subsuu_s_df(DF_BYTE, pws->b[0],  pwt->b[0]);
+    pwd->b[1]  = msa_subsuu_s_df(DF_BYTE, pws->b[1],  pwt->b[1]);
+    pwd->b[2]  = msa_subsuu_s_df(DF_BYTE, pws->b[2],  pwt->b[2]);
+    pwd->b[3]  = msa_subsuu_s_df(DF_BYTE, pws->b[3],  pwt->b[3]);
+    pwd->b[4]  = msa_subsuu_s_df(DF_BYTE, pws->b[4],  pwt->b[4]);
+    pwd->b[5]  = msa_subsuu_s_df(DF_BYTE, pws->b[5],  pwt->b[5]);
+    pwd->b[6]  = msa_subsuu_s_df(DF_BYTE, pws->b[6],  pwt->b[6]);
+    pwd->b[7]  = msa_subsuu_s_df(DF_BYTE, pws->b[7],  pwt->b[7]);
+    pwd->b[8]  = msa_subsuu_s_df(DF_BYTE, pws->b[8],  pwt->b[8]);
+    pwd->b[9]  = msa_subsuu_s_df(DF_BYTE, pws->b[9],  pwt->b[9]);
+    pwd->b[10] = msa_subsuu_s_df(DF_BYTE, pws->b[10], pwt->b[10]);
+    pwd->b[11] = msa_subsuu_s_df(DF_BYTE, pws->b[11], pwt->b[11]);
+    pwd->b[12] = msa_subsuu_s_df(DF_BYTE, pws->b[12], pwt->b[12]);
+    pwd->b[13] = msa_subsuu_s_df(DF_BYTE, pws->b[13], pwt->b[13]);
+    pwd->b[14] = msa_subsuu_s_df(DF_BYTE, pws->b[14], pwt->b[14]);
+    pwd->b[15] = msa_subsuu_s_df(DF_BYTE, pws->b[15], pwt->b[15]);
+}
+
+void helper_msa_subsuu_s_h(CPUMIPSState *env,
+                           uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->h[0]  = msa_subsuu_s_df(DF_HALF, pws->h[0],  pwt->h[0]);
+    pwd->h[1]  = msa_subsuu_s_df(DF_HALF, pws->h[1],  pwt->h[1]);
+    pwd->h[2]  = msa_subsuu_s_df(DF_HALF, pws->h[2],  pwt->h[2]);
+    pwd->h[3]  = msa_subsuu_s_df(DF_HALF, pws->h[3],  pwt->h[3]);
+    pwd->h[4]  = msa_subsuu_s_df(DF_HALF, pws->h[4],  pwt->h[4]);
+    pwd->h[5]  = msa_subsuu_s_df(DF_HALF, pws->h[5],  pwt->h[5]);
+    pwd->h[6]  = msa_subsuu_s_df(DF_HALF, pws->h[6],  pwt->h[6]);
+    pwd->h[7]  = msa_subsuu_s_df(DF_HALF, pws->h[7],  pwt->h[7]);
+}
+
+void helper_msa_subsuu_s_w(CPUMIPSState *env,
+                           uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->w[0]  = msa_subsuu_s_df(DF_WORD, pws->w[0],  pwt->w[0]);
+    pwd->w[1]  = msa_subsuu_s_df(DF_WORD, pws->w[1],  pwt->w[1]);
+    pwd->w[2]  = msa_subsuu_s_df(DF_WORD, pws->w[2],  pwt->w[2]);
+    pwd->w[3]  = msa_subsuu_s_df(DF_WORD, pws->w[3],  pwt->w[3]);
+}
+
+void helper_msa_subsuu_s_d(CPUMIPSState *env,
+                           uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->d[0]  = msa_subsuu_s_df(DF_DOUBLE, pws->d[0],  pwt->d[0]);
+    pwd->d[1]  = msa_subsuu_s_df(DF_DOUBLE, pws->d[1],  pwt->d[1]);
+}
+
+
 /*
  * Interleave
  * ----------
@@ -5296,23 +5380,6 @@ MSA_TEROP_IMMU_DF(binsli, binsl)
 MSA_TEROP_IMMU_DF(binsri, binsr)
 #undef MSA_TEROP_IMMU_DF
 
-static inline int64_t msa_subsuu_s_df(uint32_t df, int64_t arg1, int64_t arg2)
-{
-    uint64_t u_arg1 = UNSIGNED(arg1, df);
-    uint64_t u_arg2 = UNSIGNED(arg2, df);
-    int64_t max_int = DF_MAX_INT(df);
-    int64_t min_int = DF_MIN_INT(df);
-    if (u_arg1 > u_arg2) {
-        return u_arg1 - u_arg2 < (uint64_t)max_int ?
-            (int64_t)(u_arg1 - u_arg2) :
-            max_int;
-    } else {
-        return u_arg2 - u_arg1 < (uint64_t)(-min_int) ?
-            (int64_t)(u_arg1 - u_arg2) :
-            min_int;
-    }
-}
-
 static inline int64_t msa_mulv_df(uint32_t df, int64_t arg1, int64_t arg2)
 {
     return arg1 * arg2;
@@ -5436,7 +5503,6 @@ void helper_msa_ ## func ## _df(CPUMIPSState *env, uint32_t df,         \
 }
 
 MSA_BINOP_DF(subv)
-MSA_BINOP_DF(subsuu_s)
 MSA_BINOP_DF(mulv)
 
 MSA_BINOP_DF(mul_q)
diff --git a/target/mips/translate.c b/target/mips/translate.c
index b273f4d..6d9ae1d 100644
--- a/target/mips/translate.c
+++ b/target/mips/translate.c
@@ -29362,7 +29362,20 @@ static void gen_msa_3r(CPUMIPSState *env, DisasContext *ctx)
         }
         break;
     case OPC_SUBSUU_S_df:
-        gen_helper_msa_subsuu_s_df(cpu_env, tdf, twd, tws, twt);
+        switch (df) {
+        case DF_BYTE:
+            gen_helper_msa_subsuu_s_b(cpu_env, twd, tws, twt);
+            break;
+        case DF_HALF:
+            gen_helper_msa_subsuu_s_h(cpu_env, twd, tws, twt);
+            break;
+        case DF_WORD:
+            gen_helper_msa_subsuu_s_w(cpu_env, twd, tws, twt);
+            break;
+        case DF_DOUBLE:
+            gen_helper_msa_subsuu_s_d(cpu_env, twd, tws, twt);
+            break;
+        }
         break;
 
     case OPC_DOTP_S_df:
-- 
2.7.4