diff mbox series

[v3,01/16] target/arm: Create gen_gvec_[us]sra

Message ID 20200508152200.6547-2-richard.henderson@linaro.org
State Superseded
Headers show
Series target/arm: partial vector cleanup | expand

Commit Message

Richard Henderson May 8, 2020, 3:21 p.m. UTC
The functions eliminate duplication of the special cases for
this operation.  They match up with the GVecGen2iFn typedef.

Add out-of-line helpers.  We got away with only having inline
expanders because the neon vector size is only 16 bytes, and
we know that the inline expansion will always succeed.
When we reuse this for SVE, tcg-gvec-op may decide to use an
out-of-line helper due to longer vector lengths.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

---
 target/arm/helper.h        |  10 +++
 target/arm/translate.h     |   7 +-
 target/arm/translate-a64.c |  15 +---
 target/arm/translate.c     | 161 ++++++++++++++++++++++---------------
 target/arm/vec_helper.c    |  25 ++++++
 5 files changed, 139 insertions(+), 79 deletions(-)

-- 
2.20.1

Comments

Peter Maydell May 12, 2020, 1:20 p.m. UTC | #1
On Fri, 8 May 2020 at 16:22, Richard Henderson
<richard.henderson@linaro.org> wrote:
>

> The functions eliminate duplication of the special cases for

> this operation.  They match up with the GVecGen2iFn typedef.

>

> Add out-of-line helpers.  We got away with only having inline

> expanders because the neon vector size is only 16 bytes, and

> we know that the inline expansion will always succeed.

> When we reuse this for SVE, tcg-gvec-op may decide to use an

> out-of-line helper due to longer vector lengths.

>

> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>


Reviewed-by: Peter Maydell <peter.maydell@linaro.org>


thanks
-- PMM
diff mbox series

Patch

diff --git a/target/arm/helper.h b/target/arm/helper.h
index 5817626b20..9bc162345c 100644
--- a/target/arm/helper.h
+++ b/target/arm/helper.h
@@ -691,6 +691,16 @@  DEF_HELPER_FLAGS_4(gvec_pmull_q, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32)
 
 DEF_HELPER_FLAGS_4(neon_pmull_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32)
 
+DEF_HELPER_FLAGS_3(gvec_ssra_b, TCG_CALL_NO_RWG, void, ptr, ptr, i32)
+DEF_HELPER_FLAGS_3(gvec_ssra_h, TCG_CALL_NO_RWG, void, ptr, ptr, i32)
+DEF_HELPER_FLAGS_3(gvec_ssra_s, TCG_CALL_NO_RWG, void, ptr, ptr, i32)
+DEF_HELPER_FLAGS_3(gvec_ssra_d, TCG_CALL_NO_RWG, void, ptr, ptr, i32)
+
+DEF_HELPER_FLAGS_3(gvec_usra_b, TCG_CALL_NO_RWG, void, ptr, ptr, i32)
+DEF_HELPER_FLAGS_3(gvec_usra_h, TCG_CALL_NO_RWG, void, ptr, ptr, i32)
+DEF_HELPER_FLAGS_3(gvec_usra_s, TCG_CALL_NO_RWG, void, ptr, ptr, i32)
+DEF_HELPER_FLAGS_3(gvec_usra_d, TCG_CALL_NO_RWG, void, ptr, ptr, i32)
+
 #ifdef TARGET_AARCH64
 #include "helper-a64.h"
 #include "helper-sve.h"
diff --git a/target/arm/translate.h b/target/arm/translate.h
index cb7925ea46..1839a59a8e 100644
--- a/target/arm/translate.h
+++ b/target/arm/translate.h
@@ -285,8 +285,6 @@  extern const GVecGen3 mls_op[4];
 extern const GVecGen3 cmtst_op[4];
 extern const GVecGen3 sshl_op[4];
 extern const GVecGen3 ushl_op[4];
-extern const GVecGen2i ssra_op[4];
-extern const GVecGen2i usra_op[4];
 extern const GVecGen2i sri_op[4];
 extern const GVecGen2i sli_op[4];
 extern const GVecGen4 uqadd_op[4];
@@ -299,6 +297,11 @@  void gen_sshl_i32(TCGv_i32 d, TCGv_i32 a, TCGv_i32 b);
 void gen_ushl_i64(TCGv_i64 d, TCGv_i64 a, TCGv_i64 b);
 void gen_sshl_i64(TCGv_i64 d, TCGv_i64 a, TCGv_i64 b);
 
+void gen_gvec_ssra(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs,
+                   int64_t shift, uint32_t opr_sz, uint32_t max_sz);
+void gen_gvec_usra(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs,
+                   int64_t shift, uint32_t opr_sz, uint32_t max_sz);
+
 /*
  * Forward to the isar_feature_* tests given a DisasContext pointer.
  */
diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c
index 62e5729904..315de9a9b6 100644
--- a/target/arm/translate-a64.c
+++ b/target/arm/translate-a64.c
@@ -10188,19 +10188,8 @@  static void handle_vec_simd_shri(DisasContext *s, bool is_q, bool is_u,
 
     switch (opcode) {
     case 0x02: /* SSRA / USRA (accumulate) */
-        if (is_u) {
-            /* Shift count same as element size produces zero to add.  */
-            if (shift == 8 << size) {
-                goto done;
-            }
-            gen_gvec_op2i(s, is_q, rd, rn, shift, &usra_op[size]);
-        } else {
-            /* Shift count same as element size produces all sign to add.  */
-            if (shift == 8 << size) {
-                shift -= 1;
-            }
-            gen_gvec_op2i(s, is_q, rd, rn, shift, &ssra_op[size]);
-        }
+        gen_gvec_fn2i(s, is_q, rd, rn, shift,
+                      is_u ? gen_gvec_usra : gen_gvec_ssra, size);
         return;
     case 0x08: /* SRI */
         /* Shift count same as element size is valid but does nothing.  */
diff --git a/target/arm/translate.c b/target/arm/translate.c
index 74fac1d09c..c18140f2e6 100644
--- a/target/arm/translate.c
+++ b/target/arm/translate.c
@@ -3874,33 +3874,51 @@  static void gen_ssra_vec(unsigned vece, TCGv_vec d, TCGv_vec a, int64_t sh)
     tcg_gen_add_vec(vece, d, d, a);
 }
 
-static const TCGOpcode vecop_list_ssra[] = {
-    INDEX_op_sari_vec, INDEX_op_add_vec, 0
-};
+void gen_gvec_ssra(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs,
+                   int64_t shift, uint32_t opr_sz, uint32_t max_sz)
+{
+    static const TCGOpcode vecop_list[] = {
+        INDEX_op_sari_vec, INDEX_op_add_vec, 0
+    };
+    static const GVecGen2i ops[4] = {
+        { .fni8 = gen_ssra8_i64,
+          .fniv = gen_ssra_vec,
+          .fno = gen_helper_gvec_ssra_b,
+          .load_dest = true,
+          .opt_opc = vecop_list,
+          .vece = MO_8 },
+        { .fni8 = gen_ssra16_i64,
+          .fniv = gen_ssra_vec,
+          .fno = gen_helper_gvec_ssra_h,
+          .load_dest = true,
+          .opt_opc = vecop_list,
+          .vece = MO_16 },
+        { .fni4 = gen_ssra32_i32,
+          .fniv = gen_ssra_vec,
+          .fno = gen_helper_gvec_ssra_s,
+          .load_dest = true,
+          .opt_opc = vecop_list,
+          .vece = MO_32 },
+        { .fni8 = gen_ssra64_i64,
+          .fniv = gen_ssra_vec,
+          .fno = gen_helper_gvec_ssra_b,
+          .prefer_i64 = TCG_TARGET_REG_BITS == 64,
+          .opt_opc = vecop_list,
+          .load_dest = true,
+          .vece = MO_64 },
+    };
 
-const GVecGen2i ssra_op[4] = {
-    { .fni8 = gen_ssra8_i64,
-      .fniv = gen_ssra_vec,
-      .load_dest = true,
-      .opt_opc = vecop_list_ssra,
-      .vece = MO_8 },
-    { .fni8 = gen_ssra16_i64,
-      .fniv = gen_ssra_vec,
-      .load_dest = true,
-      .opt_opc = vecop_list_ssra,
-      .vece = MO_16 },
-    { .fni4 = gen_ssra32_i32,
-      .fniv = gen_ssra_vec,
-      .load_dest = true,
-      .opt_opc = vecop_list_ssra,
-      .vece = MO_32 },
-    { .fni8 = gen_ssra64_i64,
-      .fniv = gen_ssra_vec,
-      .prefer_i64 = TCG_TARGET_REG_BITS == 64,
-      .opt_opc = vecop_list_ssra,
-      .load_dest = true,
-      .vece = MO_64 },
-};
+    /* tszimm encoding produces immediates in the range [1..esize]. */
+    tcg_debug_assert(shift > 0);
+    tcg_debug_assert(shift <= (8 << vece));
+
+    /*
+     * Shifts larger than the element size are architecturally valid.
+     * Signed results in all sign bits.
+     */
+    shift = MIN(shift, (8 << vece) - 1);
+    tcg_gen_gvec_2i(rd_ofs, rm_ofs, opr_sz, max_sz, shift, &ops[vece]);
+}
 
 static void gen_usra8_i64(TCGv_i64 d, TCGv_i64 a, int64_t shift)
 {
@@ -3932,33 +3950,55 @@  static void gen_usra_vec(unsigned vece, TCGv_vec d, TCGv_vec a, int64_t sh)
     tcg_gen_add_vec(vece, d, d, a);
 }
 
-static const TCGOpcode vecop_list_usra[] = {
-    INDEX_op_shri_vec, INDEX_op_add_vec, 0
-};
+void gen_gvec_usra(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs,
+                   int64_t shift, uint32_t opr_sz, uint32_t max_sz)
+{
+    static const TCGOpcode vecop_list[] = {
+        INDEX_op_shri_vec, INDEX_op_add_vec, 0
+    };
+    static const GVecGen2i ops[4] = {
+        { .fni8 = gen_usra8_i64,
+          .fniv = gen_usra_vec,
+          .fno = gen_helper_gvec_usra_b,
+          .load_dest = true,
+          .opt_opc = vecop_list,
+          .vece = MO_8, },
+        { .fni8 = gen_usra16_i64,
+          .fniv = gen_usra_vec,
+          .fno = gen_helper_gvec_usra_h,
+          .load_dest = true,
+          .opt_opc = vecop_list,
+          .vece = MO_16, },
+        { .fni4 = gen_usra32_i32,
+          .fniv = gen_usra_vec,
+          .fno = gen_helper_gvec_usra_s,
+          .load_dest = true,
+          .opt_opc = vecop_list,
+          .vece = MO_32, },
+        { .fni8 = gen_usra64_i64,
+          .fniv = gen_usra_vec,
+          .fno = gen_helper_gvec_usra_d,
+          .prefer_i64 = TCG_TARGET_REG_BITS == 64,
+          .load_dest = true,
+          .opt_opc = vecop_list,
+          .vece = MO_64, },
+    };
 
-const GVecGen2i usra_op[4] = {
-    { .fni8 = gen_usra8_i64,
-      .fniv = gen_usra_vec,
-      .load_dest = true,
-      .opt_opc = vecop_list_usra,
-      .vece = MO_8, },
-    { .fni8 = gen_usra16_i64,
-      .fniv = gen_usra_vec,
-      .load_dest = true,
-      .opt_opc = vecop_list_usra,
-      .vece = MO_16, },
-    { .fni4 = gen_usra32_i32,
-      .fniv = gen_usra_vec,
-      .load_dest = true,
-      .opt_opc = vecop_list_usra,
-      .vece = MO_32, },
-    { .fni8 = gen_usra64_i64,
-      .fniv = gen_usra_vec,
-      .prefer_i64 = TCG_TARGET_REG_BITS == 64,
-      .load_dest = true,
-      .opt_opc = vecop_list_usra,
-      .vece = MO_64, },
-};
+    /* tszimm encoding produces immediates in the range [1..esize]. */
+    tcg_debug_assert(shift > 0);
+    tcg_debug_assert(shift <= (8 << vece));
+
+    /*
+     * Shifts larger than the element size are architecturally valid.
+     * Unsigned results in all zeros as input to accumulate: nop.
+     */
+    if (shift < (8 << vece)) {
+        tcg_gen_gvec_2i(rd_ofs, rm_ofs, opr_sz, max_sz, shift, &ops[vece]);
+    } else {
+        /* Nop, but we do need to clear the tail. */
+        tcg_gen_gvec_mov(vece, rd_ofs, rd_ofs, opr_sz, max_sz);
+    }
+}
 
 static void gen_shr8_ins_i64(TCGv_i64 d, TCGv_i64 a, int64_t shift)
 {
@@ -5220,19 +5260,12 @@  static int disas_neon_data_insn(DisasContext *s, uint32_t insn)
                 case 1:  /* VSRA */
                     /* Right shift comes here negative.  */
                     shift = -shift;
-                    /* Shifts larger than the element size are architecturally
-                     * valid.  Unsigned results in all zeros; signed results
-                     * in all sign bits.
-                     */
-                    if (!u) {
-                        tcg_gen_gvec_2i(rd_ofs, rm_ofs, vec_size, vec_size,
-                                        MIN(shift, (8 << size) - 1),
-                                        &ssra_op[size]);
-                    } else if (shift >= 8 << size) {
-                        /* rd += 0 */
+                    if (u) {
+                        gen_gvec_usra(size, rd_ofs, rm_ofs, shift,
+                                      vec_size, vec_size);
                     } else {
-                        tcg_gen_gvec_2i(rd_ofs, rm_ofs, vec_size, vec_size,
-                                        shift, &usra_op[size]);
+                        gen_gvec_ssra(size, rd_ofs, rm_ofs, shift,
+                                      vec_size, vec_size);
                     }
                     return 0;
 
diff --git a/target/arm/vec_helper.c b/target/arm/vec_helper.c
index 3d534188a8..230085b35e 100644
--- a/target/arm/vec_helper.c
+++ b/target/arm/vec_helper.c
@@ -899,6 +899,31 @@  void HELPER(gvec_sqsub_d)(void *vd, void *vq, void *vn,
     clear_tail(d, oprsz, simd_maxsz(desc));
 }
 
+
+#define DO_SRA(NAME, TYPE)                              \
+void HELPER(NAME)(void *vd, void *vn, uint32_t desc)    \
+{                                                       \
+    intptr_t i, oprsz = simd_oprsz(desc);               \
+    int shift = simd_data(desc);                        \
+    TYPE *d = vd, *n = vn;                              \
+    for (i = 0; i < oprsz / sizeof(TYPE); i++) {        \
+        d[i] += n[i] >> shift;                          \
+    }                                                   \
+    clear_tail(d, oprsz, simd_maxsz(desc));             \
+}
+
+DO_SRA(gvec_ssra_b, int8_t)
+DO_SRA(gvec_ssra_h, int16_t)
+DO_SRA(gvec_ssra_s, int32_t)
+DO_SRA(gvec_ssra_d, int64_t)
+
+DO_SRA(gvec_usra_b, uint8_t)
+DO_SRA(gvec_usra_h, uint16_t)
+DO_SRA(gvec_usra_s, uint32_t)
+DO_SRA(gvec_usra_d, uint64_t)
+
+#undef DO_SRA
+
 /*
  * Convert float16 to float32, raising no exceptions and
  * preserving exceptional values, including SNaN.