Merge tag 'pull-tcg-20240922' of https://gitlab.com/rth7680/qemu into staging

target/ppc: Fix lxvx/stxvx facility check
linux-user: update syscall_nr.h to Linux v6.10
linux-user: update syscall.tbl to Linux v6.11
tcg: Fix iteration step in 32-bit gvec operation
tcg: Propagate new TCGOp to add_as_label_use
tcg/*: Do not expand cmp_vec, cmpsel_vec early
tcg/optimize: Fold movcond with true and false values identical
tcg/optimize: Optimize cmp_vec and cmpsel_vec
tcg/optimize: Optimize bitsel_vec
tcg/i386: Optimize cmpsel with constant 0 operand 3.
tcg/i386: Implement cmp_vec with avx512 insns
tcg/i386: Implement cmpsel_vec with avx512 insns
tcg/i386: Implement vector TST{EQ,NE} for avx512
tcg/ppc: Implement cmpsel_vec and optimize with constant 0/-1 arguments
tcg/s390x: Implement cmpsel_vec and optimize with constant 0/-1 arguments

# -----BEGIN PGP SIGNATURE-----
#
# iQFQBAABCgA7FiEEekgeeIaLTbaoWgXAZN846K9+IV8FAmbwBsIdHHJpY2hhcmQu
# aGVuZGVyc29uQGxpbmFyby5vcmcACgkQZN846K9+IV/bzwf40V8fsRhfc8u/k2Xw
# +bXyfyX7ydgB+82YoO71+Wy15ntmYmbL/6O9kGuJXWX8HRDrAR8Js9zDgveysw5m
# m/EG+XAbVjYhjaoaaz2nfr+5auQoos9NoCji3s3UJln6sCuZ2Enl0DTFluoHZfgr
# /YVAghJ4dwH0hfzO9kmsKmJ9I19HugMMN4dNvEcjQyDDUhgGGjkUaRtGNxwuiwuf
# ArL2XhFauXgloryDmpUKQ0BJJ0t8rdPlaRJ8dHqJUnJwav0TW3QVXbrs8gOkmmbg
# 9plYTC9DlTguQobvapCDVjiGy/6yuSF/AUpcdQbTRtTi3BrzhTydfLlOtI64GtaQ
# OBKd
# =ID/g
# -----END PGP SIGNATURE-----
# gpg: Signature made Sun 22 Sep 2024 13:00:02 BST
# gpg:                using RSA key 7A481E78868B4DB6A85A05C064DF38E8AF7E215F
# gpg:                issuer "richard.henderson@linaro.org"
# gpg: Good signature from "Richard Henderson <richard.henderson@linaro.org>" [full]
# Primary key fingerprint: 7A48 1E78 868B 4DB6 A85A  05C0 64DF 38E8 AF7E 215F

* tag 'pull-tcg-20240922' of https://gitlab.com/rth7680/qemu: (31 commits)
  linux-user: update syscall.tbl to Linux v6.11
  linux-user,loongarch: move to syscalltbl file
  linux-user,hexagon: move to syscalltbl file
  linux-user,riscv: move to syscalltbl file
  linux-user,openrisc: move to syscalltbl file
  linux-user,aarch64: move to syscalltbl file
  linux-user: update syscall.tbl to Linux v6.10
  linux-user, mips: update syscall-args-o32.c.inc to Linux v6.10
  linux-user: update syscall_nr.h to Linux v6.10
  target/ppc: Fix lxvx/stxvx facility check
  tcg/s390x: Optimize cmpsel with constant 0/-1 arguments
  tcg/s390x: Implement cmpsel_vec
  tcg/ppc: Optimize cmpsel with constant 0/-1 arguments
  tcg/ppc: Implement cmpsel_vec
  tcg/i386: Implement vector TST{EQ,NE} for avx512
  tcg/i386: Implement cmpsel_vec with avx512 insns
  tcg/i386: Add predicate parameters to tcg_out_evex_opc
  tcg/i386: Implement cmp_vec with avx512 insns
  tcg/i386: Optimize cmpsel with constant 0 operand 3.
  tcg/optimize: Optimize bitsel_vec
  ...

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

1 files changed, 99 insertions, 0 deletions

diff --git a/tcg/optimize.c b/tcg/optimize.c
index ba16ec27e2..e9ef16b3c6 100644
--- a/tcg/optimize.c
+++ b/tcg/optimize.c
@@ -1851,6 +1851,11 @@ static bool fold_movcond(OptContext *ctx, TCGOp *op)
 {
     int i;
 
+    /* If true and false values are the same, eliminate the cmp. */
+    if (args_are_copies(op->args[3], op->args[4])) {
+        return tcg_opt_gen_mov(ctx, op, op->args[0], op->args[3]);
+    }
+
     /*
      * Canonicalize the "false" input reg to match the destination reg so
      * that the tcg backend can implement a "move if true" operation.
@@ -2417,6 +2422,36 @@ static bool fold_setcond2(OptContext *ctx, TCGOp *op)
     return tcg_opt_gen_movi(ctx, op, op->args[0], i);
 }
 
+static bool fold_cmp_vec(OptContext *ctx, TCGOp *op)
+{
+    /* Canonicalize the comparison to put immediate second. */
+    if (swap_commutative(NO_DEST, &op->args[1], &op->args[2])) {
+        op->args[3] = tcg_swap_cond(op->args[3]);
+    }
+    return false;
+}
+
+static bool fold_cmpsel_vec(OptContext *ctx, TCGOp *op)
+{
+    /* If true and false values are the same, eliminate the cmp. */
+    if (args_are_copies(op->args[3], op->args[4])) {
+        return tcg_opt_gen_mov(ctx, op, op->args[0], op->args[3]);
+    }
+
+    /* Canonicalize the comparison to put immediate second. */
+    if (swap_commutative(NO_DEST, &op->args[1], &op->args[2])) {
+        op->args[5] = tcg_swap_cond(op->args[5]);
+    }
+    /*
+     * Canonicalize the "false" input reg to match the destination,
+     * so that the tcg backend can implement "move if true".
+     */
+    if (swap_commutative(op->args[0], &op->args[4], &op->args[3])) {
+        op->args[5] = tcg_invert_cond(op->args[5]);
+    }
+    return false;
+}
+
 static bool fold_sextract(OptContext *ctx, TCGOp *op)
 {
     uint64_t z_mask, s_mask, s_mask_old;
@@ -2702,6 +2737,61 @@ static bool fold_xor(OptContext *ctx, TCGOp *op)
     return fold_masks(ctx, op);
 }
 
+static bool fold_bitsel_vec(OptContext *ctx, TCGOp *op)
+{
+    /* If true and false values are the same, eliminate the cmp. */
+    if (args_are_copies(op->args[2], op->args[3])) {
+        return tcg_opt_gen_mov(ctx, op, op->args[0], op->args[2]);
+    }
+
+    if (arg_is_const(op->args[2]) && arg_is_const(op->args[3])) {
+        uint64_t tv = arg_info(op->args[2])->val;
+        uint64_t fv = arg_info(op->args[3])->val;
+
+        if (tv == -1 && fv == 0) {
+            return tcg_opt_gen_mov(ctx, op, op->args[0], op->args[1]);
+        }
+        if (tv == 0 && fv == -1) {
+            if (TCG_TARGET_HAS_not_vec) {
+                op->opc = INDEX_op_not_vec;
+                return fold_not(ctx, op);
+            } else {
+                op->opc = INDEX_op_xor_vec;
+                op->args[2] = arg_new_constant(ctx, -1);
+                return fold_xor(ctx, op);
+            }
+        }
+    }
+    if (arg_is_const(op->args[2])) {
+        uint64_t tv = arg_info(op->args[2])->val;
+        if (tv == -1) {
+            op->opc = INDEX_op_or_vec;
+            op->args[2] = op->args[3];
+            return fold_or(ctx, op);
+        }
+        if (tv == 0 && TCG_TARGET_HAS_andc_vec) {
+            op->opc = INDEX_op_andc_vec;
+            op->args[2] = op->args[1];
+            op->args[1] = op->args[3];
+            return fold_andc(ctx, op);
+        }
+    }
+    if (arg_is_const(op->args[3])) {
+        uint64_t fv = arg_info(op->args[3])->val;
+        if (fv == 0) {
+            op->opc = INDEX_op_and_vec;
+            return fold_and(ctx, op);
+        }
+        if (fv == -1 && TCG_TARGET_HAS_orc_vec) {
+            op->opc = INDEX_op_orc_vec;
+            op->args[2] = op->args[1];
+            op->args[1] = op->args[3];
+            return fold_orc(ctx, op);
+        }
+    }
+    return false;
+}
+
 /* Propagate constants and copies, fold constant expressions. */
 void tcg_optimize(TCGContext *s)
 {
@@ -2923,6 +3013,15 @@ void tcg_optimize(TCGContext *s)
         case INDEX_op_setcond2_i32:
             done = fold_setcond2(&ctx, op);
             break;
+        case INDEX_op_cmp_vec:
+            done = fold_cmp_vec(&ctx, op);
+            break;
+        case INDEX_op_cmpsel_vec:
+            done = fold_cmpsel_vec(&ctx, op);
+            break;
+        case INDEX_op_bitsel_vec:
+            done = fold_bitsel_vec(&ctx, op);
+            break;
         CASE_OP_32_64(sextract):
             done = fold_sextract(&ctx, op);
             break;