[Qemu-devel] [PATCH v3 16/25] tcg-ppc64: Merge ppc32 brcond2, setcond2, muluh

[Richard Henderson]

Now passes tcg_add_target_add_op_defs assertions, but
not complete enough to function.

Signed-off-by: Richard Henderson <address@hidden>
---
 tcg/ppc64/tcg-target.c | 96 +++++++++++++++++++++++++++++++++++++++++++++++++-
 tcg/ppc64/tcg-target.h |  2 +-
 2 files changed, 96 insertions(+), 2 deletions(-)

diff --git a/tcg/ppc64/tcg-target.c b/tcg/ppc64/tcg-target.c
index 8152ccf..4b846a2 100644
--- a/tcg/ppc64/tcg-target.c
+++ b/tcg/ppc64/tcg-target.c
@@ -796,6 +796,8 @@ static void tcg_out_cmp(TCGContext *s, int cond, TCGArg 
arg1, TCGArg arg2,
     int imm;
     uint32_t op;
 
+    tcg_debug_assert(TCG_TARGET_REG_BITS == 64 || type == TCG_TYPE_I32);
+
     /* Simplify the comparisons below wrt CMPI.  */
     if (type == TCG_TYPE_I32) {
         arg2 = (int32_t)arg2;
@@ -1083,6 +1085,83 @@ static void tcg_out_movcond(TCGContext *s, TCGType type, 
TCGCond cond,
     }
 }
 
+static void tcg_out_cmp2(TCGContext *s, const TCGArg *args,
+                         const int *const_args)
+{
+    static const struct { uint8_t bit1, bit2; } bits[] = {
+        [TCG_COND_LT ] = { CR_LT, CR_LT },
+        [TCG_COND_LE ] = { CR_LT, CR_GT },
+        [TCG_COND_GT ] = { CR_GT, CR_GT },
+        [TCG_COND_GE ] = { CR_GT, CR_LT },
+        [TCG_COND_LTU] = { CR_LT, CR_LT },
+        [TCG_COND_LEU] = { CR_LT, CR_GT },
+        [TCG_COND_GTU] = { CR_GT, CR_GT },
+        [TCG_COND_GEU] = { CR_GT, CR_LT },
+    };
+
+    TCGCond cond = args[4], cond2;
+    TCGArg al, ah, bl, bh;
+    int blconst, bhconst;
+    int op, bit1, bit2;
+
+    al = args[0];
+    ah = args[1];
+    bl = args[2];
+    bh = args[3];
+    blconst = const_args[2];
+    bhconst = const_args[3];
+
+    switch (cond) {
+    case TCG_COND_EQ:
+        op = CRAND;
+        goto do_equality;
+    case TCG_COND_NE:
+        op = CRNAND;
+    do_equality:
+        tcg_out_cmp(s, cond, al, bl, blconst, 6, TCG_TYPE_I32);
+        tcg_out_cmp(s, cond, ah, bh, bhconst, 7, TCG_TYPE_I32);
+        tcg_out32(s, op | BT(7, CR_EQ) | BA(6, CR_EQ) | BB(7, CR_EQ));
+        break;
+
+    case TCG_COND_LT:
+    case TCG_COND_LE:
+    case TCG_COND_GT:
+    case TCG_COND_GE:
+    case TCG_COND_LTU:
+    case TCG_COND_LEU:
+    case TCG_COND_GTU:
+    case TCG_COND_GEU:
+        bit1 = bits[cond].bit1;
+        bit2 = bits[cond].bit2;
+        op = (bit1 != bit2 ? CRANDC : CRAND);
+        cond2 = tcg_unsigned_cond(cond);
+
+        tcg_out_cmp(s, cond, ah, bh, bhconst, 6, TCG_TYPE_I32);
+        tcg_out_cmp(s, cond2, al, bl, blconst, 7, TCG_TYPE_I32);
+        tcg_out32(s, op | BT(7, CR_EQ) | BA(6, CR_EQ) | BB(7, bit2));
+        tcg_out32(s, CROR | BT(7, CR_EQ) | BA(6, bit1) | BB(7, CR_EQ));
+        break;
+
+    default:
+        tcg_abort();
+    }
+}
+
+static void tcg_out_setcond2(TCGContext *s, const TCGArg *args,
+                             const int *const_args)
+{
+    tcg_out_cmp2(s, args + 1, const_args + 1);
+    tcg_out32(s, MFOCRF | RT(TCG_REG_R0) | FXM(7));
+    tcg_out_rlw(s, RLWINM, args[0], TCG_REG_R0, 31, 31, 31);
+}
+
+static void tcg_out_brcond2 (TCGContext *s, const TCGArg *args,
+                             const int *const_args)
+{
+    tcg_out_cmp2(s, args, const_args);
+    tcg_out_bc(s, BC | BI(7, CR_EQ) | BO_COND_TRUE, args[5]);
+}
+
 void ppc_tb_set_jmp_target(uintptr_t jmp_addr, uintptr_t addr)
 {
     TCGContext s;
@@ -1760,11 +1839,13 @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, 
const TCGArg *args,
         tcg_out_brcond(s, args[2], args[0], args[1], const_args[1],
                        args[3], TCG_TYPE_I32);
         break;
-
     case INDEX_op_brcond_i64:
         tcg_out_brcond(s, args[2], args[0], args[1], const_args[1],
                        args[3], TCG_TYPE_I64);
         break;
+    case INDEX_op_brcond2_i32:
+        tcg_out_brcond2(s, args, const_args);
+        break;
 
     case INDEX_op_neg_i32:
     case INDEX_op_neg_i64:
@@ -1886,6 +1967,9 @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, 
const TCGArg *args,
         tcg_out_setcond(s, TCG_TYPE_I64, args[3], args[0], args[1], args[2],
                         const_args[2]);
         break;
+    case INDEX_op_setcond2_i32:
+        tcg_out_setcond2(s, args, const_args);
+        break;
 
     case INDEX_op_bswap16_i32:
     case INDEX_op_bswap16_i64:
@@ -2036,6 +2120,9 @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, 
const TCGArg *args,
         }
         break;
 
+    case INDEX_op_muluh_i32:
+        tcg_out32(s, MULHWU | TAB(args[0], args[1], args[2]));
+        break;
     case INDEX_op_muluh_i64:
         tcg_out32(s, MULHDU | TAB(args[0], args[1], args[2]));
         break;
@@ -2101,6 +2188,8 @@ static const TCGTargetOpDef ppc_op_defs[] = {
 
     { INDEX_op_deposit_i32, { "r", "0", "rZ" } },
 
+    { INDEX_op_muluh_i32, { "r", "r", "r" } },
+
 #if TCG_TARGET_REG_BITS == 64
     { INDEX_op_ld8u_i64, { "r", "r" } },
     { INDEX_op_ld8s_i64, { "r", "r" } },
@@ -2155,6 +2244,11 @@ static const TCGTargetOpDef ppc_op_defs[] = {
     { INDEX_op_muluh_i64, { "r", "r", "r" } },
 #endif
 
+#if TCG_TARGET_REG_BITS == 32
+    { INDEX_op_brcond2_i32, { "r", "r", "ri", "ri" } },
+    { INDEX_op_setcond2_i32, { "r", "r", "r", "ri", "ri" } },
+#endif
+
 #if TCG_TARGET_REG_BITS == 64
     { INDEX_op_add2_i64, { "r", "r", "r", "r", "rI", "rZM" } },
     { INDEX_op_sub2_i64, { "r", "r", "rI", "rZM", "r", "r" } },
diff --git a/tcg/ppc64/tcg-target.h b/tcg/ppc64/tcg-target.h
index 726635f..7b90087 100644
--- a/tcg/ppc64/tcg-target.h
+++ b/tcg/ppc64/tcg-target.h
@@ -70,7 +70,7 @@ typedef enum {
 #define TCG_TARGET_HAS_movcond_i32      1
 #define TCG_TARGET_HAS_mulu2_i32        0
 #define TCG_TARGET_HAS_muls2_i32        0
-#define TCG_TARGET_HAS_muluh_i32        0
+#define TCG_TARGET_HAS_muluh_i32        1
 #define TCG_TARGET_HAS_mulsh_i32        0
 
 #if TCG_TARGET_REG_BITS == 64
-- 
1.9.3