[PATCH v4 63/71] tcg/tci: Implement movcond

[Richard Henderson]

When this opcode is not available in the backend, tcg middle-end
will expand this as a series of 5 opcodes.  So implementing this
saves bytecode space.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
 tcg/tci/tcg-target.h     |  4 ++--
 tcg/tci.c                | 16 +++++++++++++++-
 tcg/tci/tcg-target.c.inc | 10 +++++++---
 3 files changed, 24 insertions(+), 6 deletions(-)

diff --git a/tcg/tci/tcg-target.h b/tcg/tci/tcg-target.h
index 17911d3297..f53773a555 100644
--- a/tcg/tci/tcg-target.h
+++ b/tcg/tci/tcg-target.h
@@ -82,7 +82,7 @@
 #define TCG_TARGET_HAS_not_i32          1
 #define TCG_TARGET_HAS_orc_i32          0
 #define TCG_TARGET_HAS_rot_i32          1
-#define TCG_TARGET_HAS_movcond_i32      0
+#define TCG_TARGET_HAS_movcond_i32      1
 #define TCG_TARGET_HAS_muls2_i32        0
 #define TCG_TARGET_HAS_muluh_i32        0
 #define TCG_TARGET_HAS_mulsh_i32        0
@@ -119,7 +119,7 @@
 #define TCG_TARGET_HAS_not_i64          1
 #define TCG_TARGET_HAS_orc_i64          0
 #define TCG_TARGET_HAS_rot_i64          1
-#define TCG_TARGET_HAS_movcond_i64      0
+#define TCG_TARGET_HAS_movcond_i64      1
 #define TCG_TARGET_HAS_muls2_i64        0
 #define TCG_TARGET_HAS_add2_i32         0
 #define TCG_TARGET_HAS_sub2_i32         0
diff --git a/tcg/tci.c b/tcg/tci.c
index c229050c66..2391dd4d3b 100644
--- a/tcg/tci.c
+++ b/tcg/tci.c
@@ -169,6 +169,7 @@ static void tci_args_rrrr(uint32_t insn,
     *r2 = extract32(insn, 16, 4);
     *r3 = extract32(insn, 20, 4);
 }
+#endif
 
 static void tci_args_rrrrrc(uint32_t insn, TCGReg *r0, TCGReg *r1,
                             TCGReg *r2, TCGReg *r3, TCGReg *r4, TCGCond *c5)
@@ -181,6 +182,7 @@ static void tci_args_rrrrrc(uint32_t insn, TCGReg *r0, 
TCGReg *r1,
     *c5 = extract32(insn, 28, 4);
 }
 
+#if TCG_TARGET_REG_BITS == 32
 static void tci_args_rrrrrr(uint32_t insn, TCGReg *r0, TCGReg *r1,
                             TCGReg *r2, TCGReg *r3, TCGReg *r4, TCGReg *r5)
 {
@@ -421,6 +423,11 @@ uintptr_t QEMU_DISABLE_CFI tcg_qemu_tb_exec(CPUArchState 
*env,
             tci_args_rrrc(insn, &r0, &r1, &r2, &condition);
             regs[r0] = tci_compare32(regs[r1], regs[r2], condition);
             break;
+        case INDEX_op_movcond_i32:
+            tci_args_rrrrrc(insn, &r0, &r1, &r2, &r3, &r4, &condition);
+            tmp32 = tci_compare32(regs[r1], regs[r2], condition);
+            regs[r0] = regs[tmp32 ? r3 : r4];
+            break;
 #if TCG_TARGET_REG_BITS == 32
         case INDEX_op_setcond2_i32:
             tci_args_rrrrrc(insn, &r0, &r1, &r2, &r3, &r4, &condition);
@@ -433,6 +440,11 @@ uintptr_t QEMU_DISABLE_CFI tcg_qemu_tb_exec(CPUArchState 
*env,
             tci_args_rrrc(insn, &r0, &r1, &r2, &condition);
             regs[r0] = tci_compare64(regs[r1], regs[r2], condition);
             break;
+        case INDEX_op_movcond_i64:
+            tci_args_rrrrrc(insn, &r0, &r1, &r2, &r3, &r4, &condition);
+            tmp32 = tci_compare64(regs[r1], regs[r2], condition);
+            regs[r0] = regs[tmp32 ? r3 : r4];
+            break;
 #endif
         CASE_32_64(mov)
             tci_args_rr(insn, &r0, &r1);
@@ -1138,7 +1150,8 @@ int print_insn_tci(bfd_vma addr, disassemble_info *info)
                            op_name, str_r(r0), str_r(r1), str_r(r2), pos, len);
         break;
 
-#if TCG_TARGET_REG_BITS == 32
+    case INDEX_op_movcond_i32:
+    case INDEX_op_movcond_i64:
     case INDEX_op_setcond2_i32:
         tci_args_rrrrrc(insn, &r0, &r1, &r2, &r3, &r4, &c);
         info->fprintf_func(info->stream, "%-12s  %s,%s,%s,%s,%s,%s",
@@ -1146,6 +1159,7 @@ int print_insn_tci(bfd_vma addr, disassemble_info *info)
                            str_r(r3), str_r(r4), str_c(c));
         break;
 
+#if TCG_TARGET_REG_BITS == 32
     case INDEX_op_mulu2_i32:
         tci_args_rrrr(insn, &r0, &r1, &r2, &r3);
         info->fprintf_func(info->stream, "%-12s  %s,%s,%s,%s",
diff --git a/tcg/tci/tcg-target.c.inc b/tcg/tci/tcg-target.c.inc
index db29bc6e54..a0c458a60a 100644
--- a/tcg/tci/tcg-target.c.inc
+++ b/tcg/tci/tcg-target.c.inc
@@ -133,9 +133,12 @@ static TCGConstraintSetIndex tcg_target_op_def(TCGOpcode 
op)
         return C_O0_I4(r, r, r, r);
     case INDEX_op_mulu2_i32:
         return C_O2_I2(r, r, r, r);
+#endif
+
+    case INDEX_op_movcond_i32:
+    case INDEX_op_movcond_i64:
     case INDEX_op_setcond2_i32:
         return C_O1_I4(r, r, r, r, r);
-#endif
 
     case INDEX_op_qemu_ld_i32:
         return (TARGET_LONG_BITS <= TCG_TARGET_REG_BITS
@@ -419,6 +422,7 @@ static void tcg_out_op_rrrr(TCGContext *s, TCGOpcode op,
     insn = deposit32(insn, 20, 4, r3);
     tcg_out32(s, insn);
 }
+#endif
 
 static void tcg_out_op_rrrrrc(TCGContext *s, TCGOpcode op,
                               TCGReg r0, TCGReg r1, TCGReg r2,
@@ -436,6 +440,7 @@ static void tcg_out_op_rrrrrc(TCGContext *s, TCGOpcode op,
     tcg_out32(s, insn);
 }
 
+#if TCG_TARGET_REG_BITS == 32
 static void tcg_out_op_rrrrrr(TCGContext *s, TCGOpcode op,
                               TCGReg r0, TCGReg r1, TCGReg r2,
                               TCGReg r3, TCGReg r4, TCGReg r5)
@@ -591,12 +596,11 @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, 
const TCGArg *args,
         tcg_out_op_rrrc(s, opc, args[0], args[1], args[2], args[3]);
         break;
 
-#if TCG_TARGET_REG_BITS == 32
+    CASE_32_64(movcond)
     case INDEX_op_setcond2_i32:
         tcg_out_op_rrrrrc(s, opc, args[0], args[1], args[2],
                           args[3], args[4], args[5]);
         break;
-#endif
 
     CASE_32_64(ld8u)
     CASE_32_64(ld8s)
-- 
2.25.1