1 files changed, 47 insertions, 25 deletions

diff --git a/tcg/arm/tcg-target.c b/tcg/arm/tcg-target.c
index dee1ebc..8139da1 100644
--- a/tcg/arm/tcg-target.c
+++ b/tcg/arm/tcg-target.c
@@ -21,7 +21,9 @@
  * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
  * THE SOFTWARE.
  */
-const char *tcg_target_reg_names[TCG_TARGET_NB_REGS] = {
+
+#ifndef NDEBUG
+static const char * const tcg_target_reg_names[TCG_TARGET_NB_REGS] = {
     "%r0",
     "%r1",
     "%r2",
@@ -38,8 +40,9 @@ const char *tcg_target_reg_names[TCG_TARGET_NB_REGS] = {
     "%r13",
     "%r14",
 };
+#endif
 
-int tcg_target_reg_alloc_order[] = {
+static const int tcg_target_reg_alloc_order[] = {
     TCG_REG_R0,
     TCG_REG_R1,
     TCG_REG_R2,
@@ -57,10 +60,10 @@ int tcg_target_reg_alloc_order[] = {
     TCG_REG_R14,
 };
 
-const int tcg_target_call_iarg_regs[4] = {
+static const int tcg_target_call_iarg_regs[4] = {
     TCG_REG_R0, TCG_REG_R1, TCG_REG_R2, TCG_REG_R3
 };
-const int tcg_target_call_oarg_regs[2] = {
+static const int tcg_target_call_oarg_regs[2] = {
     TCG_REG_R0, TCG_REG_R1
 };
 
@@ -91,7 +94,7 @@ static inline int tcg_target_get_call_iarg_regs_count(int flags)
 }
 
 /* parse target specific constraints */
-int target_parse_constraint(TCGArgConstraint *ct, const char **pct_str)
+static int target_parse_constraint(TCGArgConstraint *ct, const char **pct_str)
 {
     const char *ct_str;
 
@@ -292,10 +295,19 @@ static inline void tcg_out_dat_reg2(TCGContext *s,
                 int cond, int opc0, int opc1, int rd0, int rd1,
                 int rn0, int rn1, int rm0, int rm1, int shift)
 {
-    tcg_out32(s, (cond << 28) | (0 << 25) | (opc0 << 21) | (1 << 20) |
-                    (rn0 << 16) | (rd0 << 12) | shift | rm0);
-    tcg_out32(s, (cond << 28) | (0 << 25) | (opc1 << 21) |
-                    (rn1 << 16) | (rd1 << 12) | shift | rm1);
+    if (rd0 == rn1 || rd0 == rm1) {
+        tcg_out32(s, (cond << 28) | (0 << 25) | (opc0 << 21) | (1 << 20) |
+                        (rn0 << 16) | (8 << 12) | shift | rm0);
+        tcg_out32(s, (cond << 28) | (0 << 25) | (opc1 << 21) |
+                        (rn1 << 16) | (rd1 << 12) | shift | rm1);
+        tcg_out_dat_reg(s, cond, ARITH_MOV,
+                        rd0, 0, TCG_REG_R8, SHIFT_IMM_LSL(0));
+    } else {
+        tcg_out32(s, (cond << 28) | (0 << 25) | (opc0 << 21) | (1 << 20) |
+                        (rn0 << 16) | (rd0 << 12) | shift | rm0);
+        tcg_out32(s, (cond << 28) | (0 << 25) | (opc1 << 21) |
+                        (rn1 << 16) | (rd1 << 12) | shift | rm1);
+    }
 }
 
 static inline void tcg_out_dat_imm(TCGContext *s,
@@ -846,10 +858,10 @@ static inline void tcg_out_qemu_ld(TCGContext *s, int cond,
     else
         data_reg2 = 0; /* surpress warning */
     addr_reg = *args++;
-#if TARGET_LONG_BITS == 64
-    addr_reg2 = *args++;
-#endif
 #ifdef CONFIG_SOFTMMU
+# if TARGET_LONG_BITS == 64
+    addr_reg2 = *args++;
+# endif
     mem_index = *args;
     s_bits = opc & 3;
 
@@ -999,8 +1011,13 @@ static inline void tcg_out_qemu_ld(TCGContext *s, int cond,
     case 3:
         /* TODO: use block load -
          * check that data_reg2 > data_reg or the other way */
-        tcg_out_ld32_12(s, COND_AL, data_reg, addr_reg, 0);
-        tcg_out_ld32_12(s, COND_AL, data_reg2, addr_reg, 4);
+        if (data_reg == addr_reg) {
+            tcg_out_ld32_12(s, COND_AL, data_reg2, addr_reg, 4);
+            tcg_out_ld32_12(s, COND_AL, data_reg, addr_reg, 0);
+        } else {
+            tcg_out_ld32_12(s, COND_AL, data_reg, addr_reg, 0);
+            tcg_out_ld32_12(s, COND_AL, data_reg2, addr_reg, 4);
+        }
         break;
     }
 #endif
@@ -1024,10 +1041,10 @@ static inline void tcg_out_qemu_st(TCGContext *s, int cond,
     else
         data_reg2 = 0; /* surpress warning */
     addr_reg = *args++;
-#if TARGET_LONG_BITS == 64
-    addr_reg2 = *args++;
-#endif
 #ifdef CONFIG_SOFTMMU
+# if TARGET_LONG_BITS == 64
+    addr_reg2 = *args++;
+# endif
     mem_index = *args;
     s_bits = opc & 3;
 
@@ -1229,19 +1246,24 @@ static inline void tcg_out_op(TCGContext *s, int opc,
         if (args[0] >> 8)
             tcg_out32(s, args[0]);
 #else
-        if (args[0] >> 8)
-            tcg_out_ld32_12(s, COND_AL, 0, 15, 0);
-        else
-            tcg_out_dat_imm(s, COND_AL, ARITH_MOV, 0, 0, args[0]);
-        tcg_out_goto(s, COND_AL, (tcg_target_ulong) tb_ret_addr);
-        if (args[0] >> 8)
-            tcg_out32(s, args[0]);
+        {
+            uint8_t *ld_ptr = s->code_ptr;
+            if (args[0] >> 8)
+                tcg_out_ld32_12(s, COND_AL, 0, 15, 0);
+            else
+                tcg_out_dat_imm(s, COND_AL, ARITH_MOV, 0, 0, args[0]);
+            tcg_out_goto(s, COND_AL, (tcg_target_ulong) tb_ret_addr);
+            if (args[0] >> 8) {
+                *ld_ptr = (uint8_t) (s->code_ptr - ld_ptr) - 8;
+                tcg_out32(s, args[0]);
+            }
+        }
 #endif
         break;
     case INDEX_op_goto_tb:
         if (s->tb_jmp_offset) {
             /* Direct jump method */
-#if 1
+#if defined(USE_DIRECT_JUMP)
             s->tb_jmp_offset[args[0]] = s->code_ptr - s->code_buf;
             tcg_out_b(s, COND_AL, 8);
 #else