[Qemu-devel] [PATCH 11/16] tcg: Add qemu_ld_st_i32/64

[Richard Henderson]

Step two in the transition, adding the new ldst opcodes.  Keep the old
opcodes around until all backends support the new opcodes.

Signed-off-by: Richard Henderson <address@hidden>
---
 tcg/README               |  43 ++++-----
 tcg/aarch64/tcg-target.h |   2 +
 tcg/arm/tcg-target.h     |   2 +
 tcg/hppa/tcg-target.h    |   2 +
 tcg/i386/tcg-target.h    |   2 +
 tcg/ia64/tcg-target.h    |   2 +
 tcg/mips/tcg-target.h    |   2 +
 tcg/ppc/tcg-target.h     |   2 +
 tcg/ppc64/tcg-target.h   |   2 +
 tcg/s390/tcg-target.h    |   2 +
 tcg/sparc/tcg-target.h   |   2 +
 tcg/tcg-op.h             | 239 ++++++++++++-----------------------------------
 tcg/tcg-opc.h            |  96 ++++++++++++-------
 tcg/tcg.c                | 209 +++++++++++++++++++++++++++++++++++++++++
 tcg/tci/tcg-target.h     |   2 +
 15 files changed, 370 insertions(+), 239 deletions(-)

diff --git a/tcg/README b/tcg/README
index 063aeb9..f178212 100644
--- a/tcg/README
+++ b/tcg/README
@@ -412,30 +412,25 @@ current TB was linked to this TB. Otherwise execute the 
next
 instructions. Only indices 0 and 1 are valid and tcg_gen_goto_tb may be issued
 at most once with each slot index per TB.
 
-* qemu_ld8u t0, t1, flags
-qemu_ld8s t0, t1, flags
-qemu_ld16u t0, t1, flags
-qemu_ld16s t0, t1, flags
-qemu_ld32 t0, t1, flags
-qemu_ld32u t0, t1, flags
-qemu_ld32s t0, t1, flags
-qemu_ld64 t0, t1, flags
-
-Load data at the QEMU CPU address t1 into t0. t1 has the QEMU CPU address
-type. 'flags' contains the QEMU memory index (selects user or kernel access)
-for example.
-
-Note that "qemu_ld32" implies a 32-bit result, while "qemu_ld32u" and
-"qemu_ld32s" imply a 64-bit result appropriately extended from 32 bits.
-
-* qemu_st8 t0, t1, flags
-qemu_st16 t0, t1, flags
-qemu_st32 t0, t1, flags
-qemu_st64 t0, t1, flags
-
-Store the data t0 at the QEMU CPU Address t1. t1 has the QEMU CPU
-address type. 'flags' contains the QEMU memory index (selects user or
-kernel access) for example.
+* qemu_ld_i32/i64 t0, t1, flags, memidx
+* qemu_st_i32/i64 t0, t1, flags, memidx
+
+Load data at the guest address t1 into t0, or store data in t0 at guest
+address t1.  The _i32/_i64 size applies to the size of the input/output
+register t0 only.  The address t1 is always sized according to the guest,
+and the width of the memory operation is controlled by flags.
+
+Both t0 and t1 may be split into little-endian ordered pairs of registers
+if dealing with 64-bit quantities on a 32-bit host.
+
+The memidx selects the qemu tlb index to use (e.g. user or kernel access).
+The flags are the TCGMemOp bits, selecting the sign, width, and endianness
+of the memory access.
+
+For a 32-bit host, qemu_ld/st_i64 is guaranteed to only be used with a
+64-bit memory access specified in flags.
+
+*********
 
 Note 1: Some shortcuts are defined when the last operand is known to be
 a constant (e.g. addi for add, movi for mov).
diff --git a/tcg/aarch64/tcg-target.h b/tcg/aarch64/tcg-target.h
index e80bf78..3723320 100644
--- a/tcg/aarch64/tcg-target.h
+++ b/tcg/aarch64/tcg-target.h
@@ -99,6 +99,8 @@ typedef enum {
 #define TCG_TARGET_HAS_muluh_i64        1
 #define TCG_TARGET_HAS_mulsh_i64        1
 
+#define TCG_TARGET_HAS_new_ldst         0
+
 static inline void flush_icache_range(uintptr_t start, uintptr_t stop)
 {
     __builtin___clear_cache((char *)start, (char *)stop);
diff --git a/tcg/arm/tcg-target.h b/tcg/arm/tcg-target.h
index 01b9de0..a8311c2 100644
--- a/tcg/arm/tcg-target.h
+++ b/tcg/arm/tcg-target.h
@@ -86,6 +86,8 @@ extern bool use_idiv_instructions;
 #define TCG_TARGET_HAS_div_i32          use_idiv_instructions
 #define TCG_TARGET_HAS_rem_i32          0
 
+#define TCG_TARGET_HAS_new_ldst         0
+
 extern bool tcg_target_deposit_valid(int ofs, int len);
 #define TCG_TARGET_deposit_i32_valid  tcg_target_deposit_valid
 
diff --git a/tcg/hppa/tcg-target.h b/tcg/hppa/tcg-target.h
index dbb1cc8..bd409a3 100644
--- a/tcg/hppa/tcg-target.h
+++ b/tcg/hppa/tcg-target.h
@@ -105,6 +105,8 @@ typedef enum {
 #define TCG_TARGET_HAS_ext8u_i32        0 /* and rd, rs, 0xff */
 #define TCG_TARGET_HAS_ext16u_i32       0 /* and rd, rs, 0xffff */
 
+#define TCG_TARGET_HAS_new_ldst         0
+
 #define TCG_AREG0 TCG_REG_R17
 
 
diff --git a/tcg/i386/tcg-target.h b/tcg/i386/tcg-target.h
index 487a092..47fdb81 100644
--- a/tcg/i386/tcg-target.h
+++ b/tcg/i386/tcg-target.h
@@ -131,6 +131,8 @@ typedef enum {
 #define TCG_TARGET_HAS_mulsh_i64        0
 #endif
 
+#define TCG_TARGET_HAS_new_ldst         0
+
 #define TCG_TARGET_deposit_i32_valid(ofs, len) \
     (((ofs) == 0 && (len) == 8) || ((ofs) == 8 && (len) == 8) || \
      ((ofs) == 0 && (len) == 16))
diff --git a/tcg/ia64/tcg-target.h b/tcg/ia64/tcg-target.h
index 8e0ce35..9fe3bf6 100644
--- a/tcg/ia64/tcg-target.h
+++ b/tcg/ia64/tcg-target.h
@@ -153,6 +153,8 @@ typedef enum {
 #define TCG_TARGET_HAS_mulsh_i32        0
 #define TCG_TARGET_HAS_mulsh_i64        0
 
+#define TCG_TARGET_HAS_new_ldst         0
+
 #define TCG_TARGET_deposit_i32_valid(ofs, len) ((len) <= 16)
 #define TCG_TARGET_deposit_i64_valid(ofs, len) ((len) <= 16)
 
diff --git a/tcg/mips/tcg-target.h b/tcg/mips/tcg-target.h
index 2eb266f..8f85d49 100644
--- a/tcg/mips/tcg-target.h
+++ b/tcg/mips/tcg-target.h
@@ -123,6 +123,8 @@ extern bool use_mips32r2_instructions;
 #define TCG_TARGET_HAS_ext16s_i32       use_mips32r2_instructions
 #define TCG_TARGET_HAS_rot_i32          use_mips32r2_instructions
 
+#define TCG_TARGET_HAS_new_ldst         0
+
 /* optional instructions automatically implemented */
 #define TCG_TARGET_HAS_neg_i32          0 /* sub  rd, zero, rt   */
 #define TCG_TARGET_HAS_ext8u_i32        0 /* andi rt, rs, 0xff   */
diff --git a/tcg/ppc/tcg-target.h b/tcg/ppc/tcg-target.h
index 758e9b6..2662689 100644
--- a/tcg/ppc/tcg-target.h
+++ b/tcg/ppc/tcg-target.h
@@ -101,6 +101,8 @@ typedef enum {
 #define TCG_TARGET_HAS_muluh_i32        0
 #define TCG_TARGET_HAS_mulsh_i32        0
 
+#define TCG_TARGET_HAS_new_ldst         0
+
 #define TCG_AREG0 TCG_REG_R27
 
 #define tcg_qemu_tb_exec(env, tb_ptr) \
diff --git a/tcg/ppc64/tcg-target.h b/tcg/ppc64/tcg-target.h
index 8f490e1..0b75170 100644
--- a/tcg/ppc64/tcg-target.h
+++ b/tcg/ppc64/tcg-target.h
@@ -125,6 +125,8 @@ typedef enum {
 #define TCG_TARGET_HAS_muluh_i64        1
 #define TCG_TARGET_HAS_mulsh_i64        1
 
+#define TCG_TARGET_HAS_new_ldst         0
+
 #define TCG_AREG0 TCG_REG_R27
 
 #define TCG_TARGET_EXTEND_ARGS 1
diff --git a/tcg/s390/tcg-target.h b/tcg/s390/tcg-target.h
index eb691b3..a57aad5 100644
--- a/tcg/s390/tcg-target.h
+++ b/tcg/s390/tcg-target.h
@@ -100,6 +100,8 @@ typedef enum TCGReg {
 #define TCG_TARGET_HAS_muluh_i64        0
 #define TCG_TARGET_HAS_mulsh_i64        0
 
+#define TCG_TARGET_HAS_new_ldst         0
+
 extern bool tcg_target_deposit_valid(int ofs, int len);
 #define TCG_TARGET_deposit_i32_valid  tcg_target_deposit_valid
 #define TCG_TARGET_deposit_i64_valid  tcg_target_deposit_valid
diff --git a/tcg/sparc/tcg-target.h b/tcg/sparc/tcg-target.h
index 3b65c8e..2bee75a 100644
--- a/tcg/sparc/tcg-target.h
+++ b/tcg/sparc/tcg-target.h
@@ -149,6 +149,8 @@ typedef enum {
 #define TCG_TARGET_HAS_mulsh_i64        0
 #endif
 
+#define TCG_TARGET_HAS_new_ldst         0
+
 #define TCG_AREG0 TCG_REG_I0
 
 static inline void flush_icache_range(uintptr_t start, uintptr_t stop)
diff --git a/tcg/tcg-op.h b/tcg/tcg-op.h
index bb30a7c..7eabf22 100644
--- a/tcg/tcg-op.h
+++ b/tcg/tcg-op.h
@@ -137,24 +137,6 @@ static inline void tcg_gen_ldst_op_i64(TCGOpcode opc, 
TCGv_i64 val,
     *tcg_ctx.gen_opparam_ptr++ = offset;
 }
 
-static inline void tcg_gen_qemu_ldst_op_i64_i32(TCGOpcode opc, TCGv_i64 val,
-                                                TCGv_i32 addr, TCGArg 
mem_index)
-{
-    *tcg_ctx.gen_opc_ptr++ = opc;
-    *tcg_ctx.gen_opparam_ptr++ = GET_TCGV_I64(val);
-    *tcg_ctx.gen_opparam_ptr++ = GET_TCGV_I32(addr);
-    *tcg_ctx.gen_opparam_ptr++ = mem_index;
-}
-
-static inline void tcg_gen_qemu_ldst_op_i64_i64(TCGOpcode opc, TCGv_i64 val,
-                                                TCGv_i64 addr, TCGArg 
mem_index)
-{
-    *tcg_ctx.gen_opc_ptr++ = opc;
-    *tcg_ctx.gen_opparam_ptr++ = GET_TCGV_I64(val);
-    *tcg_ctx.gen_opparam_ptr++ = GET_TCGV_I64(addr);
-    *tcg_ctx.gen_opparam_ptr++ = mem_index;
-}
-
 static inline void tcg_gen_op4_i32(TCGOpcode opc, TCGv_i32 arg1, TCGv_i32 arg2,
                                    TCGv_i32 arg3, TCGv_i32 arg4)
 {
@@ -361,6 +343,21 @@ static inline void tcg_gen_op6ii_i64(TCGOpcode opc, 
TCGv_i64 arg1,
     *tcg_ctx.gen_opparam_ptr++ = arg6;
 }
 
+static inline void tcg_add_param_i32(TCGv_i32 val)
+{
+    *tcg_ctx.gen_opparam_ptr++ = GET_TCGV_I32(val);
+}
+
+static inline void tcg_add_param_i64(TCGv_i64 val)
+{
+#if TCG_TARGET_REG_BITS == 32
+    *tcg_ctx.gen_opparam_ptr++ = GET_TCGV_I32(TCGV_LOW(val));
+    *tcg_ctx.gen_opparam_ptr++ = GET_TCGV_I32(TCGV_HIGH(val));
+#else
+    *tcg_ctx.gen_opparam_ptr++ = GET_TCGV_I64(val);
+#endif
+}
+
 static inline void gen_set_label(int n)
 {
     tcg_gen_op1i(INDEX_op_set_label, n);
@@ -2600,11 +2597,12 @@ static inline void tcg_gen_muls2_i64(TCGv_i64 rl, 
TCGv_i64 rh,
 #define tcg_global_mem_new tcg_global_mem_new_i32
 #define tcg_temp_local_new() tcg_temp_local_new_i32()
 #define tcg_temp_free tcg_temp_free_i32
-#define tcg_gen_qemu_ldst_op tcg_gen_op3i_i32
-#define tcg_gen_qemu_ldst_op_i64 tcg_gen_qemu_ldst_op_i64_i32
 #define TCGV_UNUSED(x) TCGV_UNUSED_I32(x)
 #define TCGV_IS_UNUSED(x) TCGV_IS_UNUSED_I32(x)
 #define TCGV_EQUAL(a, b) TCGV_EQUAL_I32(a, b)
+#define tcg_add_param_tl tcg_add_param_i32
+#define tcg_gen_qemu_ld_tl tcg_gen_qemu_ld_i32
+#define tcg_gen_qemu_st_tl tcg_gen_qemu_st_i32
 #else
 #define TCGv TCGv_i64
 #define tcg_temp_new() tcg_temp_new_i64()
@@ -2612,11 +2610,12 @@ static inline void tcg_gen_muls2_i64(TCGv_i64 rl, 
TCGv_i64 rh,
 #define tcg_global_mem_new tcg_global_mem_new_i64
 #define tcg_temp_local_new() tcg_temp_local_new_i64()
 #define tcg_temp_free tcg_temp_free_i64
-#define tcg_gen_qemu_ldst_op tcg_gen_op3i_i64
-#define tcg_gen_qemu_ldst_op_i64 tcg_gen_qemu_ldst_op_i64_i64
 #define TCGV_UNUSED(x) TCGV_UNUSED_I64(x)
 #define TCGV_IS_UNUSED(x) TCGV_IS_UNUSED_I64(x)
 #define TCGV_EQUAL(a, b) TCGV_EQUAL_I64(a, b)
+#define tcg_add_param_tl tcg_add_param_i64
+#define tcg_gen_qemu_ld_tl tcg_gen_qemu_ld_i64
+#define tcg_gen_qemu_st_tl tcg_gen_qemu_st_i64
 #endif
 
 /* debug info: write the PC of the corresponding QEMU CPU instruction */
@@ -2648,197 +2647,67 @@ static inline void tcg_gen_goto_tb(unsigned idx)
     tcg_gen_op1i(INDEX_op_goto_tb, idx);
 }
 
-#if TCG_TARGET_REG_BITS == 32
-static inline void tcg_gen_qemu_ld8u(TCGv ret, TCGv addr, int mem_index)
-{
-#if TARGET_LONG_BITS == 32
-    tcg_gen_op3i_i32(INDEX_op_qemu_ld8u, ret, addr, mem_index);
-#else
-    tcg_gen_op4i_i32(INDEX_op_qemu_ld8u, TCGV_LOW(ret), TCGV_LOW(addr),
-                     TCGV_HIGH(addr), mem_index);
-    tcg_gen_movi_i32(TCGV_HIGH(ret), 0);
-#endif
-}
-
-static inline void tcg_gen_qemu_ld8s(TCGv ret, TCGv addr, int mem_index)
-{
-#if TARGET_LONG_BITS == 32
-    tcg_gen_op3i_i32(INDEX_op_qemu_ld8s, ret, addr, mem_index);
-#else
-    tcg_gen_op4i_i32(INDEX_op_qemu_ld8s, TCGV_LOW(ret), TCGV_LOW(addr),
-                     TCGV_HIGH(addr), mem_index);
-    tcg_gen_sari_i32(TCGV_HIGH(ret), TCGV_LOW(ret), 31);
-#endif
-}
 
-static inline void tcg_gen_qemu_ld16u(TCGv ret, TCGv addr, int mem_index)
-{
-#if TARGET_LONG_BITS == 32
-    tcg_gen_op3i_i32(INDEX_op_qemu_ld16u, ret, addr, mem_index);
-#else
-    tcg_gen_op4i_i32(INDEX_op_qemu_ld16u, TCGV_LOW(ret), TCGV_LOW(addr),
-                     TCGV_HIGH(addr), mem_index);
-    tcg_gen_movi_i32(TCGV_HIGH(ret), 0);
-#endif
-}
-
-static inline void tcg_gen_qemu_ld16s(TCGv ret, TCGv addr, int mem_index)
-{
-#if TARGET_LONG_BITS == 32
-    tcg_gen_op3i_i32(INDEX_op_qemu_ld16s, ret, addr, mem_index);
-#else
-    tcg_gen_op4i_i32(INDEX_op_qemu_ld16s, TCGV_LOW(ret), TCGV_LOW(addr),
-                     TCGV_HIGH(addr), mem_index);
-    tcg_gen_sari_i32(TCGV_HIGH(ret), TCGV_LOW(ret), 31);
-#endif
-}
-
-static inline void tcg_gen_qemu_ld32u(TCGv ret, TCGv addr, int mem_index)
-{
-#if TARGET_LONG_BITS == 32
-    tcg_gen_op3i_i32(INDEX_op_qemu_ld32, ret, addr, mem_index);
-#else
-    tcg_gen_op4i_i32(INDEX_op_qemu_ld32, TCGV_LOW(ret), TCGV_LOW(addr),
-                     TCGV_HIGH(addr), mem_index);
-    tcg_gen_movi_i32(TCGV_HIGH(ret), 0);
-#endif
-}
-
-static inline void tcg_gen_qemu_ld32s(TCGv ret, TCGv addr, int mem_index)
-{
-#if TARGET_LONG_BITS == 32
-    tcg_gen_op3i_i32(INDEX_op_qemu_ld32, ret, addr, mem_index);
-#else
-    tcg_gen_op4i_i32(INDEX_op_qemu_ld32, TCGV_LOW(ret), TCGV_LOW(addr),
-                     TCGV_HIGH(addr), mem_index);
-    tcg_gen_sari_i32(TCGV_HIGH(ret), TCGV_LOW(ret), 31);
-#endif
-}
-
-static inline void tcg_gen_qemu_ld64(TCGv_i64 ret, TCGv addr, int mem_index)
-{
-#if TARGET_LONG_BITS == 32
-    tcg_gen_op4i_i32(INDEX_op_qemu_ld64, TCGV_LOW(ret), TCGV_HIGH(ret), addr, 
mem_index);
-#else
-    tcg_gen_op5i_i32(INDEX_op_qemu_ld64, TCGV_LOW(ret), TCGV_HIGH(ret),
-                     TCGV_LOW(addr), TCGV_HIGH(addr), mem_index);
-#endif
-}
-
-static inline void tcg_gen_qemu_st8(TCGv arg, TCGv addr, int mem_index)
-{
-#if TARGET_LONG_BITS == 32
-    tcg_gen_op3i_i32(INDEX_op_qemu_st8, arg, addr, mem_index);
-#else
-    tcg_gen_op4i_i32(INDEX_op_qemu_st8, TCGV_LOW(arg), TCGV_LOW(addr),
-                     TCGV_HIGH(addr), mem_index);
-#endif
-}
-
-static inline void tcg_gen_qemu_st16(TCGv arg, TCGv addr, int mem_index)
-{
-#if TARGET_LONG_BITS == 32
-    tcg_gen_op3i_i32(INDEX_op_qemu_st16, arg, addr, mem_index);
-#else
-    tcg_gen_op4i_i32(INDEX_op_qemu_st16, TCGV_LOW(arg), TCGV_LOW(addr),
-                     TCGV_HIGH(addr), mem_index);
-#endif
-}
-
-static inline void tcg_gen_qemu_st32(TCGv arg, TCGv addr, int mem_index)
-{
-#if TARGET_LONG_BITS == 32
-    tcg_gen_op3i_i32(INDEX_op_qemu_st32, arg, addr, mem_index);
-#else
-    tcg_gen_op4i_i32(INDEX_op_qemu_st32, TCGV_LOW(arg), TCGV_LOW(addr),
-                     TCGV_HIGH(addr), mem_index);
-#endif
-}
-
-static inline void tcg_gen_qemu_st64(TCGv_i64 arg, TCGv addr, int mem_index)
-{
-#if TARGET_LONG_BITS == 32
-    tcg_gen_op4i_i32(INDEX_op_qemu_st64, TCGV_LOW(arg), TCGV_HIGH(arg), addr,
-                     mem_index);
-#else
-    tcg_gen_op5i_i32(INDEX_op_qemu_st64, TCGV_LOW(arg), TCGV_HIGH(arg),
-                     TCGV_LOW(addr), TCGV_HIGH(addr), mem_index);
-#endif
-}
-
-#define tcg_gen_ld_ptr(R, A, O) tcg_gen_ld_i32(TCGV_PTR_TO_NAT(R), (A), (O))
-#define tcg_gen_discard_ptr(A) tcg_gen_discard_i32(TCGV_PTR_TO_NAT(A))
-
-#else /* TCG_TARGET_REG_BITS == 32 */
+void tcg_gen_qemu_ld_i32(TCGv_i32, TCGv, TCGArg, TCGMemOp);
+void tcg_gen_qemu_st_i32(TCGv_i32, TCGv, TCGArg, TCGMemOp);
+void tcg_gen_qemu_ld_i64(TCGv_i64, TCGv, TCGArg, TCGMemOp);
+void tcg_gen_qemu_st_i64(TCGv_i64, TCGv, TCGArg, TCGMemOp);
 
 static inline void tcg_gen_qemu_ld8u(TCGv ret, TCGv addr, int mem_index)
 {
-    tcg_gen_qemu_ldst_op(INDEX_op_qemu_ld8u, ret, addr, mem_index);
+    tcg_gen_qemu_ld_tl(ret, addr, mem_index, MO_UB);
 }
 
 static inline void tcg_gen_qemu_ld8s(TCGv ret, TCGv addr, int mem_index)
 {
-    tcg_gen_qemu_ldst_op(INDEX_op_qemu_ld8s, ret, addr, mem_index);
+    tcg_gen_qemu_ld_tl(ret, addr, mem_index, MO_SB);
 }
 
 static inline void tcg_gen_qemu_ld16u(TCGv ret, TCGv addr, int mem_index)
 {
-    tcg_gen_qemu_ldst_op(INDEX_op_qemu_ld16u, ret, addr, mem_index);
+    tcg_gen_qemu_ld_tl(ret, addr, mem_index, MO_TEUW);
 }
 
 static inline void tcg_gen_qemu_ld16s(TCGv ret, TCGv addr, int mem_index)
 {
-    tcg_gen_qemu_ldst_op(INDEX_op_qemu_ld16s, ret, addr, mem_index);
+    tcg_gen_qemu_ld_tl(ret, addr, mem_index, MO_TESW);
 }
 
 static inline void tcg_gen_qemu_ld32u(TCGv ret, TCGv addr, int mem_index)
 {
-#if TARGET_LONG_BITS == 32
-    tcg_gen_qemu_ldst_op(INDEX_op_qemu_ld32, ret, addr, mem_index);
-#else
-    tcg_gen_qemu_ldst_op(INDEX_op_qemu_ld32u, ret, addr, mem_index);
-#endif
+    tcg_gen_qemu_ld_tl(ret, addr, mem_index, MO_TEUL);
 }
 
 static inline void tcg_gen_qemu_ld32s(TCGv ret, TCGv addr, int mem_index)
 {
-#if TARGET_LONG_BITS == 32
-    tcg_gen_qemu_ldst_op(INDEX_op_qemu_ld32, ret, addr, mem_index);
-#else
-    tcg_gen_qemu_ldst_op(INDEX_op_qemu_ld32s, ret, addr, mem_index);
-#endif
+    tcg_gen_qemu_ld_tl(ret, addr, mem_index, MO_TESL);
 }
 
 static inline void tcg_gen_qemu_ld64(TCGv_i64 ret, TCGv addr, int mem_index)
 {
-    tcg_gen_qemu_ldst_op_i64(INDEX_op_qemu_ld64, ret, addr, mem_index);
+    tcg_gen_qemu_ld_i64(ret, addr, mem_index, MO_TEQ);
 }
 
 static inline void tcg_gen_qemu_st8(TCGv arg, TCGv addr, int mem_index)
 {
-    tcg_gen_qemu_ldst_op(INDEX_op_qemu_st8, arg, addr, mem_index);
+    tcg_gen_qemu_st_tl(arg, addr, mem_index, MO_UB);
 }
 
 static inline void tcg_gen_qemu_st16(TCGv arg, TCGv addr, int mem_index)
 {
-    tcg_gen_qemu_ldst_op(INDEX_op_qemu_st16, arg, addr, mem_index);
+    tcg_gen_qemu_st_tl(arg, addr, mem_index, MO_TEUW);
 }
 
 static inline void tcg_gen_qemu_st32(TCGv arg, TCGv addr, int mem_index)
 {
-    tcg_gen_qemu_ldst_op(INDEX_op_qemu_st32, arg, addr, mem_index);
+    tcg_gen_qemu_st_tl(arg, addr, mem_index, MO_TEUL);
 }
 
 static inline void tcg_gen_qemu_st64(TCGv_i64 arg, TCGv addr, int mem_index)
 {
-    tcg_gen_qemu_ldst_op_i64(INDEX_op_qemu_st64, arg, addr, mem_index);
+    tcg_gen_qemu_st_i64(arg, addr, mem_index, MO_TEQ);
 }
 
-#define tcg_gen_ld_ptr(R, A, O) tcg_gen_ld_i64(TCGV_PTR_TO_NAT(R), (A), (O))
-#define tcg_gen_discard_ptr(A) tcg_gen_discard_i64(TCGV_PTR_TO_NAT(A))
-
-#endif /* TCG_TARGET_REG_BITS != 32 */
-
 #if TARGET_LONG_BITS == 64
 #define tcg_gen_movi_tl tcg_gen_movi_i64
 #define tcg_gen_mov_tl tcg_gen_mov_i64
@@ -2997,17 +2866,25 @@ static inline void tcg_gen_qemu_st64(TCGv_i64 arg, TCGv 
addr, int mem_index)
 #endif
 
 #if TCG_TARGET_REG_BITS == 32
-#define tcg_gen_add_ptr(R, A, B) tcg_gen_add_i32(TCGV_PTR_TO_NAT(R), \
-                                               TCGV_PTR_TO_NAT(A), \
-                                               TCGV_PTR_TO_NAT(B))
-#define tcg_gen_addi_ptr(R, A, B) tcg_gen_addi_i32(TCGV_PTR_TO_NAT(R), \
-                                                 TCGV_PTR_TO_NAT(A), (B))
-#define tcg_gen_ext_i32_ptr(R, A) tcg_gen_mov_i32(TCGV_PTR_TO_NAT(R), (A))
-#else /* TCG_TARGET_REG_BITS == 32 */
-#define tcg_gen_add_ptr(R, A, B) tcg_gen_add_i64(TCGV_PTR_TO_NAT(R), \
-                                               TCGV_PTR_TO_NAT(A), \
-                                               TCGV_PTR_TO_NAT(B))
-#define tcg_gen_addi_ptr(R, A, B) tcg_gen_addi_i64(TCGV_PTR_TO_NAT(R),   \
-                                                 TCGV_PTR_TO_NAT(A), (B))
-#define tcg_gen_ext_i32_ptr(R, A) tcg_gen_ext_i32_i64(TCGV_PTR_TO_NAT(R), (A))
-#endif /* TCG_TARGET_REG_BITS != 32 */
+# define tcg_gen_ld_ptr(R, A, O) \
+    tcg_gen_ld_i32(TCGV_PTR_TO_NAT(R), (A), (O))
+# define tcg_gen_discard_ptr(A) \
+    tcg_gen_discard_i32(TCGV_PTR_TO_NAT(A))
+# define tcg_gen_add_ptr(R, A, B) \
+    tcg_gen_add_i32(TCGV_PTR_TO_NAT(R), TCGV_PTR_TO_NAT(A), TCGV_PTR_TO_NAT(B))
+# define tcg_gen_addi_ptr(R, A, B) \
+    tcg_gen_addi_i32(TCGV_PTR_TO_NAT(R), TCGV_PTR_TO_NAT(A), (B))
+# define tcg_gen_ext_i32_ptr(R, A) \
+    tcg_gen_mov_i32(TCGV_PTR_TO_NAT(R), (A))
+#else
+# define tcg_gen_ld_ptr(R, A, O) \
+    tcg_gen_ld_i64(TCGV_PTR_TO_NAT(R), (A), (O))
+# define tcg_gen_discard_ptr(A) \
+    tcg_gen_discard_i64(TCGV_PTR_TO_NAT(A))
+# define tcg_gen_add_ptr(R, A, B) \
+    tcg_gen_add_i64(TCGV_PTR_TO_NAT(R), TCGV_PTR_TO_NAT(A), TCGV_PTR_TO_NAT(B))
+# define tcg_gen_addi_ptr(R, A, B) \
+    tcg_gen_addi_i64(TCGV_PTR_TO_NAT(R), TCGV_PTR_TO_NAT(A), (B))
+# define tcg_gen_ext_i32_ptr(R, A) \
+    tcg_gen_ext_i32_i64(TCGV_PTR_TO_NAT(R), (A))
+#endif /* TCG_TARGET_REG_BITS == 32 */
diff --git a/tcg/tcg-opc.h b/tcg/tcg-opc.h
index a75c29d..d71707d 100644
--- a/tcg/tcg-opc.h
+++ b/tcg/tcg-opc.h
@@ -180,79 +180,107 @@ DEF(debug_insn_start, 0, 0, 1, TCG_OPF_NOT_PRESENT)
 #endif
 DEF(exit_tb, 0, 0, 1, TCG_OPF_BB_END)
 DEF(goto_tb, 0, 0, 1, TCG_OPF_BB_END)
-/* Note: even if TARGET_LONG_BITS is not defined, the INDEX_op
-   constants must be defined */
+
+#define IMPL_NEW_LDST \
+    (TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS \
+     | IMPL(TCG_TARGET_HAS_new_ldst))
+
+#if TARGET_LONG_BITS <= TCG_TARGET_REG_BITS
+DEF(qemu_ld_i32, 1, 1, 2, IMPL_NEW_LDST)
+DEF(qemu_st_i32, 0, 2, 2, IMPL_NEW_LDST)
+# if TCG_TARGET_REG_BITS == 64
+DEF(qemu_ld_i64, 1, 1, 2, IMPL_NEW_LDST | TCG_OPF_64BIT)
+DEF(qemu_st_i64, 0, 2, 2, IMPL_NEW_LDST | TCG_OPF_64BIT)
+# else
+DEF(qemu_ld_i64, 2, 1, 2, IMPL_NEW_LDST | TCG_OPF_64BIT)
+DEF(qemu_st_i64, 0, 3, 2, IMPL_NEW_LDST | TCG_OPF_64BIT)
+# endif
+#else
+DEF(qemu_ld_i32, 1, 2, 2, IMPL_NEW_LDST)
+DEF(qemu_st_i32, 0, 3, 2, IMPL_NEW_LDST)
+DEF(qemu_ld_i64, 2, 2, 2, IMPL_NEW_LDST | TCG_OPF_64BIT)
+DEF(qemu_st_i64, 0, 4, 2, IMPL_NEW_LDST | TCG_OPF_64BIT)
+#endif
+
+#undef IMPL_NEW_LDST
+
+#define IMPL_OLD_LDST \
+    (TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS \
+     | IMPL(!TCG_TARGET_HAS_new_ldst))
+
 #if TCG_TARGET_REG_BITS == 32
 #if TARGET_LONG_BITS == 32
-DEF(qemu_ld8u, 1, 1, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
+DEF(qemu_ld8u, 1, 1, 1, IMPL_OLD_LDST)
 #else
-DEF(qemu_ld8u, 1, 2, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
+DEF(qemu_ld8u, 1, 2, 1, IMPL_OLD_LDST)
 #endif
 #if TARGET_LONG_BITS == 32
-DEF(qemu_ld8s, 1, 1, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
+DEF(qemu_ld8s, 1, 1, 1, IMPL_OLD_LDST)
 #else
-DEF(qemu_ld8s, 1, 2, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
+DEF(qemu_ld8s, 1, 2, 1, IMPL_OLD_LDST)
 #endif
 #if TARGET_LONG_BITS == 32
-DEF(qemu_ld16u, 1, 1, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
+DEF(qemu_ld16u, 1, 1, 1, IMPL_OLD_LDST)
 #else
-DEF(qemu_ld16u, 1, 2, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
+DEF(qemu_ld16u, 1, 2, 1, IMPL_OLD_LDST)
 #endif
 #if TARGET_LONG_BITS == 32
-DEF(qemu_ld16s, 1, 1, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
+DEF(qemu_ld16s, 1, 1, 1, IMPL_OLD_LDST)
 #else
-DEF(qemu_ld16s, 1, 2, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
+DEF(qemu_ld16s, 1, 2, 1, IMPL_OLD_LDST)
 #endif
 #if TARGET_LONG_BITS == 32
-DEF(qemu_ld32, 1, 1, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
+DEF(qemu_ld32, 1, 1, 1, IMPL_OLD_LDST)
 #else
-DEF(qemu_ld32, 1, 2, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
+DEF(qemu_ld32, 1, 2, 1, IMPL_OLD_LDST)
 #endif
 #if TARGET_LONG_BITS == 32
-DEF(qemu_ld64, 2, 1, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
+DEF(qemu_ld64, 2, 1, 1, IMPL_OLD_LDST | TCG_OPF_64BIT)
 #else
-DEF(qemu_ld64, 2, 2, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
+DEF(qemu_ld64, 2, 2, 1, IMPL_OLD_LDST | TCG_OPF_64BIT)
 #endif
 
 #if TARGET_LONG_BITS == 32
-DEF(qemu_st8, 0, 2, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
+DEF(qemu_st8, 0, 2, 1, IMPL_OLD_LDST)
 #else
-DEF(qemu_st8, 0, 3, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
+DEF(qemu_st8, 0, 3, 1, IMPL_OLD_LDST)
 #endif
 #if TARGET_LONG_BITS == 32
-DEF(qemu_st16, 0, 2, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
+DEF(qemu_st16, 0, 2, 1, IMPL_OLD_LDST)
 #else
-DEF(qemu_st16, 0, 3, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
+DEF(qemu_st16, 0, 3, 1, IMPL_OLD_LDST)
 #endif
 #if TARGET_LONG_BITS == 32
-DEF(qemu_st32, 0, 2, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
+DEF(qemu_st32, 0, 2, 1, IMPL_OLD_LDST)
 #else
-DEF(qemu_st32, 0, 3, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
+DEF(qemu_st32, 0, 3, 1, IMPL_OLD_LDST)
 #endif
 #if TARGET_LONG_BITS == 32
-DEF(qemu_st64, 0, 3, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
+DEF(qemu_st64, 0, 3, 1, IMPL_OLD_LDST | TCG_OPF_64BIT)
 #else
-DEF(qemu_st64, 0, 4, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
+DEF(qemu_st64, 0, 4, 1, IMPL_OLD_LDST | TCG_OPF_64BIT)
 #endif
 
 #else /* TCG_TARGET_REG_BITS == 32 */
 
-DEF(qemu_ld8u, 1, 1, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
-DEF(qemu_ld8s, 1, 1, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
-DEF(qemu_ld16u, 1, 1, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
-DEF(qemu_ld16s, 1, 1, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
-DEF(qemu_ld32, 1, 1, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
-DEF(qemu_ld32u, 1, 1, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
-DEF(qemu_ld32s, 1, 1, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
-DEF(qemu_ld64, 1, 1, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
+DEF(qemu_ld8u, 1, 1, 1, IMPL_OLD_LDST | TCG_OPF_64BIT)
+DEF(qemu_ld8s, 1, 1, 1, IMPL_OLD_LDST | TCG_OPF_64BIT)
+DEF(qemu_ld16u, 1, 1, 1, IMPL_OLD_LDST | TCG_OPF_64BIT)
+DEF(qemu_ld16s, 1, 1, 1, IMPL_OLD_LDST | TCG_OPF_64BIT)
+DEF(qemu_ld32, 1, 1, 1, IMPL_OLD_LDST | TCG_OPF_64BIT)
+DEF(qemu_ld32u, 1, 1, 1, IMPL_OLD_LDST | TCG_OPF_64BIT)
+DEF(qemu_ld32s, 1, 1, 1, IMPL_OLD_LDST | TCG_OPF_64BIT)
+DEF(qemu_ld64, 1, 1, 1, IMPL_OLD_LDST | TCG_OPF_64BIT)
 
-DEF(qemu_st8, 0, 2, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
-DEF(qemu_st16, 0, 2, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
-DEF(qemu_st32, 0, 2, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
-DEF(qemu_st64, 0, 2, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS)
+DEF(qemu_st8, 0, 2, 1, IMPL_OLD_LDST | TCG_OPF_64BIT)
+DEF(qemu_st16, 0, 2, 1, IMPL_OLD_LDST | TCG_OPF_64BIT)
+DEF(qemu_st32, 0, 2, 1, IMPL_OLD_LDST | TCG_OPF_64BIT)
+DEF(qemu_st64, 0, 2, 1, IMPL_OLD_LDST | TCG_OPF_64BIT)
 
 #endif /* TCG_TARGET_REG_BITS != 32 */
 
+#undef IMPL_OLD_LDST
+
 #undef IMPL
 #undef IMPL64
 #undef DEF
diff --git a/tcg/tcg.c b/tcg/tcg.c
index f11b231..99a27e5 100644
--- a/tcg/tcg.c
+++ b/tcg/tcg.c
@@ -795,6 +795,188 @@ void tcg_gen_shifti_i64(TCGv_i64 ret, TCGv_i64 arg1,
 }
 #endif
 
+static inline TCGMemOp tcg_canonicalize_memop(TCGMemOp op, bool is64, bool st)
+{
+    switch (op & MO_SIZE) {
+    case MO_8:
+        op &= ~MO_BSWAP;
+        break;
+    case MO_16:
+        break;
+    case MO_32:
+        if (!is64) {
+            op &= ~MO_SIGN;
+        }
+        break;
+    case MO_64:
+        if (!is64) {
+            tcg_abort();
+        }
+        break;
+    }
+    if (st) {
+        op &= ~MO_SIGN;
+    }
+    return op;
+}
+
+static const TCGOpcode old_ld_opc[8] = {
+    [MO_UB] = INDEX_op_qemu_ld8u,
+    [MO_SB] = INDEX_op_qemu_ld8s,
+    [MO_UW] = INDEX_op_qemu_ld16u,
+    [MO_SW] = INDEX_op_qemu_ld16s,
+#if TCG_TARGET_REG_BITS == 32
+    [MO_UL] = INDEX_op_qemu_ld32,
+    [MO_SL] = INDEX_op_qemu_ld32,
+#else
+    [MO_UL] = INDEX_op_qemu_ld32u,
+    [MO_SL] = INDEX_op_qemu_ld32s,
+#endif
+    [MO_Q]  = INDEX_op_qemu_ld64,
+};
+
+static const TCGOpcode old_st_opc[4] = {
+    [MO_UB] = INDEX_op_qemu_st8,
+    [MO_UW] = INDEX_op_qemu_st16,
+    [MO_UL] = INDEX_op_qemu_st32,
+    [MO_Q]  = INDEX_op_qemu_st64,
+};
+
+void tcg_gen_qemu_ld_i32(TCGv_i32 val, TCGv addr, TCGArg idx, TCGMemOp memop)
+{
+    memop = tcg_canonicalize_memop(memop, 0, 0);
+
+    if (TCG_TARGET_HAS_new_ldst) {
+        *tcg_ctx.gen_opc_ptr++ = INDEX_op_qemu_ld_i32;
+        tcg_add_param_i32(val);
+        tcg_add_param_tl(addr);
+        *tcg_ctx.gen_opparam_ptr++ = memop;
+        *tcg_ctx.gen_opparam_ptr++ = idx;
+        return;
+    }
+
+    /* The old opcodes only support target-endian memory operations.  */
+    assert((memop & MO_BSWAP) == MO_TE || (memop & MO_SIZE) == MO_8);
+    assert(old_ld_opc[memop & MO_SSIZE] != 0);
+
+    if (TCG_TARGET_REG_BITS == 32) {
+        *tcg_ctx.gen_opc_ptr++ = old_ld_opc[memop & MO_SSIZE];
+        tcg_add_param_i32(val);
+        tcg_add_param_tl(addr);
+        *tcg_ctx.gen_opparam_ptr++ = idx;
+    } else {
+        TCGv_i64 val64 = tcg_temp_new_i64();
+
+        *tcg_ctx.gen_opc_ptr++ = old_ld_opc[memop & MO_SSIZE];
+        tcg_add_param_i64(val64);
+        tcg_add_param_tl(addr);
+        *tcg_ctx.gen_opparam_ptr++ = idx;
+
+        tcg_gen_trunc_i64_i32(val, val64);
+        tcg_temp_free_i64(val64);
+    }
+}
+
+void tcg_gen_qemu_st_i32(TCGv_i32 val, TCGv addr, TCGArg idx, TCGMemOp memop)
+{
+    memop = tcg_canonicalize_memop(memop, 0, 1);
+
+    if (TCG_TARGET_HAS_new_ldst) {
+        *tcg_ctx.gen_opc_ptr++ = INDEX_op_qemu_st_i32;
+        tcg_add_param_i32(val);
+        tcg_add_param_tl(addr);
+        *tcg_ctx.gen_opparam_ptr++ = memop;
+        *tcg_ctx.gen_opparam_ptr++ = idx;
+        return;
+    }
+
+    /* The old opcodes only support target-endian memory operations.  */
+    assert((memop & MO_BSWAP) == MO_TE || (memop & MO_SIZE) == MO_8);
+    assert(old_st_opc[memop & MO_SIZE] != 0);
+
+    if (TCG_TARGET_REG_BITS == 32) {
+        *tcg_ctx.gen_opc_ptr++ = old_st_opc[memop & MO_SIZE];
+        tcg_add_param_i32(val);
+        tcg_add_param_tl(addr);
+        *tcg_ctx.gen_opparam_ptr++ = idx;
+    } else {
+        TCGv_i64 val64 = tcg_temp_new_i64();
+
+        tcg_gen_extu_i32_i64(val64, val);
+
+        *tcg_ctx.gen_opc_ptr++ = old_st_opc[memop & MO_SIZE];
+        tcg_add_param_i64(val64);
+        tcg_add_param_tl(addr);
+        *tcg_ctx.gen_opparam_ptr++ = idx;
+
+        tcg_temp_free_i64(val64);
+    }
+}
+
+void tcg_gen_qemu_ld_i64(TCGv_i64 val, TCGv addr, TCGArg idx, TCGMemOp memop)
+{
+    memop = tcg_canonicalize_memop(memop, 1, 0);
+
+#if TCG_TARGET_REG_BITS == 32
+    if ((memop & MO_SIZE) < MO_64) {
+        tcg_gen_qemu_ld_i32(TCGV_LOW(val), addr, idx, memop);
+        if (memop & MO_SIGN) {
+            tcg_gen_sari_i32(TCGV_HIGH(val), TCGV_LOW(val), 31);
+        } else {
+            tcg_gen_movi_i32(TCGV_HIGH(val), 0);
+        }
+        return;
+    }
+#endif
+
+    if (TCG_TARGET_HAS_new_ldst) {
+        *tcg_ctx.gen_opc_ptr++ = INDEX_op_qemu_ld_i64;
+        tcg_add_param_i64(val);
+        tcg_add_param_tl(addr);
+        *tcg_ctx.gen_opparam_ptr++ = memop;
+        *tcg_ctx.gen_opparam_ptr++ = idx;
+        return;
+    }
+
+    /* The old opcodes only support target-endian memory operations.  */
+    assert((memop & MO_BSWAP) == MO_TE || (memop & MO_SIZE) == MO_8);
+    assert(old_ld_opc[memop & MO_SSIZE] != 0);
+
+    *tcg_ctx.gen_opc_ptr++ = old_ld_opc[memop & MO_SSIZE];
+    tcg_add_param_i64(val);
+    tcg_add_param_tl(addr);
+    *tcg_ctx.gen_opparam_ptr++ = idx;
+}
+
+void tcg_gen_qemu_st_i64(TCGv_i64 val, TCGv addr, TCGArg idx, TCGMemOp memop)
+{
+    memop = tcg_canonicalize_memop(memop, 1, 1);
+
+#if TCG_TARGET_REG_BITS == 32
+    if ((memop & MO_SIZE) < MO_64) {
+        tcg_gen_qemu_st_i32(TCGV_LOW(val), addr, idx, memop);
+        return;
+    }
+#endif
+
+    if (TCG_TARGET_HAS_new_ldst) {
+        *tcg_ctx.gen_opc_ptr++ = INDEX_op_qemu_st_i64;
+        tcg_add_param_i64(val);
+        tcg_add_param_tl(addr);
+        *tcg_ctx.gen_opparam_ptr++ = memop;
+        *tcg_ctx.gen_opparam_ptr++ = idx;
+        return;
+    }
+
+    /* The old opcodes only support target-endian memory operations.  */
+    assert((memop & MO_BSWAP) == MO_TE || (memop & MO_SIZE) == MO_8);
+    assert(old_st_opc[memop & MO_SIZE] != 0);
+
+    *tcg_ctx.gen_opc_ptr++ = old_st_opc[memop & MO_SIZE];
+    tcg_add_param_i64(val);
+    tcg_add_param_tl(addr);
+    *tcg_ctx.gen_opparam_ptr++ = idx;
+}
 
 static void tcg_reg_alloc_start(TCGContext *s)
 {
@@ -910,6 +1092,22 @@ static const char * const cond_name[] =
     [TCG_COND_GTU] = "gtu"
 };
 
+static const char * const ldst_name[] =
+{
+    [MO_UB]   = "ub",
+    [MO_SB]   = "sb",
+    [MO_LEUW] = "leuw",
+    [MO_LESW] = "lesw",
+    [MO_LEUL] = "leul",
+    [MO_LESL] = "lesl",
+    [MO_LEQ]  = "leq",
+    [MO_BEUW] = "beuw",
+    [MO_BESW] = "besw",
+    [MO_BEUL] = "beul",
+    [MO_BESL] = "besl",
+    [MO_BEQ]  = "beq",
+};
+
 void tcg_dump_ops(TCGContext *s)
 {
     const uint16_t *opc_ptr;
@@ -1038,6 +1236,17 @@ void tcg_dump_ops(TCGContext *s)
                 }
                 i = 1;
                 break;
+            case INDEX_op_qemu_ld_i32:
+            case INDEX_op_qemu_st_i32:
+            case INDEX_op_qemu_ld_i64:
+            case INDEX_op_qemu_st_i64:
+                if (args[k] < ARRAY_SIZE(ldst_name) && ldst_name[args[k]]) {
+                    qemu_log(",%s", ldst_name[args[k++]]);
+                } else {
+                    qemu_log(",$0x%" TCG_PRIlx, args[k++]);
+                }
+                i = 1;
+                break;
             default:
                 i = 0;
                 break;
diff --git a/tcg/tci/tcg-target.h b/tcg/tci/tcg-target.h
index 1dc069c..bd16035 100644
--- a/tcg/tci/tcg-target.h
+++ b/tcg/tci/tcg-target.h
@@ -122,6 +122,8 @@
 #define TCG_TARGET_HAS_mulsh_i64        0
 #endif /* TCG_TARGET_REG_BITS == 64 */
 
+#define TCG_TARGET_HAS_new_ldst         0
+
 /* Number of registers available.
    For 32 bit hosts, we need more than 8 registers (call arguments). */
 /* #define TCG_TARGET_NB_REGS 8 */
-- 
1.8.1.4