softmmu templates: optionally pass CPUState to memory access functions

Optionally, make memory access helpers take a parameter for CPUState
instead of relying on global env.

On most targets, perform simple moves to reorder registers. On i386,
switch from regparm(3) calling convention to standard stack-based
version.

Signed-off-by: Blue Swirl <blauwirbel@gmail.com>

16 files changed, 576 insertions, 44 deletions

diff --git a/cpu-all.h b/cpu-all.h
index 8ad1ec7a4..9621c3c92 100644
--- a/cpu-all.h
+++ b/cpu-all.h
@@ -259,12 +259,21 @@ extern unsigned long reserved_va;
 #define stfl(p, v) stfl_raw(p, v)
 #define stfq(p, v) stfq_raw(p, v)
 
+#ifndef CONFIG_TCG_PASS_AREG0
 #define ldub_code(p) ldub_raw(p)
 #define ldsb_code(p) ldsb_raw(p)
 #define lduw_code(p) lduw_raw(p)
 #define ldsw_code(p) ldsw_raw(p)
 #define ldl_code(p) ldl_raw(p)
 #define ldq_code(p) ldq_raw(p)
+#else
+#define cpu_ldub_code(env1, p) ldub_raw(p)
+#define cpu_ldsb_code(env1, p) ldsb_raw(p)
+#define cpu_lduw_code(env1, p) lduw_raw(p)
+#define cpu_ldsw_code(env1, p) ldsw_raw(p)
+#define cpu_ldl_code(env1, p) ldl_raw(p)
+#define cpu_ldq_code(env1, p) ldq_raw(p)
+#endif
 
 #define ldub_kernel(p) ldub_raw(p)
 #define ldsb_kernel(p) ldsb_raw(p)
diff --git a/exec-all.h b/exec-all.h
index 8fca67b11..93a5b22c1 100644
--- a/exec-all.h
+++ b/exec-all.h
@@ -312,7 +312,9 @@ void tlb_fill(CPUArchState *env1, target_ulong addr, int is_write, int mmu_idx,
 
 #define ACCESS_TYPE (NB_MMU_MODES + 1)
 #define MEMSUFFIX _code
+#ifndef CONFIG_TCG_PASS_AREG0
 #define env cpu_single_env
+#endif
 
 #define DATA_SIZE 1
 #include "softmmu_header.h"
diff --git a/exec.c b/exec.c
index 8fd50a1c4..be392e276 100644
--- a/exec.c
+++ b/exec.c
@@ -4595,7 +4595,11 @@ tb_page_addr_t get_page_addr_code(CPUArchState *env1, target_ulong addr)
     mmu_idx = cpu_mmu_index(env1);
     if (unlikely(env1->tlb_table[mmu_idx][page_index].addr_code !=
                  (addr & TARGET_PAGE_MASK))) {
+#ifdef CONFIG_TCG_PASS_AREG0
+        cpu_ldub_code(env1, addr);
+#else
         ldub_code(addr);
+#endif
     }
     pd = env1->iotlb[mmu_idx][page_index] & ~TARGET_PAGE_MASK;
     mr = iotlb_to_region(pd);
diff --git a/softmmu_defs.h b/softmmu_defs.h
index d47d30d71..8d59f9d4f 100644
--- a/softmmu_defs.h
+++ b/softmmu_defs.h
@@ -9,6 +9,7 @@
 #ifndef SOFTMMU_DEFS_H
 #define SOFTMMU_DEFS_H
 
+#ifndef CONFIG_TCG_PASS_AREG0
 uint8_t __ldb_mmu(target_ulong addr, int mmu_idx);
 void __stb_mmu(target_ulong addr, uint8_t val, int mmu_idx);
 uint16_t __ldw_mmu(target_ulong addr, int mmu_idx);
@@ -26,5 +27,32 @@ uint32_t __ldl_cmmu(target_ulong addr, int mmu_idx);
 void __stl_cmmu(target_ulong addr, uint32_t val, int mmu_idx);
 uint64_t __ldq_cmmu(target_ulong addr, int mmu_idx);
 void __stq_cmmu(target_ulong addr, uint64_t val, int mmu_idx);
+#else
+uint8_t helper_ldb_mmu(CPUArchState *env, target_ulong addr, int mmu_idx);
+void helper_stb_mmu(CPUArchState *env, target_ulong addr, uint8_t val,
+                    int mmu_idx);
+uint16_t helper_ldw_mmu(CPUArchState *env, target_ulong addr, int mmu_idx);
+void helper_stw_mmu(CPUArchState *env, target_ulong addr, uint16_t val,
+                    int mmu_idx);
+uint32_t helper_ldl_mmu(CPUArchState *env, target_ulong addr, int mmu_idx);
+void helper_stl_mmu(CPUArchState *env, target_ulong addr, uint32_t val,
+                    int mmu_idx);
+uint64_t helper_ldq_mmu(CPUArchState *env, target_ulong addr, int mmu_idx);
+void helper_stq_mmu(CPUArchState *env, target_ulong addr, uint64_t val,
+                    int mmu_idx);
+
+uint8_t helper_ldb_cmmu(CPUArchState *env, target_ulong addr, int mmu_idx);
+void helper_stb_cmmu(CPUArchState *env, target_ulong addr, uint8_t val,
+int mmu_idx);
+uint16_t helper_ldw_cmmu(CPUArchState *env, target_ulong addr, int mmu_idx);
+void helper_stw_cmmu(CPUArchState *env, target_ulong addr, uint16_t val,
+                     int mmu_idx);
+uint32_t helper_ldl_cmmu(CPUArchState *env, target_ulong addr, int mmu_idx);
+void helper_stl_cmmu(CPUArchState *env, target_ulong addr, uint32_t val,
+                     int mmu_idx);
+uint64_t helper_ldq_cmmu(CPUArchState *env, target_ulong addr, int mmu_idx);
+void helper_stq_cmmu(CPUArchState *env, target_ulong addr, uint64_t val,
+                     int mmu_idx);
+#endif
 
 #endif
diff --git a/softmmu_header.h b/softmmu_header.h
index 818d7b662..6b72093a0 100644
--- a/softmmu_header.h
+++ b/softmmu_header.h
@@ -78,9 +78,23 @@
 #define ADDR_READ addr_read
 #endif
 
+#ifndef CONFIG_TCG_PASS_AREG0
+#define ENV_PARAM
+#define ENV_VAR
+#define CPU_PREFIX
+#define HELPER_PREFIX __
+#else
+#define ENV_PARAM CPUArchState *env,
+#define ENV_VAR env,
+#define CPU_PREFIX cpu_
+#define HELPER_PREFIX helper_
+#endif
+
 /* generic load/store macros */
 
-static inline RES_TYPE glue(glue(ld, USUFFIX), MEMSUFFIX)(target_ulong ptr)
+static inline RES_TYPE
+glue(glue(glue(CPU_PREFIX, ld), USUFFIX), MEMSUFFIX)(ENV_PARAM
+                                                     target_ulong ptr)
 {
     int page_index;
     RES_TYPE res;
@@ -93,7 +107,9 @@ static inline RES_TYPE glue(glue(ld, USUFFIX), MEMSUFFIX)(target_ulong ptr)
     mmu_idx = CPU_MMU_INDEX;
     if (unlikely(env->tlb_table[mmu_idx][page_index].ADDR_READ !=
                  (addr & (TARGET_PAGE_MASK | (DATA_SIZE - 1))))) {
-        res = glue(glue(__ld, SUFFIX), MMUSUFFIX)(addr, mmu_idx);
+        res = glue(glue(glue(HELPER_PREFIX, ld), SUFFIX), MMUSUFFIX)(ENV_VAR
+                                                                     addr,
+                                                                     mmu_idx);
     } else {
         physaddr = addr + env->tlb_table[mmu_idx][page_index].addend;
         res = glue(glue(ld, USUFFIX), _raw)((uint8_t *)physaddr);
@@ -102,7 +118,9 @@ static inline RES_TYPE glue(glue(ld, USUFFIX), MEMSUFFIX)(target_ulong ptr)
 }
 
 #if DATA_SIZE <= 2
-static inline int glue(glue(lds, SUFFIX), MEMSUFFIX)(target_ulong ptr)
+static inline int
+glue(glue(glue(CPU_PREFIX, lds), SUFFIX), MEMSUFFIX)(ENV_PARAM
+                                                     target_ulong ptr)
 {
     int res, page_index;
     target_ulong addr;
@@ -114,7 +132,8 @@ static inline int glue(glue(lds, SUFFIX), MEMSUFFIX)(target_ulong ptr)
     mmu_idx = CPU_MMU_INDEX;
     if (unlikely(env->tlb_table[mmu_idx][page_index].ADDR_READ !=
                  (addr & (TARGET_PAGE_MASK | (DATA_SIZE - 1))))) {
-        res = (DATA_STYPE)glue(glue(__ld, SUFFIX), MMUSUFFIX)(addr, mmu_idx);
+        res = (DATA_STYPE)glue(glue(glue(HELPER_PREFIX, ld), SUFFIX),
+                               MMUSUFFIX)(ENV_VAR addr, mmu_idx);
     } else {
         physaddr = addr + env->tlb_table[mmu_idx][page_index].addend;
         res = glue(glue(lds, SUFFIX), _raw)((uint8_t *)physaddr);
@@ -127,7 +146,9 @@ static inline int glue(glue(lds, SUFFIX), MEMSUFFIX)(target_ulong ptr)
 
 /* generic store macro */
 
-static inline void glue(glue(st, SUFFIX), MEMSUFFIX)(target_ulong ptr, RES_TYPE v)
+static inline void
+glue(glue(glue(CPU_PREFIX, st), SUFFIX), MEMSUFFIX)(ENV_PARAM target_ulong ptr,
+                                                    RES_TYPE v)
 {
     int page_index;
     target_ulong addr;
@@ -139,7 +160,8 @@ static inline void glue(glue(st, SUFFIX), MEMSUFFIX)(target_ulong ptr, RES_TYPE
     mmu_idx = CPU_MMU_INDEX;
     if (unlikely(env->tlb_table[mmu_idx][page_index].addr_write !=
                  (addr & (TARGET_PAGE_MASK | (DATA_SIZE - 1))))) {
-        glue(glue(__st, SUFFIX), MMUSUFFIX)(addr, v, mmu_idx);
+        glue(glue(glue(HELPER_PREFIX, st), SUFFIX), MMUSUFFIX)(ENV_VAR addr, v,
+                                                               mmu_idx);
     } else {
         physaddr = addr + env->tlb_table[mmu_idx][page_index].addend;
         glue(glue(st, SUFFIX), _raw)((uint8_t *)physaddr, v);
@@ -151,46 +173,52 @@ static inline void glue(glue(st, SUFFIX), MEMSUFFIX)(target_ulong ptr, RES_TYPE
 #if ACCESS_TYPE != (NB_MMU_MODES + 1)
 
 #if DATA_SIZE == 8
-static inline float64 glue(ldfq, MEMSUFFIX)(target_ulong ptr)
+static inline float64 glue(glue(CPU_PREFIX, ldfq), MEMSUFFIX)(ENV_PARAM
+                                                              target_ulong ptr)
 {
     union {
         float64 d;
         uint64_t i;
     } u;
-    u.i = glue(ldq, MEMSUFFIX)(ptr);
+    u.i = glue(glue(CPU_PREFIX, ldq), MEMSUFFIX)(ENV_VAR ptr);
     return u.d;
 }
 
-static inline void glue(stfq, MEMSUFFIX)(target_ulong ptr, float64 v)
+static inline void glue(glue(CPU_PREFIX, stfq), MEMSUFFIX)(ENV_PARAM
+                                                           target_ulong ptr,
+                                                           float64 v)
 {
     union {
         float64 d;
         uint64_t i;
     } u;
     u.d = v;
-    glue(stq, MEMSUFFIX)(ptr, u.i);
+    glue(glue(CPU_PREFIX, stq), MEMSUFFIX)(ENV_VAR ptr, u.i);
 }
 #endif /* DATA_SIZE == 8 */
 
 #if DATA_SIZE == 4
-static inline float32 glue(ldfl, MEMSUFFIX)(target_ulong ptr)
+static inline float32 glue(glue(CPU_PREFIX, ldfl), MEMSUFFIX)(ENV_PARAM
+                                                              target_ulong ptr)
 {
     union {
         float32 f;
         uint32_t i;
     } u;
-    u.i = glue(ldl, MEMSUFFIX)(ptr);
+    u.i = glue(glue(CPU_PREFIX, ldl), MEMSUFFIX)(ENV_VAR ptr);
     return u.f;
 }
 
-static inline void glue(stfl, MEMSUFFIX)(target_ulong ptr, float32 v)
+static inline void glue(glue(CPU_PREFIX, stfl), MEMSUFFIX)(ENV_PARAM
+                                                           target_ulong ptr,
+                                                           float32 v)
 {
     union {
         float32 f;
         uint32_t i;
     } u;
     u.f = v;
-    glue(stl, MEMSUFFIX)(ptr, u.i);
+    glue(glue(CPU_PREFIX, stl), MEMSUFFIX)(ENV_VAR ptr, u.i);
 }
 #endif /* DATA_SIZE == 4 */
 
@@ -205,3 +233,7 @@ static inline void glue(stfl, MEMSUFFIX)(target_ulong ptr, float32 v)
 #undef CPU_MMU_INDEX
 #undef MMUSUFFIX
 #undef ADDR_READ
+#undef ENV_PARAM
+#undef ENV_VAR
+#undef CPU_PREFIX
+#undef HELPER_PREFIX
diff --git a/softmmu_template.h b/softmmu_template.h
index d633bb555..afcab1e6a 100644
--- a/softmmu_template.h
+++ b/softmmu_template.h
@@ -54,10 +54,24 @@
 #define ADDR_READ addr_read
 #endif
 
-static DATA_TYPE glue(glue(slow_ld, SUFFIX), MMUSUFFIX)(target_ulong addr,
+#ifndef CONFIG_TCG_PASS_AREG0
+#define ENV_PARAM
+#define ENV_VAR
+#define CPU_PREFIX
+#define HELPER_PREFIX __
+#else
+#define ENV_PARAM CPUArchState *env,
+#define ENV_VAR env,
+#define CPU_PREFIX cpu_
+#define HELPER_PREFIX helper_
+#endif
+
+static DATA_TYPE glue(glue(slow_ld, SUFFIX), MMUSUFFIX)(ENV_PARAM
+                                                        target_ulong addr,
                                                         int mmu_idx,
                                                         void *retaddr);
-static inline DATA_TYPE glue(io_read, SUFFIX)(target_phys_addr_t physaddr,
+static inline DATA_TYPE glue(io_read, SUFFIX)(ENV_PARAM
+                                              target_phys_addr_t physaddr,
                                               target_ulong addr,
                                               void *retaddr)
 {
@@ -89,7 +103,10 @@ static inline DATA_TYPE glue(io_read, SUFFIX)(target_phys_addr_t physaddr,
 }
 
 /* handle all cases except unaligned access which span two pages */
-DATA_TYPE glue(glue(__ld, SUFFIX), MMUSUFFIX)(target_ulong addr, int mmu_idx)
+DATA_TYPE
+glue(glue(glue(HELPER_PREFIX, ld), SUFFIX), MMUSUFFIX)(ENV_PARAM
+                                                       target_ulong addr,
+                                                       int mmu_idx)
 {
     DATA_TYPE res;
     int index;
@@ -110,22 +127,22 @@ DATA_TYPE glue(glue(__ld, SUFFIX), MMUSUFFIX)(target_ulong addr, int mmu_idx)
                 goto do_unaligned_access;
             retaddr = GETPC();
             ioaddr = env->iotlb[mmu_idx][index];
-            res = glue(io_read, SUFFIX)(ioaddr, addr, retaddr);
+            res = glue(io_read, SUFFIX)(ENV_VAR ioaddr, addr, retaddr);
         } else if (((addr & ~TARGET_PAGE_MASK) + DATA_SIZE - 1) >= TARGET_PAGE_SIZE) {
             /* slow unaligned access (it spans two pages or IO) */
         do_unaligned_access:
             retaddr = GETPC();
 #ifdef ALIGNED_ONLY
-            do_unaligned_access(addr, READ_ACCESS_TYPE, mmu_idx, retaddr);
+            do_unaligned_access(ENV_VAR addr, READ_ACCESS_TYPE, mmu_idx, retaddr);
 #endif
-            res = glue(glue(slow_ld, SUFFIX), MMUSUFFIX)(addr,
+            res = glue(glue(slow_ld, SUFFIX), MMUSUFFIX)(ENV_VAR addr,
                                                          mmu_idx, retaddr);
         } else {
             /* unaligned/aligned access in the same page */
 #ifdef ALIGNED_ONLY
             if ((addr & (DATA_SIZE - 1)) != 0) {
                 retaddr = GETPC();
-                do_unaligned_access(addr, READ_ACCESS_TYPE, mmu_idx, retaddr);
+                do_unaligned_access(ENV_VAR addr, READ_ACCESS_TYPE, mmu_idx, retaddr);
             }
 #endif
             addend = env->tlb_table[mmu_idx][index].addend;
@@ -136,7 +153,7 @@ DATA_TYPE glue(glue(__ld, SUFFIX), MMUSUFFIX)(target_ulong addr, int mmu_idx)
         retaddr = GETPC();
 #ifdef ALIGNED_ONLY
         if ((addr & (DATA_SIZE - 1)) != 0)
-            do_unaligned_access(addr, READ_ACCESS_TYPE, mmu_idx, retaddr);
+            do_unaligned_access(ENV_VAR addr, READ_ACCESS_TYPE, mmu_idx, retaddr);
 #endif
         tlb_fill(env, addr, READ_ACCESS_TYPE, mmu_idx, retaddr);
         goto redo;
@@ -145,9 +162,11 @@ DATA_TYPE glue(glue(__ld, SUFFIX), MMUSUFFIX)(target_ulong addr, int mmu_idx)
 }
 
 /* handle all unaligned cases */
-static DATA_TYPE glue(glue(slow_ld, SUFFIX), MMUSUFFIX)(target_ulong addr,
-                                                        int mmu_idx,
-                                                        void *retaddr)
+static DATA_TYPE
+glue(glue(slow_ld, SUFFIX), MMUSUFFIX)(ENV_PARAM
+                                       target_ulong addr,
+                                       int mmu_idx,
+                                       void *retaddr)
 {
     DATA_TYPE res, res1, res2;
     int index, shift;
@@ -164,15 +183,15 @@ static DATA_TYPE glue(glue(slow_ld, SUFFIX), MMUSUFFIX)(target_ulong addr,
             if ((addr & (DATA_SIZE - 1)) != 0)
                 goto do_unaligned_access;
             ioaddr = env->iotlb[mmu_idx][index];
-            res = glue(io_read, SUFFIX)(ioaddr, addr, retaddr);
+            res = glue(io_read, SUFFIX)(ENV_VAR ioaddr, addr, retaddr);
         } else if (((addr & ~TARGET_PAGE_MASK) + DATA_SIZE - 1) >= TARGET_PAGE_SIZE) {
         do_unaligned_access:
             /* slow unaligned access (it spans two pages) */
             addr1 = addr & ~(DATA_SIZE - 1);
             addr2 = addr1 + DATA_SIZE;
-            res1 = glue(glue(slow_ld, SUFFIX), MMUSUFFIX)(addr1,
+            res1 = glue(glue(slow_ld, SUFFIX), MMUSUFFIX)(ENV_VAR addr1,
                                                           mmu_idx, retaddr);
-            res2 = glue(glue(slow_ld, SUFFIX), MMUSUFFIX)(addr2,
+            res2 = glue(glue(slow_ld, SUFFIX), MMUSUFFIX)(ENV_VAR addr2,
                                                           mmu_idx, retaddr);
             shift = (addr & (DATA_SIZE - 1)) * 8;
 #ifdef TARGET_WORDS_BIGENDIAN
@@ -196,12 +215,14 @@ static DATA_TYPE glue(glue(slow_ld, SUFFIX), MMUSUFFIX)(target_ulong addr,
 
 #ifndef SOFTMMU_CODE_ACCESS
 
-static void glue(glue(slow_st, SUFFIX), MMUSUFFIX)(target_ulong addr,
+static void glue(glue(slow_st, SUFFIX), MMUSUFFIX)(ENV_PARAM
+                                                   target_ulong addr,
                                                    DATA_TYPE val,
                                                    int mmu_idx,
                                                    void *retaddr);
 
-static inline void glue(io_write, SUFFIX)(target_phys_addr_t physaddr,
+static inline void glue(io_write, SUFFIX)(ENV_PARAM
+                                          target_phys_addr_t physaddr,
                                           DATA_TYPE val,
                                           target_ulong addr,
                                           void *retaddr)
@@ -231,8 +252,10 @@ static inline void glue(io_write, SUFFIX)(target_phys_addr_t physaddr,
 #endif /* SHIFT > 2 */
 }
 
-void glue(glue(__st, SUFFIX), MMUSUFFIX)(target_ulong addr, DATA_TYPE val,
-                                         int mmu_idx)
+void glue(glue(glue(HELPER_PREFIX, st), SUFFIX), MMUSUFFIX)(ENV_PARAM
+                                                            target_ulong addr,
+                                                            DATA_TYPE val,
+                                                            int mmu_idx)
 {
     target_phys_addr_t ioaddr;
     unsigned long addend;
@@ -250,21 +273,21 @@ void glue(glue(__st, SUFFIX), MMUSUFFIX)(target_ulong addr, DATA_TYPE val,
                 goto do_unaligned_access;
             retaddr = GETPC();
             ioaddr = env->iotlb[mmu_idx][index];
-            glue(io_write, SUFFIX)(ioaddr, val, addr, retaddr);
+            glue(io_write, SUFFIX)(ENV_VAR ioaddr, val, addr, retaddr);
         } else if (((addr & ~TARGET_PAGE_MASK) + DATA_SIZE - 1) >= TARGET_PAGE_SIZE) {
         do_unaligned_access:
             retaddr = GETPC();
 #ifdef ALIGNED_ONLY
-            do_unaligned_access(addr, 1, mmu_idx, retaddr);
+            do_unaligned_access(ENV_VAR addr, 1, mmu_idx, retaddr);
 #endif
-            glue(glue(slow_st, SUFFIX), MMUSUFFIX)(addr, val,
+            glue(glue(slow_st, SUFFIX), MMUSUFFIX)(ENV_VAR addr, val,
                                                    mmu_idx, retaddr);
         } else {
             /* aligned/unaligned access in the same page */
 #ifdef ALIGNED_ONLY
             if ((addr & (DATA_SIZE - 1)) != 0) {
                 retaddr = GETPC();
-                do_unaligned_access(addr, 1, mmu_idx, retaddr);
+                do_unaligned_access(ENV_VAR addr, 1, mmu_idx, retaddr);
             }
 #endif
             addend = env->tlb_table[mmu_idx][index].addend;
@@ -275,7 +298,7 @@ void glue(glue(__st, SUFFIX), MMUSUFFIX)(target_ulong addr, DATA_TYPE val,
         retaddr = GETPC();
 #ifdef ALIGNED_ONLY
         if ((addr & (DATA_SIZE - 1)) != 0)
-            do_unaligned_access(addr, 1, mmu_idx, retaddr);
+            do_unaligned_access(ENV_VAR addr, 1, mmu_idx, retaddr);
 #endif
         tlb_fill(env, addr, 1, mmu_idx, retaddr);
         goto redo;
@@ -283,7 +306,8 @@ void glue(glue(__st, SUFFIX), MMUSUFFIX)(target_ulong addr, DATA_TYPE val,
 }
 
 /* handles all unaligned cases */
-static void glue(glue(slow_st, SUFFIX), MMUSUFFIX)(target_ulong addr,
+static void glue(glue(slow_st, SUFFIX), MMUSUFFIX)(ENV_PARAM
+                                                   target_ulong addr,
                                                    DATA_TYPE val,
                                                    int mmu_idx,
                                                    void *retaddr)
@@ -302,7 +326,7 @@ static void glue(glue(slow_st, SUFFIX), MMUSUFFIX)(target_ulong addr,
             if ((addr & (DATA_SIZE - 1)) != 0)
                 goto do_unaligned_access;
             ioaddr = env->iotlb[mmu_idx][index];
-            glue(io_write, SUFFIX)(ioaddr, val, addr, retaddr);
+            glue(io_write, SUFFIX)(ENV_VAR ioaddr, val, addr, retaddr);
         } else if (((addr & ~TARGET_PAGE_MASK) + DATA_SIZE - 1) >= TARGET_PAGE_SIZE) {
         do_unaligned_access:
             /* XXX: not efficient, but simple */
@@ -310,10 +334,12 @@ static void glue(glue(slow_st, SUFFIX), MMUSUFFIX)(target_ulong addr,
              * previous page from the TLB cache.  */
             for(i = DATA_SIZE - 1; i >= 0; i--) {
 #ifdef TARGET_WORDS_BIGENDIAN
-                glue(slow_stb, MMUSUFFIX)(addr + i, val >> (((DATA_SIZE - 1) * 8) - (i * 8)),
+                glue(slow_stb, MMUSUFFIX)(ENV_VAR addr + i,
+                                          val >> (((DATA_SIZE - 1) * 8) - (i * 8)),
                                           mmu_idx, retaddr);
 #else
-                glue(slow_stb, MMUSUFFIX)(addr + i, val >> (i * 8),
+                glue(slow_stb, MMUSUFFIX)(ENV_VAR addr + i,
+                                          val >> (i * 8),
                                           mmu_idx, retaddr);
 #endif
             }
@@ -338,3 +364,7 @@ static void glue(glue(slow_st, SUFFIX), MMUSUFFIX)(target_ulong addr,
 #undef USUFFIX
 #undef DATA_SIZE
 #undef ADDR_READ
+#undef ENV_PARAM
+#undef ENV_VAR
+#undef CPU_PREFIX
+#undef HELPER_PREFIX
diff --git a/tcg/arm/tcg-target.c b/tcg/arm/tcg-target.c
index 5af21b3f5..4d59a6385 100644
--- a/tcg/arm/tcg-target.c
+++ b/tcg/arm/tcg-target.c
@@ -929,6 +929,27 @@ static inline void tcg_out_goto_label(TCGContext *s, int cond, int label_index)
 
 #include "../../softmmu_defs.h"
 
+#ifdef CONFIG_TCG_PASS_AREG0
+/* helper signature: helper_ld_mmu(CPUState *env, target_ulong addr,
+   int mmu_idx) */
+static const void * const qemu_ld_helpers[4] = {
+    helper_ldb_mmu,
+    helper_ldw_mmu,
+    helper_ldl_mmu,
+    helper_ldq_mmu,
+};
+
+/* helper signature: helper_st_mmu(CPUState *env, target_ulong addr,
+   uintxx_t val, int mmu_idx) */
+static const void * const qemu_st_helpers[4] = {
+    helper_stb_mmu,
+    helper_stw_mmu,
+    helper_stl_mmu,
+    helper_stq_mmu,
+};
+#else
+/* legacy helper signature: __ld_mmu(target_ulong addr, int
+   mmu_idx) */
 static void *qemu_ld_helpers[4] = {
     __ldb_mmu,
     __ldw_mmu,
@@ -936,6 +957,8 @@ static void *qemu_ld_helpers[4] = {
     __ldq_mmu,
 };
 
+/* legacy helper signature: __st_mmu(target_ulong addr, uintxx_t val,
+   int mmu_idx) */
 static void *qemu_st_helpers[4] = {
     __stb_mmu,
     __stw_mmu,
@@ -943,6 +966,7 @@ static void *qemu_st_helpers[4] = {
     __stq_mmu,
 };
 #endif
+#endif
 
 #define TLB_SHIFT	(CPU_TLB_ENTRY_BITS + CPU_TLB_BITS)
 
@@ -1075,6 +1099,19 @@ static inline void tcg_out_qemu_ld(TCGContext *s, const TCGArg *args, int opc)
                     TCG_REG_R1, 0, addr_reg2, SHIFT_IMM_LSL(0));
     tcg_out_dat_imm(s, COND_AL, ARITH_MOV, TCG_REG_R2, 0, mem_index);
 # endif
+#ifdef CONFIG_TCG_PASS_AREG0
+    /* XXX/FIXME: suboptimal and incorrect for 64 bit */
+    tcg_out_dat_reg(s, COND_AL, ARITH_MOV,
+                    tcg_target_call_iarg_regs[2], 0,
+                    tcg_target_call_iarg_regs[1], SHIFT_IMM_LSL(0));
+    tcg_out_dat_reg(s, COND_AL, ARITH_MOV,
+                    tcg_target_call_iarg_regs[1], 0,
+                    tcg_target_call_iarg_regs[0], SHIFT_IMM_LSL(0));
+
+    tcg_out_dat_reg(s, COND_AL, ARITH_MOV,
+                    tcg_target_call_iarg_regs[0], 0, TCG_AREG0,
+                    SHIFT_IMM_LSL(0));
+#endif
     tcg_out_call(s, (tcg_target_long) qemu_ld_helpers[s_bits]);
 
     switch (opc) {
@@ -1341,6 +1378,22 @@ static inline void tcg_out_qemu_st(TCGContext *s, const TCGArg *args, int opc)
     }
 # endif
 
+#ifdef CONFIG_TCG_PASS_AREG0
+    /* XXX/FIXME: suboptimal and incorrect for 64 bit */
+    tcg_out_dat_reg(s, COND_AL, ARITH_MOV,
+                    tcg_target_call_iarg_regs[3], 0,
+                    tcg_target_call_iarg_regs[2], SHIFT_IMM_LSL(0));
+    tcg_out_dat_reg(s, COND_AL, ARITH_MOV,
+                    tcg_target_call_iarg_regs[2], 0,
+                    tcg_target_call_iarg_regs[1], SHIFT_IMM_LSL(0));
+    tcg_out_dat_reg(s, COND_AL, ARITH_MOV,
+                    tcg_target_call_iarg_regs[1], 0,
+                    tcg_target_call_iarg_regs[0], SHIFT_IMM_LSL(0));
+
+    tcg_out_dat_reg(s, COND_AL, ARITH_MOV,
+                    tcg_target_call_iarg_regs[0], 0, TCG_AREG0,
+                    SHIFT_IMM_LSL(0));
+#endif
     tcg_out_call(s, (tcg_target_long) qemu_st_helpers[s_bits]);
     if (opc == 3)
         tcg_out_dat_imm(s, COND_AL, ARITH_ADD, TCG_REG_R13, TCG_REG_R13, 0x10);
diff --git a/tcg/hppa/tcg-target.c b/tcg/hppa/tcg-target.c
index c5a3730a2..e579ef06a 100644
--- a/tcg/hppa/tcg-target.c
+++ b/tcg/hppa/tcg-target.c
@@ -882,6 +882,27 @@ static void tcg_out_setcond2(TCGContext *s, int cond, TCGArg ret,
 #if defined(CONFIG_SOFTMMU)
 #include "../../softmmu_defs.h"
 
+#ifdef CONFIG_TCG_PASS_AREG0
+/* helper signature: helper_ld_mmu(CPUState *env, target_ulong addr,
+   int mmu_idx) */
+static const void * const qemu_ld_helpers[4] = {
+    helper_ldb_mmu,
+    helper_ldw_mmu,
+    helper_ldl_mmu,
+    helper_ldq_mmu,
+};
+
+/* helper signature: helper_st_mmu(CPUState *env, target_ulong addr,
+   uintxx_t val, int mmu_idx) */
+static const void * const qemu_st_helpers[4] = {
+    helper_stb_mmu,
+    helper_stw_mmu,
+    helper_stl_mmu,
+    helper_stq_mmu,
+};
+#else
+/* legacy helper signature: __ld_mmu(target_ulong addr, int
+   mmu_idx) */
 static void *qemu_ld_helpers[4] = {
     __ldb_mmu,
     __ldw_mmu,
@@ -889,12 +910,15 @@ static void *qemu_ld_helpers[4] = {
     __ldq_mmu,
 };
 
+/* legacy helper signature: __st_mmu(target_ulong addr, uintxx_t val,
+   int mmu_idx) */
 static void *qemu_st_helpers[4] = {
     __stb_mmu,
     __stw_mmu,
     __stl_mmu,
     __stq_mmu,
 };
+#endif
 
 /* Load and compare a TLB entry, and branch if TLB miss.  OFFSET is set to
    the offset of the first ADDR_READ or ADDR_WRITE member of the appropriate
@@ -1061,6 +1085,15 @@ static void tcg_out_qemu_ld(TCGContext *s, const TCGArg *args, int opc)
     }
     tcg_out_movi(s, TCG_TYPE_I32, argreg, mem_index);
 
+#ifdef CONFIG_TCG_PASS_AREG0
+    /* XXX/FIXME: suboptimal */
+    tcg_out_mov(s, TCG_TYPE_I32, tcg_target_call_iarg_regs[2],
+                tcg_target_call_iarg_regs[1]);
+    tcg_out_mov(s, TCG_TYPE_TL, tcg_target_call_iarg_regs[1],
+                tcg_target_call_iarg_regs[0]);
+    tcg_out_mov(s, TCG_TYPE_PTR, tcg_target_call_iarg_regs[0],
+                TCG_AREG0);
+#endif
     tcg_out_call(s, qemu_ld_helpers[opc & 3]);
 
     switch (opc) {
@@ -1212,6 +1245,17 @@ static void tcg_out_qemu_st(TCGContext *s, const TCGArg *args, int opc)
         tcg_abort();
     }
 
+#ifdef CONFIG_TCG_PASS_AREG0
+    /* XXX/FIXME: suboptimal */
+    tcg_out_mov(s, TCG_TYPE_I32, tcg_target_call_iarg_regs[3],
+                tcg_target_call_iarg_regs[2]);
+    tcg_out_mov(s, TCG_TYPE_I64, tcg_target_call_iarg_regs[2],
+                tcg_target_call_iarg_regs[1]);
+    tcg_out_mov(s, TCG_TYPE_TL, tcg_target_call_iarg_regs[1],
+                tcg_target_call_iarg_regs[0]);
+    tcg_out_mov(s, TCG_TYPE_PTR, tcg_target_call_iarg_regs[0],
+                TCG_AREG0);
+#endif
     tcg_out_call(s, qemu_st_helpers[opc]);
 
     /* label2: */
diff --git a/tcg/i386/tcg-target.c b/tcg/i386/tcg-target.c
index c4e940da5..43a51a1c5 100644
--- a/tcg/i386/tcg-target.c
+++ b/tcg/i386/tcg-target.c
@@ -178,6 +178,9 @@ static int target_parse_constraint(TCGArgConstraint *ct, const char **pct_str)
             tcg_regset_set32(ct->u.regs, 0, 0xffff);
             tcg_regset_reset_reg(ct->u.regs, TCG_REG_RSI);
             tcg_regset_reset_reg(ct->u.regs, TCG_REG_RDI);
+#ifdef CONFIG_TCG_PASS_AREG0
+            tcg_regset_reset_reg(ct->u.regs, TCG_REG_RDX);
+#endif
         } else {
             tcg_regset_set32(ct->u.regs, 0, 0xff);
             tcg_regset_reset_reg(ct->u.regs, TCG_REG_EAX);
@@ -957,6 +960,27 @@ static void tcg_out_jmp(TCGContext *s, tcg_target_long dest)
 
 #include "../../softmmu_defs.h"
 
+#ifdef CONFIG_TCG_PASS_AREG0
+/* helper signature: helper_ld_mmu(CPUState *env, target_ulong addr,
+   int mmu_idx) */
+static const void *qemu_ld_helpers[4] = {
+    helper_ldb_mmu,
+    helper_ldw_mmu,
+    helper_ldl_mmu,
+    helper_ldq_mmu,
+};
+
+/* helper signature: helper_st_mmu(CPUState *env, target_ulong addr,
+   uintxx_t val, int mmu_idx) */
+static const void *qemu_st_helpers[4] = {
+    helper_stb_mmu,
+    helper_stw_mmu,
+    helper_stl_mmu,
+    helper_stq_mmu,
+};
+#else
+/* legacy helper signature: __ld_mmu(target_ulong addr, int
+   mmu_idx) */
 static void *qemu_ld_helpers[4] = {
     __ldb_mmu,
     __ldw_mmu,
@@ -964,12 +988,15 @@ static void *qemu_ld_helpers[4] = {
     __ldq_mmu,
 };
 
+/* legacy helper signature: __st_mmu(target_ulong addr, uintxx_t val,
+   int mmu_idx) */
 static void *qemu_st_helpers[4] = {
     __stb_mmu,
     __stw_mmu,
     __stl_mmu,
     __stq_mmu,
 };
+#endif
 
 /* Perform the TLB load and compare.
 
@@ -1188,11 +1215,26 @@ static void tcg_out_qemu_ld(TCGContext *s, const TCGArg *args,
     }
     tcg_out_push(s, args[addrlo_idx]);
     stack_adjust += 4;
+#ifdef CONFIG_TCG_PASS_AREG0
+    tcg_out_push(s, TCG_AREG0);
+    stack_adjust += 4;
+#endif
 #else
     /* The first argument is already loaded with addrlo.  */
     arg_idx = 1;
     tcg_out_movi(s, TCG_TYPE_I32, tcg_target_call_iarg_regs[arg_idx],
                  mem_index);
+#ifdef CONFIG_TCG_PASS_AREG0
+    /* XXX/FIXME: suboptimal */
+    tcg_out_mov(s, TCG_TYPE_I64, tcg_target_call_iarg_regs[3],
+                tcg_target_call_iarg_regs[2]);
+    tcg_out_mov(s, TCG_TYPE_I64, tcg_target_call_iarg_regs[2],
+                tcg_target_call_iarg_regs[1]);
+    tcg_out_mov(s, TCG_TYPE_I64, tcg_target_call_iarg_regs[1],
+                tcg_target_call_iarg_regs[0]);
+    tcg_out_mov(s, TCG_TYPE_I64, tcg_target_call_iarg_regs[0],
+                TCG_AREG0);
+#endif
 #endif
 
     tcg_out_calli(s, (tcg_target_long)qemu_ld_helpers[s_bits]);
@@ -1386,11 +1428,26 @@ static void tcg_out_qemu_st(TCGContext *s, const TCGArg *args,
     }
     tcg_out_push(s, args[addrlo_idx]);
     stack_adjust += 4;
+#ifdef CONFIG_TCG_PASS_AREG0
+    tcg_out_push(s, TCG_AREG0);
+    stack_adjust += 4;
+#endif
 #else
     tcg_out_mov(s, (opc == 3 ? TCG_TYPE_I64 : TCG_TYPE_I32),
                 TCG_REG_RSI, data_reg);
     tcg_out_movi(s, TCG_TYPE_I32, TCG_REG_RDX, mem_index);
     stack_adjust = 0;
+#ifdef CONFIG_TCG_PASS_AREG0
+    /* XXX/FIXME: suboptimal */
+    tcg_out_mov(s, TCG_TYPE_I64, tcg_target_call_iarg_regs[3],
+                tcg_target_call_iarg_regs[2]);
+    tcg_out_mov(s, TCG_TYPE_I64, tcg_target_call_iarg_regs[2],
+                tcg_target_call_iarg_regs[1]);
+    tcg_out_mov(s, TCG_TYPE_I64, tcg_target_call_iarg_regs[1],
+                tcg_target_call_iarg_regs[0]);
+    tcg_out_mov(s, TCG_TYPE_I64, tcg_target_call_iarg_regs[0],
+                TCG_AREG0);
+#endif
 #endif
 
     tcg_out_calli(s, (tcg_target_long)qemu_st_helpers[s_bits]);
diff --git a/tcg/ia64/tcg-target.c b/tcg/ia64/tcg-target.c
index f90252a44..e02dacc84 100644
--- a/tcg/ia64/tcg-target.c
+++ b/tcg/ia64/tcg-target.c
@@ -1452,12 +1452,25 @@ static inline void tcg_out_qemu_tlb(TCGContext *s, TCGArg addr_reg,
                                TCG_REG_P7, TCG_REG_R3, TCG_REG_R57));
 }
 
+#ifdef CONFIG_TCG_PASS_AREG0
+/* helper signature: helper_ld_mmu(CPUState *env, target_ulong addr,
+   int mmu_idx) */
+static const void * const qemu_ld_helpers[4] = {
+    helper_ldb_mmu,
+    helper_ldw_mmu,
+    helper_ldl_mmu,
+    helper_ldq_mmu,
+};
+#else
+/* legacy helper signature: __ld_mmu(target_ulong addr, int
+   mmu_idx) */
 static void *qemu_ld_helpers[4] = {
     __ldb_mmu,
     __ldw_mmu,
     __ldl_mmu,
     __ldq_mmu,
 };
+#endif
 
 static inline void tcg_out_qemu_ld(TCGContext *s, const TCGArg *args, int opc)
 {
@@ -1517,6 +1530,15 @@ static inline void tcg_out_qemu_ld(TCGContext *s, const TCGArg *args, int opc)
                        tcg_opc_m1 (TCG_REG_P7, OPC_LD8_M1, TCG_REG_R1, TCG_REG_R2),
                        tcg_opc_i18(TCG_REG_P0, OPC_NOP_I18, 0));
     }
+#ifdef CONFIG_TCG_PASS_AREG0
+    /* XXX/FIXME: suboptimal */
+    tcg_out_mov(s, TCG_TYPE_I32, tcg_target_call_iarg_regs[2],
+                tcg_target_call_iarg_regs[1]);
+    tcg_out_mov(s, TCG_TYPE_TL, tcg_target_call_iarg_regs[1],
+                tcg_target_call_iarg_regs[0]);
+    tcg_out_mov(s, TCG_TYPE_PTR, tcg_target_call_iarg_regs[0],
+                TCG_AREG0);
+#endif
     if (!bswap || s_bits == 0) {
         tcg_out_bundle(s, miB,
                        tcg_opc_m48(TCG_REG_P0, OPC_NOP_M48, 0),
@@ -1547,12 +1569,25 @@ static inline void tcg_out_qemu_ld(TCGContext *s, const TCGArg *args, int opc)
     }
 }
 
+#ifdef CONFIG_TCG_PASS_AREG0
+/* helper signature: helper_st_mmu(CPUState *env, target_ulong addr,
+   uintxx_t val, int mmu_idx) */
+static const void * const qemu_st_helpers[4] = {
+    helper_stb_mmu,
+    helper_stw_mmu,
+    helper_stl_mmu,
+    helper_stq_mmu,
+};
+#else
+/* legacy helper signature: __st_mmu(target_ulong addr, uintxx_t val,
+   int mmu_idx) */
 static void *qemu_st_helpers[4] = {
     __stb_mmu,
     __stw_mmu,
     __stl_mmu,
     __stq_mmu,
 };
+#endif
 
 static inline void tcg_out_qemu_st(TCGContext *s, const TCGArg *args, int opc)
 {
@@ -1622,6 +1657,17 @@ static inline void tcg_out_qemu_st(TCGContext *s, const TCGArg *args, int opc)
         data_reg = TCG_REG_R2;
     }
 
+#ifdef CONFIG_TCG_PASS_AREG0
+    /* XXX/FIXME: suboptimal */
+    tcg_out_mov(s, TCG_TYPE_I32, tcg_target_call_iarg_regs[3],
+                tcg_target_call_iarg_regs[2]);
+    tcg_out_mov(s, TCG_TYPE_I64, tcg_target_call_iarg_regs[2],
+                tcg_target_call_iarg_regs[1]);
+    tcg_out_mov(s, TCG_TYPE_TL, tcg_target_call_iarg_regs[1],
+                tcg_target_call_iarg_regs[0]);
+    tcg_out_mov(s, TCG_TYPE_PTR, tcg_target_call_iarg_regs[0],
+                TCG_AREG0);
+#endif
     tcg_out_bundle(s, miB,
                    tcg_opc_m4 (TCG_REG_P6, opc_st_m4[opc],
                                data_reg, TCG_REG_R3),
diff --git a/tcg/mips/tcg-target.c b/tcg/mips/tcg-target.c
index c6aa5bced..393ba07f2 100644
--- a/tcg/mips/tcg-target.c
+++ b/tcg/mips/tcg-target.c
@@ -750,6 +750,27 @@ static void tcg_out_setcond2(TCGContext *s, TCGCond cond, int ret,
 
 #include "../../softmmu_defs.h"
 
+#ifdef CONFIG_TCG_PASS_AREG0
+/* helper signature: helper_ld_mmu(CPUState *env, target_ulong addr,
+   int mmu_idx) */
+static const void * const qemu_ld_helpers[4] = {
+    helper_ldb_mmu,
+    helper_ldw_mmu,
+    helper_ldl_mmu,
+    helper_ldq_mmu,
+};
+
+/* helper signature: helper_st_mmu(CPUState *env, target_ulong addr,
+   uintxx_t val, int mmu_idx) */
+static const void * const qemu_st_helpers[4] = {
+    helper_stb_mmu,
+    helper_stw_mmu,
+    helper_stl_mmu,
+    helper_stq_mmu,
+};
+#else
+/* legacy helper signature: __ld_mmu(target_ulong addr, int
+   mmu_idx) */
 static void *qemu_ld_helpers[4] = {
     __ldb_mmu,
     __ldw_mmu,
@@ -757,6 +778,8 @@ static void *qemu_ld_helpers[4] = {
     __ldq_mmu,
 };
 
+/* legacy helper signature: __st_mmu(target_ulong addr, uintxx_t val,
+   int mmu_idx) */
 static void *qemu_st_helpers[4] = {
     __stb_mmu,
     __stw_mmu,
@@ -764,6 +787,7 @@ static void *qemu_st_helpers[4] = {
     __stq_mmu,
 };
 #endif
+#endif
 
 static void tcg_out_qemu_ld(TCGContext *s, const TCGArg *args,
                             int opc)
@@ -858,6 +882,15 @@ static void tcg_out_qemu_ld(TCGContext *s, const TCGArg *args,
 # endif
     tcg_out_movi(s, TCG_TYPE_I32, sp_args++, mem_index);
     tcg_out_movi(s, TCG_TYPE_I32, TCG_REG_T9, (tcg_target_long)qemu_ld_helpers[s_bits]);
+#ifdef CONFIG_TCG_PASS_AREG0
+    /* XXX/FIXME: suboptimal and incorrect for 64 on 32 bit */
+    tcg_out_mov(s, TCG_TYPE_I32, tcg_target_call_iarg_regs[2],
+                tcg_target_call_iarg_regs[1]);
+    tcg_out_mov(s, TCG_TYPE_TL, tcg_target_call_iarg_regs[1],
+                tcg_target_call_iarg_regs[0]);
+    tcg_out_mov(s, TCG_TYPE_PTR, tcg_target_call_iarg_regs[0],
+                TCG_AREG0);
+#endif
     tcg_out_opc_reg(s, OPC_JALR, TCG_REG_RA, TCG_REG_T9, 0);
     tcg_out_nop(s);
 
@@ -1069,6 +1102,17 @@ static void tcg_out_qemu_st(TCGContext *s, const TCGArg *args,
     }
 
     tcg_out_movi(s, TCG_TYPE_I32, TCG_REG_T9, (tcg_target_long)qemu_st_helpers[s_bits]);
+#ifdef CONFIG_TCG_PASS_AREG0
+    /* XXX/FIXME: suboptimal and incorrect for 64 on 32 bit */
+    tcg_out_mov(s, TCG_TYPE_I32, tcg_target_call_iarg_regs[3],
+                tcg_target_call_iarg_regs[2]);
+    tcg_out_mov(s, TCG_TYPE_I64, tcg_target_call_iarg_regs[2],
+                tcg_target_call_iarg_regs[1]);
+    tcg_out_mov(s, TCG_TYPE_TL, tcg_target_call_iarg_regs[1],
+                tcg_target_call_iarg_regs[0]);
+    tcg_out_mov(s, TCG_TYPE_PTR, tcg_target_call_iarg_regs[0],
+                TCG_AREG0);
+#endif
     tcg_out_opc_reg(s, OPC_JALR, TCG_REG_RA, TCG_REG_T9, 0);
     tcg_out_nop(s);
 
diff --git a/tcg/ppc/tcg-target.c b/tcg/ppc/tcg-target.c
index 6a34cab5f..b0aa91479 100644
--- a/tcg/ppc/tcg-target.c
+++ b/tcg/ppc/tcg-target.c
@@ -508,6 +508,27 @@ static void tcg_out_call (TCGContext *s, tcg_target_long arg, int const_arg)
 
 #include "../../softmmu_defs.h"
 
+#ifdef CONFIG_TCG_PASS_AREG0
+/* helper signature: helper_ld_mmu(CPUState *env, target_ulong addr,
+   int mmu_idx) */
+static const void * const qemu_ld_helpers[4] = {
+    helper_ldb_mmu,
+    helper_ldw_mmu,
+    helper_ldl_mmu,
+    helper_ldq_mmu,
+};
+
+/* helper signature: helper_st_mmu(CPUState *env, target_ulong addr,
+   uintxx_t val, int mmu_idx) */
+static const void * const qemu_st_helpers[4] = {
+    helper_stb_mmu,
+    helper_stw_mmu,
+    helper_stl_mmu,
+    helper_stq_mmu,
+};
+#else
+/* legacy helper signature: __ld_mmu(target_ulong addr, int
+   mmu_idx) */
 static void *qemu_ld_helpers[4] = {
     __ldb_mmu,
     __ldw_mmu,
@@ -515,6 +536,8 @@ static void *qemu_ld_helpers[4] = {
     __ldq_mmu,
 };
 
+/* legacy helper signature: __ld_mmu(target_ulong addr, int
+   mmu_idx) */
 static void *qemu_st_helpers[4] = {
     __stb_mmu,
     __stw_mmu,
@@ -522,6 +545,7 @@ static void *qemu_st_helpers[4] = {
     __stq_mmu,
 };
 #endif
+#endif
 
 static void tcg_out_qemu_ld (TCGContext *s, const TCGArg *args, int opc)
 {
@@ -598,6 +622,16 @@ static void tcg_out_qemu_ld (TCGContext *s, const TCGArg *args, int opc)
     tcg_out_movi (s, TCG_TYPE_I32, 5, mem_index);
 #endif
 
+#ifdef CONFIG_TCG_PASS_AREG0
+    /* XXX/FIXME: suboptimal */
+    tcg_out_mov(s, TCG_TYPE_I32, tcg_target_call_iarg_regs[2],
+                tcg_target_call_iarg_regs[1]);
+    tcg_out_mov(s, TCG_TYPE_TL, tcg_target_call_iarg_regs[1],
+                tcg_target_call_iarg_regs[0]);
+    tcg_out_mov(s, TCG_TYPE_PTR, tcg_target_call_iarg_regs[0],
+                TCG_AREG0);
+#endif
+
     tcg_out_call (s, (tcg_target_long) qemu_ld_helpers[s_bits], 1);
     switch (opc) {
     case 0|4:
@@ -829,6 +863,17 @@ static void tcg_out_qemu_st (TCGContext *s, const TCGArg *args, int opc)
     ir++;
 
     tcg_out_movi (s, TCG_TYPE_I32, ir, mem_index);
+#ifdef CONFIG_TCG_PASS_AREG0
+    /* XXX/FIXME: suboptimal */
+    tcg_out_mov(s, TCG_TYPE_I32, tcg_target_call_iarg_regs[3],
+                tcg_target_call_iarg_regs[2]);
+    tcg_out_mov(s, TCG_TYPE_I64, tcg_target_call_iarg_regs[2],
+                tcg_target_call_iarg_regs[1]);
+    tcg_out_mov(s, TCG_TYPE_TL, tcg_target_call_iarg_regs[1],
+                tcg_target_call_iarg_regs[0]);
+    tcg_out_mov(s, TCG_TYPE_PTR, tcg_target_call_iarg_regs[0],
+                TCG_AREG0);
+#endif
     tcg_out_call (s, (tcg_target_long) qemu_st_helpers[opc], 1);
     label2_ptr = s->code_ptr;
     tcg_out32 (s, B);
diff --git a/tcg/ppc64/tcg-target.c b/tcg/ppc64/tcg-target.c
index 7f723b5c2..409a1ac1c 100644
--- a/tcg/ppc64/tcg-target.c
+++ b/tcg/ppc64/tcg-target.c
@@ -552,6 +552,27 @@ static void tcg_out_ldsta (TCGContext *s, int ret, int addr,
 
 #include "../../softmmu_defs.h"
 
+#ifdef CONFIG_TCG_PASS_AREG0
+/* helper signature: helper_ld_mmu(CPUState *env, target_ulong addr,
+   int mmu_idx) */
+static const void * const qemu_ld_helpers[4] = {
+    helper_ldb_mmu,
+    helper_ldw_mmu,
+    helper_ldl_mmu,
+    helper_ldq_mmu,
+};
+
+/* helper signature: helper_st_mmu(CPUState *env, target_ulong addr,
+   uintxx_t val, int mmu_idx) */
+static const void * const qemu_st_helpers[4] = {
+    helper_stb_mmu,
+    helper_stw_mmu,
+    helper_stl_mmu,
+    helper_stq_mmu,
+};
+#else
+/* legacy helper signature: __ld_mmu(target_ulong addr, int
+   mmu_idx) */
 static void *qemu_ld_helpers[4] = {
     __ldb_mmu,
     __ldw_mmu,
@@ -559,12 +580,15 @@ static void *qemu_ld_helpers[4] = {
     __ldq_mmu,
 };
 
+/* legacy helper signature: __st_mmu(target_ulong addr, uintxx_t val,
+   int mmu_idx) */
 static void *qemu_st_helpers[4] = {
     __stb_mmu,
     __stw_mmu,
     __stl_mmu,
     __stq_mmu,
 };
+#endif
 
 static void tcg_out_tlb_read (TCGContext *s, int r0, int r1, int r2,
                               int addr_reg, int s_bits, int offset)
@@ -648,6 +672,15 @@ static void tcg_out_qemu_ld (TCGContext *s, const TCGArg *args, int opc)
     tcg_out_mov (s, TCG_TYPE_I64, 3, addr_reg);
     tcg_out_movi (s, TCG_TYPE_I64, 4, mem_index);
 
+#ifdef CONFIG_TCG_PASS_AREG0
+    /* XXX/FIXME: suboptimal */
+    tcg_out_mov(s, TCG_TYPE_I32, tcg_target_call_iarg_regs[2],
+                tcg_target_call_iarg_regs[1]);
+    tcg_out_mov(s, TCG_TYPE_TL, tcg_target_call_iarg_regs[1],
+                tcg_target_call_iarg_regs[0]);
+    tcg_out_mov(s, TCG_TYPE_PTR, tcg_target_call_iarg_regs[0],
+                TCG_AREG0);
+#endif
     tcg_out_call (s, (tcg_target_long) qemu_ld_helpers[s_bits], 1);
 
     switch (opc) {
@@ -796,6 +829,17 @@ static void tcg_out_qemu_st (TCGContext *s, const TCGArg *args, int opc)
     tcg_out_rld (s, RLDICL, 4, data_reg, 0, 64 - (1 << (3 + opc)));
     tcg_out_movi (s, TCG_TYPE_I64, 5, mem_index);
 
+#ifdef CONFIG_TCG_PASS_AREG0
+    /* XXX/FIXME: suboptimal */
+    tcg_out_mov(s, TCG_TYPE_I32, tcg_target_call_iarg_regs[3],
+                tcg_target_call_iarg_regs[2]);
+    tcg_out_mov(s, TCG_TYPE_I64, tcg_target_call_iarg_regs[2],
+                tcg_target_call_iarg_regs[1]);
+    tcg_out_mov(s, TCG_TYPE_TL, tcg_target_call_iarg_regs[1],
+                tcg_target_call_iarg_regs[0]);
+    tcg_out_mov(s, TCG_TYPE_PTR, tcg_target_call_iarg_regs[0],
+                TCG_AREG0);
+#endif
     tcg_out_call (s, (tcg_target_long) qemu_st_helpers[opc], 1);
 
     label2_ptr = s->code_ptr;
diff --git a/tcg/s390/tcg-target.c b/tcg/s390/tcg-target.c
index 47ffcc1f5..04662c15f 100644
--- a/tcg/s390/tcg-target.c
+++ b/tcg/s390/tcg-target.c
@@ -301,6 +301,27 @@ static const uint8_t tcg_cond_to_ltr_cond[10] = {
 
 #include "../../softmmu_defs.h"
 
+#ifdef CONFIG_TCG_PASS_AREG0
+/* helper signature: helper_ld_mmu(CPUState *env, target_ulong addr,
+   int mmu_idx) */
+static const void * const qemu_ld_helpers[4] = {
+    helper_ldb_mmu,
+    helper_ldw_mmu,
+    helper_ldl_mmu,
+    helper_ldq_mmu,
+};
+
+/* helper signature: helper_st_mmu(CPUState *env, target_ulong addr,
+   uintxx_t val, int mmu_idx) */
+static const void * const qemu_st_helpers[4] = {
+    helper_stb_mmu,
+    helper_stw_mmu,
+    helper_stl_mmu,
+    helper_stq_mmu,
+};
+#else
+/* legacy helper signature: __ld_mmu(target_ulong addr, int
+   mmu_idx) */
 static void *qemu_ld_helpers[4] = {
     __ldb_mmu,
     __ldw_mmu,
@@ -308,6 +329,8 @@ static void *qemu_ld_helpers[4] = {
     __ldq_mmu,
 };
 
+/* legacy helper signature: __st_mmu(target_ulong addr, uintxx_t val,
+   int mmu_idx) */
 static void *qemu_st_helpers[4] = {
     __stb_mmu,
     __stw_mmu,
@@ -315,6 +338,7 @@ static void *qemu_st_helpers[4] = {
     __stq_mmu,
 };
 #endif
+#endif
 
 static uint8_t *tb_ret_addr;
 
@@ -1483,9 +1507,29 @@ static void tcg_prepare_qemu_ldst(TCGContext* s, TCGReg data_reg,
             tcg_abort();
         }
         tcg_out_movi(s, TCG_TYPE_I32, TCG_REG_R4, mem_index);
+#ifdef CONFIG_TCG_PASS_AREG0
+        /* XXX/FIXME: suboptimal */
+        tcg_out_mov(s, TCG_TYPE_I32, tcg_target_call_iarg_regs[2],
+                    tcg_target_call_iarg_regs[1]);
+        tcg_out_mov(s, TCG_TYPE_TL, tcg_target_call_iarg_regs[1],
+                    tcg_target_call_iarg_regs[0]);
+        tcg_out_mov(s, TCG_TYPE_PTR, tcg_target_call_iarg_regs[0],
+                    TCG_AREG0);
+#endif
         tgen_calli(s, (tcg_target_ulong)qemu_st_helpers[s_bits]);
     } else {
         tcg_out_movi(s, TCG_TYPE_I32, arg1, mem_index);
+#ifdef CONFIG_TCG_PASS_AREG0
+        /* XXX/FIXME: suboptimal */
+        tcg_out_mov(s, TCG_TYPE_I32, tcg_target_call_iarg_regs[3],
+                    tcg_target_call_iarg_regs[2]);
+        tcg_out_mov(s, TCG_TYPE_I64, tcg_target_call_iarg_regs[2],
+                    tcg_target_call_iarg_regs[1]);
+        tcg_out_mov(s, TCG_TYPE_TL, tcg_target_call_iarg_regs[1],
+                    tcg_target_call_iarg_regs[0]);
+        tcg_out_mov(s, TCG_TYPE_PTR, tcg_target_call_iarg_regs[0],
+                    TCG_AREG0);
+#endif
         tgen_calli(s, (tcg_target_ulong)qemu_ld_helpers[s_bits]);
 
         /* sign extension */
diff --git a/tcg/sparc/tcg-target.c b/tcg/sparc/tcg-target.c
index b287122df..80f081867 100644
--- a/tcg/sparc/tcg-target.c
+++ b/tcg/sparc/tcg-target.c
@@ -59,6 +59,12 @@ static const char * const tcg_target_reg_names[TCG_TARGET_NB_REGS] = {
 };
 #endif
 
+#ifdef CONFIG_TCG_PASS_AREG0
+#define ARG_OFFSET 1
+#else
+#define ARG_OFFSET 0
+#endif
+
 static const int tcg_target_reg_alloc_order[] = {
     TCG_REG_L0,
     TCG_REG_L1,
@@ -86,9 +92,9 @@ static const int tcg_target_call_iarg_regs[6] = {
 
 static const int tcg_target_call_oarg_regs[] = {
     TCG_REG_O0,
-#if TCG_TARGET_REG_BITS == 32
-    TCG_REG_O1
-#endif
+    TCG_REG_O1,
+    TCG_REG_O2,
+    TCG_REG_O3,
 };
 
 static inline int check_fit_tl(tcg_target_long val, unsigned int bits)
@@ -155,6 +161,9 @@ static int target_parse_constraint(TCGArgConstraint *ct, const char **pct_str)
         tcg_regset_reset_reg(ct->u.regs, TCG_REG_O0);
         tcg_regset_reset_reg(ct->u.regs, TCG_REG_O1);
         tcg_regset_reset_reg(ct->u.regs, TCG_REG_O2);
+#ifdef CONFIG_TCG_PASS_AREG0
+        tcg_regset_reset_reg(ct->u.regs, TCG_REG_O3);
+#endif
         break;
     case 'I':
         ct->ct |= TCG_CT_CONST_S11;
@@ -706,6 +715,27 @@ static void tcg_target_qemu_prologue(TCGContext *s)
 
 #include "../../softmmu_defs.h"
 
+#ifdef CONFIG_TCG_PASS_AREG0
+/* helper signature: helper_ld_mmu(CPUState *env, target_ulong addr,
+   int mmu_idx) */
+static const void * const qemu_ld_helpers[4] = {
+    helper_ldb_mmu,
+    helper_ldw_mmu,
+    helper_ldl_mmu,
+    helper_ldq_mmu,
+};
+
+/* helper signature: helper_st_mmu(CPUState *env, target_ulong addr,
+   uintxx_t val, int mmu_idx) */
+static const void * const qemu_st_helpers[4] = {
+    helper_stb_mmu,
+    helper_stw_mmu,
+    helper_stl_mmu,
+    helper_stq_mmu,
+};
+#else
+/* legacy helper signature: __ld_mmu(target_ulong addr, int
+   mmu_idx) */
 static const void * const qemu_ld_helpers[4] = {
     __ldb_mmu,
     __ldw_mmu,
@@ -713,6 +743,8 @@ static const void * const qemu_ld_helpers[4] = {
     __ldq_mmu,
 };
 
+/* legacy helper signature: __st_mmu(target_ulong addr, uintxx_t val,
+   int mmu_idx) */
 static const void * const qemu_st_helpers[4] = {
     __stb_mmu,
     __stw_mmu,
@@ -720,6 +752,7 @@ static const void * const qemu_st_helpers[4] = {
     __stq_mmu,
 };
 #endif
+#endif
 
 #if TARGET_LONG_BITS == 32
 #define TARGET_LD_OP LDUW
@@ -801,6 +834,17 @@ static void tcg_out_qemu_ld(TCGContext *s, const TCGArg *args,
 
     /* mov */
     tcg_out_movi(s, TCG_TYPE_I32, arg1, mem_index);
+#ifdef CONFIG_TCG_PASS_AREG0
+    /* XXX/FIXME: suboptimal */
+    tcg_out_mov(s, TCG_TYPE_I32, tcg_target_call_iarg_regs[3],
+                tcg_target_call_iarg_regs[2]);
+    tcg_out_mov(s, TCG_TYPE_I64, tcg_target_call_iarg_regs[2],
+                tcg_target_call_iarg_regs[1]);
+    tcg_out_mov(s, TCG_TYPE_TL, tcg_target_call_iarg_regs[1],
+                tcg_target_call_iarg_regs[0]);
+    tcg_out_mov(s, TCG_TYPE_PTR, tcg_target_call_iarg_regs[0],
+                TCG_AREG0);
+#endif
 
     /* XXX: move that code at the end of the TB */
     /* qemu_ld_helper[s_bits](arg0, arg1) */
diff --git a/tcg/tci/tcg-target.c b/tcg/tci/tcg-target.c
index bd8507366..453f1875e 100644
--- a/tcg/tci/tcg-target.c
+++ b/tcg/tci/tcg-target.c
@@ -798,6 +798,9 @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, const TCGArg *args,
     case INDEX_op_qemu_st8:
     case INDEX_op_qemu_st16:
     case INDEX_op_qemu_st32:
+#ifdef CONFIG_TCG_PASS_AREG0
+        tcg_out_r(s, TCG_AREG0);
+#endif
         tcg_out_r(s, *args++);
         tcg_out_r(s, *args++);
 #if TARGET_LONG_BITS > TCG_TARGET_REG_BITS
@@ -808,6 +811,9 @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, const TCGArg *args,
 #endif
         break;
     case INDEX_op_qemu_st64:
+#ifdef CONFIG_TCG_PASS_AREG0
+        tcg_out_r(s, TCG_AREG0);
+#endif
         tcg_out_r(s, *args++);
 #if TCG_TARGET_REG_BITS == 32
         tcg_out_r(s, *args++);