diff --git a/include/exec/cpu_ldst.h b/include/exec/cpu_ldst.h
index a4dad0772f..a878fd0105 100644
--- a/include/exec/cpu_ldst.h
+++ b/include/exec/cpu_ldst.h
@@ -63,6 +63,7 @@
 #define CPU_LDST_H
 
 #include "exec/memopidx.h"
+#include "qemu/int128.h"
 
 #if defined(CONFIG_USER_ONLY)
 /* sparc32plus has 64bit long but 32bit space address
@@ -233,6 +234,92 @@ void cpu_stl_le_mmu(CPUArchState *env, abi_ptr ptr, uint32_t val,
 void cpu_stq_le_mmu(CPUArchState *env, abi_ptr ptr, uint64_t val,
                     MemOpIdx oi, uintptr_t ra);
 
+uint32_t cpu_atomic_cmpxchgb_mmu(CPUArchState *env, target_ulong addr,
+                                 uint32_t cmpv, uint32_t newv,
+                                 MemOpIdx oi, uintptr_t retaddr);
+uint32_t cpu_atomic_cmpxchgw_le_mmu(CPUArchState *env, target_ulong addr,
+                                    uint32_t cmpv, uint32_t newv,
+                                    MemOpIdx oi, uintptr_t retaddr);
+uint32_t cpu_atomic_cmpxchgl_le_mmu(CPUArchState *env, target_ulong addr,
+                                    uint32_t cmpv, uint32_t newv,
+                                    MemOpIdx oi, uintptr_t retaddr);
+uint64_t cpu_atomic_cmpxchgq_le_mmu(CPUArchState *env, target_ulong addr,
+                                    uint64_t cmpv, uint64_t newv,
+                                    MemOpIdx oi, uintptr_t retaddr);
+uint32_t cpu_atomic_cmpxchgw_be_mmu(CPUArchState *env, target_ulong addr,
+                                    uint32_t cmpv, uint32_t newv,
+                                    MemOpIdx oi, uintptr_t retaddr);
+uint32_t cpu_atomic_cmpxchgl_be_mmu(CPUArchState *env, target_ulong addr,
+                                    uint32_t cmpv, uint32_t newv,
+                                    MemOpIdx oi, uintptr_t retaddr);
+uint64_t cpu_atomic_cmpxchgq_be_mmu(CPUArchState *env, target_ulong addr,
+                                    uint64_t cmpv, uint64_t newv,
+                                    MemOpIdx oi, uintptr_t retaddr);
+
+#define GEN_ATOMIC_HELPER(NAME, TYPE, SUFFIX)         \
+TYPE cpu_atomic_ ## NAME ## SUFFIX ## _mmu            \
+    (CPUArchState *env, target_ulong addr, TYPE val,  \
+     MemOpIdx oi, uintptr_t retaddr);
+
+#ifdef CONFIG_ATOMIC64
+#define GEN_ATOMIC_HELPER_ALL(NAME)          \
+    GEN_ATOMIC_HELPER(NAME, uint32_t, b)     \
+    GEN_ATOMIC_HELPER(NAME, uint32_t, w_le)  \
+    GEN_ATOMIC_HELPER(NAME, uint32_t, w_be)  \
+    GEN_ATOMIC_HELPER(NAME, uint32_t, l_le)  \
+    GEN_ATOMIC_HELPER(NAME, uint32_t, l_be)  \
+    GEN_ATOMIC_HELPER(NAME, uint64_t, q_le)  \
+    GEN_ATOMIC_HELPER(NAME, uint64_t, q_be)
+#else
+#define GEN_ATOMIC_HELPER_ALL(NAME)          \
+    GEN_ATOMIC_HELPER(NAME, uint32_t, b)     \
+    GEN_ATOMIC_HELPER(NAME, uint32_t, w_le)  \
+    GEN_ATOMIC_HELPER(NAME, uint32_t, w_be)  \
+    GEN_ATOMIC_HELPER(NAME, uint32_t, l_le)  \
+    GEN_ATOMIC_HELPER(NAME, uint32_t, l_be)
+#endif
+
+GEN_ATOMIC_HELPER_ALL(fetch_add)
+GEN_ATOMIC_HELPER_ALL(fetch_sub)
+GEN_ATOMIC_HELPER_ALL(fetch_and)
+GEN_ATOMIC_HELPER_ALL(fetch_or)
+GEN_ATOMIC_HELPER_ALL(fetch_xor)
+GEN_ATOMIC_HELPER_ALL(fetch_smin)
+GEN_ATOMIC_HELPER_ALL(fetch_umin)
+GEN_ATOMIC_HELPER_ALL(fetch_smax)
+GEN_ATOMIC_HELPER_ALL(fetch_umax)
+
+GEN_ATOMIC_HELPER_ALL(add_fetch)
+GEN_ATOMIC_HELPER_ALL(sub_fetch)
+GEN_ATOMIC_HELPER_ALL(and_fetch)
+GEN_ATOMIC_HELPER_ALL(or_fetch)
+GEN_ATOMIC_HELPER_ALL(xor_fetch)
+GEN_ATOMIC_HELPER_ALL(smin_fetch)
+GEN_ATOMIC_HELPER_ALL(umin_fetch)
+GEN_ATOMIC_HELPER_ALL(smax_fetch)
+GEN_ATOMIC_HELPER_ALL(umax_fetch)
+
+GEN_ATOMIC_HELPER_ALL(xchg)
+
+#undef GEN_ATOMIC_HELPER_ALL
+#undef GEN_ATOMIC_HELPER
+
+Int128 cpu_atomic_cmpxchgo_le_mmu(CPUArchState *env, target_ulong addr,
+                                  Int128 cmpv, Int128 newv,
+                                  MemOpIdx oi, uintptr_t retaddr);
+Int128 cpu_atomic_cmpxchgo_be_mmu(CPUArchState *env, target_ulong addr,
+                                  Int128 cmpv, Int128 newv,
+                                  MemOpIdx oi, uintptr_t retaddr);
+
+Int128 cpu_atomic_ldo_le_mmu(CPUArchState *env, target_ulong addr,
+                             MemOpIdx oi, uintptr_t retaddr);
+Int128 cpu_atomic_ldo_be_mmu(CPUArchState *env, target_ulong addr,
+                             MemOpIdx oi, uintptr_t retaddr);
+void cpu_atomic_sto_le_mmu(CPUArchState *env, target_ulong addr, Int128 val,
+                           MemOpIdx oi, uintptr_t retaddr);
+void cpu_atomic_sto_be_mmu(CPUArchState *env, target_ulong addr, Int128 val,
+                           MemOpIdx oi, uintptr_t retaddr);
+
 #if defined(CONFIG_USER_ONLY)
 
 extern __thread uintptr_t helper_retaddr;
diff --git a/include/tcg/tcg.h b/include/tcg/tcg.h
index 9f398b9afe..83e38487cf 100644
--- a/include/tcg/tcg.h
+++ b/include/tcg/tcg.h
@@ -33,7 +33,6 @@
 #include "qemu/queue.h"
 #include "tcg/tcg-mo.h"
 #include "tcg-target.h"
-#include "qemu/int128.h"
 #include "tcg/tcg-cond.h"
 
 /* XXX: make safe guess about sizes */
@@ -1312,92 +1311,6 @@ void helper_be_stq_mmu(CPUArchState *env, target_ulong addr, uint64_t val,
 #endif
 #endif /* CONFIG_SOFTMMU */
 
-uint32_t cpu_atomic_cmpxchgb_mmu(CPUArchState *env, target_ulong addr,
-                                 uint32_t cmpv, uint32_t newv,
-                                 MemOpIdx oi, uintptr_t retaddr);
-uint32_t cpu_atomic_cmpxchgw_le_mmu(CPUArchState *env, target_ulong addr,
-                                    uint32_t cmpv, uint32_t newv,
-                                    MemOpIdx oi, uintptr_t retaddr);
-uint32_t cpu_atomic_cmpxchgl_le_mmu(CPUArchState *env, target_ulong addr,
-                                    uint32_t cmpv, uint32_t newv,
-                                    MemOpIdx oi, uintptr_t retaddr);
-uint64_t cpu_atomic_cmpxchgq_le_mmu(CPUArchState *env, target_ulong addr,
-                                    uint64_t cmpv, uint64_t newv,
-                                    MemOpIdx oi, uintptr_t retaddr);
-uint32_t cpu_atomic_cmpxchgw_be_mmu(CPUArchState *env, target_ulong addr,
-                                    uint32_t cmpv, uint32_t newv,
-                                    MemOpIdx oi, uintptr_t retaddr);
-uint32_t cpu_atomic_cmpxchgl_be_mmu(CPUArchState *env, target_ulong addr,
-                                    uint32_t cmpv, uint32_t newv,
-                                    MemOpIdx oi, uintptr_t retaddr);
-uint64_t cpu_atomic_cmpxchgq_be_mmu(CPUArchState *env, target_ulong addr,
-                                    uint64_t cmpv, uint64_t newv,
-                                    MemOpIdx oi, uintptr_t retaddr);
-
-#define GEN_ATOMIC_HELPER(NAME, TYPE, SUFFIX)         \
-TYPE cpu_atomic_ ## NAME ## SUFFIX ## _mmu            \
-    (CPUArchState *env, target_ulong addr, TYPE val,  \
-     MemOpIdx oi, uintptr_t retaddr);
-
-#ifdef CONFIG_ATOMIC64
-#define GEN_ATOMIC_HELPER_ALL(NAME)          \
-    GEN_ATOMIC_HELPER(NAME, uint32_t, b)     \
-    GEN_ATOMIC_HELPER(NAME, uint32_t, w_le)  \
-    GEN_ATOMIC_HELPER(NAME, uint32_t, w_be)  \
-    GEN_ATOMIC_HELPER(NAME, uint32_t, l_le)  \
-    GEN_ATOMIC_HELPER(NAME, uint32_t, l_be)  \
-    GEN_ATOMIC_HELPER(NAME, uint64_t, q_le)  \
-    GEN_ATOMIC_HELPER(NAME, uint64_t, q_be)
-#else
-#define GEN_ATOMIC_HELPER_ALL(NAME)          \
-    GEN_ATOMIC_HELPER(NAME, uint32_t, b)     \
-    GEN_ATOMIC_HELPER(NAME, uint32_t, w_le)  \
-    GEN_ATOMIC_HELPER(NAME, uint32_t, w_be)  \
-    GEN_ATOMIC_HELPER(NAME, uint32_t, l_le)  \
-    GEN_ATOMIC_HELPER(NAME, uint32_t, l_be)
-#endif
-
-GEN_ATOMIC_HELPER_ALL(fetch_add)
-GEN_ATOMIC_HELPER_ALL(fetch_sub)
-GEN_ATOMIC_HELPER_ALL(fetch_and)
-GEN_ATOMIC_HELPER_ALL(fetch_or)
-GEN_ATOMIC_HELPER_ALL(fetch_xor)
-GEN_ATOMIC_HELPER_ALL(fetch_smin)
-GEN_ATOMIC_HELPER_ALL(fetch_umin)
-GEN_ATOMIC_HELPER_ALL(fetch_smax)
-GEN_ATOMIC_HELPER_ALL(fetch_umax)
-
-GEN_ATOMIC_HELPER_ALL(add_fetch)
-GEN_ATOMIC_HELPER_ALL(sub_fetch)
-GEN_ATOMIC_HELPER_ALL(and_fetch)
-GEN_ATOMIC_HELPER_ALL(or_fetch)
-GEN_ATOMIC_HELPER_ALL(xor_fetch)
-GEN_ATOMIC_HELPER_ALL(smin_fetch)
-GEN_ATOMIC_HELPER_ALL(umin_fetch)
-GEN_ATOMIC_HELPER_ALL(smax_fetch)
-GEN_ATOMIC_HELPER_ALL(umax_fetch)
-
-GEN_ATOMIC_HELPER_ALL(xchg)
-
-#undef GEN_ATOMIC_HELPER_ALL
-#undef GEN_ATOMIC_HELPER
-
-Int128 cpu_atomic_cmpxchgo_le_mmu(CPUArchState *env, target_ulong addr,
-                                  Int128 cmpv, Int128 newv,
-                                  MemOpIdx oi, uintptr_t retaddr);
-Int128 cpu_atomic_cmpxchgo_be_mmu(CPUArchState *env, target_ulong addr,
-                                  Int128 cmpv, Int128 newv,
-                                  MemOpIdx oi, uintptr_t retaddr);
-
-Int128 cpu_atomic_ldo_le_mmu(CPUArchState *env, target_ulong addr,
-                             MemOpIdx oi, uintptr_t retaddr);
-Int128 cpu_atomic_ldo_be_mmu(CPUArchState *env, target_ulong addr,
-                             MemOpIdx oi, uintptr_t retaddr);
-void cpu_atomic_sto_le_mmu(CPUArchState *env, target_ulong addr, Int128 val,
-                           MemOpIdx oi, uintptr_t retaddr);
-void cpu_atomic_sto_be_mmu(CPUArchState *env, target_ulong addr, Int128 val,
-                           MemOpIdx oi, uintptr_t retaddr);
-
 #ifdef CONFIG_DEBUG_TCG
 void tcg_assert_listed_vecop(TCGOpcode);
 #else
diff --git a/target/arm/helper-a64.c b/target/arm/helper-a64.c
index 4cafd3c11a..b110c57956 100644
--- a/target/arm/helper-a64.c
+++ b/target/arm/helper-a64.c
@@ -32,7 +32,6 @@
 #include "exec/cpu_ldst.h"
 #include "qemu/int128.h"
 #include "qemu/atomic128.h"
-#include "tcg/tcg.h"
 #include "fpu/softfloat.h"
 #include <zlib.h> /* For crc32 */
 
diff --git a/target/m68k/op_helper.c b/target/m68k/op_helper.c
index c1bf73b6f9..cfbc987ba6 100644
--- a/target/m68k/op_helper.c
+++ b/target/m68k/op_helper.c
@@ -22,7 +22,6 @@
 #include "exec/exec-all.h"
 #include "exec/cpu_ldst.h"
 #include "semihosting/semihost.h"
-#include "tcg/tcg.h"
 
 #if !defined(CONFIG_USER_ONLY)
 
diff --git a/target/ppc/mem_helper.c b/target/ppc/mem_helper.c
index e2282baa8d..39945d9ea5 100644
--- a/target/ppc/mem_helper.c
+++ b/target/ppc/mem_helper.c
@@ -25,7 +25,6 @@
 #include "exec/helper-proto.h"
 #include "helper_regs.h"
 #include "exec/cpu_ldst.h"
-#include "tcg/tcg.h"
 #include "internal.h"
 #include "qemu/atomic128.h"
 
diff --git a/target/s390x/tcg/mem_helper.c b/target/s390x/tcg/mem_helper.c
index e64d1bc725..251d4acf55 100644
--- a/target/s390x/tcg/mem_helper.c
+++ b/target/s390x/tcg/mem_helper.c
@@ -27,7 +27,6 @@
 #include "exec/cpu_ldst.h"
 #include "qemu/int128.h"
 #include "qemu/atomic128.h"
-#include "tcg/tcg.h"
 #include "trace.h"
 
 #if !defined(CONFIG_USER_ONLY)