diff --git a/include/exec/exec-all.h b/include/exec/exec-all.h
index 6c095e8636..30cdd699f4 100644
--- a/include/exec/exec-all.h
+++ b/include/exec/exec-all.h
@@ -328,29 +328,8 @@ static inline void tb_set_jmp_target1(uintptr_t jmp_addr, uintptr_t addr)
 void aarch64_tb_set_jmp_target(uintptr_t jmp_addr, uintptr_t addr);
 #define tb_set_jmp_target1 aarch64_tb_set_jmp_target
 #elif defined(__arm__)
-static inline void tb_set_jmp_target1(uintptr_t jmp_addr, uintptr_t addr)
-{
-#if !QEMU_GNUC_PREREQ(4, 1)
-    register unsigned long _beg __asm ("a1");
-    register unsigned long _end __asm ("a2");
-    register unsigned long _flg __asm ("a3");
-#endif
-
-    /* we could use a ldr pc, [pc, #-4] kind of branch and avoid the flush */
-    *(uint32_t *)jmp_addr =
-        (*(uint32_t *)jmp_addr & ~0xffffff)
-        | (((addr - (jmp_addr + 8)) >> 2) & 0xffffff);
-
-#if QEMU_GNUC_PREREQ(4, 1)
-    __builtin___clear_cache((char *) jmp_addr, (char *) jmp_addr + 4);
-#else
-    /* flush icache */
-    _beg = jmp_addr;
-    _end = jmp_addr + 4;
-    _flg = 0;
-    __asm __volatile__ ("swi 0x9f0002" : : "r" (_beg), "r" (_end), "r" (_flg));
-#endif
-}
+void arm_tb_set_jmp_target(uintptr_t jmp_addr, uintptr_t addr);
+#define tb_set_jmp_target1 arm_tb_set_jmp_target
 #elif defined(__sparc__) || defined(__mips__)
 void tb_set_jmp_target1(uintptr_t jmp_addr, uintptr_t addr);
 #else
diff --git a/tcg/arm/tcg-target.inc.c b/tcg/arm/tcg-target.inc.c
index 2b7fbddbf0..977baa05c5 100644
--- a/tcg/arm/tcg-target.inc.c
+++ b/tcg/arm/tcg-target.inc.c
@@ -121,6 +121,14 @@ static inline void reloc_pc24(tcg_insn_unit *code_ptr, tcg_insn_unit *target)
     *code_ptr = (*code_ptr & ~0xffffff) | (offset & 0xffffff);
 }
 
+static inline void reloc_pc24_atomic(tcg_insn_unit *code_ptr, tcg_insn_unit *target)
+{
+    ptrdiff_t offset = (tcg_ptr_byte_diff(target, code_ptr) - 8) >> 2;
+    tcg_insn_unit insn = atomic_read(code_ptr);
+    tcg_debug_assert(offset == sextract32(offset, 0, 24));
+    atomic_set(code_ptr, deposit32(insn, 0, 24, offset));
+}
+
 static void patch_reloc(tcg_insn_unit *code_ptr, int type,
                         intptr_t value, intptr_t addend)
 {
@@ -1038,6 +1046,16 @@ static void tcg_out_call(TCGContext *s, tcg_insn_unit *addr)
     }
 }
 
+void arm_tb_set_jmp_target(uintptr_t jmp_addr, uintptr_t addr)
+{
+    tcg_insn_unit *code_ptr = (tcg_insn_unit *)jmp_addr;
+    tcg_insn_unit *target = (tcg_insn_unit *)addr;
+
+    /* we could use a ldr pc, [pc, #-4] kind of branch and avoid the flush */
+    reloc_pc24_atomic(code_ptr, target);
+    flush_icache_range(jmp_addr, jmp_addr + 4);
+}
+
 static inline void tcg_out_goto_label(TCGContext *s, int cond, TCGLabel *l)
 {
     if (l->has_value) {