From 9b751356a74577baa8be2d1bdbb41cf1e7208573 Mon Sep 17 00:00:00 2001
From: Frode Solheim <frode@fs-uae.net>
Date: Sun, 6 Sep 2015 22:13:27 +0200
Subject: [PATCH] JIT: use align_target(align_jumps)

---
 jit/codegen_x86.cpp     |  3 +++
 jit/compemu_support.cpp | 27 ++++++++++++++++++---------
 2 files changed, 21 insertions(+), 9 deletions(-)

diff --git a/jit/codegen_x86.cpp b/jit/codegen_x86.cpp
index bef0ebb2..5a38cc2e 100644
--- a/jit/codegen_x86.cpp
+++ b/jit/codegen_x86.cpp
@@ -3402,7 +3402,10 @@ raw_init_cpu(void)
 	/* Alignments */
 	if (tune_alignment) {
 		align_loops = x86_alignments[c->x86_processor].align_loop;
+#ifdef UAE
+#else
 		align_jumps = x86_alignments[c->x86_processor].align_jump;
+#endif
 	}
 	{ 
 		TCHAR *s = au (c->x86_vendor_id);
diff --git a/jit/compemu_support.cpp b/jit/compemu_support.cpp
index ff618b60..8ef35a4b 100644
--- a/jit/compemu_support.cpp
+++ b/jit/compemu_support.cpp
@@ -3366,6 +3366,15 @@ static inline void create_popalls(void)
 	current_compile_p=popallspace;
 	set_target(current_compile_p);
 
+	/* We need to guarantee 16-byte stack alignment on x86 at any point
+	   within the JIT generated code. We have multiple exit points
+	   possible but a single entry. A "jmp" is used so that we don't
+	   have to generate stack alignment in generated code that has to
+	   call external functions (e.g. a generic instruction handler).
+
+	   In summary, JIT generated code is not leaf so we have to deal
+	   with it here to maintain correct stack alignment. */
+	align_target(align_jumps);
 	current_compile_p=get_target();
 	pushall_call_handler=get_target();
 	for (i=N_REGS;i--;) {
@@ -3378,47 +3387,47 @@ static inline void create_popalls(void)
 	raw_jmp_m_indexed((uintptr)cache_tags,r,SIZEOF_VOID_P);
 
 	/* now the exit points */
-	align_target(32);
+	align_target(align_jumps);
 	popall_do_nothing=get_target();
 	for (i=0;i<N_REGS;i++) {
 		if (need_to_preserve[i])
 			raw_pop_l_r(i);
 	}
 	raw_jmp((uintptr)do_nothing);
-	align_target(32);
 
+	align_target(align_jumps);
 	popall_execute_normal=get_target();
 	for (i=0;i<N_REGS;i++) {
 		if (need_to_preserve[i])
 			raw_pop_l_r(i);
 	}
 	raw_jmp((uintptr)execute_normal);
-	align_target(32);
 
+	align_target(align_jumps);
 	popall_cache_miss=get_target();
 	for (i=0;i<N_REGS;i++) {
 		if (need_to_preserve[i])
 			raw_pop_l_r(i);
 	}
 	raw_jmp((uintptr)cache_miss);
-	align_target(32);
 
+	align_target(align_jumps);
 	popall_recompile_block=get_target();
 	for (i=0;i<N_REGS;i++) {
 		if (need_to_preserve[i])
 			raw_pop_l_r(i);
 	}
 	raw_jmp((uintptr)recompile_block);
-	align_target(32);
 
+	align_target(align_jumps);
 	popall_exec_nostats=get_target();
 	for (i=0;i<N_REGS;i++) {
 		if (need_to_preserve[i])
 			raw_pop_l_r(i);
 	}
 	raw_jmp((uintptr)exec_nostats);
-	align_target(32);
 
+	align_target(align_jumps);
 	popall_check_checksum=get_target();
 	for (i=0;i<N_REGS;i++) {
 		if (need_to_preserve[i])
@@ -3450,13 +3459,13 @@ static void prepare_block(blockinfo* bi)
 	int i;
 
 	set_target(current_compile_p);
-	align_target(32);
+	align_target(align_jumps);
 	bi->direct_pen=(cpuop_func*)get_target();
 	raw_mov_l_rm(0,(uintptr)&(bi->pc_p));
 	raw_mov_l_mr((uintptr)&regs.pc_p,0);
 	raw_jmp((uintptr)popall_execute_normal);
 
-	align_target(32);
+	align_target(align_jumps);
 	bi->direct_pcc=(cpuop_func*)get_target();
 	raw_mov_l_rm(0,(uintptr)&(bi->pc_p));
 	raw_mov_l_mr((uintptr)&regs.pc_p,0);
@@ -4288,7 +4297,7 @@ void compile_block(cpu_history* pc_hist, int blocklen, int totcycles)
 #endif
 
 		log_dump();
-		align_target(32);
+		align_target(align_jumps);
 		current_compile_p=get_target();
 
 		raise_in_cl_list(bi);
-- 
2.47.3