6502/llvm-6502
1
0
Fork 0
mirror of https://github.com/c64scene-ar/llvm-6502.git synced 2024-12-16 11:30:51 +00:00
Code Issues Projects Releases Wiki Activity
87483ed180
llvm-6502/test/CodeGen/X86/stackmap-shadow-optimization.ll

29 lines
1.0 KiB
LLVM
Raw Normal View History

[Stackmaps] Make ithe frame-pointer required for stackmaps. Do not eliminate the frame pointer if there is a stackmap or patchpoint in the function. All stackmap references should be FP relative. This fixes PR21107. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@218920 91177308-0d34-0410-b5e6-96231b3b80d8
2014-10-02 22:21:49 +00:00
; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=corei7 | FileCheck %s
[X86] Optimize stackmap shadows on X86. This patch minimizes the number of nops that must be emitted on X86 to satisfy stackmap shadow constraints. To minimize the number of nops inserted, the X86AsmPrinter now records the size of the most recent stackmap's shadow in the StackMapShadowTracker class, and tracks the number of instruction bytes emitted since the that stackmap instruction was encountered. Padding is emitted (if it is required at all) immediately before the next stackmap/patchpoint instruction, or at the end of the basic block. This optimization should reduce code-size and improve performance for people using the llvm stackmap intrinsic on X86. <rdar://problem/14959522> git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213892 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-24 20:40:55 +00:00
Stackmap shadows should consider call returns a branch target. To avoid emitting too many nops, a stackmap shadow can include emitted instructions in the shadow, but these must not include branch targets. A return from a call should count as a branch target as patching over the instructions after the call would lead to incorrect behaviour for threads currently making that call, when they return. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@220710 91177308-0d34-0410-b5e6-96231b3b80d8
2014-10-27 19:40:35 +00:00
; Check that the X86 stackmap shadow optimization is only outputting a 3-byte
; nop here. 8-bytes are requested, but 5 are covered by the code for the call to
; bar. However, the frame teardown and the return do not count towards the
; stackmap shadow as the call return counts as a branch target so must flush
; the shadow.
Fix a stackmap bug introduced in r220710. For a call to not return in to the stackmap shadow, the shadow must end with the call. To do this, we must insert any required nops *before* the call, and not after it. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@220728 91177308-0d34-0410-b5e6-96231b3b80d8
2014-10-27 22:38:45 +00:00
; Note that in order for a thread to not return in to the patched space
; the call must be at the end of the shadow, so the required nop must be
; before the call, not after.
[X86] Optimize stackmap shadows on X86. This patch minimizes the number of nops that must be emitted on X86 to satisfy stackmap shadow constraints. To minimize the number of nops inserted, the X86AsmPrinter now records the size of the most recent stackmap's shadow in the StackMapShadowTracker class, and tracks the number of instruction bytes emitted since the that stackmap instruction was encountered. Padding is emitted (if it is required at all) immediately before the next stackmap/patchpoint instruction, or at the end of the basic block. This optimization should reduce code-size and improve performance for people using the llvm stackmap intrinsic on X86. <rdar://problem/14959522> git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213892 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-24 20:40:55 +00:00
define void @shadow_optimization_test() {
entry:
; CHECK-LABEL: shadow_optimization_test:
; CHECK: callq _bar
Fix a stackmap bug introduced in r220710. For a call to not return in to the stackmap shadow, the shadow must end with the call. To do this, we must insert any required nops *before* the call, and not after it. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@220728 91177308-0d34-0410-b5e6-96231b3b80d8
2014-10-27 22:38:45 +00:00
; CHECK: nop
; CHECK: callq _bar
[X86] Optimize stackmap shadows on X86. This patch minimizes the number of nops that must be emitted on X86 to satisfy stackmap shadow constraints. To minimize the number of nops inserted, the X86AsmPrinter now records the size of the most recent stackmap's shadow in the StackMapShadowTracker class, and tracks the number of instruction bytes emitted since the that stackmap instruction was encountered. Padding is emitted (if it is required at all) immediately before the next stackmap/patchpoint instruction, or at the end of the basic block. This optimization should reduce code-size and improve performance for people using the llvm stackmap intrinsic on X86. <rdar://problem/14959522> git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213892 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-24 20:40:55 +00:00
; CHECK-NOT: nop
; CHECK: callq _bar
Fix a stackmap bug introduced in r220710. For a call to not return in to the stackmap shadow, the shadow must end with the call. To do this, we must insert any required nops *before* the call, and not after it. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@220728 91177308-0d34-0410-b5e6-96231b3b80d8
2014-10-27 22:38:45 +00:00
; CHECK-NOT: nop
[X86] Optimize stackmap shadows on X86. This patch minimizes the number of nops that must be emitted on X86 to satisfy stackmap shadow constraints. To minimize the number of nops inserted, the X86AsmPrinter now records the size of the most recent stackmap's shadow in the StackMapShadowTracker class, and tracks the number of instruction bytes emitted since the that stackmap instruction was encountered. Padding is emitted (if it is required at all) immediately before the next stackmap/patchpoint instruction, or at the end of the basic block. This optimization should reduce code-size and improve performance for people using the llvm stackmap intrinsic on X86. <rdar://problem/14959522> git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213892 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-24 20:40:55 +00:00
call void @bar()
tail call void (i64, i32, ...)* @llvm.experimental.stackmap(i64 0, i32 8)
call void @bar()
Fix a stackmap bug introduced in r220710. For a call to not return in to the stackmap shadow, the shadow must end with the call. To do this, we must insert any required nops *before* the call, and not after it. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@220728 91177308-0d34-0410-b5e6-96231b3b80d8
2014-10-27 22:38:45 +00:00
call void @bar()
[X86] Optimize stackmap shadows on X86. This patch minimizes the number of nops that must be emitted on X86 to satisfy stackmap shadow constraints. To minimize the number of nops inserted, the X86AsmPrinter now records the size of the most recent stackmap's shadow in the StackMapShadowTracker class, and tracks the number of instruction bytes emitted since the that stackmap instruction was encountered. Padding is emitted (if it is required at all) immediately before the next stackmap/patchpoint instruction, or at the end of the basic block. This optimization should reduce code-size and improve performance for people using the llvm stackmap intrinsic on X86. <rdar://problem/14959522> git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213892 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-24 20:40:55 +00:00
ret void
}
declare void @llvm.experimental.stackmap(i64, i32, ...)
[Stackmaps] Make ithe frame-pointer required for stackmaps. Do not eliminate the frame pointer if there is a stackmap or patchpoint in the function. All stackmap references should be FP relative. This fixes PR21107. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@218920 91177308-0d34-0410-b5e6-96231b3b80d8
2014-10-02 22:21:49 +00:00
declare void @bar()
Reference in New Issue Copy Permalink