[AArch64] Fix variadic tail-calls on ARM64EC #79774
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters. Learn more about bidirectional Unicode characters
Member
| @llvm/pr-subscribers-backend-aarch64 Author: Billy Laws (bylaws) ChangesARM64EC varargs calls expect that x4 = sp at entry, special handling is needed to ensure this with tail calls since they occur after the epilogue and the x4 write happens before. Full diff: https://github.com/llvm/llvm-project/pull/79774.diff 2 Files Affected:
diff --git a/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp b/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp index 332fb37655288ce..18ff1167e1679c7 100644 --- a/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp +++ b/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp @@ -7991,11 +7991,22 @@ AArch64TargetLowering::LowerCall(CallLoweringInfo &CLI, } if (IsVarArg && Subtarget->isWindowsArm64EC()) { + SDValue ParamPtr = StackPtr; + if (IsTailCall) { + const AArch64FrameLowering *TFI = + MF.getSubtarget<AArch64Subtarget>().getFrameLowering(); + + // Create a dummy object at the top of the stack that can be used to get + // the SP after the epilogue + int FI = MF.getFrameInfo().CreateFixedObject(1, FPDiff, true); + ParamPtr = DAG.getFrameIndex(FI, PtrVT); + } + // For vararg calls, the Arm64EC ABI requires values in x4 and x5 // describing the argument list. x4 contains the address of the // first stack parameter. x5 contains the size in bytes of all parameters // passed on the stack. - RegsToPass.emplace_back(AArch64::X4, StackPtr); + RegsToPass.emplace_back(AArch64::X4, ParamPtr); RegsToPass.emplace_back(AArch64::X5, DAG.getConstant(NumBytes, DL, MVT::i64)); } diff --git a/llvm/test/CodeGen/AArch64/vararg-tallcall.ll b/llvm/test/CodeGen/AArch64/vararg-tallcall.ll index 2d6db1642247d72..812837639196e61 100644 --- a/llvm/test/CodeGen/AArch64/vararg-tallcall.ll +++ b/llvm/test/CodeGen/AArch64/vararg-tallcall.ll @@ -1,5 +1,6 @@ ; RUN: llc -mtriple=aarch64-windows-msvc %s -o - | FileCheck %s ; RUN: llc -mtriple=aarch64-linux-gnu %s -o - | FileCheck %s +; RUN: llc -mtriple=arm64ec-windows-msvc %s -o - | FileCheck %s --check-prefixes=CHECK-EC ; RUN: llc -global-isel -global-isel-abort=2 -verify-machineinstrs -mtriple=aarch64-windows-msvc %s -o - | FileCheck %s ; RUN: llc -global-isel -global-isel-abort=2 -verify-machineinstrs -mtriple=aarch64-linux-gnu %s -o - | FileCheck %s @@ -32,3 +33,10 @@ attributes #1 = { noinline optnone "thunk" } ; CHECK: ldr x9, [x9] ; CHECK: mov v0.16b, v16.16b ; CHECK: br x9 +; CHECK-EC: mov v7.16b, v0.16b +; CHECK-EC: ldr x9, [x0] +; CHECK-EC: ldr x11, [x9] +; CHECK-EC: mov v0.16b, v7.16b +; CHECK-EC: add x4, sp, #64 +; CHECK-EC: add sp, sp, #64 +; CHECK-EC: br x11 |
Collaborator
efriedma-quic left a comment
efriedma-quic left a comment There was a problem hiding this comment.
The "1-byte" fixed object is maybe a little weird, but shouldn't cause any practical issue; really, it's a pointer to the end of the shadow, but that would just be more confusing.
I'd like to see a regression test that includes an opportunistic sibcall, in addition to the musttail call test. (Maybe makes sense to test in a separate file, to make it easier to find stuff related to EC calling convention.)
ARM64EC varargs calls expect that x4 = sp at entry, special handling is needed to ensure this with tail calls since they occur after the epilogue and the X4 write happens before.
Contributor Author
| Addressed |
llvmbot pushed a commit to llvmbot/llvm-project that referenced this pull request Feb 5, 2024
ARM64EC varargs calls expect that x4 = sp at entry, special handling is needed to ensure this with tail calls since they occur after the epilogue and the x4 write happens before. I tried going through AArch64MachineFrameLowering for this, hoping to avoid creating the dummy object but this was the best I could do since the stack info that uses isn't populated at this stage, CreateFixedObject also explicitly forbids 0 sized objects. (cherry picked from commit c761b4a)
dpaoliello pushed a commit to dpaoliello/llvm-project that referenced this pull request Feb 14, 2024
ARM64EC varargs calls expect that x4 = sp at entry, special handling is needed to ensure this with tail calls since they occur after the epilogue and the x4 write happens before. I tried going through AArch64MachineFrameLowering for this, hoping to avoid creating the dummy object but this was the best I could do since the stack info that uses isn't populated at this stage, CreateFixedObject also explicitly forbids 0 sized objects.
dpaoliello pushed a commit to dpaoliello/llvm-project that referenced this pull request Feb 27, 2024
ARM64EC varargs calls expect that x4 = sp at entry, special handling is needed to ensure this with tail calls since they occur after the epilogue and the x4 write happens before. I tried going through AArch64MachineFrameLowering for this, hoping to avoid creating the dummy object but this was the best I could do since the stack info that uses isn't populated at this stage, CreateFixedObject also explicitly forbids 0 sized objects.
dpaoliello pushed a commit to dpaoliello/llvm-project that referenced this pull request Mar 8, 2024
ARM64EC varargs calls expect that x4 = sp at entry, special handling is needed to ensure this with tail calls since they occur after the epilogue and the x4 write happens before. I tried going through AArch64MachineFrameLowering for this, hoping to avoid creating the dummy object but this was the best I could do since the stack info that uses isn't populated at this stage, CreateFixedObject also explicitly forbids 0 sized objects.
dpaoliello pushed a commit to dpaoliello/llvm-project that referenced this pull request Mar 11, 2024
ARM64EC varargs calls expect that x4 = sp at entry, special handling is needed to ensure this with tail calls since they occur after the epilogue and the x4 write happens before. I tried going through AArch64MachineFrameLowering for this, hoping to avoid creating the dummy object but this was the best I could do since the stack info that uses isn't populated at this stage, CreateFixedObject also explicitly forbids 0 sized objects.
Closed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit. This suggestion is invalid because no changes were made to the code. Suggestions cannot be applied while the pull request is closed. Suggestions cannot be applied while viewing a subset of changes. Only one suggestion per line can be applied in a batch. Add this suggestion to a batch that can be applied as a single commit. Applying suggestions on deleted lines is not supported. You must change the existing code in this line in order to create a valid suggestion. Outdated suggestions cannot be applied. This suggestion has been applied or marked resolved. Suggestions cannot be applied from pending reviews. Suggestions cannot be applied on multi-line comments. Suggestions cannot be applied while the pull request is queued to merge. Suggestion cannot be applied right now. Please check back later.
ARM64EC varargs calls expect that x4 = sp at entry, special handling is needed to ensure this with tail calls since they occur after the epilogue and the x4 write happens before.
I tried going through AArch64MachineFrameLowering for this, hoping to avoid creating the dummy object but this was the best I could do since the stack info that uses isn't populated at this stage, CreateFixedObject also explicitly forbids 0 sized objects.
cc: @efriedma-quic