[llvm] [AArch64] Stack probing for function prologues (PR #66524)
Momchil Velikov via llvm-commits
llvm-commits at lists.llvm.org
Fri Sep 15 09:11:52 PDT 2023
https://github.com/momchil-velikov created https://github.com/llvm/llvm-project/pull/66524
This adds code to AArch64 function prologues to protect against stack clash attacks by probing (writing to) the stack at regular enough intervals to ensure that the guard page cannot be skipped over.
The patch depends on and maintains the following invariants:
Upon function entry the caller guarantees that it has probed the stack (e.g. performed a store) at some address [sp, #N], where`0 <= N <= 1024`. This invariant comes from a requirement for compatibility with GCC. Any address range in the allocated stack, no smaller than stack-probe-size bytes contains at least one probe At any time the stack pointer is above or in the guard page Probes are performed in descreasing address order
The stack-probe-size is a function attribute that can be set by a platform to correspond to the guard page size.
By default, the stack probe size is 4KiB, which is a safe default as this is the smallest possible page size for AArch64. Linux uses a 64KiB guard for AArch64, so this can be overridden by the stack-probe-size function attribute.
For small frames without a frame pointer (<= 240 bytes), no probes are needed.
For larger frame sizes, LLVM always stores x29 to the stack. This serves as an implicit stack probe. Thus, while allocating stack objects the compiler assumes that the stack has been probed at [sp].
There are multiple probing sequences that can be emitted, depending on the size of the stack allocation:
A straight-line sequence of subtracts and stores, used when the allocation size is smaller than 5 guard pages. A loop allocating and probing one page size per iteration, plus at most a single probe to deal with the remainder, used when the allocation size is larger but still known at compile time. A loop which moves the SP down to the target value held in a register (or a loop, moving a scratch register to the target value help in SP), used when the allocation size is not known at compile-time, such as when allocating space for SVE values, or when over-aligning the stack. This is emitted in AArch64InstrInfo because it will also be used for dynamic allocas in a future patch. A single probe where the amount of stack adjustment is unknown, but is known to be less than or equal to a page size.
>From 5e25a2ff47e1838e4bca5ec6c87cc72b9491d6f3 Mon Sep 17 00:00:00 2001
From: Momchil Velikov <momchil.velikov at arm.com>
Date: Fri, 15 Sep 2023 12:46:44 +0100
Subject: [PATCH] [AArch64] Stack probing for function prologues
This adds code to AArch64 function prologues to protect against stack
clash attacks by probing (writing to) the stack at regular enough
intervals to ensure that the guard page cannot be skipped over.
The patch depends on and maintains the following invariants:
Upon function entry the caller guarantees that it has probed the stack (e.g. performed a store) at some address [sp, #N], where`0 <= N <= 1024`. This invariant comes from a requirement for compatibility with GCC.
Any address range in the allocated stack, no smaller than stack-probe-size bytes contains at least one probe
At any time the stack pointer is above or in the guard page
Probes are performed in descreasing address order
The stack-probe-size is a function attribute that can be set by
a platform to correspond to the guard page size.
By default, the stack probe size is 4KiB, which is a safe default as
this is the smallest possible page size for AArch64. Linux uses a
64KiB guard for AArch64, so this can be overridden by the
stack-probe-size function attribute.
For small frames without a frame pointer (<= 240 bytes), no probes are
needed.
For larger frame sizes, LLVM always stores x29 to the stack. This
serves as an implicit stack probe. Thus, while allocating stack
objects the compiler assumes that the stack has been probed at [sp].
There are multiple probing sequences that can be emitted, depending on
the size of the stack allocation:
A straight-line sequence of subtracts and stores, used when the allocation size is smaller than 5 guard pages.
A loop allocating and probing one page size per iteration, plus at most a single probe to deal with the remainder, used when the allocation size is larger but still known at compile time.
A loop which moves the SP down to the target value held in a register (or a loop, moving a scratch register to the target value help in SP), used when the allocation size is not known at compile-time, such as when allocating space for SVE values, or when over-aligning the stack. This is emitted in AArch64InstrInfo because it will also be used for dynamic allocas in a future patch.
A single probe where the amount of stack adjustment is unknown, but is known to be less than or equal to a page size.
Change-Id: Ib31bc23d2806835bc6da822599f2bd2ecc18e4bf
Co-authored-by: Oliver Stannard <oliver.stannard at linaro.org>
---
.../Target/AArch64/AArch64FrameLowering.cpp | 392 ++++++++++-
.../lib/Target/AArch64/AArch64FrameLowering.h | 23 +
.../Target/AArch64/AArch64ISelLowering.cpp | 34 +
llvm/lib/Target/AArch64/AArch64ISelLowering.h | 14 +
llvm/lib/Target/AArch64/AArch64InstrInfo.cpp | 93 ++-
llvm/lib/Target/AArch64/AArch64InstrInfo.h | 5 +
llvm/lib/Target/AArch64/AArch64InstrInfo.td | 21 +-
llvm/test/CodeGen/AArch64/framelayout-sve.mir | 4 +-
.../AArch64/spill-stack-realignment.mir | 2 +-
.../test/CodeGen/AArch64/stack-probing-64k.ll | 392 +++++++++++
.../test/CodeGen/AArch64/stack-probing-sve.ll | 661 ++++++++++++++++++
llvm/test/CodeGen/AArch64/stack-probing.ll | 474 +++++++++++++
12 files changed, 2078 insertions(+), 37 deletions(-)
create mode 100644 llvm/test/CodeGen/AArch64/stack-probing-64k.ll
create mode 100644 llvm/test/CodeGen/AArch64/stack-probing-sve.ll
create mode 100644 llvm/test/CodeGen/AArch64/stack-probing.ll
diff --git a/llvm/lib/Target/AArch64/AArch64FrameLowering.cpp b/llvm/lib/Target/AArch64/AArch64FrameLowering.cpp
index 68e68449d4073b2..2bf61bdddf3d511 100644
--- a/llvm/lib/Target/AArch64/AArch64FrameLowering.cpp
+++ b/llvm/lib/Target/AArch64/AArch64FrameLowering.cpp
@@ -301,6 +301,7 @@ static bool produceCompactUnwindFrame(MachineFunction &MF);
static bool needsWinCFI(const MachineFunction &MF);
static StackOffset getSVEStackSize(const MachineFunction &MF);
static bool needsShadowCallStackPrologueEpilogue(MachineFunction &MF);
+static unsigned findScratchNonCalleeSaveRegister(MachineBasicBlock *MBB);
/// Returns true if a homogeneous prolog or epilog code can be emitted
/// for the size optimization. If possible, a frame helper call is injected.
@@ -672,6 +673,74 @@ void AArch64FrameLowering::emitCalleeSavedSVERestores(
emitCalleeSavedRestores(MBB, MBBI, true);
}
+void AArch64FrameLowering::allocateSVEStackSpace(
+ MachineBasicBlock &MBB, MachineBasicBlock::iterator MBBI,
+ StackOffset AllocSize, StackOffset InitialOffset, bool EmitCFI) const {
+ DebugLoc DL;
+ MachineFunction &MF = *MBB.getParent();
+ const AArch64Subtarget &Subtarget = MF.getSubtarget<AArch64Subtarget>();
+ const AArch64RegisterInfo &RegInfo = *Subtarget.getRegisterInfo();
+ const AArch64TargetLowering &TLI = *Subtarget.getTargetLowering();
+ const TargetInstrInfo &TII = *Subtarget.getInstrInfo();
+
+ // If not probing the stack or the (uknown) allocation size is less than the
+ // probe size decrement the stack pointer right away. This avoids having to
+ // emit a probing loop when allocating space for up to 16 SVE registers when
+ // using 4k probes.
+
+ // The bit-length of SVE registers is architecturally limited.
+ const int64_t MAX_BYTES_PER_SCALABLE_BYTE = 16;
+ int64_t ProbeSize = TLI.getStackProbeSize(MF);
+ if (!TLI.hasInlineStackProbe(MF) ||
+ AllocSize.getScalable() * MAX_BYTES_PER_SCALABLE_BYTE +
+ AllocSize.getFixed() <=
+ ProbeSize) {
+ emitFrameOffset(MBB, MBBI, DL, AArch64::SP, AArch64::SP, -AllocSize, &TII,
+ MachineInstr::FrameSetup, false, false, nullptr, EmitCFI,
+ InitialOffset);
+ if (TLI.hasInlineStackProbe(MF)) {
+ // Issue a probe at the top of the stack to prepare for subsequent
+ // allocations.
+ // STR XZR, [TargetReg]
+ BuildMI(MBB, MBBI, DL, TII.get(AArch64::STRXui))
+ .addReg(AArch64::XZR)
+ .addReg(AArch64::SP)
+ .addImm(0)
+ .setMIFlags(MachineInstr::FrameSetup);
+ }
+ return;
+ }
+
+ // If we can't be sure the allocation size if less than the probe size, we
+ // have to emit a stack probing loop.
+ Register ScratchReg = findScratchNonCalleeSaveRegister(&MBB);
+ assert(ScratchReg != AArch64::NoRegister);
+ // Get the new top of the stack into a scratch register.
+ emitFrameOffset(MBB, MBBI, DL, ScratchReg, AArch64::SP, -AllocSize, &TII,
+ MachineInstr::FrameSetup, false, false, nullptr, EmitCFI,
+ InitialOffset);
+ // Arrange to emit a probing loop by decrementing SP until it reaches that
+ // new top of the stack.
+ BuildMI(MBB, MBBI, DL, TII.get(AArch64::PROBED_STACKALLOC_VAR), AArch64::SP)
+ .addReg(ScratchReg);
+ // Set SP to its new value.
+ // MOV SP, Xs
+ BuildMI(MBB, MBBI, DL, TII.get(AArch64::ADDXri), AArch64::SP)
+ .addReg(ScratchReg)
+ .addImm(0)
+ .addImm(AArch64_AM::getShifterImm(AArch64_AM::LSL, 0))
+ .setMIFlags(MachineInstr::FrameSetup);
+ if (EmitCFI) {
+ // Set the CFA register back to SP.
+ unsigned Reg = RegInfo.getDwarfRegNum(AArch64::SP, true);
+ unsigned CFIIndex =
+ MF.addFrameInst(MCCFIInstruction::createDefCfaRegister(nullptr, Reg));
+ BuildMI(MBB, MBBI, DL, TII.get(TargetOpcode::CFI_INSTRUCTION))
+ .addCFIIndex(CFIIndex)
+ .setMIFlags(MachineInstr::FrameSetup);
+ }
+}
+
static MCRegister getRegisterOrZero(MCRegister Reg, bool HasSVE) {
switch (Reg.id()) {
default:
@@ -855,9 +924,11 @@ bool AArch64FrameLowering::canUseAsPrologue(
MachineBasicBlock *TmpMBB = const_cast<MachineBasicBlock *>(&MBB);
const AArch64Subtarget &Subtarget = MF->getSubtarget<AArch64Subtarget>();
const AArch64RegisterInfo *RegInfo = Subtarget.getRegisterInfo();
+ const AArch64TargetLowering *TLI = Subtarget.getTargetLowering();
- // Don't need a scratch register if we're not going to re-align the stack.
- if (!RegInfo->hasStackRealignment(*MF))
+ // Don't need a scratch register if we're not going to re-align the stack or
+ // emit stack probes.
+ if (!RegInfo->hasStackRealignment(*MF) && TLI->hasInlineStackProbe(*MF))
return true;
// Otherwise, we can use any block as long as it has a scratch register
// available.
@@ -1429,6 +1500,7 @@ void AArch64FrameLowering::emitPrologue(MachineFunction &MF,
const Function &F = MF.getFunction();
const AArch64Subtarget &Subtarget = MF.getSubtarget<AArch64Subtarget>();
const AArch64RegisterInfo *RegInfo = Subtarget.getRegisterInfo();
+ const AArch64TargetLowering &TLI = *Subtarget.getTargetLowering();
const TargetInstrInfo *TII = Subtarget.getInstrInfo();
MachineModuleInfo &MMI = MF.getMMI();
AArch64FunctionInfo *AFI = MF.getInfo<AArch64FunctionInfo>();
@@ -1784,12 +1856,14 @@ void AArch64FrameLowering::emitPrologue(MachineFunction &MF,
}
}
- StackOffset AllocateBefore = SVEStackSize, AllocateAfter = {};
+ StackOffset SVECalleeSavedSize = {}, SVELocalsSize = SVEStackSize;
MachineBasicBlock::iterator CalleeSavesBegin = MBBI, CalleeSavesEnd = MBBI;
// Process the SVE callee-saves to determine what space needs to be
// allocated.
if (int64_t CalleeSavedSize = AFI->getSVECalleeSavedStackSize()) {
+ LLVM_DEBUG(dbgs() << "SVECalleeSavedStackSize = " << CalleeSavedSize
+ << "\n");
// Find callee save instructions in frame.
CalleeSavesBegin = MBBI;
assert(IsSVECalleeSave(CalleeSavesBegin) && "Unexpected instruction");
@@ -1797,33 +1871,40 @@ void AArch64FrameLowering::emitPrologue(MachineFunction &MF,
++MBBI;
CalleeSavesEnd = MBBI;
- AllocateBefore = StackOffset::getScalable(CalleeSavedSize);
- AllocateAfter = SVEStackSize - AllocateBefore;
+ SVECalleeSavedSize = StackOffset::getScalable(CalleeSavedSize);
+ SVELocalsSize = SVEStackSize - SVECalleeSavedSize;
+
+ // Allocate space for the SVE callee saves.
+ if (SVECalleeSavedSize) {
+ allocateSVEStackSpace(
+ MBB, CalleeSavesBegin, SVECalleeSavedSize,
+ StackOffset::getFixed((int64_t)MFI.getStackSize() - NumBytes),
+ EmitAsyncCFI && !HasFP);
+ if (EmitAsyncCFI)
+ emitCalleeSavedSVELocations(MBB, CalleeSavesEnd);
+ }
}
- // Allocate space for the callee saves (if any).
- emitFrameOffset(
- MBB, CalleeSavesBegin, DL, AArch64::SP, AArch64::SP, -AllocateBefore, TII,
- MachineInstr::FrameSetup, false, false, nullptr,
- EmitAsyncCFI && !HasFP && AllocateBefore,
- StackOffset::getFixed((int64_t)MFI.getStackSize() - NumBytes));
-
- if (EmitAsyncCFI)
- emitCalleeSavedSVELocations(MBB, CalleeSavesEnd);
-
- // Finally allocate remaining SVE stack space.
- emitFrameOffset(MBB, CalleeSavesEnd, DL, AArch64::SP, AArch64::SP,
- -AllocateAfter, TII, MachineInstr::FrameSetup, false, false,
- nullptr, EmitAsyncCFI && !HasFP && AllocateAfter,
- AllocateBefore + StackOffset::getFixed(
- (int64_t)MFI.getStackSize() - NumBytes));
+ // Allocate stack space for the local SVE objects.
+ if (SVELocalsSize)
+ allocateSVEStackSpace(
+ MBB, CalleeSavesEnd, SVELocalsSize,
+ SVECalleeSavedSize +
+ StackOffset::getFixed((int64_t)MFI.getStackSize() - NumBytes),
+ EmitAsyncCFI && !HasFP);
// Allocate space for the rest of the frame.
if (NumBytes) {
unsigned scratchSPReg = AArch64::SP;
+ bool NeedsStackProbe = TLI.hasInlineStackProbe(MF) &&
+ (NumBytes > AArch64::StackProbeMaxUnprobedStack ||
+ MFI.hasVarSizedObjects());
if (NeedsRealignment) {
scratchSPReg = findScratchNonCalleeSaveRegister(&MBB);
+ NeedsStackProbe |= TLI.hasInlineStackProbe(MF) &&
+ (NumBytes + MFI.getMaxAlign().value()) >
+ AArch64::StackProbeMaxUnprobedStack;
assert(scratchSPReg != AArch64::NoRegister);
}
@@ -1832,12 +1913,36 @@ void AArch64FrameLowering::emitPrologue(MachineFunction &MF,
// FIXME: in the case of dynamic re-alignment, NumBytes doesn't have
// the correct value here, as NumBytes also includes padding bytes,
// which shouldn't be counted here.
- emitFrameOffset(
- MBB, MBBI, DL, scratchSPReg, AArch64::SP,
- StackOffset::getFixed(-NumBytes), TII, MachineInstr::FrameSetup,
- false, NeedsWinCFI, &HasWinCFI, EmitAsyncCFI && !HasFP,
+ StackOffset CFAOffset =
SVEStackSize +
- StackOffset::getFixed((int64_t)MFI.getStackSize() - NumBytes));
+ StackOffset::getFixed((int64_t)MFI.getStackSize() - NumBytes);
+ if (NeedsStackProbe && !NeedsRealignment) {
+ // If we don't need to re-align the stack, we can use a more efficient
+ // sequence for stack probing.
+ Register ScratchReg = findScratchNonCalleeSaveRegister(&MBB);
+ assert(ScratchReg != AArch64::NoRegister);
+ BuildMI(MBB, MBBI, DL, TII->get(AArch64::PROBED_STACKALLOC))
+ .addDef(ScratchReg)
+ .addImm(NumBytes)
+ .addImm(CFAOffset.getFixed())
+ .addImm(CFAOffset.getScalable());
+ // The fixed allocation may leave unprobed bytes at the top of the
+ // stack. If we have variable-sized objects, we need to issue an extra
+ // probe, so their allocations starts in a known state.
+ if (MFI.hasVarSizedObjects()) {
+ // STR XZR, [SP]
+ BuildMI(MBB, MBBI, DL, TII->get(AArch64::STRXui))
+ .addReg(AArch64::XZR)
+ .addReg(AArch64::SP)
+ .addImm(0)
+ .setMIFlags(MachineInstr::FrameSetup);
+ }
+ } else {
+ emitFrameOffset(MBB, MBBI, DL, scratchSPReg, AArch64::SP,
+ StackOffset::getFixed(-NumBytes), TII,
+ MachineInstr::FrameSetup, false, NeedsWinCFI,
+ &HasWinCFI, EmitAsyncCFI && !HasFP, CFAOffset);
+ }
}
if (NeedsRealignment) {
assert(MFI.getMaxAlign() > Align(1));
@@ -1846,12 +1951,48 @@ void AArch64FrameLowering::emitPrologue(MachineFunction &MF,
// SUB X9, SP, NumBytes
// -- X9 is temporary register, so shouldn't contain any live data here,
// -- free to use. This is already produced by emitFrameOffset above.
- // AND SP, X9, 0b11111...0000
- uint64_t AndMask = ~(MFI.getMaxAlign().value() - 1);
- BuildMI(MBB, MBBI, DL, TII->get(AArch64::ANDXri), AArch64::SP)
- .addReg(scratchSPReg, RegState::Kill)
- .addImm(AArch64_AM::encodeLogicalImmediate(AndMask, 64));
+ const uint64_t MaxAlign = MFI.getMaxAlign().value();
+ const uint64_t AndMask = ~(MaxAlign - 1);
+
+ if (NeedsStackProbe) {
+ // If allocation size is known to not exceed the probe size, don't emit
+ // a probing loop.
+ if (NumBytes + MaxAlign - 1 <= TLI.getStackProbeSize(MF)) {
+ // AND SP, X9, 0b11111...0000
+ BuildMI(MBB, MBBI, DL, TII->get(AArch64::ANDXri), AArch64::SP)
+ .addReg(scratchSPReg, RegState::Kill)
+ .addImm(AArch64_AM::encodeLogicalImmediate(AndMask, 64))
+ .setMIFlags(MachineInstr::FrameSetup);
+ // STR XZR, [SP]
+ BuildMI(MBB, MBBI, DL, TII->get(AArch64::STRXui))
+ .addReg(AArch64::XZR)
+ .addReg(AArch64::SP)
+ .addImm(0)
+ .setMIFlags(MachineInstr::FrameSetup);
+ } else {
+ // AND X9, X9, 0b11111...0000
+ BuildMI(MBB, MBBI, DL, TII->get(AArch64::ANDXri), scratchSPReg)
+ .addReg(scratchSPReg, RegState::Kill)
+ .addImm(AArch64_AM::encodeLogicalImmediate(AndMask, 64))
+ .setMIFlags(MachineInstr::FrameSetup);
+ BuildMI(MBB, MBBI, DL, TII->get(AArch64::PROBED_STACKALLOC_VAR),
+ AArch64::SP)
+ .addReg(scratchSPReg);
+ // MOV SP, X9
+ BuildMI(MBB, MBBI, DL, TII->get(AArch64::ADDXri), AArch64::SP)
+ .addReg(scratchSPReg)
+ .addImm(0)
+ .addImm(AArch64_AM::getShifterImm(AArch64_AM::LSL, 0))
+ .setMIFlags(MachineInstr::FrameSetup);
+ }
+ } else {
+ // AND SP, X9, 0b11111...0000
+ BuildMI(MBB, MBBI, DL, TII->get(AArch64::ANDXri), AArch64::SP)
+ .addReg(scratchSPReg, RegState::Kill)
+ .addImm(AArch64_AM::encodeLogicalImmediate(AndMask, 64))
+ .setMIFlags(MachineInstr::FrameSetup);
+ }
AFI->setStackRealigned(true);
// No need for SEH instructions here; if we're realigning the stack,
@@ -4057,3 +4198,192 @@ void AArch64FrameLowering::orderFrameObjects(
dbgs() << "\n";
});
}
+
+/// Emit a loop to decrement SP until it is equal to TargetReg, with probes at
+/// least every ProbeSize bytes. Returns an iterator of the first instruction
+/// after the loop. The difference between SP and TargetReg must be an exact
+/// multiple of ProbeSize.
+MachineBasicBlock::iterator
+AArch64FrameLowering::inlineStackProbeLoopExactMultiple(
+ MachineBasicBlock::iterator MBBI, int64_t ProbeSize,
+ Register TargetReg) const {
+ MachineBasicBlock &MBB = *MBBI->getParent();
+ MachineFunction &MF = *MBB.getParent();
+ const AArch64InstrInfo *TII =
+ MF.getSubtarget<AArch64Subtarget>().getInstrInfo();
+ DebugLoc DL = MBB.findDebugLoc(MBBI);
+
+ MachineFunction::iterator MBBInsertPoint = std::next(MBB.getIterator());
+ MachineBasicBlock *LoopMBB = MF.CreateMachineBasicBlock(MBB.getBasicBlock());
+ MF.insert(MBBInsertPoint, LoopMBB);
+ MachineBasicBlock *ExitMBB = MF.CreateMachineBasicBlock(MBB.getBasicBlock());
+ MF.insert(MBBInsertPoint, ExitMBB);
+
+ // SUB SP, SP, #ProbeSize (or equivalent if ProbeSize is not encodable
+ // in SUB).
+ emitFrameOffset(*LoopMBB, LoopMBB->end(), DL, AArch64::SP, AArch64::SP,
+ StackOffset::getFixed(-ProbeSize), TII,
+ MachineInstr::FrameSetup);
+ // STR XZR, [SP]
+ BuildMI(*LoopMBB, LoopMBB->end(), DL, TII->get(AArch64::STRXui))
+ .addReg(AArch64::XZR)
+ .addReg(AArch64::SP)
+ .addImm(0)
+ .setMIFlags(MachineInstr::FrameSetup);
+ // CMP SP, TargetReg
+ BuildMI(*LoopMBB, LoopMBB->end(), DL, TII->get(AArch64::SUBSXrx64),
+ AArch64::XZR)
+ .addReg(AArch64::SP)
+ .addReg(TargetReg)
+ .addImm(AArch64_AM::getArithExtendImm(AArch64_AM::UXTX, 0))
+ .setMIFlags(MachineInstr::FrameSetup);
+ // B.CC Loop
+ BuildMI(*LoopMBB, LoopMBB->end(), DL, TII->get(AArch64::Bcc))
+ .addImm(AArch64CC::NE)
+ .addMBB(LoopMBB)
+ .setMIFlags(MachineInstr::FrameSetup);
+
+ LoopMBB->addSuccessor(ExitMBB);
+ LoopMBB->addSuccessor(LoopMBB);
+ // Synthesize the exit MBB.
+ ExitMBB->splice(ExitMBB->end(), &MBB, MBBI, MBB.end());
+ ExitMBB->transferSuccessorsAndUpdatePHIs(&MBB);
+ MBB.addSuccessor(LoopMBB);
+ // Update liveins.
+ recomputeLiveIns(*LoopMBB);
+ recomputeLiveIns(*ExitMBB);
+
+ return ExitMBB->begin();
+}
+
+MachineBasicBlock::iterator AArch64FrameLowering::inlineStackProbeFixed(
+ MachineBasicBlock::iterator MBBI, Register ScratchReg, int64_t FrameSize,
+ StackOffset CFAOffset) const {
+ MachineBasicBlock *MBB = MBBI->getParent();
+ MachineFunction &MF = *MBB->getParent();
+ const AArch64TargetLowering *TLI =
+ MF.getSubtarget<AArch64Subtarget>().getTargetLowering();
+ const AArch64InstrInfo *TII =
+ MF.getSubtarget<AArch64Subtarget>().getInstrInfo();
+ AArch64FunctionInfo *AFI = MF.getInfo<AArch64FunctionInfo>();
+ bool EmitAsyncCFI = AFI->needsAsyncDwarfUnwindInfo(MF);
+ bool HasFP = hasFP(MF);
+
+ DebugLoc DL;
+ int64_t ProbeSize = TLI->getStackProbeSize(MF);
+ int64_t NumBlocks = FrameSize / ProbeSize;
+ int64_t ResidualSize = FrameSize % ProbeSize;
+
+ LLVM_DEBUG(dbgs() << "Stack probing: total " << FrameSize << " bytes, "
+ << NumBlocks << " blocks of " << ProbeSize
+ << " bytes, plus " << ResidualSize << " bytes\n");
+
+ // Decrement SP by NumBlock * ProbeSize bytes, with either unrolled or
+ // ordinary loop.
+ if (NumBlocks <= AArch64::StackProbeMaxLoopUnroll) {
+ for (int i = 0; i < NumBlocks; ++i) {
+ // SUB SP, SP, #FrameSize (or equivalent if FrameSize is not
+ // encodable in a SUB).
+ emitFrameOffset(*MBB, MBBI, DL, AArch64::SP, AArch64::SP,
+ StackOffset::getFixed(-ProbeSize), TII,
+ MachineInstr::FrameSetup, false, false, nullptr,
+ EmitAsyncCFI && !HasFP, CFAOffset);
+ CFAOffset += StackOffset::getFixed(ProbeSize);
+ // STR XZR, [SP]
+ BuildMI(*MBB, MBBI, DL, TII->get(AArch64::STRXui))
+ .addReg(AArch64::XZR)
+ .addReg(AArch64::SP)
+ .addImm(0)
+ .setMIFlags(MachineInstr::FrameSetup);
+ }
+ } else if (NumBlocks != 0) {
+ // SUB ScratchReg, SP, #FrameSize (or equivalent if FrameSize is not
+ // encodable in ADD). ScrathReg may temporarily become the CFA register.
+ emitFrameOffset(*MBB, MBBI, DL, ScratchReg, AArch64::SP,
+ StackOffset::getFixed(-ProbeSize * NumBlocks), TII,
+ MachineInstr::FrameSetup, false, false, nullptr,
+ EmitAsyncCFI && !HasFP, CFAOffset);
+ CFAOffset += StackOffset::getFixed(ProbeSize * NumBlocks);
+ MBBI = inlineStackProbeLoopExactMultiple(MBBI, ProbeSize, ScratchReg);
+ MBB = MBBI->getParent();
+ if (EmitAsyncCFI && !HasFP) {
+ // Set the CFA register back to SP.
+ const AArch64RegisterInfo &RegInfo =
+ *MF.getSubtarget<AArch64Subtarget>().getRegisterInfo();
+ unsigned Reg = RegInfo.getDwarfRegNum(AArch64::SP, true);
+ unsigned CFIIndex =
+ MF.addFrameInst(MCCFIInstruction::createDefCfaRegister(nullptr, Reg));
+ BuildMI(*MBB, MBBI, DL, TII->get(TargetOpcode::CFI_INSTRUCTION))
+ .addCFIIndex(CFIIndex)
+ .setMIFlags(MachineInstr::FrameSetup);
+ }
+ }
+
+ if (ResidualSize != 0) {
+ // SUB SP, SP, #ResidualSize (or equivalent if ResidualSize is not encodable
+ // in SUB).
+ emitFrameOffset(*MBB, MBBI, DL, AArch64::SP, AArch64::SP,
+ StackOffset::getFixed(-ResidualSize), TII,
+ MachineInstr::FrameSetup, false, false, nullptr,
+ EmitAsyncCFI && !HasFP, CFAOffset);
+ if (ResidualSize > AArch64::StackProbeMaxUnprobedStack) {
+ // STR XZR, [SP]
+ BuildMI(*MBB, MBBI, DL, TII->get(AArch64::STRXui))
+ .addReg(AArch64::XZR)
+ .addReg(AArch64::SP)
+ .addImm(0)
+ .setMIFlags(MachineInstr::FrameSetup);
+ }
+ }
+
+ MachineBasicBlock::iterator Next = std::next(MBBI);
+ return Next;
+}
+
+MachineBasicBlock::iterator AArch64FrameLowering::inlineStackProbeFixed(
+ MachineBasicBlock::iterator MBBI) const {
+
+ Register ScratchReg = MBBI->getOperand(0).getReg();
+ int64_t FrameSize = MBBI->getOperand(1).getImm();
+ StackOffset CFAOffset = StackOffset::get(MBBI->getOperand(2).getImm(),
+ MBBI->getOperand(3).getImm());
+
+ MachineBasicBlock::iterator NextInst =
+ inlineStackProbeFixed(MBBI, ScratchReg, FrameSize, CFAOffset);
+
+ MBBI->eraseFromParent();
+ return NextInst;
+}
+
+MachineBasicBlock::iterator AArch64FrameLowering::inlineStackProbeVar(
+ MachineBasicBlock::iterator MBBI) const {
+ MachineBasicBlock &MBB = *MBBI->getParent();
+ MachineFunction &MF = *MBB.getParent();
+ const AArch64InstrInfo *TII =
+ MF.getSubtarget<AArch64Subtarget>().getInstrInfo();
+
+ DebugLoc DL = MBB.findDebugLoc(MBBI);
+ Register ScratchReg = MBBI->getOperand(0).getReg();
+ Register TargetReg = MBBI->getOperand(1).getReg();
+
+ MachineBasicBlock::iterator NextInst =
+ TII->insertStackProbingLoop(MBBI, ScratchReg, TargetReg);
+
+ MBBI->eraseFromParent();
+ return NextInst;
+}
+
+void AArch64FrameLowering::inlineStackProbe(MachineFunction &MF,
+ MachineBasicBlock &MBB) const {
+ for (auto MBBI = MBB.begin(), E = MBB.end(); MBBI != E;) {
+ if (MBBI->getOpcode() == AArch64::PROBED_STACKALLOC) {
+ MBBI = inlineStackProbeFixed(MBBI);
+ E = MBBI->getParent()->end();
+ } else if (MBBI->getOpcode() == AArch64::PROBED_STACKALLOC_VAR) {
+ MBBI = inlineStackProbeVar(MBBI);
+ E = MBBI->getParent()->end();
+ } else {
+ ++MBBI;
+ }
+ }
+}
diff --git a/llvm/lib/Target/AArch64/AArch64FrameLowering.h b/llvm/lib/Target/AArch64/AArch64FrameLowering.h
index 2d5529aa2fd8cb9..a50e220114f90a2 100644
--- a/llvm/lib/Target/AArch64/AArch64FrameLowering.h
+++ b/llvm/lib/Target/AArch64/AArch64FrameLowering.h
@@ -157,10 +157,33 @@ class AArch64FrameLowering : public TargetFrameLowering {
MachineBasicBlock::iterator MBBI) const;
void emitCalleeSavedSVERestores(MachineBasicBlock &MBB,
MachineBasicBlock::iterator MBBI) const;
+ void allocateSVEStackSpace(MachineBasicBlock &MBB,
+ MachineBasicBlock::iterator MBBI,
+ StackOffset AllocSize, StackOffset InitialOffset,
+ bool EmitCFI) const;
/// Emit target zero call-used regs.
void emitZeroCallUsedRegs(BitVector RegsToZero,
MachineBasicBlock &MBB) const override;
+
+ /// Replace a StackProbe stub (if any) with the actual probe code inline
+ void inlineStackProbe(MachineFunction &MF,
+ MachineBasicBlock &PrologueMBB) const override;
+
+ MachineBasicBlock::iterator
+ inlineStackProbeFixed(MachineBasicBlock::iterator MBBI, Register ScratchReg,
+ int64_t FrameSize, StackOffset CFAOffset) const;
+
+ MachineBasicBlock::iterator
+ inlineStackProbeFixed(MachineBasicBlock::iterator MBBI) const;
+
+ MachineBasicBlock::iterator
+ inlineStackProbeVar(MachineBasicBlock::iterator MBBI) const;
+
+ MachineBasicBlock::iterator
+ inlineStackProbeLoopExactMultiple(MachineBasicBlock::iterator MBBI,
+ int64_t NegProbeSize,
+ Register TargetReg) const;
};
} // End llvm namespace
diff --git a/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp b/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
index 30a66c7ae4be845..4609e0f7081407d 100644
--- a/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
+++ b/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
@@ -26273,3 +26273,37 @@ bool AArch64TargetLowering::preferScalarizeSplat(SDNode *N) const {
}
return true;
}
+
+bool AArch64TargetLowering::hasInlineStackProbe(
+ const MachineFunction &MF) const {
+ // If the function specifically requests inline stack probes, emit them.
+ if (MF.getFunction().hasFnAttribute("probe-stack")) {
+ if (MF.getFunction().getFnAttribute("probe-stack").getValueAsString() ==
+ "inline-asm")
+ return true;
+ else
+ llvm_unreachable("Unsupported stack probing method");
+ }
+
+ return false;
+}
+
+unsigned
+AArch64TargetLowering::getStackProbeSize(const MachineFunction &MF) const {
+ const TargetFrameLowering *TFI = Subtarget->getFrameLowering();
+ unsigned StackAlign = TFI->getStackAlignment();
+ assert(StackAlign >= 1 && isPowerOf2_32(StackAlign) &&
+ "Unexpected stack alignment");
+ // The default stack probe size is 4096 if the function has no
+ // stack-probe-size attribute. This is a safe default because it is the
+ // smallest possible guard page size.
+ unsigned StackProbeSize = 4096;
+ const Function &Fn = MF.getFunction();
+ if (Fn.hasFnAttribute("stack-probe-size"))
+ Fn.getFnAttribute("stack-probe-size")
+ .getValueAsString()
+ .getAsInteger(0, StackProbeSize);
+ // Round down to the stack alignment.
+ StackProbeSize &= ~(StackAlign - 1);
+ return StackProbeSize ? StackProbeSize : StackAlign;
+}
diff --git a/llvm/lib/Target/AArch64/AArch64ISelLowering.h b/llvm/lib/Target/AArch64/AArch64ISelLowering.h
index 48b5bc16b36537b..fa97821edb9a4a7 100644
--- a/llvm/lib/Target/AArch64/AArch64ISelLowering.h
+++ b/llvm/lib/Target/AArch64/AArch64ISelLowering.h
@@ -504,6 +504,13 @@ const unsigned RoundingBitsPos = 22;
ArrayRef<MCPhysReg> getGPRArgRegs();
ArrayRef<MCPhysReg> getFPRArgRegs();
+/// Maximum allowed number of unprobed bytes above SP at an ABI
+/// boundary.
+const unsigned StackProbeMaxUnprobedStack = 1024;
+
+/// Maximum number of iterations to unroll for a constant size probing loop.
+const unsigned StackProbeMaxLoopUnroll = 4;
+
} // namespace AArch64
class AArch64Subtarget;
@@ -938,6 +945,13 @@ class AArch64TargetLowering : public TargetLowering {
// used for 64bit and 128bit vectors as well.
bool useSVEForFixedLengthVectorVT(EVT VT, bool OverrideNEON = false) const;
+ /// True if stack clash protection is enabled for this functions.
+ bool hasInlineStackProbe(const MachineFunction &MF) const override;
+
+ /// Get the interval between stack-clash probes, which is equal to the stack
+ /// guard size, in bytes.
+ unsigned getStackProbeSize(const MachineFunction &MF) const;
+
private:
/// Keep a pointer to the AArch64Subtarget around so that we can
/// make the right decision when generating code for different targets.
diff --git a/llvm/lib/Target/AArch64/AArch64InstrInfo.cpp b/llvm/lib/Target/AArch64/AArch64InstrInfo.cpp
index 5ad2e4837e3db4e..1ff4b480a679df1 100644
--- a/llvm/lib/Target/AArch64/AArch64InstrInfo.cpp
+++ b/llvm/lib/Target/AArch64/AArch64InstrInfo.cpp
@@ -10,8 +10,8 @@
//
//===----------------------------------------------------------------------===//
-#include "AArch64ExpandImm.h"
#include "AArch64InstrInfo.h"
+#include "AArch64ExpandImm.h"
#include "AArch64FrameLowering.h"
#include "AArch64MachineFunctionInfo.h"
#include "AArch64Subtarget.h"
@@ -20,6 +20,7 @@
#include "llvm/ADT/ArrayRef.h"
#include "llvm/ADT/STLExtras.h"
#include "llvm/ADT/SmallVector.h"
+#include "llvm/CodeGen/LivePhysRegs.h"
#include "llvm/CodeGen/MachineBasicBlock.h"
#include "llvm/CodeGen/MachineCombinerPattern.h"
#include "llvm/CodeGen/MachineFrameInfo.h"
@@ -8596,6 +8597,96 @@ bool AArch64InstrInfo::isReallyTriviallyReMaterializable(
return TargetInstrInfo::isReallyTriviallyReMaterializable(MI);
}
+MachineBasicBlock::iterator
+AArch64InstrInfo::insertStackProbingLoop(MachineBasicBlock::iterator MBBI,
+ Register ScratchReg,
+ Register TargetReg) const {
+ MachineBasicBlock &MBB = *MBBI->getParent();
+ MachineFunction &MF = *MBB.getParent();
+ const AArch64TargetLowering *TLI =
+ MF.getSubtarget<AArch64Subtarget>().getTargetLowering();
+ const AArch64InstrInfo *TII =
+ MF.getSubtarget<AArch64Subtarget>().getInstrInfo();
+ int64_t ProbeSize = (int64_t)TLI->getStackProbeSize(MF);
+ DebugLoc DL = MBB.findDebugLoc(MBBI);
+
+ MachineFunction::iterator MBBInsertPoint = std::next(MBB.getIterator());
+ MachineBasicBlock *LoopTestMBB =
+ MF.CreateMachineBasicBlock(MBB.getBasicBlock());
+ MF.insert(MBBInsertPoint, LoopTestMBB);
+ MachineBasicBlock *LoopBodyMBB =
+ MF.CreateMachineBasicBlock(MBB.getBasicBlock());
+ MF.insert(MBBInsertPoint, LoopBodyMBB);
+ MachineBasicBlock *ExitMBB = MF.CreateMachineBasicBlock(MBB.getBasicBlock());
+ MF.insert(MBBInsertPoint, ExitMBB);
+
+ // LoopTest:
+ // SUB ScratchReg, ScratchReg, #ProbeSize
+ emitFrameOffset(*LoopTestMBB, LoopTestMBB->end(), DL, ScratchReg, ScratchReg,
+ StackOffset::getFixed(-ProbeSize), TII,
+ MachineInstr::FrameSetup);
+
+ // CMP ScratchReg, TargetReg
+ AArch64CC::CondCode Cond = AArch64CC::LE;
+ Register Op1 = ScratchReg;
+ Register Op2 = TargetReg;
+ if (Op2 == AArch64::SP) {
+ assert(Op1 != AArch64::SP && "At most one of the registers can be SP");
+ // CMP TargetReg, ScratchReg
+ std::swap(Op1, Op2);
+ Cond = AArch64CC::GT;
+ }
+ BuildMI(*LoopTestMBB, LoopTestMBB->end(), DL, TII->get(AArch64::SUBSXrx64),
+ AArch64::XZR)
+ .addReg(Op1)
+ .addReg(Op2)
+ .addImm(AArch64_AM::getArithExtendImm(AArch64_AM::UXTX, 0))
+ .setMIFlags(MachineInstr::FrameSetup);
+
+ // B.<Cond> LoopExit
+ BuildMI(*LoopTestMBB, LoopTestMBB->end(), DL, TII->get(AArch64::Bcc))
+ .addImm(Cond)
+ .addMBB(ExitMBB)
+ .setMIFlags(MachineInstr::FrameSetup);
+
+ // STR XZR, [ScratchReg]
+ BuildMI(*LoopBodyMBB, LoopBodyMBB->end(), DL, TII->get(AArch64::STRXui))
+ .addReg(AArch64::XZR)
+ .addReg(ScratchReg)
+ .addImm(0)
+ .setMIFlags(MachineInstr::FrameSetup);
+
+ // B loop
+ BuildMI(*LoopBodyMBB, LoopBodyMBB->end(), DL, TII->get(AArch64::B))
+ .addMBB(LoopTestMBB)
+ .setMIFlags(MachineInstr::FrameSetup);
+
+ // LoopExit:
+ // STR XZR, [TargetReg]
+ BuildMI(*ExitMBB, ExitMBB->begin(), DL, TII->get(AArch64::STRXui))
+ .addReg(AArch64::XZR)
+ .addReg(TargetReg)
+ .addImm(0)
+ .setMIFlags(MachineInstr::FrameSetup);
+
+ ExitMBB->splice(ExitMBB->end(), &MBB, std::next(MBBI), MBB.end());
+ ExitMBB->transferSuccessorsAndUpdatePHIs(&MBB);
+
+ LoopTestMBB->addSuccessor(ExitMBB);
+ LoopTestMBB->addSuccessor(LoopBodyMBB);
+ LoopBodyMBB->addSuccessor(LoopTestMBB);
+ MBB.addSuccessor(LoopTestMBB);
+
+ // Update liveins.
+ if (MF.getRegInfo().reservedRegsFrozen()) {
+ recomputeLiveIns(*LoopTestMBB);
+ recomputeLiveIns(*LoopBodyMBB);
+ recomputeLiveIns(*ExitMBB);
+ }
+
+ return ExitMBB->begin();
+}
+
#define GET_INSTRINFO_HELPERS
#define GET_INSTRMAP_INFO
#include "AArch64GenInstrInfo.inc"
diff --git a/llvm/lib/Target/AArch64/AArch64InstrInfo.h b/llvm/lib/Target/AArch64/AArch64InstrInfo.h
index 2cd028d263f694d..ee9fb07ea18c5c9 100644
--- a/llvm/lib/Target/AArch64/AArch64InstrInfo.h
+++ b/llvm/lib/Target/AArch64/AArch64InstrInfo.h
@@ -357,6 +357,11 @@ class AArch64InstrInfo final : public AArch64GenInstrInfo {
int64_t &VGSized);
bool isReallyTriviallyReMaterializable(const MachineInstr &MI) const override;
+
+ MachineBasicBlock::iterator
+ insertStackProbingLoop(MachineBasicBlock::iterator MBBI, Register ScratchReg,
+ Register TargetReg) const;
+
#define GET_INSTRINFO_HELPER_DECLS
#include "AArch64GenInstrInfo.inc"
diff --git a/llvm/lib/Target/AArch64/AArch64InstrInfo.td b/llvm/lib/Target/AArch64/AArch64InstrInfo.td
index 82b79cd7232cc90..1d16a6ab6aad815 100644
--- a/llvm/lib/Target/AArch64/AArch64InstrInfo.td
+++ b/llvm/lib/Target/AArch64/AArch64InstrInfo.td
@@ -887,7 +887,8 @@ include "SMEInstrFormats.td"
// Miscellaneous instructions.
//===----------------------------------------------------------------------===//
-let Defs = [SP], Uses = [SP], hasSideEffects = 1, isCodeGenOnly = 1 in {
+let hasSideEffects = 1, isCodeGenOnly = 1 in {
+let Defs = [SP], Uses = [SP] in {
// We set Sched to empty list because we expect these instructions to simply get
// removed in most cases.
def ADJCALLSTACKDOWN : Pseudo<(outs), (ins i32imm:$amt1, i32imm:$amt2),
@@ -896,7 +897,23 @@ def ADJCALLSTACKDOWN : Pseudo<(outs), (ins i32imm:$amt1, i32imm:$amt2),
def ADJCALLSTACKUP : Pseudo<(outs), (ins i32imm:$amt1, i32imm:$amt2),
[(AArch64callseq_end timm:$amt1, timm:$amt2)]>,
Sched<[]>;
-} // Defs = [SP], Uses = [SP], hasSideEffects = 1, isCodeGenOnly = 1
+
+// Probed stack allocation of a constant size, used in function prologues when
+// stack-clash protection is enabled.
+def PROBED_STACKALLOC : Pseudo<(outs GPR64:$scratch),
+ (ins i64imm:$stacksize, i64imm:$fixed_offset,
+ i64imm:$scalable_offset),
+ []>,
+ Sched<[]>;
+} // Defs = [SP], Uses = [SP]
+
+// Probed stack allocation of a variable size, used in function prologues when
+// stack-clash protection is enabled.
+def PROBED_STACKALLOC_VAR : Pseudo<(outs GPR64sp:$scratch),
+ (ins GPR64sp:$target),
+ []>,
+ Sched<[]>;
+} //hasSideEffects = 1, isCodeGenOnly = 1
let isReMaterializable = 1, isCodeGenOnly = 1 in {
// FIXME: The following pseudo instructions are only needed because remat
diff --git a/llvm/test/CodeGen/AArch64/framelayout-sve.mir b/llvm/test/CodeGen/AArch64/framelayout-sve.mir
index 54d6a5fc1997660..e4544cbe96fccac 100644
--- a/llvm/test/CodeGen/AArch64/framelayout-sve.mir
+++ b/llvm/test/CodeGen/AArch64/framelayout-sve.mir
@@ -206,7 +206,7 @@ body: |
# CHECK-NEXT: frame-setup CFI_INSTRUCTION offset $w29, -16
# CHECK-NEXT: $sp = frame-setup ADDVL_XXI $sp, -2
# CHECK-NEXT: $[[TMP:x[0-9]+]] = frame-setup SUBXri $sp, 16, 0
-# CHECK-NEXT: $sp = ANDXri killed $[[TMP]]
+# CHECK-NEXT: $sp = frame-setup ANDXri killed $[[TMP]]
# CHECK-NEXT: $sp = frame-destroy ADDXri $fp, 0, 0
# CHECK-NEXT: frame-destroy CFI_INSTRUCTION def_cfa $wsp, 16
# CHECK-NEXT: $sp, $fp, $lr = frame-destroy LDPXpost $sp, 2
@@ -1032,7 +1032,7 @@ body: |
# CHECK-NEXT: frame-setup CFI_INSTRUCTION escape 0x10, 0x4f, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x40, 0x92, 0x2e, 0x00, 0x1e, 0x22
# CHECK-NEXT: $sp = frame-setup ADDVL_XXI $sp, -1
# CHECK-NEXT: $[[TMP:x[0-9]+]] = frame-setup SUBXri $sp, 16, 0
-# CHECK-NEXT: $sp = ANDXri killed $[[TMP]]
+# CHECK-NEXT: $sp = frame-setup ANDXri killed $[[TMP]]
# CHECK: $sp = frame-destroy ADDVL_XXI $fp, -18
# CHECK-NEXT: $p15 = frame-destroy LDR_PXI $sp, 4
diff --git a/llvm/test/CodeGen/AArch64/spill-stack-realignment.mir b/llvm/test/CodeGen/AArch64/spill-stack-realignment.mir
index 1b9411d07f433ab..f6fc627ac2d3d87 100644
--- a/llvm/test/CodeGen/AArch64/spill-stack-realignment.mir
+++ b/llvm/test/CodeGen/AArch64/spill-stack-realignment.mir
@@ -21,7 +21,7 @@ stack:
- { id: 1, size: 4, alignment: 4, local-offset: -68 }
# CHECK: body:
-# CHECK: $sp = ANDXri killed ${{x[0-9]+}}, 7865
+# CHECK: $sp = frame-setup ANDXri killed ${{x[0-9]+}}, 7865
# CHECK: STRSui $s0, $sp, 0
# CHECK: STRSui $s0, $fp, 7
body: |
diff --git a/llvm/test/CodeGen/AArch64/stack-probing-64k.ll b/llvm/test/CodeGen/AArch64/stack-probing-64k.ll
new file mode 100644
index 000000000000000..912d0c273c66180
--- /dev/null
+++ b/llvm/test/CodeGen/AArch64/stack-probing-64k.ll
@@ -0,0 +1,392 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple aarch64-none-eabi < %s -verify-machineinstrs -enable-post-misched=false | FileCheck %s
+; RUN: llc -mtriple aarch64-none-eabi < %s -verify-machineinstrs -enable-post-misched=false -global-isel | FileCheck %s
+
+; Tests for prolog sequences for stack probing, when using a 64KiB stack guard.
+
+; 64k bytes is the largest frame we can probe in one go.
+define void @static_65536(ptr %out) #0 {
+; CHECK-LABEL: static_65536:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: sub sp, sp, #16, lsl #12 // =65536
+; CHECK-NEXT: .cfi_def_cfa_offset 65552
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: mov x8, sp
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: add sp, sp, #16, lsl #12 // =65536
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %v = alloca i8, i64 65536, align 1
+ store i8* %v, ptr %out, align 8
+ ret void
+}
+
+; 64k+16 bytes, still needs just one probe.
+define void @static_65552(ptr %out) #0 {
+; CHECK-LABEL: static_65552:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: sub sp, sp, #16, lsl #12 // =65536
+; CHECK-NEXT: .cfi_def_cfa_offset 65552
+; CHECK-NEXT: str xzr, [sp], #-16
+; CHECK-NEXT: .cfi_def_cfa_offset 65568
+; CHECK-NEXT: mov x8, sp
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: add sp, sp, #16, lsl #12 // =65536
+; CHECK-NEXT: .cfi_def_cfa_offset 32
+; CHECK-NEXT: add sp, sp, #16
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %v = alloca i8, i64 65552, align 1
+ store i8* %v, ptr %out, align 8
+ ret void
+}
+
+; 64k+1024 bytes, the largest frame which needs just one probe.
+define void @static_66560(ptr %out) #0 {
+; CHECK-LABEL: static_66560:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: sub sp, sp, #16, lsl #12 // =65536
+; CHECK-NEXT: .cfi_def_cfa_offset 65552
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: sub sp, sp, #1024
+; CHECK-NEXT: .cfi_def_cfa_offset 66576
+; CHECK-NEXT: mov x8, sp
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: add sp, sp, #16, lsl #12 // =65536
+; CHECK-NEXT: .cfi_def_cfa_offset 1040
+; CHECK-NEXT: add sp, sp, #1024
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %v = alloca i8, i64 66560, align 1
+ store i8* %v, ptr %out, align 8
+ ret void
+}
+
+; 64k+1024+16 bytes, the smallest frame which needs two probes.
+define void @static_66576(ptr %out) #0 {
+; CHECK-LABEL: static_66576:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: sub sp, sp, #16, lsl #12 // =65536
+; CHECK-NEXT: .cfi_def_cfa_offset 65552
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: sub sp, sp, #1040
+; CHECK-NEXT: .cfi_def_cfa_offset 66592
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: mov x8, sp
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: add sp, sp, #16, lsl #12 // =65536
+; CHECK-NEXT: .cfi_def_cfa_offset 1056
+; CHECK-NEXT: add sp, sp, #1040
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %v = alloca i8, i64 66576, align 1
+ store i8* %v, ptr %out, align 8
+ ret void
+}
+
+; 2*64k+1024, the largest frame needing two probes.
+define void @static_132096(ptr %out) #0 {
+; CHECK-LABEL: static_132096:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: sub sp, sp, #16, lsl #12 // =65536
+; CHECK-NEXT: .cfi_def_cfa_offset 65552
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: sub sp, sp, #16, lsl #12 // =65536
+; CHECK-NEXT: .cfi_def_cfa_offset 131088
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: sub sp, sp, #1024
+; CHECK-NEXT: .cfi_def_cfa_offset 132112
+; CHECK-NEXT: mov x8, sp
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: add sp, sp, #32, lsl #12 // =131072
+; CHECK-NEXT: .cfi_def_cfa_offset 1040
+; CHECK-NEXT: add sp, sp, #1024
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %v = alloca i8, i64 132096, align 1
+ store i8* %v, ptr %out, align 8
+ ret void
+}
+
+; 5*64k-16, the largest frame probed without a loop.
+define void @static_327664(ptr %out) #0 {
+; CHECK-LABEL: static_327664:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: sub sp, sp, #16, lsl #12 // =65536
+; CHECK-NEXT: .cfi_def_cfa_offset 65552
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: sub sp, sp, #16, lsl #12 // =65536
+; CHECK-NEXT: .cfi_def_cfa_offset 131088
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: sub sp, sp, #16, lsl #12 // =65536
+; CHECK-NEXT: .cfi_def_cfa_offset 196624
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: sub sp, sp, #16, lsl #12 // =65536
+; CHECK-NEXT: .cfi_def_cfa_offset 262160
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: sub sp, sp, #15, lsl #12 // =61440
+; CHECK-NEXT: .cfi_def_cfa_offset 323600
+; CHECK-NEXT: sub sp, sp, #4080
+; CHECK-NEXT: .cfi_def_cfa_offset 327680
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: mov x8, sp
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: add sp, sp, #79, lsl #12 // =323584
+; CHECK-NEXT: .cfi_def_cfa_offset 4096
+; CHECK-NEXT: add sp, sp, #4080
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %v = alloca i8, i64 327664, align 1
+ store i8* %v, ptr %out, align 8
+ ret void
+}
+
+; 5*64k, smallest frame probed with a loop.
+define void @static_327680(ptr %out) #0 {
+; CHECK-LABEL: static_327680:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: sub x9, sp, #80, lsl #12 // =327680
+; CHECK-NEXT: .cfi_def_cfa w9, 327696
+; CHECK-NEXT: .LBB6_1: // %entry
+; CHECK-NEXT: // =>This Inner Loop Header: Depth=1
+; CHECK-NEXT: sub sp, sp, #16, lsl #12 // =65536
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: cmp sp, x9
+; CHECK-NEXT: b.ne .LBB6_1
+; CHECK-NEXT: // %bb.2: // %entry
+; CHECK-NEXT: .cfi_def_cfa_register wsp
+; CHECK-NEXT: mov x8, sp
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: add sp, sp, #80, lsl #12 // =327680
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %v = alloca i8, i64 327680, align 1
+ store i8* %v, ptr %out, align 8
+ ret void
+}
+
+; 5*64k+1024, large enough to use a loop, but not a multiple of 64KiB
+; so has a reminder, but no extra probe.
+define void @static_328704(ptr %out) #0 {
+; CHECK-LABEL: static_328704:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: sub x9, sp, #80, lsl #12 // =327680
+; CHECK-NEXT: .cfi_def_cfa w9, 327696
+; CHECK-NEXT: .LBB7_1: // %entry
+; CHECK-NEXT: // =>This Inner Loop Header: Depth=1
+; CHECK-NEXT: sub sp, sp, #16, lsl #12 // =65536
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: cmp sp, x9
+; CHECK-NEXT: b.ne .LBB7_1
+; CHECK-NEXT: // %bb.2: // %entry
+; CHECK-NEXT: .cfi_def_cfa_register wsp
+; CHECK-NEXT: sub sp, sp, #1024
+; CHECK-NEXT: .cfi_def_cfa_offset 328720
+; CHECK-NEXT: mov x8, sp
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: add sp, sp, #80, lsl #12 // =327680
+; CHECK-NEXT: .cfi_def_cfa_offset 1040
+; CHECK-NEXT: add sp, sp, #1024
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %v = alloca i8, i64 328704, align 1
+ store i8* %v, ptr %out, align 8
+ ret void
+}
+
+; 5*64k+1040, large enough to use a loop, has a reminder and
+; an extra probe.
+define void @static_328720(ptr %out) #0 {
+; CHECK-LABEL: static_328720:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: sub x9, sp, #80, lsl #12 // =327680
+; CHECK-NEXT: .cfi_def_cfa w9, 327696
+; CHECK-NEXT: .LBB8_1: // %entry
+; CHECK-NEXT: // =>This Inner Loop Header: Depth=1
+; CHECK-NEXT: sub sp, sp, #16, lsl #12 // =65536
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: cmp sp, x9
+; CHECK-NEXT: b.ne .LBB8_1
+; CHECK-NEXT: // %bb.2: // %entry
+; CHECK-NEXT: .cfi_def_cfa_register wsp
+; CHECK-NEXT: sub sp, sp, #1040
+; CHECK-NEXT: .cfi_def_cfa_offset 328736
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: mov x8, sp
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: add sp, sp, #80, lsl #12 // =327680
+; CHECK-NEXT: .cfi_def_cfa_offset 1056
+; CHECK-NEXT: add sp, sp, #1040
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %v = alloca i8, i64 328720, align 1
+ store i8* %v, ptr %out, align 8
+ ret void
+}
+
+; A small allocation, but with a very large alignment requirement. We do this
+; by moving SP far enough that a sufficiently-aligned block will exist
+; somewhere in the stack frame, so must probe the whole of that larger SP move.
+define void @static_16_align_131072(ptr %out) #0 {
+; CHECK-LABEL: static_16_align_131072:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: stp x29, x30, [sp, #-16]! // 16-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: mov x29, sp
+; CHECK-NEXT: .cfi_def_cfa w29, 16
+; CHECK-NEXT: .cfi_offset w30, -8
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: sub x9, sp, #31, lsl #12 // =126976
+; CHECK-NEXT: sub x9, x9, #4080
+; CHECK-NEXT: and x9, x9, #0xfffffffffffe0000
+; CHECK-NEXT: .LBB9_1: // %entry
+; CHECK-NEXT: // =>This Inner Loop Header: Depth=1
+; CHECK-NEXT: sub sp, sp, #16, lsl #12 // =65536
+; CHECK-NEXT: cmp sp, x9
+; CHECK-NEXT: b.le .LBB9_3
+; CHECK-NEXT: // %bb.2: // %entry
+; CHECK-NEXT: // in Loop: Header=BB9_1 Depth=1
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: b .LBB9_1
+; CHECK-NEXT: .LBB9_3: // %entry
+; CHECK-NEXT: str xzr, [x9]
+; CHECK-NEXT: mov sp, x9
+; CHECK-NEXT: mov x8, sp
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: mov sp, x29
+; CHECK-NEXT: .cfi_def_cfa wsp, 16
+; CHECK-NEXT: ldp x29, x30, [sp], #16 // 16-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w30
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %v = alloca i8, i64 16, align 131072
+ store i8* %v, ptr %out, align 8
+ ret void
+}
+
+; A small allocation, but with a very large alignment requirement which
+; is nevertheless small enough as to not need a loop.
+define void @static_16_align_8192(ptr %out) #0 {
+; CHECK-LABEL: static_16_align_8192:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: stp x29, x30, [sp, #-16]! // 16-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: mov x29, sp
+; CHECK-NEXT: .cfi_def_cfa w29, 16
+; CHECK-NEXT: .cfi_offset w30, -8
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: sub x9, sp, #1, lsl #12 // =4096
+; CHECK-NEXT: sub x9, x9, #4080
+; CHECK-NEXT: and sp, x9, #0xffffffffffffe000
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: mov x8, sp
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: mov sp, x29
+; CHECK-NEXT: .cfi_def_cfa wsp, 16
+; CHECK-NEXT: ldp x29, x30, [sp], #16 // 16-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w30
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %v = alloca i8, i64 16, align 8192
+ store i8* %v, ptr %out, align 8
+ ret void
+}
+
+; A large allocation with a very large alignment requirement which
+; is nevertheless small enough as to not need a loop.
+define void @static_32752_align_32k(ptr %out) #0 {
+; CHECK-LABEL: static_32752_align_32k:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: stp x29, x30, [sp, #-16]! // 16-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: mov x29, sp
+; CHECK-NEXT: .cfi_def_cfa w29, 16
+; CHECK-NEXT: .cfi_offset w30, -8
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: sub x9, sp, #7, lsl #12 // =28672
+; CHECK-NEXT: sub x9, x9, #4080
+; CHECK-NEXT: and sp, x9, #0xffffffffffff8000
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: mov x8, sp
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: mov sp, x29
+; CHECK-NEXT: .cfi_def_cfa wsp, 16
+; CHECK-NEXT: ldp x29, x30, [sp], #16 // 16-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w30
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %v = alloca i8, i64 32752, align 32768
+ store i8* %v, ptr %out, align 8
+ ret void
+}
+
+attributes #0 = { uwtable(async) "probe-stack"="inline-asm" "stack-probe-size"="65536" "frame-pointer"="none" }
diff --git a/llvm/test/CodeGen/AArch64/stack-probing-sve.ll b/llvm/test/CodeGen/AArch64/stack-probing-sve.ll
new file mode 100644
index 000000000000000..6f7420d04d2830f
--- /dev/null
+++ b/llvm/test/CodeGen/AArch64/stack-probing-sve.ll
@@ -0,0 +1,661 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple aarch64-none-eabi < %s -verify-machineinstrs | FileCheck %s
+; RUN: llc -mtriple aarch64-none-eabi < %s -verify-machineinstrs -global-isel -global-isel-abort=2 | FileCheck %s
+
+; Test prolog sequences for stack probing when SVE objects are involved.
+
+; The space for SVE objects needs probing in the general case, because
+; the stack adjustment may happen to be too big (i.e. greater than the
+; probe size) to allocate with a single `addvl`.
+; When we do know that the stack adjustment cannot exceed the probe size
+; we can avoid emitting a probe loop and emit a simple `addvl; str`
+; sequence instead.
+
+define void @sve_1_vector(ptr %out) #0 {
+; CHECK-LABEL: sve_1_vector:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: addvl sp, sp, #-1
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0c, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 8 * VG
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: addvl sp, sp, #1
+; CHECK-NEXT: .cfi_def_cfa wsp, 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %vec = alloca <vscale x 4 x float>, align 16
+ ret void
+}
+
+; As above, but with 4 SVE vectors of stack space.
+define void @sve_4_vector(ptr %out) #0 {
+; CHECK-LABEL: sve_4_vector:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: addvl sp, sp, #-4
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0c, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0x20, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 32 * VG
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: addvl sp, sp, #4
+; CHECK-NEXT: .cfi_def_cfa wsp, 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %vec1 = alloca <vscale x 4 x float>, align 16
+ %vec2 = alloca <vscale x 4 x float>, align 16
+ %vec3 = alloca <vscale x 4 x float>, align 16
+ %vec4 = alloca <vscale x 4 x float>, align 16
+ ret void
+}
+
+; As above, but with 16 SVE vectors of stack space.
+; The stack adjustment is less than or equal to 16 x 256 = 4096, so
+; we can allocate the locals at once.
+define void @sve_16_vector(ptr %out) #0 {
+; CHECK-LABEL: sve_16_vector:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: addvl sp, sp, #-16
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0x80, 0x01, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 128 * VG
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: addvl sp, sp, #16
+; CHECK-NEXT: .cfi_def_cfa wsp, 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %vec1 = alloca <vscale x 4 x float>, align 16
+ %vec2 = alloca <vscale x 4 x float>, align 16
+ %vec3 = alloca <vscale x 4 x float>, align 16
+ %vec4 = alloca <vscale x 4 x float>, align 16
+ %vec5 = alloca <vscale x 4 x float>, align 16
+ %vec6 = alloca <vscale x 4 x float>, align 16
+ %vec7 = alloca <vscale x 4 x float>, align 16
+ %vec8 = alloca <vscale x 4 x float>, align 16
+ %vec9 = alloca <vscale x 4 x float>, align 16
+ %vec10 = alloca <vscale x 4 x float>, align 16
+ %vec11 = alloca <vscale x 4 x float>, align 16
+ %vec12 = alloca <vscale x 4 x float>, align 16
+ %vec13 = alloca <vscale x 4 x float>, align 16
+ %vec14 = alloca <vscale x 4 x float>, align 16
+ %vec15 = alloca <vscale x 4 x float>, align 16
+ %vec16 = alloca <vscale x 4 x float>, align 16
+ ret void
+}
+
+; As above, but with 17 SVE vectors of stack space. Now we need
+; a probing loops since stack adjustment may be greater than
+; the probe size (17 x 256 = 4354 bytes)
+define void @sve_17_vector(ptr %out) #0 {
+; CHECK-LABEL: sve_17_vector:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: addvl x9, sp, #-17
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x79, 0x00, 0x11, 0x10, 0x22, 0x11, 0x88, 0x01, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $x9 + 16 + 136 * VG
+; CHECK-NEXT: .LBB3_1: // %entry
+; CHECK-NEXT: // =>This Inner Loop Header: Depth=1
+; CHECK-NEXT: sub sp, sp, #1, lsl #12 // =4096
+; CHECK-NEXT: cmp sp, x9
+; CHECK-NEXT: b.le .LBB3_3
+; CHECK-NEXT: // %bb.2: // %entry
+; CHECK-NEXT: // in Loop: Header=BB3_1 Depth=1
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: b .LBB3_1
+; CHECK-NEXT: .LBB3_3: // %entry
+; CHECK-NEXT: str xzr, [x9]
+; CHECK-NEXT: mov sp, x9
+; CHECK-NEXT: .cfi_def_cfa_register wsp
+; CHECK-NEXT: addvl sp, sp, #17
+; CHECK-NEXT: .cfi_def_cfa wsp, 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %vec1 = alloca <vscale x 4 x float>, align 16
+ %vec2 = alloca <vscale x 4 x float>, align 16
+ %vec3 = alloca <vscale x 4 x float>, align 16
+ %vec4 = alloca <vscale x 4 x float>, align 16
+ %vec5 = alloca <vscale x 4 x float>, align 16
+ %vec6 = alloca <vscale x 4 x float>, align 16
+ %vec7 = alloca <vscale x 4 x float>, align 16
+ %vec8 = alloca <vscale x 4 x float>, align 16
+ %vec9 = alloca <vscale x 4 x float>, align 16
+ %vec10 = alloca <vscale x 4 x float>, align 16
+ %vec11 = alloca <vscale x 4 x float>, align 16
+ %vec12 = alloca <vscale x 4 x float>, align 16
+ %vec13 = alloca <vscale x 4 x float>, align 16
+ %vec14 = alloca <vscale x 4 x float>, align 16
+ %vec15 = alloca <vscale x 4 x float>, align 16
+ %vec16 = alloca <vscale x 4 x float>, align 16
+ %vec17 = alloca <vscale x 4 x float>, align 16
+ ret void
+}
+
+; Space for callee-saved SVE register is allocated similarly to allocating
+; space for SVE locals. When we know the stack adjustment cannot exceed the
+; probe size we can skip the explict probe, since saving SVE registers serves
+; as an implicit probe.
+define void @sve_1v_csr(<vscale x 4 x float> %a) #0 {
+; CHECK-LABEL: sve_1v_csr:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: addvl sp, sp, #-1
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0c, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 8 * VG
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: str z8, [sp] // 16-byte Folded Spill
+; CHECK-NEXT: .cfi_escape 0x10, 0x48, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x78, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $d8 @ cfa - 16 - 8 * VG
+; CHECK-NEXT: //APP
+; CHECK-NEXT: //NO_APP
+; CHECK-NEXT: ldr z8, [sp] // 16-byte Folded Reload
+; CHECK-NEXT: addvl sp, sp, #1
+; CHECK-NEXT: .cfi_def_cfa wsp, 16
+; CHECK-NEXT: .cfi_restore z8
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ call void asm sideeffect "", "~{z8}" ()
+ ret void
+}
+
+define void @sve_4v_csr(<vscale x 4 x float> %a) #0 {
+; CHECK-LABEL: sve_4v_csr:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: addvl sp, sp, #-4
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0c, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0x20, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 32 * VG
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: str z11, [sp] // 16-byte Folded Spill
+; CHECK-NEXT: str z10, [sp, #1, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z9, [sp, #2, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z8, [sp, #3, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: .cfi_escape 0x10, 0x48, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x78, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $d8 @ cfa - 16 - 8 * VG
+; CHECK-NEXT: .cfi_escape 0x10, 0x49, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x70, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $d9 @ cfa - 16 - 16 * VG
+; CHECK-NEXT: .cfi_escape 0x10, 0x4a, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x68, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $d10 @ cfa - 16 - 24 * VG
+; CHECK-NEXT: .cfi_escape 0x10, 0x4b, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x60, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $d11 @ cfa - 16 - 32 * VG
+; CHECK-NEXT: //APP
+; CHECK-NEXT: //NO_APP
+; CHECK-NEXT: ldr z11, [sp] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z10, [sp, #1, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z9, [sp, #2, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z8, [sp, #3, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: addvl sp, sp, #4
+; CHECK-NEXT: .cfi_def_cfa wsp, 16
+; CHECK-NEXT: .cfi_restore z8
+; CHECK-NEXT: .cfi_restore z9
+; CHECK-NEXT: .cfi_restore z10
+; CHECK-NEXT: .cfi_restore z11
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ call void asm sideeffect "", "~{z8},~{z9},~{z10},~{z11}" ()
+ ret void
+}
+
+define void @sve_16v_csr(<vscale x 4 x float> %a) #0 {
+; CHECK-LABEL: sve_16v_csr:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: addvl sp, sp, #-16
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0x80, 0x01, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 128 * VG
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: str z23, [sp] // 16-byte Folded Spill
+; CHECK-NEXT: str z22, [sp, #1, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z21, [sp, #2, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z20, [sp, #3, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z19, [sp, #4, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z18, [sp, #5, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z17, [sp, #6, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z16, [sp, #7, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z15, [sp, #8, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z14, [sp, #9, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z13, [sp, #10, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z12, [sp, #11, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z11, [sp, #12, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z10, [sp, #13, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z9, [sp, #14, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z8, [sp, #15, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: .cfi_escape 0x10, 0x48, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x78, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $d8 @ cfa - 16 - 8 * VG
+; CHECK-NEXT: .cfi_escape 0x10, 0x49, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x70, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $d9 @ cfa - 16 - 16 * VG
+; CHECK-NEXT: .cfi_escape 0x10, 0x4a, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x68, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $d10 @ cfa - 16 - 24 * VG
+; CHECK-NEXT: .cfi_escape 0x10, 0x4b, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x60, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $d11 @ cfa - 16 - 32 * VG
+; CHECK-NEXT: .cfi_escape 0x10, 0x4c, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x58, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $d12 @ cfa - 16 - 40 * VG
+; CHECK-NEXT: .cfi_escape 0x10, 0x4d, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x50, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $d13 @ cfa - 16 - 48 * VG
+; CHECK-NEXT: .cfi_escape 0x10, 0x4e, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x48, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $d14 @ cfa - 16 - 56 * VG
+; CHECK-NEXT: .cfi_escape 0x10, 0x4f, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x40, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $d15 @ cfa - 16 - 64 * VG
+; CHECK-NEXT: //APP
+; CHECK-NEXT: //NO_APP
+; CHECK-NEXT: ldr z23, [sp] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z22, [sp, #1, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z21, [sp, #2, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z20, [sp, #3, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z19, [sp, #4, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z18, [sp, #5, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z17, [sp, #6, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z16, [sp, #7, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z15, [sp, #8, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z14, [sp, #9, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z13, [sp, #10, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z12, [sp, #11, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z11, [sp, #12, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z10, [sp, #13, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z9, [sp, #14, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z8, [sp, #15, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: addvl sp, sp, #16
+; CHECK-NEXT: .cfi_def_cfa wsp, 16
+; CHECK-NEXT: .cfi_restore z8
+; CHECK-NEXT: .cfi_restore z9
+; CHECK-NEXT: .cfi_restore z10
+; CHECK-NEXT: .cfi_restore z11
+; CHECK-NEXT: .cfi_restore z12
+; CHECK-NEXT: .cfi_restore z13
+; CHECK-NEXT: .cfi_restore z14
+; CHECK-NEXT: .cfi_restore z15
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ call void asm sideeffect "", "~{z8},~{z9},~{z10},~{z11},~{z12},~{z13},~{z14},~{z15},~{z16},~{z17},~{z18},~{z19},~{z20},~{z21},~{z22},~{z23}" ()
+ ret void
+}
+
+define void @sve_1p_csr(<vscale x 4 x float> %a) #0 {
+; CHECK-LABEL: sve_1p_csr:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: addvl sp, sp, #-1
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0c, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 8 * VG
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: str p8, [sp, #7, mul vl] // 2-byte Folded Spill
+; CHECK-NEXT: //APP
+; CHECK-NEXT: //NO_APP
+; CHECK-NEXT: ldr p8, [sp, #7, mul vl] // 2-byte Folded Reload
+; CHECK-NEXT: addvl sp, sp, #1
+; CHECK-NEXT: .cfi_def_cfa wsp, 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ call void asm sideeffect "", "~{p8}" ()
+ ret void
+}
+
+define void @sve_4p_csr(<vscale x 4 x float> %a) #0 {
+; CHECK-LABEL: sve_4p_csr:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: addvl sp, sp, #-1
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0c, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 8 * VG
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: str p11, [sp, #4, mul vl] // 2-byte Folded Spill
+; CHECK-NEXT: str p10, [sp, #5, mul vl] // 2-byte Folded Spill
+; CHECK-NEXT: str p9, [sp, #6, mul vl] // 2-byte Folded Spill
+; CHECK-NEXT: str p8, [sp, #7, mul vl] // 2-byte Folded Spill
+; CHECK-NEXT: //APP
+; CHECK-NEXT: //NO_APP
+; CHECK-NEXT: ldr p11, [sp, #4, mul vl] // 2-byte Folded Reload
+; CHECK-NEXT: ldr p10, [sp, #5, mul vl] // 2-byte Folded Reload
+; CHECK-NEXT: ldr p9, [sp, #6, mul vl] // 2-byte Folded Reload
+; CHECK-NEXT: ldr p8, [sp, #7, mul vl] // 2-byte Folded Reload
+; CHECK-NEXT: addvl sp, sp, #1
+; CHECK-NEXT: .cfi_def_cfa wsp, 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ call void asm sideeffect "", "~{p8},~{p9},~{p10},~{p11}" ()
+ ret void
+}
+
+define void @sve_16v_1p_csr(<vscale x 4 x float> %a) #0 {
+; CHECK-LABEL: sve_16v_1p_csr:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: addvl x9, sp, #-17
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x79, 0x00, 0x11, 0x10, 0x22, 0x11, 0x88, 0x01, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $x9 + 16 + 136 * VG
+; CHECK-NEXT: .LBB9_1: // %entry
+; CHECK-NEXT: // =>This Inner Loop Header: Depth=1
+; CHECK-NEXT: sub sp, sp, #1, lsl #12 // =4096
+; CHECK-NEXT: cmp sp, x9
+; CHECK-NEXT: b.le .LBB9_3
+; CHECK-NEXT: // %bb.2: // %entry
+; CHECK-NEXT: // in Loop: Header=BB9_1 Depth=1
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: b .LBB9_1
+; CHECK-NEXT: .LBB9_3: // %entry
+; CHECK-NEXT: str xzr, [x9]
+; CHECK-NEXT: mov sp, x9
+; CHECK-NEXT: .cfi_def_cfa_register wsp
+; CHECK-NEXT: str p8, [sp, #7, mul vl] // 2-byte Folded Spill
+; CHECK-NEXT: str z23, [sp, #1, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z22, [sp, #2, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z21, [sp, #3, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z20, [sp, #4, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z19, [sp, #5, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z18, [sp, #6, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z17, [sp, #7, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z16, [sp, #8, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z15, [sp, #9, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z14, [sp, #10, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z13, [sp, #11, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z12, [sp, #12, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z11, [sp, #13, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z10, [sp, #14, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z9, [sp, #15, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: str z8, [sp, #16, mul vl] // 16-byte Folded Spill
+; CHECK-NEXT: .cfi_escape 0x10, 0x48, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x78, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $d8 @ cfa - 16 - 8 * VG
+; CHECK-NEXT: .cfi_escape 0x10, 0x49, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x70, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $d9 @ cfa - 16 - 16 * VG
+; CHECK-NEXT: .cfi_escape 0x10, 0x4a, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x68, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $d10 @ cfa - 16 - 24 * VG
+; CHECK-NEXT: .cfi_escape 0x10, 0x4b, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x60, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $d11 @ cfa - 16 - 32 * VG
+; CHECK-NEXT: .cfi_escape 0x10, 0x4c, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x58, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $d12 @ cfa - 16 - 40 * VG
+; CHECK-NEXT: .cfi_escape 0x10, 0x4d, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x50, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $d13 @ cfa - 16 - 48 * VG
+; CHECK-NEXT: .cfi_escape 0x10, 0x4e, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x48, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $d14 @ cfa - 16 - 56 * VG
+; CHECK-NEXT: .cfi_escape 0x10, 0x4f, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x40, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $d15 @ cfa - 16 - 64 * VG
+; CHECK-NEXT: //APP
+; CHECK-NEXT: //NO_APP
+; CHECK-NEXT: ldr p8, [sp, #7, mul vl] // 2-byte Folded Reload
+; CHECK-NEXT: ldr z23, [sp, #1, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z22, [sp, #2, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z21, [sp, #3, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z20, [sp, #4, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z19, [sp, #5, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z18, [sp, #6, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z17, [sp, #7, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z16, [sp, #8, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z15, [sp, #9, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z14, [sp, #10, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z13, [sp, #11, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z12, [sp, #12, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z11, [sp, #13, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z10, [sp, #14, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z9, [sp, #15, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: ldr z8, [sp, #16, mul vl] // 16-byte Folded Reload
+; CHECK-NEXT: addvl sp, sp, #17
+; CHECK-NEXT: .cfi_def_cfa wsp, 16
+; CHECK-NEXT: .cfi_restore z8
+; CHECK-NEXT: .cfi_restore z9
+; CHECK-NEXT: .cfi_restore z10
+; CHECK-NEXT: .cfi_restore z11
+; CHECK-NEXT: .cfi_restore z12
+; CHECK-NEXT: .cfi_restore z13
+; CHECK-NEXT: .cfi_restore z14
+; CHECK-NEXT: .cfi_restore z15
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ call void asm sideeffect "", "~{p8},~{z8},~{z9},~{z10},~{z11},~{z12},~{z13},~{z14},~{z15},~{z16},~{z17},~{z18},~{z19},~{z20},~{z21},~{z22},~{z23}" ()
+ ret void
+}
+
+; A SVE vector and a 16-byte fixed size object.
+define void @sve_1_vector_16_arr(ptr %out) #0 {
+; CHECK-LABEL: sve_1_vector_16_arr:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: addvl sp, sp, #-1
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0c, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 8 * VG
+; CHECK-NEXT: str xzr, [sp], #-16
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0c, 0x8f, 0x00, 0x11, 0x20, 0x22, 0x11, 0x08, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 32 + 8 * VG
+; CHECK-NEXT: addvl sp, sp, #1
+; CHECK-NEXT: .cfi_def_cfa wsp, 32
+; CHECK-NEXT: add sp, sp, #16
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %vec = alloca <vscale x 4 x float>, align 16
+ %arr = alloca i8, i64 16, align 1
+ ret void
+}
+
+; A large SVE stack object and a large stack slot, both of which need probing.
+; TODO: This could be optimised by combining the fixed-size offset into the
+; loop.
+define void @sve_1_vector_4096_arr(ptr %out) #0 {
+; CHECK-LABEL: sve_1_vector_4096_arr:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: addvl x9, sp, #-32
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x79, 0x00, 0x11, 0x10, 0x22, 0x11, 0x80, 0x02, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $x9 + 16 + 256 * VG
+; CHECK-NEXT: addvl x9, x9, #-32
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x79, 0x00, 0x11, 0x10, 0x22, 0x11, 0x80, 0x04, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $x9 + 16 + 512 * VG
+; CHECK-NEXT: .LBB11_1: // %entry
+; CHECK-NEXT: // =>This Inner Loop Header: Depth=1
+; CHECK-NEXT: sub sp, sp, #1, lsl #12 // =4096
+; CHECK-NEXT: cmp sp, x9
+; CHECK-NEXT: b.le .LBB11_3
+; CHECK-NEXT: // %bb.2: // %entry
+; CHECK-NEXT: // in Loop: Header=BB11_1 Depth=1
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: b .LBB11_1
+; CHECK-NEXT: .LBB11_3: // %entry
+; CHECK-NEXT: str xzr, [x9]
+; CHECK-NEXT: mov sp, x9
+; CHECK-NEXT: .cfi_def_cfa_register wsp
+; CHECK-NEXT: sub sp, sp, #1, lsl #12 // =4096
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0e, 0x8f, 0x00, 0x11, 0x90, 0x20, 0x22, 0x11, 0x80, 0x04, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 4112 + 512 * VG
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: sub sp, sp, #1, lsl #12 // =4096
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0f, 0x8f, 0x00, 0x11, 0x90, 0xc0, 0x00, 0x22, 0x11, 0x80, 0x04, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 8208 + 512 * VG
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: sub sp, sp, #1, lsl #12 // =4096
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0f, 0x8f, 0x00, 0x11, 0x90, 0xe0, 0x00, 0x22, 0x11, 0x80, 0x04, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 12304 + 512 * VG
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: addvl sp, sp, #31
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0f, 0x8f, 0x00, 0x11, 0x90, 0xe0, 0x00, 0x22, 0x11, 0x88, 0x02, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 12304 + 264 * VG
+; CHECK-NEXT: addvl sp, sp, #31
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0e, 0x8f, 0x00, 0x11, 0x90, 0xe0, 0x00, 0x22, 0x11, 0x10, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 12304 + 16 * VG
+; CHECK-NEXT: addvl sp, sp, #2
+; CHECK-NEXT: .cfi_def_cfa wsp, 12304
+; CHECK-NEXT: add sp, sp, #3, lsl #12 // =12288
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %vec = alloca <vscale x 256 x float>, align 16
+ %arr = alloca i8, i64 12288, align 1
+ ret void
+}
+
+; Not tested: SVE stack objects with alignment >16 bytes, which isn't currently
+; supported even without stack-probing.
+
+; An SVE vector, and a 16-byte fixed size object, which
+; has a large alignment requirement.
+define void @sve_1_vector_16_arr_align_8192(ptr %out) #0 {
+; CHECK-LABEL: sve_1_vector_16_arr_align_8192:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: stp x29, x30, [sp, #-16]! // 16-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: mov x29, sp
+; CHECK-NEXT: .cfi_def_cfa w29, 16
+; CHECK-NEXT: .cfi_offset w30, -8
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: addvl sp, sp, #-1
+; CHECK-NEXT: sub x9, sp, #1, lsl #12 // =4096
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: sub x9, x9, #4080
+; CHECK-NEXT: and x9, x9, #0xffffffffffffe000
+; CHECK-NEXT: .LBB12_1: // %entry
+; CHECK-NEXT: // =>This Inner Loop Header: Depth=1
+; CHECK-NEXT: sub sp, sp, #1, lsl #12 // =4096
+; CHECK-NEXT: cmp sp, x9
+; CHECK-NEXT: b.le .LBB12_3
+; CHECK-NEXT: // %bb.2: // %entry
+; CHECK-NEXT: // in Loop: Header=BB12_1 Depth=1
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: b .LBB12_1
+; CHECK-NEXT: .LBB12_3: // %entry
+; CHECK-NEXT: str xzr, [x9]
+; CHECK-NEXT: mov sp, x9
+; CHECK-NEXT: mov sp, x29
+; CHECK-NEXT: .cfi_def_cfa wsp, 16
+; CHECK-NEXT: ldp x29, x30, [sp], #16 // 16-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w30
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %vec = alloca <vscale x 4 x float>, align 16
+ %arr = alloca i8, i64 16, align 8192
+ ret void
+}
+
+; With 64k guard pages, we can allocate bigger SVE space without a probing loop.
+define void @sve_1024_64k_guard(ptr %out) #0 "stack-probe-size"="65536" {
+; CHECK-LABEL: sve_1024_64k_guard:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: addvl sp, sp, #-32
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0x80, 0x02, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 256 * VG
+; CHECK-NEXT: addvl sp, sp, #-32
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0x80, 0x04, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 512 * VG
+; CHECK-NEXT: addvl sp, sp, #-32
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0x80, 0x06, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 768 * VG
+; CHECK-NEXT: addvl sp, sp, #-32
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0x80, 0x08, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 1024 * VG
+; CHECK-NEXT: addvl sp, sp, #-32
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0x80, 0x0a, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 1280 * VG
+; CHECK-NEXT: addvl sp, sp, #-32
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0x80, 0x0c, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 1536 * VG
+; CHECK-NEXT: addvl sp, sp, #-32
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0x80, 0x0e, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 1792 * VG
+; CHECK-NEXT: addvl sp, sp, #-32
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0x80, 0x10, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 2048 * VG
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: addvl sp, sp, #31
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0x88, 0x0e, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 1800 * VG
+; CHECK-NEXT: addvl sp, sp, #31
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0x90, 0x0c, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 1552 * VG
+; CHECK-NEXT: addvl sp, sp, #31
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0x98, 0x0a, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 1304 * VG
+; CHECK-NEXT: addvl sp, sp, #31
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0xa0, 0x08, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 1056 * VG
+; CHECK-NEXT: addvl sp, sp, #31
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0xa8, 0x06, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 808 * VG
+; CHECK-NEXT: addvl sp, sp, #31
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0xb0, 0x04, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 560 * VG
+; CHECK-NEXT: addvl sp, sp, #31
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0xb8, 0x02, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 312 * VG
+; CHECK-NEXT: addvl sp, sp, #31
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0xc0, 0x00, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 64 * VG
+; CHECK-NEXT: addvl sp, sp, #8
+; CHECK-NEXT: .cfi_def_cfa wsp, 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %vec = alloca <vscale x 1024 x float>, align 16
+ ret void
+}
+
+define void @sve_1028_64k_guard(ptr %out) #0 "stack-probe-size"="65536" {
+; CHECK-LABEL: sve_1028_64k_guard:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: addvl x9, sp, #-32
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x79, 0x00, 0x11, 0x10, 0x22, 0x11, 0x80, 0x02, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $x9 + 16 + 256 * VG
+; CHECK-NEXT: addvl x9, x9, #-32
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x79, 0x00, 0x11, 0x10, 0x22, 0x11, 0x80, 0x04, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $x9 + 16 + 512 * VG
+; CHECK-NEXT: addvl x9, x9, #-32
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x79, 0x00, 0x11, 0x10, 0x22, 0x11, 0x80, 0x06, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $x9 + 16 + 768 * VG
+; CHECK-NEXT: addvl x9, x9, #-32
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x79, 0x00, 0x11, 0x10, 0x22, 0x11, 0x80, 0x08, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $x9 + 16 + 1024 * VG
+; CHECK-NEXT: addvl x9, x9, #-32
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x79, 0x00, 0x11, 0x10, 0x22, 0x11, 0x80, 0x0a, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $x9 + 16 + 1280 * VG
+; CHECK-NEXT: addvl x9, x9, #-32
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x79, 0x00, 0x11, 0x10, 0x22, 0x11, 0x80, 0x0c, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $x9 + 16 + 1536 * VG
+; CHECK-NEXT: addvl x9, x9, #-32
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x79, 0x00, 0x11, 0x10, 0x22, 0x11, 0x80, 0x0e, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $x9 + 16 + 1792 * VG
+; CHECK-NEXT: addvl x9, x9, #-32
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x79, 0x00, 0x11, 0x10, 0x22, 0x11, 0x80, 0x10, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $x9 + 16 + 2048 * VG
+; CHECK-NEXT: addvl x9, x9, #-1
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x79, 0x00, 0x11, 0x10, 0x22, 0x11, 0x88, 0x10, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $x9 + 16 + 2056 * VG
+; CHECK-NEXT: .LBB14_1: // %entry
+; CHECK-NEXT: // =>This Inner Loop Header: Depth=1
+; CHECK-NEXT: sub sp, sp, #16, lsl #12 // =65536
+; CHECK-NEXT: cmp sp, x9
+; CHECK-NEXT: b.le .LBB14_3
+; CHECK-NEXT: // %bb.2: // %entry
+; CHECK-NEXT: // in Loop: Header=BB14_1 Depth=1
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: b .LBB14_1
+; CHECK-NEXT: .LBB14_3: // %entry
+; CHECK-NEXT: str xzr, [x9]
+; CHECK-NEXT: mov sp, x9
+; CHECK-NEXT: .cfi_def_cfa_register wsp
+; CHECK-NEXT: addvl sp, sp, #31
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0x90, 0x0e, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 1808 * VG
+; CHECK-NEXT: addvl sp, sp, #31
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0x98, 0x0c, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 1560 * VG
+; CHECK-NEXT: addvl sp, sp, #31
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0xa0, 0x0a, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 1312 * VG
+; CHECK-NEXT: addvl sp, sp, #31
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0xa8, 0x08, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 1064 * VG
+; CHECK-NEXT: addvl sp, sp, #31
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0xb0, 0x06, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 816 * VG
+; CHECK-NEXT: addvl sp, sp, #31
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0xb8, 0x04, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 568 * VG
+; CHECK-NEXT: addvl sp, sp, #31
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0xc0, 0x02, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 320 * VG
+; CHECK-NEXT: addvl sp, sp, #31
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0xc8, 0x00, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 72 * VG
+; CHECK-NEXT: addvl sp, sp, #9
+; CHECK-NEXT: .cfi_def_cfa wsp, 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %vec = alloca <vscale x 1024 x float>, align 16
+ %vec1 = alloca <vscale x 4 x float>, align 16
+ ret void
+}
+
+attributes #0 = { uwtable(async) "probe-stack"="inline-asm" "frame-pointer"="none" "target-features"="+sve" }
diff --git a/llvm/test/CodeGen/AArch64/stack-probing.ll b/llvm/test/CodeGen/AArch64/stack-probing.ll
new file mode 100644
index 000000000000000..fb4c4e00b719197
--- /dev/null
+++ b/llvm/test/CodeGen/AArch64/stack-probing.ll
@@ -0,0 +1,474 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple aarch64-none-eabi < %s -verify-machineinstrs -enable-post-misched=false | FileCheck %s
+; RUN: llc -mtriple aarch64-none-eabi < %s -verify-machineinstrs -enable-post-misched=false -global-isel | FileCheck %s
+
+; Tests for prolog sequences for stack probing, when using a 4KiB stack guard.
+
+; Small stack frame, no probing required.
+define void @static_64(ptr %out) #0 {
+; CHECK-LABEL: static_64:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: sub sp, sp, #64
+; CHECK-NEXT: .cfi_def_cfa_offset 64
+; CHECK-NEXT: mov x8, sp
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: add sp, sp, #64
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: ret
+entry:
+ %v = alloca i8, i64 64, align 1
+ store ptr %v, ptr %out, align 8
+ ret void
+}
+
+; At 256 bytes we start to always create a frame pointer. No frame smaller then
+; this needs a probe, so we can use the saving of at least one CSR as a probe
+; at the top of our frame.
+define void @static_256(ptr %out) #0 {
+; CHECK-LABEL: static_256:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: sub sp, sp, #272
+; CHECK-NEXT: .cfi_def_cfa_offset 272
+; CHECK-NEXT: str x29, [sp, #256] // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: mov x8, sp
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: add sp, sp, #272
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %v = alloca i8, i64 256, align 1
+ store ptr %v, ptr %out, align 8
+ ret void
+}
+
+; At 1024 bytes, this is the largest frame which doesn't need probing.
+define void @static_1024(ptr %out) #0 {
+; CHECK-LABEL: static_1024:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: sub sp, sp, #1024
+; CHECK-NEXT: .cfi_def_cfa_offset 1040
+; CHECK-NEXT: mov x8, sp
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: add sp, sp, #1024
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %v = alloca i8, i64 1024, align 1
+ store ptr %v, ptr %out, align 8
+ ret void
+}
+
+; At 1024+16 bytes, this is the smallest frame which needs probing.
+define void @static_1040(ptr %out) #0 {
+; CHECK-LABEL: static_1040:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: sub sp, sp, #1040
+; CHECK-NEXT: .cfi_def_cfa_offset 1056
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: mov x8, sp
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: add sp, sp, #1040
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %v = alloca i8, i64 1040, align 1
+ store ptr %v, ptr %out, align 8
+ ret void
+}
+
+; 4k bytes is the largest frame we can probe in one go.
+define void @static_4096(ptr %out) #0 {
+; CHECK-LABEL: static_4096:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: sub sp, sp, #1, lsl #12 // =4096
+; CHECK-NEXT: .cfi_def_cfa_offset 4112
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: mov x8, sp
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: add sp, sp, #1, lsl #12 // =4096
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %v = alloca i8, i64 4096, align 1
+ store ptr %v, ptr %out, align 8
+ ret void
+}
+
+; 4k+16 bytes, still needs just one probe.
+define void @static_4112(ptr %out) #0 {
+; CHECK-LABEL: static_4112:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: sub sp, sp, #1, lsl #12 // =4096
+; CHECK-NEXT: .cfi_def_cfa_offset 4112
+; CHECK-NEXT: str xzr, [sp], #-16
+; CHECK-NEXT: .cfi_def_cfa_offset 4128
+; CHECK-NEXT: mov x8, sp
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: add sp, sp, #1, lsl #12 // =4096
+; CHECK-NEXT: .cfi_def_cfa_offset 32
+; CHECK-NEXT: add sp, sp, #16
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %v = alloca i8, i64 4112, align 1
+ store ptr %v, ptr %out, align 8
+ ret void
+}
+
+; 4k+1024 bytes, the largest frame which needs just one probe.
+define void @static_5120(ptr %out) #0 {
+; CHECK-LABEL: static_5120:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: sub sp, sp, #1, lsl #12 // =4096
+; CHECK-NEXT: .cfi_def_cfa_offset 4112
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: sub sp, sp, #1024
+; CHECK-NEXT: .cfi_def_cfa_offset 5136
+; CHECK-NEXT: mov x8, sp
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: add sp, sp, #1, lsl #12 // =4096
+; CHECK-NEXT: .cfi_def_cfa_offset 1040
+; CHECK-NEXT: add sp, sp, #1024
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %v = alloca i8, i64 5120, align 1
+ store ptr %v, ptr %out, align 8
+ ret void
+}
+
+; 4k+1024+16, the smallest frame which needs two probes.
+define void @static_5136(ptr %out) #0 {
+; CHECK-LABEL: static_5136:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: sub sp, sp, #1, lsl #12 // =4096
+; CHECK-NEXT: .cfi_def_cfa_offset 4112
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: sub sp, sp, #1040
+; CHECK-NEXT: .cfi_def_cfa_offset 5152
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: mov x8, sp
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: add sp, sp, #1, lsl #12 // =4096
+; CHECK-NEXT: .cfi_def_cfa_offset 1056
+; CHECK-NEXT: add sp, sp, #1040
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %v = alloca i8, i64 5136, align 1
+ store ptr %v, ptr %out, align 8
+ ret void
+}
+
+; 2*4k+1024, the largest frame needing two probes
+define void @static_9216(ptr %out) #0 {
+; CHECK-LABEL: static_9216:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: sub sp, sp, #1, lsl #12 // =4096
+; CHECK-NEXT: .cfi_def_cfa_offset 4112
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: sub sp, sp, #1, lsl #12 // =4096
+; CHECK-NEXT: .cfi_def_cfa_offset 8208
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: sub sp, sp, #1024
+; CHECK-NEXT: .cfi_def_cfa_offset 9232
+; CHECK-NEXT: mov x8, sp
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: add sp, sp, #2, lsl #12 // =8192
+; CHECK-NEXT: .cfi_def_cfa_offset 1040
+; CHECK-NEXT: add sp, sp, #1024
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %v = alloca i8, i64 9216, align 1
+ store ptr %v, ptr %out, align 8
+ ret void
+}
+
+; 5*4k-16, the largest frame probed without a loop
+define void @static_20464(ptr %out) #0 {
+; CHECK-LABEL: static_20464:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: sub sp, sp, #1, lsl #12 // =4096
+; CHECK-NEXT: .cfi_def_cfa_offset 4112
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: sub sp, sp, #1, lsl #12 // =4096
+; CHECK-NEXT: .cfi_def_cfa_offset 8208
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: sub sp, sp, #1, lsl #12 // =4096
+; CHECK-NEXT: .cfi_def_cfa_offset 12304
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: sub sp, sp, #1, lsl #12 // =4096
+; CHECK-NEXT: .cfi_def_cfa_offset 16400
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: sub sp, sp, #4080
+; CHECK-NEXT: .cfi_def_cfa_offset 20480
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: mov x8, sp
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: add sp, sp, #4, lsl #12 // =16384
+; CHECK-NEXT: .cfi_def_cfa_offset 4096
+; CHECK-NEXT: add sp, sp, #4080
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %v = alloca i8, i64 20464, align 1
+ store ptr %v, ptr %out, align 8
+ ret void
+}
+
+; 5*4k, the smallest frame probed with a loop
+define void @static_20480(ptr %out) #0 {
+; CHECK-LABEL: static_20480:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: sub x9, sp, #5, lsl #12 // =20480
+; CHECK-NEXT: .cfi_def_cfa w9, 20496
+; CHECK-NEXT: .LBB10_1: // %entry
+; CHECK-NEXT: // =>This Inner Loop Header: Depth=1
+; CHECK-NEXT: sub sp, sp, #1, lsl #12 // =4096
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: cmp sp, x9
+; CHECK-NEXT: b.ne .LBB10_1
+; CHECK-NEXT: // %bb.2: // %entry
+; CHECK-NEXT: .cfi_def_cfa_register wsp
+; CHECK-NEXT: mov x8, sp
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: add sp, sp, #5, lsl #12 // =20480
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %v = alloca i8, i64 20480, align 1
+ store ptr %v, ptr %out, align 8
+ ret void
+}
+
+; 5*4k + 1024, large enough to use a loop, but not a multiple of 4KiB
+; so has a reminder, but no extra probe.
+define void @static_21504(ptr %out) #0 {
+; CHECK-LABEL: static_21504:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: sub x9, sp, #5, lsl #12 // =20480
+; CHECK-NEXT: .cfi_def_cfa w9, 20496
+; CHECK-NEXT: .LBB11_1: // %entry
+; CHECK-NEXT: // =>This Inner Loop Header: Depth=1
+; CHECK-NEXT: sub sp, sp, #1, lsl #12 // =4096
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: cmp sp, x9
+; CHECK-NEXT: b.ne .LBB11_1
+; CHECK-NEXT: // %bb.2: // %entry
+; CHECK-NEXT: .cfi_def_cfa_register wsp
+; CHECK-NEXT: sub sp, sp, #1024
+; CHECK-NEXT: .cfi_def_cfa_offset 21520
+; CHECK-NEXT: mov x8, sp
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: add sp, sp, #5, lsl #12 // =20480
+; CHECK-NEXT: .cfi_def_cfa_offset 1040
+; CHECK-NEXT: add sp, sp, #1024
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %v = alloca i8, i64 21504, align 1
+ store ptr %v, ptr %out, align 8
+ ret void
+}
+
+; 5*4k+1040, large enough to use a loop, has a reminder and
+; an extra probe.
+define void @static_21520(ptr %out) #0 {
+; CHECK-LABEL: static_21520:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: sub x9, sp, #5, lsl #12 // =20480
+; CHECK-NEXT: .cfi_def_cfa w9, 20496
+; CHECK-NEXT: .LBB12_1: // %entry
+; CHECK-NEXT: // =>This Inner Loop Header: Depth=1
+; CHECK-NEXT: sub sp, sp, #1, lsl #12 // =4096
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: cmp sp, x9
+; CHECK-NEXT: b.ne .LBB12_1
+; CHECK-NEXT: // %bb.2: // %entry
+; CHECK-NEXT: .cfi_def_cfa_register wsp
+; CHECK-NEXT: sub sp, sp, #1040
+; CHECK-NEXT: .cfi_def_cfa_offset 21536
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: mov x8, sp
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: add sp, sp, #5, lsl #12 // =20480
+; CHECK-NEXT: .cfi_def_cfa_offset 1056
+; CHECK-NEXT: add sp, sp, #1040
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %v = alloca i8, i64 21520, align 1
+ store ptr %v, ptr %out, align 8
+ ret void
+}
+
+; A small allocation, but with a very large alignment requirement. We do this
+; by moving SP far enough that a sufficiently-aligned block will exist
+; somewhere in the stack frame, so must probe the whole of that larger SP move.
+define void @static_16_align_8192(ptr %out) #0 {
+; CHECK-LABEL: static_16_align_8192:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: stp x29, x30, [sp, #-16]! // 16-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: mov x29, sp
+; CHECK-NEXT: .cfi_def_cfa w29, 16
+; CHECK-NEXT: .cfi_offset w30, -8
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: sub x9, sp, #1, lsl #12 // =4096
+; CHECK-NEXT: sub x9, x9, #4080
+; CHECK-NEXT: and x9, x9, #0xffffffffffffe000
+; CHECK-NEXT: .LBB13_1: // %entry
+; CHECK-NEXT: // =>This Inner Loop Header: Depth=1
+; CHECK-NEXT: sub sp, sp, #1, lsl #12 // =4096
+; CHECK-NEXT: cmp sp, x9
+; CHECK-NEXT: b.le .LBB13_3
+; CHECK-NEXT: // %bb.2: // %entry
+; CHECK-NEXT: // in Loop: Header=BB13_1 Depth=1
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: b .LBB13_1
+; CHECK-NEXT: .LBB13_3: // %entry
+; CHECK-NEXT: str xzr, [x9]
+; CHECK-NEXT: mov sp, x9
+; CHECK-NEXT: mov x8, sp
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: mov sp, x29
+; CHECK-NEXT: .cfi_def_cfa wsp, 16
+; CHECK-NEXT: ldp x29, x30, [sp], #16 // 16-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w30
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %v = alloca i8, i64 16, align 8192
+ store ptr %v, ptr %out, align 8
+ ret void
+}
+
+; A small allocation with a very large alignment requirement, but
+; nevertheless small enough as to not need a loop.
+define void @static_16_align_2048(ptr %out) #0 {
+; CHECK-LABEL: static_16_align_2048:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: stp x29, x30, [sp, #-16]! // 16-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: mov x29, sp
+; CHECK-NEXT: .cfi_def_cfa w29, 16
+; CHECK-NEXT: .cfi_offset w30, -8
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: sub x9, sp, #2032
+; CHECK-NEXT: and sp, x9, #0xfffffffffffff800
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: mov x8, sp
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: mov sp, x29
+; CHECK-NEXT: .cfi_def_cfa wsp, 16
+; CHECK-NEXT: ldp x29, x30, [sp], #16 // 16-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w30
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %v = alloca i8, i64 16, align 2048
+ store ptr %v, ptr %out, align 8
+ ret void
+}
+
+; A large(-ish) allocation with a very large alignment requirement, but
+; nevertheless small enough as to not need a loop.
+define void @static_2032_align_2048(ptr %out) #0 {
+; CHECK-LABEL: static_2032_align_2048:
+; CHECK: // %bb.0: // %entry
+; CHECK-NEXT: stp x29, x30, [sp, #-16]! // 16-byte Folded Spill
+; CHECK-NEXT: .cfi_def_cfa_offset 16
+; CHECK-NEXT: mov x29, sp
+; CHECK-NEXT: .cfi_def_cfa w29, 16
+; CHECK-NEXT: .cfi_offset w30, -8
+; CHECK-NEXT: .cfi_offset w29, -16
+; CHECK-NEXT: sub x9, sp, #2032
+; CHECK-NEXT: and sp, x9, #0xfffffffffffff800
+; CHECK-NEXT: str xzr, [sp]
+; CHECK-NEXT: mov x8, sp
+; CHECK-NEXT: str x8, [x0]
+; CHECK-NEXT: mov sp, x29
+; CHECK-NEXT: .cfi_def_cfa wsp, 16
+; CHECK-NEXT: ldp x29, x30, [sp], #16 // 16-byte Folded Reload
+; CHECK-NEXT: .cfi_def_cfa_offset 0
+; CHECK-NEXT: .cfi_restore w30
+; CHECK-NEXT: .cfi_restore w29
+; CHECK-NEXT: ret
+entry:
+ %v = alloca i8, i64 2032, align 2048
+ store ptr %v, ptr %out, align 8
+ ret void
+}
+
+attributes #0 = { uwtable(async) "probe-stack"="inline-asm" "frame-pointer"="none" }
More information about the llvm-commits
mailing list