This is an archive of the discontinued LLVM Phabricator instance.

Differential D106185

[Attributor] Introduce getPotentialCopiesOfStoredValue and use it
ClosedPublic

Authored by jdoerfert on Jul 16 2021, 1:23 PM.

Details

Reviewers
kuter
ggeorgakoudis
wsmoses
uenoku
homerdin
sstefan1
baziotis
Commits
rGadddd3dbdae0: [Attributor] Introduce getPotentialCopiesOfStoredValue and use it
Summary

This patch introduces getPotentialCopiesOfStoredValue which uses
AAPointerInfo to determine all "aliases" or "potential copies" of a
value that is stored into memory. This operation can fail but if it
succeeds it means we can visit all "uses" of a value even if it is
temporarily stored in memory.

There are two users for the function:

  1. Attributor::checkForAllUses which will now ignore the value use in a store if all "potential copies" can be identified and instead be visited. This allows various AAs, including AAPointerInfo itself, to look through memory.
  2. AANoCapture which uses a custom use tracking through the CaptureTracker interface and therefore needs to be thought explicitly.

Diff Detail

Event Timeline

jdoerfert created this revision.Jul 16 2021, 1:23 PM
Herald added a reviewer: uenoku. · View Herald TranscriptJul 16 2021, 1:23 PM
Herald added a reviewer: homerdin. · View Herald Transcript
Herald added subscribers: ormris, okura, uenoku and 2 others. · View Herald Transcript
jdoerfert requested review of this revision.Jul 16 2021, 1:23 PM
Herald added a reviewer: sstefan1. · View Herald TranscriptJul 16 2021, 1:23 PM
Herald added a reviewer: baziotis. · View Herald Transcript
Herald added a project: Restricted Project. · View Herald Transcript
Herald added subscribers: bbn, sstefan1. · View Herald Transcript
Harbormaster completed remote builds in B114581: Diff 359430.Jul 16 2021, 1:24 PM
jdoerfert mentioned this in D106188: [Attributor] Delete dead stores.Jul 16 2021, 1:44 PM
kuter added inline comments.Jul 16 2021, 4:48 PM
llvm/lib/Transforms/IPO/Attributor.cpp
348

Nit: typo

llvm/lib/Transforms/IPO/AttributorAttributes.cpp
3090–3095

From what I can see checkForAllUses can now return uses from different functions.
I am worried about some code not handling this case correctly.

For example, do you think this piece of code can cause a problem ?

jdoerfert added inline comments.Jul 16 2021, 5:48 PM
llvm/lib/Transforms/IPO/AttributorAttributes.cpp
3090–3095

Assuming AAReachability would handle interprocedural queries right (it will assert now) this would be fine.

That said, I find the UserI check above very suspect and I can not remember why this is supposed to be valid without following the user at least.

jdoerfert added a child revision: D106188: [Attributor] Delete dead stores.Jul 19 2021, 8:58 PM
jdoerfert updated this revision to Diff 360005.Jul 19 2021, 9:01 PM

Fixed spelling

Harbormaster completed remote builds in B115014: Diff 360005.Jul 19 2021, 9:01 PM
kuter accepted this revision.Jul 20 2021, 9:59 AM

LGTM

This revision is now accepted and ready to land.Jul 20 2021, 9:59 AM
This revision was landed with ongoing or failed builds.Jul 26 2021, 9:34 PM
Closed by commit rGadddd3dbdae0: [Attributor] Introduce getPotentialCopiesOfStoredValue and use it (authored by jdoerfert). · Explain Why
This revision was automatically updated to reflect the committed changes.
jdoerfert added a commit: rGadddd3dbdae0: [Attributor] Introduce getPotentialCopiesOfStoredValue and use it.
jdoerfert mentioned this in rG41bd26dff986: [Attributor] Delete dead stores.

Revision Contents

PathSize
llvm/
include/
llvm/
Transforms/
IPO/
14 lines
lib/
Transforms/
IPO/
98 lines
23 lines
test/
Transforms/
Attributor/
ArgumentPromotion/
11 lines
13 lines
IPConstantProp/
30 lines
21 lines
72 lines
OpenMP/
132 lines
36 lines

Diff 361896

llvm/include/llvm/Transforms/IPO/Attributor.h

Show First 20 Lines • Show All 179 Lines • ▼ Show 20 Lines
/// ///
/// \returns True if \p Objects contains all assumed underlying objects, and /// \returns True if \p Objects contains all assumed underlying objects, and
/// false if something went wrong and the objects could not be /// false if something went wrong and the objects could not be
/// determined. /// determined.
bool getAssumedUnderlyingObjects(Attributor &A, const Value &Ptr, bool getAssumedUnderlyingObjects(Attributor &A, const Value &Ptr,
SmallVectorImpl<Value *> &Objects, SmallVectorImpl<Value *> &Objects,
const AbstractAttribute &QueryingAA, const AbstractAttribute &QueryingAA,
const Instruction *CtxI); const Instruction *CtxI);
/// Collect all potential values of the one stored by \p SI into
/// \p PotentialCopies. That is, the only copies that were made via the
/// store are assumed to be known and all in \p PotentialCopies. Dependences
/// onto \p QueryingAA are properly tracked, \p UsedAssumedInformation will
/// inform the caller if assumed information was used.
///
/// \returns True if the assumed potential copies are all in \p PotentialCopies,
/// false if something went wrong and the copies could not be
/// determined.
bool getPotentialCopiesOfStoredValue(
Attributor &A, StoreInst &SI, SmallSetVector<Value *, 4> &PotentialCopies,
const AbstractAttribute &QueryingAA, bool &UsedAssumedInformation);
} // namespace AA } // namespace AA
/// The value passed to the line option that defines the maximal initialization /// The value passed to the line option that defines the maximal initialization
/// chain length. /// chain length.
extern unsigned MaxInitializationChainLength; extern unsigned MaxInitializationChainLength;
///{ ///{
enum class ChangeStatus { enum class ChangeStatus {
▲ Show 20 Lines • Show All 4,377 Lines • Show Last 20 Lines

llvm/lib/Transforms/IPO/Attributor.cpp

Show All 23 Lines
#include "llvm/Analysis/LazyValueInfo.h" #include "llvm/Analysis/LazyValueInfo.h"
#include "llvm/Analysis/MemorySSAUpdater.h" #include "llvm/Analysis/MemorySSAUpdater.h"
#include "llvm/Analysis/MustExecute.h" #include "llvm/Analysis/MustExecute.h"
#include "llvm/Analysis/ValueTracking.h" #include "llvm/Analysis/ValueTracking.h"
#include "llvm/IR/Attributes.h" #include "llvm/IR/Attributes.h"
#include "llvm/IR/Constant.h" #include "llvm/IR/Constant.h"
#include "llvm/IR/Constants.h" #include "llvm/IR/Constants.h"
#include "llvm/IR/GlobalValue.h" #include "llvm/IR/GlobalValue.h"
#include "llvm/IR/GlobalVariable.h"
#include "llvm/IR/IRBuilder.h" #include "llvm/IR/IRBuilder.h"
#include "llvm/IR/Instruction.h" #include "llvm/IR/Instruction.h"
#include "llvm/IR/IntrinsicInst.h" #include "llvm/IR/IntrinsicInst.h"
#include "llvm/IR/NoFolder.h" #include "llvm/IR/NoFolder.h"
#include "llvm/IR/ValueHandle.h" #include "llvm/IR/ValueHandle.h"
#include "llvm/IR/Verifier.h" #include "llvm/IR/Verifier.h"
#include "llvm/InitializePasses.h" #include "llvm/InitializePasses.h"
#include "llvm/Support/Casting.h" #include "llvm/Support/Casting.h"
▲ Show 20 Lines • Show All 236 Lines • ▼ Show 20 Lines if (isa_and_nonnull<UndefValue>(*A))
return getWithType(**B, *Ty); return getWithType(**B, *Ty);
if (isa<UndefValue>(*B)) if (isa<UndefValue>(*B))
return A; return A;
if (*A && *B && *A == getWithType(**B, *Ty)) if (*A && *B && *A == getWithType(**B, *Ty))
return A; return A;
return nullptr; return nullptr;
} }
bool AA::getPotentialCopiesOfStoredValue(
Attributor &A, StoreInst &SI, SmallSetVector<Value *, 4> &PotentialCopies,
const AbstractAttribute &QueryingAA, bool &UsedAssumedInformation) {
Value &Ptr = *SI.getPointerOperand();
SmallVector<Value *, 8> Objects;
if (!AA::getAssumedUnderlyingObjects(A, Ptr, Objects, QueryingAA, &SI)) {
LLVM_DEBUG(
dbgs() << "Underlying objects stored into could not be determined\n";);
return false;
}
SmallVector<const AAPointerInfo *> PIs;
SmallVector<Value *> NewCopies;
for (Value *Obj : Objects) {
LLVM_DEBUG(dbgs() << "Visit underlying object " << *Obj << "\n");
if (isa<UndefValue>(Obj))
continue;
if (isa<ConstantPointerNull>(Obj)) {
// A null pointer access can be undefined but any offset from null may
// be OK. We do not try to optimize the latter.
if (!NullPointerIsDefined(SI.getFunction(),
Ptr.getType()->getPointerAddressSpace()) &&
A.getAssumedSimplified(Ptr, QueryingAA, UsedAssumedInformation) ==
Obj)
continue;
LLVM_DEBUG(
dbgs() << "Underlying object is a valid nullptr, giving up.\n";);
return false;
}
if (!isa<AllocaInst>(Obj) && !isa<GlobalVariable>(Obj)) {
LLVM_DEBUG(dbgs() << "Underlying object is not supported yet: " << *Obj
<< "\n";);
return false;
}
if (auto *GV = dyn_cast<GlobalVariable>(Obj))
if (!GV->hasLocalLinkage()) {
LLVM_DEBUG(dbgs() << "Underlying object is global with external "
"linkage, not supported yet: "
<< *Obj << "\n";);
return false;
}
auto CheckAccess = [&](const AAPointerInfo::Access &Acc, bool IsExact) {
if (!Acc.isRead())
return true;
auto *LI = dyn_cast<LoadInst>(Acc.getRemoteInst());
if (!LI) {
LLVM_DEBUG(dbgs() << "Underlying object read through a non-load "
"instruction not supported yet: "
<< *Acc.getRemoteInst() << "\n";);
return false;
}
NewCopies.push_back(LI);
return true;
};
auto &PI = A.getAAFor<AAPointerInfo>(QueryingAA, IRPosition::value(*Obj),
DepClassTy::NONE);
if (!PI.forallInterferingAccesses(SI, CheckAccess)) {
LLVM_DEBUG(
dbgs()
<< "Failed to verify all interfering accesses for underlying object: "
kuterUnsubmitted
Not Done
ReplyInline Actions

Nit: typo

kuter: Nit: typo
<< *Obj << "\n");
return false;
}
PIs.push_back(&PI);
}
for (auto *PI : PIs) {
if (!PI->getState().isAtFixpoint())
UsedAssumedInformation = true;
A.recordDependence(*PI, QueryingAA, DepClassTy::OPTIONAL);
}
PotentialCopies.insert(NewCopies.begin(), NewCopies.end());
return true;
}
/// Return true if \p New is equal or worse than \p Old. /// Return true if \p New is equal or worse than \p Old.
static bool isEqualOrWorse(const Attribute &New, const Attribute &Old) { static bool isEqualOrWorse(const Attribute &New, const Attribute &Old) {
if (!Old.isIntAttribute()) if (!Old.isIntAttribute())
return true; return true;
return Old.getValueAsInt() >= New.getValueAsInt(); return Old.getValueAsInt() >= New.getValueAsInt();
} }
▲ Show 20 Lines • Show All 660 Lines • ▼ Show 20 Lines if (isAssumedDead(*U, &QueryingAA, LivenessAA, UsedAssumedInformation,
LLVM_DEBUG(dbgs() << "[Attributor] Dead use, skip!\n"); LLVM_DEBUG(dbgs() << "[Attributor] Dead use, skip!\n");
continue; continue;
} }
if (U->getUser()->isDroppable()) { if (U->getUser()->isDroppable()) {
LLVM_DEBUG(dbgs() << "[Attributor] Droppable user, skip!\n"); LLVM_DEBUG(dbgs() << "[Attributor] Droppable user, skip!\n");
continue; continue;
} }
if (auto *SI = dyn_cast<StoreInst>(U->getUser())) {
if (&SI->getOperandUse(0) == U) {
SmallSetVector<Value *, 4> PotentialCopies;
if (AA::getPotentialCopiesOfStoredValue(*this, *SI, PotentialCopies,
QueryingAA,
UsedAssumedInformation)) {
LLVM_DEBUG(dbgs() << "[Attributor] Value is stored, continue with "
<< PotentialCopies.size()
<< " potential copies instead!\n");
for (Value *PotentialCopy : PotentialCopies)
for (const Use &U : PotentialCopy->uses())
Worklist.push_back(&U);
continue;
}
}
}
bool Follow = false; bool Follow = false;
if (!Pred(*U, Follow)) if (!Pred(*U, Follow))
return false; return false;
if (!Follow) if (!Follow)
continue; continue;
for (const Use &UU : U->getUser()->uses()) for (const Use &UU : U->getUser()->uses())
Worklist.push_back(&UU); Worklist.push_back(&UU);
} }
▲ Show 20 Lines • Show All 1,972 Lines • Show Last 20 Lines

llvm/lib/Transforms/IPO/AttributorAttributes.cpp

  • This file is larger than 256 KB, so syntax highlighting is disabled by default.
Show First 20 Lines • Show All 3,081 Lines • ▼ Show 20 Lines auto UsePred = [&](const Use &U, bool &Follow) -> bool {
// If UserI is the curr instruction and there is a single potential use of // If UserI is the curr instruction and there is a single potential use of
// the value in UserI we allow the use. // the value in UserI we allow the use.
// TODO: We should inspect the operands and allow those that cannot alias // TODO: We should inspect the operands and allow those that cannot alias
// with the value. // with the value.
if (UserI == getCtxI() && UserI->getNumOperands() == 1) if (UserI == getCtxI() && UserI->getNumOperands() == 1)
return true; return true;
if (ScopeFn) { if (ScopeFn) {
const auto &ReachabilityAA = A.getAAFor<AAReachability>( const auto &ReachabilityAA = A.getAAFor<AAReachability>(
*this, IRPosition::function(*ScopeFn), DepClassTy::OPTIONAL); *this, IRPosition::function(*ScopeFn), DepClassTy::OPTIONAL);
if (!ReachabilityAA.isAssumedReachable(A, *UserI, *getCtxI())) if (!ReachabilityAA.isAssumedReachable(A, *UserI, *getCtxI()))
return true; return true;
kuterUnsubmitted
Not Done
ReplyInline Actions

From what I can see checkForAllUses can now return uses from different functions.
I am worried about some code not handling this case correctly.

For example, do you think this piece of code can cause a problem ?

kuter: From what I can see `checkForAllUses` can now return uses from different functions. I am…
jdoerfertAuthorUnsubmitted
Done
ReplyInline Actions

Assuming AAReachability would handle interprocedural queries right (it will assert now) this would be fine.

That said, I find the UserI check above very suspect and I can not remember why this is supposed to be valid without following the user at least.

jdoerfert: Assuming `AAReachability` would handle interprocedural queries right (it will assert now) this…
if (auto *CB = dyn_cast<CallBase>(UserI)) { if (auto *CB = dyn_cast<CallBase>(UserI)) {
if (CB->isArgOperand(&U)) { if (CB->isArgOperand(&U)) {
unsigned ArgNo = CB->getArgOperandNo(&U); unsigned ArgNo = CB->getArgOperandNo(&U);
const auto &NoCaptureAA = A.getAAFor<AANoCapture>( const auto &NoCaptureAA = A.getAAFor<AANoCapture>(
*this, IRPosition::callsite_argument(*CB, ArgNo), *this, IRPosition::callsite_argument(*CB, ArgNo),
DepClassTy::OPTIONAL); DepClassTy::OPTIONAL);
▲ Show 20 Lines • Show All 1,627 Lines • ▼ Show 20 Lines bool captured(const Use *U) override {
} }
// Deal with ptr2int by following uses. // Deal with ptr2int by following uses.
if (isa<PtrToIntInst>(UInst)) { if (isa<PtrToIntInst>(UInst)) {
LLVM_DEBUG(dbgs() << " - ptr2int assume the worst!\n"); LLVM_DEBUG(dbgs() << " - ptr2int assume the worst!\n");
return valueMayBeCaptured(UInst); return valueMayBeCaptured(UInst);
} }
// For stores we check if we can follow the value through memory or not.
if (auto *SI = dyn_cast<StoreInst>(UInst)) {
if (SI->isVolatile())
return isCapturedIn(/* Memory */ true, /* Integer */ false,
/* Return */ false);
bool UsedAssumedInformation = false;
if (!AA::getPotentialCopiesOfStoredValue(
A, *SI, PotentialCopies, NoCaptureAA, UsedAssumedInformation))
return isCapturedIn(/* Memory */ true, /* Integer */ false,
/* Return */ false);
// Not captured directly, potential copies will be checked.
return isCapturedIn(/* Memory */ false, /* Integer */ false,
/* Return */ false);
}
// Explicitly catch return instructions. // Explicitly catch return instructions.
if (isa<ReturnInst>(UInst)) if (isa<ReturnInst>(UInst)) {
if (UInst->getFunction() == NoCaptureAA.getAnchorScope())
return isCapturedIn(/* Memory */ false, /* Integer */ false, return isCapturedIn(/* Memory */ false, /* Integer */ false,
/* Return */ true); /* Return */ true);
return isCapturedIn(/* Memory */ true, /* Integer */ true,
/* Return */ true);
}
// For now we only use special logic for call sites. However, the tracker // For now we only use special logic for call sites. However, the tracker
// itself knows about a lot of other non-capturing cases already. // itself knows about a lot of other non-capturing cases already.
auto *CB = dyn_cast<CallBase>(UInst); auto *CB = dyn_cast<CallBase>(UInst);
if (!CB || !CB->isArgOperand(U)) if (!CB || !CB->isArgOperand(U))
return isCapturedIn(/* Memory */ true, /* Integer */ true, return isCapturedIn(/* Memory */ true, /* Integer */ true,
/* Return */ true); /* Return */ true);
▲ Show 20 Lines • Show All 4,814 Lines • Show Last 20 Lines

llvm/test/Transforms/Attributor/ArgumentPromotion/byval-2.ll

Show All 17 Lines
; IS__CGSCC_OPM-NEXT: store i32 0, i32* [[X]], align 4 ; IS__CGSCC_OPM-NEXT: store i32 0, i32* [[X]], align 4
; IS__CGSCC_OPM-NEXT: ret void ; IS__CGSCC_OPM-NEXT: ret void
; ;
; IS__CGSCC_NPM: Function Attrs: nofree norecurse nosync nounwind readnone willreturn ; IS__CGSCC_NPM: Function Attrs: nofree norecurse nosync nounwind readnone willreturn
; IS__CGSCC_NPM-LABEL: define {{[^@]+}}@f ; IS__CGSCC_NPM-LABEL: define {{[^@]+}}@f
; IS__CGSCC_NPM-SAME: (i32 [[TMP0:%.*]], i64 [[TMP1:%.*]], i32 [[TMP2:%.*]]) #[[ATTR0:[0-9]+]] { ; IS__CGSCC_NPM-SAME: (i32 [[TMP0:%.*]], i64 [[TMP1:%.*]], i32 [[TMP2:%.*]]) #[[ATTR0:[0-9]+]] {
; IS__CGSCC_NPM-NEXT: entry: ; IS__CGSCC_NPM-NEXT: entry:
; IS__CGSCC_NPM-NEXT: [[X_PRIV:%.*]] = alloca i32, align 4 ; IS__CGSCC_NPM-NEXT: [[X_PRIV:%.*]] = alloca i32, align 4
; IS__CGSCC_NPM-NEXT: store i32 [[TMP2]], i32* [[X_PRIV]], align 4 ; IS__CGSCC_NPM-NEXT: store i32 undef, i32* [[X_PRIV]], align 4
; IS__CGSCC_NPM-NEXT: [[B_PRIV:%.*]] = alloca [[STRUCT_SS:%.*]], align 8 ; IS__CGSCC_NPM-NEXT: [[B_PRIV:%.*]] = alloca [[STRUCT_SS:%.*]], align 8
; IS__CGSCC_NPM-NEXT: [[B_PRIV_CAST:%.*]] = bitcast %struct.ss* [[B_PRIV]] to i32* ; IS__CGSCC_NPM-NEXT: [[B_PRIV_CAST:%.*]] = bitcast %struct.ss* [[B_PRIV]] to i32*
; IS__CGSCC_NPM-NEXT: store i32 1, i32* [[B_PRIV_CAST]], align 8 ; IS__CGSCC_NPM-NEXT: store i32 1, i32* [[B_PRIV_CAST]], align 8
; IS__CGSCC_NPM-NEXT: [[B_PRIV_0_1:%.*]] = getelementptr [[STRUCT_SS]], %struct.ss* [[B_PRIV]], i32 0, i32 1 ; IS__CGSCC_NPM-NEXT: [[B_PRIV_0_1:%.*]] = getelementptr [[STRUCT_SS]], %struct.ss* [[B_PRIV]], i32 0, i32 1
; IS__CGSCC_NPM-NEXT: store i64 2, i64* [[B_PRIV_0_1]], align 4 ; IS__CGSCC_NPM-NEXT: store i64 2, i64* [[B_PRIV_0_1]], align 4
; IS__CGSCC_NPM-NEXT: [[TMP:%.*]] = getelementptr [[STRUCT_SS]], %struct.ss* [[B_PRIV]], i32 0, i32 0 ; IS__CGSCC_NPM-NEXT: [[TMP:%.*]] = getelementptr [[STRUCT_SS]], %struct.ss* [[B_PRIV]], i32 0, i32 0
; IS__CGSCC_NPM-NEXT: [[TMP1:%.*]] = load i32, i32* [[TMP]], align 8 ; IS__CGSCC_NPM-NEXT: [[TMP1:%.*]] = load i32, i32* [[TMP]], align 8
; IS__CGSCC_NPM-NEXT: [[TMP2:%.*]] = add i32 [[TMP1]], 1 ; IS__CGSCC_NPM-NEXT: [[TMP2:%.*]] = add i32 [[TMP1]], 1
Show All 30 Lines
; IS__CGSCC_OPM-NEXT: entry: ; IS__CGSCC_OPM-NEXT: entry:
; IS__CGSCC_OPM-NEXT: [[S:%.*]] = alloca [[STRUCT_SS:%.*]], align 8 ; IS__CGSCC_OPM-NEXT: [[S:%.*]] = alloca [[STRUCT_SS:%.*]], align 8
; IS__CGSCC_OPM-NEXT: [[TMP1:%.*]] = getelementptr [[STRUCT_SS]], %struct.ss* [[S]], i32 0, i32 0 ; IS__CGSCC_OPM-NEXT: [[TMP1:%.*]] = getelementptr [[STRUCT_SS]], %struct.ss* [[S]], i32 0, i32 0
; IS__CGSCC_OPM-NEXT: store i32 1, i32* [[TMP1]], align 8 ; IS__CGSCC_OPM-NEXT: store i32 1, i32* [[TMP1]], align 8
; IS__CGSCC_OPM-NEXT: [[TMP4:%.*]] = getelementptr [[STRUCT_SS]], %struct.ss* [[S]], i32 0, i32 1 ; IS__CGSCC_OPM-NEXT: [[TMP4:%.*]] = getelementptr [[STRUCT_SS]], %struct.ss* [[S]], i32 0, i32 1
; IS__CGSCC_OPM-NEXT: store i64 2, i64* [[TMP4]], align 4 ; IS__CGSCC_OPM-NEXT: store i64 2, i64* [[TMP4]], align 4
; IS__CGSCC_OPM-NEXT: ret i32 0 ; IS__CGSCC_OPM-NEXT: ret i32 0
; ;
; IS__CGSCC_NPM: Function Attrs: argmemonly nofree norecurse nosync nounwind willreturn ; IS__CGSCC_NPM: Function Attrs: nofree norecurse nosync nounwind readnone willreturn
; IS__CGSCC_NPM-LABEL: define {{[^@]+}}@test ; IS__CGSCC_NPM-LABEL: define {{[^@]+}}@test
; IS__CGSCC_NPM-SAME: (i32* nocapture nofree noundef nonnull readonly align 4 dereferenceable(4) [[X:%.*]]) #[[ATTR1:[0-9]+]] { ; IS__CGSCC_NPM-SAME: (i32* nocapture nofree nonnull readnone align 4 dereferenceable(4) [[X:%.*]]) #[[ATTR0]] {
; IS__CGSCC_NPM-NEXT: entry: ; IS__CGSCC_NPM-NEXT: entry:
; IS__CGSCC_NPM-NEXT: [[S:%.*]] = alloca [[STRUCT_SS:%.*]], align 8 ; IS__CGSCC_NPM-NEXT: [[S:%.*]] = alloca [[STRUCT_SS:%.*]], align 8
; IS__CGSCC_NPM-NEXT: [[TMP1:%.*]] = getelementptr [[STRUCT_SS]], %struct.ss* [[S]], i32 0, i32 0 ; IS__CGSCC_NPM-NEXT: [[TMP1:%.*]] = getelementptr [[STRUCT_SS]], %struct.ss* [[S]], i32 0, i32 0
; IS__CGSCC_NPM-NEXT: store i32 1, i32* [[TMP1]], align 8 ; IS__CGSCC_NPM-NEXT: store i32 1, i32* [[TMP1]], align 8
; IS__CGSCC_NPM-NEXT: [[TMP4:%.*]] = getelementptr [[STRUCT_SS]], %struct.ss* [[S]], i32 0, i32 1 ; IS__CGSCC_NPM-NEXT: [[TMP4:%.*]] = getelementptr [[STRUCT_SS]], %struct.ss* [[S]], i32 0, i32 1
; IS__CGSCC_NPM-NEXT: store i64 2, i64* [[TMP4]], align 4 ; IS__CGSCC_NPM-NEXT: store i64 2, i64* [[TMP4]], align 4
; IS__CGSCC_NPM-NEXT: ret i32 0 ; IS__CGSCC_NPM-NEXT: ret i32 0
; ;
entry: entry:
%S = alloca %struct.ss %S = alloca %struct.ss
%tmp1 = getelementptr %struct.ss, %struct.ss* %S, i32 0, i32 0 %tmp1 = getelementptr %struct.ss, %struct.ss* %S, i32 0, i32 0
store i32 1, i32* %tmp1, align 8 store i32 1, i32* %tmp1, align 8
%tmp4 = getelementptr %struct.ss, %struct.ss* %S, i32 0, i32 1 %tmp4 = getelementptr %struct.ss, %struct.ss* %S, i32 0, i32 1
store i64 2, i64* %tmp4, align 4 store i64 2, i64* %tmp4, align 4
call void @f(%struct.ss* byval(%struct.ss) %S, i32* byval(i32) %X) call void @f(%struct.ss* byval(%struct.ss) %S, i32* byval(i32) %X)
ret i32 0 ret i32 0
} }
;. ;.
; IS__TUNIT____: attributes #[[ATTR0]] = { nofree nosync nounwind readnone willreturn } ; IS__TUNIT____: attributes #[[ATTR0]] = { nofree nosync nounwind readnone willreturn }
;. ;.
; IS__CGSCC_OPM: attributes #[[ATTR0]] = { nofree norecurse nosync nounwind readnone willreturn } ; IS__CGSCC____: attributes #[[ATTR0:[0-9]+]] = { nofree norecurse nosync nounwind readnone willreturn }
;.
; IS__CGSCC_NPM: attributes #[[ATTR0]] = { nofree norecurse nosync nounwind readnone willreturn }
; IS__CGSCC_NPM: attributes #[[ATTR1]] = { argmemonly nofree norecurse nosync nounwind willreturn }
;. ;.

llvm/test/Transforms/Attributor/ArgumentPromotion/fp80.ll

Show All 34 Lines
; IS__TUNIT_NPM-NEXT: entry: ; IS__TUNIT_NPM-NEXT: entry:
; IS__TUNIT_NPM-NEXT: [[A_CAST:%.*]] = bitcast %struct.Foo* @a to i32* ; IS__TUNIT_NPM-NEXT: [[A_CAST:%.*]] = bitcast %struct.Foo* @a to i32*
; IS__TUNIT_NPM-NEXT: [[TMP0:%.*]] = load i32, i32* [[A_CAST]], align 8 ; IS__TUNIT_NPM-NEXT: [[TMP0:%.*]] = load i32, i32* [[A_CAST]], align 8
; IS__TUNIT_NPM-NEXT: [[A_0_1:%.*]] = getelementptr [[STRUCT_FOO:%.*]], %struct.Foo* @a, i32 0, i32 1 ; IS__TUNIT_NPM-NEXT: [[A_0_1:%.*]] = getelementptr [[STRUCT_FOO:%.*]], %struct.Foo* @a, i32 0, i32 1
; IS__TUNIT_NPM-NEXT: [[TMP1:%.*]] = load i64, i64* [[A_0_1]], align 8 ; IS__TUNIT_NPM-NEXT: [[TMP1:%.*]] = load i64, i64* [[A_0_1]], align 8
; IS__TUNIT_NPM-NEXT: [[TMP2:%.*]] = call i64 @CaptureAStruct(i32 [[TMP0]], i64 [[TMP1]]) #[[ATTR0]] ; IS__TUNIT_NPM-NEXT: [[TMP2:%.*]] = call i64 @CaptureAStruct(i32 [[TMP0]], i64 [[TMP1]]) #[[ATTR0]]
; IS__TUNIT_NPM-NEXT: unreachable ; IS__TUNIT_NPM-NEXT: unreachable
; ;
; IS__CGSCC____: Function Attrs: nofree norecurse noreturn nosync nounwind readonly willreturn ; IS__CGSCC____: Function Attrs: nofree norecurse noreturn nosync nounwind readnone willreturn
; IS__CGSCC____-LABEL: define {{[^@]+}}@run ; IS__CGSCC____-LABEL: define {{[^@]+}}@run
; IS__CGSCC____-SAME: () #[[ATTR0:[0-9]+]] { ; IS__CGSCC____-SAME: () #[[ATTR0:[0-9]+]] {
; IS__CGSCC____-NEXT: entry: ; IS__CGSCC____-NEXT: entry:
; IS__CGSCC____-NEXT: [[TMP0:%.*]] = load i32, i32* getelementptr inbounds ([[STRUCT_FOO:%.*]], %struct.Foo* @a, i32 0, i32 0), align 8
; IS__CGSCC____-NEXT: [[A_0_1:%.*]] = getelementptr [[STRUCT_FOO]], %struct.Foo* @a, i32 0, i32 1
; IS__CGSCC____-NEXT: [[TMP1:%.*]] = load i64, i64* [[A_0_1]], align 8
; IS__CGSCC____-NEXT: unreachable ; IS__CGSCC____-NEXT: unreachable
; ;
entry: entry:
tail call i8 @UseLongDoubleUnsafely(%union.u* byval(%union.u) align 16 bitcast (%struct.s* @b to %union.u*)) tail call i8 @UseLongDoubleUnsafely(%union.u* byval(%union.u) align 16 bitcast (%struct.s* @b to %union.u*))
tail call x86_fp80 @UseLongDoubleSafely(%union.u* byval(%union.u) align 16 bitcast (%struct.s* @b to %union.u*)) tail call x86_fp80 @UseLongDoubleSafely(%union.u* byval(%union.u) align 16 bitcast (%struct.s* @b to %union.u*))
call i64 @AccessPaddingOfStruct(%struct.Foo* byval(%struct.Foo) @a) call i64 @AccessPaddingOfStruct(%struct.Foo* byval(%struct.Foo) @a)
call i64 @CaptureAStruct(%struct.Foo* byval(%struct.Foo) @a) call i64 @CaptureAStruct(%struct.Foo* byval(%struct.Foo) @a)
ret void ret void
▲ Show 20 Lines • Show All 45 Lines • ▼ Show 20 Lines
; IS__CGSCC_OPM-NEXT: [[PHI:%.*]] = phi %struct.Foo* [ null, [[ENTRY:%.*]] ], [ [[GEP:%.*]], [[LOOP]] ] ; IS__CGSCC_OPM-NEXT: [[PHI:%.*]] = phi %struct.Foo* [ null, [[ENTRY:%.*]] ], [ [[GEP:%.*]], [[LOOP]] ]
; IS__CGSCC_OPM-NEXT: [[TMP0:%.*]] = phi %struct.Foo* [ [[A]], [[ENTRY]] ], [ [[TMP0]], [[LOOP]] ] ; IS__CGSCC_OPM-NEXT: [[TMP0:%.*]] = phi %struct.Foo* [ [[A]], [[ENTRY]] ], [ [[TMP0]], [[LOOP]] ]
; IS__CGSCC_OPM-NEXT: store %struct.Foo* [[PHI]], %struct.Foo** [[A_PTR]], align 8 ; IS__CGSCC_OPM-NEXT: store %struct.Foo* [[PHI]], %struct.Foo** [[A_PTR]], align 8
; IS__CGSCC_OPM-NEXT: [[GEP]] = getelementptr [[STRUCT_FOO:%.*]], %struct.Foo* [[A]], i64 0 ; IS__CGSCC_OPM-NEXT: [[GEP]] = getelementptr [[STRUCT_FOO:%.*]], %struct.Foo* [[A]], i64 0
; IS__CGSCC_OPM-NEXT: br label [[LOOP]] ; IS__CGSCC_OPM-NEXT: br label [[LOOP]]
; ;
; IS________OPM: Function Attrs: nofree noreturn nosync nounwind readnone ; IS________OPM: Function Attrs: nofree noreturn nosync nounwind readnone
; IS________OPM-LABEL: define {{[^@]+}}@CaptureAStruct ; IS________OPM-LABEL: define {{[^@]+}}@CaptureAStruct
; IS________OPM-SAME: (%struct.Foo* noalias nofree noundef nonnull byval([[STRUCT_FOO:%.*]]) align 8 dereferenceable(16) [[A:%.*]]) #[[ATTR0]] { ; IS________OPM-SAME: (%struct.Foo* noalias nocapture nofree noundef nonnull readnone byval([[STRUCT_FOO:%.*]]) align 8 dereferenceable(16) [[A:%.*]]) #[[ATTR0]] {
; IS________OPM-NEXT: entry: ; IS________OPM-NEXT: entry:
; IS________OPM-NEXT: [[A_PTR:%.*]] = alloca %struct.Foo*, align 8 ; IS________OPM-NEXT: [[A_PTR:%.*]] = alloca %struct.Foo*, align 8
; IS________OPM-NEXT: br label [[LOOP:%.*]] ; IS________OPM-NEXT: br label [[LOOP:%.*]]
; IS________OPM: loop: ; IS________OPM: loop:
; IS________OPM-NEXT: [[PHI:%.*]] = phi %struct.Foo* [ null, [[ENTRY:%.*]] ], [ [[A]], [[LOOP]] ] ; IS________OPM-NEXT: [[PHI:%.*]] = phi %struct.Foo* [ null, [[ENTRY:%.*]] ], [ [[A]], [[LOOP]] ]
; IS________OPM-NEXT: [[TMP0:%.*]] = phi %struct.Foo* [ [[A]], [[ENTRY]] ], [ [[TMP0]], [[LOOP]] ] ; IS________OPM-NEXT: [[TMP0:%.*]] = phi %struct.Foo* [ [[A]], [[ENTRY]] ], [ [[TMP0]], [[LOOP]] ]
; IS________OPM-NEXT: store %struct.Foo* [[PHI]], %struct.Foo** [[A_PTR]], align 8 ; IS________OPM-NEXT: store %struct.Foo* [[PHI]], %struct.Foo** [[A_PTR]], align 8
; IS________OPM-NEXT: br label [[LOOP]] ; IS________OPM-NEXT: br label [[LOOP]]
Show All 16 Lines
; IS__TUNIT_NPM-NEXT: br label [[LOOP]] ; IS__TUNIT_NPM-NEXT: br label [[LOOP]]
; ;
; IS__CGSCC____: Function Attrs: nofree norecurse noreturn nosync nounwind readnone ; IS__CGSCC____: Function Attrs: nofree norecurse noreturn nosync nounwind readnone
; IS__CGSCC____-LABEL: define {{[^@]+}}@CaptureAStruct ; IS__CGSCC____-LABEL: define {{[^@]+}}@CaptureAStruct
; IS__CGSCC____-SAME: (i32 [[TMP0:%.*]], i64 [[TMP1:%.*]]) #[[ATTR2:[0-9]+]] { ; IS__CGSCC____-SAME: (i32 [[TMP0:%.*]], i64 [[TMP1:%.*]]) #[[ATTR2:[0-9]+]] {
; IS__CGSCC____-NEXT: entry: ; IS__CGSCC____-NEXT: entry:
; IS__CGSCC____-NEXT: [[A_PRIV:%.*]] = alloca [[STRUCT_FOO:%.*]], align 8 ; IS__CGSCC____-NEXT: [[A_PRIV:%.*]] = alloca [[STRUCT_FOO:%.*]], align 8
; IS__CGSCC____-NEXT: [[A_PRIV_CAST:%.*]] = bitcast %struct.Foo* [[A_PRIV]] to i32* ; IS__CGSCC____-NEXT: [[A_PRIV_CAST:%.*]] = bitcast %struct.Foo* [[A_PRIV]] to i32*
; IS__CGSCC____-NEXT: store i32 [[TMP0]], i32* [[A_PRIV_CAST]], align 8 ; IS__CGSCC____-NEXT: store i32 undef, i32* [[A_PRIV_CAST]], align 8
; IS__CGSCC____-NEXT: [[A_PRIV_0_1:%.*]] = getelementptr [[STRUCT_FOO]], %struct.Foo* [[A_PRIV]], i32 0, i32 1 ; IS__CGSCC____-NEXT: [[A_PRIV_0_1:%.*]] = getelementptr [[STRUCT_FOO]], %struct.Foo* [[A_PRIV]], i32 0, i32 1
; IS__CGSCC____-NEXT: store i64 [[TMP1]], i64* [[A_PRIV_0_1]], align 8 ; IS__CGSCC____-NEXT: store i64 undef, i64* [[A_PRIV_0_1]], align 8
; IS__CGSCC____-NEXT: [[A_PTR:%.*]] = alloca %struct.Foo*, align 8 ; IS__CGSCC____-NEXT: [[A_PTR:%.*]] = alloca %struct.Foo*, align 8
; IS__CGSCC____-NEXT: br label [[LOOP:%.*]] ; IS__CGSCC____-NEXT: br label [[LOOP:%.*]]
; IS__CGSCC____: loop: ; IS__CGSCC____: loop:
; IS__CGSCC____-NEXT: [[PHI:%.*]] = phi %struct.Foo* [ null, [[ENTRY:%.*]] ], [ [[A_PRIV]], [[LOOP]] ] ; IS__CGSCC____-NEXT: [[PHI:%.*]] = phi %struct.Foo* [ null, [[ENTRY:%.*]] ], [ [[A_PRIV]], [[LOOP]] ]
; IS__CGSCC____-NEXT: [[TMP2:%.*]] = phi %struct.Foo* [ [[A_PRIV]], [[ENTRY]] ], [ [[TMP2]], [[LOOP]] ] ; IS__CGSCC____-NEXT: [[TMP2:%.*]] = phi %struct.Foo* [ [[A_PRIV]], [[ENTRY]] ], [ [[TMP2]], [[LOOP]] ]
; IS__CGSCC____-NEXT: store %struct.Foo* [[PHI]], %struct.Foo** [[A_PTR]], align 8 ; IS__CGSCC____-NEXT: store %struct.Foo* [[PHI]], %struct.Foo** [[A_PTR]], align 8
; IS__CGSCC____-NEXT: br label [[LOOP]] ; IS__CGSCC____-NEXT: br label [[LOOP]]
; ;
entry: entry:
%a_ptr = alloca %struct.Foo* %a_ptr = alloca %struct.Foo*
br label %loop br label %loop
loop: loop:
%phi = phi %struct.Foo* [ null, %entry ], [ %gep, %loop ] %phi = phi %struct.Foo* [ null, %entry ], [ %gep, %loop ]
%0 = phi %struct.Foo* [ %a, %entry ], [ %0, %loop ] %0 = phi %struct.Foo* [ %a, %entry ], [ %0, %loop ]
store %struct.Foo* %phi, %struct.Foo** %a_ptr store %struct.Foo* %phi, %struct.Foo** %a_ptr
%gep = getelementptr %struct.Foo, %struct.Foo* %a, i64 0 %gep = getelementptr %struct.Foo, %struct.Foo* %a, i64 0
br label %loop br label %loop
} }
;. ;.
; NOT_CGSCC_NPM: attributes #[[ATTR0:[0-9]+]] = { nofree noreturn nosync nounwind readnone } ; NOT_CGSCC_NPM: attributes #[[ATTR0:[0-9]+]] = { nofree noreturn nosync nounwind readnone }
;. ;.
; IS__CGSCC____: attributes #[[ATTR0]] = { nofree norecurse noreturn nosync nounwind readonly willreturn } ; IS__CGSCC____: attributes #[[ATTR0]] = { nofree norecurse noreturn nosync nounwind readnone willreturn }
; IS__CGSCC____: attributes #[[ATTR1]] = { nofree norecurse nosync nounwind readnone willreturn } ; IS__CGSCC____: attributes #[[ATTR1]] = { nofree norecurse nosync nounwind readnone willreturn }
; IS__CGSCC____: attributes #[[ATTR2]] = { nofree norecurse noreturn nosync nounwind readnone } ; IS__CGSCC____: attributes #[[ATTR2]] = { nofree norecurse noreturn nosync nounwind readnone }
;. ;.

llvm/test/Transforms/Attributor/IPConstantProp/2009-09-24-byval-ptr.ll

Show First 20 Lines • Show All 49 Lines • ▼ Show 20 Lines
; IS__CGSCC_NPM: Function Attrs: nounwind ; IS__CGSCC_NPM: Function Attrs: nounwind
; IS__CGSCC_NPM-LABEL: define {{[^@]+}}@vfu1 ; IS__CGSCC_NPM-LABEL: define {{[^@]+}}@vfu1
; IS__CGSCC_NPM-SAME: (i8 [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR0:[0-9]+]] { ; IS__CGSCC_NPM-SAME: (i8 [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR0:[0-9]+]] {
; IS__CGSCC_NPM-NEXT: entry: ; IS__CGSCC_NPM-NEXT: entry:
; IS__CGSCC_NPM-NEXT: [[U_PRIV:%.*]] = alloca [[STRUCT_MYSTR:%.*]], align 8 ; IS__CGSCC_NPM-NEXT: [[U_PRIV:%.*]] = alloca [[STRUCT_MYSTR:%.*]], align 8
; IS__CGSCC_NPM-NEXT: [[U_PRIV_CAST:%.*]] = bitcast %struct.MYstr* [[U_PRIV]] to i8* ; IS__CGSCC_NPM-NEXT: [[U_PRIV_CAST:%.*]] = bitcast %struct.MYstr* [[U_PRIV]] to i8*
; IS__CGSCC_NPM-NEXT: store i8 [[TMP0]], i8* [[U_PRIV_CAST]], align 8 ; IS__CGSCC_NPM-NEXT: store i8 [[TMP0]], i8* [[U_PRIV_CAST]], align 8
; IS__CGSCC_NPM-NEXT: [[U_PRIV_0_1:%.*]] = getelementptr [[STRUCT_MYSTR]], %struct.MYstr* [[U_PRIV]], i32 0, i32 1 ; IS__CGSCC_NPM-NEXT: [[U_PRIV_0_1:%.*]] = getelementptr [[STRUCT_MYSTR]], %struct.MYstr* [[U_PRIV]], i32 0, i32 1
; IS__CGSCC_NPM-NEXT: store i32 [[TMP1]], i32* [[U_PRIV_0_1]], align 4 ; IS__CGSCC_NPM-NEXT: store i32 undef, i32* [[U_PRIV_0_1]], align 4
; IS__CGSCC_NPM-NEXT: [[TMP2:%.*]] = getelementptr [[STRUCT_MYSTR]], %struct.MYstr* [[U_PRIV]], i32 0, i32 1 ; IS__CGSCC_NPM-NEXT: [[TMP2:%.*]] = getelementptr [[STRUCT_MYSTR]], %struct.MYstr* [[U_PRIV]], i32 0, i32 1
; IS__CGSCC_NPM-NEXT: store i32 99, i32* [[TMP2]], align 4 ; IS__CGSCC_NPM-NEXT: store i32 99, i32* [[TMP2]], align 4
; IS__CGSCC_NPM-NEXT: [[TMP3:%.*]] = getelementptr [[STRUCT_MYSTR]], %struct.MYstr* [[U_PRIV]], i32 0, i32 0 ; IS__CGSCC_NPM-NEXT: [[TMP3:%.*]] = getelementptr [[STRUCT_MYSTR]], %struct.MYstr* [[U_PRIV]], i32 0, i32 0
; IS__CGSCC_NPM-NEXT: store i8 97, i8* [[TMP3]], align 8 ; IS__CGSCC_NPM-NEXT: store i8 97, i8* [[TMP3]], align 8
; IS__CGSCC_NPM-NEXT: [[L:%.*]] = load i8, i8* [[TMP3]], align 8 ; IS__CGSCC_NPM-NEXT: [[L:%.*]] = load i8, i8* [[TMP3]], align 8
; IS__CGSCC_NPM-NEXT: call void @use(i8 [[L]]) ; IS__CGSCC_NPM-NEXT: call void @use(i8 [[L]])
; IS__CGSCC_NPM-NEXT: br label [[RETURN:%.*]] ; IS__CGSCC_NPM-NEXT: br label [[RETURN:%.*]]
; IS__CGSCC_NPM: return: ; IS__CGSCC_NPM: return:
▲ Show 20 Lines • Show All 114 Lines • ▼ Show 20 Lines
; IS__CGSCC_OPM-NEXT: [[RESULT:%.*]] = call i32 @vfu2(%struct.MYstr* noalias nocapture nofree noundef nonnull readnone byval([[STRUCT_MYSTR]]) align 8 dereferenceable(8) @mystr) #[[ATTR0]] ; IS__CGSCC_OPM-NEXT: [[RESULT:%.*]] = call i32 @vfu2(%struct.MYstr* noalias nocapture nofree noundef nonnull readnone byval([[STRUCT_MYSTR]]) align 8 dereferenceable(8) @mystr) #[[ATTR0]]
; IS__CGSCC_OPM-NEXT: ret i32 [[RESULT]] ; IS__CGSCC_OPM-NEXT: ret i32 [[RESULT]]
; ;
; IS__CGSCC_NPM: Function Attrs: nounwind ; IS__CGSCC_NPM: Function Attrs: nounwind
; IS__CGSCC_NPM-LABEL: define {{[^@]+}}@unions ; IS__CGSCC_NPM-LABEL: define {{[^@]+}}@unions
; IS__CGSCC_NPM-SAME: () #[[ATTR0]] { ; IS__CGSCC_NPM-SAME: () #[[ATTR0]] {
; IS__CGSCC_NPM-NEXT: entry: ; IS__CGSCC_NPM-NEXT: entry:
; IS__CGSCC_NPM-NEXT: [[TMP0:%.*]] = load i8, i8* getelementptr inbounds ([[STRUCT_MYSTR:%.*]], %struct.MYstr* @mystr, i32 0, i32 0), align 8 ; IS__CGSCC_NPM-NEXT: [[TMP0:%.*]] = load i8, i8* getelementptr inbounds ([[STRUCT_MYSTR:%.*]], %struct.MYstr* @mystr, i32 0, i32 0), align 8
; IS__CGSCC_NPM-NEXT: [[MYSTR_0_1:%.*]] = getelementptr [[STRUCT_MYSTR]], %struct.MYstr* @mystr, i32 0, i32 1 ; IS__CGSCC_NPM-NEXT: call void @vfu1(i8 [[TMP0]], i32 undef) #[[ATTR0]]
; IS__CGSCC_NPM-NEXT: [[TMP1:%.*]] = load i32, i32* [[MYSTR_0_1]], align 8 ; IS__CGSCC_NPM-NEXT: [[TMP1:%.*]] = load i8, i8* getelementptr inbounds ([[STRUCT_MYSTR]], %struct.MYstr* @mystr, i32 0, i32 0), align 8
; IS__CGSCC_NPM-NEXT: call void @vfu1(i8 [[TMP0]], i32 [[TMP1]]) #[[ATTR0]]
; IS__CGSCC_NPM-NEXT: [[TMP2:%.*]] = load i8, i8* getelementptr inbounds ([[STRUCT_MYSTR]], %struct.MYstr* @mystr, i32 0, i32 0), align 8
; IS__CGSCC_NPM-NEXT: [[MYSTR_0_12:%.*]] = getelementptr [[STRUCT_MYSTR]], %struct.MYstr* @mystr, i32 0, i32 1 ; IS__CGSCC_NPM-NEXT: [[MYSTR_0_12:%.*]] = getelementptr [[STRUCT_MYSTR]], %struct.MYstr* @mystr, i32 0, i32 1
; IS__CGSCC_NPM-NEXT: [[TMP3:%.*]] = load i32, i32* [[MYSTR_0_12]], align 8 ; IS__CGSCC_NPM-NEXT: [[TMP2:%.*]] = load i32, i32* [[MYSTR_0_12]], align 8
; IS__CGSCC_NPM-NEXT: [[RESULT:%.*]] = call i32 @vfu2(i8 [[TMP2]], i32 [[TMP3]]) #[[ATTR0]] ; IS__CGSCC_NPM-NEXT: [[RESULT:%.*]] = call i32 @vfu2(i8 [[TMP1]], i32 [[TMP2]]) #[[ATTR0]]
; IS__CGSCC_NPM-NEXT: ret i32 [[RESULT]] ; IS__CGSCC_NPM-NEXT: ret i32 [[RESULT]]
; ;
entry: entry:
call void @vfu1(%struct.MYstr* byval(%struct.MYstr) align 4 @mystr) nounwind call void @vfu1(%struct.MYstr* byval(%struct.MYstr) align 4 @mystr) nounwind
%result = call i32 @vfu2(%struct.MYstr* byval(%struct.MYstr) align 4 @mystr) nounwind %result = call i32 @vfu2(%struct.MYstr* byval(%struct.MYstr) align 4 @mystr) nounwind
ret i32 %result ret i32 %result
} }
▲ Show 20 Lines • Show All 84 Lines • ▼ Show 20 Lines
; IS__TUNIT_OPM-NEXT: call void @vfu1(%struct.MYstr* nocapture nofree noundef nonnull readonly byval([[STRUCT_MYSTR:%.*]]) align 8 dereferenceable(8) @mystr) #[[ATTR0]] ; IS__TUNIT_OPM-NEXT: call void @vfu1(%struct.MYstr* nocapture nofree noundef nonnull readonly byval([[STRUCT_MYSTR:%.*]]) align 8 dereferenceable(8) @mystr) #[[ATTR0]]
; IS__TUNIT_OPM-NEXT: [[RESULT:%.*]] = call i32 @vfu2_v2(%struct.MYstr* nocapture nofree noundef nonnull readonly byval([[STRUCT_MYSTR]]) align 8 dereferenceable(8) @mystr) #[[ATTR2]] ; IS__TUNIT_OPM-NEXT: [[RESULT:%.*]] = call i32 @vfu2_v2(%struct.MYstr* nocapture nofree noundef nonnull readonly byval([[STRUCT_MYSTR]]) align 8 dereferenceable(8) @mystr) #[[ATTR2]]
; IS__TUNIT_OPM-NEXT: ret i32 [[RESULT]] ; IS__TUNIT_OPM-NEXT: ret i32 [[RESULT]]
; ;
; IS__TUNIT_NPM: Function Attrs: nounwind ; IS__TUNIT_NPM: Function Attrs: nounwind
; IS__TUNIT_NPM-LABEL: define {{[^@]+}}@unions_v2 ; IS__TUNIT_NPM-LABEL: define {{[^@]+}}@unions_v2
; IS__TUNIT_NPM-SAME: () #[[ATTR0]] { ; IS__TUNIT_NPM-SAME: () #[[ATTR0]] {
; IS__TUNIT_NPM-NEXT: entry: ; IS__TUNIT_NPM-NEXT: entry:
; IS__TUNIT_NPM-NEXT: [[MYSTR_CAST:%.*]] = bitcast %struct.MYstr* @mystr to i8*
; IS__TUNIT_NPM-NEXT: [[TMP0:%.*]] = load i8, i8* [[MYSTR_CAST]], align 8
; IS__TUNIT_NPM-NEXT: [[MYSTR_0_1:%.*]] = getelementptr [[STRUCT_MYSTR:%.*]], %struct.MYstr* @mystr, i32 0, i32 1
; IS__TUNIT_NPM-NEXT: [[TMP1:%.*]] = load i32, i32* [[MYSTR_0_1]], align 8
; IS__TUNIT_NPM-NEXT: call void @vfu1(i8 [[TMP0]], i32 [[TMP1]]) #[[ATTR0]]
; IS__TUNIT_NPM-NEXT: [[MYSTR_CAST1:%.*]] = bitcast %struct.MYstr* @mystr to i8* ; IS__TUNIT_NPM-NEXT: [[MYSTR_CAST1:%.*]] = bitcast %struct.MYstr* @mystr to i8*
; IS__TUNIT_NPM-NEXT: [[TMP2:%.*]] = load i8, i8* [[MYSTR_CAST1]], align 8 ; IS__TUNIT_NPM-NEXT: [[TMP0:%.*]] = load i8, i8* [[MYSTR_CAST1]], align 8
; IS__TUNIT_NPM-NEXT: [[MYSTR_0_12:%.*]] = getelementptr [[STRUCT_MYSTR]], %struct.MYstr* @mystr, i32 0, i32 1 ; IS__TUNIT_NPM-NEXT: [[MYSTR_0_12:%.*]] = getelementptr [[STRUCT_MYSTR:%.*]], %struct.MYstr* @mystr, i32 0, i32 1
; IS__TUNIT_NPM-NEXT: [[TMP3:%.*]] = load i32, i32* [[MYSTR_0_12]], align 8 ; IS__TUNIT_NPM-NEXT: [[TMP1:%.*]] = load i32, i32* [[MYSTR_0_12]], align 8
; IS__TUNIT_NPM-NEXT: call void @vfu1(i8 [[TMP0]], i32 [[TMP1]]) #[[ATTR0]]
; IS__TUNIT_NPM-NEXT: [[MYSTR_CAST:%.*]] = bitcast %struct.MYstr* @mystr to i8*
; IS__TUNIT_NPM-NEXT: [[TMP2:%.*]] = load i8, i8* [[MYSTR_CAST]], align 8
; IS__TUNIT_NPM-NEXT: [[MYSTR_0_1:%.*]] = getelementptr [[STRUCT_MYSTR]], %struct.MYstr* @mystr, i32 0, i32 1
; IS__TUNIT_NPM-NEXT: [[TMP3:%.*]] = load i32, i32* [[MYSTR_0_1]], align 8
; IS__TUNIT_NPM-NEXT: [[RESULT:%.*]] = call i32 @vfu2_v2(i8 [[TMP2]], i32 [[TMP3]]) #[[ATTR2]] ; IS__TUNIT_NPM-NEXT: [[RESULT:%.*]] = call i32 @vfu2_v2(i8 [[TMP2]], i32 [[TMP3]]) #[[ATTR2]]
; IS__TUNIT_NPM-NEXT: ret i32 [[RESULT]] ; IS__TUNIT_NPM-NEXT: ret i32 [[RESULT]]
; ;
; IS__CGSCC_OPM: Function Attrs: nounwind ; IS__CGSCC_OPM: Function Attrs: nounwind
; IS__CGSCC_OPM-LABEL: define {{[^@]+}}@unions_v2 ; IS__CGSCC_OPM-LABEL: define {{[^@]+}}@unions_v2
; IS__CGSCC_OPM-SAME: () #[[ATTR0]] { ; IS__CGSCC_OPM-SAME: () #[[ATTR0]] {
; IS__CGSCC_OPM-NEXT: entry: ; IS__CGSCC_OPM-NEXT: entry:
; IS__CGSCC_OPM-NEXT: call void @vfu1(%struct.MYstr* noalias nocapture nofree noundef nonnull readonly byval([[STRUCT_MYSTR:%.*]]) align 8 dereferenceable(8) @mystr) #[[ATTR0]] ; IS__CGSCC_OPM-NEXT: call void @vfu1(%struct.MYstr* noalias nocapture nofree noundef nonnull readonly byval([[STRUCT_MYSTR:%.*]]) align 8 dereferenceable(8) @mystr) #[[ATTR0]]
; IS__CGSCC_OPM-NEXT: [[RESULT:%.*]] = call i32 @vfu2_v2(%struct.MYstr* noalias nocapture nofree noundef nonnull readnone byval([[STRUCT_MYSTR]]) align 8 dereferenceable(8) @mystr) #[[ATTR0]] ; IS__CGSCC_OPM-NEXT: [[RESULT:%.*]] = call i32 @vfu2_v2(%struct.MYstr* noalias nocapture nofree noundef nonnull readnone byval([[STRUCT_MYSTR]]) align 8 dereferenceable(8) @mystr) #[[ATTR0]]
; IS__CGSCC_OPM-NEXT: ret i32 [[RESULT]] ; IS__CGSCC_OPM-NEXT: ret i32 [[RESULT]]
; ;
; IS__CGSCC_NPM: Function Attrs: nounwind ; IS__CGSCC_NPM: Function Attrs: nounwind
; IS__CGSCC_NPM-LABEL: define {{[^@]+}}@unions_v2 ; IS__CGSCC_NPM-LABEL: define {{[^@]+}}@unions_v2
; IS__CGSCC_NPM-SAME: () #[[ATTR0]] { ; IS__CGSCC_NPM-SAME: () #[[ATTR0]] {
; IS__CGSCC_NPM-NEXT: entry: ; IS__CGSCC_NPM-NEXT: entry:
; IS__CGSCC_NPM-NEXT: call void @vfu1(i8 noundef 0, i32 noundef 0) #[[ATTR0]] ; IS__CGSCC_NPM-NEXT: call void @vfu1(i8 noundef 0, i32 undef) #[[ATTR0]]
; IS__CGSCC_NPM-NEXT: [[RESULT:%.*]] = call i32 @vfu2_v2(i8 undef, i32 undef) #[[ATTR0]] ; IS__CGSCC_NPM-NEXT: [[RESULT:%.*]] = call i32 @vfu2_v2(i8 undef, i32 undef) #[[ATTR0]]
; IS__CGSCC_NPM-NEXT: ret i32 [[RESULT]] ; IS__CGSCC_NPM-NEXT: ret i32 [[RESULT]]
; ;
entry: entry:
call void @vfu1(%struct.MYstr* byval(%struct.MYstr) align 4 @mystr) nounwind call void @vfu1(%struct.MYstr* byval(%struct.MYstr) align 4 @mystr) nounwind
%result = call i32 @vfu2_v2(%struct.MYstr* byval(%struct.MYstr) align 4 @mystr) nounwind %result = call i32 @vfu2_v2(%struct.MYstr* byval(%struct.MYstr) align 4 @mystr) nounwind
ret i32 %result ret i32 %result
} }
;. ;.
; IS__TUNIT____: attributes #[[ATTR0:[0-9]+]] = { nounwind } ; IS__TUNIT____: attributes #[[ATTR0:[0-9]+]] = { nounwind }
; IS__TUNIT____: attributes #[[ATTR1:[0-9]+]] = { nofree nosync nounwind readnone willreturn } ; IS__TUNIT____: attributes #[[ATTR1:[0-9]+]] = { nofree nosync nounwind readnone willreturn }
; IS__TUNIT____: attributes #[[ATTR2:[0-9]+]] = { nounwind readonly } ; IS__TUNIT____: attributes #[[ATTR2:[0-9]+]] = { nounwind readonly }
;. ;.
; IS__CGSCC____: attributes #[[ATTR0:[0-9]+]] = { nounwind } ; IS__CGSCC____: attributes #[[ATTR0:[0-9]+]] = { nounwind }
; IS__CGSCC____: attributes #[[ATTR1:[0-9]+]] = { nofree norecurse nosync nounwind readnone willreturn } ; IS__CGSCC____: attributes #[[ATTR1:[0-9]+]] = { nofree norecurse nosync nounwind readnone willreturn }
;. ;.

llvm/test/Transforms/Attributor/nodelete.ll

; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --function-signature --check-attributes --check-globals ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --function-signature --check-attributes --check-globals
; RUN: opt -attributor -enable-new-pm=0 -attributor-manifest-internal -attributor-max-iterations-verify -attributor-annotate-decl-cs -attributor-max-iterations=8 -S < %s | FileCheck %s --check-prefixes=CHECK,NOT_CGSCC_NPM,NOT_CGSCC_OPM,NOT_TUNIT_NPM,IS__TUNIT____,IS________OPM,IS__TUNIT_OPM ; RUN: opt -attributor -enable-new-pm=0 -attributor-manifest-internal -attributor-max-iterations-verify -attributor-annotate-decl-cs -attributor-max-iterations=6 -S < %s | FileCheck %s --check-prefixes=CHECK,NOT_CGSCC_NPM,NOT_CGSCC_OPM,NOT_TUNIT_NPM,IS__TUNIT____,IS________OPM,IS__TUNIT_OPM
; RUN: opt -aa-pipeline=basic-aa -passes=attributor -attributor-manifest-internal -attributor-max-iterations-verify -attributor-annotate-decl-cs -attributor-max-iterations=8 -S < %s | FileCheck %s --check-prefixes=CHECK,NOT_CGSCC_OPM,NOT_CGSCC_NPM,NOT_TUNIT_OPM,IS__TUNIT____,IS________NPM,IS__TUNIT_NPM ; RUN: opt -aa-pipeline=basic-aa -passes=attributor -attributor-manifest-internal -attributor-max-iterations-verify -attributor-annotate-decl-cs -attributor-max-iterations=6 -S < %s | FileCheck %s --check-prefixes=CHECK,NOT_CGSCC_OPM,NOT_CGSCC_NPM,NOT_TUNIT_OPM,IS__TUNIT____,IS________NPM,IS__TUNIT_NPM
; RUN: opt -attributor-cgscc -enable-new-pm=0 -attributor-manifest-internal -attributor-annotate-decl-cs -S < %s | FileCheck %s --check-prefixes=CHECK,NOT_TUNIT_NPM,NOT_TUNIT_OPM,NOT_CGSCC_NPM,IS__CGSCC____,IS________OPM,IS__CGSCC_OPM ; RUN: opt -attributor-cgscc -enable-new-pm=0 -attributor-manifest-internal -attributor-annotate-decl-cs -S < %s | FileCheck %s --check-prefixes=CHECK,NOT_TUNIT_NPM,NOT_TUNIT_OPM,NOT_CGSCC_NPM,IS__CGSCC____,IS________OPM,IS__CGSCC_OPM
; RUN: opt -aa-pipeline=basic-aa -passes=attributor-cgscc -attributor-manifest-internal -attributor-annotate-decl-cs -S < %s | FileCheck %s --check-prefixes=CHECK,NOT_TUNIT_NPM,NOT_TUNIT_OPM,NOT_CGSCC_OPM,IS__CGSCC____,IS________NPM,IS__CGSCC_NPM ; RUN: opt -aa-pipeline=basic-aa -passes=attributor-cgscc -attributor-manifest-internal -attributor-annotate-decl-cs -S < %s | FileCheck %s --check-prefixes=CHECK,NOT_TUNIT_NPM,NOT_TUNIT_OPM,NOT_CGSCC_OPM,IS__CGSCC____,IS________NPM,IS__CGSCC_NPM
%"a" = type { i64 } %"a" = type { i64 }
%"b" = type { i8 } %"b" = type { i8 }
define hidden i64 @f1() align 2 { define hidden i64 @f1() align 2 {
; IS__TUNIT____: Function Attrs: nofree nosync nounwind readnone willreturn ; IS__TUNIT____: Function Attrs: nofree nosync nounwind readnone willreturn
; IS__TUNIT____-LABEL: define {{[^@]+}}@f1 ; IS__TUNIT____-LABEL: define {{[^@]+}}@f1
; IS__TUNIT____-SAME: () #[[ATTR0:[0-9]+]] align 2 { ; IS__TUNIT____-SAME: () #[[ATTR0:[0-9]+]] align 2 {
; IS__TUNIT____-NEXT: entry: ; IS__TUNIT____-NEXT: entry:
; IS__TUNIT____-NEXT: ret i64 undef ; IS__TUNIT____-NEXT: ret i64 undef
; ;
; IS__CGSCC____: Function Attrs: nofree norecurse nosync nounwind readnone willreturn ; IS__CGSCC____: Function Attrs: nofree norecurse nosync nounwind readnone willreturn
; IS__CGSCC____-LABEL: define {{[^@]+}}@f1 ; IS__CGSCC____-LABEL: define {{[^@]+}}@f1
; IS__CGSCC____-SAME: () #[[ATTR0:[0-9]+]] align 2 { ; IS__CGSCC____-SAME: () #[[ATTR0:[0-9]+]] align 2 {
; IS__CGSCC____-NEXT: entry: ; IS__CGSCC____-NEXT: entry:
; IS__CGSCC____-NEXT: [[REF_TMP:%.*]] = alloca [[A:%.*]], align 8
; IS__CGSCC____-NEXT: ret i64 undef ; IS__CGSCC____-NEXT: ret i64 undef
; ;
entry: entry:
%ref.tmp = alloca %"a", align 8 %ref.tmp = alloca %"a", align 8
%call2 = call i64 @f2(%"a"* %ref.tmp) %call2 = call i64 @f2(%"a"* %ref.tmp)
ret i64 %call2 ret i64 %call2
} }
define internal i64 @f2(%"a"* %this) align 2 { define internal i64 @f2(%"a"* %this) align 2 {
; IS__CGSCC____: Function Attrs: nofree norecurse nosync nounwind readnone willreturn ; IS__CGSCC____: Function Attrs: nofree norecurse nosync nounwind readnone willreturn
; IS__CGSCC____-LABEL: define {{[^@]+}}@f2 ; IS__CGSCC____-LABEL: define {{[^@]+}}@f2
; IS__CGSCC____-SAME: (%a* noalias nocapture nofree noundef nonnull readnone align 8 dereferenceable(8) [[THIS:%.*]]) #[[ATTR0]] align 2 { ; IS__CGSCC____-SAME: () #[[ATTR0]] align 2 {
; IS__CGSCC____-NEXT: entry: ; IS__CGSCC____-NEXT: entry:
; IS__CGSCC____-NEXT: [[THIS_ADDR:%.*]] = alloca %a*, align 8 ; IS__CGSCC____-NEXT: [[THIS_ADDR:%.*]] = alloca %a*, align 8
; IS__CGSCC____-NEXT: store %a* [[THIS]], %a** [[THIS_ADDR]], align 8 ; IS__CGSCC____-NEXT: store %a* undef, %a** [[THIS_ADDR]], align 8
; IS__CGSCC____-NEXT: ret i64 undef ; IS__CGSCC____-NEXT: ret i64 undef
; ;
entry: entry:
%this.addr = alloca %"a"*, align 8 %this.addr = alloca %"a"*, align 8
store %"a"* %this, %"a"** %this.addr, align 8 store %"a"* %this, %"a"** %this.addr, align 8
%this1 = load %"a"*, %"a"** %this.addr, align 8 %this1 = load %"a"*, %"a"** %this.addr, align 8
%0 = bitcast %"a"* %this1 to %"b"* %0 = bitcast %"a"* %this1 to %"b"*
call void @f3(%"b"* %0) call void @f3(%"b"* %0)
ret i64 undef ret i64 undef
} }
define internal void @f3(%"b"* %this) align 2 { define internal void @f3(%"b"* %this) align 2 {
; IS__CGSCC____: Function Attrs: nofree norecurse nosync nounwind readnone willreturn ; IS__CGSCC____: Function Attrs: nofree norecurse nosync nounwind readnone willreturn
; IS__CGSCC____-LABEL: define {{[^@]+}}@f3 ; IS__CGSCC____-LABEL: define {{[^@]+}}@f3
; IS__CGSCC____-SAME: (%b* noalias nocapture nofree noundef nonnull readnone align 8 dereferenceable(8) [[THIS:%.*]]) #[[ATTR0]] align 2 { ; IS__CGSCC____-SAME: () #[[ATTR0]] align 2 {
; IS__CGSCC____-NEXT: entry: ; IS__CGSCC____-NEXT: entry:
; IS__CGSCC____-NEXT: [[THIS_ADDR:%.*]] = alloca %b*, align 8 ; IS__CGSCC____-NEXT: [[THIS_ADDR:%.*]] = alloca %b*, align 8
; IS__CGSCC____-NEXT: store %b* [[THIS]], %b** [[THIS_ADDR]], align 8 ; IS__CGSCC____-NEXT: store %b* undef, %b** [[THIS_ADDR]], align 8
; IS__CGSCC____-NEXT: ret void ; IS__CGSCC____-NEXT: ret void
; ;
entry: entry:
%this.addr = alloca %"b"*, align 8 %this.addr = alloca %"b"*, align 8
store %"b"* %this, %"b"** %this.addr, align 8 store %"b"* %this, %"b"** %this.addr, align 8
%this1 = load %"b"*, %"b"** %this.addr, align 8 %this1 = load %"b"*, %"b"** %this.addr, align 8
%call = call i1 @f4(%"b"* %this1) %call = call i1 @f4(%"b"* %this1)
ret void ret void
} }
define internal i1 @f4(%"b"* %this) align 2 { define internal i1 @f4(%"b"* %this) align 2 {
; IS__CGSCC____: Function Attrs: nofree norecurse nosync nounwind readnone willreturn ; IS__CGSCC____: Function Attrs: nofree norecurse nosync nounwind readnone willreturn
; IS__CGSCC____-LABEL: define {{[^@]+}}@f4 ; IS__CGSCC____-LABEL: define {{[^@]+}}@f4
; IS__CGSCC____-SAME: (%b* noalias nocapture nofree noundef nonnull readnone align 8 dereferenceable(8) [[THIS:%.*]]) #[[ATTR0]] align 2 { ; IS__CGSCC____-SAME: () #[[ATTR0]] align 2 {
; IS__CGSCC____-NEXT: entry: ; IS__CGSCC____-NEXT: entry:
; IS__CGSCC____-NEXT: [[THIS_ADDR:%.*]] = alloca %b*, align 8 ; IS__CGSCC____-NEXT: [[THIS_ADDR:%.*]] = alloca %b*, align 8
; IS__CGSCC____-NEXT: store %b* [[THIS]], %b** [[THIS_ADDR]], align 8 ; IS__CGSCC____-NEXT: store %b* undef, %b** [[THIS_ADDR]], align 8
; IS__CGSCC____-NEXT: ret i1 undef ; IS__CGSCC____-NEXT: ret i1 undef
; ;
entry: entry:
%this.addr = alloca %"b"*, align 8 %this.addr = alloca %"b"*, align 8
store %"b"* %this, %"b"** %this.addr, align 8 store %"b"* %this, %"b"** %this.addr, align 8
%this1 = load %"b"*, %"b"** %this.addr, align 8 %this1 = load %"b"*, %"b"** %this.addr, align 8
%call = call %"a"* @f5(%"b"* %this1) %call = call %"a"* @f5(%"b"* %this1)
ret i1 undef ret i1 undef
} }
define internal %"a"* @f5(%"b"* %this) align 2 { define internal %"a"* @f5(%"b"* %this) align 2 {
; IS__CGSCC____: Function Attrs: nofree norecurse nosync nounwind readnone willreturn ; IS__CGSCC____: Function Attrs: nofree norecurse nosync nounwind readnone willreturn
; IS__CGSCC____-LABEL: define {{[^@]+}}@f5 ; IS__CGSCC____-LABEL: define {{[^@]+}}@f5
; IS__CGSCC____-SAME: (%b* noalias nocapture nofree noundef nonnull readnone align 8 dereferenceable(8) [[THIS:%.*]]) #[[ATTR0]] align 2 { ; IS__CGSCC____-SAME: () #[[ATTR0]] align 2 {
; IS__CGSCC____-NEXT: entry: ; IS__CGSCC____-NEXT: entry:
; IS__CGSCC____-NEXT: [[THIS_ADDR:%.*]] = alloca %b*, align 8 ; IS__CGSCC____-NEXT: [[THIS_ADDR:%.*]] = alloca %b*, align 8
; IS__CGSCC____-NEXT: store %b* [[THIS]], %b** [[THIS_ADDR]], align 8 ; IS__CGSCC____-NEXT: store %b* undef, %b** [[THIS_ADDR]], align 8
; IS__CGSCC____-NEXT: ret %a* undef ; IS__CGSCC____-NEXT: ret %a* undef
; ;
entry: entry:
%this.addr = alloca %"b"*, align 8 %this.addr = alloca %"b"*, align 8
store %"b"* %this, %"b"** %this.addr, align 8 store %"b"* %this, %"b"** %this.addr, align 8
%this1 = load %"b"*, %"b"** %this.addr, align 8 %this1 = load %"b"*, %"b"** %this.addr, align 8
%0 = bitcast %"b"* %this1 to %"a"* %0 = bitcast %"b"* %this1 to %"a"*
ret %"a"* %0 ret %"a"* %0
} }
;. ;.
; IS__TUNIT____: attributes #[[ATTR0]] = { nofree nosync nounwind readnone willreturn } ; IS__TUNIT____: attributes #[[ATTR0]] = { nofree nosync nounwind readnone willreturn }
;. ;.
; IS__CGSCC____: attributes #[[ATTR0]] = { nofree norecurse nosync nounwind readnone willreturn } ; IS__CGSCC____: attributes #[[ATTR0]] = { nofree norecurse nosync nounwind readnone willreturn }
;. ;.

llvm/test/Transforms/Attributor/value-simplify.ll

Show First 20 Lines • Show All 468 Lines • ▼ Show 20 Lines
; ;
; IS__TUNIT_NPM: Function Attrs: nofree nosync nounwind willreturn ; IS__TUNIT_NPM: Function Attrs: nofree nosync nounwind willreturn
; IS__TUNIT_NPM-LABEL: define {{[^@]+}}@complicated_args_preallocated ; IS__TUNIT_NPM-LABEL: define {{[^@]+}}@complicated_args_preallocated
; IS__TUNIT_NPM-SAME: () #[[ATTR0:[0-9]+]] { ; IS__TUNIT_NPM-SAME: () #[[ATTR0:[0-9]+]] {
; IS__TUNIT_NPM-NEXT: [[C:%.*]] = call token @llvm.call.preallocated.setup(i32 noundef 1) #[[ATTR4:[0-9]+]] ; IS__TUNIT_NPM-NEXT: [[C:%.*]] = call token @llvm.call.preallocated.setup(i32 noundef 1) #[[ATTR4:[0-9]+]]
; IS__TUNIT_NPM-NEXT: [[CALL:%.*]] = call noundef nonnull align 536870912 dereferenceable(4) i32* @test_preallocated(i32* noalias nocapture nofree noundef writeonly preallocated(i32) align 536870912 null) #[[ATTR1]] [ "preallocated"(token [[C]]) ] ; IS__TUNIT_NPM-NEXT: [[CALL:%.*]] = call noundef nonnull align 536870912 dereferenceable(4) i32* @test_preallocated(i32* noalias nocapture nofree noundef writeonly preallocated(i32) align 536870912 null) #[[ATTR1]] [ "preallocated"(token [[C]]) ]
; IS__TUNIT_NPM-NEXT: ret i32* [[CALL]] ; IS__TUNIT_NPM-NEXT: ret i32* [[CALL]]
; ;
; IS__CGSCC____: Function Attrs: nofree nosync nounwind willreturn ; IS__CGSCC_OPM: Function Attrs: nofree nosync nounwind willreturn
; IS__CGSCC____-LABEL: define {{[^@]+}}@complicated_args_preallocated ; IS__CGSCC_OPM-LABEL: define {{[^@]+}}@complicated_args_preallocated
; IS__CGSCC____-SAME: () #[[ATTR0:[0-9]+]] { ; IS__CGSCC_OPM-SAME: () #[[ATTR0:[0-9]+]] {
; IS__CGSCC____-NEXT: [[C:%.*]] = call token @llvm.call.preallocated.setup(i32 noundef 1) #[[ATTR5:[0-9]+]] ; IS__CGSCC_OPM-NEXT: [[C:%.*]] = call token @llvm.call.preallocated.setup(i32 noundef 1) #[[ATTR5:[0-9]+]]
; IS__CGSCC____-NEXT: ret i32* null ; IS__CGSCC_OPM-NEXT: ret i32* null
;
; IS__CGSCC_NPM: Function Attrs: nofree nosync nounwind willreturn
; IS__CGSCC_NPM-LABEL: define {{[^@]+}}@complicated_args_preallocated
; IS__CGSCC_NPM-SAME: () #[[ATTR0:[0-9]+]] {
; IS__CGSCC_NPM-NEXT: [[C:%.*]] = call token @llvm.call.preallocated.setup(i32 noundef 1) #[[ATTR4:[0-9]+]]
; IS__CGSCC_NPM-NEXT: ret i32* null
; ;
%c = call token @llvm.call.preallocated.setup(i32 1) %c = call token @llvm.call.preallocated.setup(i32 1)
%call = call i32* @test_preallocated(i32* preallocated(i32) null) ["preallocated"(token %c)] %call = call i32* @test_preallocated(i32* preallocated(i32) null) ["preallocated"(token %c)]
ret i32* %call ret i32* %call
} }
define internal void @test_sret(%struct.X* sret(%struct.X) %a, %struct.X** %b) { define internal void @test_sret(%struct.X* sret(%struct.X) %a, %struct.X** %b) {
; ;
▲ Show 20 Lines • Show All 68 Lines • ▼ Show 20 Lines
; IS__CGSCC_OPM-NEXT: store i8* null, i8** [[G0]], align 8 ; IS__CGSCC_OPM-NEXT: store i8* null, i8** [[G0]], align 8
; IS__CGSCC_OPM-NEXT: ret void ; IS__CGSCC_OPM-NEXT: ret void
; ;
; IS__CGSCC_NPM: Function Attrs: nofree norecurse nosync nounwind readnone willreturn ; IS__CGSCC_NPM: Function Attrs: nofree norecurse nosync nounwind readnone willreturn
; IS__CGSCC_NPM-LABEL: define {{[^@]+}}@test_byval ; IS__CGSCC_NPM-LABEL: define {{[^@]+}}@test_byval
; IS__CGSCC_NPM-SAME: (i8* noalias nocapture nofree readnone [[TMP0:%.*]]) #[[ATTR1]] { ; IS__CGSCC_NPM-SAME: (i8* noalias nocapture nofree readnone [[TMP0:%.*]]) #[[ATTR1]] {
; IS__CGSCC_NPM-NEXT: [[A_PRIV:%.*]] = alloca [[STRUCT_X:%.*]], align 8 ; IS__CGSCC_NPM-NEXT: [[A_PRIV:%.*]] = alloca [[STRUCT_X:%.*]], align 8
; IS__CGSCC_NPM-NEXT: [[A_PRIV_CAST:%.*]] = bitcast %struct.X* [[A_PRIV]] to i8** ; IS__CGSCC_NPM-NEXT: [[A_PRIV_CAST:%.*]] = bitcast %struct.X* [[A_PRIV]] to i8**
; IS__CGSCC_NPM-NEXT: store i8* [[TMP0]], i8** [[A_PRIV_CAST]], align 8 ; IS__CGSCC_NPM-NEXT: store i8* undef, i8** [[A_PRIV_CAST]], align 8
; IS__CGSCC_NPM-NEXT: [[G0:%.*]] = getelementptr [[STRUCT_X]], %struct.X* [[A_PRIV]], i32 0, i32 0 ; IS__CGSCC_NPM-NEXT: [[G0:%.*]] = getelementptr [[STRUCT_X]], %struct.X* [[A_PRIV]], i32 0, i32 0
; IS__CGSCC_NPM-NEXT: store i8* null, i8** [[G0]], align 8 ; IS__CGSCC_NPM-NEXT: store i8* null, i8** [[G0]], align 8
; IS__CGSCC_NPM-NEXT: ret void ; IS__CGSCC_NPM-NEXT: ret void
; ;
%g0 = getelementptr %struct.X, %struct.X* %a, i32 0, i32 0 %g0 = getelementptr %struct.X, %struct.X* %a, i32 0, i32 0
store i8* null, i8** %g0 store i8* null, i8** %g0
ret void ret void
} }
define void @complicated_args_byval() { define void @complicated_args_byval() {
; IS__TUNIT____: Function Attrs: nofree nosync nounwind readnone willreturn ; IS__TUNIT____: Function Attrs: nofree nosync nounwind readnone willreturn
; IS__TUNIT____-LABEL: define {{[^@]+}}@complicated_args_byval ; IS__TUNIT____-LABEL: define {{[^@]+}}@complicated_args_byval
; IS__TUNIT____-SAME: () #[[ATTR1]] { ; IS__TUNIT____-SAME: () #[[ATTR1]] {
; IS__TUNIT____-NEXT: ret void ; IS__TUNIT____-NEXT: ret void
; ;
; IS__CGSCC_OPM: Function Attrs: nofree norecurse nosync nounwind readnone willreturn ; IS__CGSCC____: Function Attrs: nofree norecurse nosync nounwind readnone willreturn
; IS__CGSCC_OPM-LABEL: define {{[^@]+}}@complicated_args_byval ; IS__CGSCC____-LABEL: define {{[^@]+}}@complicated_args_byval
; IS__CGSCC_OPM-SAME: () #[[ATTR1]] { ; IS__CGSCC____-SAME: () #[[ATTR1]] {
; IS__CGSCC_OPM-NEXT: ret void ; IS__CGSCC____-NEXT: ret void
;
; IS__CGSCC_NPM: Function Attrs: nofree norecurse nosync nounwind readonly willreturn
; IS__CGSCC_NPM-LABEL: define {{[^@]+}}@complicated_args_byval
; IS__CGSCC_NPM-SAME: () #[[ATTR3:[0-9]+]] {
; IS__CGSCC_NPM-NEXT: ret void
; ;
call void @test_byval(%struct.X* byval(%struct.X) @S) call void @test_byval(%struct.X* byval(%struct.X) @S)
ret void ret void
} }
declare void @sync() declare void @sync()
; Make sure we *do not* load @S here! ; Make sure we *do not* load @S here!
define internal i8*@test_byval2(%struct.X* byval(%struct.X) %a) { define internal i8*@test_byval2(%struct.X* byval(%struct.X) %a) {
Show All 10 Lines
; IS__TUNIT_NPM-NEXT: [[A_PRIV_CAST:%.*]] = bitcast %struct.X* [[A_PRIV]] to i8** ; IS__TUNIT_NPM-NEXT: [[A_PRIV_CAST:%.*]] = bitcast %struct.X* [[A_PRIV]] to i8**
; IS__TUNIT_NPM-NEXT: store i8* [[TMP0]], i8** [[A_PRIV_CAST]], align 8 ; IS__TUNIT_NPM-NEXT: store i8* [[TMP0]], i8** [[A_PRIV_CAST]], align 8
; IS__TUNIT_NPM-NEXT: call void @sync() ; IS__TUNIT_NPM-NEXT: call void @sync()
; IS__TUNIT_NPM-NEXT: [[G0:%.*]] = getelementptr [[STRUCT_X]], %struct.X* [[A_PRIV]], i32 0, i32 0 ; IS__TUNIT_NPM-NEXT: [[G0:%.*]] = getelementptr [[STRUCT_X]], %struct.X* [[A_PRIV]], i32 0, i32 0
; IS__TUNIT_NPM-NEXT: [[L:%.*]] = load i8*, i8** [[G0]], align 8 ; IS__TUNIT_NPM-NEXT: [[L:%.*]] = load i8*, i8** [[G0]], align 8
; IS__TUNIT_NPM-NEXT: ret i8* [[L]] ; IS__TUNIT_NPM-NEXT: ret i8* [[L]]
; ;
; IS__CGSCC_NPM-LABEL: define {{[^@]+}}@test_byval2 ; IS__CGSCC_NPM-LABEL: define {{[^@]+}}@test_byval2
; IS__CGSCC_NPM-SAME: (i8* nofree [[TMP0:%.*]]) { ; IS__CGSCC_NPM-SAME: (i8* noalias nofree readnone "no-capture-maybe-returned" [[TMP0:%.*]]) {
; IS__CGSCC_NPM-NEXT: [[A_PRIV:%.*]] = alloca [[STRUCT_X:%.*]], align 8 ; IS__CGSCC_NPM-NEXT: [[A_PRIV:%.*]] = alloca [[STRUCT_X:%.*]], align 8
; IS__CGSCC_NPM-NEXT: [[A_PRIV_CAST:%.*]] = bitcast %struct.X* [[A_PRIV]] to i8** ; IS__CGSCC_NPM-NEXT: [[A_PRIV_CAST:%.*]] = bitcast %struct.X* [[A_PRIV]] to i8**
; IS__CGSCC_NPM-NEXT: store i8* [[TMP0]], i8** [[A_PRIV_CAST]], align 8 ; IS__CGSCC_NPM-NEXT: store i8* [[TMP0]], i8** [[A_PRIV_CAST]], align 8
; IS__CGSCC_NPM-NEXT: call void @sync() ; IS__CGSCC_NPM-NEXT: call void @sync()
; IS__CGSCC_NPM-NEXT: [[G0:%.*]] = getelementptr [[STRUCT_X]], %struct.X* [[A_PRIV]], i32 0, i32 0 ; IS__CGSCC_NPM-NEXT: [[G0:%.*]] = getelementptr [[STRUCT_X]], %struct.X* [[A_PRIV]], i32 0, i32 0
; IS__CGSCC_NPM-NEXT: [[L:%.*]] = load i8*, i8** [[G0]], align 8 ; IS__CGSCC_NPM-NEXT: [[L:%.*]] = load i8*, i8** [[G0]], align 8
; IS__CGSCC_NPM-NEXT: ret i8* [[L]] ; IS__CGSCC_NPM-NEXT: ret i8* [[L]]
; ;
Show All 15 Lines
; IS__TUNIT_NPM-NEXT: ret i8* [[C]] ; IS__TUNIT_NPM-NEXT: ret i8* [[C]]
; ;
; IS__CGSCC_OPM-LABEL: define {{[^@]+}}@complicated_args_byval2() { ; IS__CGSCC_OPM-LABEL: define {{[^@]+}}@complicated_args_byval2() {
; IS__CGSCC_OPM-NEXT: [[C:%.*]] = call i8* @test_byval2(%struct.X* noalias nocapture nofree noundef nonnull readonly byval([[STRUCT_X:%.*]]) align 8 dereferenceable(8) @S) ; IS__CGSCC_OPM-NEXT: [[C:%.*]] = call i8* @test_byval2(%struct.X* noalias nocapture nofree noundef nonnull readonly byval([[STRUCT_X:%.*]]) align 8 dereferenceable(8) @S)
; IS__CGSCC_OPM-NEXT: ret i8* [[C]] ; IS__CGSCC_OPM-NEXT: ret i8* [[C]]
; ;
; IS__CGSCC_NPM-LABEL: define {{[^@]+}}@complicated_args_byval2() { ; IS__CGSCC_NPM-LABEL: define {{[^@]+}}@complicated_args_byval2() {
; IS__CGSCC_NPM-NEXT: [[TMP1:%.*]] = load i8*, i8** getelementptr inbounds ([[STRUCT_X:%.*]], %struct.X* @S, i32 0, i32 0), align 8 ; IS__CGSCC_NPM-NEXT: [[TMP1:%.*]] = load i8*, i8** getelementptr inbounds ([[STRUCT_X:%.*]], %struct.X* @S, i32 0, i32 0), align 8
; IS__CGSCC_NPM-NEXT: [[C:%.*]] = call i8* @test_byval2(i8* nofree [[TMP1]]) ; IS__CGSCC_NPM-NEXT: [[C:%.*]] = call i8* @test_byval2(i8* noalias nofree readnone "no-capture-maybe-returned" [[TMP1]])
; IS__CGSCC_NPM-NEXT: ret i8* [[C]] ; IS__CGSCC_NPM-NEXT: ret i8* [[C]]
; ;
%c = call i8* @test_byval2(%struct.X* byval(%struct.X) @S) %c = call i8* @test_byval2(%struct.X* byval(%struct.X) @S)
ret i8* %c ret i8* %c
} }
define void @fixpoint_changed(i32* %p) { define void @fixpoint_changed(i32* %p) {
; IS__TUNIT_OPM: Function Attrs: argmemonly nofree nosync nounwind writeonly ; IS__TUNIT_OPM: Function Attrs: argmemonly nofree nosync nounwind writeonly
▲ Show 20 Lines • Show All 217 Lines • ▼ Show 20 Lines
; IS__TUNIT_OPM-NEXT: ret void ; IS__TUNIT_OPM-NEXT: ret void
; ;
; IS__TUNIT_NPM: Function Attrs: nofree nosync nounwind willreturn writeonly ; IS__TUNIT_NPM: Function Attrs: nofree nosync nounwind willreturn writeonly
; IS__TUNIT_NPM-LABEL: define {{[^@]+}}@user_as3 ; IS__TUNIT_NPM-LABEL: define {{[^@]+}}@user_as3
; IS__TUNIT_NPM-SAME: () #[[ATTR3]] { ; IS__TUNIT_NPM-SAME: () #[[ATTR3]] {
; IS__TUNIT_NPM-NEXT: store i32 0, i32 addrspace(3)* @ConstAS3Ptr, align 4 ; IS__TUNIT_NPM-NEXT: store i32 0, i32 addrspace(3)* @ConstAS3Ptr, align 4
; IS__TUNIT_NPM-NEXT: ret void ; IS__TUNIT_NPM-NEXT: ret void
; ;
; IS__CGSCC____: Function Attrs: nofree norecurse nosync nounwind willreturn writeonly ; IS__CGSCC_OPM: Function Attrs: nofree norecurse nosync nounwind willreturn writeonly
; IS__CGSCC____-LABEL: define {{[^@]+}}@user_as3 ; IS__CGSCC_OPM-LABEL: define {{[^@]+}}@user_as3
; IS__CGSCC____-SAME: () #[[ATTR4:[0-9]+]] { ; IS__CGSCC_OPM-SAME: () #[[ATTR4:[0-9]+]] {
; IS__CGSCC____-NEXT: store i32 0, i32 addrspace(3)* @ConstAS3Ptr, align 4 ; IS__CGSCC_OPM-NEXT: store i32 0, i32 addrspace(3)* @ConstAS3Ptr, align 4
; IS__CGSCC____-NEXT: ret void ; IS__CGSCC_OPM-NEXT: ret void
;
; IS__CGSCC_NPM: Function Attrs: nofree norecurse nosync nounwind willreturn writeonly
; IS__CGSCC_NPM-LABEL: define {{[^@]+}}@user_as3
; IS__CGSCC_NPM-SAME: () #[[ATTR3:[0-9]+]] {
; IS__CGSCC_NPM-NEXT: store i32 0, i32 addrspace(3)* @ConstAS3Ptr, align 4
; IS__CGSCC_NPM-NEXT: ret void
; ;
%call = call fastcc i32 addrspace(3)* @const_ptr_return_as3() %call = call fastcc i32 addrspace(3)* @const_ptr_return_as3()
store i32 0, i32 addrspace(3)* %call store i32 0, i32 addrspace(3)* %call
ret void ret void
} }
define void @user() { define void @user() {
; IS__TUNIT_OPM: Function Attrs: nofree nosync nounwind willreturn writeonly ; IS__TUNIT_OPM: Function Attrs: nofree nosync nounwind willreturn writeonly
; IS__TUNIT_OPM-LABEL: define {{[^@]+}}@user ; IS__TUNIT_OPM-LABEL: define {{[^@]+}}@user
; IS__TUNIT_OPM-SAME: () #[[ATTR4]] { ; IS__TUNIT_OPM-SAME: () #[[ATTR4]] {
; IS__TUNIT_OPM-NEXT: store i32 0, i32* addrspacecast (i32 addrspace(3)* @ConstAS3Ptr to i32*), align 4 ; IS__TUNIT_OPM-NEXT: store i32 0, i32* addrspacecast (i32 addrspace(3)* @ConstAS3Ptr to i32*), align 4
; IS__TUNIT_OPM-NEXT: ret void ; IS__TUNIT_OPM-NEXT: ret void
; ;
; IS__TUNIT_NPM: Function Attrs: nofree nosync nounwind willreturn writeonly ; IS__TUNIT_NPM: Function Attrs: nofree nosync nounwind willreturn writeonly
; IS__TUNIT_NPM-LABEL: define {{[^@]+}}@user ; IS__TUNIT_NPM-LABEL: define {{[^@]+}}@user
; IS__TUNIT_NPM-SAME: () #[[ATTR3]] { ; IS__TUNIT_NPM-SAME: () #[[ATTR3]] {
; IS__TUNIT_NPM-NEXT: store i32 0, i32* addrspacecast (i32 addrspace(3)* @ConstAS3Ptr to i32*), align 4 ; IS__TUNIT_NPM-NEXT: store i32 0, i32* addrspacecast (i32 addrspace(3)* @ConstAS3Ptr to i32*), align 4
; IS__TUNIT_NPM-NEXT: ret void ; IS__TUNIT_NPM-NEXT: ret void
; ;
; IS__CGSCC____: Function Attrs: nofree norecurse nosync nounwind willreturn writeonly ; IS__CGSCC_OPM: Function Attrs: nofree norecurse nosync nounwind willreturn writeonly
; IS__CGSCC____-LABEL: define {{[^@]+}}@user ; IS__CGSCC_OPM-LABEL: define {{[^@]+}}@user
; IS__CGSCC____-SAME: () #[[ATTR4]] { ; IS__CGSCC_OPM-SAME: () #[[ATTR4]] {
; IS__CGSCC____-NEXT: store i32 0, i32* addrspacecast (i32 addrspace(3)* @ConstAS3Ptr to i32*), align 4 ; IS__CGSCC_OPM-NEXT: store i32 0, i32* addrspacecast (i32 addrspace(3)* @ConstAS3Ptr to i32*), align 4
; IS__CGSCC____-NEXT: ret void ; IS__CGSCC_OPM-NEXT: ret void
;
; IS__CGSCC_NPM: Function Attrs: nofree norecurse nosync nounwind willreturn writeonly
; IS__CGSCC_NPM-LABEL: define {{[^@]+}}@user
; IS__CGSCC_NPM-SAME: () #[[ATTR3]] {
; IS__CGSCC_NPM-NEXT: store i32 0, i32* addrspacecast (i32 addrspace(3)* @ConstAS3Ptr to i32*), align 4
; IS__CGSCC_NPM-NEXT: ret void
; ;
%call = call fastcc i32* @const_ptr_return() %call = call fastcc i32* @const_ptr_return()
store i32 0, i32* %call store i32 0, i32* %call
ret void ret void
} }
define i1 @test_merge_with_undef_values_ptr(i1 %c) { define i1 @test_merge_with_undef_values_ptr(i1 %c) {
▲ Show 20 Lines • Show All 346 Lines • ▼ Show 20 Lines
; IS__CGSCC_OPM: attributes #[[ATTR2]] = { argmemonly nofree norecurse nosync nounwind willreturn writeonly } ; IS__CGSCC_OPM: attributes #[[ATTR2]] = { argmemonly nofree norecurse nosync nounwind willreturn writeonly }
; IS__CGSCC_OPM: attributes #[[ATTR3]] = { argmemonly nofree norecurse nosync nounwind writeonly } ; IS__CGSCC_OPM: attributes #[[ATTR3]] = { argmemonly nofree norecurse nosync nounwind writeonly }
; IS__CGSCC_OPM: attributes #[[ATTR4]] = { nofree norecurse nosync nounwind willreturn writeonly } ; IS__CGSCC_OPM: attributes #[[ATTR4]] = { nofree norecurse nosync nounwind willreturn writeonly }
; IS__CGSCC_OPM: attributes #[[ATTR5]] = { willreturn } ; IS__CGSCC_OPM: attributes #[[ATTR5]] = { willreturn }
;. ;.
; IS__CGSCC_NPM: attributes #[[ATTR0]] = { nofree nosync nounwind willreturn } ; IS__CGSCC_NPM: attributes #[[ATTR0]] = { nofree nosync nounwind willreturn }
; IS__CGSCC_NPM: attributes #[[ATTR1]] = { nofree norecurse nosync nounwind readnone willreturn } ; IS__CGSCC_NPM: attributes #[[ATTR1]] = { nofree norecurse nosync nounwind readnone willreturn }
; IS__CGSCC_NPM: attributes #[[ATTR2]] = { argmemonly nofree norecurse nosync nounwind willreturn writeonly } ; IS__CGSCC_NPM: attributes #[[ATTR2]] = { argmemonly nofree norecurse nosync nounwind willreturn writeonly }
; IS__CGSCC_NPM: attributes #[[ATTR3]] = { nofree norecurse nosync nounwind readonly willreturn } ; IS__CGSCC_NPM: attributes #[[ATTR3]] = { nofree norecurse nosync nounwind willreturn writeonly }
; IS__CGSCC_NPM: attributes #[[ATTR4]] = { nofree norecurse nosync nounwind willreturn writeonly } ; IS__CGSCC_NPM: attributes #[[ATTR4]] = { willreturn }
; IS__CGSCC_NPM: attributes #[[ATTR5]] = { willreturn }
;. ;.
; CHECK: [[RNG0]] = !{i32 0, i32 -2147483648} ; CHECK: [[RNG0]] = !{i32 0, i32 -2147483648}
;. ;.

llvm/test/Transforms/OpenMP/custom_state_machines.ll

Show First 20 Lines • Show All 913 Lines • ▼ Show 20 Lines
; CHECK-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; CHECK-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4
; CHECK-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 ; CHECK-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4
; CHECK-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 ; CHECK-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4
; CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* noalias noundef nonnull readnone align 8 dereferenceable(24) @[[GLOB1:[0-9]+]], i1 noundef false, i1 noundef false, i1 noundef true) ; CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* noalias noundef nonnull readnone align 8 dereferenceable(24) @[[GLOB1:[0-9]+]], i1 noundef false, i1 noundef false, i1 noundef true)
; CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1
; CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]]
; CHECK: user_code.entry: ; CHECK: user_code.entry:
; CHECK-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR2:[0-9]+]] ; CHECK-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR2:[0-9]+]]
; CHECK-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4 ; CHECK-NEXT: store i32 undef, i32* [[DOTTHREADID_TEMP_]], align 4
; CHECK-NEXT: call void @__omp_outlined__(i32* noundef nonnull align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noundef nonnull align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR2]] ; CHECK-NEXT: call void @__omp_outlined__(i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR2]]
; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true) ; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; CHECK: worker.exit: ; CHECK: worker.exit:
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK-LABEL: define {{[^@]+}}@__kmpc_target_init ; CHECK-LABEL: define {{[^@]+}}@__kmpc_target_init
; CHECK-SAME: (%struct.ident_t* noalias nocapture nofree nonnull readnone align 8 dereferenceable(24) [[TMP0:%.*]], i1 [[TMP1:%.*]], i1 [[USE_GENERIC_STATE_MACHINE:%.*]], i1 [[TMP2:%.*]]) { ; CHECK-SAME: (%struct.ident_t* noalias nocapture nofree nonnull readnone align 8 dereferenceable(24) [[TMP0:%.*]], i1 [[TMP1:%.*]], i1 [[USE_GENERIC_STATE_MACHINE:%.*]], i1 [[TMP2:%.*]]) {
; CHECK-NEXT: store i1 false, i1* @V, align 4 ; CHECK-NEXT: store i1 false, i1* @V, align 4
; CHECK-NEXT: [[CALL:%.*]] = call i32 @unknown() ; CHECK-NEXT: [[CALL:%.*]] = call i32 @unknown()
; CHECK-NEXT: ret i32 [[CALL]] ; CHECK-NEXT: ret i32 [[CALL]]
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__ ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__
; CHECK-SAME: (i32* noalias nofree noundef nonnull align 4 dereferenceable(4) [[DOTGLOBAL_TID_:%.*]], i32* noalias nofree noundef nonnull align 4 dereferenceable(4) [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i32* noalias nocapture nofree nonnull readnone align 4 dereferenceable(4) [[DOTGLOBAL_TID_:%.*]], i32* noalias nocapture nofree nonnull readnone align 4 dereferenceable(4) [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTGLOBAL_TID__ADDR]], align 8
; CHECK-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTBOUND_TID__ADDR]], align 8
; CHECK-NEXT: call void @no_parallel_region_in_here.internalized() #[[ATTR7:[0-9]+]] ; CHECK-NEXT: call void @no_parallel_region_in_here.internalized() #[[ATTR7:[0-9]+]]
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent nounwind ; CHECK: Function Attrs: convergent nounwind
; CHECK-LABEL: define {{[^@]+}}@no_parallel_region_in_here.internalized ; CHECK-LABEL: define {{[^@]+}}@no_parallel_region_in_here.internalized
; CHECK-SAME: () #[[ATTR1:[0-9]+]] { ; CHECK-SAME: () #[[ATTR1:[0-9]+]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
▲ Show 20 Lines • Show All 69 Lines • ▼ Show 20 Lines
; CHECK-NEXT: call void @__kmpc_barrier_simple_spmd(%struct.ident_t* @[[GLOB1]], i32 [[TMP0]]) ; CHECK-NEXT: call void @__kmpc_barrier_simple_spmd(%struct.ident_t* @[[GLOB1]], i32 [[TMP0]])
; CHECK-NEXT: br label [[WORKER_STATE_MACHINE_BEGIN]] ; CHECK-NEXT: br label [[WORKER_STATE_MACHINE_BEGIN]]
; CHECK: thread.user_code.check: ; CHECK: thread.user_code.check:
; CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1
; CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]]
; CHECK: user_code.entry: ; CHECK: user_code.entry:
; CHECK-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR2]] ; CHECK-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR2]]
; CHECK-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4 ; CHECK-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; CHECK-NEXT: call void @__omp_outlined__1(i32* noundef nonnull align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noundef nonnull align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR2]] ; CHECK-NEXT: call void @__omp_outlined__1(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR2]]
; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true) ; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; CHECK: worker.exit: ; CHECK: worker.exit:
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__1 ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__1
; CHECK-SAME: (i32* noalias nofree noundef nonnull align 4 dereferenceable(4) [[DOTGLOBAL_TID_:%.*]], i32* noalias nofree noundef nonnull align 4 dereferenceable(4) [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i32* noalias nocapture nofree noundef nonnull readonly align 4 dereferenceable(4) [[DOTGLOBAL_TID_:%.*]], i32* noalias nocapture nofree nonnull readnone align 4 dereferenceable(4) [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [0 x i8*], align 8 ; CHECK-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [0 x i8*], align 8
; CHECK-NEXT: [[CAPTURED_VARS_ADDRS1:%.*]] = alloca [0 x i8*], align 8 ; CHECK-NEXT: [[CAPTURED_VARS_ADDRS1:%.*]] = alloca [0 x i8*], align 8
; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8
; CHECK-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTBOUND_TID__ADDR]], align 8
; CHECK-NEXT: [[TMP0:%.*]] = load i32, i32* [[DOTGLOBAL_TID_]], align 4 ; CHECK-NEXT: [[TMP0:%.*]] = load i32, i32* [[DOTGLOBAL_TID_]], align 4
; CHECK-NEXT: [[TMP1:%.*]] = bitcast [0 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8** ; CHECK-NEXT: [[TMP1:%.*]] = bitcast [0 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
; CHECK-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* noundef @[[GLOB1]], i32 [[TMP0]], i32 noundef 1, i32 noundef -1, i32 noundef -1, i8* noundef bitcast (void (i32*, i32*)* @__omp_outlined__2 to i8*), i8* noundef @__omp_outlined__2_wrapper.ID, i8** noundef [[TMP1]], i64 noundef 0) ; CHECK-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* noundef @[[GLOB1]], i32 [[TMP0]], i32 noundef 1, i32 noundef -1, i32 noundef -1, i8* noundef bitcast (void (i32*, i32*)* @__omp_outlined__2 to i8*), i8* noundef @__omp_outlined__2_wrapper.ID, i8** noundef [[TMP1]], i64 noundef 0)
; CHECK-NEXT: call void @no_parallel_region_in_here.internalized() #[[ATTR7]] ; CHECK-NEXT: call void @no_parallel_region_in_here.internalized() #[[ATTR7]]
; CHECK-NEXT: [[TMP2:%.*]] = bitcast [0 x i8*]* [[CAPTURED_VARS_ADDRS1]] to i8** ; CHECK-NEXT: [[TMP2:%.*]] = bitcast [0 x i8*]* [[CAPTURED_VARS_ADDRS1]] to i8**
; CHECK-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* noundef @[[GLOB1]], i32 [[TMP0]], i32 noundef 1, i32 noundef -1, i32 noundef -1, i8* noundef bitcast (void (i32*, i32*)* @__omp_outlined__3 to i8*), i8* noundef @__omp_outlined__3_wrapper.ID, i8** noundef [[TMP2]], i64 noundef 0) ; CHECK-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* noundef @[[GLOB1]], i32 [[TMP0]], i32 noundef 1, i32 noundef -1, i32 noundef -1, i8* noundef bitcast (void (i32*, i32*)* @__omp_outlined__3 to i8*), i8* noundef @__omp_outlined__3_wrapper.ID, i8** noundef [[TMP2]], i64 noundef 0)
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__2 ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__2
; CHECK-SAME: (i32* noalias nofree [[DOTGLOBAL_TID_:%.*]], i32* noalias nofree [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i32* noalias nocapture nofree readnone [[DOTGLOBAL_TID_:%.*]], i32* noalias nocapture nofree readnone [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTGLOBAL_TID__ADDR]], align 8
; CHECK-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTBOUND_TID__ADDR]], align 8
; CHECK-NEXT: call void @p0() #[[ATTR8:[0-9]+]] ; CHECK-NEXT: call void @p0() #[[ATTR8:[0-9]+]]
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__2_wrapper ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__2_wrapper
; CHECK-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTADDR:%.*]] = alloca i16, align 2 ; CHECK-NEXT: [[DOTADDR:%.*]] = alloca i16, align 2
; CHECK-NEXT: [[DOTADDR1:%.*]] = alloca i32, align 4 ; CHECK-NEXT: [[DOTADDR1:%.*]] = alloca i32, align 4
; CHECK-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; CHECK-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4
; CHECK-NEXT: [[GLOBAL_ARGS:%.*]] = alloca i8**, align 8 ; CHECK-NEXT: [[GLOBAL_ARGS:%.*]] = alloca i8**, align 8
; CHECK-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 ; CHECK-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4
; CHECK-NEXT: store i16 [[TMP0]], i16* [[DOTADDR]], align 2 ; CHECK-NEXT: store i16 [[TMP0]], i16* [[DOTADDR]], align 2
; CHECK-NEXT: store i32 [[TMP1]], i32* [[DOTADDR1]], align 4 ; CHECK-NEXT: store i32 [[TMP1]], i32* [[DOTADDR1]], align 4
; CHECK-NEXT: call void @__kmpc_get_shared_variables(i8*** [[GLOBAL_ARGS]]) ; CHECK-NEXT: call void @__kmpc_get_shared_variables(i8*** [[GLOBAL_ARGS]])
; CHECK-NEXT: call void @__omp_outlined__2(i32* [[DOTADDR1]], i32* [[DOTZERO_ADDR]]) #[[ATTR2]] ; CHECK-NEXT: call void @__omp_outlined__2(i32* [[DOTADDR1]], i32* [[DOTZERO_ADDR]]) #[[ATTR2]]
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__3 ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__3
; CHECK-SAME: (i32* noalias nofree [[DOTGLOBAL_TID_:%.*]], i32* noalias nofree [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i32* noalias nocapture nofree readnone [[DOTGLOBAL_TID_:%.*]], i32* noalias nocapture nofree readnone [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTGLOBAL_TID__ADDR]], align 8
; CHECK-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTBOUND_TID__ADDR]], align 8
; CHECK-NEXT: call void @p1() #[[ATTR8]] ; CHECK-NEXT: call void @p1() #[[ATTR8]]
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__3_wrapper ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__3_wrapper
; CHECK-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
▲ Show 20 Lines • Show All 57 Lines • ▼ Show 20 Lines
; CHECK-NEXT: call void @__kmpc_barrier_simple_spmd(%struct.ident_t* @[[GLOB1]], i32 [[TMP0]]) ; CHECK-NEXT: call void @__kmpc_barrier_simple_spmd(%struct.ident_t* @[[GLOB1]], i32 [[TMP0]])
; CHECK-NEXT: br label [[WORKER_STATE_MACHINE_BEGIN]] ; CHECK-NEXT: br label [[WORKER_STATE_MACHINE_BEGIN]]
; CHECK: thread.user_code.check: ; CHECK: thread.user_code.check:
; CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1
; CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]]
; CHECK: user_code.entry: ; CHECK: user_code.entry:
; CHECK-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR2]] ; CHECK-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR2]]
; CHECK-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4 ; CHECK-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; CHECK-NEXT: call void @__omp_outlined__4(i32* noundef nonnull align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noundef nonnull align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR2]] ; CHECK-NEXT: call void @__omp_outlined__4(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR2]]
; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true) ; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; CHECK: worker.exit: ; CHECK: worker.exit:
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__4 ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__4
; CHECK-SAME: (i32* noalias nofree noundef nonnull align 4 dereferenceable(4) [[DOTGLOBAL_TID_:%.*]], i32* noalias nofree noundef nonnull align 4 dereferenceable(4) [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i32* noalias nocapture nofree noundef nonnull readonly align 4 dereferenceable(4) [[DOTGLOBAL_TID_:%.*]], i32* noalias nocapture nofree nonnull readnone align 4 dereferenceable(4) [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [0 x i8*], align 8 ; CHECK-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [0 x i8*], align 8
; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8
; CHECK-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTBOUND_TID__ADDR]], align 8
; CHECK-NEXT: call void @simple_state_machine_interprocedural_before.internalized() #[[ATTR7]] ; CHECK-NEXT: call void @simple_state_machine_interprocedural_before.internalized() #[[ATTR7]]
; CHECK-NEXT: call void @no_parallel_region_in_here.internalized() #[[ATTR7]] ; CHECK-NEXT: call void @no_parallel_region_in_here.internalized() #[[ATTR7]]
; CHECK-NEXT: [[TMP0:%.*]] = load i32, i32* [[DOTGLOBAL_TID_]], align 4 ; CHECK-NEXT: [[TMP0:%.*]] = load i32, i32* [[DOTGLOBAL_TID_]], align 4
; CHECK-NEXT: [[TMP1:%.*]] = bitcast [0 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8** ; CHECK-NEXT: [[TMP1:%.*]] = bitcast [0 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
; CHECK-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* noundef @[[GLOB1]], i32 [[TMP0]], i32 noundef 1, i32 noundef -1, i32 noundef -1, i8* noundef bitcast (void (i32*, i32*)* @__omp_outlined__5 to i8*), i8* noundef @__omp_outlined__5_wrapper.ID, i8** noundef [[TMP1]], i64 noundef 0) ; CHECK-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* noundef @[[GLOB1]], i32 [[TMP0]], i32 noundef 1, i32 noundef -1, i32 noundef -1, i8* noundef bitcast (void (i32*, i32*)* @__omp_outlined__5 to i8*), i8* noundef @__omp_outlined__5_wrapper.ID, i8** noundef [[TMP1]], i64 noundef 0)
; CHECK-NEXT: call void @simple_state_machine_interprocedural_after.internalized() #[[ATTR7]] ; CHECK-NEXT: call void @simple_state_machine_interprocedural_after.internalized() #[[ATTR7]]
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
Show All 17 Lines
; CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB2]]) ; CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB2]])
; CHECK-NEXT: [[TMP1:%.*]] = bitcast [0 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8** ; CHECK-NEXT: [[TMP1:%.*]] = bitcast [0 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
; CHECK-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP0]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*)* @__omp_outlined__17 to i8*), i8* bitcast (void (i16, i32)* @__omp_outlined__17_wrapper to i8*), i8** [[TMP1]], i64 0) ; CHECK-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP0]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*)* @__omp_outlined__17 to i8*), i8* bitcast (void (i16, i32)* @__omp_outlined__17_wrapper to i8*), i8** [[TMP1]], i64 0)
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__5 ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__5
; CHECK-SAME: (i32* noalias nofree [[DOTGLOBAL_TID_:%.*]], i32* noalias nofree [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i32* noalias nocapture nofree readnone [[DOTGLOBAL_TID_:%.*]], i32* noalias nocapture nofree readnone [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTGLOBAL_TID__ADDR]], align 8
; CHECK-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTBOUND_TID__ADDR]], align 8
; CHECK-NEXT: call void @p1() #[[ATTR8]] ; CHECK-NEXT: call void @p1() #[[ATTR8]]
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__5_wrapper ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__5_wrapper
; CHECK-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
▲ Show 20 Lines • Show All 75 Lines • ▼ Show 20 Lines
; CHECK-NEXT: call void @__kmpc_barrier_simple_spmd(%struct.ident_t* @[[GLOB1]], i32 [[TMP0]]) ; CHECK-NEXT: call void @__kmpc_barrier_simple_spmd(%struct.ident_t* @[[GLOB1]], i32 [[TMP0]])
; CHECK-NEXT: br label [[WORKER_STATE_MACHINE_BEGIN]] ; CHECK-NEXT: br label [[WORKER_STATE_MACHINE_BEGIN]]
; CHECK: thread.user_code.check: ; CHECK: thread.user_code.check:
; CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1
; CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]]
; CHECK: user_code.entry: ; CHECK: user_code.entry:
; CHECK-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR2]] ; CHECK-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR2]]
; CHECK-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4 ; CHECK-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; CHECK-NEXT: call void @__omp_outlined__6(i32* noundef nonnull align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noundef nonnull align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR2]] ; CHECK-NEXT: call void @__omp_outlined__6(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR2]]
; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true) ; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; CHECK: worker.exit: ; CHECK: worker.exit:
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__6 ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__6
; CHECK-SAME: (i32* noalias nofree noundef nonnull align 4 dereferenceable(4) [[DOTGLOBAL_TID_:%.*]], i32* noalias nofree noundef nonnull align 4 dereferenceable(4) [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i32* noalias nocapture nofree noundef nonnull readonly align 4 dereferenceable(4) [[DOTGLOBAL_TID_:%.*]], i32* noalias nocapture nofree nonnull readnone align 4 dereferenceable(4) [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [0 x i8*], align 8 ; CHECK-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [0 x i8*], align 8
; CHECK-NEXT: [[CAPTURED_VARS_ADDRS1:%.*]] = alloca [0 x i8*], align 8 ; CHECK-NEXT: [[CAPTURED_VARS_ADDRS1:%.*]] = alloca [0 x i8*], align 8
; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8
; CHECK-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTBOUND_TID__ADDR]], align 8
; CHECK-NEXT: [[TMP0:%.*]] = load i32, i32* [[DOTGLOBAL_TID_]], align 4 ; CHECK-NEXT: [[TMP0:%.*]] = load i32, i32* [[DOTGLOBAL_TID_]], align 4
; CHECK-NEXT: [[TMP1:%.*]] = bitcast [0 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8** ; CHECK-NEXT: [[TMP1:%.*]] = bitcast [0 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
; CHECK-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* noundef @[[GLOB1]], i32 [[TMP0]], i32 noundef 1, i32 noundef -1, i32 noundef -1, i8* noundef bitcast (void (i32*, i32*)* @__omp_outlined__7 to i8*), i8* noundef @__omp_outlined__7_wrapper.ID, i8** noundef [[TMP1]], i64 noundef 0) ; CHECK-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* noundef @[[GLOB1]], i32 [[TMP0]], i32 noundef 1, i32 noundef -1, i32 noundef -1, i8* noundef bitcast (void (i32*, i32*)* @__omp_outlined__7 to i8*), i8* noundef @__omp_outlined__7_wrapper.ID, i8** noundef [[TMP1]], i64 noundef 0)
; CHECK-NEXT: [[TMP2:%.*]] = call i32 @unknown() #[[ATTR8]] ; CHECK-NEXT: [[TMP2:%.*]] = call i32 @unknown() #[[ATTR8]]
; CHECK-NEXT: [[TMP3:%.*]] = bitcast [0 x i8*]* [[CAPTURED_VARS_ADDRS1]] to i8** ; CHECK-NEXT: [[TMP3:%.*]] = bitcast [0 x i8*]* [[CAPTURED_VARS_ADDRS1]] to i8**
; CHECK-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* noundef @[[GLOB1]], i32 [[TMP0]], i32 noundef 1, i32 noundef -1, i32 noundef -1, i8* noundef bitcast (void (i32*, i32*)* @__omp_outlined__8 to i8*), i8* noundef @__omp_outlined__8_wrapper.ID, i8** noundef [[TMP3]], i64 noundef 0) ; CHECK-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* noundef @[[GLOB1]], i32 [[TMP0]], i32 noundef 1, i32 noundef -1, i32 noundef -1, i8* noundef bitcast (void (i32*, i32*)* @__omp_outlined__8 to i8*), i8* noundef @__omp_outlined__8_wrapper.ID, i8** noundef [[TMP3]], i64 noundef 0)
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__7 ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__7
; CHECK-SAME: (i32* noalias nofree [[DOTGLOBAL_TID_:%.*]], i32* noalias nofree [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i32* noalias nocapture nofree readnone [[DOTGLOBAL_TID_:%.*]], i32* noalias nocapture nofree readnone [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTGLOBAL_TID__ADDR]], align 8
; CHECK-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTBOUND_TID__ADDR]], align 8
; CHECK-NEXT: call void @p0() #[[ATTR8]] ; CHECK-NEXT: call void @p0() #[[ATTR8]]
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__7_wrapper ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__7_wrapper
; CHECK-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTADDR:%.*]] = alloca i16, align 2 ; CHECK-NEXT: [[DOTADDR:%.*]] = alloca i16, align 2
; CHECK-NEXT: [[DOTADDR1:%.*]] = alloca i32, align 4 ; CHECK-NEXT: [[DOTADDR1:%.*]] = alloca i32, align 4
; CHECK-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; CHECK-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4
; CHECK-NEXT: [[GLOBAL_ARGS:%.*]] = alloca i8**, align 8 ; CHECK-NEXT: [[GLOBAL_ARGS:%.*]] = alloca i8**, align 8
; CHECK-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 ; CHECK-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4
; CHECK-NEXT: store i16 [[TMP0]], i16* [[DOTADDR]], align 2 ; CHECK-NEXT: store i16 [[TMP0]], i16* [[DOTADDR]], align 2
; CHECK-NEXT: store i32 [[TMP1]], i32* [[DOTADDR1]], align 4 ; CHECK-NEXT: store i32 [[TMP1]], i32* [[DOTADDR1]], align 4
; CHECK-NEXT: call void @__kmpc_get_shared_variables(i8*** [[GLOBAL_ARGS]]) ; CHECK-NEXT: call void @__kmpc_get_shared_variables(i8*** [[GLOBAL_ARGS]])
; CHECK-NEXT: call void @__omp_outlined__7(i32* [[DOTADDR1]], i32* [[DOTZERO_ADDR]]) #[[ATTR2]] ; CHECK-NEXT: call void @__omp_outlined__7(i32* [[DOTADDR1]], i32* [[DOTZERO_ADDR]]) #[[ATTR2]]
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__8 ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__8
; CHECK-SAME: (i32* noalias nofree [[DOTGLOBAL_TID_:%.*]], i32* noalias nofree [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i32* noalias nocapture nofree readnone [[DOTGLOBAL_TID_:%.*]], i32* noalias nocapture nofree readnone [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTGLOBAL_TID__ADDR]], align 8
; CHECK-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTBOUND_TID__ADDR]], align 8
; CHECK-NEXT: call void @p1() #[[ATTR8]] ; CHECK-NEXT: call void @p1() #[[ATTR8]]
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__8_wrapper ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__8_wrapper
; CHECK-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
▲ Show 20 Lines • Show All 51 Lines • ▼ Show 20 Lines
; CHECK-NEXT: call void @__kmpc_barrier_simple_spmd(%struct.ident_t* @[[GLOB1]], i32 [[TMP0]]) ; CHECK-NEXT: call void @__kmpc_barrier_simple_spmd(%struct.ident_t* @[[GLOB1]], i32 [[TMP0]])
; CHECK-NEXT: br label [[WORKER_STATE_MACHINE_BEGIN]] ; CHECK-NEXT: br label [[WORKER_STATE_MACHINE_BEGIN]]
; CHECK: thread.user_code.check: ; CHECK: thread.user_code.check:
; CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1
; CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]]
; CHECK: user_code.entry: ; CHECK: user_code.entry:
; CHECK-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR2]] ; CHECK-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR2]]
; CHECK-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4 ; CHECK-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; CHECK-NEXT: call void @__omp_outlined__9(i32* noundef nonnull align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noundef nonnull align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR2]] ; CHECK-NEXT: call void @__omp_outlined__9(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR2]]
; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true) ; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; CHECK: worker.exit: ; CHECK: worker.exit:
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__9 ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__9
; CHECK-SAME: (i32* noalias nofree noundef nonnull align 4 dereferenceable(4) [[DOTGLOBAL_TID_:%.*]], i32* noalias nofree noundef nonnull align 4 dereferenceable(4) [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i32* noalias nocapture nofree noundef nonnull readonly align 4 dereferenceable(4) [[DOTGLOBAL_TID_:%.*]], i32* noalias nocapture nofree nonnull readnone align 4 dereferenceable(4) [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [0 x i8*], align 8 ; CHECK-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [0 x i8*], align 8
; CHECK-NEXT: [[CAPTURED_VARS_ADDRS1:%.*]] = alloca [0 x i8*], align 8 ; CHECK-NEXT: [[CAPTURED_VARS_ADDRS1:%.*]] = alloca [0 x i8*], align 8
; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8
; CHECK-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTBOUND_TID__ADDR]], align 8
; CHECK-NEXT: [[TMP0:%.*]] = load i32, i32* [[DOTGLOBAL_TID_]], align 4 ; CHECK-NEXT: [[TMP0:%.*]] = load i32, i32* [[DOTGLOBAL_TID_]], align 4
; CHECK-NEXT: [[TMP1:%.*]] = bitcast [0 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8** ; CHECK-NEXT: [[TMP1:%.*]] = bitcast [0 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
; CHECK-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* noundef @[[GLOB1]], i32 [[TMP0]], i32 noundef 1, i32 noundef -1, i32 noundef -1, i8* noundef bitcast (void (i32*, i32*)* @__omp_outlined__10 to i8*), i8* noundef @__omp_outlined__10_wrapper.ID, i8** noundef [[TMP1]], i64 noundef 0) ; CHECK-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* noundef @[[GLOB1]], i32 [[TMP0]], i32 noundef 1, i32 noundef -1, i32 noundef -1, i8* noundef bitcast (void (i32*, i32*)* @__omp_outlined__10 to i8*), i8* noundef @__omp_outlined__10_wrapper.ID, i8** noundef [[TMP1]], i64 noundef 0)
; CHECK-NEXT: call void @unknown_no_openmp() #[[ATTR9:[0-9]+]] ; CHECK-NEXT: call void @unknown_no_openmp() #[[ATTR9:[0-9]+]]
; CHECK-NEXT: [[TMP2:%.*]] = bitcast [0 x i8*]* [[CAPTURED_VARS_ADDRS1]] to i8** ; CHECK-NEXT: [[TMP2:%.*]] = bitcast [0 x i8*]* [[CAPTURED_VARS_ADDRS1]] to i8**
; CHECK-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* noundef @[[GLOB1]], i32 [[TMP0]], i32 noundef 1, i32 noundef -1, i32 noundef -1, i8* noundef bitcast (void (i32*, i32*)* @__omp_outlined__11 to i8*), i8* noundef @__omp_outlined__11_wrapper.ID, i8** noundef [[TMP2]], i64 noundef 0) ; CHECK-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* noundef @[[GLOB1]], i32 [[TMP0]], i32 noundef 1, i32 noundef -1, i32 noundef -1, i8* noundef bitcast (void (i32*, i32*)* @__omp_outlined__11 to i8*), i8* noundef @__omp_outlined__11_wrapper.ID, i8** noundef [[TMP2]], i64 noundef 0)
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__10 ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__10
; CHECK-SAME: (i32* noalias nofree [[DOTGLOBAL_TID_:%.*]], i32* noalias nofree [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i32* noalias nocapture nofree readnone [[DOTGLOBAL_TID_:%.*]], i32* noalias nocapture nofree readnone [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTGLOBAL_TID__ADDR]], align 8
; CHECK-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTBOUND_TID__ADDR]], align 8
; CHECK-NEXT: call void @p0() #[[ATTR8]] ; CHECK-NEXT: call void @p0() #[[ATTR8]]
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__10_wrapper ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__10_wrapper
; CHECK-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTADDR:%.*]] = alloca i16, align 2 ; CHECK-NEXT: [[DOTADDR:%.*]] = alloca i16, align 2
; CHECK-NEXT: [[DOTADDR1:%.*]] = alloca i32, align 4 ; CHECK-NEXT: [[DOTADDR1:%.*]] = alloca i32, align 4
; CHECK-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; CHECK-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4
; CHECK-NEXT: [[GLOBAL_ARGS:%.*]] = alloca i8**, align 8 ; CHECK-NEXT: [[GLOBAL_ARGS:%.*]] = alloca i8**, align 8
; CHECK-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 ; CHECK-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4
; CHECK-NEXT: store i16 [[TMP0]], i16* [[DOTADDR]], align 2 ; CHECK-NEXT: store i16 [[TMP0]], i16* [[DOTADDR]], align 2
; CHECK-NEXT: store i32 [[TMP1]], i32* [[DOTADDR1]], align 4 ; CHECK-NEXT: store i32 [[TMP1]], i32* [[DOTADDR1]], align 4
; CHECK-NEXT: call void @__kmpc_get_shared_variables(i8*** [[GLOBAL_ARGS]]) ; CHECK-NEXT: call void @__kmpc_get_shared_variables(i8*** [[GLOBAL_ARGS]])
; CHECK-NEXT: call void @__omp_outlined__10(i32* [[DOTADDR1]], i32* [[DOTZERO_ADDR]]) #[[ATTR2]] ; CHECK-NEXT: call void @__omp_outlined__10(i32* [[DOTADDR1]], i32* [[DOTZERO_ADDR]]) #[[ATTR2]]
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__11 ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__11
; CHECK-SAME: (i32* noalias nofree [[DOTGLOBAL_TID_:%.*]], i32* noalias nofree [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i32* noalias nocapture nofree readnone [[DOTGLOBAL_TID_:%.*]], i32* noalias nocapture nofree readnone [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTGLOBAL_TID__ADDR]], align 8
; CHECK-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTBOUND_TID__ADDR]], align 8
; CHECK-NEXT: call void @p1() #[[ATTR8]] ; CHECK-NEXT: call void @p1() #[[ATTR8]]
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__11_wrapper ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__11_wrapper
; CHECK-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
Show All 17 Lines
; CHECK-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 ; CHECK-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4
; CHECK-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 ; CHECK-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4
; CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* noalias noundef nonnull readnone align 8 dereferenceable(24) @[[GLOB1]], i1 noundef true, i1 noundef false, i1 noundef false) ; CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* noalias noundef nonnull readnone align 8 dereferenceable(24) @[[GLOB1]], i1 noundef true, i1 noundef false, i1 noundef false)
; CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1
; CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]]
; CHECK: user_code.entry: ; CHECK: user_code.entry:
; CHECK-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR2]] ; CHECK-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR2]]
; CHECK-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4 ; CHECK-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; CHECK-NEXT: call void @__omp_outlined__12(i32* noundef nonnull align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noundef nonnull align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR2]] ; CHECK-NEXT: call void @__omp_outlined__12(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR2]]
; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false) ; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; CHECK: worker.exit: ; CHECK: worker.exit:
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__12 ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__12
; CHECK-SAME: (i32* noalias nofree noundef nonnull align 4 dereferenceable(4) [[DOTGLOBAL_TID_:%.*]], i32* noalias nofree noundef nonnull align 4 dereferenceable(4) [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i32* noalias nocapture nofree noundef nonnull readonly align 4 dereferenceable(4) [[DOTGLOBAL_TID_:%.*]], i32* noalias nocapture nofree nonnull readnone align 4 dereferenceable(4) [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [0 x i8*], align 8 ; CHECK-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [0 x i8*], align 8
; CHECK-NEXT: [[CAPTURED_VARS_ADDRS1:%.*]] = alloca [0 x i8*], align 8 ; CHECK-NEXT: [[CAPTURED_VARS_ADDRS1:%.*]] = alloca [0 x i8*], align 8
; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8
; CHECK-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTBOUND_TID__ADDR]], align 8
; CHECK-NEXT: [[TMP0:%.*]] = load i32, i32* [[DOTGLOBAL_TID_]], align 4 ; CHECK-NEXT: [[TMP0:%.*]] = load i32, i32* [[DOTGLOBAL_TID_]], align 4
; CHECK-NEXT: [[TMP1:%.*]] = bitcast [0 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8** ; CHECK-NEXT: [[TMP1:%.*]] = bitcast [0 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
; CHECK-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* noundef @[[GLOB1]], i32 [[TMP0]], i32 noundef 1, i32 noundef -1, i32 noundef -1, i8* noundef bitcast (void (i32*, i32*)* @__omp_outlined__13 to i8*), i8* noundef bitcast (void (i16, i32)* @__omp_outlined__13_wrapper to i8*), i8** noundef [[TMP1]], i64 noundef 0) ; CHECK-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* noundef @[[GLOB1]], i32 [[TMP0]], i32 noundef 1, i32 noundef -1, i32 noundef -1, i8* noundef bitcast (void (i32*, i32*)* @__omp_outlined__13 to i8*), i8* noundef bitcast (void (i16, i32)* @__omp_outlined__13_wrapper to i8*), i8** noundef [[TMP1]], i64 noundef 0)
; CHECK-NEXT: [[TMP2:%.*]] = bitcast [0 x i8*]* [[CAPTURED_VARS_ADDRS1]] to i8** ; CHECK-NEXT: [[TMP2:%.*]] = bitcast [0 x i8*]* [[CAPTURED_VARS_ADDRS1]] to i8**
; CHECK-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* noundef @[[GLOB1]], i32 [[TMP0]], i32 noundef 1, i32 noundef -1, i32 noundef -1, i8* noundef bitcast (void (i32*, i32*)* @__omp_outlined__14 to i8*), i8* noundef bitcast (void (i16, i32)* @__omp_outlined__14_wrapper to i8*), i8** noundef [[TMP2]], i64 noundef 0) ; CHECK-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* noundef @[[GLOB1]], i32 [[TMP0]], i32 noundef 1, i32 noundef -1, i32 noundef -1, i8* noundef bitcast (void (i32*, i32*)* @__omp_outlined__14 to i8*), i8* noundef bitcast (void (i16, i32)* @__omp_outlined__14_wrapper to i8*), i8** noundef [[TMP2]], i64 noundef 0)
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__13 ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__13
; CHECK-SAME: (i32* noalias nofree [[DOTGLOBAL_TID_:%.*]], i32* noalias nofree [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i32* noalias nocapture nofree readnone [[DOTGLOBAL_TID_:%.*]], i32* noalias nocapture nofree readnone [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTGLOBAL_TID__ADDR]], align 8
; CHECK-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTBOUND_TID__ADDR]], align 8
; CHECK-NEXT: call void @p0() #[[ATTR8]] ; CHECK-NEXT: call void @p0() #[[ATTR8]]
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__13_wrapper ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__13_wrapper
; CHECK-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTADDR:%.*]] = alloca i16, align 2 ; CHECK-NEXT: [[DOTADDR:%.*]] = alloca i16, align 2
; CHECK-NEXT: [[DOTADDR1:%.*]] = alloca i32, align 4 ; CHECK-NEXT: [[DOTADDR1:%.*]] = alloca i32, align 4
; CHECK-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; CHECK-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4
; CHECK-NEXT: [[GLOBAL_ARGS:%.*]] = alloca i8**, align 8 ; CHECK-NEXT: [[GLOBAL_ARGS:%.*]] = alloca i8**, align 8
; CHECK-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 ; CHECK-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4
; CHECK-NEXT: store i16 [[TMP0]], i16* [[DOTADDR]], align 2 ; CHECK-NEXT: store i16 [[TMP0]], i16* [[DOTADDR]], align 2
; CHECK-NEXT: store i32 [[TMP1]], i32* [[DOTADDR1]], align 4 ; CHECK-NEXT: store i32 [[TMP1]], i32* [[DOTADDR1]], align 4
; CHECK-NEXT: call void @__kmpc_get_shared_variables(i8*** [[GLOBAL_ARGS]]) ; CHECK-NEXT: call void @__kmpc_get_shared_variables(i8*** [[GLOBAL_ARGS]])
; CHECK-NEXT: call void @__omp_outlined__13(i32* [[DOTADDR1]], i32* [[DOTZERO_ADDR]]) #[[ATTR2]] ; CHECK-NEXT: call void @__omp_outlined__13(i32* [[DOTADDR1]], i32* [[DOTZERO_ADDR]]) #[[ATTR2]]
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__14 ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__14
; CHECK-SAME: (i32* noalias nofree [[DOTGLOBAL_TID_:%.*]], i32* noalias nofree [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i32* noalias nocapture nofree readnone [[DOTGLOBAL_TID_:%.*]], i32* noalias nocapture nofree readnone [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTGLOBAL_TID__ADDR]], align 8
; CHECK-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTBOUND_TID__ADDR]], align 8
; CHECK-NEXT: call void @p1() #[[ATTR8]] ; CHECK-NEXT: call void @p1() #[[ATTR8]]
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__14_wrapper ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__14_wrapper
; CHECK-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
▲ Show 20 Lines • Show All 44 Lines • ▼ Show 20 Lines
; CHECK: worker_state_machine.done.barrier: ; CHECK: worker_state_machine.done.barrier:
; CHECK-NEXT: call void @__kmpc_barrier_simple_spmd(%struct.ident_t* @[[GLOB1]], i32 [[TMP0]]) ; CHECK-NEXT: call void @__kmpc_barrier_simple_spmd(%struct.ident_t* @[[GLOB1]], i32 [[TMP0]])
; CHECK-NEXT: br label [[WORKER_STATE_MACHINE_BEGIN]] ; CHECK-NEXT: br label [[WORKER_STATE_MACHINE_BEGIN]]
; CHECK: thread.user_code.check: ; CHECK: thread.user_code.check:
; CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1
; CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]]
; CHECK: user_code.entry: ; CHECK: user_code.entry:
; CHECK-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR2]] ; CHECK-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR2]]
; CHECK-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4 ; CHECK-NEXT: store i32 undef, i32* [[DOTTHREADID_TEMP_]], align 4
; CHECK-NEXT: call void @__omp_outlined__15(i32* noundef nonnull align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noundef nonnull align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR2]] ; CHECK-NEXT: call void @__omp_outlined__15(i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR2]]
; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true) ; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; CHECK: worker.exit: ; CHECK: worker.exit:
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__15 ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__15
; CHECK-SAME: (i32* noalias nofree noundef nonnull align 4 dereferenceable(4) [[DOTGLOBAL_TID_:%.*]], i32* noalias nofree noundef nonnull align 4 dereferenceable(4) [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i32* noalias nocapture nofree nonnull readnone align 4 dereferenceable(4) [[DOTGLOBAL_TID_:%.*]], i32* noalias nocapture nofree nonnull readnone align 4 dereferenceable(4) [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTGLOBAL_TID__ADDR]], align 8
; CHECK-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTBOUND_TID__ADDR]], align 8
; CHECK-NEXT: [[CALL:%.*]] = call i32 @omp_get_thread_num() #[[ATTR2]] ; CHECK-NEXT: [[CALL:%.*]] = call i32 @omp_get_thread_num() #[[ATTR2]]
; CHECK-NEXT: call void @simple_state_machine_interprocedural_nested_recursive_after.internalized(i32 [[CALL]]) #[[ATTR7]] ; CHECK-NEXT: call void @simple_state_machine_interprocedural_nested_recursive_after.internalized(i32 [[CALL]]) #[[ATTR7]]
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent nounwind ; CHECK: Function Attrs: convergent nounwind
; CHECK-LABEL: define {{[^@]+}}@simple_state_machine_interprocedural_nested_recursive_after.internalized ; CHECK-LABEL: define {{[^@]+}}@simple_state_machine_interprocedural_nested_recursive_after.internalized
; CHECK-SAME: (i32 [[A:%.*]]) #[[ATTR1]] { ; CHECK-SAME: (i32 [[A:%.*]]) #[[ATTR1]] {
▲ Show 20 Lines • Show All 65 Lines • ▼ Show 20 Lines
; CHECK: worker_state_machine.done.barrier: ; CHECK: worker_state_machine.done.barrier:
; CHECK-NEXT: call void @__kmpc_barrier_simple_spmd(%struct.ident_t* @[[GLOB1]], i32 [[TMP0]]) ; CHECK-NEXT: call void @__kmpc_barrier_simple_spmd(%struct.ident_t* @[[GLOB1]], i32 [[TMP0]])
; CHECK-NEXT: br label [[WORKER_STATE_MACHINE_BEGIN]] ; CHECK-NEXT: br label [[WORKER_STATE_MACHINE_BEGIN]]
; CHECK: thread.user_code.check: ; CHECK: thread.user_code.check:
; CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1
; CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]]
; CHECK: user_code.entry: ; CHECK: user_code.entry:
; CHECK-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR2]] ; CHECK-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR2]]
; CHECK-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4 ; CHECK-NEXT: store i32 undef, i32* [[DOTTHREADID_TEMP_]], align 4
; CHECK-NEXT: call void @__omp_outlined__16(i32* noundef nonnull align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noundef nonnull align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR2]] ; CHECK-NEXT: call void @__omp_outlined__16(i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR2]]
; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true) ; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; CHECK: worker.exit: ; CHECK: worker.exit:
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__16 ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__16
; CHECK-SAME: (i32* noalias nofree noundef nonnull align 4 dereferenceable(4) [[DOTGLOBAL_TID_:%.*]], i32* noalias nofree noundef nonnull align 4 dereferenceable(4) [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i32* noalias nocapture nofree nonnull readnone align 4 dereferenceable(4) [[DOTGLOBAL_TID_:%.*]], i32* noalias nocapture nofree nonnull readnone align 4 dereferenceable(4) [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTGLOBAL_TID__ADDR]], align 8
; CHECK-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTBOUND_TID__ADDR]], align 8
; CHECK-NEXT: call void @weak_callee_empty() #[[ATTR7]] ; CHECK-NEXT: call void @weak_callee_empty() #[[ATTR7]]
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent nounwind ; CHECK: Function Attrs: convergent nounwind
; CHECK-LABEL: define {{[^@]+}}@weak_callee_empty ; CHECK-LABEL: define {{[^@]+}}@weak_callee_empty
; CHECK-SAME: () #[[ATTR1]] { ; CHECK-SAME: () #[[ATTR1]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__17 ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__17
; CHECK-SAME: (i32* noalias nofree [[DOTGLOBAL_TID_:%.*]], i32* noalias nofree [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i32* noalias nocapture nofree readnone [[DOTGLOBAL_TID_:%.*]], i32* noalias nocapture nofree readnone [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTGLOBAL_TID__ADDR]], align 8
; CHECK-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTBOUND_TID__ADDR]], align 8
; CHECK-NEXT: call void @p0() #[[ATTR8]] ; CHECK-NEXT: call void @p0() #[[ATTR8]]
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__17_wrapper ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__17_wrapper
; CHECK-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTADDR:%.*]] = alloca i16, align 2 ; CHECK-NEXT: [[DOTADDR:%.*]] = alloca i16, align 2
; CHECK-NEXT: [[DOTADDR1:%.*]] = alloca i32, align 4 ; CHECK-NEXT: [[DOTADDR1:%.*]] = alloca i32, align 4
; CHECK-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; CHECK-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4
; CHECK-NEXT: [[GLOBAL_ARGS:%.*]] = alloca i8**, align 8 ; CHECK-NEXT: [[GLOBAL_ARGS:%.*]] = alloca i8**, align 8
; CHECK-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 ; CHECK-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4
; CHECK-NEXT: store i16 [[TMP0]], i16* [[DOTADDR]], align 2 ; CHECK-NEXT: store i16 [[TMP0]], i16* [[DOTADDR]], align 2
; CHECK-NEXT: store i32 [[TMP1]], i32* [[DOTADDR1]], align 4 ; CHECK-NEXT: store i32 [[TMP1]], i32* [[DOTADDR1]], align 4
; CHECK-NEXT: call void @__kmpc_get_shared_variables(i8*** [[GLOBAL_ARGS]]) ; CHECK-NEXT: call void @__kmpc_get_shared_variables(i8*** [[GLOBAL_ARGS]])
; CHECK-NEXT: call void @__omp_outlined__17(i32* [[DOTADDR1]], i32* [[DOTZERO_ADDR]]) #[[ATTR2]] ; CHECK-NEXT: call void @__omp_outlined__17(i32* [[DOTADDR1]], i32* [[DOTZERO_ADDR]]) #[[ATTR2]]
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__18 ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__18
; CHECK-SAME: (i32* noalias nofree [[DOTGLOBAL_TID_:%.*]], i32* noalias nofree [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i32* noalias nocapture nofree readnone [[DOTGLOBAL_TID_:%.*]], i32* noalias nocapture nofree readnone [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTGLOBAL_TID__ADDR]], align 8
; CHECK-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTBOUND_TID__ADDR]], align 8
; CHECK-NEXT: call void @p0() #[[ATTR8]] ; CHECK-NEXT: call void @p0() #[[ATTR8]]
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__18_wrapper ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__18_wrapper
; CHECK-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
Show All 28 Lines
; CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB2]]) ; CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB2]])
; CHECK-NEXT: [[TMP1:%.*]] = bitcast [0 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8** ; CHECK-NEXT: [[TMP1:%.*]] = bitcast [0 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
; CHECK-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP0]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*)* @__omp_outlined__19 to i8*), i8* bitcast (void (i16, i32)* @__omp_outlined__19_wrapper to i8*), i8** [[TMP1]], i64 0) ; CHECK-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP0]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*)* @__omp_outlined__19 to i8*), i8* bitcast (void (i16, i32)* @__omp_outlined__19_wrapper to i8*), i8** [[TMP1]], i64 0)
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__19 ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__19
; CHECK-SAME: (i32* noalias nofree [[DOTGLOBAL_TID_:%.*]], i32* noalias nofree [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i32* noalias nocapture nofree readnone [[DOTGLOBAL_TID_:%.*]], i32* noalias nocapture nofree readnone [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTGLOBAL_TID__ADDR]], align 8
; CHECK-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTBOUND_TID__ADDR]], align 8
; CHECK-NEXT: call void @p0() #[[ATTR8]] ; CHECK-NEXT: call void @p0() #[[ATTR8]]
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
; ;
; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK: Function Attrs: convergent norecurse nounwind
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__19_wrapper ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__19_wrapper
; CHECK-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
Show All 11 Lines

llvm/test/Transforms/OpenMP/spmdization.ll

Show First 20 Lines • Show All 76 Lines • ▼ Show 20 Lines
; CHECK-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 ; CHECK-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4
; CHECK-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 ; CHECK-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4
; CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i1 true, i1 false, i1 false) ; CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i1 true, i1 false, i1 false)
; CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1
; CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]]
; CHECK: user_code.entry: ; CHECK: user_code.entry:
; CHECK-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3:[0-9]+]] ; CHECK-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3:[0-9]+]]
; CHECK-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4 ; CHECK-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; CHECK-NEXT: call void @__omp_outlined__(i32* noundef nonnull align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noundef nonnull align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]] ; CHECK-NEXT: call void @__omp_outlined__(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false) ; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; CHECK: worker.exit: ; CHECK: worker.exit:
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
entry: entry:
%.zero.addr = alloca i32, align 4 %.zero.addr = alloca i32, align 4
%.threadid_temp. = alloca i32, align 4 %.threadid_temp. = alloca i32, align 4
Show All 12 Lines
worker.exit: ; preds = %entry worker.exit: ; preds = %entry
ret void ret void
} }
declare i32 @__kmpc_target_init(%struct.ident_t*, i1, i1, i1) declare i32 @__kmpc_target_init(%struct.ident_t*, i1, i1, i1)
define internal void @__omp_outlined__(i32* noalias %.global_tid., i32* noalias %.bound_tid.) #0 { define internal void @__omp_outlined__(i32* noalias %.global_tid., i32* noalias %.bound_tid.) #0 {
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__ ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__
; CHECK-SAME: (i32* noalias nofree noundef nonnull align 4 dereferenceable(4) [[DOTGLOBAL_TID_:%.*]], i32* noalias nofree noundef nonnull align 4 dereferenceable(4) [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i32* noalias nocapture nofree noundef nonnull readonly align 4 dereferenceable(4) [[DOTGLOBAL_TID_:%.*]], i32* noalias nocapture nofree nonnull readnone align 4 dereferenceable(4) [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[I:%.*]] = alloca i32, align 4 ; CHECK-NEXT: [[I:%.*]] = alloca i32, align 4
; CHECK-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [0 x i8*], align 8 ; CHECK-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [0 x i8*], align 8
; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8
; CHECK-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTBOUND_TID__ADDR]], align 8
; CHECK-NEXT: store i32 0, i32* [[I]], align 4 ; CHECK-NEXT: store i32 0, i32* [[I]], align 4
; CHECK-NEXT: br label [[FOR_COND:%.*]] ; CHECK-NEXT: br label [[FOR_COND:%.*]]
; CHECK: for.cond: ; CHECK: for.cond:
; CHECK-NEXT: [[TMP0:%.*]] = load i32, i32* [[I]], align 4 ; CHECK-NEXT: [[TMP0:%.*]] = load i32, i32* [[I]], align 4
; CHECK-NEXT: [[CMP:%.*]] = icmp slt i32 [[TMP0]], 100 ; CHECK-NEXT: [[CMP:%.*]] = icmp slt i32 [[TMP0]], 100
; CHECK-NEXT: br i1 [[CMP]], label [[FOR_BODY:%.*]], label [[FOR_END:%.*]] ; CHECK-NEXT: br i1 [[CMP]], label [[FOR_BODY:%.*]], label [[FOR_END:%.*]]
; CHECK: for.body: ; CHECK: for.body:
; CHECK-NEXT: [[TMP1:%.*]] = load i32, i32* [[DOTGLOBAL_TID_]], align 4 ; CHECK-NEXT: [[TMP1:%.*]] = load i32, i32* [[DOTGLOBAL_TID_]], align 4
Show All 39 Lines
for.end: ; preds = %for.cond for.end: ; preds = %for.cond
call void @spmd_amenable() #4 call void @spmd_amenable() #4
ret void ret void
} }
define internal void @__omp_outlined__1(i32* noalias %.global_tid., i32* noalias %.bound_tid.) #0 { define internal void @__omp_outlined__1(i32* noalias %.global_tid., i32* noalias %.bound_tid.) #0 {
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__1 ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__1
; CHECK-SAME: (i32* noalias nofree [[DOTGLOBAL_TID_:%.*]], i32* noalias nofree [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i32* noalias nocapture nofree readnone [[DOTGLOBAL_TID_:%.*]], i32* noalias nocapture nofree readnone [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTGLOBAL_TID__ADDR]], align 8
; CHECK-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTBOUND_TID__ADDR]], align 8
; CHECK-NEXT: call void @unknown() #[[ATTR5:[0-9]+]] ; CHECK-NEXT: call void @unknown() #[[ATTR5:[0-9]+]]
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
entry: entry:
%.global_tid..addr = alloca i32*, align 8 %.global_tid..addr = alloca i32*, align 8
%.bound_tid..addr = alloca i32*, align 8 %.bound_tid..addr = alloca i32*, align 8
store i32* %.global_tid., i32** %.global_tid..addr, align 8 store i32* %.global_tid., i32** %.global_tid..addr, align 8
store i32* %.bound_tid., i32** %.bound_tid..addr, align 8 store i32* %.bound_tid., i32** %.bound_tid..addr, align 8
▲ Show 20 Lines • Show All 49 Lines • ▼ Show 20 Lines
; CHECK-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 ; CHECK-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4
; CHECK-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 ; CHECK-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4
; CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i1 true, i1 false, i1 false) ; CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i1 true, i1 false, i1 false)
; CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1
; CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]]
; CHECK: user_code.entry: ; CHECK: user_code.entry:
; CHECK-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]] ; CHECK-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; CHECK-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4 ; CHECK-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; CHECK-NEXT: call void @__omp_outlined__2(i32* noundef nonnull align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noundef nonnull align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]] ; CHECK-NEXT: call void @__omp_outlined__2(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false) ; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; CHECK: worker.exit: ; CHECK: worker.exit:
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
entry: entry:
%.zero.addr = alloca i32, align 4 %.zero.addr = alloca i32, align 4
%.threadid_temp. = alloca i32, align 4 %.threadid_temp. = alloca i32, align 4
Show All 10 Linesuser_code.entry: ; preds = %entry
ret void ret void
worker.exit: ; preds = %entry worker.exit: ; preds = %entry
ret void ret void
} }
define internal void @__omp_outlined__2(i32* noalias %.global_tid., i32* noalias %.bound_tid.) #0 { define internal void @__omp_outlined__2(i32* noalias %.global_tid., i32* noalias %.bound_tid.) #0 {
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__2 ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__2
; CHECK-SAME: (i32* noalias nofree noundef nonnull align 4 dereferenceable(4) [[DOTGLOBAL_TID_:%.*]], i32* noalias nofree noundef nonnull align 4 dereferenceable(4) [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i32* noalias nocapture nofree noundef nonnull readonly align 4 dereferenceable(4) [[DOTGLOBAL_TID_:%.*]], i32* noalias nocapture nofree nonnull readnone align 4 dereferenceable(4) [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[I:%.*]] = alloca i32, align 4 ; CHECK-NEXT: [[I:%.*]] = alloca i32, align 4
; CHECK-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [0 x i8*], align 8 ; CHECK-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [0 x i8*], align 8
; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8
; CHECK-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTBOUND_TID__ADDR]], align 8
; CHECK-NEXT: [[TMP0:%.*]] = alloca i8, i64 4, align 1 ; CHECK-NEXT: [[TMP0:%.*]] = alloca i8, i64 4, align 1
; CHECK-NEXT: [[X_ON_STACK:%.*]] = bitcast i8* [[TMP0]] to i32* ; CHECK-NEXT: [[X_ON_STACK:%.*]] = bitcast i8* [[TMP0]] to i32*
; CHECK-NEXT: call void @use(i32* nocapture [[X_ON_STACK]]) #[[ATTR4]] ; CHECK-NEXT: call void @use(i32* nocapture [[X_ON_STACK]]) #[[ATTR4]]
; CHECK-NEXT: store i32 0, i32* [[I]], align 4 ; CHECK-NEXT: store i32 0, i32* [[I]], align 4
; CHECK-NEXT: br label [[FOR_COND:%.*]] ; CHECK-NEXT: br label [[FOR_COND:%.*]]
; CHECK: for.cond: ; CHECK: for.cond:
; CHECK-NEXT: [[TMP1:%.*]] = load i32, i32* [[I]], align 4 ; CHECK-NEXT: [[TMP1:%.*]] = load i32, i32* [[I]], align 4
; CHECK-NEXT: [[CMP:%.*]] = icmp slt i32 [[TMP1]], 100 ; CHECK-NEXT: [[CMP:%.*]] = icmp slt i32 [[TMP1]], 100
▲ Show 20 Lines • Show All 50 Lines • ▼ Show 20 Lines
} }
declare i8* @__kmpc_alloc_shared(i64) #3 declare i8* @__kmpc_alloc_shared(i64) #3
declare void @use(i32* nocapture) #2 declare void @use(i32* nocapture) #2
define internal void @__omp_outlined__3(i32* noalias %.global_tid., i32* noalias %.bound_tid.) #0 { define internal void @__omp_outlined__3(i32* noalias %.global_tid., i32* noalias %.bound_tid.) #0 {
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__3 ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__3
; CHECK-SAME: (i32* noalias nofree [[DOTGLOBAL_TID_:%.*]], i32* noalias nofree [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i32* noalias nocapture nofree readnone [[DOTGLOBAL_TID_:%.*]], i32* noalias nocapture nofree readnone [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTGLOBAL_TID__ADDR]], align 8
; CHECK-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTBOUND_TID__ADDR]], align 8
; CHECK-NEXT: call void @unknown() #[[ATTR5]] ; CHECK-NEXT: call void @unknown() #[[ATTR5]]
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
entry: entry:
%.global_tid..addr = alloca i32*, align 8 %.global_tid..addr = alloca i32*, align 8
%.bound_tid..addr = alloca i32*, align 8 %.bound_tid..addr = alloca i32*, align 8
store i32* %.global_tid., i32** %.global_tid..addr, align 8 store i32* %.global_tid., i32** %.global_tid..addr, align 8
store i32* %.bound_tid., i32** %.bound_tid..addr, align 8 store i32* %.bound_tid., i32** %.bound_tid..addr, align 8
Show All 39 Lines
; CHECK-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 ; CHECK-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4
; CHECK-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 ; CHECK-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4
; CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i1 true, i1 false, i1 false) ; CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i1 true, i1 false, i1 false)
; CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1
; CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]]
; CHECK: user_code.entry: ; CHECK: user_code.entry:
; CHECK-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]] ; CHECK-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; CHECK-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4 ; CHECK-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; CHECK-NEXT: call void @__omp_outlined__4(i32* noundef nonnull align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noundef nonnull align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]] ; CHECK-NEXT: call void @__omp_outlined__4(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false) ; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; CHECK: worker.exit: ; CHECK: worker.exit:
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
entry: entry:
%.zero.addr = alloca i32, align 4 %.zero.addr = alloca i32, align 4
%.threadid_temp. = alloca i32, align 4 %.threadid_temp. = alloca i32, align 4
Show All 10 Linesuser_code.entry: ; preds = %entry
ret void ret void
worker.exit: ; preds = %entry worker.exit: ; preds = %entry
ret void ret void
} }
define internal void @__omp_outlined__4(i32* noalias %.global_tid., i32* noalias %.bound_tid.) #0 { define internal void @__omp_outlined__4(i32* noalias %.global_tid., i32* noalias %.bound_tid.) #0 {
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__4 ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__4
; CHECK-SAME: (i32* noalias nofree noundef nonnull align 4 dereferenceable(4) [[DOTGLOBAL_TID_:%.*]], i32* noalias nofree noundef nonnull align 4 dereferenceable(4) [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i32* noalias nocapture nofree noundef nonnull readonly align 4 dereferenceable(4) [[DOTGLOBAL_TID_:%.*]], i32* noalias nocapture nofree nonnull readnone align 4 dereferenceable(4) [[DOTBOUND_TID_:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[I:%.*]] = alloca i32, align 4 ; CHECK-NEXT: [[I:%.*]] = alloca i32, align 4
; CHECK-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 8 ; CHECK-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 8
; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8
; CHECK-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTBOUND_TID__ADDR]], align 8
; CHECK-NEXT: store i32 0, i32* [[I]], align 4 ; CHECK-NEXT: store i32 0, i32* [[I]], align 4
; CHECK-NEXT: br label [[FOR_COND:%.*]] ; CHECK-NEXT: br label [[FOR_COND:%.*]]
; CHECK: for.cond: ; CHECK: for.cond:
; CHECK-NEXT: [[TMP0:%.*]] = load i32, i32* [[I]], align 4 ; CHECK-NEXT: [[TMP0:%.*]] = load i32, i32* [[I]], align 4
; CHECK-NEXT: [[CMP:%.*]] = icmp slt i32 [[TMP0]], 100 ; CHECK-NEXT: [[CMP:%.*]] = icmp slt i32 [[TMP0]], 100
; CHECK-NEXT: br i1 [[CMP]], label [[FOR_BODY:%.*]], label [[FOR_END:%.*]] ; CHECK-NEXT: br i1 [[CMP]], label [[FOR_BODY:%.*]], label [[FOR_END:%.*]]
; CHECK: for.body: ; CHECK: for.body:
; CHECK-NEXT: [[TMP1:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[CAPTURED_VARS_ADDRS]], i64 0, i64 0 ; CHECK-NEXT: [[TMP1:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[CAPTURED_VARS_ADDRS]], i64 0, i64 0
▲ Show 20 Lines • Show All 47 Lines • ▼ Show 20 Lines
for.end: ; preds = %for.cond for.end: ; preds = %for.cond
call void @spmd_amenable() #4 call void @spmd_amenable() #4
call void @__kmpc_free_shared(i8* %x, i64 4) call void @__kmpc_free_shared(i8* %x, i64 4)
ret void ret void
} }
define internal void @__omp_outlined__5(i32* noalias %.global_tid., i32* noalias %.bound_tid., i32* nonnull align 4 dereferenceable(4) %x) #0 { define internal void @__omp_outlined__5(i32* noalias %.global_tid., i32* noalias %.bound_tid., i32* nonnull align 4 dereferenceable(4) %x) #0 {
; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__5 ; CHECK-LABEL: define {{[^@]+}}@__omp_outlined__5
; CHECK-SAME: (i32* noalias nofree [[DOTGLOBAL_TID_:%.*]], i32* noalias nofree [[DOTBOUND_TID_:%.*]], i32* nofree nonnull align 4 dereferenceable(4) [[X:%.*]]) #[[ATTR0]] { ; CHECK-SAME: (i32* noalias nocapture nofree readnone [[DOTGLOBAL_TID_:%.*]], i32* noalias nocapture nofree readnone [[DOTBOUND_TID_:%.*]], i32* nocapture nofree nonnull align 4 dereferenceable(4) [[X:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry: ; CHECK-NEXT: entry:
; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: [[X_ADDR:%.*]] = alloca i32*, align 8 ; CHECK-NEXT: [[X_ADDR:%.*]] = alloca i32*, align 8
; CHECK-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTGLOBAL_TID__ADDR]], align 8
; CHECK-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 ; CHECK-NEXT: store i32* undef, i32** [[DOTBOUND_TID__ADDR]], align 8
; CHECK-NEXT: store i32* [[X]], i32** [[X_ADDR]], align 8 ; CHECK-NEXT: store i32* [[X]], i32** [[X_ADDR]], align 8
; CHECK-NEXT: [[TMP0:%.*]] = load i32, i32* [[X]], align 4 ; CHECK-NEXT: [[TMP0:%.*]] = load i32, i32* [[X]], align 4
; CHECK-NEXT: [[INC:%.*]] = add nsw i32 [[TMP0]], 1 ; CHECK-NEXT: [[INC:%.*]] = add nsw i32 [[TMP0]], 1
; CHECK-NEXT: store i32 [[INC]], i32* [[X]], align 4 ; CHECK-NEXT: store i32 [[INC]], i32* [[X]], align 4
; CHECK-NEXT: call void @unknown() #[[ATTR5]] ; CHECK-NEXT: call void @unknown() #[[ATTR5]]
; CHECK-NEXT: ret void ; CHECK-NEXT: ret void
; ;
entry: entry:
▲ Show 20 Lines • Show All 96 Lines • Show Last 20 Lines