[llvm-branch-commits] [llvm] 6c43564 - [Coroutine] Improve coro-elide-musttail.ll test
Author: Xun Li
Date: 2021-01-22T20:23:48-08:00
New Revision: 6c43564530365ac2c074d7515d4eada294d4ca0c
URL:
https://github.com/llvm/llvm-project/commit/6c43564530365ac2c074d7515d4eada294d4ca0c
DIFF:
https://github.com/llvm/llvm-project/commit/6c43564530365ac2c074d7515d4eada294d4ca0c.diff
LOG: [Coroutine] Improve coro-elide-musttail.ll test
The test wasn't sensitive to alias analysis. As you can seen from D95117 when
AA is added by default this is affected.
Updating the test so that it coveres both cases for AA analysis.
Note that this patch depends on D95117 to land first.
Differential Revision: https://reviews.llvm.org/D95247
Added:
Modified:
llvm/test/Transforms/Coroutines/coro-elide-musttail.ll
Removed:
diff --git a/llvm/test/Transforms/Coroutines/coro-elide-musttail.ll
b/llvm/test/Transforms/Coroutines/coro-elide-musttail.ll
index 26f9b5826920..2920bacb5635 100644
--- a/llvm/test/Transforms/Coroutines/coro-elide-musttail.ll
+++ b/llvm/test/Transforms/Coroutines/coro-elide-musttail.ll
@@ -2,7 +2,10 @@
; Only run with new pass manager since old pass manager's alias analysis isn't
; powerful enough to tell that the tailcall's arguments don't alias the frame.
;
+; RUN: opt < %s -coro-elide -S | FileCheck %s
+; RUN: opt < %s -disable-basic-aa -coro-elide -S | FileCheck %s
-check-prefix=NOAA
; RUN: opt < %s -passes='coro-elide' -S | FileCheck %s
+; RUN: opt < %s -aa-pipeline= -passes='coro-elide' -S | FileCheck %s
-check-prefix=NOAA
%"bar.Frame" = type { void (%"bar.Frame"*)*, void (%"bar.Frame"*)*,
%"struct.coroutine::promise_type", i1 }
%"struct.coroutine::promise_type" = type { i32 }
@@ -14,6 +17,7 @@
declare dso_local void @"bar"() align 2
declare dso_local fastcc void @"bar.resume"(%"bar.Frame"*) align 2
+; There is a musttail call. CoroElide won't happen.
define internal fastcc void @foo.resume_musttail(%"foo.Frame"* %FramePtr) {
; CHECK-LABEL: @foo.resume_musttail(
; CHECK-NEXT: entry:
@@ -34,8 +38,32 @@ entry:
ret void
}
-define internal fastcc void @foo.resume_no_musttail(%"foo.Frame"* %FramePtr) {
-; CHECK-LABEL: @foo.resume_no_musttail(
+; The new frame (TMP0) could potentially alias CALL34, the tailcall attribute
on that call must be removed
+define internal fastcc void @foo.resume_no_musttail_with_alias(%"foo.Frame"*
%FramePtr) {
+; CHECK-LABEL: @foo.resume_no_musttail_with_alias(
+; CHECK-NEXT: entry:
+; CHECK-NEXT:[[TMP0:%.*]] = alloca [24 x i8], align 8
+; CHECK-NEXT:[[VFRAME:%.*]] = bitcast [24 x i8]* [[TMP0]] to i8*
+; CHECK-NEXT:[[TMP1:%.*]] = tail call token @llvm.coro.id(i32 16, i8*
null, i8* bitcast (void ()* @bar to i8*), i8* bitcast ([3 x void
(%bar.Frame*)*]* @bar.resumers to i8*))
+; CHECK-NEXT:call fastcc void undef(i8* [[VFRAME]])
+; CHECK-NEXT:[[CALL34:%.*]] = call i8* undef()
+; CHECK-NEXT:call fastcc void undef(i8* [[CALL34]])
+; CHECK-NEXT:ret void
+;
+entry:
+ %0 = tail call token @llvm.coro.id(i32 16, i8* null, i8* bitcast (void ()*
@"bar" to i8*), i8* bitcast ([3 x void (%"bar.Frame"*)*]* @"bar.resumers" to
i8*))
+ %1 = tail call i1 @llvm.coro.alloc(token %0)
+ %2 = tail call i8* @llvm.coro.begin(token %0, i8* null)
+ call i8* @llvm.coro.subfn.addr(i8* %2, i8 1)
+ call fastcc void undef(i8* %2)
+ %call34 = call i8* undef()
+ tail call fastcc void undef(i8* %call34)
+ ret void
+}
+
+; The new frame (TMP0) does not alias CALL34, tailcall attribute can reimain.
This analysis is only available when alias analysis is enabled.
+define internal fastcc void @foo.resume_no_musttail_no_alias(%"foo.Frame"*
%FramePtr) {
+; CHECK-LABEL: @foo.resume_no_musttail_no_alias(
; CHECK-NEXT: entry:
; CHECK-NEXT:[[TMP0:%.*]] = alloca [24 x i8], align 8
; CHECK-NEXT:[[VFRAME:%.*]] = bitcast [24 x i8]* [[TMP0]] to i8*
@@ -44,6 +72,15 @@ define internal fastcc void
@foo.resume_no_musttail(%"foo.Frame"* %FramePtr) {
; CHECK-NEXT:tail call fastcc void undef(i8* [[CALL34]])
; CHECK-NEXT:ret void
;
+; NOAA-LABEL: @foo.resume_no_musttail_no_alias(
+; NOAA-NEXT: entry:
+; NOAA-NEXT:[[TMP0:%.*]] = alloca [24 x i8], align 8
+; NOAA-NEXT:[[VFRAME:%.*]] = bitcast [24 x i8]* [[TMP0]] to i8*
+; NOAA-NEXT:[[TMP1:%.*]] = call token @llvm.coro.id(i32 16, i8* null, i8*
bitcast (void ()* @bar to i8*), i8* bitcast ([3 x void (%bar.Frame*)*]*
@bar.resumers to i8*))
+; NOAA-NEXT:[[CALL34:%.*]] = call i8* undef()
+; NOAA-NEXT:call fastcc void undef(i8* [[CALL34]])
+; NOAA-NEXT:ret void
+;
entry:
%0 = tail call token @llvm.coro.id(i32 16, i8* null, i8* bitcast (void ()*
@"bar" to i8*), i8* bitcast ([3 x void (%"bar.Frame"*)*]* @"bar.resumers" to
i8*))
%1 = tail call i1 @llvm.coro.alloc(token %0)
@@ -54,6 +91,7 @@ entry:
ret void
}
+
; Function Attrs: argmemonly nofree nosync nounwind willreturn
declare void @llvm.lifetime.start.p0i8(i64 immarg, i8* nocapture) #0
[llvm-branch-commits] [llvm] 17c3538 - Revert "Fix unused variable in CoroFrame.cpp when building Release with GCC 10"
Author: Xun Li
Date: 2021-01-25T08:37:45-08:00
New Revision: 17c3538aef656178b342573043eff328f5cf2673
URL:
https://github.com/llvm/llvm-project/commit/17c3538aef656178b342573043eff328f5cf2673
DIFF:
https://github.com/llvm/llvm-project/commit/17c3538aef656178b342573043eff328f5cf2673.diff
LOG: Revert "Fix unused variable in CoroFrame.cpp when building Release with
GCC 10"
This reverts commit ff5e896425577f445ed080d88b582aab0896fba0.
Added:
Modified:
llvm/lib/Transforms/Coroutines/CoroFrame.cpp
Removed:
diff --git a/llvm/lib/Transforms/Coroutines/CoroFrame.cpp
b/llvm/lib/Transforms/Coroutines/CoroFrame.cpp
index 3f290f2f087b..cd5c1863bb3c 100644
--- a/llvm/lib/Transforms/Coroutines/CoroFrame.cpp
+++ b/llvm/lib/Transforms/Coroutines/CoroFrame.cpp
@@ -1549,7 +1549,6 @@ static void rewritePHIs(BasicBlock &BB) {
for (BasicBlock *Pred : Preds) {
if (CatchSwitchInst *CS =
dyn_cast(Pred->getTerminator())) {
-(void)CS;
// CleanupPad with a CatchSwitch predecessor: therefore this is an
// unwind destination that needs to be handle specially.
assert(CS->getUnwindDest() == &BB);
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [libc] af2439f - [libc][Obvious] Add header guards for the generated linux syscall header file.
Author: Siva Chandra Reddy
Date: 2021-08-30T17:41:45-07:00
New Revision: af2439f559cb6fcb8db797e58adde6b9ec0457cb
URL:
https://github.com/llvm/llvm-project/commit/af2439f559cb6fcb8db797e58adde6b9ec0457cb
DIFF:
https://github.com/llvm/llvm-project/commit/af2439f559cb6fcb8db797e58adde6b9ec0457cb.diff
LOG: [libc][Obvious] Add header guards for the generated linux syscall header
file.
Added:
Modified:
libc/config/linux/syscall.h.def
Removed:
diff --git a/libc/config/linux/syscall.h.def b/libc/config/linux/syscall.h.def
index 7f52a26ff0e7b..0daceaf00dded 100644
--- a/libc/config/linux/syscall.h.def
+++ b/libc/config/linux/syscall.h.def
@@ -6,4 +6,9 @@
//
//===--===//
+#ifndef LLVM_LIBC_CONFIG_LINUX_SYSCALL_H
+#define LLVM_LIBC_CONFIG_LINUX_SYSCALL_H
+
%%include_file(${inline_syscalls})
+
+#endif // LLVM_LIBC_CONFIG_LINUX_SYSCALL_H
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [llvm] ff5e896 - Fix unused variable in CoroFrame.cpp when building Release with GCC 10
Author: Daniel Paoliello
Date: 2021-01-13T22:53:25-08:00
New Revision: ff5e896425577f445ed080d88b582aab0896fba0
URL:
https://github.com/llvm/llvm-project/commit/ff5e896425577f445ed080d88b582aab0896fba0
DIFF:
https://github.com/llvm/llvm-project/commit/ff5e896425577f445ed080d88b582aab0896fba0.diff
LOG: Fix unused variable in CoroFrame.cpp when building Release with GCC 10
When building with GCC 10, the following warning is reported:
```
/llvm-project/llvm/lib/Transforms/Coroutines/CoroFrame.cpp:1527:28: warning:
unused variable ‘CS’ [-Wunused-variable]
1527 | if (CatchSwitchInst *CS =
```
This change adds a cast to `void` to avoid the warning.
Reviewed By: lxfind
Differential Revision: https://reviews.llvm.org/D94456
Added:
Modified:
llvm/lib/Transforms/Coroutines/CoroFrame.cpp
Removed:
diff --git a/llvm/lib/Transforms/Coroutines/CoroFrame.cpp
b/llvm/lib/Transforms/Coroutines/CoroFrame.cpp
index 84b78fce3f44..16e7e2c2ceaf 100644
--- a/llvm/lib/Transforms/Coroutines/CoroFrame.cpp
+++ b/llvm/lib/Transforms/Coroutines/CoroFrame.cpp
@@ -1535,6 +1535,7 @@ static void rewritePHIs(BasicBlock &BB) {
for (BasicBlock *Pred : Preds) {
if (CatchSwitchInst *CS =
dyn_cast(Pred->getTerminator())) {
+(void)CS;
// CleanupPad with a CatchSwitch predecessor: therefore this is an
// unwind destination that needs to be handle specially.
assert(CS->getUnwindDest() == &BB);
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [llvm] 1d04dc5 - [Coroutine] Do not CoroElide if there are musttail calls
Author: Xun Li
Date: 2021-01-18T09:06:21-08:00
New Revision: 1d04dc52dd24d791970e56053cdd67fe149b0554
URL:
https://github.com/llvm/llvm-project/commit/1d04dc52dd24d791970e56053cdd67fe149b0554
DIFF:
https://github.com/llvm/llvm-project/commit/1d04dc52dd24d791970e56053cdd67fe149b0554.diff
LOG: [Coroutine] Do not CoroElide if there are musttail calls
This is to address https://bugs.llvm.org/show_bug.cgi?id=48626.
When there are musttail calls that use parameters aliasing the newly created
coroutine frame, the existing implementation will fatal.
We simply cannot perform CoroElide in such cases. In theory a precise analysis
can be done to check whether the parameters of the musttail call
actually alias the frame, but it's very hard to do it before the transformation
happens. Also in most cases the existence of musttail call is
generated due to symmetric transfers, and in those cases alias analysis won't
be able to tell that they don't alias anyway.
Differential Revision: https://reviews.llvm.org/D94834
Added:
llvm/test/Transforms/Coroutines/coro-elide-musttail.ll
Modified:
llvm/lib/Transforms/Coroutines/CoroElide.cpp
Removed:
diff --git a/llvm/lib/Transforms/Coroutines/CoroElide.cpp
b/llvm/lib/Transforms/Coroutines/CoroElide.cpp
index a0b8168f8717..07a183cfc66b 100644
--- a/llvm/lib/Transforms/Coroutines/CoroElide.cpp
+++ b/llvm/lib/Transforms/Coroutines/CoroElide.cpp
@@ -83,14 +83,8 @@ static void removeTailCallAttribute(AllocaInst *Frame,
AAResults &AA) {
Function &F = *Frame->getFunction();
for (Instruction &I : instructions(F))
if (auto *Call = dyn_cast(&I))
- if (Call->isTailCall() && operandReferences(Call, Frame, AA)) {
-// FIXME: If we ever hit this check. Evaluate whether it is more
-// appropriate to retain musttail and allow the code to compile.
-if (Call->isMustTailCall())
- report_fatal_error("Call referring to the coroutine frame cannot be "
- "marked as musttail");
+ if (Call->isTailCall() && operandReferences(Call, Frame, AA))
Call->setTailCall(false);
- }
}
// Given a resume function @f.resume(%f.frame* %frame), returns the size
@@ -252,7 +246,20 @@ bool Lowerer::shouldElide(Function *F, DominatorTree &DT)
const {
// If size of the set is the same as total number of coro.begin, that means
we
// found a coro.free or coro.destroy referencing each coro.begin, so we can
// perform heap elision.
- return ReferencedCoroBegins.size() == CoroBegins.size();
+ if (ReferencedCoroBegins.size() != CoroBegins.size())
+return false;
+
+ // If any call in the function is a musttail call, it usually won't work
+ // because we cannot drop the tailcall attribute, and a tail call will reuse
+ // the entire stack where we are going to put the new frame. In theory a more
+ // precise analysis can be done to check whether the new frame aliases with
+ // the call, however it's challenging to do so before the elision actually
+ // happened.
+ for (BasicBlock &BB : *F)
+if (BB.getTerminatingMustTailCall())
+ return false;
+
+ return true;
}
void Lowerer::collectPostSplitCoroIds(Function *F) {
diff --git a/llvm/test/Transforms/Coroutines/coro-elide-musttail.ll
b/llvm/test/Transforms/Coroutines/coro-elide-musttail.ll
new file mode 100644
index ..751190413b03
--- /dev/null
+++ b/llvm/test/Transforms/Coroutines/coro-elide-musttail.ll
@@ -0,0 +1,74 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; Only run with new pass manager since old pass manager's alias analysis isn't
+; powerful enough to tell that the tailcall's arguments don't alias the frame.
+;
+; RUN: opt < %s -passes='coro-elide' -S | FileCheck %s
+
+%"bar.Frame" = type { void (%"bar.Frame"*)*, void (%"bar.Frame"*)*,
%"struct.coroutine::promise_type", i1 }
+%"struct.coroutine::promise_type" = type { i32 }
+%"foo.Frame" = type { void (%"foo.Frame"*)*, void (%"foo.Frame"*)*,
%"struct.coroutine::promise_type", i1, %"struct.coroutine::promise_type::final_awaitable" }
+%"struct.coroutine::promise_type" = type { i32 }
+%"struct.coroutine::promise_type::final_awaitable" = type { i8 }
+@"bar.resumers" = private constant [3 x void (%"bar.Frame"*)*] [void
(%"bar.Frame"*)* @"bar.resume", void (%"bar.Frame"*)* undef, void
(%"bar.Frame"*)* undef]
+
+declare dso_local void @"bar"() align 2
+declare dso_local fastcc void @"bar.resume"(%"bar.Frame"*) align 2
+
+define internal fastcc void @foo.resume_musttail(%"foo.Frame"* %FramePtr) {
+; CHECK-LABEL: @foo.resume_musttail(
+; CHECK-NEXT: entry:
+; CHECK-NEXT:[[TMP0:%.*]] = tail call token @llvm.coro.id
+; CHECK-NEXT:[[TMP1:%.*]] = tail call i1 @llvm.coro.alloc(token [[TMP0]])
+; CHECK-NEXT:[[TMP2:%.*]] = tail call i8* @llvm.coro.begin(token [[TMP0]],
i8* null)
+; CHECK-NEXT:[[CALL34:%.*]] = call i8* undef()
+; CH
[llvm-branch-commits] [llvm] bd3ca66 - [Inlining] Delete redundant optnone/alwaysinline check
Author: Xun Li
Date: 2021-01-21T18:38:10-08:00
New Revision: bd3cad14464b1bb7eecbd3cc227ee0614799
URL:
https://github.com/llvm/llvm-project/commit/bd3cad14464b1bb7eecbd3cc227ee0614799
DIFF:
https://github.com/llvm/llvm-project/commit/bd3cad14464b1bb7eecbd3cc227ee0614799.diff
LOG: [Inlining] Delete redundant optnone/alwaysinline check
The same check is done in InlineCost:
https://github.com/llvm/llvm-project/blob/8b0bd54d0ec968df28ccc58bbb537a7b7c074ef2/llvm/lib/Analysis/InlineCost.cpp#L2537-L2552
Also, doing a check on the callee here is confusing, because anything that
deals with callee should be done in the inner loop where we proecss all calls
from the same caller.
Differential Revision: https://reviews.llvm.org/D95186
Added:
Modified:
llvm/lib/Transforms/IPO/Inliner.cpp
Removed:
diff --git a/llvm/lib/Transforms/IPO/Inliner.cpp
b/llvm/lib/Transforms/IPO/Inliner.cpp
index 574d870bccd8..a7d7594c00b3 100644
--- a/llvm/lib/Transforms/IPO/Inliner.cpp
+++ b/llvm/lib/Transforms/IPO/Inliner.cpp
@@ -761,12 +761,6 @@ PreservedAnalyses InlinerPass::run(LazyCallGraph::SCC
&InitialC,
LazyCallGraph::Node &N = *CG.lookup(F);
if (CG.lookupSCC(N) != C)
continue;
-if (!Calls[I].first->getCalledFunction()->hasFnAttribute(
-Attribute::AlwaysInline) &&
-F.hasOptNone()) {
- setInlineRemark(*Calls[I].first, "optnone attribute");
- continue;
-}
LLVM_DEBUG(dbgs() << "Inlining calls in: " << F.getName() << "\n");
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [llvm] 3e2b424 - Remove RefSCC::handleTrivialEdgeInsertion
Author: Xun Li
Date: 2021-01-04T20:21:01-08:00
New Revision: 3e2b42489f897ededae1d3269dcaf084da692111
URL:
https://github.com/llvm/llvm-project/commit/3e2b42489f897ededae1d3269dcaf084da692111
DIFF:
https://github.com/llvm/llvm-project/commit/3e2b42489f897ededae1d3269dcaf084da692111.diff
LOG: Remove RefSCC::handleTrivialEdgeInsertion
This function no longer does anything useful. It probably did something
originally but latter changes removed them and didn't clean up this function.
The checks are already done in the callers as well.
Differential Revision: https://reviews.llvm.org/D94055
Added:
Modified:
llvm/include/llvm/Analysis/LazyCallGraph.h
llvm/lib/Analysis/LazyCallGraph.cpp
Removed:
diff --git a/llvm/include/llvm/Analysis/LazyCallGraph.h
b/llvm/include/llvm/Analysis/LazyCallGraph.h
index 7478e1726366..e92134d074e5 100644
--- a/llvm/include/llvm/Analysis/LazyCallGraph.h
+++ b/llvm/include/llvm/Analysis/LazyCallGraph.h
@@ -598,10 +598,6 @@ class LazyCallGraph {
void verify();
#endif
-/// Handle any necessary parent set updates after inserting a trivial ref
-/// or call edge.
-void handleTrivialEdgeInsertion(Node &SourceN, Node &TargetN);
-
public:
using iterator = pointee_iterator::const_iterator>;
using range = iterator_range;
diff --git a/llvm/lib/Analysis/LazyCallGraph.cpp
b/llvm/lib/Analysis/LazyCallGraph.cpp
index 26529891bc03..ab6946ee96be 100644
--- a/llvm/lib/Analysis/LazyCallGraph.cpp
+++ b/llvm/lib/Analysis/LazyCallGraph.cpp
@@ -1373,23 +1373,6 @@ LazyCallGraph::RefSCC::removeInternalRefEdge(Node
&SourceN,
return Result;
}
-void LazyCallGraph::RefSCC::handleTrivialEdgeInsertion(Node &SourceN,
- Node &TargetN) {
- // The only trivial case that requires any graph updates is when we add new
- // ref edge and may connect
diff erent RefSCCs along that path. This is only
- // because of the parents set. Every other part of the graph remains constant
- // after this edge insertion.
- assert(G->lookupRefSCC(SourceN) == this && "Source must be in this RefSCC.");
- RefSCC &TargetRC = *G->lookupRefSCC(TargetN);
- if (&TargetRC == this)
-return;
-
-#ifdef EXPENSIVE_CHECKS
- assert(TargetRC.isDescendantOf(*this) &&
- "Target must be a descendant of the Source.");
-#endif
-}
-
void LazyCallGraph::RefSCC::insertTrivialCallEdge(Node &SourceN,
Node &TargetN) {
#ifndef NDEBUG
@@ -1420,9 +1403,6 @@ void LazyCallGraph::RefSCC::insertTrivialCallEdge(Node
&SourceN,
// Create the new edge.
SourceN->Edges.emplace_back(TargetN, Edge::Call);
}
-
- // Now that we have the edge, handle the graph fallout.
- handleTrivialEdgeInsertion(SourceN, TargetN);
}
void LazyCallGraph::RefSCC::insertTrivialRefEdge(Node &SourceN, Node &TargetN)
{
@@ -1449,9 +1429,6 @@ void LazyCallGraph::RefSCC::insertTrivialRefEdge(Node
&SourceN, Node &TargetN) {
// Create the new edge.
SourceN->Edges.emplace_back(TargetN, Edge::Ref);
-
- // Now that we have the edge, handle the graph fallout.
- handleTrivialEdgeInsertion(SourceN, TargetN);
}
void LazyCallGraph::RefSCC::replaceNodeFunction(Node &N, Function &NewF) {
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [llvm] 4652718 - Cleanup coro-inline.ll
Author: Xun Li
Date: 2020-12-18T08:05:04-08:00
New Revision: 4652718ee38c519c6bb9470758d07430b0de02dd
URL:
https://github.com/llvm/llvm-project/commit/4652718ee38c519c6bb9470758d07430b0de02dd
DIFF:
https://github.com/llvm/llvm-project/commit/4652718ee38c519c6bb9470758d07430b0de02dd.diff
LOG: Cleanup coro-inline.ll
Following up with the comments in D92706.
- Use -passes instead of -enable-new-pm
- CoroEarly should happen before AlwaysInliner, adjust it.
- Remove some unnecessary barriers (still kept one)
- Cleanup unnecessary debug info
Differential Revision: https://reviews.llvm.org/D93342
Added:
Modified:
llvm/test/Transforms/Coroutines/coro-inline.ll
Removed:
diff --git a/llvm/test/Transforms/Coroutines/coro-inline.ll
b/llvm/test/Transforms/Coroutines/coro-inline.ll
index fc5c1d0d6b1f..2cc679a0127d 100644
--- a/llvm/test/Transforms/Coroutines/coro-inline.ll
+++ b/llvm/test/Transforms/Coroutines/coro-inline.ll
@@ -1,7 +1,7 @@
-; RUN: opt < %s -always-inline -barrier -coro-early -barrier -coro-split -S |
FileCheck %s
-; RUN: opt < %s -enable-new-pm -always-inline -coro-early -coro-split -S |
FileCheck %s
-; RUN: opt < %s -sample-profile-file=%S/Inputs/sample.text.prof
-pgo-kind=pgo-sample-use-pipeline -coro-early -barrier -sample-profile -barrier
-coro-split -disable-inlining=true -S | FileCheck %s
-; RUN: opt < %s -enable-new-pm -sample-profile-file=%S/Inputs/sample.text.prof
-pgo-kind=pgo-sample-use-pipeline -coro-early -sample-profile -coro-split
-disable-inlining=true -S | FileCheck %s
+; RUN: opt < %s -always-inline -barrier -coro-split -S | FileCheck %s
+; RUN: opt < %s -passes='always-inline,cgscc(coro-split)' -S | FileCheck %s
+; RUN: opt < %s -sample-profile-file=%S/Inputs/sample.text.prof
-pgo-kind=pgo-sample-use-pipeline -sample-profile -coro-split
-disable-inlining=true -S | FileCheck %s
+; RUN: opt < %s -sample-profile-file=%S/Inputs/sample.text.prof
-pgo-kind=pgo-sample-use-pipeline -passes='sample-profile,cgscc(coro-split)'
-disable-inlining=true -S | FileCheck %s
; Function Attrs: alwaysinline ssp uwtable
define void @ff() #0 !dbg !12 {
@@ -11,7 +11,6 @@ entry:
ret void
}
-; CHECK: call void @ff()
; Function Attrs: alwaysinline ssp uwtable
define void @foo() #0 !dbg !8 {
entry:
@@ -20,11 +19,14 @@ entry:
call void @ff(), !dbg !11
ret void
}
+; CHECK-LABEL: define void @foo()
+; CHECK: call void @ff()
+
declare token @llvm.coro.id(i32, i8* readnone, i8* nocapture readonly, i8*)
declare i8* @llvm.coro.begin(token, i8* writeonly)
-attributes #0 = { alwaysinline ssp uwtable "coroutine.presplit"="1"
"correctly-rounded-divide-sqrt-fp-math"="false" "disable-tail-calls"="false"
"frame-pointer"="all" "less-precise-fpmad"="false" "min-legal-vector-width"="0"
"no-infs-fp-math"="false" "no-jump-tables"="false" "no-nans-fp-math"="false"
"no-signed-zeros-fp-math"="false" "no-trapping-math"="true"
"stack-protector-buffer-size"="8" "target-cpu"="penryn"
"target-features"="+cx16,+cx8,+fxsr,+mmx,+sahf,+sse,+sse2,+sse3,+sse4.1,+ssse3,+x87"
"unsafe-fp-math"="false" "use-sample-profile" "use-soft-float"="false" }
+attributes #0 = { alwaysinline ssp uwtable "coroutine.presplit"="1"
"use-sample-profile" }
!llvm.dbg.cu = !{!0}
!llvm.module.flags = !{!3, !4, !5, !6}
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [llvm] 80b0f74 - Small improvements to Intrinsic::getName
Author: Xun Li
Date: 2020-12-02T16:49:12-08:00
New Revision: 80b0f74c8c539b2385f897467aa99b2b6b298c04
URL:
https://github.com/llvm/llvm-project/commit/80b0f74c8c539b2385f897467aa99b2b6b298c04
DIFF:
https://github.com/llvm/llvm-project/commit/80b0f74c8c539b2385f897467aa99b2b6b298c04.diff
LOG: Small improvements to Intrinsic::getName
While I was adding a new intrinsic instruction (not overloaded), I accidentally
used CreateUnaryIntrinsic to create the intrinsics, which turns out to be
passing the type list to getName, and ended up naming the intrinsics function
with type suffix, which leads to wierd bugs latter on. It took me a long time
to debug.
It seems a good idea to add an assertion in getName so that it fails if types
are passed but it's not a overloaded function.
Also, the overloade version of getName is less efficient because it creates an
std::string. We should avoid calling it if we know that there are no types
provided.
Differential Revision: https://reviews.llvm.org/D92523
Added:
Modified:
llvm/lib/IR/Function.cpp
Removed:
diff --git a/llvm/lib/IR/Function.cpp b/llvm/lib/IR/Function.cpp
index a2857707..37fa3d272781 100644
--- a/llvm/lib/IR/Function.cpp
+++ b/llvm/lib/IR/Function.cpp
@@ -781,6 +781,8 @@ StringRef Intrinsic::getName(ID id) {
std::string Intrinsic::getName(ID id, ArrayRef Tys) {
assert(id < num_intrinsics && "Invalid intrinsic ID!");
+ assert((Tys.empty() || Intrinsic::isOverloaded(id)) &&
+ "This version of getName is for overloaded intrinsics only");
std::string Result(IntrinsicNameTable[id]);
for (Type *Ty : Tys) {
Result += "." + getMangledTypeStr(Ty);
@@ -1243,7 +1245,7 @@ Function *Intrinsic::getDeclaration(Module *M, ID id,
ArrayRef Tys) {
// There can never be multiple globals with the same name of
diff erent types,
// because intrinsics must be a specific type.
return cast(
- M->getOrInsertFunction(getName(id, Tys),
+ M->getOrInsertFunction(Tys.empty() ? getName(id) : getName(id, Tys),
getType(M->getContext(), id, Tys))
.getCallee());
}
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [llvm] 31e60b9 - [coroutine] should disable inline before calling coro split
Author: Xun Li
Date: 2020-12-08T08:53:08-08:00
New Revision: 31e60b9133596c185e6daac7f1761dc17e464826
URL:
https://github.com/llvm/llvm-project/commit/31e60b9133596c185e6daac7f1761dc17e464826
DIFF:
https://github.com/llvm/llvm-project/commit/31e60b9133596c185e6daac7f1761dc17e464826.diff
LOG: [coroutine] should disable inline before calling coro split
This is a rework of D85812, which didn't land.
When callee coroutine function is inlined into caller coroutine function before
coro-split pass, llvm will emits "coroutine should have exactly one defining
@llvm.coro.begin". It seems that coro-early pass can not handle this quiet well.
So we believe that unsplited coroutine function should not be inlined.
This patch fix such issue by not inlining function if it has attribute
"coroutine.presplit" (it means the function has not been splited) to fix this
issue
test plan: check-llvm, check-clang
In D85812, there was suggestions on moving the macros to Attributes.td to avoid
circular header dependency issue.
I believe it's not worth doing just to be able to use one constant string in
one place.
Today, there are already 3 possible attribute values for "coroutine.presplit":
https://github.com/llvm/llvm-project/blob/c6543cc6b8f107b58e7205d8fc64865a508bacba/llvm/lib/Transforms/Coroutines/CoroInternal.h#L40-L42
If we move them into Attributes.td, we would be adding 3 new attributes to
EnumAttr, just to support this, which I think is an overkill.
Instead, I think the best way to do this is to add an API in Function class
that checks whether this function is a coroutine, by checking the attribute by
name directly.
Differential Revision: https://reviews.llvm.org/D92706
Added:
llvm/test/Transforms/Coroutines/Inputs/sample.text.prof
llvm/test/Transforms/Coroutines/coro-inline.ll
Modified:
llvm/include/llvm/IR/Function.h
llvm/lib/Analysis/InlineCost.cpp
llvm/lib/Transforms/IPO/AlwaysInliner.cpp
Removed:
diff --git a/llvm/include/llvm/IR/Function.h b/llvm/include/llvm/IR/Function.h
index cf61db1d33479..019e3a98a1af5 100644
--- a/llvm/include/llvm/IR/Function.h
+++ b/llvm/include/llvm/IR/Function.h
@@ -268,6 +268,12 @@ class Function : public GlobalObject, public
ilist_node {
getContext(), AttributeList::FunctionIndex, Kind));
}
+ /// A function will have the "coroutine.presplit" attribute if it's
+ /// a coroutine and has not gone through full CoroSplit pass.
+ bool isPresplitCoroutine() const {
+return hasFnAttribute("coroutine.presplit");
+ }
+
enum ProfileCountType { PCT_Invalid, PCT_Real, PCT_Synthetic };
/// Class to represent profile counts.
diff --git a/llvm/lib/Analysis/InlineCost.cpp
b/llvm/lib/Analysis/InlineCost.cpp
index 077b05edcd136..5ee12818e44cc 100644
--- a/llvm/lib/Analysis/InlineCost.cpp
+++ b/llvm/lib/Analysis/InlineCost.cpp
@@ -2332,6 +2332,13 @@ Optional
llvm::getAttributeBasedInliningDecision(
if (!Callee)
return InlineResult::failure("indirect call");
+ // When callee coroutine function is inlined into caller coroutine function
+ // before coro-split pass,
+ // coro-early pass can not handle this quiet well.
+ // So we won't inline the coroutine function if it have not been unsplited
+ if (Callee->isPresplitCoroutine())
+return InlineResult::failure("unsplited coroutine call");
+
// Never inline calls with byval arguments that does not have the alloca
// address space. Since byval arguments can be replaced with a copy to an
// alloca, the inlined code would need to be adjusted to handle that the
diff --git a/llvm/lib/Transforms/IPO/AlwaysInliner.cpp
b/llvm/lib/Transforms/IPO/AlwaysInliner.cpp
index cf810dbc39523..532599b42e0dc 100644
--- a/llvm/lib/Transforms/IPO/AlwaysInliner.cpp
+++ b/llvm/lib/Transforms/IPO/AlwaysInliner.cpp
@@ -46,7 +46,14 @@ PreservedAnalyses AlwaysInlinerPass::run(Module &M,
SmallSetVector Calls;
bool Changed = false;
SmallVector InlinedFunctions;
- for (Function &F : M)
+ for (Function &F : M) {
+// When callee coroutine function is inlined into caller coroutine function
+// before coro-split pass,
+// coro-early pass can not handle this quiet well.
+// So we won't inline the coroutine function if it have not been unsplited
+if (F.isPresplitCoroutine())
+ continue;
+
if (!F.isDeclaration() && F.hasFnAttribute(Attribute::AlwaysInline) &&
isInlineViable(F).isSuccess()) {
Calls.clear();
@@ -90,6 +97,7 @@ PreservedAnalyses AlwaysInlinerPass::run(Module &M,
// invalidation issues while deleting functions.
InlinedFunctions.push_back(&F);
}
+ }
// Remove any live functions.
erase_if(InlinedFunctions, [&](Function *F) {
@@ -182,6 +190,13 @@ InlineCost AlwaysInlinerLegacyPass::getInlineCost(CallBase
&CB) {
if (!Callee)
return InlineCost::getNever("indirect call");
+ // When callee coroutine function
