llvm-project/clang/test/CodeGenObjCXX/lambda-expressions.mm
Chandler Carruth fcd33149b4 Cleanup the handling of noinline function attributes, -fno-inline,
-fno-inline-functions, -O0, and optnone.

These were really, really tangled together:
- We used the noinline LLVM attribute for -fno-inline
  - But not for -fno-inline-functions (breaking LTO)
  - But we did use it for -finline-hint-functions (yay, LTO is happy!)
  - But we didn't for -O0 (LTO is sad yet again...)
- We had weird structuring of CodeGenOpts with both an inlining
  enumeration and a boolean. They interacted in weird ways and
  needlessly.
- A *lot* of set smashing went on with setting these, and then got worse
  when we considered optnone and other inlining-effecting attributes.
- A bunch of inline affecting attributes were managed in a completely
  different place from -fno-inline.
- Even with -fno-inline we failed to put the LLVM noinline attribute
  onto many generated function definitions because they didn't show up
  as AST-level functions.
- If you passed -O0 but -finline-functions we would run the normal
  inliner pass in LLVM despite it being in the O0 pipeline, which really
  doesn't make much sense.
- Lastly, we used things like '-fno-inline' to manipulate the pass
  pipeline which forced the pass pipeline to be much more
  parameterizable than it really needs to be. Instead we can *just* use
  the optimization level to select a pipeline and control the rest via
  attributes.

Sadly, this causes a bunch of churn in tests because we don't run the
optimizer in the tests and check the contents of attribute sets. It
would be awesome if attribute sets were a bit more FileCheck friendly,
but oh well.

I think this is a significant improvement and should remove the semantic
need to change what inliner pass we run in order to comply with the
requested inlining semantics by relying completely on attributes. It
also cleans up tho optnone and related handling a bit.

One unfortunate aspect of this is that for generating alwaysinline
routines like those in OpenMP we end up removing noinline and then
adding alwaysinline. I tried a bunch of other approaches, but because we
recompute function attributes from scratch and don't have a declaration
here I couldn't find anything substantially cleaner than this.

Differential Revision: https://reviews.llvm.org/D28053

llvm-svn: 290398
2016-12-23 01:24:49 +00:00

130 lines
4.9 KiB
Plaintext

// RUN: %clang_cc1 -triple x86_64-apple-darwin10.0.0 -emit-llvm -o - %s -fexceptions -std=c++11 -fblocks -fobjc-arc | FileCheck -check-prefix=ARC %s
// RUN: %clang_cc1 -triple x86_64-apple-darwin10.0.0 -emit-llvm -o - %s -fexceptions -std=c++11 -fblocks | FileCheck -check-prefix=MRC %s
typedef int (^fp)();
fp f() { auto x = []{ return 3; }; return x; }
// ARC: %[[LAMBDACLASS:.*]] = type { i32 }
// MRC: @OBJC_METH_VAR_NAME{{.*}} = private unnamed_addr constant [5 x i8] c"copy\00"
// MRC: @OBJC_METH_VAR_NAME{{.*}} = private unnamed_addr constant [12 x i8] c"autorelease\00"
// MRC-LABEL: define i32 ()* @_Z1fv(
// MRC-LABEL: define internal i32 ()* @"_ZZ1fvENK3$_0cvU13block_pointerFivEEv"
// MRC: store i8* bitcast (i8** @_NSConcreteStackBlock to i8*)
// MRC: store i8* bitcast (i32 (i8*)* @"___ZZ1fvENK3$_0cvU13block_pointerFivEEv_block_invoke" to i8*)
// MRC: call i32 ()* bitcast (i8* (i8*, i8*, ...)* @objc_msgSend to i32 ()* (i8*, i8*)*)
// MRC: call i32 ()* bitcast (i8* (i8*, i8*, ...)* @objc_msgSend to i32 ()* (i8*, i8*)*)
// MRC: ret i32 ()*
// ARC-LABEL: define i32 ()* @_Z1fv(
// ARC-LABEL: define internal i32 ()* @"_ZZ1fvENK3$_0cvU13block_pointerFivEEv"
// ARC: store i8* bitcast (i8** @_NSConcreteStackBlock to i8*)
// ARC: store i8* bitcast (i32 (i8*)* @"___ZZ1fvENK3$_0cvU13block_pointerFivEEv_block_invoke" to i8*)
// ARC: call i8* @objc_retainBlock
// ARC: call i8* @objc_autoreleaseReturnValue
typedef int (^fp)();
fp global;
void f2() { global = []{ return 3; }; }
// MRC: define void @_Z2f2v() [[NUW:#[0-9]+]] {
// MRC: store i8* bitcast (i32 (i8*)* @___Z2f2v_block_invoke to i8*),
// MRC-NOT: call
// MRC: ret void
// ("global" contains a dangling pointer after this function runs.)
// ARC: define void @_Z2f2v() [[NUW:#[0-9]+]] {
// ARC: store i8* bitcast (i32 (i8*)* @___Z2f2v_block_invoke to i8*),
// ARC: call i8* @objc_retainBlock
// ARC: call void @objc_release
// ARC-LABEL: define internal i32 @___Z2f2v_block_invoke
// ARC: call i32 @"_ZZ2f2vENK3$_1clEv
template <class T> void take_lambda(T &&lambda) { lambda(); }
void take_block(void (^block)()) { block(); }
// rdar://13800041
@interface A
- (void) test;
@end
@interface B : A @end
@implementation B
- (void) test {
take_block(^{
take_lambda([=]{
take_block(^{
take_lambda([=] {
[super test];
});
});
});
});
}
@end
// ARC: define void @_ZN13LambdaCapture4foo1ERi(i32* dereferenceable(4) %{{.*}})
// ARC: %[[CAPTURE0:.*]] = getelementptr inbounds %[[LAMBDACLASS]], %[[LAMBDACLASS]]* %{{.*}}, i32 0, i32 0
// ARC: store i32 %{{.*}}, i32* %[[CAPTURE0]]
// ARC: define internal void @"_ZZN13LambdaCapture4foo1ERiENK3$_3clEv"(%[[LAMBDACLASS]]* %{{.*}})
// ARC: %[[BLOCK:.*]] = alloca <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, i32 }>
// ARC: %[[CAPTURE1:.*]] = getelementptr inbounds <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, i32 }>, <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, i32 }>* %[[BLOCK]], i32 0, i32 5
// ARC: store i32 %{{.*}}, i32* %[[CAPTURE1]]
// ARC: define internal void @"___ZZN13LambdaCapture4foo1ERiENK3$_3clEv_block_invoke"
// ARC: %[[CAPTURE2:.*]] = getelementptr inbounds <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, i32 }>, <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, i32 }>* %{{.*}}, i32 0, i32 5
// ARC: store i32 %{{.*}}, i32* %[[CAPTURE2]]
// ARC: define internal void @"___ZZN13LambdaCapture4foo1ERiENK3$_3clEv_block_invoke_2"(i8* %{{.*}})
// ARC: %[[CAPTURE3:.*]] = getelementptr inbounds <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, i32 }>, <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, i32 }>* %{{.*}}, i32 0, i32 5
// ARC: %[[V1:.*]] = load i32, i32* %[[CAPTURE3]]
// ARC: store i32 %[[V1]], i32* @_ZN13LambdaCapture1iE
namespace LambdaCapture {
int i;
void foo1(int &a) {
auto lambda = [a]{
auto block1 = ^{
auto block2 = ^{
i = a;
};
block2();
};
block1();
};
lambda();
}
}
// ARC-LABEL: define linkonce_odr i32 ()* @_ZZNK13StaticMembersIfE1fMUlvE_clEvENKUlvE_cvU13block_pointerFivEEv
// Check lines for BlockInLambda test below
// ARC-LABEL: define internal i32 @___ZZN13BlockInLambda1X1fEvENKUlvE_clEv_block_invoke
// ARC: [[Y:%.*]] = getelementptr inbounds %"struct.BlockInLambda::X", %"struct.BlockInLambda::X"* {{.*}}, i32 0, i32 1
// ARC-NEXT: [[YVAL:%.*]] = load i32, i32* [[Y]], align 4
// ARC-NEXT: ret i32 [[YVAL]]
typedef int (^fptr)();
template<typename T> struct StaticMembers {
static fptr f;
};
template<typename T>
fptr StaticMembers<T>::f = [] { auto f = []{return 5;}; return fptr(f); }();
template fptr StaticMembers<float>::f;
namespace BlockInLambda {
struct X {
int x,y;
void f() {
[this]{return ^{return y;}();}();
};
};
void g(X& x) {
x.f();
};
}
// ARC: attributes [[NUW]] = { noinline nounwind{{.*}} }
// MRC: attributes [[NUW]] = { noinline nounwind{{.*}} }