2015-04-24 07:22:33 +08:00
|
|
|
; RUN: llc < %s | FileCheck %s
|
2015-04-22 02:23:57 +08:00
|
|
|
|
|
|
|
|
; Test case based on this source:
|
|
|
|
|
; int puts(const char*);
|
|
|
|
|
; __declspec(noinline) void crash() {
|
|
|
|
|
; *(volatile int*)0 = 42;
|
|
|
|
|
; }
|
|
|
|
|
; int filt();
|
|
|
|
|
; void use_both() {
|
|
|
|
|
; __try {
|
|
|
|
|
; __try {
|
|
|
|
|
; crash();
|
|
|
|
|
; } __finally {
|
|
|
|
|
; puts("__finally");
|
|
|
|
|
; }
|
|
|
|
|
; } __except (filt()) {
|
|
|
|
|
; puts("__except");
|
|
|
|
|
; }
|
|
|
|
|
; }
|
|
|
|
|
|
|
|
|
|
target datalayout = "e-m:w-i64:64-f80:128-n8:16:32:64-S128"
|
|
|
|
|
target triple = "x86_64-pc-windows-msvc"
|
|
|
|
|
|
|
|
|
|
$"\01??_C@_09KJEHOMHG@__finally?$AA@" = comdat any
|
|
|
|
|
|
|
|
|
|
$"\01??_C@_08MLCMLGHM@__except?$AA@" = comdat any
|
|
|
|
|
|
|
|
|
|
@"\01??_C@_09KJEHOMHG@__finally?$AA@" = linkonce_odr unnamed_addr constant [10 x i8] c"__finally\00", comdat, align 1
|
|
|
|
|
@"\01??_C@_08MLCMLGHM@__except?$AA@" = linkonce_odr unnamed_addr constant [9 x i8] c"__except\00", comdat, align 1
|
|
|
|
|
|
|
|
|
|
declare void @crash()
|
|
|
|
|
|
|
|
|
|
declare i32 @filt()
|
|
|
|
|
|
|
|
|
|
; Function Attrs: nounwind uwtable
|
2015-06-18 04:52:32 +08:00
|
|
|
define void @use_both() #1 personality i8* bitcast (i32 (...)* @__C_specific_handler to i8*) {
|
2015-04-22 02:23:57 +08:00
|
|
|
entry:
|
|
|
|
|
%exn.slot = alloca i8*
|
|
|
|
|
%ehselector.slot = alloca i32
|
|
|
|
|
invoke void @crash() #5
|
2015-10-10 07:05:54 +08:00
|
|
|
to label %invoke.cont unwind label %__finally
|
2015-04-22 02:23:57 +08:00
|
|
|
|
|
|
|
|
invoke.cont: ; preds = %entry
|
2015-07-08 07:23:03 +08:00
|
|
|
%0 = call i8* @llvm.localaddress()
|
2015-04-22 02:23:57 +08:00
|
|
|
invoke void @"\01?fin$0@0@use_both@@"(i1 zeroext false, i8* %0) #5
|
2015-10-10 07:05:54 +08:00
|
|
|
to label %invoke.cont2 unwind label %catch.dispatch
|
2015-04-22 02:23:57 +08:00
|
|
|
|
|
|
|
|
invoke.cont2: ; preds = %invoke.cont
|
|
|
|
|
br label %__try.cont
|
|
|
|
|
|
2015-10-10 07:05:54 +08:00
|
|
|
__finally: ; preds = %entry
|
[IR] Reformulate LLVM's EH funclet IR
While we have successfully implemented a funclet-oriented EH scheme on
top of LLVM IR, our scheme has some notable deficiencies:
- catchendpad and cleanupendpad are necessary in the current design
but they are difficult to explain to others, even to seasoned LLVM
experts.
- catchendpad and cleanupendpad are optimization barriers. They cannot
be split and force all potentially throwing call-sites to be invokes.
This has a noticable effect on the quality of our code generation.
- catchpad, while similar in some aspects to invoke, is fairly awkward.
It is unsplittable, starts a funclet, and has control flow to other
funclets.
- The nesting relationship between funclets is currently a property of
control flow edges. Because of this, we are forced to carefully
analyze the flow graph to see if there might potentially exist illegal
nesting among funclets. While we have logic to clone funclets when
they are illegally nested, it would be nicer if we had a
representation which forbade them upfront.
Let's clean this up a bit by doing the following:
- Instead, make catchpad more like cleanuppad and landingpad: no control
flow, just a bunch of simple operands; catchpad would be splittable.
- Introduce catchswitch, a control flow instruction designed to model
the constraints of funclet oriented EH.
- Make funclet scoping explicit by having funclet instructions consume
the token produced by the funclet which contains them.
- Remove catchendpad and cleanupendpad. Their presence can be inferred
implicitly using coloring information.
N.B. The state numbering code for the CLR has been updated but the
veracity of it's output cannot be spoken for. An expert should take a
look to make sure the results are reasonable.
Reviewers: rnk, JosephTremoulet, andrew.w.kaylor
Differential Revision: http://reviews.llvm.org/D15139
llvm-svn: 255422
2015-12-12 13:38:55 +08:00
|
|
|
%cleanuppad = cleanuppad within none []
|
2015-10-10 07:05:54 +08:00
|
|
|
%locals = call i8* @llvm.localaddress()
|
2015-12-16 05:27:27 +08:00
|
|
|
invoke void @"\01?fin$0@0@use_both@@"(i1 zeroext true, i8* %locals) #5 [ "funclet"(token %cleanuppad) ]
|
[IR] Reformulate LLVM's EH funclet IR
While we have successfully implemented a funclet-oriented EH scheme on
top of LLVM IR, our scheme has some notable deficiencies:
- catchendpad and cleanupendpad are necessary in the current design
but they are difficult to explain to others, even to seasoned LLVM
experts.
- catchendpad and cleanupendpad are optimization barriers. They cannot
be split and force all potentially throwing call-sites to be invokes.
This has a noticable effect on the quality of our code generation.
- catchpad, while similar in some aspects to invoke, is fairly awkward.
It is unsplittable, starts a funclet, and has control flow to other
funclets.
- The nesting relationship between funclets is currently a property of
control flow edges. Because of this, we are forced to carefully
analyze the flow graph to see if there might potentially exist illegal
nesting among funclets. While we have logic to clone funclets when
they are illegally nested, it would be nicer if we had a
representation which forbade them upfront.
Let's clean this up a bit by doing the following:
- Instead, make catchpad more like cleanuppad and landingpad: no control
flow, just a bunch of simple operands; catchpad would be splittable.
- Introduce catchswitch, a control flow instruction designed to model
the constraints of funclet oriented EH.
- Make funclet scoping explicit by having funclet instructions consume
the token produced by the funclet which contains them.
- Remove catchendpad and cleanupendpad. Their presence can be inferred
implicitly using coloring information.
N.B. The state numbering code for the CLR has been updated but the
veracity of it's output cannot be spoken for. An expert should take a
look to make sure the results are reasonable.
Reviewers: rnk, JosephTremoulet, andrew.w.kaylor
Differential Revision: http://reviews.llvm.org/D15139
llvm-svn: 255422
2015-12-12 13:38:55 +08:00
|
|
|
to label %invoke.cont3 unwind label %catch.dispatch
|
2015-10-10 07:05:54 +08:00
|
|
|
|
|
|
|
|
invoke.cont3: ; preds = %__finally
|
[IR] Reformulate LLVM's EH funclet IR
While we have successfully implemented a funclet-oriented EH scheme on
top of LLVM IR, our scheme has some notable deficiencies:
- catchendpad and cleanupendpad are necessary in the current design
but they are difficult to explain to others, even to seasoned LLVM
experts.
- catchendpad and cleanupendpad are optimization barriers. They cannot
be split and force all potentially throwing call-sites to be invokes.
This has a noticable effect on the quality of our code generation.
- catchpad, while similar in some aspects to invoke, is fairly awkward.
It is unsplittable, starts a funclet, and has control flow to other
funclets.
- The nesting relationship between funclets is currently a property of
control flow edges. Because of this, we are forced to carefully
analyze the flow graph to see if there might potentially exist illegal
nesting among funclets. While we have logic to clone funclets when
they are illegally nested, it would be nicer if we had a
representation which forbade them upfront.
Let's clean this up a bit by doing the following:
- Instead, make catchpad more like cleanuppad and landingpad: no control
flow, just a bunch of simple operands; catchpad would be splittable.
- Introduce catchswitch, a control flow instruction designed to model
the constraints of funclet oriented EH.
- Make funclet scoping explicit by having funclet instructions consume
the token produced by the funclet which contains them.
- Remove catchendpad and cleanupendpad. Their presence can be inferred
implicitly using coloring information.
N.B. The state numbering code for the CLR has been updated but the
veracity of it's output cannot be spoken for. An expert should take a
look to make sure the results are reasonable.
Reviewers: rnk, JosephTremoulet, andrew.w.kaylor
Differential Revision: http://reviews.llvm.org/D15139
llvm-svn: 255422
2015-12-12 13:38:55 +08:00
|
|
|
cleanupret from %cleanuppad unwind label %catch.dispatch
|
2015-04-22 02:23:57 +08:00
|
|
|
|
|
|
|
|
catch.dispatch: ; preds = %invoke.cont3, %lpad1
|
[IR] Reformulate LLVM's EH funclet IR
While we have successfully implemented a funclet-oriented EH scheme on
top of LLVM IR, our scheme has some notable deficiencies:
- catchendpad and cleanupendpad are necessary in the current design
but they are difficult to explain to others, even to seasoned LLVM
experts.
- catchendpad and cleanupendpad are optimization barriers. They cannot
be split and force all potentially throwing call-sites to be invokes.
This has a noticable effect on the quality of our code generation.
- catchpad, while similar in some aspects to invoke, is fairly awkward.
It is unsplittable, starts a funclet, and has control flow to other
funclets.
- The nesting relationship between funclets is currently a property of
control flow edges. Because of this, we are forced to carefully
analyze the flow graph to see if there might potentially exist illegal
nesting among funclets. While we have logic to clone funclets when
they are illegally nested, it would be nicer if we had a
representation which forbade them upfront.
Let's clean this up a bit by doing the following:
- Instead, make catchpad more like cleanuppad and landingpad: no control
flow, just a bunch of simple operands; catchpad would be splittable.
- Introduce catchswitch, a control flow instruction designed to model
the constraints of funclet oriented EH.
- Make funclet scoping explicit by having funclet instructions consume
the token produced by the funclet which contains them.
- Remove catchendpad and cleanupendpad. Their presence can be inferred
implicitly using coloring information.
N.B. The state numbering code for the CLR has been updated but the
veracity of it's output cannot be spoken for. An expert should take a
look to make sure the results are reasonable.
Reviewers: rnk, JosephTremoulet, andrew.w.kaylor
Differential Revision: http://reviews.llvm.org/D15139
llvm-svn: 255422
2015-12-12 13:38:55 +08:00
|
|
|
%cs1 = catchswitch within none [label %__except] unwind to caller
|
2015-04-22 02:23:57 +08:00
|
|
|
|
|
|
|
|
__except: ; preds = %catch.dispatch
|
[IR] Reformulate LLVM's EH funclet IR
While we have successfully implemented a funclet-oriented EH scheme on
top of LLVM IR, our scheme has some notable deficiencies:
- catchendpad and cleanupendpad are necessary in the current design
but they are difficult to explain to others, even to seasoned LLVM
experts.
- catchendpad and cleanupendpad are optimization barriers. They cannot
be split and force all potentially throwing call-sites to be invokes.
This has a noticable effect on the quality of our code generation.
- catchpad, while similar in some aspects to invoke, is fairly awkward.
It is unsplittable, starts a funclet, and has control flow to other
funclets.
- The nesting relationship between funclets is currently a property of
control flow edges. Because of this, we are forced to carefully
analyze the flow graph to see if there might potentially exist illegal
nesting among funclets. While we have logic to clone funclets when
they are illegally nested, it would be nicer if we had a
representation which forbade them upfront.
Let's clean this up a bit by doing the following:
- Instead, make catchpad more like cleanuppad and landingpad: no control
flow, just a bunch of simple operands; catchpad would be splittable.
- Introduce catchswitch, a control flow instruction designed to model
the constraints of funclet oriented EH.
- Make funclet scoping explicit by having funclet instructions consume
the token produced by the funclet which contains them.
- Remove catchendpad and cleanupendpad. Their presence can be inferred
implicitly using coloring information.
N.B. The state numbering code for the CLR has been updated but the
veracity of it's output cannot be spoken for. An expert should take a
look to make sure the results are reasonable.
Reviewers: rnk, JosephTremoulet, andrew.w.kaylor
Differential Revision: http://reviews.llvm.org/D15139
llvm-svn: 255422
2015-12-12 13:38:55 +08:00
|
|
|
%catchpad = catchpad within %cs1 [i8* bitcast (i32 (i8*, i8*)* @"\01?filt$0@0@use_both@@" to i8*)]
|
2015-12-16 05:27:27 +08:00
|
|
|
%call = call i32 @puts(i8* getelementptr inbounds ([9 x i8], [9 x i8]* @"\01??_C@_08MLCMLGHM@__except?$AA@", i32 0, i32 0)) [ "funclet"(token %catchpad) ]
|
[IR] Reformulate LLVM's EH funclet IR
While we have successfully implemented a funclet-oriented EH scheme on
top of LLVM IR, our scheme has some notable deficiencies:
- catchendpad and cleanupendpad are necessary in the current design
but they are difficult to explain to others, even to seasoned LLVM
experts.
- catchendpad and cleanupendpad are optimization barriers. They cannot
be split and force all potentially throwing call-sites to be invokes.
This has a noticable effect on the quality of our code generation.
- catchpad, while similar in some aspects to invoke, is fairly awkward.
It is unsplittable, starts a funclet, and has control flow to other
funclets.
- The nesting relationship between funclets is currently a property of
control flow edges. Because of this, we are forced to carefully
analyze the flow graph to see if there might potentially exist illegal
nesting among funclets. While we have logic to clone funclets when
they are illegally nested, it would be nicer if we had a
representation which forbade them upfront.
Let's clean this up a bit by doing the following:
- Instead, make catchpad more like cleanuppad and landingpad: no control
flow, just a bunch of simple operands; catchpad would be splittable.
- Introduce catchswitch, a control flow instruction designed to model
the constraints of funclet oriented EH.
- Make funclet scoping explicit by having funclet instructions consume
the token produced by the funclet which contains them.
- Remove catchendpad and cleanupendpad. Their presence can be inferred
implicitly using coloring information.
N.B. The state numbering code for the CLR has been updated but the
veracity of it's output cannot be spoken for. An expert should take a
look to make sure the results are reasonable.
Reviewers: rnk, JosephTremoulet, andrew.w.kaylor
Differential Revision: http://reviews.llvm.org/D15139
llvm-svn: 255422
2015-12-12 13:38:55 +08:00
|
|
|
catchret from %catchpad to label %__try.cont
|
2015-04-22 02:23:57 +08:00
|
|
|
|
|
|
|
|
__try.cont: ; preds = %__except, %invoke.cont2
|
|
|
|
|
ret void
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
; CHECK-LABEL: use_both:
|
|
|
|
|
; CHECK: .Ltmp0
|
|
|
|
|
; CHECK: callq crash
|
|
|
|
|
; CHECK: .Ltmp1
|
|
|
|
|
; CHECK: .Ltmp4
|
2015-10-10 07:05:54 +08:00
|
|
|
; CHECK: callq "?fin$0@0@use_both@@"
|
|
|
|
|
; CHECK: .Ltmp5
|
2015-04-22 02:23:57 +08:00
|
|
|
; CHECK: retq
|
|
|
|
|
;
|
|
|
|
|
; CHECK: .seh_handlerdata
|
2015-12-16 07:40:58 +08:00
|
|
|
; CHECK-NEXT: .Luse_both$parent_frame_offset
|
2015-10-10 07:05:54 +08:00
|
|
|
; CHECK-NEXT: .long (.Llsda_end0-.Llsda_begin0)/16
|
|
|
|
|
; CHECK-NEXT: .Llsda_begin0:
|
|
|
|
|
; CHECK-NEXT: .long .Ltmp0@IMGREL+1
|
2015-04-22 02:23:57 +08:00
|
|
|
; CHECK-NEXT: .long .Ltmp1@IMGREL+1
|
2015-10-10 07:05:54 +08:00
|
|
|
; CHECK-NEXT: .long "?dtor$2@?0?use_both@4HA"@IMGREL
|
2015-04-22 02:23:57 +08:00
|
|
|
; CHECK-NEXT: .long 0
|
2015-10-10 07:05:54 +08:00
|
|
|
; CHECK-NEXT: .long .Ltmp0@IMGREL+1
|
2015-04-22 02:23:57 +08:00
|
|
|
; CHECK-NEXT: .long .Ltmp1@IMGREL+1
|
|
|
|
|
; CHECK-NEXT: .long "?filt$0@0@use_both@@"@IMGREL
|
2015-10-10 07:05:54 +08:00
|
|
|
; CHECK-NEXT: .long .LBB0_{{[0-9]+}}@IMGREL
|
2015-04-22 02:23:57 +08:00
|
|
|
; CHECK-NEXT: .long .Ltmp4@IMGREL+1
|
2015-10-10 07:05:54 +08:00
|
|
|
; CHECK-NEXT: .long .Ltmp5@IMGREL+1
|
2015-04-22 02:23:57 +08:00
|
|
|
; CHECK-NEXT: .long "?filt$0@0@use_both@@"@IMGREL
|
2015-10-10 07:05:54 +08:00
|
|
|
; CHECK-NEXT: .long .LBB0_{{[0-9]+}}@IMGREL
|
|
|
|
|
; CHECK-NEXT: .Llsda_end0:
|
2015-04-22 02:23:57 +08:00
|
|
|
|
|
|
|
|
; Function Attrs: noinline nounwind
|
|
|
|
|
define internal i32 @"\01?filt$0@0@use_both@@"(i8* %exception_pointers, i8* %frame_pointer) #2 {
|
|
|
|
|
entry:
|
|
|
|
|
%frame_pointer.addr = alloca i8*, align 8
|
|
|
|
|
%exception_pointers.addr = alloca i8*, align 8
|
|
|
|
|
%exn.slot = alloca i8*
|
|
|
|
|
store i8* %frame_pointer, i8** %frame_pointer.addr, align 8
|
|
|
|
|
store i8* %exception_pointers, i8** %exception_pointers.addr, align 8
|
|
|
|
|
%0 = load i8*, i8** %exception_pointers.addr
|
|
|
|
|
%1 = bitcast i8* %0 to { i32*, i8* }*
|
|
|
|
|
%2 = getelementptr inbounds { i32*, i8* }, { i32*, i8* }* %1, i32 0, i32 0
|
|
|
|
|
%3 = load i32*, i32** %2
|
|
|
|
|
%4 = load i32, i32* %3
|
|
|
|
|
%5 = zext i32 %4 to i64
|
|
|
|
|
%6 = inttoptr i64 %5 to i8*
|
|
|
|
|
store i8* %6, i8** %exn.slot
|
|
|
|
|
%call = call i32 @filt()
|
|
|
|
|
ret i32 %call
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
define internal void @"\01?fin$0@0@use_both@@"(i1 zeroext %abnormal_termination, i8* %frame_pointer) #3 {
|
|
|
|
|
entry:
|
|
|
|
|
%frame_pointer.addr = alloca i8*, align 8
|
|
|
|
|
%abnormal_termination.addr = alloca i8, align 1
|
|
|
|
|
store i8* %frame_pointer, i8** %frame_pointer.addr, align 8
|
|
|
|
|
%frombool = zext i1 %abnormal_termination to i8
|
|
|
|
|
store i8 %frombool, i8* %abnormal_termination.addr, align 1
|
|
|
|
|
%call = call i32 @puts(i8* getelementptr inbounds ([10 x i8], [10 x i8]* @"\01??_C@_09KJEHOMHG@__finally?$AA@", i32 0, i32 0))
|
|
|
|
|
ret void
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
declare i32 @puts(i8*) #3
|
|
|
|
|
|
|
|
|
|
declare i32 @__C_specific_handler(...)
|
|
|
|
|
|
|
|
|
|
; Function Attrs: nounwind readnone
|
2015-07-08 07:23:03 +08:00
|
|
|
declare i8* @llvm.localaddress() #4
|
2015-04-22 02:23:57 +08:00
|
|
|
|
|
|
|
|
; Function Attrs: nounwind readnone
|
|
|
|
|
declare i32 @llvm.eh.typeid.for(i8*) #4
|
|
|
|
|
|
|
|
|
|
attributes #0 = { noinline nounwind uwtable "less-precise-fpmad"="false" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "stack-protector-buffer-size"="8" "target-cpu"="x86-64" "unsafe-fp-math"="false" "use-soft-float"="false" }
|
|
|
|
|
attributes #1 = { nounwind uwtable "less-precise-fpmad"="false" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "stack-protector-buffer-size"="8" "target-cpu"="x86-64" "unsafe-fp-math"="false" "use-soft-float"="false" }
|
|
|
|
|
attributes #2 = { noinline nounwind "less-precise-fpmad"="false" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "stack-protector-buffer-size"="8" "target-cpu"="x86-64" "unsafe-fp-math"="false" "use-soft-float"="false" }
|
|
|
|
|
attributes #3 = { "less-precise-fpmad"="false" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "stack-protector-buffer-size"="8" "target-cpu"="x86-64" "unsafe-fp-math"="false" "use-soft-float"="false" }
|
|
|
|
|
attributes #4 = { nounwind readnone }
|
|
|
|
|
attributes #5 = { noinline }
|
|
|
|
|
attributes #6 = { nounwind }
|