maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/clang/test/CodeGen/exceptions-seh-finally.c

233 lines
6.5 KiB
C
Raw Normal View History

[SEH] Delete the 32-bit IR lowering for __finally blocks and use x64 32-bit finally funclets are intended to be called both directly from the parent function and indirectly from the EH runtime. Because we aren't contorting LLVM's X86 prologue to match MSVC's, calling the finally block directly passes in a different value of EBP than the one that the runtime provides. We need an adapter thunk to adjust EBP to the expected value. However, WinEHPrepare already has to solve this problem when cleanups are not pre-outlined, so we can go ahead and rely on it rather than duplicating work. Now we only do the llvm.x86.seh.recoverfp dance for 32-bit SEH filter functions. llvm-svn: 241187
2015-07-02 05:00:00 +08:00
// RUN: %clang_cc1 %s -triple x86_64-pc-win32 -fms-extensions -emit-llvm -o - | FileCheck %s
Re-enable 32-bit SEH after the alignment fix llvm-svn: 241878
2015-07-10 08:16:25 +08:00
// RUN: %clang_cc1 %s -triple i686-pc-win32 -fms-extensions -emit-llvm -o - | FileCheck %s
Initial support for Win64 SEH IR emission The lowering looks a lot like normal EH lowering, with the exception that the exceptions are caught by executing filter expression code instead of matching typeinfo globals. The filter expressions are outlined into functions which are used in landingpad clauses where typeinfo would normally go. Major aspects that still need work: - Non-call exceptions in __try bodies won't work yet. The plan is to outline the __try block in the frontend to keep things simple. - Filter expressions cannot use local variables until capturing is implemented. - __finally blocks will not run after exceptions. Fixing this requires work in the LLVM SEH preparation pass. The IR lowering looks like this: // C code: bool safe_div(int n, int d, int *r) { __try { *r = normal_div(n, d); } __except(_exception_code() == EXCEPTION_INT_DIVIDE_BY_ZERO) { return false; } return true; } ; LLVM IR: define i32 @filter(i8* %e, i8* %fp) { %ehptrs = bitcast i8* %e to i32** %ehrec = load i32** %ehptrs %code = load i32* %ehrec %matches = icmp eq i32 %code, i32 u0xC0000094 %matches.i32 = zext i1 %matches to i32 ret i32 %matches.i32 } define i1 zeroext @safe_div(i32 %n, i32 %d, i32* %r) { %rr = invoke i32 @normal_div(i32 %n, i32 %d) to label %normal unwind to label %lpad normal: store i32 %rr, i32* %r ret i1 1 lpad: %ehvals = landingpad {i8*, i32} personality i32 (...)* @__C_specific_handler catch i8* bitcast (i32 (i8*, i8*)* @filter to i8*) %ehptr = extractvalue {i8*, i32} %ehvals, i32 0 %sel = extractvalue {i8*, i32} %ehvals, i32 1 %filter_sel = call i32 @llvm.eh.seh.typeid.for(i8* bitcast (i32 (i8*, i8*)* @filter to i8*)) %matches = icmp eq i32 %sel, %filter_sel br i1 %matches, label %eh.except, label %eh.resume eh.except: ret i1 false eh.resume: resume } Reviewers: rjmccall, rsmith, majnemer Differential Revision: http://reviews.llvm.org/D5607 llvm-svn: 226760
2015-01-22 09:36:17 +08:00
Fix crash on finally blocks that don't fall through llvm-svn: 228243
2015-02-05 08:58:46 +08:00
void abort(void) __attribute__((noreturn));
Initial support for Win64 SEH IR emission The lowering looks a lot like normal EH lowering, with the exception that the exceptions are caught by executing filter expression code instead of matching typeinfo globals. The filter expressions are outlined into functions which are used in landingpad clauses where typeinfo would normally go. Major aspects that still need work: - Non-call exceptions in __try bodies won't work yet. The plan is to outline the __try block in the frontend to keep things simple. - Filter expressions cannot use local variables until capturing is implemented. - __finally blocks will not run after exceptions. Fixing this requires work in the LLVM SEH preparation pass. The IR lowering looks like this: // C code: bool safe_div(int n, int d, int *r) { __try { *r = normal_div(n, d); } __except(_exception_code() == EXCEPTION_INT_DIVIDE_BY_ZERO) { return false; } return true; } ; LLVM IR: define i32 @filter(i8* %e, i8* %fp) { %ehptrs = bitcast i8* %e to i32** %ehrec = load i32** %ehptrs %code = load i32* %ehrec %matches = icmp eq i32 %code, i32 u0xC0000094 %matches.i32 = zext i1 %matches to i32 ret i32 %matches.i32 } define i1 zeroext @safe_div(i32 %n, i32 %d, i32* %r) { %rr = invoke i32 @normal_div(i32 %n, i32 %d) to label %normal unwind to label %lpad normal: store i32 %rr, i32* %r ret i1 1 lpad: %ehvals = landingpad {i8*, i32} personality i32 (...)* @__C_specific_handler catch i8* bitcast (i32 (i8*, i8*)* @filter to i8*) %ehptr = extractvalue {i8*, i32} %ehvals, i32 0 %sel = extractvalue {i8*, i32} %ehvals, i32 1 %filter_sel = call i32 @llvm.eh.seh.typeid.for(i8* bitcast (i32 (i8*, i8*)* @filter to i8*)) %matches = icmp eq i32 %sel, %filter_sel br i1 %matches, label %eh.except, label %eh.resume eh.except: ret i1 false eh.resume: resume } Reviewers: rjmccall, rsmith, majnemer Differential Revision: http://reviews.llvm.org/D5607 llvm-svn: 226760
2015-01-22 09:36:17 +08:00
void might_crash(void);
void cleanup(void);
int check_condition(void);
void basic_finally(void) {
__try {
might_crash();
} __finally {
cleanup();
}
}
// CHECK-LABEL: define void @basic_finally()
// CHECK: invoke void @might_crash()
Implement IRGen for SEH __finally and AbnormalTermination Previously we would simply double-emit the body of the __finally block, but that doesn't work when it contains any kind of Decl, which we can't double emit. This fixes that by emitting the block once and branching into a shared code region and then branching back out. llvm-svn: 228222
2015-02-05 06:37:07 +08:00
// CHECK: to label %[[invoke_cont:[^ ]*]] unwind label %[[lpad:[^ ]*]]
//
// CHECK: [[invoke_cont]]
[SEH] Switch from frameaddress(0) to localaddress This should do the right thing for stack realignment prologues. llvm-svn: 241644
2015-07-08 07:23:31 +08:00
// CHECK: %[[fp:[^ ]*]] = call i8* @llvm.localaddress()
Revert "Revert 241171, 241187, 241199 (32-bit SEH)." This reverts commit r241244, but restricts SEH support to Win64. This way, Chromium builds will still fall back on TUs with SEH, and Clang developers can work on this incrementally upstream while patching this small predicate locally. It'll also make it easier to review small fixes. llvm-svn: 241533
2015-07-07 08:36:30 +08:00
// CHECK: call void @"\01?fin$0@0@basic_finally@@"({{i8( zeroext)?}} 0, i8* %[[fp]])
Implement IRGen for SEH __finally and AbnormalTermination Previously we would simply double-emit the body of the __finally block, but that doesn't work when it contains any kind of Decl, which we can't double emit. This fixes that by emitting the block once and branching into a shared code region and then branching back out. llvm-svn: 228222
2015-02-05 06:37:07 +08:00
// CHECK-NEXT: ret void
//
// CHECK: [[lpad]]
// CHECK-NEXT: landingpad
Initial support for Win64 SEH IR emission The lowering looks a lot like normal EH lowering, with the exception that the exceptions are caught by executing filter expression code instead of matching typeinfo globals. The filter expressions are outlined into functions which are used in landingpad clauses where typeinfo would normally go. Major aspects that still need work: - Non-call exceptions in __try bodies won't work yet. The plan is to outline the __try block in the frontend to keep things simple. - Filter expressions cannot use local variables until capturing is implemented. - __finally blocks will not run after exceptions. Fixing this requires work in the LLVM SEH preparation pass. The IR lowering looks like this: // C code: bool safe_div(int n, int d, int *r) { __try { *r = normal_div(n, d); } __except(_exception_code() == EXCEPTION_INT_DIVIDE_BY_ZERO) { return false; } return true; } ; LLVM IR: define i32 @filter(i8* %e, i8* %fp) { %ehptrs = bitcast i8* %e to i32** %ehrec = load i32** %ehptrs %code = load i32* %ehrec %matches = icmp eq i32 %code, i32 u0xC0000094 %matches.i32 = zext i1 %matches to i32 ret i32 %matches.i32 } define i1 zeroext @safe_div(i32 %n, i32 %d, i32* %r) { %rr = invoke i32 @normal_div(i32 %n, i32 %d) to label %normal unwind to label %lpad normal: store i32 %rr, i32* %r ret i1 1 lpad: %ehvals = landingpad {i8*, i32} personality i32 (...)* @__C_specific_handler catch i8* bitcast (i32 (i8*, i8*)* @filter to i8*) %ehptr = extractvalue {i8*, i32} %ehvals, i32 0 %sel = extractvalue {i8*, i32} %ehvals, i32 1 %filter_sel = call i32 @llvm.eh.seh.typeid.for(i8* bitcast (i32 (i8*, i8*)* @filter to i8*)) %matches = icmp eq i32 %sel, %filter_sel br i1 %matches, label %eh.except, label %eh.resume eh.except: ret i1 false eh.resume: resume } Reviewers: rjmccall, rsmith, majnemer Differential Revision: http://reviews.llvm.org/D5607 llvm-svn: 226760
2015-01-22 09:36:17 +08:00
// CHECK-NEXT: cleanup
[SEH] Switch from frameaddress(0) to localaddress This should do the right thing for stack realignment prologues. llvm-svn: 241644
2015-07-08 07:23:31 +08:00
// CHECK: %[[fp:[^ ]*]] = call i8* @llvm.localaddress()
Revert "Revert 241171, 241187, 241199 (32-bit SEH)." This reverts commit r241244, but restricts SEH support to Win64. This way, Chromium builds will still fall back on TUs with SEH, and Clang developers can work on this incrementally upstream while patching this small predicate locally. It'll also make it easier to review small fixes. llvm-svn: 241533
2015-07-07 08:36:30 +08:00
// CHECK: call void @"\01?fin$0@0@basic_finally@@"({{i8( zeroext)?}} 1, i8* %[[fp]])
Implement IRGen for SEH __finally and AbnormalTermination Previously we would simply double-emit the body of the __finally block, but that doesn't work when it contains any kind of Decl, which we can't double emit. This fixes that by emitting the block once and branching into a shared code region and then branching back out. llvm-svn: 228222
2015-02-05 06:37:07 +08:00
// CHECK: resume { i8*, i32 }
Initial support for Win64 SEH IR emission The lowering looks a lot like normal EH lowering, with the exception that the exceptions are caught by executing filter expression code instead of matching typeinfo globals. The filter expressions are outlined into functions which are used in landingpad clauses where typeinfo would normally go. Major aspects that still need work: - Non-call exceptions in __try bodies won't work yet. The plan is to outline the __try block in the frontend to keep things simple. - Filter expressions cannot use local variables until capturing is implemented. - __finally blocks will not run after exceptions. Fixing this requires work in the LLVM SEH preparation pass. The IR lowering looks like this: // C code: bool safe_div(int n, int d, int *r) { __try { *r = normal_div(n, d); } __except(_exception_code() == EXCEPTION_INT_DIVIDE_BY_ZERO) { return false; } return true; } ; LLVM IR: define i32 @filter(i8* %e, i8* %fp) { %ehptrs = bitcast i8* %e to i32** %ehrec = load i32** %ehptrs %code = load i32* %ehrec %matches = icmp eq i32 %code, i32 u0xC0000094 %matches.i32 = zext i1 %matches to i32 ret i32 %matches.i32 } define i1 zeroext @safe_div(i32 %n, i32 %d, i32* %r) { %rr = invoke i32 @normal_div(i32 %n, i32 %d) to label %normal unwind to label %lpad normal: store i32 %rr, i32* %r ret i1 1 lpad: %ehvals = landingpad {i8*, i32} personality i32 (...)* @__C_specific_handler catch i8* bitcast (i32 (i8*, i8*)* @filter to i8*) %ehptr = extractvalue {i8*, i32} %ehvals, i32 0 %sel = extractvalue {i8*, i32} %ehvals, i32 1 %filter_sel = call i32 @llvm.eh.seh.typeid.for(i8* bitcast (i32 (i8*, i8*)* @filter to i8*)) %matches = icmp eq i32 %sel, %filter_sel br i1 %matches, label %eh.except, label %eh.resume eh.except: ret i1 false eh.resume: resume } Reviewers: rjmccall, rsmith, majnemer Differential Revision: http://reviews.llvm.org/D5607 llvm-svn: 226760
2015-01-22 09:36:17 +08:00
Revert "Revert 241171, 241187, 241199 (32-bit SEH)." This reverts commit r241244, but restricts SEH support to Win64. This way, Chromium builds will still fall back on TUs with SEH, and Clang developers can work on this incrementally upstream while patching this small predicate locally. It'll also make it easier to review small fixes. llvm-svn: 241533
2015-07-07 08:36:30 +08:00
// CHECK: define internal void @"\01?fin$0@0@basic_finally@@"({{.*}})
Reland r234613 (and follow-ups 234614, 234616, 234618) The frameescape intrinsic cannot be inlined, so I fixed the inliner in r234937. This should address PR23216. llvm-svn: 234942
2015-04-15 04:59:00 +08:00
// CHECK: call void @cleanup()
Implement IRGen for SEH __finally and AbnormalTermination Previously we would simply double-emit the body of the __finally block, but that doesn't work when it contains any kind of Decl, which we can't double emit. This fixes that by emitting the block once and branching into a shared code region and then branching back out. llvm-svn: 228222
2015-02-05 06:37:07 +08:00
// Mostly check that we don't double emit 'r' which would crash.
void decl_in_finally(void) {
__try {
might_crash();
} __finally {
int r;
}
}
// Ditto, don't crash double emitting 'l'.
void label_in_finally(void) {
Initial support for Win64 SEH IR emission The lowering looks a lot like normal EH lowering, with the exception that the exceptions are caught by executing filter expression code instead of matching typeinfo globals. The filter expressions are outlined into functions which are used in landingpad clauses where typeinfo would normally go. Major aspects that still need work: - Non-call exceptions in __try bodies won't work yet. The plan is to outline the __try block in the frontend to keep things simple. - Filter expressions cannot use local variables until capturing is implemented. - __finally blocks will not run after exceptions. Fixing this requires work in the LLVM SEH preparation pass. The IR lowering looks like this: // C code: bool safe_div(int n, int d, int *r) { __try { *r = normal_div(n, d); } __except(_exception_code() == EXCEPTION_INT_DIVIDE_BY_ZERO) { return false; } return true; } ; LLVM IR: define i32 @filter(i8* %e, i8* %fp) { %ehptrs = bitcast i8* %e to i32** %ehrec = load i32** %ehptrs %code = load i32* %ehrec %matches = icmp eq i32 %code, i32 u0xC0000094 %matches.i32 = zext i1 %matches to i32 ret i32 %matches.i32 } define i1 zeroext @safe_div(i32 %n, i32 %d, i32* %r) { %rr = invoke i32 @normal_div(i32 %n, i32 %d) to label %normal unwind to label %lpad normal: store i32 %rr, i32* %r ret i1 1 lpad: %ehvals = landingpad {i8*, i32} personality i32 (...)* @__C_specific_handler catch i8* bitcast (i32 (i8*, i8*)* @filter to i8*) %ehptr = extractvalue {i8*, i32} %ehvals, i32 0 %sel = extractvalue {i8*, i32} %ehvals, i32 1 %filter_sel = call i32 @llvm.eh.seh.typeid.for(i8* bitcast (i32 (i8*, i8*)* @filter to i8*)) %matches = icmp eq i32 %sel, %filter_sel br i1 %matches, label %eh.except, label %eh.resume eh.except: ret i1 false eh.resume: resume } Reviewers: rjmccall, rsmith, majnemer Differential Revision: http://reviews.llvm.org/D5607 llvm-svn: 226760
2015-01-22 09:36:17 +08:00
__try {
might_crash();
} __finally {
l:
cleanup();
if (check_condition())
goto l;
}
}
Implement IRGen for SEH __finally and AbnormalTermination Previously we would simply double-emit the body of the __finally block, but that doesn't work when it contains any kind of Decl, which we can't double emit. This fixes that by emitting the block once and branching into a shared code region and then branching back out. llvm-svn: 228222
2015-02-05 06:37:07 +08:00
// CHECK-LABEL: define void @label_in_finally()
// CHECK: invoke void @might_crash()
// CHECK: to label %[[invoke_cont:[^ ]*]] unwind label %[[lpad:[^ ]*]]
//
// CHECK: [[invoke_cont]]
[SEH] Switch from frameaddress(0) to localaddress This should do the right thing for stack realignment prologues. llvm-svn: 241644
2015-07-08 07:23:31 +08:00
// CHECK: %[[fp:[^ ]*]] = call i8* @llvm.localaddress()
Revert "Revert 241171, 241187, 241199 (32-bit SEH)." This reverts commit r241244, but restricts SEH support to Win64. This way, Chromium builds will still fall back on TUs with SEH, and Clang developers can work on this incrementally upstream while patching this small predicate locally. It'll also make it easier to review small fixes. llvm-svn: 241533
2015-07-07 08:36:30 +08:00
// CHECK: call void @"\01?fin$0@0@label_in_finally@@"({{i8( zeroext)?}} 0, i8* %[[fp]])
Reland r234613 (and follow-ups 234614, 234616, 234618) The frameescape intrinsic cannot be inlined, so I fixed the inliner in r234937. This should address PR23216. llvm-svn: 234942
2015-04-15 04:59:00 +08:00
// CHECK: ret void
Revert "Revert 241171, 241187, 241199 (32-bit SEH)." This reverts commit r241244, but restricts SEH support to Win64. This way, Chromium builds will still fall back on TUs with SEH, and Clang developers can work on this incrementally upstream while patching this small predicate locally. It'll also make it easier to review small fixes. llvm-svn: 241533
2015-07-07 08:36:30 +08:00
// CHECK: define internal void @"\01?fin$0@0@label_in_finally@@"({{.*}})
Implement IRGen for SEH __finally and AbnormalTermination Previously we would simply double-emit the body of the __finally block, but that doesn't work when it contains any kind of Decl, which we can't double emit. This fixes that by emitting the block once and branching into a shared code region and then branching back out. llvm-svn: 228222
2015-02-05 06:37:07 +08:00
// CHECK: br label %[[l:[^ ]*]]
//
// CHECK: [[l]]
// CHECK: call void @cleanup()
// CHECK: call i32 @check_condition()
// CHECK: br i1 {{.*}}, label
// CHECK: br label %[[l]]
int crashed;
void use_abnormal_termination(void) {
__try {
might_crash();
} __finally {
crashed = __abnormal_termination();
}
}
// CHECK-LABEL: define void @use_abnormal_termination()
// CHECK: invoke void @might_crash()
// CHECK: to label %[[invoke_cont:[^ ]*]] unwind label %[[lpad:[^ ]*]]
//
// CHECK: [[invoke_cont]]
[SEH] Switch from frameaddress(0) to localaddress This should do the right thing for stack realignment prologues. llvm-svn: 241644
2015-07-08 07:23:31 +08:00
// CHECK: %[[fp:[^ ]*]] = call i8* @llvm.localaddress()
Revert "Revert 241171, 241187, 241199 (32-bit SEH)." This reverts commit r241244, but restricts SEH support to Win64. This way, Chromium builds will still fall back on TUs with SEH, and Clang developers can work on this incrementally upstream while patching this small predicate locally. It'll also make it easier to review small fixes. llvm-svn: 241533
2015-07-07 08:36:30 +08:00
// CHECK: call void @"\01?fin$0@0@use_abnormal_termination@@"({{i8( zeroext)?}} 0, i8* %[[fp]])
Reland r234613 (and follow-ups 234614, 234616, 234618) The frameescape intrinsic cannot be inlined, so I fixed the inliner in r234937. This should address PR23216. llvm-svn: 234942
2015-04-15 04:59:00 +08:00
// CHECK: ret void
Implement IRGen for SEH __finally and AbnormalTermination Previously we would simply double-emit the body of the __finally block, but that doesn't work when it contains any kind of Decl, which we can't double emit. This fixes that by emitting the block once and branching into a shared code region and then branching back out. llvm-svn: 228222
2015-02-05 06:37:07 +08:00
//
// CHECK: [[lpad]]
// CHECK-NEXT: landingpad
// CHECK-NEXT: cleanup
[SEH] Switch from frameaddress(0) to localaddress This should do the right thing for stack realignment prologues. llvm-svn: 241644
2015-07-08 07:23:31 +08:00
// CHECK: %[[fp:[^ ]*]] = call i8* @llvm.localaddress()
Revert "Revert 241171, 241187, 241199 (32-bit SEH)." This reverts commit r241244, but restricts SEH support to Win64. This way, Chromium builds will still fall back on TUs with SEH, and Clang developers can work on this incrementally upstream while patching this small predicate locally. It'll also make it easier to review small fixes. llvm-svn: 241533
2015-07-07 08:36:30 +08:00
// CHECK: call void @"\01?fin$0@0@use_abnormal_termination@@"({{i8( zeroext)?}} 1, i8* %[[fp]])
Implement IRGen for SEH __finally and AbnormalTermination Previously we would simply double-emit the body of the __finally block, but that doesn't work when it contains any kind of Decl, which we can't double emit. This fixes that by emitting the block once and branching into a shared code region and then branching back out. llvm-svn: 228222
2015-02-05 06:37:07 +08:00
// CHECK: resume { i8*, i32 }
Fix crash on finally blocks that don't fall through llvm-svn: 228243
2015-02-05 08:58:46 +08:00
Revert "Revert 241171, 241187, 241199 (32-bit SEH)." This reverts commit r241244, but restricts SEH support to Win64. This way, Chromium builds will still fall back on TUs with SEH, and Clang developers can work on this incrementally upstream while patching this small predicate locally. It'll also make it easier to review small fixes. llvm-svn: 241533
2015-07-07 08:36:30 +08:00
// CHECK: define internal void @"\01?fin$0@0@use_abnormal_termination@@"({{i8( zeroext)?}} %[[abnormal:abnormal_termination]], i8* %frame_pointer)
// CHECK: %[[abnormal_zext:[^ ]*]] = zext i8 %[[abnormal]] to i32
Reland r234613 (and follow-ups 234614, 234616, 234618) The frameescape intrinsic cannot be inlined, so I fixed the inliner in r234937. This should address PR23216. llvm-svn: 234942
2015-04-15 04:59:00 +08:00
// CHECK: store i32 %[[abnormal_zext]], i32* @crashed
// CHECK-NEXT: ret void
Fix crash on finally blocks that don't fall through llvm-svn: 228243
2015-02-05 08:58:46 +08:00
void noreturn_noop_finally() {
__try {
__noop();
} __finally {
abort();
}
}
// CHECK-LABEL: define void @noreturn_noop_finally()
Revert "Revert 241171, 241187, 241199 (32-bit SEH)." This reverts commit r241244, but restricts SEH support to Win64. This way, Chromium builds will still fall back on TUs with SEH, and Clang developers can work on this incrementally upstream while patching this small predicate locally. It'll also make it easier to review small fixes. llvm-svn: 241533
2015-07-07 08:36:30 +08:00
// CHECK: call void @"\01?fin$0@0@noreturn_noop_finally@@"({{.*}})
Reland r234613 (and follow-ups 234614, 234616, 234618) The frameescape intrinsic cannot be inlined, so I fixed the inliner in r234937. This should address PR23216. llvm-svn: 234942
2015-04-15 04:59:00 +08:00
// CHECK: ret void
Revert "Revert 241171, 241187, 241199 (32-bit SEH)." This reverts commit r241244, but restricts SEH support to Win64. This way, Chromium builds will still fall back on TUs with SEH, and Clang developers can work on this incrementally upstream while patching this small predicate locally. It'll also make it easier to review small fixes. llvm-svn: 241533
2015-07-07 08:36:30 +08:00
// CHECK: define internal void @"\01?fin$0@0@noreturn_noop_finally@@"({{.*}})
Fix crash on finally blocks that don't fall through llvm-svn: 228243
2015-02-05 08:58:46 +08:00
// CHECK: call void @abort()
Reland r234613 (and follow-ups 234614, 234616, 234618) The frameescape intrinsic cannot be inlined, so I fixed the inliner in r234937. This should address PR23216. llvm-svn: 234942
2015-04-15 04:59:00 +08:00
// CHECK: unreachable
Fix crash on finally blocks that don't fall through llvm-svn: 228243
2015-02-05 08:58:46 +08:00
void noreturn_finally() {
__try {
might_crash();
} __finally {
abort();
}
}
// CHECK-LABEL: define void @noreturn_finally()
// CHECK: invoke void @might_crash()
// CHECK: to label %[[cont:[^ ]*]] unwind label %[[lpad:[^ ]*]]
//
// CHECK: [[cont]]
Revert "Revert 241171, 241187, 241199 (32-bit SEH)." This reverts commit r241244, but restricts SEH support to Win64. This way, Chromium builds will still fall back on TUs with SEH, and Clang developers can work on this incrementally upstream while patching this small predicate locally. It'll also make it easier to review small fixes. llvm-svn: 241533
2015-07-07 08:36:30 +08:00
// CHECK: call void @"\01?fin$0@0@noreturn_finally@@"({{.*}})
Reland r234613 (and follow-ups 234614, 234616, 234618) The frameescape intrinsic cannot be inlined, so I fixed the inliner in r234937. This should address PR23216. llvm-svn: 234942
2015-04-15 04:59:00 +08:00
// CHECK: ret void
Fix crash on finally blocks that don't fall through llvm-svn: 228243
2015-02-05 08:58:46 +08:00
//
// CHECK: [[lpad]]
// CHECK: landingpad
// CHECK-NEXT: cleanup
Revert "Revert 241171, 241187, 241199 (32-bit SEH)." This reverts commit r241244, but restricts SEH support to Win64. This way, Chromium builds will still fall back on TUs with SEH, and Clang developers can work on this incrementally upstream while patching this small predicate locally. It'll also make it easier to review small fixes. llvm-svn: 241533
2015-07-07 08:36:30 +08:00
// CHECK: call void @"\01?fin$0@0@noreturn_finally@@"({{.*}})
Reland r234613 (and follow-ups 234614, 234616, 234618) The frameescape intrinsic cannot be inlined, so I fixed the inliner in r234937. This should address PR23216. llvm-svn: 234942
2015-04-15 04:59:00 +08:00
// CHECK: resume { i8*, i32 }
Revert "Revert 241171, 241187, 241199 (32-bit SEH)." This reverts commit r241244, but restricts SEH support to Win64. This way, Chromium builds will still fall back on TUs with SEH, and Clang developers can work on this incrementally upstream while patching this small predicate locally. It'll also make it easier to review small fixes. llvm-svn: 241533
2015-07-07 08:36:30 +08:00
// CHECK: define internal void @"\01?fin$0@0@noreturn_finally@@"({{.*}})
Reland r234613 (and follow-ups 234614, 234616, 234618) The frameescape intrinsic cannot be inlined, so I fixed the inliner in r234937. This should address PR23216. llvm-svn: 234942
2015-04-15 04:59:00 +08:00
// CHECK: call void @abort()
// CHECK: unreachable
Reland r230460 with a test fix for -Asserts builds. Original CL description: Produce less broken basic block sequences for __finally blocks. The way cleanups (such as PerformSEHFinally) get emitted is that codegen generates some initialization code, then calls the cleanup's Emit() with the insertion point set to a good place, then the cleanup is supposed to emit its stuff, and then codegen might tack in a jump or similar to where the insertion point is after the cleanup. The PerformSEHFinally cleanup tries to just stash away the block it's supposed to codegen into, and then does codegen later, into that stashed block. However, after codegen'ing the __finally block, it used to set the insertion point to the finally's continuation block (where the __finally cleanup goes when its body is completed after regular, non-exceptional control flow). That's not correct, as that block can (and generally does) already ends in a jump. Instead, remember the insertion point that was current before the __finally got emitted, and restore that. Fixes two of the crashes in PR22553. llvm-svn: 230503
2015-02-26 00:25:00 +08:00
int finally_with_return() {
__try {
return 42;
} __finally {
}
}
// CHECK-LABEL: define i32 @finally_with_return()
Revert "Revert 241171, 241187, 241199 (32-bit SEH)." This reverts commit r241244, but restricts SEH support to Win64. This way, Chromium builds will still fall back on TUs with SEH, and Clang developers can work on this incrementally upstream while patching this small predicate locally. It'll also make it easier to review small fixes. llvm-svn: 241533
2015-07-07 08:36:30 +08:00
// CHECK: call void @"\01?fin$0@0@finally_with_return@@"({{.*}})
Reland r230460 with a test fix for -Asserts builds. Original CL description: Produce less broken basic block sequences for __finally blocks. The way cleanups (such as PerformSEHFinally) get emitted is that codegen generates some initialization code, then calls the cleanup's Emit() with the insertion point set to a good place, then the cleanup is supposed to emit its stuff, and then codegen might tack in a jump or similar to where the insertion point is after the cleanup. The PerformSEHFinally cleanup tries to just stash away the block it's supposed to codegen into, and then does codegen later, into that stashed block. However, after codegen'ing the __finally block, it used to set the insertion point to the finally's continuation block (where the __finally cleanup goes when its body is completed after regular, non-exceptional control flow). That's not correct, as that block can (and generally does) already ends in a jump. Instead, remember the insertion point that was current before the __finally got emitted, and restore that. Fixes two of the crashes in PR22553. llvm-svn: 230503
2015-02-26 00:25:00 +08:00
// CHECK-NEXT: ret i32 42
Revert "Revert 241171, 241187, 241199 (32-bit SEH)." This reverts commit r241244, but restricts SEH support to Win64. This way, Chromium builds will still fall back on TUs with SEH, and Clang developers can work on this incrementally upstream while patching this small predicate locally. It'll also make it easier to review small fixes. llvm-svn: 241533
2015-07-07 08:36:30 +08:00
// CHECK: define internal void @"\01?fin$0@0@finally_with_return@@"({{.*}})
Reland r234613 (and follow-ups 234614, 234616, 234618) The frameescape intrinsic cannot be inlined, so I fixed the inliner in r234937. This should address PR23216. llvm-svn: 234942
2015-04-15 04:59:00 +08:00
// CHECK-NOT: br i1
// CHECK-NOT: br label
// CHECK: ret void
Reland r230460 with a test fix for -Asserts builds. Original CL description: Produce less broken basic block sequences for __finally blocks. The way cleanups (such as PerformSEHFinally) get emitted is that codegen generates some initialization code, then calls the cleanup's Emit() with the insertion point set to a good place, then the cleanup is supposed to emit its stuff, and then codegen might tack in a jump or similar to where the insertion point is after the cleanup. The PerformSEHFinally cleanup tries to just stash away the block it's supposed to codegen into, and then does codegen later, into that stashed block. However, after codegen'ing the __finally block, it used to set the insertion point to the finally's continuation block (where the __finally cleanup goes when its body is completed after regular, non-exceptional control flow). That's not correct, as that block can (and generally does) already ends in a jump. Instead, remember the insertion point that was current before the __finally got emitted, and restore that. Fixes two of the crashes in PR22553. llvm-svn: 230503
2015-02-26 00:25:00 +08:00
int nested___finally___finally() {
__try {
__try {
} __finally {
return 1;
}
} __finally {
// Intentionally no return here.
}
return 0;
}
Reland r234613 (and follow-ups 234614, 234616, 234618) The frameescape intrinsic cannot be inlined, so I fixed the inliner in r234937. This should address PR23216. llvm-svn: 234942
2015-04-15 04:59:00 +08:00
Reland r230460 with a test fix for -Asserts builds. Original CL description: Produce less broken basic block sequences for __finally blocks. The way cleanups (such as PerformSEHFinally) get emitted is that codegen generates some initialization code, then calls the cleanup's Emit() with the insertion point set to a good place, then the cleanup is supposed to emit its stuff, and then codegen might tack in a jump or similar to where the insertion point is after the cleanup. The PerformSEHFinally cleanup tries to just stash away the block it's supposed to codegen into, and then does codegen later, into that stashed block. However, after codegen'ing the __finally block, it used to set the insertion point to the finally's continuation block (where the __finally cleanup goes when its body is completed after regular, non-exceptional control flow). That's not correct, as that block can (and generally does) already ends in a jump. Instead, remember the insertion point that was current before the __finally got emitted, and restore that. Fixes two of the crashes in PR22553. llvm-svn: 230503
2015-02-26 00:25:00 +08:00
// CHECK-LABEL: define i32 @nested___finally___finally
Revert "Revert 241171, 241187, 241199 (32-bit SEH)." This reverts commit r241244, but restricts SEH support to Win64. This way, Chromium builds will still fall back on TUs with SEH, and Clang developers can work on this incrementally upstream while patching this small predicate locally. It'll also make it easier to review small fixes. llvm-svn: 241533
2015-07-07 08:36:30 +08:00
// CHECK: invoke void @"\01?fin$1@0@nested___finally___finally@@"({{.*}})
Reland r234613 (and follow-ups 234614, 234616, 234618) The frameescape intrinsic cannot be inlined, so I fixed the inliner in r234937. This should address PR23216. llvm-svn: 234942
2015-04-15 04:59:00 +08:00
// CHECK: to label %[[outercont:[^ ]*]] unwind label %[[lpad:[^ ]*]]
Reland r230460 with a test fix for -Asserts builds. Original CL description: Produce less broken basic block sequences for __finally blocks. The way cleanups (such as PerformSEHFinally) get emitted is that codegen generates some initialization code, then calls the cleanup's Emit() with the insertion point set to a good place, then the cleanup is supposed to emit its stuff, and then codegen might tack in a jump or similar to where the insertion point is after the cleanup. The PerformSEHFinally cleanup tries to just stash away the block it's supposed to codegen into, and then does codegen later, into that stashed block. However, after codegen'ing the __finally block, it used to set the insertion point to the finally's continuation block (where the __finally cleanup goes when its body is completed after regular, non-exceptional control flow). That's not correct, as that block can (and generally does) already ends in a jump. Instead, remember the insertion point that was current before the __finally got emitted, and restore that. Fixes two of the crashes in PR22553. llvm-svn: 230503
2015-02-26 00:25:00 +08:00
//
Reland r234613 (and follow-ups 234614, 234616, 234618) The frameescape intrinsic cannot be inlined, so I fixed the inliner in r234937. This should address PR23216. llvm-svn: 234942
2015-04-15 04:59:00 +08:00
// CHECK: [[outercont]]
Revert "Revert 241171, 241187, 241199 (32-bit SEH)." This reverts commit r241244, but restricts SEH support to Win64. This way, Chromium builds will still fall back on TUs with SEH, and Clang developers can work on this incrementally upstream while patching this small predicate locally. It'll also make it easier to review small fixes. llvm-svn: 241533
2015-07-07 08:36:30 +08:00
// CHECK: call void @"\01?fin$0@0@nested___finally___finally@@"({{.*}})
Reland r234613 (and follow-ups 234614, 234616, 234618) The frameescape intrinsic cannot be inlined, so I fixed the inliner in r234937. This should address PR23216. llvm-svn: 234942
2015-04-15 04:59:00 +08:00
// CHECK-NEXT: ret i32 0
Don't crash on leaving nested __finally blocks through an EH edge. The __finally emission block tries to be clever by removing unused continuation edges if there's an unconditional jump out of the __finally block. With exception edges, the EH continuation edge isn't always unused though and we'd crash in a few places. Just don't be clever. That makes the IR for __finally blocks a bit longer in some cases (hence small and behavior-preserving changes to existing tests), but it makes no difference in general and it fixes the last crash from PR22553. http://reviews.llvm.org/D7918 llvm-svn: 230697
2015-02-27 06:34:33 +08:00
//
Reland r234613 (and follow-ups 234614, 234616, 234618) The frameescape intrinsic cannot be inlined, so I fixed the inliner in r234937. This should address PR23216. llvm-svn: 234942
2015-04-15 04:59:00 +08:00
// CHECK: [[lpad]]
// CHECK-NEXT: landingpad
// CHECK-NEXT: cleanup
Revert "Revert 241171, 241187, 241199 (32-bit SEH)." This reverts commit r241244, but restricts SEH support to Win64. This way, Chromium builds will still fall back on TUs with SEH, and Clang developers can work on this incrementally upstream while patching this small predicate locally. It'll also make it easier to review small fixes. llvm-svn: 241533
2015-07-07 08:36:30 +08:00
// CHECK: call void @"\01?fin$0@0@nested___finally___finally@@"({{.*}})
Reland r234613 (and follow-ups 234614, 234616, 234618) The frameescape intrinsic cannot be inlined, so I fixed the inliner in r234937. This should address PR23216. llvm-svn: 234942
2015-04-15 04:59:00 +08:00
Revert "Revert 241171, 241187, 241199 (32-bit SEH)." This reverts commit r241244, but restricts SEH support to Win64. This way, Chromium builds will still fall back on TUs with SEH, and Clang developers can work on this incrementally upstream while patching this small predicate locally. It'll also make it easier to review small fixes. llvm-svn: 241533
2015-07-07 08:36:30 +08:00
// CHECK-LABEL: define internal void @"\01?fin$0@0@nested___finally___finally@@"({{.*}})
Reland r234613 (and follow-ups 234614, 234616, 234618) The frameescape intrinsic cannot be inlined, so I fixed the inliner in r234937. This should address PR23216. llvm-svn: 234942
2015-04-15 04:59:00 +08:00
// CHECK: ret void
Revert "Revert 241171, 241187, 241199 (32-bit SEH)." This reverts commit r241244, but restricts SEH support to Win64. This way, Chromium builds will still fall back on TUs with SEH, and Clang developers can work on this incrementally upstream while patching this small predicate locally. It'll also make it easier to review small fixes. llvm-svn: 241533
2015-07-07 08:36:30 +08:00
// CHECK-LABEL: define internal void @"\01?fin$1@0@nested___finally___finally@@"({{.*}})
Reland r234613 (and follow-ups 234614, 234616, 234618) The frameescape intrinsic cannot be inlined, so I fixed the inliner in r234937. This should address PR23216. llvm-svn: 234942
2015-04-15 04:59:00 +08:00
// CHECK: unreachable
Don't crash on leaving nested __finally blocks through an EH edge. The __finally emission block tries to be clever by removing unused continuation edges if there's an unconditional jump out of the __finally block. With exception edges, the EH continuation edge isn't always unused though and we'd crash in a few places. Just don't be clever. That makes the IR for __finally blocks a bit longer in some cases (hence small and behavior-preserving changes to existing tests), but it makes no difference in general and it fixes the last crash from PR22553. http://reviews.llvm.org/D7918 llvm-svn: 230697
2015-02-27 06:34:33 +08:00
int nested___finally___finally_with_eh_edge() {
__try {
__try {
might_crash();
} __finally {
return 899;
}
} __finally {
// Intentionally no return here.
}
return 912;
}
// CHECK-LABEL: define i32 @nested___finally___finally_with_eh_edge
Reland r234613 (and follow-ups 234614, 234616, 234618) The frameescape intrinsic cannot be inlined, so I fixed the inliner in r234937. This should address PR23216. llvm-svn: 234942
2015-04-15 04:59:00 +08:00
// CHECK: invoke void @might_crash()
// CHECK-NEXT: to label %[[invokecont:[^ ]*]] unwind label %[[lpad1:[^ ]*]]
Revert r234613 (and follow-ups 234614, 234616, 234618), it caused PR23216. llvm-svn: 234786
2015-04-14 04:01:20 +08:00
//
Reland r234613 (and follow-ups 234614, 234616, 234618) The frameescape intrinsic cannot be inlined, so I fixed the inliner in r234937. This should address PR23216. llvm-svn: 234942
2015-04-15 04:59:00 +08:00
// [[invokecont]]
Revert "Revert 241171, 241187, 241199 (32-bit SEH)." This reverts commit r241244, but restricts SEH support to Win64. This way, Chromium builds will still fall back on TUs with SEH, and Clang developers can work on this incrementally upstream while patching this small predicate locally. It'll also make it easier to review small fixes. llvm-svn: 241533
2015-07-07 08:36:30 +08:00
// CHECK: invoke void @"\01?fin$1@0@nested___finally___finally_with_eh_edge@@"({{.*}})
Reland r234613 (and follow-ups 234614, 234616, 234618) The frameescape intrinsic cannot be inlined, so I fixed the inliner in r234937. This should address PR23216. llvm-svn: 234942
2015-04-15 04:59:00 +08:00
// CHECK: to label %[[outercont:[^ ]*]] unwind label %[[lpad2:[^ ]*]]
Revert r234613 (and follow-ups 234614, 234616, 234618), it caused PR23216. llvm-svn: 234789
2015-04-14 04:04:22 +08:00
//
Reland r234613 (and follow-ups 234614, 234616, 234618) The frameescape intrinsic cannot be inlined, so I fixed the inliner in r234937. This should address PR23216. llvm-svn: 234942
2015-04-15 04:59:00 +08:00
// CHECK: [[outercont]]
Revert "Revert 241171, 241187, 241199 (32-bit SEH)." This reverts commit r241244, but restricts SEH support to Win64. This way, Chromium builds will still fall back on TUs with SEH, and Clang developers can work on this incrementally upstream while patching this small predicate locally. It'll also make it easier to review small fixes. llvm-svn: 241533
2015-07-07 08:36:30 +08:00
// CHECK: call void @"\01?fin$0@0@nested___finally___finally_with_eh_edge@@"({{.*}})
Reland r234613 (and follow-ups 234614, 234616, 234618) The frameescape intrinsic cannot be inlined, so I fixed the inliner in r234937. This should address PR23216. llvm-svn: 234942
2015-04-15 04:59:00 +08:00
// CHECK-NEXT: ret i32 912
Revert r234613 (and follow-ups 234614, 234616, 234618), it caused PR23216. llvm-svn: 234789
2015-04-14 04:04:22 +08:00
//
Reland r234613 (and follow-ups 234614, 234616, 234618) The frameescape intrinsic cannot be inlined, so I fixed the inliner in r234937. This should address PR23216. llvm-svn: 234942
2015-04-15 04:59:00 +08:00
// CHECK: [[lpad1]]
Revert r234786, it contained a bunch of stuff I did not mean to commit. llvm-svn: 234787
2015-04-14 04:03:03 +08:00
// CHECK-NEXT: landingpad
// CHECK-NEXT: cleanup
Revert "Revert 241171, 241187, 241199 (32-bit SEH)." This reverts commit r241244, but restricts SEH support to Win64. This way, Chromium builds will still fall back on TUs with SEH, and Clang developers can work on this incrementally upstream while patching this small predicate locally. It'll also make it easier to review small fixes. llvm-svn: 241533
2015-07-07 08:36:30 +08:00
// CHECK: invoke void @"\01?fin$1@0@nested___finally___finally_with_eh_edge@@"({{.*}})
Reland r234613 (and follow-ups 234614, 234616, 234618) The frameescape intrinsic cannot be inlined, so I fixed the inliner in r234937. This should address PR23216. llvm-svn: 234942
2015-04-15 04:59:00 +08:00
// CHECK: to label %[[outercont:[^ ]*]] unwind label %[[lpad2]]
Revert r234613 (and follow-ups 234614, 234616, 234618), it caused PR23216. llvm-svn: 234789
2015-04-14 04:04:22 +08:00
//
Reland r234613 (and follow-ups 234614, 234616, 234618) The frameescape intrinsic cannot be inlined, so I fixed the inliner in r234937. This should address PR23216. llvm-svn: 234942
2015-04-15 04:59:00 +08:00
// CHECK: [[lpad2]]
// CHECK-NEXT: landingpad
// CHECK-NEXT: cleanup
Revert "Revert 241171, 241187, 241199 (32-bit SEH)." This reverts commit r241244, but restricts SEH support to Win64. This way, Chromium builds will still fall back on TUs with SEH, and Clang developers can work on this incrementally upstream while patching this small predicate locally. It'll also make it easier to review small fixes. llvm-svn: 241533
2015-07-07 08:36:30 +08:00
// CHECK: call void @"\01?fin$0@0@nested___finally___finally_with_eh_edge@@"({{.*}})
Don't crash on leaving nested __finally blocks through an EH edge. The __finally emission block tries to be clever by removing unused continuation edges if there's an unconditional jump out of the __finally block. With exception edges, the EH continuation edge isn't always unused though and we'd crash in a few places. Just don't be clever. That makes the IR for __finally blocks a bit longer in some cases (hence small and behavior-preserving changes to existing tests), but it makes no difference in general and it fixes the last crash from PR22553. http://reviews.llvm.org/D7918 llvm-svn: 230697
2015-02-27 06:34:33 +08:00
// CHECK: resume
Reland r234613 (and follow-ups 234614, 234616, 234618) The frameescape intrinsic cannot be inlined, so I fixed the inliner in r234937. This should address PR23216. llvm-svn: 234942
2015-04-15 04:59:00 +08:00
Revert "Revert 241171, 241187, 241199 (32-bit SEH)." This reverts commit r241244, but restricts SEH support to Win64. This way, Chromium builds will still fall back on TUs with SEH, and Clang developers can work on this incrementally upstream while patching this small predicate locally. It'll also make it easier to review small fixes. llvm-svn: 241533
2015-07-07 08:36:30 +08:00
// CHECK-LABEL: define internal void @"\01?fin$0@0@nested___finally___finally_with_eh_edge@@"({{.*}})
Reland r234613 (and follow-ups 234614, 234616, 234618) The frameescape intrinsic cannot be inlined, so I fixed the inliner in r234937. This should address PR23216. llvm-svn: 234942
2015-04-15 04:59:00 +08:00
// CHECK: ret void
Revert "Revert 241171, 241187, 241199 (32-bit SEH)." This reverts commit r241244, but restricts SEH support to Win64. This way, Chromium builds will still fall back on TUs with SEH, and Clang developers can work on this incrementally upstream while patching this small predicate locally. It'll also make it easier to review small fixes. llvm-svn: 241533
2015-07-07 08:36:30 +08:00
// CHECK-LABEL: define internal void @"\01?fin$1@0@nested___finally___finally_with_eh_edge@@"({{.*}})
Reland r234613 (and follow-ups 234614, 234616, 234618) The frameescape intrinsic cannot be inlined, so I fixed the inliner in r234937. This should address PR23216. llvm-svn: 234942
2015-04-15 04:59:00 +08:00
// CHECK: unreachable