maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/llvm/test/SafepointIRVerifier/constant-bases.ll

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

71 lines
3.5 KiB
LLVM
Raw Normal View History

[SafepointIRVerifier] Add verifier pass for finding GC relocation bugs Original Patch and summary by Philip Reames. RewriteStatepointsForGC tries to rewrite a function in a manner where the optimizer can't end up using a pointer value after it might have been relocated by a safepoint. This pass checks the invariant that RSForGC is supposed to establish and that (if we constructed semantics correctly) later passes must preserve. This has been a really useful diagnostic tool when initially developing the rewriting scheme and has found numerous bugs. Differential Revision: https://reviews.llvm.org/D15940 Reviewed by: swaroop.sridhar, mjacob Subscribers: llvm-commits llvm-svn: 307112
2017-07-05 09:16:29 +08:00
; RUN: opt -safepoint-ir-verifier-print-only -verify-safepoint-ir -S %s 2>&1 | FileCheck %s
define i8 addrspace(1)* @test1(i64 %arg) gc "statepoint-example" {
; CHECK: No illegal uses found by SafepointIRVerifier in: test1
entry:
[Tests] Remove deopt operands from SafepointIRVerfier tests This linter has nothing to do with deopt, and the operands had clearly been copied blindly from another source. Rather than migrate to deopt operand bundle, let's just simplify the tests.
2020-05-29 05:34:47 +08:00
%safepoint_token = call token (i64, i32, void ()*, i32, i32, ...) @llvm.experimental.gc.statepoint.p0f_isVoidf(i64 0, i32 0, void ()* undef, i32 0, i32 0, i32 0, i32 0)
[SafepointIRVerifier] Add verifier pass for finding GC relocation bugs Original Patch and summary by Philip Reames. RewriteStatepointsForGC tries to rewrite a function in a manner where the optimizer can't end up using a pointer value after it might have been relocated by a safepoint. This pass checks the invariant that RSForGC is supposed to establish and that (if we constructed semantics correctly) later passes must preserve. This has been a really useful diagnostic tool when initially developing the rewriting scheme and has found numerous bugs. Differential Revision: https://reviews.llvm.org/D15940 Reviewed by: swaroop.sridhar, mjacob Subscribers: llvm-commits llvm-svn: 307112
2017-07-05 09:16:29 +08:00
ret i8 addrspace(1)* null
}
define i8 addrspace(1)* @test2(i64 %arg) gc "statepoint-example" {
; CHECK: No illegal uses found by SafepointIRVerifier in: test2
entry:
%load_addr = getelementptr i8, i8 addrspace(1)* inttoptr (i64 15 to i8 addrspace(1)*), i64 %arg
[Tests] Remove deopt operands from SafepointIRVerfier tests This linter has nothing to do with deopt, and the operands had clearly been copied blindly from another source. Rather than migrate to deopt operand bundle, let's just simplify the tests.
2020-05-29 05:34:47 +08:00
%safepoint_token = call token (i64, i32, void ()*, i32, i32, ...) @llvm.experimental.gc.statepoint.p0f_isVoidf(i64 0, i32 0, void ()* undef, i32 0, i32 0, i32 0, i32 0)
[SafepointIRVerifier] Add verifier pass for finding GC relocation bugs Original Patch and summary by Philip Reames. RewriteStatepointsForGC tries to rewrite a function in a manner where the optimizer can't end up using a pointer value after it might have been relocated by a safepoint. This pass checks the invariant that RSForGC is supposed to establish and that (if we constructed semantics correctly) later passes must preserve. This has been a really useful diagnostic tool when initially developing the rewriting scheme and has found numerous bugs. Differential Revision: https://reviews.llvm.org/D15940 Reviewed by: swaroop.sridhar, mjacob Subscribers: llvm-commits llvm-svn: 307112
2017-07-05 09:16:29 +08:00
ret i8 addrspace(1)* %load_addr
}
define i8 addrspace(1)* @test3(i64 %arg) gc "statepoint-example" {
; CHECK: No illegal uses found by SafepointIRVerifier in: test3
entry:
%load_addr = getelementptr i32, i32 addrspace(1)* inttoptr (i64 15 to i32 addrspace(1)*), i64 %arg
%load_addr.cast = bitcast i32 addrspace(1)* %load_addr to i8 addrspace(1)*
[Tests] Remove deopt operands from SafepointIRVerfier tests This linter has nothing to do with deopt, and the operands had clearly been copied blindly from another source. Rather than migrate to deopt operand bundle, let's just simplify the tests.
2020-05-29 05:34:47 +08:00
%safepoint_token = call token (i64, i32, void ()*, i32, i32, ...) @llvm.experimental.gc.statepoint.p0f_isVoidf(i64 0, i32 0, void ()* undef, i32 0, i32 0, i32 0, i32 0)
[SafepointIRVerifier] Add verifier pass for finding GC relocation bugs Original Patch and summary by Philip Reames. RewriteStatepointsForGC tries to rewrite a function in a manner where the optimizer can't end up using a pointer value after it might have been relocated by a safepoint. This pass checks the invariant that RSForGC is supposed to establish and that (if we constructed semantics correctly) later passes must preserve. This has been a really useful diagnostic tool when initially developing the rewriting scheme and has found numerous bugs. Differential Revision: https://reviews.llvm.org/D15940 Reviewed by: swaroop.sridhar, mjacob Subscribers: llvm-commits llvm-svn: 307112
2017-07-05 09:16:29 +08:00
ret i8 addrspace(1)* %load_addr.cast
}
define i8 addrspace(1)* @test4(i64 %arg, i1 %cond) gc "statepoint-example" {
; CHECK: No illegal uses found by SafepointIRVerifier in: test4
entry:
%load_addr.1 = getelementptr i8, i8 addrspace(1)* inttoptr (i64 15 to i8 addrspace(1)*), i64 %arg
br i1 %cond, label %split, label %join
split:
%load_addr.2 = getelementptr i8, i8 addrspace(1)* inttoptr (i64 30 to i8 addrspace(1)*), i64 %arg
br label %join
join:
%load_addr = phi i8 addrspace(1)* [%load_addr.1, %entry], [%load_addr.2, %split]
[Tests] Remove deopt operands from SafepointIRVerfier tests This linter has nothing to do with deopt, and the operands had clearly been copied blindly from another source. Rather than migrate to deopt operand bundle, let's just simplify the tests.
2020-05-29 05:34:47 +08:00
%safepoint_token = call token (i64, i32, void ()*, i32, i32, ...) @llvm.experimental.gc.statepoint.p0f_isVoidf(i64 0, i32 0, void ()* undef, i32 0, i32 0, i32 0, i32 0)
[SafepointIRVerifier] Add verifier pass for finding GC relocation bugs Original Patch and summary by Philip Reames. RewriteStatepointsForGC tries to rewrite a function in a manner where the optimizer can't end up using a pointer value after it might have been relocated by a safepoint. This pass checks the invariant that RSForGC is supposed to establish and that (if we constructed semantics correctly) later passes must preserve. This has been a really useful diagnostic tool when initially developing the rewriting scheme and has found numerous bugs. Differential Revision: https://reviews.llvm.org/D15940 Reviewed by: swaroop.sridhar, mjacob Subscribers: llvm-commits llvm-svn: 307112
2017-07-05 09:16:29 +08:00
ret i8 addrspace(1)* %load_addr
}
define i8 addrspace(1)* @test5(i64 %arg, i1 %cond) gc "statepoint-example" {
; CHECK: No illegal uses found by SafepointIRVerifier in: test5
entry:
%load_addr.1 = getelementptr i8, i8 addrspace(1)* inttoptr (i64 15 to i8 addrspace(1)*), i64 %arg
%load_addr.2 = getelementptr i8, i8 addrspace(1)* inttoptr (i64 30 to i8 addrspace(1)*), i64 %arg
%load_addr = select i1 %cond, i8 addrspace(1)* %load_addr.1, i8 addrspace(1)* %load_addr.2
[Tests] Remove deopt operands from SafepointIRVerfier tests This linter has nothing to do with deopt, and the operands had clearly been copied blindly from another source. Rather than migrate to deopt operand bundle, let's just simplify the tests.
2020-05-29 05:34:47 +08:00
%safepoint_token = call token (i64, i32, void ()*, i32, i32, ...) @llvm.experimental.gc.statepoint.p0f_isVoidf(i64 0, i32 0, void ()* undef, i32 0, i32 0, i32 0, i32 0)
[SafepointIRVerifier] Add verifier pass for finding GC relocation bugs Original Patch and summary by Philip Reames. RewriteStatepointsForGC tries to rewrite a function in a manner where the optimizer can't end up using a pointer value after it might have been relocated by a safepoint. This pass checks the invariant that RSForGC is supposed to establish and that (if we constructed semantics correctly) later passes must preserve. This has been a really useful diagnostic tool when initially developing the rewriting scheme and has found numerous bugs. Differential Revision: https://reviews.llvm.org/D15940 Reviewed by: swaroop.sridhar, mjacob Subscribers: llvm-commits llvm-svn: 307112
2017-07-05 09:16:29 +08:00
ret i8 addrspace(1)* %load_addr
}
define i8 addrspace(1)* @test6(i64 %arg, i1 %cond, i8 addrspace(1)* %base) gc "statepoint-example" {
; CHECK-LABEL: Verifying gc pointers in function: test6
; CHECK: Illegal use of unrelocated value found!
entry:
%load_addr.1 = getelementptr i8, i8 addrspace(1)* %base, i64 %arg
br i1 %cond, label %split, label %join
split:
%load_addr.2 = getelementptr i8, i8 addrspace(1)* inttoptr (i64 30 to i8 addrspace(1)*), i64 %arg
br label %join
join:
%load_addr = phi i8 addrspace(1)* [%load_addr.1, %entry], [%load_addr.2, %split]
[Tests] Remove deopt operands from SafepointIRVerfier tests This linter has nothing to do with deopt, and the operands had clearly been copied blindly from another source. Rather than migrate to deopt operand bundle, let's just simplify the tests.
2020-05-29 05:34:47 +08:00
%safepoint_token = call token (i64, i32, void ()*, i32, i32, ...) @llvm.experimental.gc.statepoint.p0f_isVoidf(i64 0, i32 0, void ()* undef, i32 0, i32 0, i32 0, i32 0)
[SafepointIRVerifier] Add verifier pass for finding GC relocation bugs Original Patch and summary by Philip Reames. RewriteStatepointsForGC tries to rewrite a function in a manner where the optimizer can't end up using a pointer value after it might have been relocated by a safepoint. This pass checks the invariant that RSForGC is supposed to establish and that (if we constructed semantics correctly) later passes must preserve. This has been a really useful diagnostic tool when initially developing the rewriting scheme and has found numerous bugs. Differential Revision: https://reviews.llvm.org/D15940 Reviewed by: swaroop.sridhar, mjacob Subscribers: llvm-commits llvm-svn: 307112
2017-07-05 09:16:29 +08:00
ret i8 addrspace(1)* %load_addr
}
declare token @llvm.experimental.gc.statepoint.p0f_isVoidf(i64, i32, void ()*, i32, i32, ...)