user0/rust
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Always use llvm.used for coverage symbols

Browse source 
This follows what clang does in CoverageMappingGen. Using just
llvm.compiler.used is insufficient at least for MSVC targets.
This commit is contained in:
Nikita Popov 2021-08-20 21:13:18 +02:00
parent f3ae726f30
commit 306259c645
6 changed files with 59 additions and 35 deletions
Download patch file Download diff file Expand all files Collapse all files

5
src/test/run-make-fulldeps/coverage-llvmir/filecheck.testprog.txt
View file

@ -28,11 +28,8 @@ CHECK-SAME: section "[[INSTR_PROF_DATA]]"{{.*}}, align 8
CHECK: @__llvm_prf_nm = private constant
CHECK-SAME: section "[[INSTR_PROF_NAME]]", align 1
CHECK: @llvm.compiler.used = appending global
CHECK-SAME: i8* bitcast ({ {{.*}} }* @__llvm_coverage_mapping to i8*)
WINDOWS-SAME: i8* bitcast (i32 ()* @__llvm_profile_runtime_user to i8*)
CHECK-SAME: section "llvm.metadata"
CHECK: @llvm.used = appending global
CHECK-SAME: i8* bitcast ({ {{.*}} }* @__llvm_coverage_mapping to i8*)
CHECK-SAME: i8* getelementptr inbounds ({{.*}}* @__llvm_prf_nm, i32 0, i32 0)
CHECK-SAME: section "llvm.metadata"