llvm-project/clang/test/CoverageMapping/ir.c

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

32 lines
2.3 KiB
C
Raw Normal View History

// Check the data structures emitted by coverage mapping
// RUN: %clang_cc1 -mllvm -emptyline-comment-coverage=false -triple x86_64-apple-macosx10.9 -main-file-name ir.c %s -o - -emit-llvm -fprofile-instrument=clang -fcoverage-mapping -mllvm -enable-name-compression=false | FileCheck %s -check-prefixes=COMMON,DARWIN
// RUN: %clang_cc1 -mllvm -emptyline-comment-coverage=false -triple x86_64--windows-msvc -main-file-name ir.c %s -o - -emit-llvm -fprofile-instrument=clang -fcoverage-mapping -mllvm -enable-name-compression=false | FileCheck %s -check-prefixes=COMMON,WINDOWS
Reland: [Coverage] Revise format to reduce binary size Try again with an up-to-date version of D69471 (99317124 was a stale revision). --- Revise the coverage mapping format to reduce binary size by: 1. Naming function records and marking them `linkonce_odr`, and 2. Compressing filenames. This shrinks the size of llc's coverage segment by 82% (334MB -> 62MB) and speeds up end-to-end single-threaded report generation by 10%. For reference the compressed name data in llc is 81MB (__llvm_prf_names). Rationale for changes to the format: - With the current format, most coverage function records are discarded. E.g., more than 97% of the records in llc are *duplicate* placeholders for functions visible-but-not-used in TUs. Placeholders *are* used to show under-covered functions, but duplicate placeholders waste space. - We reached general consensus about giving (1) a try at the 2017 code coverage BoF [1]. The thinking was that using `linkonce_odr` to merge duplicates is simpler than alternatives like teaching build systems about a coverage-aware database/module/etc on the side. - Revising the format is expensive due to the backwards compatibility requirement, so we might as well compress filenames while we're at it. This shrinks the encoded filenames in llc by 86% (12MB -> 1.6MB). See CoverageMappingFormat.rst for the details on what exactly has changed. Fixes PR34533 [2], hopefully. [1] http://lists.llvm.org/pipermail/llvm-dev/2017-October/118428.html [2] https://bugs.llvm.org/show_bug.cgi?id=34533 Differential Revision: https://reviews.llvm.org/D69471
2019-10-22 02:48:38 +08:00
static inline void unused() {}
Reland: [Coverage] Revise format to reduce binary size Try again with an up-to-date version of D69471 (99317124 was a stale revision). --- Revise the coverage mapping format to reduce binary size by: 1. Naming function records and marking them `linkonce_odr`, and 2. Compressing filenames. This shrinks the size of llc's coverage segment by 82% (334MB -> 62MB) and speeds up end-to-end single-threaded report generation by 10%. For reference the compressed name data in llc is 81MB (__llvm_prf_names). Rationale for changes to the format: - With the current format, most coverage function records are discarded. E.g., more than 97% of the records in llc are *duplicate* placeholders for functions visible-but-not-used in TUs. Placeholders *are* used to show under-covered functions, but duplicate placeholders waste space. - We reached general consensus about giving (1) a try at the 2017 code coverage BoF [1]. The thinking was that using `linkonce_odr` to merge duplicates is simpler than alternatives like teaching build systems about a coverage-aware database/module/etc on the side. - Revising the format is expensive due to the backwards compatibility requirement, so we might as well compress filenames while we're at it. This shrinks the encoded filenames in llc by 86% (12MB -> 1.6MB). See CoverageMappingFormat.rst for the details on what exactly has changed. Fixes PR34533 [2], hopefully. [1] http://lists.llvm.org/pipermail/llvm-dev/2017-October/118428.html [2] https://bugs.llvm.org/show_bug.cgi?id=34533 Differential Revision: https://reviews.llvm.org/D69471
2019-10-22 02:48:38 +08:00
void foo(void) {}
int main(void) {
foo();
return 0;
}
Reland: [Coverage] Revise format to reduce binary size Try again with an up-to-date version of D69471 (99317124 was a stale revision). --- Revise the coverage mapping format to reduce binary size by: 1. Naming function records and marking them `linkonce_odr`, and 2. Compressing filenames. This shrinks the size of llc's coverage segment by 82% (334MB -> 62MB) and speeds up end-to-end single-threaded report generation by 10%. For reference the compressed name data in llc is 81MB (__llvm_prf_names). Rationale for changes to the format: - With the current format, most coverage function records are discarded. E.g., more than 97% of the records in llc are *duplicate* placeholders for functions visible-but-not-used in TUs. Placeholders *are* used to show under-covered functions, but duplicate placeholders waste space. - We reached general consensus about giving (1) a try at the 2017 code coverage BoF [1]. The thinking was that using `linkonce_odr` to merge duplicates is simpler than alternatives like teaching build systems about a coverage-aware database/module/etc on the side. - Revising the format is expensive due to the backwards compatibility requirement, so we might as well compress filenames while we're at it. This shrinks the encoded filenames in llc by 86% (12MB -> 1.6MB). See CoverageMappingFormat.rst for the details on what exactly has changed. Fixes PR34533 [2], hopefully. [1] http://lists.llvm.org/pipermail/llvm-dev/2017-October/118428.html [2] https://bugs.llvm.org/show_bug.cgi?id=34533 Differential Revision: https://reviews.llvm.org/D69471
2019-10-22 02:48:38 +08:00
// Check the function records. Two of the record names should come in the 'used'
// flavor, and one should not.
// DARWIN: [[FuncRecord1:@__covrec_[0-9A-F]+u]] = linkonce_odr hidden constant <{ i64, i32, i64, i64, [{{.*}} x i8] }> <{ {{.*}} }>, section "__LLVM_COV,__llvm_covfun", align 8
// DARWIN: [[FuncRecord2:@__covrec_[0-9A-F]+u]] = linkonce_odr hidden constant <{ i64, i32, i64, i64, [{{.*}} x i8] }> <{ {{.*}} }>, section "__LLVM_COV,__llvm_covfun", align 8
// DARWIN: [[FuncRecord3:@__covrec_[0-9A-F]+]] = linkonce_odr hidden constant <{ i64, i32, i64, i64, [{{.*}} x i8] }> <{ {{.*}} }>, section "__LLVM_COV,__llvm_covfun", align 8
// DARWIN: @__llvm_coverage_mapping = private constant { { i32, i32, i32, i32 }, [{{.*}} x i8] } { {{.*}} }, section "__LLVM_COV,__llvm_covmap", align 8
// WINDOWS: [[FuncRecord1:@__covrec_[0-9A-F]+u]] = linkonce_odr hidden constant <{ i64, i32, i64, i64, [{{.*}} x i8] }> <{ {{.*}} }>, section ".lcovfun$M", comdat, align 8
// WINDOWS: [[FuncRecord2:@__covrec_[0-9A-F]+u]] = linkonce_odr hidden constant <{ i64, i32, i64, i64, [{{.*}} x i8] }> <{ {{.*}} }>, section ".lcovfun$M", comdat, align 8
// WINDOWS: [[FuncRecord3:@__covrec_[0-9A-F]+]] = linkonce_odr hidden constant <{ i64, i32, i64, i64, [{{.*}} x i8] }> <{ {{.*}} }>, section ".lcovfun$M", comdat, align 8
// WINDOWS: @__llvm_coverage_mapping = private constant { { i32, i32, i32, i32 }, [{{.*}} x i8] } { {{.*}} }, section ".lcovmap$M", align 8
// COMMON: @llvm.used = appending global [{{.*}} x i8*]
// COMMON-SAME: [[FuncRecord1]]
// COMMON-SAME: [[FuncRecord2]]
// COMMON-SAME: [[FuncRecord3]]
// COMMON-SAME: @__llvm_coverage_mapping