[LLVMdev] GCC compatibility code coverage issue .

[Justin Bogner]

Umesh Kalappa <umesh.kalappa0 at gmail.com> writes:
> Hi Justin ,
>
> Thank you for the confirmation and we would like to know that ,going
> forward the clang has the support the gcc gcov format or use the
> -fprofile-instr-generate -fcoverage-mapping  and get ride of gcov
> format .

Going forward, the -fprofile-instr-generate -fcoverage-mapping (which
I'll refer to as instrprof) work will continue to be improved and built
upon. The gcov support in llvm isn't going away, but it isn't likely to
be improved or changed much at this point.

> We are planing to customize the clang code coverage for embedded world
> ,before we start tweaking the gcov / -fprofile-instr-generate
> code-base ,we would like to take feedback from the group .
>
> Did -fprofile-instr-generate -fcoverage-mapping better than(optimized
> instrumented) gcov ? ,if so any benchmark reference will be
> appreciated.

I'm not sure what you mean by better. The instrprof approach is more
accurate and has significantly lower runtime overhead than the gcov
support in llvm. It is also quite a bit simpler and more flexible in
terms of data collection. It does currently have a pretty high overhead
in terms of binary size, especially for coverage.

> Thank you and awaiting for your suggestions.
> ~Umesh
>
> On Wed, May 13, 2015 at 1:17 PM, Justin Bogner <mail at justinbogner.com> wrote:
>> Yes, the fix for this is in the 3.6.1 release candidate.
>>
>>
>> On Wednesday, May 13, 2015, Umesh Kalappa <umesh.kalappa0 at gmail.com> wrote:
>>>
>>> Hi Justin,
>>>
>>> How  you  doing there ?
>>>
>>> Do you able to fetch the changes w.r.t PR22436 to 4.6.1.
>>>
>>> Thank you
>>> ~Umesh
>>>
>>> On Wed, Apr 29, 2015 at 7:42 PM, Justin Bogner <mail at justinbogner.com>
>>> wrote:
>>> > Oh, sorry, this is http://llvm.org/PR22436 which is fixed in trunk, but
>>> > not
>>> > in 3.6.0. I'll double-check / see about getting the fix in 3.6.1.
>>> >
>>> > On Tuesday, April 28, 2015, Umesh Kalappa <umesh.kalappa0 at gmail.com>
>>> > wrote:
>>> >>
>>> >> Hi Justin ,
>>> >>
>>> >> Thank you for the reply,but without version too ,the llvm-cov fails i.e
>>> >>
>>> >> bash-4.1$ clang -fprofile-arcs -ftest-coverage test.c
>>> >> bash-4.1$ ./a.out
>>> >> bash-4.1$ llvm-cov gcov test.c
>>> >> Unexpected number of edges (in main).
>>> >> Invalid .gcda File!
>>> >>
>>> >> bash-4.1$ clang -g -O0 -fprofile-arcs -ftest-coverage -Xclang
>>> >> -coverage-cfg-checksum -Xclang -coverage-no-function-names-in-data
>>> >> test.c
>>> >> bash-4.1$ ./a.out
>>> >> bash-4.1$ llvm-cov gcov test.c
>>> >> Unexpected end of memory buffer: 386832052.
>>> >> Invalid .gcno File!
>>> >>
>>> >> bash-4.1$ clang -v
>>> >> clang version 3.6.0 (tags/RELEASE_360/final)
>>> >> Target: x86_64-unknown-linux-gnu
>>> >> Thread model: posix
>>> >>
>>> >> Thank you
>>> >> ~Umesh
>>> >>
>>> >> On Tue, Apr 28, 2015 at 8:32 PM, Justin Bogner
>>> >> <justin at justinbogner.com>
>>> >> wrote:
>>> >> >
>>> >> > On Tuesday, April 28, 2015, Umesh Kalappa <umesh.kalappa0 at gmail.com>
>>> >> > wrote:
>>> >> >>
>>> >> >> Hi All,
>>> >> >>
>>> >> >> We trying  to use clang+llvm to generate the gcc  coverage format as
>>> >> >>
>>> >> >> clang version 3.6.0
>>> >> >>
>>> >> >> $clang  --coverage -Xclang -coverage-cfg-checksum -Xclang
>>> >> >> -coverage-no-function-names-in-data -Xclang -coverage-version='407*'
>>> >> >> test.c
>>> >> >>
>>> >> >> $a.out
>>> >> >>
>>> >> >> $llvm-cov gcov test.gcda
>>> >> >>
>>> >> >> Unexpected version: *704.
>>> >> >> Invalid .gcno File!
>>> >> >
>>> >> >
>>> >> > The llvm-cov tool supports the 402* format, which is what clang emits
>>> >> > by
>>> >> > default. If you remove the -Xclang arguments from your compilation it
>>> >> > should
>>> >> > work.
>>> >> >
>>> >> >>
>>> >> >> Debugging the above cause ,But any hints from experts here ,will
>>> >> >> help a
>>> >> >> lot .
>>> >> >>
>>> >> >> Appreciate your time ,Thank you
>>> >> >> ~Umesh
>>> >> >> _______________________________________________
>>> >> >> LLVM Developers mailing list
>>> >> >> LLVMdev at cs.uiuc.edu         http://llvm.cs.uiuc.edu
>>> >> >> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev