On Mon, May 19, 2014 at 11:01 PM, Hajime Morrita <[email protected]> wrote:
> Thanks for the feedback folks!
>
> On Sun, May 18, 2014 at 11:48 PM, Konstantin Serebryany
> <[email protected]> wrote:
>>
>> On Sat, May 17, 2014 at 11:12 AM, 'Alexander Potapenko' via
>> address-sanitizer <[email protected]> wrote:
>> > We don't have any such tools yet, but you can hack something on top of
>> > one
>> > of the existing source highlighting scripts for gcov (which I'm not
>> > familiar
>> > to, unfortunately)
>> >
>> > On May 17, 2014 12:03 AM, "Hajime Morrita" <[email protected]> wrote:
>> >>
>> >> Hi,
>> >>
>> >> I just noticed that ASAN go code coverage support [1]. This is great!
>> >> And
>> >> I'm wondering how I can generate highlighted source code from this
>> >> data, to
>> >> indicate covered/uncovered code paths. That data would be pretty useful
>> >> to
>> >> improve code coverage of automated tests.
>> >>
>> >> The generated output from sancov.py and llvm-symbolizer contains only
>> >> the
>> >> line number of the basic block beginning, but it does not contain one
>> >> of the
>> >> ending of the basic block. I'd like to know how to figure out the line
>> >> numbers of the bb ends,
>>
>
>>
>> I don't think this is currently possible.
>> The (experimental!) asan-coverage dumps PCs on disk and then
>> llvm-symbolizer
>> translates PCs to the line numbers -- but not to the line ranges.
>
>
> Thanks for the explanation. This matches my understanding.
> Do you have any plan to add such feature?
No.
> Or does any existing clang
> profiling feature support that?

The new clang's -fprofile-instr-generate may have this in future,
but I don't think it has that now.

>
> Another possibility in my mind is to extrapolate your idea on fuzzing [1],
> that is, to record all BBs somehow, and compare coverage-dumped BB and the
> full BB to see uncovered region.

Hmm? Not sure I understand your suggestion.
BB is a region of code w/o branching. The beginning of BB completely
defines the entire BB.
It's just that the current implementation does not have a way to show
the entire BB

--kcc


> I don't know if it is possible with current clang toolchain though.
>
> What do you think? Any feedback is appreciated.
> --
> morrita
>
>
>
> [1]
> http://llvm.org/devmtg/2014-04/PDFs/LightningTalks/EuroLLVM'14%20--%20ASan%20%2B%20Coverage.pdf
>
>>
>>
>> --kcc
>>
>> >> so that I can highlight the code of each bb
>> >> begin/end pair.
>> >>
>> >> Any advice would be appreciated.
>>
>> --
>> You received this message because you are subscribed to the Google Groups
>> "address-sanitizer" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to [email protected].
>> For more options, visit https://groups.google.com/d/optout.
>
>
> --
> You received this message because you are subscribed to the Google Groups
> "address-sanitizer" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> For more options, visit https://groups.google.com/d/optout.

-- 
You received this message because you are subscribed to the Google Groups 
"address-sanitizer" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to