Fwd: [PATCH 2/4] covoar: Use DWARF to map addresses to source files and lines.

Cillian O'Donnell cpodonnell8 at gmail.com
Thu May 10 17:11:24 UTC 2018


On Thu, 10 May 2018, 16:01 Vijay Kumar Banerjee, <vijaykumar9597 at gmail.com>
wrote:

>
>
> On 8 May 2018 at 00:33, Cillian O'Donnell <cpodonnell8 at gmail.com> wrote:
>
>> Here's the branch with work in progress for the tester. Have a look when
>> you can Vijay.
>>
>> https://github.com/cillianodonnell/rtems-tools/tree/tester-support
>>
>>
>> I am going through this now.
> according to your last commit message, the qemu.cfg is not generating cov
> files.
>

The commit message was just a note to myself to check if that was the case.
Sorry forgot to mention we used to search for exe.cov and with Chris'
changes its now searching for .cov. Have to change the name that's passed
to the qemu cmd to fix that. You could try and change that.

Just to see where it stands, I run the rtems-test with --coverage and
> --no-clean, seems like the cov file is being generated .
> I can see a base_sp.exe.cov file after running for base_sp.exe the reason
> why it's not running is probably because covoar is searching for
> base_sp.cov not base_sp.exe.cov
>
> after renaming the file to base_sp.cov and running covoar from the leon3
> directory I was able to reproduce the same error as you
> ...............
> ...............
> ...............
> terminate called after throwing an instance of 'rld::error'
> Aborted
> .................
> .................
> .................
>
> I think the next part of the tester support will be to load the library
> from out of the build directory. (?)
>

I wouldn't even worry about that for the moment. Just getting the tester to
run covoar with no errors and generate a report with data intact would be
good enough to throw up the patches for a first review and then we can see
what else needs to be done.

>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rtems.org/pipermail/devel/attachments/20180510/7fb3bb40/attachment-0002.html>


More information about the devel mailing list