[PATCH 2/2] libdrm: add etnaviv tests

[Emil Velikov]

On 30 August 2016 at 10:08, Daniel Vetter <daniel at ffwll.ch> wrote:
> On Tue, Aug 30, 2016 at 09:14:51AM +0200, Christian Gmeiner wrote:
>> From: The etnaviv authors <dri-devel at lists.freedesktop.org>
>>
>> This adds the following basic unit tests:
>>
>> - etnaviv_2d_test
>>   Let the 2D core render a defined pattern into a bo
>>   and store it as bmp.
>>
>> - etnaviv_bo_cache_test
>>   Basic tests to validate the bo-cache behavior.
>>
>> - etnaviv_cmd_stream_test
>>   Tests for the etna_cmd_stream API.
>
> igt (now at a new location at
> https://cgit.freedesktop.org/drm/igt-gpu-tools/) would be a nice place for
> these I think, if you want to participate there. vc4 has tests in there,
> and collabora is porting a lot of the kms tests to be generic (so that
> they can be run on any kms driver). Then you could just run those tests on
> any driver and have a reasonable assurance you didn't break the world.
>
Having things in IGT might be better indeed. For the moment it might
be better to land this as-is and move things in due time while add new
tests directly in IGT ?

Other than that - a couple of trivial questions/suggestions. Similar
to the patch 1/2 ones - can be done as follow-up if needed.

>> --- /dev/null
>> +++ b/tests/etnaviv/cmdstream.xml.h

>> +The rules-ng-ng source files this header was generated from are:
>> +- /home/orion/projects/etna_viv/rnndb/cmdstream.xml (  12589 bytes, from 2013-09-01 10:53:22)
>> +- /home/orion/projects/etna_viv/rnndb/common.xml    (  18379 bytes, from 2014-01-27 15:58:05)
Always wondered about this - can we fix rnn to not print the full path
but just files or rnndb/filename.xml ?

>> +Copyright (C) 2013
There is something funny here - currently we're not 2013 and the
copyright holders are missing. Guess something went bonkers in rnn ?

>> +int main(int argc, char *argv[])
>> +{

>> +fail:
>> +     if (stream)
>> +             etna_cmd_stream_del(stream);
>> +
>> +     if (pipe)
>> +             etna_pipe_del(pipe);
>> +
>> +     if (gpu)
>> +             etna_gpu_del(gpu);
>> +
>> +     if (dev)
>> +             etna_device_del(dev);
>> +
>> +     close(fd);
>> +
Feel free to ignore: you can use separate labels and drop all the if
checks. Also s/fail/out/ since this is not an error path.
Same comment goes the whole patch.

Thanks
Emil