[igt-dev] [PATCH i-g-t] igt: Use lib gem_execbuf where possible

Daniele Ceraolo Spurio daniele.ceraolospurio at intel.com
Fri Feb 9 18:04:38 UTC 2018



On 09/02/18 08:01, Antonio Argenziano wrote:
> 
> 
> On 09/02/18 07:25, Antonio Argenziano wrote:
>>
>>
>> On 09/02/18 01:43, Chris Wilson wrote:
>>> Quoting Antonio Argenziano (2018-02-09 00:21:57)
>>>> diff --git a/tests/gem_busy.c b/tests/gem_busy.c
>>>> index c349c291..444c1dd1 100644
>>>> --- a/tests/gem_busy.c
>>>> +++ b/tests/gem_busy.c
>>>> @@ -332,11 +332,6 @@ static void alarm_handler(int sig)
>>>>   {
>>>>   }
>>>> -static int __execbuf(int fd, struct drm_i915_gem_execbuffer2 *execbuf)
>>>> -{
>>>> -       return ioctl(fd, DRM_IOCTL_I915_GEM_EXECBUFFER2, execbuf);
>>>> -}
>>>
>>> Wrong. There's an important difference here, can you see it? It's all to
>>> do with signal handling...
>>
>> I guess we do not want to keep trying doing the same ioctl once the 
>> ringbuffer is full...
> 
> Say one wanted to put measure_ring_size() in lib, what would be the 
> right place for it?
> 
> Thanks,
> Antonio
> 

I had sent a series for that (and igt_cork) a while back:

https://patchwork.freedesktop.org/series/31858/

It's still on my list of things to come back to but I doubt I'll have 
time in the short term, so feel free to pick it up if you don't want to 
wait.

Thanks,
Daniele

>>
>> Thanks for finding this, it would have been a nasty one.
>>
>> Antonio
>>
>>> -Chris
>>>
> _______________________________________________
> igt-dev mailing list
> igt-dev at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/igt-dev


More information about the igt-dev mailing list