[Mesa-dev] [PATCH 08/12] gallium/radeon/winsyses: print CS submission error number

[Nicolai Hähnle]

On 04.08.2016 22:30, Marek Olšák wrote:
> On Tue, Aug 2, 2016 at 4:38 PM, Nicolai Hähnle <nhaehnle at gmail.com> wrote:
>> On 29.07.2016 23:42, Marek Olšák wrote:
>>>
>>> From: Marek Olšák <marek.olsak at amd.com>
>>>
>>> ---
>>>  src/gallium/winsys/amdgpu/drm/amdgpu_cs.c     | 2 +-
>>>  src/gallium/winsys/radeon/drm/radeon_drm_cs.c | 2 +-
>>>  2 files changed, 2 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/src/gallium/winsys/amdgpu/drm/amdgpu_cs.c
>>> b/src/gallium/winsys/amdgpu/drm/amdgpu_cs.c
>>> index fb517b9..35e1b48 100644
>>> --- a/src/gallium/winsys/amdgpu/drm/amdgpu_cs.c
>>> +++ b/src/gallium/winsys/amdgpu/drm/amdgpu_cs.c
>>> @@ -935,7 +935,7 @@ void amdgpu_cs_submit_ib(void *job, int thread_index)
>>>           fprintf(stderr, "amdgpu: Not enough memory for command
>>> submission.\n");
>>>        else
>>>           fprintf(stderr, "amdgpu: The CS has been rejected, "
>>> -                 "see dmesg for more information.\n");
>>> +                 "see dmesg for more information (%i).\n", r);
>>>
>>>        amdgpu_fence_signalled(cs->fence);
>>>     } else {
>>> diff --git a/src/gallium/winsys/radeon/drm/radeon_drm_cs.c
>>> b/src/gallium/winsys/radeon/drm/radeon_drm_cs.c
>>> index 606c38d..4a6f005 100644
>>> --- a/src/gallium/winsys/radeon/drm/radeon_drm_cs.c
>>> +++ b/src/gallium/winsys/radeon/drm/radeon_drm_cs.c
>>> @@ -447,7 +447,7 @@ void radeon_drm_cs_emit_ioctl_oneshot(void *job, int
>>> thread_index)
>>>              }
>>>          } else {
>>>              fprintf(stderr, "radeon: The kernel rejected CS, "
>>> -                    "see dmesg for more information.\n");
>>> +                    "see dmesg for more information (%i).\n", r);
>>>          }
>>>      }
>>>
>>>
>>
>> It would be nice to use strerror_r here:
>>
>> char buf[128];
>> strerror_r(-r, buf, sizeof(buf));
>> ... then use buf ...
>>
>> That would also get rid of the ENOMEM special case.
>
> Are you sure that's a good idea? It returns "Cannot allocate memory"
> for ENOMEM, which is slightly different from "Not enough memory for
> command submission."

I believe they're equally (un)informative for an unaware user, and as 
developers we will know the likely underlying reason regardless. But I 
admit that it's a bit of bike-shedding.

Also, while I started experimenting with the buffer-list-commit/rollback 
thing, I have other things to look at first. So whichever way you decide 
on this particular patch, I'm fine with the series as-is. For patches 1-11:

Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>

>
> Marek
>