[PATCH v2 1/2] drm: Add GPU reset sysfs event

[Sharma, Shashank]

On 3/9/2022 8:47 AM, Simon Ser wrote:
> Hi,
> 
> Maybe it would be a good idea to state the intended use-case in the
> commit message? 

It was added in the second patch, but yeah, it makes more sense to add a 
cover-letter probably.

And explain why the current (driver-specific IIRC) APIs
> aren't enough?
> 
As you mentioned, we also started with a driver specific API, and then 
realized that most of the drivers can benefit from this infrastructure, 
as most of them would be more or less interested in the similar 
information.

> Since this introduces new uAPI, can you point to a user-space patch
> which uses the new uAPI? See this link for more info on DRM user-space
> requirements:
> https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fdri.freedesktop.org%2Fdocs%2Fdrm%2Fgpu%2Fdrm-uapi.html%23open-source-userspace-requirements&data=04%7C01%7Cshashank.sharma%40amd.com%7C85e8e40c84524f3272e908da01a11b1c%7C3dd8961fe4884e608e11a82d994e183d%7C0%7C0%7C637824088716047386%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=%2BMTzBIfl3FGHdAv%2BSx%2F83a86x8eksbOhdQVqdrJ2Rq0%3D&reserved=0
> 
The userspace work is in progress, this was to get a general feedback 
from the community and the interest.

> Please use drm_dbg_* and drm_warn_* helpers instead of DRM_DEBUG and
> DRM_WARN. This allows identifying on which device the uevent happens.
> 
Well, I don't want to break the consistency across the file, as other
functions seems to be using DRM_DEBUG family of prints.
> On Tuesday, March 8th, 2022 at 19:04, Shashank Sharma <contactshashanksharma at gmail.com> wrote:
> 
>> +void drm_sysfs_reset_event(struct drm_device *dev, struct drm_reset_event *reset_info)
> 
> reset_info can be const.
Sure.

- Shashank