[Intel-gfx] Intel i915 corruption issue Gnome EOG #146

[Jonny Grant]

On 30/10/2020 10:54, Chris Wilson wrote:
> Quoting Joonas Lahtinen (2020-10-30 10:17:17)
>> + intel-gfx mailing list
>>
>> Quoting Joonas Lahtinen (2020-10-30 12:15:44)
>>> Quoting Jonny Grant (2020-10-27 22:42:19)
>>>> Hello Jani, Joonas
>>>>
>>>> https://gitlab.gnome.org/GNOME/eog/-/issues/146
>>>>
>>>> Is this issue something you could debug?
>>>
>>> Can you file a bug according to the instructions here with details
>>> about the hardware:
>>>
>>> https://gitlab.freedesktop.org/drm/intel/-/wikis/How-to-file-i915-bugs
>>>
>>>> I've got some screenshots
>>>>
>>>>
>>>>
>>>> Seems to be above 32bit.
>>>>
>>>> https://gitlab.gnome.org/GNOME/eog/-/issues/150#note_946421
>>>>
>>>> If you don't think it is an i915 issue, could you post a comment?
>>>> It only reproduces on intel i915 chipset laptops
>>>
>>> Quite a few components are involved, so it might also be a bug in the
>>> Mesa 3D side.
>>>
>>> An useful excercise would be to figure out the exact pixel width after
>>> which the contents is clipped. I noticed one reporting user is using
>>> Thinkpad X220, have any users of more modern hardware reported similar
>>> issue?
> 
> Large pixmap support was dropped from glamor, -modesetting can only
> support images up to the HW surface limit. It is a functional
> regression.
> -Chris
> 

Hell Chris, Joonas

Thank you for your reply.

May I ask if you can reproduce the issue on one of your developer machines at Intel? so we can rule in/out intel graphics...

Let me know if you have a Gnome desktop with EOG. Let me know if you'd like me to email the image
Jonny