Canvas in Firefox, corruption

Canvas in Firefox, corruption

[Sjon]

Hi,

I have been using nouveau for a few months now, and it works fine.

A few weeks ago I noticed something is broken though; since Google 
switched to using Canvas for their image search I am having some sort of 
corruption in the thumbnails that Google serves.

More information plus screenshots are available here:

https://bbs.archlinux.org/viewtopic.php?id=101469
https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-nouveau/+bug/608613

I was wondering if there is anything I can do to help debugging this? My 
setup currently is Archlinux with a GeForce 9800 GTX and these package 
versions:

nouveau-drm-lts 0.0.16_20100313-2
nouveau-firmware 20091212-4
xf86-video-nouveau 0.0.16_git20100517-1

I do not see any errors getting logged when this corruption appears, is 
there any way to make a debug-version of nouveau to get additional 
information?

Thanks,

Sjon

Re: Canvas in Firefox, corruption

[Andreas Radke]

Am Sat, 14 Aug 2010 14:49:26 +0200
schrieb Sjon <archlinux.org-pbuv2Qasjt8Gk0GAwe0YXA@public.gmane.org>:


> nouveau-drm-lts 0.0.16_20100313-2
> nouveau-firmware 20091212-4
> xf86-video-nouveau 0.0.16_git20100517-1

Are you running the .32 LTS kernel? Please confirm if the behavior is
the same with .35.x Arch stock kernel.

-Andy

Re: Canvas in Firefox, corruption

[Sjon]

On 08/14/2010 09:04 PM, Andreas Radke wrote:
> Am Sat, 14 Aug 2010 14:49:26 +0200
> schrieb Sjon<archlinux.org-pbuv2Qasjt8Gk0GAwe0YXA@public.gmane.org>:
>
>
>> nouveau-drm-lts 0.0.16_20100313-2
>> nouveau-firmware 20091212-4
>> xf86-video-nouveau 0.0.16_git20100517-1
>
> Are you running the .32 LTS kernel? Please confirm if the behavior is
> the same with .35.x Arch stock kernel.

Although I have kernel26-lts-2.6.32.18-1 installed (because 
nouveau-drm-lts depends on it), my machine currently runs on 
kernel26-2.6.34.3-1

I have never intended to install or use LTS, but it seems it was 
installed via some dependency.

Thanks,
Sjon

Re: Canvas in Firefox, corruption

[Maarten Maathuis]

On Sat, Aug 14, 2010 at 9:04 PM, Andreas Radke <a.radke-KvP5wT2u2U0@public.gmane.org> wrote:
> Am Sat, 14 Aug 2010 14:49:26 +0200
> schrieb Sjon <archlinux.org-pbuv2Qasjt8Gk0GAwe0YXA@public.gmane.org>:
>
>
>> nouveau-drm-lts 0.0.16_20100313-2
>> nouveau-firmware 20091212-4
>> xf86-video-nouveau 0.0.16_git20100517-1
>
> Are you running the .32 LTS kernel? Please confirm if the behavior is
> the same with .35.x Arch stock kernel.

I can confirm this behaviour on the latest "nouveau" kernel, but i
haven't had time to look into it.

>
> -Andy
> _______________________________________________
> Nouveau mailing list
> Nouveau-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org
> http://lists.freedesktop.org/mailman/listinfo/nouveau
>



-- 
Far away from the primal instinct, the song seems to fade away, the
river get wider between your thoughts and the things we do and say.

Re: Canvas in Firefox, corruption

[Maarten Maathuis]

On Sat, Aug 14, 2010 at 9:12 PM, Sjon <archlinux.org-pbuv2Qasjt8Gk0GAwe0YXA@public.gmane.org> wrote:
> On 08/14/2010 09:04 PM, Andreas Radke wrote:
>>
>> Am Sat, 14 Aug 2010 14:49:26 +0200
>> schrieb Sjon<archlinux.org-pbuv2Qasjt8Gk0GAwe0YXA@public.gmane.org>:
>>
>>
>>> nouveau-drm-lts 0.0.16_20100313-2
>>> nouveau-firmware 20091212-4
>>> xf86-video-nouveau 0.0.16_git20100517-1
>>
>> Are you running the .32 LTS kernel? Please confirm if the behavior is
>> the same with .35.x Arch stock kernel.
>
> Although I have kernel26-lts-2.6.32.18-1 installed (because nouveau-drm-lts
> depends on it), my machine currently runs on kernel26-2.6.34.3-1
>
> I have never intended to install or use LTS, but it seems it was installed
> via some dependency.
>
> Thanks,
> Sjon
> _______________________________________________
> Nouveau mailing list
> Nouveau-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org
> http://lists.freedesktop.org/mailman/listinfo/nouveau
>

The bug appears to be a pixman bug, fixed by commit
9897bb4eeed165b76001dfefd3a89bcb96d38a72. I have requested it to be
cherry picked for the next release. The REPEAT_NONE for XRGB source
fallback is being triggered in a composite operation.

Maarten.

-- 
Far away from the primal instinct, the song seems to fade away, the
river get wider between your thoughts and the things we do and say.

Re: Canvas in Firefox, corruption

[Maarten Maathuis]

On Sun, Aug 15, 2010 at 12:37 AM, Maarten Maathuis <madman2003-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> On Sat, Aug 14, 2010 at 9:12 PM, Sjon <archlinux.org-pbuv2Qasjt8Gk0GAwe0YXA@public.gmane.org> wrote:
>> On 08/14/2010 09:04 PM, Andreas Radke wrote:
>>>
>>> Am Sat, 14 Aug 2010 14:49:26 +0200
>>> schrieb Sjon<archlinux.org-pbuv2Qasjt8Gk0GAwe0YXA@public.gmane.org>:
>>>
>>>
>>>> nouveau-drm-lts 0.0.16_20100313-2
>>>> nouveau-firmware 20091212-4
>>>> xf86-video-nouveau 0.0.16_git20100517-1
>>>
>>> Are you running the .32 LTS kernel? Please confirm if the behavior is
>>> the same with .35.x Arch stock kernel.
>>
>> Although I have kernel26-lts-2.6.32.18-1 installed (because nouveau-drm-lts
>> depends on it), my machine currently runs on kernel26-2.6.34.3-1
>>
>> I have never intended to install or use LTS, but it seems it was installed
>> via some dependency.
>>
>> Thanks,
>> Sjon
>> _______________________________________________
>> Nouveau mailing list
>> Nouveau-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org
>> http://lists.freedesktop.org/mailman/listinfo/nouveau
>>
>
> The bug appears to be a pixman bug, fixed by commit
> 9897bb4eeed165b76001dfefd3a89bcb96d38a72. I have requested it to be
> cherry picked for the next release. The REPEAT_NONE for XRGB source
> fallback is being triggered in a composite operation.
>
> Maarten.
>
> --
> Far away from the primal instinct, the song seems to fade away, the
> river get wider between your thoughts and the things we do and say.
>

Maybe someone should tell mozilla to use ARGB instead of XRGB, because
that will be hardware accelerated (XRender and opengl disagree on what
should be sampled when going outside the texture without an alpha
channel).

-- 
Far away from the primal instinct, the song seems to fade away, the
river get wider between your thoughts and the things we do and say.

Re: Canvas in Firefox, corruption

[Xavier Chantry]

On Sun, Aug 15, 2010 at 12:48 AM, Maarten Maathuis <madman2003-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
>
> Maybe someone should tell mozilla to use ARGB instead of XRGB, because
> that will be hardware accelerated (XRender and opengl disagree on what
> should be sampled when going outside the texture without an alpha
> channel).
>

You are very welcome to add any corrections or clarifications to the report.
https://bugzilla.mozilla.org/show_bug.cgi?id=587499

Re: Canvas in Firefox, corruption

[Xavier Chantry]

On Wed, Aug 18, 2010 at 11:51 PM, Xavier Chantry
<chantry.xavier-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> On Sun, Aug 15, 2010 at 12:48 AM, Maarten Maathuis <madman2003-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
>>
>> Maybe someone should tell mozilla to use ARGB instead of XRGB, because
>> that will be hardware accelerated (XRender and opengl disagree on what
>> should be sampled when going outside the texture without an alpha
>> channel).
>>
>
> You are very welcome to add any corrections or clarifications to the report.
> https://bugzilla.mozilla.org/show_bug.cgi?id=587499
>

https://bugzilla.mozilla.org/show_bug.cgi?id=587499#c1
"Or alternatively PAD repeat can be set on the source images at least in some
cases. Everything depends on whether it is preferable (for performance reasons)
to have SRC compositing operation instead of OVER for some images which are
known to be opaque. Also see bug 568767 for more details."

What do you think of this comment, how does nouveau handle that ?

Re: Canvas in Firefox, corruption

[Maarten Maathuis]

On Mon, Oct 4, 2010 at 2:29 PM, Xavier Chantry <chantry.xavier-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> On Wed, Aug 18, 2010 at 11:51 PM, Xavier Chantry
> <chantry.xavier-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
>> On Sun, Aug 15, 2010 at 12:48 AM, Maarten Maathuis <madman2003-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
>>>
>>> Maybe someone should tell mozilla to use ARGB instead of XRGB, because
>>> that will be hardware accelerated (XRender and opengl disagree on what
>>> should be sampled when going outside the texture without an alpha
>>> channel).
>>>
>>
>> You are very welcome to add any corrections or clarifications to the report.
>> https://bugzilla.mozilla.org/show_bug.cgi?id=587499
>>
>
> https://bugzilla.mozilla.org/show_bug.cgi?id=587499#c1
> "Or alternatively PAD repeat can be set on the source images at least in some
> cases. Everything depends on whether it is preferable (for performance reasons)
> to have SRC compositing operation instead of OVER for some images which are
> known to be opaque. Also see bug 568767 for more details."
>
> What do you think of this comment, how does nouveau handle that ?
>

nv40+ will be fine, older cards can't repeat NPOT textures. So in that
sense switching to ARGB is better than RepeatPad.


-- 
Far away from the primal instinct, the song seems to fade away, the
river get wider between your thoughts and the things we do and say.

Re: Canvas in Firefox, corruption

[Xavier Chantry]

On Mon, Oct 4, 2010 at 6:27 PM, Maarten Maathuis <madman2003-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> On Mon, Oct 4, 2010 at 2:29 PM, Xavier Chantry <chantry.xavier-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
>>
>> https://bugzilla.mozilla.org/show_bug.cgi?id=587499#c1
>> "Or alternatively PAD repeat can be set on the source images at least in some
>> cases. Everything depends on whether it is preferable (for performance reasons)
>> to have SRC compositing operation instead of OVER for some images which are
>> known to be opaque. Also see bug 568767 for more details."
>>
>> What do you think of this comment, how does nouveau handle that ?
>>
>
> nv40+ will be fine, older cards can't repeat NPOT textures. So in that
> sense switching to ARGB is better than RepeatPad.
>
>

3 new comments after I posted your answer there.
It would really be awesome if you could just check directly in the bug
report what they are saying, and comment if necessary :)