2009-10-10 18:37:23

by Sergei Trofimovich

[permalink] [raw]
Subject: [bisected] i915: linux-2.6.32-rc3+ regression

Hi!

After syncing git tree to bd381934bf13ccb1af2813ae26c6fe00ec85d254
I've experienced pixel color corruption in Xorg.
Some of pixels (not all) have abnormal colors seen on images like this:
http://trofimeus.narod.ru/2009-10-09-22-00-17.png [1.5M]
(you won't see any problem until you run -rc3+ kernel).

I've bisected it[2] and found offending commit[1].
My hardware is HP Compaq 2510p laptop:
00:02.0 VGA compatible controller: Intel Corporation Mobile GM965/GL960 Integrated Graphics Controller (rev 0c)

Do you need more info?

Thanks!

[1] offending commit
> 068143d38804825d59d951a192cfadd2e22f457d is the first bad commit
> commit 068143d38804825d59d951a192cfadd2e22f457d
> Author: Dave Airlie <[email protected]>
> Date: Mon Oct 5 09:58:02 2009 +1000
>
> drm/fb: add setcmap and fix 8-bit support.
>
> This adds support for the setcmap api and fixes the 8bpp
> support at least on radeon hardware. It adds a new load_lut
> hook which can be called once the color map is setup.
>
> Signed-off-by: Dave Airlie <[email protected]>

[2] bisect log:
:040000 040000 9334e1ee9d528eaa5babd419bd53b40417e65f0f 7e06eb50488482efd38ac243080c9f27ee8682ec M drivers
:040000 040000 31c8cb95831669746c0ee2ff8e34d15dc6c49b0e 0523107fc14ca0bb6914b93ea56dab20c741d61c M include
git bisect start
# bad: [bd381934bf13ccb1af2813ae26c6fe00ec85d254] Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rol
and/infiniband
git bisect bad bd381934bf13ccb1af2813ae26c6fe00ec85d254
# good: [374576a8b6f865022c0fd1ca62396889b23d66dd] Linux 2.6.32-rc3
git bisect good 374576a8b6f865022c0fd1ca62396889b23d66dd
# bad: [1c6e6d91b22c4271e8a5dab559a08cb005a77073] Merge branch 'drm-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/air
lied/drm-2.6
git bisect bad 1c6e6d91b22c4271e8a5dab559a08cb005a77073
# good: [d8e7b2b3ac5319fcee616488c628c2bbe7e2937b] Merge branch 'omap-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/ke
rnel/git/tmlind/linux-omap-2.6
git bisect good d8e7b2b3ac5319fcee616488c628c2bbe7e2937b
# good: [185974dd596e67e73906790c2a4a4579d75911e6] drm: Delete the DRM_DEBUG_KMS in drm_mode_cursor_ioctl
git bisect good 185974dd596e67e73906790c2a4a4579d75911e6
# bad: [01ceae8edd7a0a6d8588dc103ad9f55e2c52cae9] drm/radeon/kms: Fix RS600/RV515/R520/RS690 IRQ
git bisect bad 01ceae8edd7a0a6d8588dc103ad9f55e2c52cae9
# bad: [b8c00ac5b50b54491657f8b6740db1df50149944] drm/fb: add more correct 8/16/24/32 bpp fb support.
git bisect bad b8c00ac5b50b54491657f8b6740db1df50149944
# bad: [068143d38804825d59d951a192cfadd2e22f457d] drm/fb: add setcmap and fix 8-bit support.
git bisect bad 068143d38804825d59d951a192cfadd2e22f457d
# good: [dfee5614e4d83a32cef9193a8b19bc1d8900f93d] drm/radeon/kms: respect single crtc cards, only create one crtc. (v2)
git bisect good dfee5614e4d83a32cef9193a8b19bc1d8900f93d


--

Sergei


Attachments:
signature.asc (198.00 B)

2009-10-10 21:27:40

by Maxim Levitsky

[permalink] [raw]
Subject: Re: [Intel-gfx] [bisected] i915: linux-2.6.32-rc3+ regression

On Sat, 2009-10-10 at 21:37 +0300, Sergei Trofimovich wrote:
> Hi!
>
> After syncing git tree to bd381934bf13ccb1af2813ae26c6fe00ec85d254
> I've experienced pixel color corruption in Xorg.
> Some of pixels (not all) have abnormal colors seen on images like this:
> http://trofimeus.narod.ru/2009-10-09-22-00-17.png [1.5M]
> (you won't see any problem until you run -rc3+ kernel).
>
> I've bisected it[2] and found offending commit[1].
> My hardware is HP Compaq 2510p laptop:
> 00:02.0 VGA compatible controller: Intel Corporation Mobile GM965/GL960 Integrated Graphics Controller (rev 0c)
>
> Do you need more info?
>
> Thanks!
>
> [1] offending commit
> > 068143d38804825d59d951a192cfadd2e22f457d is the first bad commit
> > commit 068143d38804825d59d951a192cfadd2e22f457d
> > Author: Dave Airlie <[email protected]>
> > Date: Mon Oct 5 09:58:02 2009 +1000
> >
> > drm/fb: add setcmap and fix 8-bit support.
> >
> > This adds support for the setcmap api and fixes the 8bpp
> > support at least on radeon hardware. It adds a new load_lut
> > hook which can be called once the color map is setup.
> >
> > Signed-off-by: Dave Airlie <[email protected]>
>
> [2] bisect log:
> :040000 040000 9334e1ee9d528eaa5babd419bd53b40417e65f0f 7e06eb50488482efd38ac243080c9f27ee8682ec M drivers
> :040000 040000 31c8cb95831669746c0ee2ff8e34d15dc6c49b0e 0523107fc14ca0bb6914b93ea56dab20c741d61c M include
> git bisect start
> # bad: [bd381934bf13ccb1af2813ae26c6fe00ec85d254] Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rol
> and/infiniband
> git bisect bad bd381934bf13ccb1af2813ae26c6fe00ec85d254
> # good: [374576a8b6f865022c0fd1ca62396889b23d66dd] Linux 2.6.32-rc3
> git bisect good 374576a8b6f865022c0fd1ca62396889b23d66dd
> # bad: [1c6e6d91b22c4271e8a5dab559a08cb005a77073] Merge branch 'drm-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/air
> lied/drm-2.6
> git bisect bad 1c6e6d91b22c4271e8a5dab559a08cb005a77073
> # good: [d8e7b2b3ac5319fcee616488c628c2bbe7e2937b] Merge branch 'omap-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/ke
> rnel/git/tmlind/linux-omap-2.6
> git bisect good d8e7b2b3ac5319fcee616488c628c2bbe7e2937b
> # good: [185974dd596e67e73906790c2a4a4579d75911e6] drm: Delete the DRM_DEBUG_KMS in drm_mode_cursor_ioctl
> git bisect good 185974dd596e67e73906790c2a4a4579d75911e6
> # bad: [01ceae8edd7a0a6d8588dc103ad9f55e2c52cae9] drm/radeon/kms: Fix RS600/RV515/R520/RS690 IRQ
> git bisect bad 01ceae8edd7a0a6d8588dc103ad9f55e2c52cae9
> # bad: [b8c00ac5b50b54491657f8b6740db1df50149944] drm/fb: add more correct 8/16/24/32 bpp fb support.
> git bisect bad b8c00ac5b50b54491657f8b6740db1df50149944
> # bad: [068143d38804825d59d951a192cfadd2e22f457d] drm/fb: add setcmap and fix 8-bit support.
> git bisect bad 068143d38804825d59d951a192cfadd2e22f457d
> # good: [dfee5614e4d83a32cef9193a8b19bc1d8900f93d] drm/radeon/kms: respect single crtc cards, only create one crtc. (v2)
> git bisect good dfee5614e4d83a32cef9193a8b19bc1d8900f93d
>
>
> _______________________________________________
> Intel-gfx mailing list
> [email protected]
> http://lists.freedesktop.org/mailman/listinfo/intel-gfx


I found that today too.
I had filled a bugreport.
https://bugs.freedesktop.org/show_bug.cgi?id=24444


Same commit is responsible.

Best regards,
Maxim Levitsky

2009-10-10 21:46:08

by Rafael J. Wysocki

[permalink] [raw]
Subject: Re: [Intel-gfx] [bisected] i915: linux-2.6.32-rc3+ regression

On Saturday 10 October 2009, Maxim Levitsky wrote:
> On Sat, 2009-10-10 at 21:37 +0300, Sergei Trofimovich wrote:
> > Hi!
> >
> > After syncing git tree to bd381934bf13ccb1af2813ae26c6fe00ec85d254
> > I've experienced pixel color corruption in Xorg.
> > Some of pixels (not all) have abnormal colors seen on images like this:
> > http://trofimeus.narod.ru/2009-10-09-22-00-17.png [1.5M]
> > (you won't see any problem until you run -rc3+ kernel).
> >
> > I've bisected it[2] and found offending commit[1].
> > My hardware is HP Compaq 2510p laptop:
> > 00:02.0 VGA compatible controller: Intel Corporation Mobile GM965/GL960 Integrated Graphics Controller (rev 0c)
> >
> > Do you need more info?
> >
> > Thanks!
> >
> > [1] offending commit
> > > 068143d38804825d59d951a192cfadd2e22f457d is the first bad commit
> > > commit 068143d38804825d59d951a192cfadd2e22f457d
> > > Author: Dave Airlie <[email protected]>
> > > Date: Mon Oct 5 09:58:02 2009 +1000
> > >
> > > drm/fb: add setcmap and fix 8-bit support.
> > >
> > > This adds support for the setcmap api and fixes the 8bpp
> > > support at least on radeon hardware. It adds a new load_lut
> > > hook which can be called once the color map is setup.
> > >
> > > Signed-off-by: Dave Airlie <[email protected]>
> >
> > [2] bisect log:
> > :040000 040000 9334e1ee9d528eaa5babd419bd53b40417e65f0f 7e06eb50488482efd38ac243080c9f27ee8682ec M drivers
> > :040000 040000 31c8cb95831669746c0ee2ff8e34d15dc6c49b0e 0523107fc14ca0bb6914b93ea56dab20c741d61c M include
> > git bisect start
> > # bad: [bd381934bf13ccb1af2813ae26c6fe00ec85d254] Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rol
> > and/infiniband
> > git bisect bad bd381934bf13ccb1af2813ae26c6fe00ec85d254
> > # good: [374576a8b6f865022c0fd1ca62396889b23d66dd] Linux 2.6.32-rc3
> > git bisect good 374576a8b6f865022c0fd1ca62396889b23d66dd
> > # bad: [1c6e6d91b22c4271e8a5dab559a08cb005a77073] Merge branch 'drm-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/air
> > lied/drm-2.6
> > git bisect bad 1c6e6d91b22c4271e8a5dab559a08cb005a77073
> > # good: [d8e7b2b3ac5319fcee616488c628c2bbe7e2937b] Merge branch 'omap-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/ke
> > rnel/git/tmlind/linux-omap-2.6
> > git bisect good d8e7b2b3ac5319fcee616488c628c2bbe7e2937b
> > # good: [185974dd596e67e73906790c2a4a4579d75911e6] drm: Delete the DRM_DEBUG_KMS in drm_mode_cursor_ioctl
> > git bisect good 185974dd596e67e73906790c2a4a4579d75911e6
> > # bad: [01ceae8edd7a0a6d8588dc103ad9f55e2c52cae9] drm/radeon/kms: Fix RS600/RV515/R520/RS690 IRQ
> > git bisect bad 01ceae8edd7a0a6d8588dc103ad9f55e2c52cae9
> > # bad: [b8c00ac5b50b54491657f8b6740db1df50149944] drm/fb: add more correct 8/16/24/32 bpp fb support.
> > git bisect bad b8c00ac5b50b54491657f8b6740db1df50149944
> > # bad: [068143d38804825d59d951a192cfadd2e22f457d] drm/fb: add setcmap and fix 8-bit support.
> > git bisect bad 068143d38804825d59d951a192cfadd2e22f457d
> > # good: [dfee5614e4d83a32cef9193a8b19bc1d8900f93d] drm/radeon/kms: respect single crtc cards, only create one crtc. (v2)
> > git bisect good dfee5614e4d83a32cef9193a8b19bc1d8900f93d
> >
> >
> > _______________________________________________
> > Intel-gfx mailing list
> > [email protected]
> > http://lists.freedesktop.org/mailman/listinfo/intel-gfx
>
>
> I found that today too.
> I had filled a bugreport.
> https://bugs.freedesktop.org/show_bug.cgi?id=24444

Would you please notify me about _kernel_ regression reports filed at
freedesktop.org?

Thanks,
Rafael

2009-10-10 21:56:40

by Dave Airlie

[permalink] [raw]
Subject: Re: [Intel-gfx] [bisected] i915: linux-2.6.32-rc3+ regression

On Sun, Oct 11, 2009 at 7:47 AM, Rafael J. Wysocki <[email protected]> wrote:
> On Saturday 10 October 2009, Maxim Levitsky wrote:
>> On Sat, 2009-10-10 at 21:37 +0300, Sergei Trofimovich wrote:
>> > Hi!
>> >
>> > After syncing git tree to bd381934bf13ccb1af2813ae26c6fe00ec85d254
>> > I've experienced pixel color corruption in Xorg.
>> > Some of pixels (not all) have abnormal colors seen on images like this:
>> > http://trofimeus.narod.ru/2009-10-09-22-00-17.png [1.5M]
>> > (you won't see any problem until you run -rc3+ kernel).

Are you running X.org server 1.7 + driver? if so there is a bug in the current
released driver between that combination, the fix is on the intel-gfx
mailing list.

The 1.6 X server + same driver doesn't show this problem.

Dave.

>> >
>> > I've bisected it[2] and found offending commit[1].
>> > My hardware is HP Compaq 2510p laptop:
>> > 00:02.0 VGA compatible controller: Intel Corporation Mobile GM965/GL960 Integrated Graphics Controller (rev 0c)
>> >
>> > Do you need more info?
>> >
>> > Thanks!
>> >
>> > [1] offending commit
>> > > 068143d38804825d59d951a192cfadd2e22f457d is the first bad commit
>> > > commit 068143d38804825d59d951a192cfadd2e22f457d
>> > > Author: Dave Airlie <[email protected]>
>> > > Date: ? Mon Oct 5 09:58:02 2009 +1000
>> > >
>> > > ? ? drm/fb: add setcmap and fix 8-bit support.
>> > >
>> > > ? ? This adds support for the setcmap api and fixes the 8bpp
>> > > ? ? support at least on radeon hardware. It adds a new load_lut
>> > > ? ? hook which can be called once the color map is setup.
>> > >
>> > > ? ? Signed-off-by: Dave Airlie <[email protected]>
>> >
>> > [2] bisect log:
>> > :040000 040000 9334e1ee9d528eaa5babd419bd53b40417e65f0f 7e06eb50488482efd38ac243080c9f27ee8682ec M ? drivers
>> > :040000 040000 31c8cb95831669746c0ee2ff8e34d15dc6c49b0e 0523107fc14ca0bb6914b93ea56dab20c741d61c M ? include
>> > git bisect start
>> > # bad: [bd381934bf13ccb1af2813ae26c6fe00ec85d254] Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rol
>> > and/infiniband
>> > git bisect bad bd381934bf13ccb1af2813ae26c6fe00ec85d254
>> > # good: [374576a8b6f865022c0fd1ca62396889b23d66dd] Linux 2.6.32-rc3
>> > git bisect good 374576a8b6f865022c0fd1ca62396889b23d66dd
>> > # bad: [1c6e6d91b22c4271e8a5dab559a08cb005a77073] Merge branch 'drm-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/air
>> > lied/drm-2.6
>> > git bisect bad 1c6e6d91b22c4271e8a5dab559a08cb005a77073
>> > # good: [d8e7b2b3ac5319fcee616488c628c2bbe7e2937b] Merge branch 'omap-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/ke
>> > rnel/git/tmlind/linux-omap-2.6
>> > git bisect good d8e7b2b3ac5319fcee616488c628c2bbe7e2937b
>> > # good: [185974dd596e67e73906790c2a4a4579d75911e6] drm: Delete the DRM_DEBUG_KMS in drm_mode_cursor_ioctl
>> > git bisect good 185974dd596e67e73906790c2a4a4579d75911e6
>> > # bad: [01ceae8edd7a0a6d8588dc103ad9f55e2c52cae9] drm/radeon/kms: Fix RS600/RV515/R520/RS690 IRQ
>> > git bisect bad 01ceae8edd7a0a6d8588dc103ad9f55e2c52cae9
>> > # bad: [b8c00ac5b50b54491657f8b6740db1df50149944] drm/fb: add more correct 8/16/24/32 bpp fb support.
>> > git bisect bad b8c00ac5b50b54491657f8b6740db1df50149944
>> > # bad: [068143d38804825d59d951a192cfadd2e22f457d] drm/fb: add setcmap and fix 8-bit support.
>> > git bisect bad 068143d38804825d59d951a192cfadd2e22f457d
>> > # good: [dfee5614e4d83a32cef9193a8b19bc1d8900f93d] drm/radeon/kms: respect single crtc cards, only create one crtc. (v2)
>> > git bisect good dfee5614e4d83a32cef9193a8b19bc1d8900f93d
>> >
>> >
>> > _______________________________________________
>> > Intel-gfx mailing list
>> > [email protected]
>> > http://lists.freedesktop.org/mailman/listinfo/intel-gfx
>>
>>
>> I found that today too.
>> I had filled a bugreport.
>> https://bugs.freedesktop.org/show_bug.cgi?id=24444
>
> Would you please notify me about _kernel_ regression reports filed at
> freedesktop.org?
>
> Thanks,
> Rafael
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at ?http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at ?http://www.tux.org/lkml/
>

2009-10-11 09:17:13

by Sergei Trofimovich

[permalink] [raw]
Subject: Re: [Intel-gfx] [bisected] i915: linux-2.6.32-rc3+ regression

On Sun, 11 Oct 2009 07:56:01 +1000
Dave Airlie <[email protected]> wrote:

> Are you running X.org server 1.7 + driver? if so there is a bug in the current
> released driver between that combination, the fix is on the intel-gfx
> mailing list.
>
> The 1.6 X server + same driver doesn't show this problem.
Yes, both driver and xorg-server are current -git masters.

Patch http://lists.freedesktop.org/archives/intel-gfx/2009-October/004488.html
helped me.

Thanks!

> Dave.
>

--

Sergei


Attachments:
signature.asc (198.00 B)

2009-10-14 23:36:37

by Maxim Levitsky

[permalink] [raw]
Subject: Re: [Intel-gfx] [bisected] i915: linux-2.6.32-rc3+ regression

On Sun, 2009-10-11 at 12:17 +0300, Sergei Trofimovich wrote:
> On Sun, 11 Oct 2009 07:56:01 +1000
> Dave Airlie <[email protected]> wrote:
>
> > Are you running X.org server 1.7 + driver? if so there is a bug in the current
> > released driver between that combination, the fix is on the intel-gfx
> > mailing list.
> >
> > The 1.6 X server + same driver doesn't show this problem.
> Yes, both driver and xorg-server are current -git masters.
>
> Patch http://lists.freedesktop.org/archives/intel-gfx/2009-October/004488.html
> helped me.
>
> Thanks!
>
> > Dave.
> >
>

Works for me too, thanks!

Best regards,
Maxim Levitsky

2009-10-15 01:35:39

by Maxim Levitsky

[permalink] [raw]
Subject: Re: [Intel-gfx] [bisected] i915: linux-2.6.32-rc3+ regression

On Thu, 2009-10-15 at 01:35 +0200, Maxim Levitsky wrote:
> On Sun, 2009-10-11 at 12:17 +0300, Sergei Trofimovich wrote:
> > On Sun, 11 Oct 2009 07:56:01 +1000
> > Dave Airlie <[email protected]> wrote:
> >
> > > Are you running X.org server 1.7 + driver? if so there is a bug in the current
> > > released driver between that combination, the fix is on the intel-gfx
> > > mailing list.
> > >
> > > The 1.6 X server + same driver doesn't show this problem.
> > Yes, both driver and xorg-server are current -git masters.
> >
> > Patch http://lists.freedesktop.org/archives/intel-gfx/2009-October/004488.html
> > helped me.
> >
> > Thanks!
> >
> > > Dave.
> > >
> >
>
> Works for me too, thanks!
>
> Best regards,
> Maxim Levitsky
>


One minor issue though, gamma table is still corrupted after X was run
(usplash shows wrong colours)


Best regards,
Maxim Levitsky