Dear Cyril, dear maintainer(s),

After upgrading to 2:2.17.0+git20120101-1 I have refresh issues using the
driver with a composited KDE SC 4.7.4 desktop. Sometimes while typing the
window of the KDE application I am typing or parts of it will become black.
This quite frequently happens with KMail for example. But I have also seen
it with browsing the web, also with Iceweasel.

If this happens when I move the mouse over the black areas they get
restored to their actual content. Sometimes the restauration happens
"automatically". In any case when I disable and enable compositing
again everything is restored as well.

Ok, I managed to produce an example screenshot. It happens quite
reproducably when I compose a new mail and then type something into
the subject line of the mail composer window of KMail. Funnily
as I continue typing the actual contents of the black areas appear
and go black again after a keystroke.

I think related it some short flickering to big black areas at times on
switching windows or activities.

I did not have this with the driver version, before the version that enabled
SNA. I used the version with disabled SNA with 3.2-rc7 debian kernel as well
already. The most recent one from today (included in above Version: line)
also has this issue.

This all happens on a Lenovo ThinkPad T520 with Sandybridge graphics and
full HD laptop display.

/etc/X11/xorg.conf.d/anzeige.conf is just displaysize:

merkaba:~> cat /etc/X11/xorg.conf.d/anzeige.conf
Section "Device"
        Identifier      "Grafikkern"

        # Monitore
        Option          "Monitor-LVDS1" "Notebook-Display"
EndSection

Section "Monitor"
        Identifier      "Notebook-Display"

        DisplaySize     340 192
EndSection

Package versions involved are:

martin@merkaba:~> apt-show-versions | egrep "(xserver-xorg|xserver-xorg-core|xserver-xorg-video-intel|libdrm|libgl1-mesa-dri|kdebase-runtime|libqtgui4)" | grep -v dbg
kdebase-runtime/experimental uptodate 4:4.7.4-1
libdrm-intel1/sid uptodate 2.4.30-1
libdrm-nouveau1a/sid uptodate 2.4.30-1
libdrm-radeon1/sid uptodate 2.4.30-1
libdrm2/sid uptodate 2.4.30-1
libgl1-mesa-dri/wheezy uptodate 7.11.2-1
libqtgui4/sid uptodate 4:4.7.4-2
xserver-xorg/wheezy uptodate 1:7.6+10
xserver-xorg-core/experimental uptodate 2:1.11.99.901-1
xserver-xorg-input-all/wheezy uptodate 1:7.6+10
xserver-xorg-input-evdev/experimental uptodate 1:2.6.99.901-1
xserver-xorg-input-synaptics/experimental uptodate 1.5.0+git20120101-1
xserver-xorg-video-intel/experimental uptodate 2:2.17.0+git20120104-1

I think this is an upstream bug. Can you forward it or should I
create a bug at freedesktop.org?

Thanks,
Martin

Well I seen a black area once, but this might have been something
different. They issue does not appear to happen when I disable compositing
in KWin with Alt-Shift-F12.

If you wish I can put this bug upstream at freedesktop.org.

Thanks,

forwarded 655017 https://bugs.freedesktop.org/44787
thanks

I just forwarded this bug to bugs.freedesktop.org:

https://bugs.freedesktop.org/44787

Thanks,

Hi Cyril!

Could you an upgrade of the xserver-xorg-video-intel package to include
at least the commit mentioned by Chris Wilson (on bugs.freedesktop.org)?

Then I could easily test whether that fixes the reported issue.

--- Comment #4 from Chris Wilson <chris@chris-wilson.co.uk> 2012-01-14 13:23:41 PST ---
If I recognize the symptoms correctly, this should be fixed with

commit 983b755313df8a0d256c59c32ec4106e35f237aa
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Thu Jan 12 21:05:16 2012 +0000

    sna/damage: Fix union of extents with dirty damage but no region

    By failing to account for certain paths which would create a damage elt
    without fully initialisating the damage region (only the damage extents),
    we would later overwrite the damage extents with only the extents for
    this operation (rather than the union of this operation with the current
    damage). This fixes a regression from 098592ca5d,
    (sna: Remove the independent tracking of elts from boxes).

    Include the associated damage migration debugging code of the callers.

    Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Thanks,
-- 
Martin 'Helios' Steigerwald - http://www.Lichtvoll.de
GPG: 03B0 0D6C 0040 0710 4AFA  B82F 991B EAAC A599 84C7

Hi Martin!

Martin Steigerwald <Martin@lichtvoll.de> (14/01/2012):

I'll try and look into it during the week-end. Thanks for taking it
upstream.

Mraw,
KiBi.

Am Samstag, 14. Januar 2012 schrieb Cyril Brulebois:

Hi Cyril,

Thanks for the updated driver package.

Hmmm,

merkaba:~> zcat /usr/share/doc/xserver-xorg-video-
intel/changelog.Debian.gz | head -4
xserver-xorg-video-intel (2:2.17.0+git20120115-1) experimental;
urgency=low

  * New upstream snapshot:
    - Merge from upstream master up to 5df7147b09.

doesn´t (completely) fix the issue.

As it contains the aforiomentioned commit

merkaba:~> zgrep 983b755313df8a0d256c59c32ec4106e35f237aa
/usr/share/doc/xserver-xorg-video-intel/changelog.gz
commit 983b755313df8a0d256c59c32ec4106e35f237aa

I´ll report upstream again.

Thanks,

Hi Cyril!

Any change I can poke you again for an even never GIT snapshot?

I consider it fair to demand "do it yourself" ;). I can have a try myself
with the instructions you gave in bug #631516…

… if that basically applies here as well.

Ciao,
Martin

----------  Weitergeleitete Nachricht  ----------
[…]
https://bugs.freedesktop.org/show_bug.cgi?id=44787

    sna/gen6: Restore the non-pipelined op after every WM binding table
update

    The hw wants it as demonstrated by the '>' in KDE's menus. Why is it
    always KDE that demonstrates coherency problems...

    Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

fixes the glitch in menus for me. The code used to do that precisely that
reason, I just forgot the reason, removed the forced flush and thought
everything was still fine. Do you want to poke Cyril for a new package and
try
again? :)
[…]
-------------------------------------------------------

Ciao,
-- 
Martin 'Helios' Steigerwald - http://www.Lichtvoll.de
GPG: 03B0 0D6C 0040 0710 4AFA  B82F 991B EAAC A599 84C7

Hi,

Martin Steigerwald <Martin@lichtvoll.de> (16/01/2012):

I don't think uploading more than a snapshot a week is a good idea, but
I could help you prepare a source package in the worst case.

That would work indeed. You can replace the cherry-pick with git merge
upstream/master, BTW, if you need more than one.

Mraw,
KiBi.

Am Montag, 16. Januar 2012 schrieben Sie:

Well I can build a package myself. No need to upload it to the archive.

From the dialog with Chris Wilson (see the fdo bug) I gather that it would
be helpful to try the newest git from intel driver.

So I need

debcheckout xorg-driver-video xorg-driver-video.git

instead of the mesa stuff?

Meanwhile the issue almost completely went away after I enabled "VSync" in
Systemsettings/Desktop Effects. But still it would be good to see whether I
can reproduce it with the newest intel driver available before further
steps. Cause the issues might have been fixed meanwhile.

Thanks,

Martin Steigerwald <Martin@lichtvoll.de> (16/01/2012):

debcheckout xserver-xorg-video-intel xserver-xorg-video-intel.git

Feel free to join #debian-x on irc.debian.org if you need more hand
holding.

OK.

Mraw,
KiBi.

Hi Cyril!

Building a new package worked marvellously thanks to your instructions.

I will now be testing this bug against:

commit 850495f956c811b1eb617d2e704e6bb7b5a86369
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Tue Jan 17 22:51:29 2012 +0000

    sna: Fix increment of damage boxes after updating for rectangles

which seems to be another fix for refresh issues and at least from the
description looks like something I might have seen.

Should some issues remain I can always build a new package, its easy
enough.

I will report back, once all of these refresh issues are fixed ;).

Thanks,