Bug 23722 - Regression in kernel 2.6.37 with regards to i915 driver (Intel Graphics)
Summary: Regression in kernel 2.6.37 with regards to i915 driver (Intel Graphics)
Status: RESOLVED CODE_FIX
Alias: None
Product: Drivers
Classification: Unclassified
Component: Video(DRI - Intel) (show other bugs)
Hardware: x86-64 Linux
: P1 high
Assignee: drivers_video-dri-intel@kernel-bugs.osdl.org
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-11-25 09:46 UTC by Raymond Wooninck
Modified: 2010-12-16 13:38 UTC (History)
1 user (show)

See Also:
Kernel Version: 2.6.37-rc3
Subsystem:
Regression: Yes
Bisected commit-id:


Attachments

Description Raymond Wooninck 2010-11-25 09:46:30 UTC
Hardware: Lenovo T410 (intel Arrandale Graphics HD 5700). Running with latest
Xorg and Mesa (7.9). Distribution: openSUSE Factory.

The desktop is working normally with kernel 2.6.36. Desktop effects are enabled
and everything is very responsive. Booting with kernel 2.6.37-rc2(3) results in
an unusable desktop if Wobbly Windows is activated. Windows are appearing very
slow and dragging them across the screen causes a very big lag if it doesn't
result in a frozen screen. 

Switching Wobbly Windows off, makes the desktop more responsive, but still you
can see some initial lag when the window is dragged from it's place. Stop
dragging the windows causes another small lag to draw over existing windows. 

The lag-ness of moving windows around is being caused by the translucency which
is set for on when moving windows. When I set this to opaque when moving
windows, I have again a fast and direct moving window. 

The plasma pop-ups are appearing slow as that the desktop theme contains some
translucent elements in there. 

Of course switching off desktop effects all together resolves the mentioned
issues.

Regards
Raymond
Comment 1 Raymond Wooninck 2010-11-25 09:48:14 UTC
I am using KDE 4.6 latest trunk version.
Comment 2 Chris Wilson 2010-12-16 13:38:29 UTC
commit de18a29e0fa3904894b4e02fae0e712cd43f740c
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Sat Nov 27 22:30:41 2010 +0100

    drm/i915: fix regression due to ba3d8d749b01548b9
    
    We don't track gpu flush request in any special way. So even with
    obj->write_domain == 0, a gpu flush might be outstanding but no
    yet executed. Even worse, the latest request might use the object
    only for reading. So and unconditional call to object_wait_rendering
    is needed for !pipelined.
    
    Hence revert that patch fully and untangle the flushing from the
    synchronization again.
    
    Reported-by: Keith Packard <keithp@keithp.com>
    Tested-by: Keith Packard <keithp@keithp.com>
    Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
    Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Note You need to log in before you can comment on or make changes to this bug.