GNOME shell broken after suspend and resume for driver r300_dri

[Christoph Frieben]

Can anybody report success for the suspend and resume cycle triggered
from within the GNOME shell when the graphics card is based on some
R300/R400 variant? In my case, the shell is still broken after resume.
The video device is some ATI Radeon X800 card which is using the R420
chip. Only a black desktop, a mouse pointer and the GNOME panel are
visible after resume. The desktop is not completely frozen since I am
still able to move the mouse pointer. It is also possible to switch to
a VT in order to reboot the system. Thanks, ~C
PS: The F15 development cycle is nearing completion, and in my
opinion, this is a major usability issue.

[Adam Williamson]

On Thu, 2011-05-05 at 09:42 +0200, Christoph Frieben wrote:
>  Can anybody report success for the suspend and resume cycle
> triggered
>  from within the GNOME shell when the graphics card is based on some
>  R300/R400 variant? In my case, the shell is still broken after resume.
>  The video device is some ATI Radeon X800 card which is using the R420
>  chip. Only a black desktop, a mouse pointer and the GNOME panel are
>  visible after resume. The desktop is not completely frozen since I am
>  still able to move the mouse pointer. It is also possible to switch to
>  a VT in order to reboot the system. Thanks, ~C
>  PS: The F15 development cycle is nearing completion, and in my
>  opinion, this is a major usability issue. 
Please just file a bug, it's not massively important right now whether
this affects similar chipsets or not; the important thing is that it
gets tracked as a bug. thanks!
-- 
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Fedora Talk: adamwill AT fedoraproject DOT org
http://www.happyassassin.net

[Christoph Frieben]

2011/5/6 Adam Williamson:
>  Please just file a bug, it's not massively important right now
> whether
>  this affects similar chipsets or not; the important thing is that it
>  gets tracked as a bug. thanks!
>  --
>  Adam Williamson 
I had reported bug 643700 back in October 2010 for Fedora 14 ("black
window when running glxgears after resume from suspend on ATI Radeon
X800 SE") but it did not get further attention and nobody but you is
on the CC list even though the hardware is mainstream, and I would
expect many other Radeon users to be affected, too, making this a more
global issue. That's why I was interested in obtaining feedback
through the list. ~C

[Christoph Frieben]

2011/5/7 Christoph Frieben:
>  and nobody but you is on the CC list 
Sorry, I meant to write "nobody but Adam" (Jackson). ~C

[Adam Williamson]

On Sat, 2011-05-07 at 06:57 +0200, Christoph Frieben wrote:
>  2011/5/6 Adam Williamson:
>  > Please just file a bug, it's not massively important right now whether
>  > this affects similar chipsets or not; the important thing is that it
>  > gets tracked as a bug. thanks!
>  > --
>  > Adam Williamson
>  
>  I had reported bug 643700 back in October 2010 for Fedora 14 ("black
>  window when running glxgears after resume from suspend on ATI Radeon
>  X800 SE") but it did not get further attention and nobody but you is
>  on the CC list even though the hardware is mainstream, and I would
>  expect many other Radeon users to be affected, too, making this a more
>  global issue. That's why I was interested in obtaining feedback
>  through the list. ~C 
well, prior to f15, breaking 3D after suspend isn't massively important
- certainly ranks behind many other radeon bugs the maintainers would
have put first in priority. Of course, Shell changes that a bit. I'd
recommend you update the bug with the new info (it's still valid against
f15, and now the impact is worse because of Shell). thanks :)
-- 
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Fedora Talk: adamwill AT fedoraproject DOT org
http://www.happyassassin.net

[Christoph Frieben]

2011/5/7 Adam Williamson:
>  well, prior to f15, breaking 3D after suspend isn't massively
> important
>  - certainly ranks behind many other radeon bugs the maintainers would
>  have put first in priority. Of course, Shell changes that a bit. I'd
>  recommend you update the bug with the new info (it's still valid against
>  f15, and now the impact is worse because of Shell). thanks :) 
I had set the version tag to 15 earlier this year and posted a message
referring to its GNOME3 relevance,

    http://lists.fedoraproject.org/pipermail/devel/2011-February/148580.html .

In my opinion, this bug should at least be assigned importance nth if
not more. However, before doing so, it would be interesting to know
about the share of user affected by this issue. Thus, I urge all
Radeon users relying on the r300 Mesa driver to check whether they can
resume from suspend successfully to a fully working GNOME shell or
compiz. Relevant hardware are all Radeon video cards sporting R3xx or
R4xx chips. Thanks! ~C

[Christoph Frieben]

2011/5/7 Adam Williamson:
>  well, prior to f15, breaking 3D after suspend isn't massively
> important
>  - certainly ranks behind many other radeon bugs the maintainers would
>  have put first in priority. Of course, Shell changes that a bit. I'd
>  recommend you update the bug with the new info (it's still valid against
>  f15, and now the impact is worse because of Shell). thanks :) 
Resume from supend is still broken for an ATI Radeon X800 SE even for
the latest F17 snapshot which is a major usability issue. The
correspoding bug https://bugzilla.redhat.com/show_bug.cgi?id=643700
was filed 18 months ago without seeing any activity. This happens for
a standard ATI video card which worked fine for releases prior or
equal to F13. If somebody else sees this issue, too, he should
complete said bug report. Thanks! ~C

[mwesten]

On 05/05/2011 03:42 AM, Christoph Frieben wrote:
>  Can anybody report success for the suspend and resume cycle
> triggered
>  from within the GNOME shell when the graphics card is based on some
>  R300/R400 variant? In my case, the shell is still broken after resume.
>  The video device is some ATI Radeon X800 card which is using the R420
>  chip. Only a black desktop, a mouse pointer and the GNOME panel are
>  visible after resume. The desktop is not completely frozen since I am
>  still able to move the mouse pointer. It is also possible to switch to
>  a VT in order to reboot the system. Thanks, ~C
>  PS: The F15 development cycle is nearing completion, and in my
>  opinion, this is a major usability issue. 
I have Fedora-15-i686-Live-Desktop.iso (Final RC3) on a LiveUSB and ATI 
Radeon X300SE (RV370) graphics.  Suspend/resume from GNOME 3 is working 
fine for me.