Start X. Switch VTs before it gets very far in initializing (say, when GDM is in
a busy state before it starts up.)
Errors in messages are:
[drm:i830_wait_ring] *ERROR* space: 131048 wanted 131064
[drm:i830_wait_ring] *ERROR* lockup
Created attachment 90020 [details]
Does this still happen Bill?
I sent Arjan and Alan the latest DRM which might possibly fix this. Assuming
the changes get into our kernel before final release, could you test them?
Arjan, if you add it in, please let us know what build to test.
Correction, these DRM changes did not get into the kernel, and wont be.
Intel DRI is broken and will remain broken in our final release.
Deferring for a future release of Red Hat Linux.
Don't have the machine in front of me. Owen, can you reproduce this on the final
This happens with i865 in Severn, FWIW.
*** Bug 102551 has been marked as a duplicate of this bug. ***
Dave, are you aware of this problem at all? Started in 4.3.0, and occurs
aparently on most i830 and higher hardware aparently.
IMHO, it's kernel DRM bug from the messages. Others say this only occurs
with the Red Hat kernel, not stock 4.3.0 DRM. DRI-CVS DRM reportedly also
works. Should I reassign to kernel for a DRM update?
I have an 845 (I think 845G), and I saw this in Severn. It does not
happen in Yarrow. It does not reproduce with either the 2.6.0-test9
kernel from Arjan, nor the 2.4 kernel shipped default with Yarrow.
I believe this bug is now fixed with our latest kernel update in
rawhide. Arjan has fixed several bugs in the i810/i830 DRM code,
which has many people reporting they no longer experience i8xx
lockups. Bug #119311 has a bit more details about which specific
kernel to try.
Changing bug state to MODIFIED. Please test and confirm wether the
problem is fixed or not, setting the bug to RAWHIDE if fixed, or
to ASSIGNED if it is still present.
Thanks in advance.
Closing. If it comes back, it should probably get a new bug report.