severity 612766 important thanks Hi Patrick, first, thanks for the tests! Bumping severity since it looks like we've got a bad issue here. Patrick Winnertz <email@example.com> (10/02/2011): > As requested from KiBi, this bug can be reproduced using the > official squeeze kernel. […] > [ 326.450] (EE) intel(0): Failed to submit batch buffer, expect rendering corruption or even a frozen display: No such device. before I forget: - Switching from libdrm/squeeze to libdrm/sid fixes this issue. - Going back to libdrm/squeeze and applying 49447a9…  (trivial conflict) on top of it changes the ENODEV issue into the EINVAL one as seen with kernels >> 2.6.32. 1. http://cgit.freedesktop.org/mesa/drm/commit/?id=49447a9b957047db2549b8a929e763bbd87808ba Next step would be trying the various versions available on: http://snapshot.debian.org/package/libdrm/ and tell us which work and which don't. That should help us narrow it down to the commit(s) we need to get this bug fixed. KiBi.
Description: Digital signature