Created attachment 14312 [details]
When resume from suspend to ram, it appears the busy mouse cursor and the system freezes. Keyboard does not work. Although it is possible to enter by SSH, but /etc/init.d/gdm restart does nothing. Unable to Recover X.
It is stochastic, a longer suspension more likely (1 hour, 95%). If you play videos or DRI applications, even more likely freezing after suspend.
Hard and Soft:
Ubuntu 7.10 Gutsy, xserver-xorg-video-intel 2:2.1.1. Thinkpad X41 Tablet.
I also tested it with the git version, latest update <firstname.lastname@example.org> 2008-02-08 20:34:04, this always freezes, but instead of a mouse cursor appear colors. (Besides With this driver is not working DRI, possibly for not change kernel modules drm, i915 ??).
Other experiences I think are related to this driver:
1.Gdm frecuently auto reboot with X error logs, Gnome-Compiz start with problems, panel and screen appearance, disappearance,.. (This seems to change with some versions driver intel I have been testing: xserver-xorg-video-intel_2.1.1-0ubuntu10~bwh4_i386.deb, xserver-xorg-video-intel_2.1.1-0ubuntu10~pcjc2.2_i386.deb ).
2. DRI apps (Celestia, Google Earth,..) crash with xrandr rotated and freezes the whole system.
3. The new Ubuntu Hardy Alpha 4 livecd, freezes on a black screen instead of starting X.
A new clue about this bug.
When X freezes after suspend (a restarting X lockup cycle and keyboard blocked), you can enter in the machine with SSH. WELL: A REBOOT WITH KEXEC DOESN'T SOLVE THE PROBLEM. Linux reboots and X with intel driver freezes again. It's necesary a hard system reboot.
#kexec -l --append="`cat /proc/cmdline`" --initrd=/boot/initrd.img-`uname -r` /boot/vmlinuz-`uname -r`
Other tests with SSH when X is freezed:
# (Kill -9 gdms if necesary)
#nano /etc/X11/xorg.conf (Change intel to vesa driver)
#/etc/init.d/gdm start (OK. X goes with VESA)
#nano /etc/X11/xorg.conf (Change again to intel or i810 driver)
#/etc/init.d/gdm start (Freezes Again, lockup cycle)
Liken, Are you able to access the system after X freeze? Would you please verify the BusMaster Bit of PCI as bug# 14031? It's also a Thinkpad X41. thanks.
Liken, can you use the intel_reg_dumper tool from src/reg_dumper to get a snapshot of the regs after a fresh boot (before starting X) then after resume (again w/o X running) so we can see if there are differences? So:
- boot up
- capture reg dump before starting X
- start X
- chvt 1
- capture reg dump
Then attach the dumps to this bug.
Created attachment 14449 [details]
Created attachment 14450 [details]
Created attachment 14451 [details]
Created attachment 14452 [details]
Xorg.0.log when freezing
Created attachment 14453 [details]
Created attachment 14454 [details]
# mv /etc/init.d/gdm /etc/init.d/gdm_not
# intel_reg_dumper > dump_freshrebootbeforex
# lspci -vvv -xxx > lspci_freshrebootbeforex
# /etc/init.d/gdm_not start
(In Gnome I play a video full screen and run glxgears. This is to almost ensure
the freezing. Althout is not mandatory, other factors are time after wakeup)
# (Resume and Suspend. OK. X is freezed in a lockup cycle. Then SSH to the
SSH# intel_reg_dumper > dump_whenfreezed
SSH# lscpi -vvv -xxx > lspci_whenfreezed
SSH# cp /var/log/Xorg.0.log . (I830WaitLpRing(), timeout for 2 seconds,
SSH# (Reboot system with Kexec. Garbage in display)
SSH Again# intel_reg_dumper > dump_afterkexecreboot_beforex
SSH# /etc/init.d/gdm_not start (X Freezes lockup cycle)
Thinkpad X41 Tablet
intel_reg_dumper compiled from same version. (Git Version is incompatible with
Ubuntu 7.10 Gutsy (With vesafb active to text boot but to remove it does not
solve the problem. LiveCD freezes equally at suspend/resume. And in Hardy
LiveCD with intel 2.2 Black Screen instead of starting X)
This Bug #192873 is ALIVE in Hardy Alpha5 LiveCD.
I rectify previous message.
In Hardy Alpha 5 (Last Updates, installed in HD) X doesn't freezes after resume from suspend to ram. This only happens with livecd, perhaps a problem with USB CD-ROM or something.
In Hardy the bug with intel-video-driver is worse, it crashs sporadically, you suspend or not.
I report this bug, log files and intel_reg_dumper in:
IBM Thinkpad X41 Tablet
Ubuntu Hardy Last Updates
-------Significative errors in X:
---When it crashes:
Error in I830WaitLpRing(), timeout for 2 seconds
Fatal server error:
--.When it wants to restart:
(WW) intel(0): PRB0_CTL (0x0001f001) indicates ring buffer enabled
(WW) intel(0): PRB0_HEAD (0xc0a1a91c) and PRB0_TAIL (0x0001aa70) indicate ring buffer not flushed
(WW) intel(0): Existing errors found in hardware state.
I see is now exactly the bug# 14031
*** This bug has been marked as a duplicate of bug 14031 ***