A few minutes after resuming from suspend or hibernate, the display suddenly
freezes. This is not triggered by anything obvious (such as starting a
particular program), just randomly after some key presses or mouse movements.
After display freezing, I can still ssh into the box. The entire user session
continues to run, I can start programs, etc.
I didn't see anything interesting in dmesg and Xorg.0.log, gdb stack trace is
totally useless, stracing the X server shows that it's usually waiting in an
ioctl(), and receives keyboard and mouse events. I'll attach detailled logs in
a minute.
The single change that I spotted was in the registers:
This consistently changes like this after such a freeze happens. Whenever I
look at MI_MODE in a working system, it is always 0x00000200. No other
registers change after the freeze.
This started happening two or three weeks ago. I am on Ubuntu jaunty
(development release), which closely tracks X.org upstream releases. It never
happened until then.
It definitively did not happen with -intel 2.4.1/X.org 1.5.2/Linux 2.6.27. Now
I have -intel 2.6.1/X.org 1.6.0/Linux 2.6.28.7.
My hardware:
Intel mobile 945GM
Intel Core 2 Duo 1.2
1 GB RAM
Internal 1280x800 LVDS (switched off)
External 1280x1024 TFT
Binary package hint: xserver- xorg-video- intel
A few minutes after resuming from suspend or hibernate, the display suddenly
freezes. This is not triggered by anything obvious (such as starting a
particular program), just randomly after some key presses or mouse movements.
After display freezing, I can still ssh into the box. The entire user session
continues to run, I can start programs, etc.
I didn't see anything interesting in dmesg and Xorg.0.log, gdb stack trace is
totally useless, stracing the X server shows that it's usually waiting in an
ioctl(), and receives keyboard and mouse events. I'll attach detailled logs in
a minute.
The single change that I spotted was in the registers:
--- regs.afterhiber nate.txt 2009-03-07 08:24:27.000000000 +0100
+++ regs.freeze.txt 2009-03-07 08:35:18.000000000 +0100
@@ -140 +140 @@
-(II): MI_MODE: 0x00000200
+(II): MI_MODE: 0x00000000
This consistently changes like this after such a freeze happens. Whenever I
look at MI_MODE in a working system, it is always 0x00000200. No other
registers change after the freeze.
This started happening two or three weeks ago. I am on Ubuntu jaunty
(development release), which closely tracks X.org upstream releases. It never
happened until then.
It definitively did not happen with -intel 2.4.1/X.org 1.5.2/Linux 2.6.27. Now
I have -intel 2.6.1/X.org 1.6.0/Linux 2.6.28.7.
My hardware:
Intel mobile 945GM
Intel Core 2 Duo 1.2
1 GB RAM
Internal 1280x800 LVDS (switched off)
External 1280x1024 TFT
I attached logs and more details in the upstream bug report at https:/ /bugs.freedeskt op.org/ show_bug. cgi?id= 20520