Hello,
I would like to report a bug that sometimes (in about 10% of cases) inhibits wakeup from suspend of my Thinkpad T61 8889-2BG. I am using Ubuntu 7.10. This bug occurs with the standard kernel as well as with a self compiled vanilla kernel (2.6.24.2) patched with thinkpad-acpi. There is another bug while waking up from suspend in the standard kernel (notebook needs several key presses and trackpoint as well as trackpad movements to wake up) but this bug seems to be solved in 2.6.24.2. Therefore, I am only reporting the bug that persists in 2.6.24.2:
The notebook suspends without any problems at all times. However, sometimes (about 10% of cases) after opening the lid or pressing the power button when the notebook is suspended, the moon stays lit, the battery LED and the power LED are turned on, the harddisk starts to spin, but then nothing else happens. The screen stays dark and there is no harddisk activity. Then, I have to power off the notebook and restart. The recommended DebuggingKernelSuspend method produces the following messages in dmesg:
[ 1.864000] Using IPI No-Shortcut mode
[ 1.864000] Magic number: 0:35:998
[ 1.864000] hash matches /build/buildd/linux-source-2.6.22-2.6.22/drivers/base/power/resume.c:57
[ 1.864000] hash matches device ttyae
[ 1.868000] hash matches device 00:00
[ 1.868000] Freeing unused kernel memory: 364k freed
With kernel version 2.6.24.2 the messages are as follows:
[ 28.217234] Using IPI No-Shortcut mode
[ 28.217367] Magic number: 0:865:403
[ 28.217368] hash matches drivers/base/power/main.c:97
[ 28.217599] Freeing unused kernel memory: 340k freed
As the unreliability of suspending renders this feature completely useless I would be very grateful if someone could help me to solve the problem. Please let me know if you need any further information.
Many thanks in advance!
Karl Martin
[lspci]
00:00.0 Host bridge [0600]: Intel Corporation Mobile PM965/GM965/GL960 Memory Controller Hub [8086:2a00] (rev 0c)
Subsystem: Lenovo Lenovo Thinkpad T61 [17aa:20b1]
01:00.0 VGA compatible controller [0300]: nVidia Corporation Quadro NVS 140M [10de:0429] (rev a1) (prog-if 00 [VGA])
Subsystem: Lenovo Unknown device [17aa:20d8]
I have tried to narrow down the problem further. This is quite time consuming as sometimes the crash does not occur in 20 suspend and resume cycles and next time it crashes after the second try.
Anyway, resume works reliably when booting with init=/bin/bash. It also works reliably if I switch to a text console and kill gdm leaving the nvidia module loaded. However, using both methods the screen stays black after resuming but the keyboard works. There were also no crashes so far using the vesa driver instead of the proprietary NVIDIA driver. I tried all combinations of the options in the acpi-support file but it did not work reliably with the proprietary NVIDIA driver. The kernel parameter acpi_sleep=s3_mode did not help too.
Nevertheless, the problem seams to be caused by the proprietary NVIDIA driver or the interaction of the kernel and this driver. Is there any way to remove this driver before suspending? Using the vesa driver is not a real alternative as it does not provide 3D acceleration and consumes much more energy from battery.