Failure to resume from S3 with 195.36.15 on Quadro NVS 140M

[bsp]

Hello everyone,

I need help with this: I have an XServer running and suspend to ram. When resuming, the primary display only shows black except some blue and red stripes at the top of the screen. A secondary display, if attached, stays black. This is happens whether I suspend directly from X or switch to another VT first. The system itself seems to be ok: The Wifi connection is established, there is hd activity, connecting by SSH is possible. Neither killing the XServer or switching to a different VT helps.

Without an XServer running, suspend/resume works flawlessly.

The same problem appears with a 190.42 driver using kernel 2.6.29. Series 185 worked just fine.

System info:
Lenovo Thinkpad T61 (7663)
Quadro NVS 140M
Kernel 2.6.33 TuxOnIce 3.1 (Gentoo)
Xorg 1.6.5
Nvidia driver 195.36.15

nvidia-bug-report-log acquired using SSH after a failed resume.

[packermann]

Hi,

seems to be a problem with your X-Server-version. I have also a NVS140M and 195.36.15 running, but my X.Org-Server is 1.7.6 (Debian) and suspend-to-RAM works here flawlessly, regardless if connected to a secondary monitor or not.

[bsp]

Aha! Found the culprit - apparently it was vbetool (not that I have an idea what that is or what it does.)

Commenting out the following lines in /etc/hibernate/ram.conf did the trick:

EnableVbetool yes
VbetoolPost yes

[AaronP]

Yes, using VBE to try to POST the GPU on resume is almost guaranteed to cause a lockup.