310.19: "NVML Unknown error" after resume from failed hibernate

Video re-init fails after the following chain of events:
0: Ensure that an app is using enough ram to require swap (i.e. RAM usage > physical RAM amount)
1: Run sudo pm-hibernate while in an xterm.

TuxOnIce fails with the message that there’s insufficient space. On resume, video re-initialization fails.

The latest version that was tested was 310.19 and I believe this has been an issue since at least 296.36, if I recall correctly.

The GPU in use is a GeForce 9600M GS (G96), on a Centrino 2 chipset.

nvidia-bug-report.sh was run from a remote terminal by logging in via ssh and relevant nvidia-bug-report.log.gz is attached.
nvidia-bug-report.log.gz (63.3 KB)