SL-5500 crash on suspend theory

Hello,

I have got an theory regarding spontaneous SL-5500 OZ 3.5.4 crash during
suspend.

Looks like it happens, when kernel have too much data in buffers to sync
back to the media (SD card).
When buffers don't manage to get on disk in time (e.g. before the general
halt), the kernel stuck.

This also explains frequency of crash-during-suspend cases while PDA is in
craddle, when the files transfer from or to computer just occured, and no
forceful sync was done.

Some time ago I proposed (not related to stuckages) to disable lighiting and
blank screen immediately after initation of suspend procedure. That would
greatly improve user perception on device's responsiveness.

--
Arieh