So obviously one of the quirks in /usr/lib/pm-utils/sleep.d works around this kernel bug. We need to find out which. Can you please move the files in that directory to some other place one by one and do a "sudo pm-suspend" / resume cycle until it starts failing? That's the "magic" hook then. Which one is it?
After that exercise, move the hooks back to /usr/lib/pm-utils/sleep.d/.
So obviously one of the quirks in /usr/lib/ pm-utils/ sleep.d works around this kernel bug. We need to find out which. Can you please move the files in that directory to some other place one by one and do a "sudo pm-suspend" / resume cycle until it starts failing? That's the "magic" hook then. Which one is it?
After that exercise, move the hooks back to /usr/lib/ pm-utils/ sleep.d/ .
Thanks!