Help identifying kernel-specific resume issues

[TJ]

Hi Leann.

(copying to kernel team)

I wonder if you could help me identify kernel bug reports that might be
the result of a specific resume scenario I've discovered:
"request_firmware() fails on resume from suspend"

https://bugs.edge.launchpad.net/ubuntu/+source/linux/+bug/331415

Maybe there is already a suitable tag I can use to locate reports, or
maybe an additional tag might be useful?

The scenario is that upon resume from suspend the PC will appear to have
locked up for 60 seconds before the screen has resumed. In this time
there will be no sign of hard-disk activity and the user might well
assume the system has locked up and restart the PC.

In fact, in this specific scenario, the kernel is caught in a deadlock
waiting for userspace to provide a firmware image whilst userspace
processes are still frozen.

As a result the system waits for 60 seconds (the default time-out)
before continuing.

After identifying the cause and successfully using an asynchronous
alternative for loading firmware in one driver I did a simple grep and
found over 100 other potential sites where the issue might also be
triggered. I'd need to analyse each one to discover exactly how many are
genuine trigger sites.

It occurs to me that many issues reported as 'failure to resume' might
well have the same root cause in other drivers and it would be useful to
try to identify them in launchpad.

Thanks,

TJ.