Re: [BUG] mutex deadlock of dpm_resume() in low memory situation

[Rafael J. Wysocki]

On Wed, Dec 27, 2023 at 5:08 PM Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
>
> On Wed, Dec 27, 2023 at 05:42:50PM +0900, Youngmin Nam wrote:
> > Could you look into this issue ?
>
> Can you submit a patch that resolves the issue for you, as you have a
> way to actually test this out?  That would be the quickest way to get it
> resolved, and to help confirm that this is even an issue at all.

This is a real problem, unfortunately, and the fix would require some
infra changes AFAICS.

To address it, we would need a variant of async_schedule_node_domain()
that would bail out on low memory instead of attempting to run the
stuff synchronously which is harmful (not just for the deadlock
reason) in the suspend-resume paths.

I'll try to cut a test patch shortly.