mbox series

[0/2] ACPI: PM: s2idle: Fix possible suspend lock-up

Message ID 2787005.CsmIKtZlk9@kreacher (mailing list archive)
Headers show
Series ACPI: PM: s2idle: Fix possible suspend lock-up | expand

Message

Rafael J. Wysocki Nov. 28, 2019, 10:42 p.m. UTC 
Hi All,

The first patch in this series is a fix for a suspend-to-idle issues introduced
in 5.4 (see its changelog for details).

The second one is more of an optimization, although some systems may need it
too (depending on how fragile their platform firmware is).

Kenneth,

This series is roughly equivalent to the patch at

https://lore.kernel.org/linux-pm/CAJZ5v0h1Ro75++4xuCznkx6GNYd+G5NpMGP96z1jdh=dm9uZbw@mail.gmail.com/T/#m11ca9a14efe4e5193bbda69767595a5fb7bd5479

The main difference is that it flushes system_wq before ec_query_wq, but that
should not really matter if the issue you saw in 5.4 is the one described
in the changelog of patch [1/2].

Please test this series and let me know if it works for you too.

In case it does work, it would be good to know if patch [1/2] is sufficient to
fix the suspend problem at least with ec_no_wakeup == 0 (it may not be
sufficient with ec_no_wakeup == 1).

Thanks!

Comments

Kenneth Crudup Nov. 29, 2019, 8:35 p.m. UTC | #1 
On Thu, 28 Nov 2019, Rafael J. Wysocki wrote:

> Please test this series and let me know if it works for you too.

I've tested this quite a few times against the issues reported on the Linux-PM
mailing list ("resume failures if the charger is plugged in while suspended
and the battery is at < 90%" and "spurious wakeups from the EC for non-power
events") and this patchset has fixed both of these issues (the latter is
especially surprising and welcome).

I'm pretty happy with it; it's been a while since I've had a consistently
cold laptop while suspended. Much thanks!

	-Kenny
Rafael J. Wysocki Dec. 1, 2019, 7:54 p.m. UTC | #2 
On Fri, Nov 29, 2019 at 9:35 PM Kenneth R. Crudup <kenny@panix.com> wrote:
>
>
> On Thu, 28 Nov 2019, Rafael J. Wysocki wrote:
>
> > Please test this series and let me know if it works for you too.
>
> I've tested this quite a few times against the issues reported on the Linux-PM
> mailing list ("resume failures if the charger is plugged in while suspended
> and the battery is at < 90%" and "spurious wakeups from the EC for non-power
> events") and this patchset has fixed both of these issues (the latter is
> especially surprising and welcome).
>
> I'm pretty happy with it; it's been a while since I've had a consistently
> cold laptop while suspended. Much thanks!

Thanks a lot for the testing and feedback, much appreciated!

I'll queue up this series for the next PM pull request.