Re: regression 5.6.4->5.6.5 at drivers/acpi/ec.c

[Greg Kroah-Hartman]

On Fri, Apr 17, 2020 at 09:54:54PM +0200, Rafael J. Wysocki wrote:
> On Fri, Apr 17, 2020 at 9:41 PM Toralf Förster <toralf.foerster@xxxxxx> wrote:
> >
> > On 4/17/20 8:52 PM, Rafael J. Wysocki wrote:
> > > On Fri, Apr 17, 2020 at 6:36 PM Toralf Förster <toralf.foerster@xxxxxx> wrote:
> > >>
> > >> On 4/17/20 5:53 PM, Rafael J. Wysocki wrote:
> > >>> Does the patch below (untested) make any difference?
> > >>>
> > >>> ---
> > >>>  drivers/acpi/ec.c |    5 ++++-
> > >>>  1 file changed, 4 insertions(+), 1 deletion(-)
> > >>>
> > >>> Index: linux-pm/drivers/acpi/ec.c
> > >>> ===================================================================
> > >>> --- linux-pm.orig/drivers/acpi/ec.c
> > >>> +++ linux-pm/drivers/acpi/ec.c
> > >>> @@ -2067,7 +2067,10 @@ static struct acpi_driver acpi_ec_driver
> > >>>               .add = acpi_ec_add,
> > >>>               .remove = acpi_ec_remove,
> > >>>               },
> > >>> -     .drv.pm = &acpi_ec_pm,
> > >>> +     .drv = {
> > >>> +             .probe_type = PROBE_FORCE_SYNCHRONOUS,
> > >>> +             .pm = &acpi_ec_pm,
> > >>> +     },
> > >>>  };
> > >>>
> > >>>  static void acpi_ec_destroy_workqueues(void)
> > >> I'd say no, but for completeness:
> > >
> > > OK, it looks like mainline commit
> > >
> > > 65a691f5f8f0 ("ACPI: EC: Do not clear boot_ec_is_ecdt in acpi_ec_add()")
> > >
> > > was backported into 5.6.5 by mistake.
> > >
> > > Can you please revert that patch and retest?
> > >
> > Yes, reverting that commit solved the issue.
> 
> OK, thanks!
> 
> Greg, I'm not sure why commit 65a691f5f8f0 from the mainline ended up in 5.6.5.
> 
> It has not been marked for -stable or otherwise requested to be
> included AFAICS.  Also it depends on other mainline commits that have
> not been included into 5.6.5.
> 
> Can you please drop it?

Will go do so right now, sorry about that.  Sasha, you might want to
adjust your tools a bit...