On Sun, Sep 25, 2022 at 11:50 PM Rafael J. Wysocki <rafael@xxxxxxxxxx> wrote:
>
> On Sat, Sep 24, 2022 at 6:19 PM Rafael J. Wysocki <rafael@xxxxxxxxxx> wrote:
> >
> > On Fri, Sep 16, 2022 at 6:33 AM Kai-Heng Feng
> > <kai.heng.feng@xxxxxxxxxxxxx> wrote:
> > >
> > > Commit d60cd06331a3 ("PM: ACPI: reboot: Use S5 for reboot") caused Dell
> > > PowerEdge r440 hangs at reboot.
> > >
> > > The issue is fixed by commit 2ca1c94ce0b6 ("tg3: Disable tg3 device on
> > > system reboot to avoid triggering AER"), so use the new sysoff API to
> > > reinstate S5 for reboot on ACPI-based systems.
> > >
> > > Using S5 for reboot is default behavior under Windows, "A full shutdown
> > > (S5) occurs when a system restart is requested" [1].
> > >
> > > [1] https://docs.microsoft.com/en-us/windows/win32/power/system-power-state
> > >
> > > Cc: Josef Bacik <josef@xxxxxxxxxxxxxx>
> > > Suggested-by: Dmitry Osipenko <dmitry.osipenko@xxxxxxxxxxxxx>
> > > Signed-off-by: Kai-Heng Feng <kai.heng.feng@xxxxxxxxxxxxx>
> > > ---
> > > v4:
> > > - Add comment and add more info to commit message.
> > > v3:
> > > - Use new API to invoke ACPI S5.
> > > v2:
> > > - Use do_kernel_power_off_prepare() instead.
> > >
> > > drivers/acpi/sleep.c | 8 ++++++++
> > > 1 file changed, 8 insertions(+)
> > >
> > > diff --git a/drivers/acpi/sleep.c b/drivers/acpi/sleep.c
> > > index ad4b2987b3d6e..0b557c0d405ef 100644
> > > --- a/drivers/acpi/sleep.c
> > > +++ b/drivers/acpi/sleep.c
> > > @@ -1088,6 +1088,14 @@ int __init acpi_sleep_init(void)
> > > register_sys_off_handler(SYS_OFF_MODE_POWER_OFF,
> > > SYS_OFF_PRIO_FIRMWARE,
> > > acpi_power_off, NULL);
> > > +
> > > + /*
> > > + * Windows uses S5 for reboot, so some BIOSes depend on it to
> > > + * perform proper reboot.
> > > + */
> > > + register_sys_off_handler(SYS_OFF_MODE_RESTART_PREPARE,
> > > + SYS_OFF_PRIO_FIRMWARE,
> > > + acpi_power_off_prepare, NULL);
> > > } else {
> > > acpi_no_s5 = true;
> > > }
> > > --
> >
> > Applied as 6.1 material with some edits in the changelog, thanks!
>
> And dropped due to a build failure.
>
> It looks like it depends on patch [1/2], but I haven't seen that one.
Sorry about that, let me resend it to linux-pm too.
Kai-Heng
