Hi  Matt,

-----Original Message-----
From: linux-pm-owner@vger.kernel.org [mailto:linux-pm-
owner@vger.kernel.org] On Behalf Of Matt Fleming
Sent: Friday, March 04, 2016 9:33 PM
To: Chen, Yu C
Cc: linux-acpi@vger.kernel.org; linux-efi@vger.kernel.org; linux-
pm@vger.kernel.org; rjw@sisk.pl; lenb@kernel.org; tglx@linutronix.de;
x86@kernel.org; Zhang, Rui; Zheng, Lv
Subject: Re: [PATCH][RFC] ACPI / PM: Fix poweroff issue on HW-full
platforms without _S5

On Wed, 02 Mar, at 09:29:17PM, Chen Yu wrote:

quoted

The problem is Linux registers pm_power_off = efi_power_off only if we
are in hardware reduced mode. Actually, what we also want is to do
this when ACPI S5 is simply not supported.
That should handle both the HW reduced mode, and the HW-full mode
where the DSDT fails to supply an _S5 object.

This patch fixes this issue by introducing a new flag acpi_no_s5 which
indicates the non-existence of _S5. The initial state of acpi_no_s5 is
false and probed in acpi_sleep_init, then we'll later see the updated
value in efi_poweroff_required, according to which we can set
pm_power_off to efi_power_off in efi_shutdown_init.

Suggested-by: Len Brown <redacted>
Signed-off-by: Chen Yu <yu.c.chen@intel.com>
---
 arch/x86/platform/efi/quirks.c | 3 ++-
 drivers/acpi/sleep.c           | 2 ++
 include/acpi/acpixf.h          | 6 ++++++
 3 files changed, 10 insertions(+), 1 deletion(-)

Are there legacy platforms without _S5 where we currently handle reboot
via some other means that are going to break if we attempt an EFI reboot?

I ask because historically performing an EFI reboot has been all kinds of buggy
on x86.

I made a double check and  yes there are other components would like to
register their customized pm_power_off if _S5 is not available.
So if EFI reboot is not stable, current code may  not be robust enough .
Hum, how about this one?
(assigned to efi reboot only when there is no other components registered pm_power_off)
 return  acpi_gbl_reduced_hardware ||
	(acpi_no_s5 && ! pm_power_off);

Also, is this what Windows does if _S5 is absent?

I'm not sure, Windows might also leverage other drivers to achieve a poweroff. 

yu