Re: ahci_softreset prevents acpi_power_off

[Tejun Heo <htejun@gmail.com>]

Jeff Garzik wrote:
> Tejun Heo wrote:
>> Arjan van de Ven wrote:
>>> I'd be interested in finding out how to best test this; if the bios is
>>> really broken I'd love to add a test to the Linux-ready Firmware
>>> Developer Kit for this, so that BIOS developers can make sure future
>>> bioses do not suffer from this bug...
>>
>> As reported, this is almost a butterfly effect.  ->softreset method is
>> only used during initialization and error recovery of ATA devices which
>> has almost nothing to do with the rest of the system.  This is almost
>> like 'changing my mixer input to line-in makes power off fail'.  (it's
>> more related due to ATA ACPI stuff and maybe that's why this happens but
>> I'm trying to make a point here.)
> 
> It's quite possible that the BIOS in question wants AHCI in some
> specific state at poweroff.

I would be surprised if this weren't an accident.  We reset the
controller during initialization, so whether softreset or hardreset is
used, the end status cannot be much different.  And, I really don't
wanna change ahci and/or libata for this.

-- 
tejun