Re: [ACPI] Re: RFC: ACPI/scsi/libata integration and hotswap

[Jeff G. <jg...@po...>]

Alan Cox wrote:
> On Iau, 2005-12-08 at 09:14 -0500, Jeff Garzik wrote:
> 
>>These are only for PATA.  We don't care about _GTM/_STM on SATA.
> 
> 
> Even your piix driver supports PATA. Put the foaming (justified ;))
> hatred for ACPI aside for a moment and take a look at the real world as
> it unfortunately is right now.

First, I clearly said "except on ata_piix ... or PATA"

Second, don't put words in my mouth.  I don't hate ACPI, and libata's 
direction for hotswap and suspend/resume has zero to do with "foaming 
hatred."

Right now, the top priority is getting SATA suspend/resume correct, and 
_hopefully_ doing it in a way that's friendly to PATA.  And as I said, 
we don't care about _GTM/_STM on SATA.

Further, all current ACPI proposed code is completely half-assed.  It's 
"hope and pray", because libata configures the device and does resets -- 
which is bound to CONFLICT WITH ACPI.

Even further, I want to support both ACPI cases (x86[-64]) and non-ACPI 
cases (other arches).  Some platforms want ACPI for passwords or other 
settings.  Some platforms don't have ACPI at all.  Locking libata into 
ACPI _only_ for suspend/resume is completely unacceptable.

I'm not a hope-n-pray kind of guy.  I want to get it right.  People are 
more than welcome to use unapplied patches floating around the 'net 
until we get there.

	Jeff