Re: [RFC]add ACPI hooks for IDE suspend/resume

[Bartlomiej Zolnierkiewicz]

On 12/7/05, Shaohua Li <shaohua.li@xxxxxxxxx> wrote:
> On Wed, 2005-12-07 at 16:49 +0100, Bartlomiej Zolnierkiewicz wrote:
> > On 12/7/05, Alan Cox <alan@xxxxxxxxxxxxxxxxxxx> wrote:
> > > On Mer, 2005-12-07 at 15:45 +0100, Bartlomiej Zolnierkiewicz wrote:
> > > > OK, I understand it now - when using 'ide-generic' host driver for IDE
> > > > PCI device, resume fails (for obvious reason - IDE PCI device is not
> > > > re-configured) and this patch fixes it through using ACPI methods.
> >
> > I was talking about bugzilla bug #5604.
> Sorry for my ignorance in IDE side. From the ACPI spec, there isn't a
> generic way to save/restore IDE's configuration. That's why ACPI
> provides such methods. I suppose all IDE drivers need call the methods,
> wrong?

>From the hardware POV:
* there is generic way to save/restores IDE device's configuration
* there is no generic way to save/restore IDE controller's configuration

>From the software POV what we only do currently is setting controller
and drive for a correct transfer mode by using host driver specific callback
(in case of using 'ide-generic' there is no such callback).

> > > Even in the piix case some devices need it because the bios wants to
> > > issue commands such as password control if the laptop is set up in
> > > secure modes.
> >
> > I completely agree.  However at the moment this patch doesn't seem
> > to issue any ATA commands (code is commented out in _GTF) so
> > this is not a case for bugzilla bug #5604.
> I actually tried to invoke ATA commands using IDE APIs, but can't find
> any available one. I'd be very happy if you can give me any hint how to
> do it or even you can fix it.

Probably do_rw_taskfile() is the method you want to use, you also need
to place invoking of ACPI provided ATA commands in the right place in
the IDE PM state machine [ ide_{start,complete}_power_step() ].

PS1 Please don't use taskfile_lib_get_identify(), drive->id
should contain valid ID - if it doesn't it is a BUG.

PS2 Have you seen libata ACPI patches by Randy?
Maybe some of the code dealing with ACPI can be put to
<linux/ata.h> and be shared between IDE and libata drivers?

Thanks,
Bartlomiej
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/