Re: sata suspend resume ... (fwd)

["Randy.Dunlap" <rdunlap@xenotime.net>]

On Tue, 25 Apr 2006 08:54:21 +0800 (SGT) Jeff Chua wrote:

> 
> On Sun, 23 Apr 2006, Mark Lord wrote:
> 
> > Try Randy Dunlop's libata-acpi patches -- I've been using variants of them
> > for a *very long time* here now, as they're the only thing that works for me.
> 
> This the one that makes does it!!! I've tried just about all patches 
> including the famous "mdelay(2000);" but none worked. ... perhaps it's 
> just something to do with IBM X60s which is dual-core.
> 
> Here's what I did to make it work...
> 
> Linux is vanilla linux-2.6.17-rc2.
> 
> Pavel suggested the right step as I was going no where with SMP suspend. 
> Switching to UP and using "suspend2" (CONFIG_SUSPEND2) didn't work either 
> with PIIX or AHCI. Patch is suspend2-2.2.5-for-2.6.17-rc2.tar.bz2
> 
> Then I tried "suspend" (CONFIG_SOFTWARE_SUSPEND) with both PIIX and 
> AHIC in UP mode ... still didn't work.
> 
> Next, I applied libata-acpi.patch with a little modification. No sure 
> what the implication is but the original patch seems to use "i" outside 
> the loop. Here's what I added to libata-acpi.patch ... can someone verify 
> this or do we need this at all?
> 
> +
> +       for (i = 0; i < ATA_MAX_DEVICES; i++)
> +               ata_acpi_push_id(ap, i);

Yes, that looks appropriate.  The function that code lives in
has changed quite a bit since this patch was written.

> Tried both PIIX and AHCI in UP mode. Failed with AHCI, but worked in 
> "suspend to disk" in PIIX mode (configured in BIOS under SATA -> 
> "COMPATIBILITY").

---
~Randy