[ltp] T60p: Kernel patches required for suspend to ram?

Kristian Berge Nessa linux-thinkpad@linux-thinkpad.org
Mon, 9 Oct 2006 21:30:59 +0200


Torsdag 05 oktober 2006 23:40 skreiv Kristian Berge Nessa:
> Onsdag 04 oktober 2006 20:35 skreiv Henrique de Moraes Holschuh:
> > On Wed, 04 Oct 2006, Kristian Berge Nessa wrote:
> > > So, in essence, the question is: What magic have Suse done to their
> > > (2.6.15 based) kernel in order for this to work, that is not yet in the
> > > mainline 2.6.18 kernel?
> >
> > 1. AHCI suspend/resume
> > 2. ACPI support for SATA ACPI handles (those hit LKML very recently)
> > 3. MSI/MSI-X fixes, PCI config space fixes (AFAIK, 2.6.18 has those)
>
> Assuming that: AHCI [1] is not needed; (I have a 'ahci' module, but it is
> not loaded, and 'ata_piix' is the only one who talks to 'libata'),
> and that 2.6.18 has the MSI/MSI-X [3] fixes,
> only [2] remains (or so I thought...).
>
> Further assuming that [2] referred to the "Return of the ACPI patches" as
> posted by Kristen Accardi to linux-ide last thursday (2006-09-28);
> these seems to have made it into the -mm3 patchset.
>
> So I compiled a 2.6.18-mm3 kernel, but got exactly the same behaviour. :(


Still at loss:
Tried a 2.6.19-rc1 kernel, with the same results...
Tried to enable AHCI (Since the AHCI patches are supposedly in 2.6.19-rc1), 
and different combinations of ec_intr and acpi_sleep settings.
Tried to remove all the modules and services I could, but still the same...

Perused through the list of patches SuSE applied to their 2.6.16.13-4 kernel 
(which works wrt. suspend), but found nothing (to my eyes) applicable at 
first sight that were not already in some way or another in 2.6.19-rc1.

Enabled all debugging options wrt. PM, but found nothing that looked 
suspicious. The laptop silently dies with the line (never to wake up again):
Oct  9 20:51:23 albion kernel: PM: Preparing system for mem sleep

According to the Suspend/resume event tracer (CONFIG_PM_TRACE), the last PM 
event point was at the last port of my Intel 82801G IHC7 PCI Bridge (00:1c.3 
0604: 8086:27d6 (rev 02)). Don't know if that rings any bells anywhere...


-- 

Mvh,
Kristian Berge Nessa
kristian@cavatina.no

Cavatina Software AS
Einervn. 2A, 4056 Tananger
Tlf.: 414 49 991