Resume failure: ata3.01: revalidation failed (errno=-2)

Resume failure: ata3.01: revalidation failed (errno=-2)

[Tino Keitel]

Hi,

I tried if I can get reliable suspend to RAM on my new Mac mini Core 2
Duo using 2.6.29, and did a "burn in" test with an endless
suspend/resume loop using RTC wakeup. After ~30 attempts, the kernel
hangs with a screen full of this message:

ata3.01: revalidation failed (errno=-2)

In the kernel log, I see _one_ occurrence of this message at each resume:

kern.debug: ata3.01: NODEV after polling detection
kern.err: ata3.01: revalidation failed (errno=-2)

I tried the same using 2.6.27.20 and got the same, after 52 attempts.

The full kernel log is here:

http://tikei.de/kernlog.txt

This is the SATA device:

00:1f.2 IDE interface [0101]: Intel Corporation 82801GBM/GHM (ICH7
Family) SATA IDE Controller [8086:27c4] (rev 02)

Any hints how I can fix or debug this? It looks like a generic message
that only shows that something is wrong in a much deeper layer. Maybe I
should add some debugging to ata_dev_revalidate().

Regards,
Tino

Re: Resume failure: ata3.01: revalidation failed (errno=-2)

[Rafael J. Wysocki]

On Wednesday 25 March 2009, Tino Keitel wrote:
> Hi,
> 
> I tried if I can get reliable suspend to RAM on my new Mac mini Core 2
> Duo using 2.6.29, and did a "burn in" test with an endless
> suspend/resume loop using RTC wakeup. After ~30 attempts, the kernel
> hangs with a screen full of this message:
> 
> ata3.01: revalidation failed (errno=-2)
> 
> In the kernel log, I see _one_ occurrence of this message at each resume:
> 
> kern.debug: ata3.01: NODEV after polling detection
> kern.err: ata3.01: revalidation failed (errno=-2)
> 
> I tried the same using 2.6.27.20 and got the same, after 52 attempts.

Can you run this test with the current mainline, please?

Rafael