Regressions in sata_nv regarding STANDYBY IMMEDIATE.

Regressions in sata_nv regarding STANDYBY IMMEDIATE.

[Peter Dons Tychsen]

Hello,

Recent kernels broke suspend/hibernate completely on one of my setups
after updating to more recent kernels (3.3). It used to work perfectly
and the machine was hardly ever powered completely off (standby).
(older M2N-E ASUS motherboard with nvidia MC55 controller).

The symptoms where that the system would hang entering hibernate. The
only complaint from the system was that command STANDBY_IMMEDIATE had
failed.

At first i assumed that the disk (an older WD Green disk) was dying, and
replaced the disk with a cheap Seagate disk. However, this turned out
not to be the case.

Looking through the ATA code i found that STANDBY had been replaced with
STANDBY_IMMEDIATE as it was supposed to be more compliant. Replacing it
with the old STANDBY made the standby procedure work again.

Now the question is: The STANDBY and STANDBY_IMMEDIATE commands are for
the disk. Why does it specifically break for disks on MCP55 and not on
other chipsets? Does it make sense?

FWIW both disks seem to suspend fine on the same controller using an
other O/S.

Later i also discovered that the new disk actually made things worse.
Using the hack suspend worked again for the old disk. The new disk could
now suspend but would mostly get stuck getting out of suspend as it
would seem that the system caused it to stop responding all together
requiring a hard reboot.

Any ideas?

Thanks,

/pedro

Re: Regressions in sata_nv regarding STANDYBY IMMEDIATE.

[Robert Hancock]

On 04/25/2012 01:12 AM, Peter Dons Tychsen wrote:
> Hello,
>
> Recent kernels broke suspend/hibernate completely on one of my setups
> after updating to more recent kernels (3.3). It used to work perfectly
> and the machine was hardly ever powered completely off (standby).
> (older M2N-E ASUS motherboard with nvidia MC55 controller).
>
> The symptoms where that the system would hang entering hibernate. The
> only complaint from the system was that command STANDBY_IMMEDIATE had
> failed.
>
> At first i assumed that the disk (an older WD Green disk) was dying, and
> replaced the disk with a cheap Seagate disk. However, this turned out
> not to be the case.
>
> Looking through the ATA code i found that STANDBY had been replaced with
> STANDBY_IMMEDIATE as it was supposed to be more compliant. Replacing it
> with the old STANDBY made the standby procedure work again.

I'm not aware of this being changed recently. If it's this commit you're 
referring to, it was back in 2007:

http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git;a=commit;h=78981a7c6c34bddbb90da72cf6ce10953e84aad8

What was the last kernel that was known to work OK?

>
> Now the question is: The STANDBY and STANDBY_IMMEDIATE commands are for
> the disk. Why does it specifically break for disks on MCP55 and not on
> other chipsets? Does it make sense?

Not really, but these NVIDIA SATA controllers (especially the pre-AHCI 
ones like this) are finicky beasts.

>
> FWIW both disks seem to suspend fine on the same controller using an
> other O/S.
>
> Later i also discovered that the new disk actually made things worse.
> Using the hack suspend worked again for the old disk. The new disk could
> now suspend but would mostly get stuck getting out of suspend as it
> would seem that the system caused it to stop responding all together
> requiring a hard reboot.
>
> Any ideas?
>
> Thanks,
>
> /pedro
>
>
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-ide" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>

Re: Regressions in sata_nv regarding STANDYBY IMMEDIATE.

[Peter Dons Tychsen]

Hi Robert,

Thanks for the comments,

On Sat, 2012-04-28 at 00:30 -0600, Robert Hancock wrote:
> What was the last kernel that was known to work OK? 

I will try to find the time to bisect it properly.

Thanks,

/pedro