Re: FW: bisected regression: v3.6-rc1: resume from s2ram does not restore ata_piix (v3.5 worked)

Re: FW: bisected regression: v3.6-rc1: resume from s2ram does not restore ata_piix (v3.5 worked)

[Aaron Lu]

[-- Attachment #1: Type: text/plain, Size: 3754 bytes --]

Hi Sergei,

The only problem I can see is the offending commit didn't do a gtm for
IDE channel during init. It was used to be done in
ata_acpi_associate_ide_port.

So can you please test if the following code fix your problem? Thanks.

diff --git a/drivers/ata/libata-acpi.c b/drivers/ata/libata-acpi.c
index 902b5a4..0f338bb 100644
--- a/drivers/ata/libata-acpi.c
+++ b/drivers/ata/libata-acpi.c
@@ -1101,6 +1101,9 @@ static int ata_acpi_bind_host(struct ata_port *ap,
acpi_handle *handle)
if (!*handle)
return -ENODEV;

+ if (ata_acpi_gtm(ap, &ap->__acpi_init_gtm) == 0)
+ ap->pflags |= ATA_PFLAG_INIT_GTM_VALID;
+
return 0;
}

Thanks,
Aaron

> ---------- Forwarded message ----------
> From: Sergei Trofimovich <slyich@gmail.com>
> Date: Mon, Aug 13, 2012 at 5:20 PM
> Subject: bisected regression: v3.6-rc1: resume from s2ram does not
restore ata_piix (v3.5 worked)
> To: linux-pm@lists.linux-foundation.org, Matthew Garrett
<mjg@redhat.com>, Holger Macht <holger@homac.de>, Lin Ming
<ming.m.lin@intel.com>, Jeff Garzik <jgarzik@redhat.com>
> Cc: linux-kernel@vger.kernel.org
>
>
> It's a laptop compaq 2510p (~5 years old core2 laptop) with a single
SATA drive:
>
> 00:1f.1 IDE interface: Intel Corporation 82801HM/HEM (ICH8M/ICH8M-E)
IDE Controller (rev 03)
> Subsystem: Hewlett-Packard Company Device 30c9
> Kernel driver in use: ata_piix
>
> kernel v3.5 worked fine. 3.6-rc1 resumes, but disk stays inaccessble.
> Seems to be 100% reproducible. Bisection gave sane result[1].
>
> I was not able to revert it as-is, thus couldn't verify the revert
helps on top of master.
>
> Do you need more info?
>
> Thanks!
>
> [1]
> commit 30dcf76acc695cbd2fa919e294670fe9552e16e7
> Author: Matthew Garrett <mjg@redhat.com>
> Date: Mon Jun 25 16:13:04 2012 +0800
>
> libata: migrate ACPI code over to new bindings
>
> Now that we have the ability to directly glue the ACPI namespace to the
> driver model in libata, we don't need the custom code to handle the same
> thing. Remove it and migrate the functions over to the new code.
>
> Signed-off-by: Matthew Garrett <mjg@redhat.com>
> Signed-off-by: Holger Macht <holger@homac.de>
> Signed-off-by: Lin Ming <ming.m.lin@intel.com>
> Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
>
> :040000 040000 6a659a9d4a92b2085f6d0b58484cb2f82cd12cfa
> 125fe5d5fa8b208a08792b03e752571d825465d2 M drivers
> :040000 040000 b7c3819be4e82ae6e2ac9688055aeb2bc1bc4ebd
> 9cbc8fd15b147a178f723bcdb9e7c34f9868400f M include
>
> git bisect good 492d542273a4859f8bf8cc7744cdf71ef50b39ea
> # bad: [354b2eac3848bddbcb111079138b907ccca70ae8] libata-acpi: fix up
for acpi_pm_device_sleep_state API git bisect bad
354b2eac3848bddbcb111079138b907ccca70ae8
> # bad: [dc7f71f486f4f5fa96f6dcf86833da020cde8a11] sata_dwc_460ex:
> device tree may specify dma_channel
> git bisect bad dc7f71f486f4f5fa96f6dcf86833da020cde8a11
> # bad: [91e4d5a1d7d11ca0b08803a11cb8dc866d2d611f] drivers/acpi/glue:
> revert accidental license-related 6b66d95895c bits git bisect bad
91e4d5a1d7d11ca0b08803a11cb8dc866d2d611f
> # bad: [3bd46600a7a7e938c54df8cdbac9910668c7dfb0] libata-acpi: add ata
port runtime D3Cold support git bisect bad
3bd46600a7a7e938c54df8cdbac9910668c7dfb0
> # bad: [30dcf76acc695cbd2fa919e294670fe9552e16e7] libata: migrate ACPI
code over to new bindings git bisect bad
30dcf76acc695cbd2fa919e294670fe9552e16e7
> # good: [6b66d95895c149cbc04d4fac5a2f5477c543a8ae] libata: bind the
Linux device tree to the ACPI device tree git bisect good
6b66d95895c149cbc04d4fac5a2f5477c543a8ae
> # good: [6b66d95895c149cbc04d4fac5a2f5477c543a8ae] libata: bind the
Linux device tree to the ACPI device tree git bisect good
6b66d95895c149cbc04d4fac5a2f5477c543a8ae
> 30dcf76acc695cbd2fa919e294670fe9552e16e7 is the first bad commit
>
> --
>
> Sergei


[-- Attachment #2: Type: text/html, Size: 6194 bytes --]