Re: Regression: kernel 2.6.24{,.1} ahci problem, does not boot (resend)

Previous thread: Re: Is there a "blackhole" /dev/null directory? by Bodo Eggert on Thursday, February 14, 2008 - 1:16 pm. (1 message)

Next thread: [PATCH] arm/pxa/spitz.h rewritten and commented by Stanislav Brabec on Thursday, February 14, 2008 - 1:50 pm. (2 messages)
To: <linux-kernel@...>, <linux-ide@...>
Date: Thursday, February 14, 2008 - 1:47 pm

--MP_/Vd/D6xBF56+nJrKUKI5XFXp
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

Hello,
on one of my machines neither 2.6.24 nor 2.6.24.1 work.
The system is 64bit on Athlon X2 and ATI-Chipset (SB600).

Extract from the kernel messages during boot:

[ 66.943103] ahci 0000:00:12.0: controller can't do 64bit DMA, forcing 32=
bit
[ 66.950374] ahci 0000:00:12.0: controller can't do PMP, turning off CAP_=
PMP
[ 67.956470] ahci 0000:00:12.0: AHCI 0001.0100 32 slots 4 ports 3 Gbps 0x=
f impl SATA mode
[ 67.964996] ahci 0000:00:12.0: flags: ncq sntf ilck pm led clo pio slum =
part
[ 67.972820] scsi0 : ahci
[ 67.975699] scsi1 : ahci
[ 67.978445] scsi2 : ahci
[ 67.981178] scsi3 : ahci
[ 67.983949] ata1: SATA max UDMA/133 abar m1024@0xfadffc00 port 0xfadffd0=
0 irq 509
[ 67.991825] ata2: SATA max UDMA/133 abar m1024@0xfadffc00 port 0xfadffd8=
0 irq 509
[ 67.999729] ata3: SATA max UDMA/133 abar m1024@0xfadffc00 port 0xfadffe0=
0 irq 509
[ 68.007619] ata4: SATA max UDMA/133 abar m1024@0xfadffc00 port 0xfadffe8=
0 irq 509
[ 68.470669] ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[ 98.431945] ata1.00: qc timeout (cmd 0xec)
[ 98.454907] ata1.00: failed to IDENTIFY (I/O error, err_mask=3D0x4)
[ 98.461296] ata1: failed to recover some devices, retrying in 5 secs
[ 103.916773] ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[ 133.878045] ata1.00: qc timeout (cmd 0xec)
[ 133.882371] ata1.00: failed to IDENTIFY (I/O error, err_mask=3D0x4)
[ 133.888797] ata1: failed to recover some devices, retrying in 5 secs
[ 139.343901] ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[ 169.305174] ata1.00: qc timeout (cmd 0xec)
[ 169.309534] ata1.00: failed to IDENTIFY (I/O error, err_mask=3D0x4)
[ 169.315926] ata1: failed to recover some devices, retrying in 5 secs
[ 174.771030] ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 300)

The complete boot-log (captured via serial co...

To: Malte Schröder <maltesch@...>
Cc: <linux-kernel@...>, <linux-ide@...>
Date: Thursday, February 14, 2008 - 7:35 pm

Does irqpoll kernel parameter help?

--
tejun
--

To: Tejun Heo <htejun@...>
Cc: <linux-kernel@...>, <linux-ide@...>
Date: Friday, February 15, 2008 - 2:12 pm

On Fri, 15 Feb 2008 08:35:10 +0900
=20
No, problem stays.

--=20
---------------------------------------
Malte Schr=C3=B6der
MalteSch@gmx.de
ICQ# 68121508
---------------------------------------

To: Malte Schröder <maltesch@...>
Cc: <linux-kernel@...>, <linux-ide@...>
Date: Wednesday, February 20, 2008 - 10:51 pm

Can you capture full boot log w/ irqpoll specified? If you have root
filesystem connected to ahci, you'll probably have to use serial or
netconsole. Also, please post full boot log from 2.6.23.11.

Thanks.

--
tejun
--

To: Tejun Heo <htejun@...>
Cc: <linux-kernel@...>, <linux-ide@...>
Date: Thursday, February 21, 2008 - 12:27 pm

On Thu, 21 Feb 2008 11:51:11 +0900

I "solved" the problem by updating the BIOS. It now works perfectly. I
thought I had mailed that .. maybe I forgot.

--=20
---------------------------------------
Malte Schr=C3=B6der
MalteSch@gmx.de
ICQ# 68121508
---------------------------------------

To: Malte <maltesch@...>
Cc: Tejun Heo <htejun@...>, <linux-kernel@...>, <linux-ide@...>
Date: Monday, March 17, 2008 - 5:00 pm

I have the same problem (ASUS M2R32-MVP board with SB600 chipset,
Athlon64 X2), but I am already using the latest BIOS for this mainboard.

2.6.22.19 works
2.6.24.3 does not work
2.6.25-rc6 does not work

I can try to bisect it (maybe tomorrow).
I use AHCI mode, maybe I should try the legacy mode as well.

-Yenya

--
| Jan "Yenya" Kasprzak <kas at {fi.muni.cz - work | yenya.net - private}> |
| GPG: ID 1024/D3498839 Fingerprint 0D99A7FB206605D7 8B35FCDE05B18A5E |
--

To: Malte <maltesch@...>
Cc: Tejun Heo <htejun@...>, <linux-kernel@...>, <linux-ide@...>
Date: Tuesday, March 18, 2008 - 10:01 am

Sorry for the noise, there is even newer BIOS (dated Mar 01),
and with this BIOS my mainboard works in AHCI mode even with 2.6.25-rc6.

-Yenya

--
| Jan "Yenya" Kasprzak <kas at {fi.muni.cz - work | yenya.net - private}> |
| GPG: ID 1024/D3498839 Fingerprint 0D99A7FB206605D7 8B35FCDE05B18A5E |
--

To: Jan Kasprzak <kas@...>
Cc: Malte Schröder <maltesch@...>, <linux-kernel@...>, <linux-ide@...>
Date: Tuesday, March 18, 2008 - 11:24 pm

Hmm... I still wanna know why it got broke in the first place. The only
thing I can think of is IRQ routing problem in which case pci=nomsi or
irqpoll should help. Any chance you can test the old BIOS?

--
tejun
--

To: Tejun Heo <htejun@...>
Cc: Malte <maltesch@...>, <linux-kernel@...>, <linux-ide@...>
Date: Wednesday, March 19, 2008 - 2:19 am

Yes, I can. What data are you interested in? Boot logs with
pci=nomsi and irqpoll ? Anything else?

-Yenya

--
| Jan "Yenya" Kasprzak <kas at {fi.muni.cz - work | yenya.net - private}> |
| GPG: ID 1024/D3498839 Fingerprint 0D99A7FB206605D7 8B35FCDE05B18A5E |
--

To: Jan Kasprzak <kas@...>
Cc: Malte Schröder <maltesch@...>, <linux-kernel@...>, <linux-ide@...>
Date: Wednesday, March 19, 2008 - 2:36 am

pci=nomsi, then, irqpoll should be enough for now.

Thanks.

--
tejun
--

To: Tejun Heo <htejun@...>
Cc: Malte <maltesch@...>, <linux-kernel@...>, <linux-ide@...>
Date: Friday, March 21, 2008 - 12:43 pm

I have tried to do this, but unfortunately ASUS BIOS flash
utility does not let me to downgrade to the 0906 bios, where the problem
occurs. It lets me only downgrade from 1101 to 100x, but not to 0906
(even when called from the 100x BIOS).

-Yenya

--
| Jan "Yenya" Kasprzak <kas at {fi.muni.cz - work | yenya.net - private}> |
| GPG: ID 1024/D3498839 Fingerprint 0D99A7FB206605D7 8B35FCDE05B18A5E |
--

To: Jan Kasprzak <kas@...>
Cc: Malte Schröder <maltesch@...>, <linux-kernel@...>, <linux-ide@...>
Date: Saturday, March 22, 2008 - 4:13 am

Eee... Let's hope someone else reports again.

Thanks for the trouble.

--
tejun
--

To: Jan Kasprzak <kas@...>
Cc: Malte Schröder <maltesch@...>, <linux-kernel@...>, <linux-ide@...>
Date: Saturday, March 22, 2008 - 4:14 am

That reads a bit strange, right? That should have been "Thanks for the
trouble you put into testing." :-)

--
tejun
--

Previous thread: Re: Is there a "blackhole" /dev/null directory? by Bodo Eggert on Thursday, February 14, 2008 - 1:16 pm. (1 message)

Next thread: [PATCH] arm/pxa/spitz.h rewritten and commented by Stanislav Brabec on Thursday, February 14, 2008 - 1:50 pm. (2 messages)