aacraid problems on 2.6.35-rc3

aacraid problems on 2.6.35-rc3

[Alex]

Hi,

I'm trying to compile a kernel with aacraid support on a dual Opteron
246 box with two 36GB disks on a Tyan S2891, using an initrd and lilo.

There is an older kernel on there that works just fine, so I suspect
it's something I'm missing in the kernel config. No matter what I try,
the kernel fails to find the disks on the controller. I've even tried
building it without an initrd and compiling everything into the
kernel, but it still panics.

I've included the aacraid support, SCSI disk, SCSI generic, and even
RAID1/5 support. The controller and motherboard also have the latest
BIOS updates.

What else is necessary to support aacraid?

It only prints the following when booting:

Adaptec aacraid driver 1.1-5[2456]-ms

Is I2O or ACPI support necessary?

Are there any boot-time options to increase debugging output to
troubleshoot this?

The dmesg output from a working system is here:

http://pastebin.com/0AnVBV7s

The kernel config for the attempt is here:

http://pastebin.com/gi9ybDwS

Thanks for any ideas.
Best regards,
Alex

Re: aacraid problems on 2.6.35-rc3

[James Bottomley]

On Wed, 2010-06-30 at 10:28 -0400, Alex wrote:
> Hi,
> 
> I'm trying to compile a kernel with aacraid support on a dual Opteron
> 246 box with two 36GB disks on a Tyan S2891, using an initrd and lilo.
> 
> There is an older kernel on there that works just fine, so I suspect
> it's something I'm missing in the kernel config. No matter what I try,
> the kernel fails to find the disks on the controller. I've even tried
> building it without an initrd and compiling everything into the
> kernel, but it still panics.
> 
> I've included the aacraid support, SCSI disk, SCSI generic, and even
> RAID1/5 support. The controller and motherboard also have the latest
> BIOS updates.
> 
> What else is necessary to support aacraid?
> 
> It only prints the following when booting:
> 
> Adaptec aacraid driver 1.1-5[2456]-ms

That driver comes from a rather old kernel (2.6.29 or before).  Could
you try with a modern version (which should print 1.1-5[26400]-ms just
in case the problem got fixed in the interim.

> Is I2O or ACPI support necessary?

ACPI may be necessary for the driver to attach its interrupt correctly,
but this is a motherboard issue.

> Are there any boot-time options to increase debugging output to
> troubleshoot this?
> 
> The dmesg output from a working system is here:
> 
> http://pastebin.com/0AnVBV7s

This trace shows your system won't configure interrupts correctly
without ACPI ... if ACPI isn't working, more than just the aacraid will
be non functional.  A dmesg of the failing system would be definitive in
diagnosing this.

James

Re: aacraid problems on 2.6.35-rc3

[Alex]

Hi,

>> It only prints the following when booting:
>>
>> Adaptec aacraid driver 1.1-5[2456]-ms
>
> That driver comes from a rather old kernel (2.6.29 or before).  Could
> you try with a modern version (which should print 1.1-5[26400]-ms just
> in case the problem got fixed in the interim.

Yes, my apologies. This was from the working kernel, which is 2.6.27.
I am working with 2.6.35-rc3.

>> The dmesg output from a working system is here:
>>
>> http://pastebin.com/0AnVBV7s
>
> This trace shows your system won't configure interrupts correctly
> without ACPI ... if ACPI isn't working, more than just the aacraid will
> be non functional.  A dmesg of the failing system would be definitive in
> diagnosing this.

How can I do this? The kernel panics without mounting a filesystem.

Which are the ACPI options that I should configure?

Thanks again,
Alex
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: aacraid problems on 2.6.35-rc3

[James Bottomley]

On Wed, 2010-06-30 at 10:57 -0400, Alex wrote:
> Hi,
> 
> >> It only prints the following when booting:
> >>
> >> Adaptec aacraid driver 1.1-5[2456]-ms
> >
> > That driver comes from a rather old kernel (2.6.29 or before).  Could
> > you try with a modern version (which should print 1.1-5[26400]-ms just
> > in case the problem got fixed in the interim.
> 
> Yes, my apologies. This was from the working kernel, which is 2.6.27.
> I am working with 2.6.35-rc3.
> 
> >> The dmesg output from a working system is here:
> >>
> >> http://pastebin.com/0AnVBV7s
> >
> > This trace shows your system won't configure interrupts correctly
> > without ACPI ... if ACPI isn't working, more than just the aacraid will
> > be non functional.  A dmesg of the failing system would be definitive in
> > diagnosing this.
> 
> How can I do this? The kernel panics without mounting a filesystem.

You need a serial or net console to capture the output.  At a guess, the
latest ACPI cockup is this one:

commit b430acbd7c4b919886fa7fd92eeb7a695f1940d3
Author: Len Brown <len.brown@intel.com>
Date:   Thu May 6 17:41:08 2010 -0400

    ACPICA: simplify SCI_EN workaround

Try reverting that and seeing if it boots.

James