Re: Fw: 2.6.24.X: SATA/AHCI related boot delay. - not with 2.6.24.3

Re: Fw: 2.6.24.X: SATA/AHCI related boot delay. - not with 2.6.24.3

[Tejun Heo]

Hello,

osb972ww-linuxczmil@yahoo.com wrote:
>  Greetings,
>   I'm having a problem with 2.6.23.1-42 (Fedora 8) properly driving a SATA 
> drive.  Although it is detected during boot as indicated by the dmesg below, 
> once the sytstem is up and running, the drive is not accessible.   The second 
> issue is with the CFdrive on the the primary IDE.  It works, but I can't enable 
> interrupts with hdparm -u 1 /dev/sda.   
> 
>  The interesting thing is when I put in a Fedora 5 system with a 2.6.17 kernel, 
> everything works as expected.  I can hdparm -u 1 /dev/hda and it works as  expected, and
>  the SATA drive is detected and works properly all on the exact  same hardware.  
>  
>  I've tried several different SATA drives also.   Thery all work on the old 
>  Fedora 5/ 2.6.17 system, and none work on Fedora 8/2.6.23 or 2.6.24 kernels.

Does irqpoll kernel parameter help?

-- 
tejun

Re: Fw: 2.6.24.X: SATA/AHCI related boot delay. - not with 2.6.24.3

[osb972ww-linuxczmil]

HI Tejun,

  setting irqpoll fixed both problems, and there are no "nasty grams" in dmesg now.   The original reason I was trying to use the hdparm -u on /dev/sda was because it was hanging for a few seconds when doing writes to the CFdisk, and in the past that was the kind of thing that hdparm -u would generally fix.   Now with the irqpoll kernel option, the /dev/sdb SATA drive was found, and I can mount it and write to it.  The CFcard /dev/sda is also responding to writes better than before.

Thanks for the suggestion to use irqpoll.    What is the problem within the driver(s) that is requiring irqpoll. ?

Best regards,
C. W. Wright



----- Original Message ----
> From: Tejun Heo <htejun@gmail.com>
> To: osb972ww-linuxczmil@yahoo.com
> Cc: linux-kernel@vger.kernel.org; IDE/ATA development list <linux-ide@vger.kernel.org>
> Sent: Thursday, March 20, 2008 10:23:50 PM
> Subject: Re: Fw: 2.6.24.X: SATA/AHCI related boot delay. - not with 2.6.24.3
> 
> Hello,
> 
> osb972ww-linuxczmil@yahoo.com wrote:
> >  Greetings,
> >   I'm having a problem with 2.6.23.1-42 (Fedora 8) properly driving a SATA 
> > drive.  Although it is detected during boot as indicated by the dmesg below, 
> > once the sytstem is up and running, the drive is not accessible.   The second 
> > issue is with the CFdrive on the the primary IDE.  It works, but I can't 
> enable 
> > interrupts with hdparm -u 1 /dev/sda.   
> > 
> >  The interesting thing is when I put in a Fedora 5 system with a 2.6.17 
> kernel, 
> > everything works as expected.  I can hdparm -u 1 /dev/hda and it works as  
> expected, and
> >  the SATA drive is detected and works properly all on the exact  same 
> hardware.  
> >  
> >  I've tried several different SATA drives also.   Thery all work on the old 
> >  Fedora 5/ 2.6.17 system, and none work on Fedora 8/2.6.23 or 2.6.24 kernels.
> 
> Does irqpoll kernel parameter help?
> 
> -- 
> tejun
> 

Re: Fw: 2.6.24.X: SATA/AHCI related boot delay. - not with 2.6.24.3

[Tejun Heo]

Hello,

osb972ww-linuxczmil@yahoo.com wrote:
> setting irqpoll fixed both problems, and there are no "nasty grams"
> in dmesg now.   The original reason I was trying to use the hdparm -u
> on /dev/sda was because it was hanging for a few seconds when doing
> writes to the CFdisk, and in the past that was the kind of thing that
> hdparm -u would generally fix.

libata drivers don't need or support hdparm -u settings.

>  Now with the irqpoll kernel option,
> the /dev/sdb SATA drive was found, and I can mount it and write to
> it.  The CFcard /dev/sda is also responding to writes better than
> before.
> 
> Thanks for the suggestion to use irqpoll.    What is the problem
> within the driver(s) that is requiring irqpoll. ?

Well, IRQ 18 is taken offline during USB initialization.  The IRQ is
shared with the SATA controller, so SATA controller doesn't work either.
 It isn't clear why the IRQ handler is taken offline.  It could be the
USB controller but can also be the SATA controller.  Does not loading
usb driver make any difference?

-- 
tejun

Re: 2.6.24.X: SATA/AHCI related boot delay. - not with 2.6.24.3

[Alan Cox]

On Thu, 20 Mar 2008 20:04:13 -0700 (PDT)
osb972ww-linuxczmil@yahoo.com wrote:

> 
> HI Tejun,
> 
>   setting irqpoll fixed both problems, and there are no "nasty grams" in dmesg now.   The original reason I was trying to use the hdparm -u on /dev/sda was because it was hanging for a few seconds when doing writes to the CFdisk, and in the past that was the kind of thing that hdparm -u would generally fix.   Now with the irqpoll kernel option, the /dev/sdb SATA drive was found, and I can mount it and write to it.  The CFcard /dev/sda is also responding to writes better than before.
> 
> Thanks for the suggestion to use irqpoll.    What is the problem within the driver(s) that is requiring irqpoll. ?

irqpoll works around BIOS bugs not driver bugs (well generally speaking
anyway).

Alan

Re: Fw: 2.6.24.X: SATA/AHCI related boot delay. - not with 2.6.24.3

[osb972ww-linuxczmil]

Tejun, Alan,

   Everything works when I disable the USB in the bios settings.  I was also able to remove the irqpoll kernel parameter after turning off bios USB. The particularly interesting thing is that the USB mouse still works, as does a USB 2.0 drive I have connected.  The Compact Flash drive at /dev/sda works fine, and the connected SATA drive works as well, though I need to do some testing to see how well these devices are performing. 

Using the irqpoll also made things work except for once.  In that case, I got the same errors, delayed bootup, and missing SATA drive.

Is there a new "hdparm like" tool for working with the new drives/drivers to be able to read/set parameters, run various tests, etc?

This motherboard is an IBT MB899 mini-ITX.

Thanks for the guidance with this!

C. W. Wright




----- Original Message ----
> From: Tejun Heo <htejun@gmail.com>
> To: osb972ww-linuxczmil@yahoo.com
> Cc: linux-kernel@vger.kernel.org; IDE/ATA development list <linux-ide@vger.kernel.org>
> Sent: Friday, March 21, 2008 5:56:10 AM
> Subject: Re: Fw: 2.6.24.X: SATA/AHCI related boot delay. - not with 2.6.24.3
> 
> Hello,
> 
> osb972ww-linuxczmil@yahoo.com wrote:
> > setting irqpoll fixed both problems, and there are no "nasty grams"
> > in dmesg now.   The original reason I was trying to use the hdparm -u
> > on /dev/sda was because it was hanging for a few seconds when doing
> > writes to the CFdisk, and in the past that was the kind of thing that
> > hdparm -u would generally fix.
> 
> libata drivers don't need or support hdparm -u settings.
> 
> >  Now with the irqpoll kernel option,
> > the /dev/sdb SATA drive was found, and I can mount it and write to
> > it.  The CFcard /dev/sda is also responding to writes better than
> > before.
> > 
> > Thanks for the suggestion to use irqpoll.    What is the problem
> > within the driver(s) that is requiring irqpoll. ?
> 
> Well, IRQ 18 is taken offline during USB initialization.  The IRQ is
> shared with the SATA controller, so SATA controller doesn't work either.
>  It isn't clear why the IRQ handler is taken offline.  It could be the
> USB controller but can also be the SATA controller.  Does not loading
> usb driver make any difference?
> 
> -- 
> tejun
> --
> 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: Fw: 2.6.24.X: SATA/AHCI related boot delay. - not with 2.6.24.3

[Tejun Heo]

Hello,

osb972ww-linuxczmil@yahoo.com wrote:
> Tejun, Alan,
> 
> Everything works when I disable the USB in the bios settings.  I was
> also able to remove the irqpoll kernel parameter after turning off
> bios USB. The particularly interesting thing is that the USB mouse
> still works, as does a USB 2.0 drive I have connected.  The Compact
> Flash drive at /dev/sda works fine, and the connected SATA drive
> works as well, though I need to do some testing to see how well these
> devices are performing.

Hmmm.. Can you please file a bug report against USB subsystem on
bugzilla.kernel.org and cc me?  USB on your system is triggering IRQ storm.

> Using the irqpoll also made things work except for once.  In that
> case, I got the same errors, delayed bootup, and missing SATA drive.

Hmmm.. Do you have logs?

> Is there a new "hdparm like" tool for working with the new
> drives/drivers to be able to read/set parameters, run various tests,
> etc?

hdparm will still be the utility of choice but libata doesn't allow many
settings IDE used to and libata probably will never implement some of
IDE's tunables.  Is there any specific setting you have in mind?

Thanks.

-- 
tejun