please patch linux-source-2.6.22 to fix ata_piix bug

Hi,

I don't know where to post but I've got almost the same problem with a P5W-DH and a SATA disk drive.
I upgraded from Feisty to Gutsy yesterday and after the upgrade, the system doesn't boot properly.
I got this kind of message : (initramfs) ata3.01 : failed to set xfermode (err_mask=0x40) right after grub list

I unpluged my SATA disk NEC and it finally works.

Are you designing this problem so I can patch my kernel ?
(I don't want to patch it if I don't have to)

Thanks

man this is cruddy unless .23 is coming to gutsy if not well i don't know what then i'd like this fixed in Gutsy before it ships. ugh! what a pain in the behind i's tell u..

See ubuntu forum http://ubuntuforums.org/showthread.php?t=586894
for more discussion of this item

I think I have a problem related to this. I recently upgraded from feisty (fresh install - wiped root partition) and the install worked fine.

I noticed my previous /dev/hda disk is now known as /dev/sda in gutsy, even though it is a PATA disk, but didn't think more about it at the time. Root partition is on another disk (which is serial ata, WDC WD2500KS-00M ) known as /dev/sdb (it was a scsi dev in feisty too).

The problem is my system won't cold boot (ie after it has been powered down for a while) - the kernel stops somewhere after grub and I'm dumped to a rescue shell with no access to disk. It complains about not finding root partition, in fact it can't find /dev/sdb at all. I can mount /dev/sda* fine in the rescue shell, but nothing from /dev/sdb)

The funny thing is if I reboot from the rescue shell, gusty will happily boot. This is reproducible EVERY time. I boot once, and as soon as it hangs after grub I hit ctrl-alt-delete and everything works the second time.

The IDE controller is a SiS 5513 (integrated on the mainboard)

FYI -- I also had no problems in Feisty, but have been having *serious* issues in Gutsy with booting. I update /etc/intiramfs-tools/modules to include ide-generic, but that means -- no DMA. I also have a SIS5513 IDE controller. I was thinking of recompiling my own kernel, when I found this:

From /boot/config-2.6.20-16-generic (feisty)
CONFIG_BLK_DEV_SIS5513=m

From /boot/config-2.6.22-14-generic (gutsy)
# CONFIG_BLK_DEV_SIS5513 is not set

Why was this module turned off for Gutsy? Please allow this as a module again. As a module, it can do no harm to folks with other IDE drivers, but I'm guessing it could sure relieve some grief for those of us with this particular driver!

FYI -- the absence of the SIS5513 module seems to be breaking those of us with SiS 5513 IDE controllers. I went a little outside the purely Ubuntu way of doing things, but not too far.

I downloaded the latest 2.6.22.12 version from the linux kernel archives. Then I took my Ubuntu config file, and mostly updated just the one line (also updated the kernel version tag, but nothing else). Then I recompiled the kernel.

My system now boots directly. The original poster does not list his IDE controller, so I cannot tell if this is the same problem -- but seems to fix SiS 5513 controllers.

Cheers!

I did the same as David Wynn, and compiled my own kernel which included the SIS5513 module. (following the instructions on https://help.ubuntu.com/community/Kernel/Compile )
My system seems to boot fine after that.

Problem is I can't build the restricted modules, so I had to go back to the standard -generic kernel again in order to get my X back again. :-(

Any chance we'll see an updated kernel package with this module included?

Apparently not. According to several people I have talked to on the
Ubuntu kernel team, this patch will stay where it is in 2.6.23, and will
not be applied to 2.6.22. This means we have to wait for Hardy.
On Fri, 2007-11-30 at 22:20 +0000, Isak Savo wrote:
> I did the same as David Wynn, and compiled my own kernel which included the SIS5513 module. (following the instructions on https://help.ubuntu.com/community/Kernel/Compile )
> My system seems to boot fine after that.
>
> Problem is I can't build the restricted modules, so I had to go back to
> the standard -generic kernel again in order to get my X back again. :-(
>
> Any chance we'll see an updated kernel package with this module
> included?
>

A careful look at this forum topic:
http://ubuntuforums.org/showthread.php?t=586894
will show that the original poster (mfajer) has an Intel Corporation 82801EB/ER (ICH5/ICH5R) IDE Controller, yet his Etch lsmod report includes the SIS5513 module. So, looks like people with the Intel 82801 controllers are also messed up by the removal of this module.

Hi Edward,

The Hardy Heron 2.6.24 kernel is out for testing. Care to try installing and report back your results. Thanks!

I tried. The latest 2.6.22 and 2.6.24 Ubuntu kernels both prevent my SiS 5513 driver from seeing my hard drives. They lock up very quickly after starting the kernel.

Can you please at least explain what the SIS5513 module is trying to do and why we're adamant that it must be removed from Ubuntu? I'm assuming that something has been coded that is supposed to take the place of the SIS5513 module. However, that "something" is evidently not working. It's not working in Gutsy, and its not working yet in Hardy.

Any chance we could create a kernel *with* the module compiled and ready for use in either Gutsy or Hardy?

David
----------------------------------------
> From: <email address hidden>
> To: <email address hidden>
> Date: Mon, 3 Dec 2007 20:04:19 +0000
> Subject: [Bug 133666] Re: please patch linux-source-2.6.22 to fix ata_piix bug
>
> Hi Edward,
>
> The Hardy Heron 2.6.24 kernel is out for testing. Care to try
> installing and report back your results. Thanks!
>
> ** Also affects: linux (Ubuntu)
> Importance: Undecided
> Status: New
>
> ** Changed in: linux (Ubuntu)
> Status: New => Incomplete
>
> --
> please patch linux-source-2.6.22 to fix ata_piix bug
> https://bugs.launchpad.net/bugs/133666
> You received this bug notification because you are a direct subscriber
> of the bug.

_________________________________________________________________
Share life as it happens with the new Windows Live.Download today it's FREE!
http://www.windowslive.com/share.html?ocid=TXT_TAGLM_Wave2_sharelife_112007

I tried. The latest 2.6.22 and 2.6.24 Ubuntu kernels both prevent my SiS 5513 driver from seeing my hard drives. They lock up very quickly after starting the kernel.

Can you please at least explain what the SIS5513 module is trying to do and why we're adamant that it must be removed from Ubuntu? I'm assuming that something has been coded that is supposed to take the place of the SIS5513 module. However, that "something" is evidently not working. It's not working in Gutsy, and its not working yet in Hardy.

Any chance we could create a kernel *with* the module compiled and ready for use in either Gutsy or Hardy?

David
----------------------------------------
> From: <email address hidden>
> To: <email address hidden>
> Date: Mon, 3 Dec 2007 20:04:19 +0000
> Subject: [Bug 133666] Re: please patch linux-source-2.6.22 to fix ata_piix bug
>
> Hi Edward,
>
> The Hardy Heron 2.6.24 kernel is out for testing. Care to try
> installing and report back your results. Thanks!
>
> ** Also affects: linux (Ubuntu)
> Importance: Undecided
> Status: New
>
> ** Changed in: linux (Ubuntu)
> Status: New => Incomplete
>
> --
> please patch linux-source-2.6.22 to fix ata_piix bug
> https://bugs.launchpad.net/bugs/133666
> You received this bug notification because you are a direct subscriber
> of the bug.

_________________________________________________________________
Your smile counts. The more smiles you share, the more we donate.  Join in.
www.windowslive.com/smile?ocid=TXT_TAGLM_Wave2_oprsmilewlhmtagline

I have an Athlon Classic 1200 with a SiS 5513 IDE driver. The 2.6.22 kernels in Gutsy and Hardy, and the 2.6.24 kernel in Hardy all show this problem by not recognizing my hard drive partitions on bootup -- causing a complete lock up early in the boot process. I have proven this is fixable in the 2.6.22 and 2.6.23 kernels by simply building the sis5513 module when recompiling the kernel. Perhaps there is some other more generally applicable code fix, but so far, none of the final release kernels for Gutsy have any such fix, and neither do the early versions of the Hardy kernels.

I have been able to do a partial fix by adding 'ide-generic' and 'ide-disk' to /etc/initramfs-tools/modules and running update-initramfs. This has "worked" in both Gutsy and Hardy kernels -- but it has its problems. Most importantly, ide-generic completely cancels dma on all drives. So, while this will allow me to use xorg and KDE, it is not a true final solution.

I have an active Gutsy partition, and now a Hardy partition (upgraded from Feisty this week). I would like to see this bug actively worked in Hardy to a point where I can actually use my hard drives *with* dma turned on.

We've now chosen to leave Gutsy with a known major failure for everyone with SiS 5513 and Intel 82801 EB/ER IDE drivers. Can we please get this failure fixed in Hardy? (Hint: The sis5513 module still works -- but it has been marked as not to be compiled for the Gutsy and Hardy kernels. So, if we are mandating that the module cannot be allowed in the kernel, we need to look at what its doing in order to support these drivers.)

Please keep us posted as new changes are put in place for this driver.

Cheers!

Tested recompile of 2.6.24. The SiS 5513 module as originally supplied by the kernel.org folks is still working. The 2.6.22 and 24 kernels with this module removed in Gutsy and Hardy are not working.

If there's some reason that we do not want this module -- then we still have not fixed the kernel to correctly support this hardware and we need to review our fix in Hardy.

With Gutsy being a stable system, I would not suggest any new kernel patches. But, because this has caused major regression for anyone with a SiS 5513-related controller (including certain Intel 82801 controllers), I would request that we restore the old fix so that Gutsy can actually be a working system. So, where Hardy may need to have new code to replace the module, Gutsy should just have the existing module compiled and made available to the kernel.

Cheers!

Amit, please see what it will take to get this fixed in gutsy.

As for Hardy, it is 2.6.24 based, so assuming that the proper code is there (as it claimed to be in 2.6.23)

Clarification: after reviewing bug 152820, I found that my issue was directly related to upgrading from Feisty to Gutsy, and some missing related upgrade notes. Feel free to review that bug to see how I fixed my problem. (Had to change the root partition specification in /etc/fstab)

That being said, I fear that we may have deviated from the original intent of this bug. This bug originally indicated a patch that is supposed to be fixed in 2.6.23 and beyond. Since my Gutsy system is now working with the stock kernel (after some /etc/fstab tweaks), I can't speak to whether the original issue is still a problem.

That being said -- I'm working, so I'm now content.

Cheers!

Apologies for the mess with the tasks, but at least now we have ended up with the relevant linux task and gotten rid of the duplicate version-specific ones.

Per a decision made by the Ubuntu Kernel Team, bugs will longer be assigned to the ubuntu-kernel-team in Launchpad as part of the bug triage process. The ubuntu-kernel-team is being unassigned from this bug report. Refer to https://wiki.ubuntu.com/KernelTeamBugPolicies for more information. Thanks.