stock 2.6.22 kernels don't detect both IDE HDDs anymore

This is the diff from 2.6.22-13-rt (stock) to my "kind of working" version. There are lots of changes, but I guess the only ones that may be relevant are the ones in the IDE section (the other potentially relevant changes in the list of partition types handled didn't change anything in practice - I tried that first because I thought it might be caused by my LDM partitions).

Affects 2.6.22, not 2.6.20 - don't know why I qualified it as regarding 2.6.20.

Perhaps related to Bug 133666, which documents a known issue with ATA / IDE drive detection. That bug also lists a patch, but the patch has not been applied to Ubuntu yet.

oh! so this is prob. why i'm having the probs i am having. I had .13 working fine until after some updates - small no kernels of course and that's when they stopped on me, well tho it could be related or not, i don't know but .12 is the only one working for me now and I'm two kernels behind of course.

I will admit, I haven't confirmed this in *exactly* a Ubuntu way -- but I have confirmed the issue, and the fix is very simple -- at least in my case, it was much easier than the diff shown. The key is right here in the diff:

-# CONFIG_BLK_DEV_SIS5513 is not set
+CONFIG_BLK_DEV_SIS5513=m

Looking back at my own system, triple boot: Windows ME (got tired of paying the tax); Feisty; Gutsy.

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

I took the absolute last version of the 2.6.22 kernels from the official linux kernel archives (as opposed to the Ubuntu source). I then took my config-2.6.22-14-generic and allowed for the SIS5513 module to be built. I rebuilt the kernel and headers and installed them. I then rebooted my computer.

With the Ubuntu 2.6.22-14, my computer always hangs very quickly -- the progress bar shows a couple millimeters before the crash. Then it just takes several minutes to finally drop to a command line. With my newly compiled kernel with the SiS 5513 module available, the computer boots all the way to a normal command line (I'd probably have X, but I don't have the restricted drivers yet for the new kernel).

I am convinced that the solution here, and for everyone else dealing with SiS 5513 controllers is that the current compile of the kernel does not make the needed module available to us.

This will be retargeted towards the Hardy kernel once it is released. I've tagged this as "hardy-kernel-candidate" so that we make sure to retarget this report once the new release is out. However against the linux-source-2.6.22 and linux-source-2.6.20 packages this is being marked as "Won't Fix" as it does not meet the criteria for a stable release update. To learn more about the stable release update process please refer to https://wiki.ubuntu.com/StableReleaseUpdates . Thanks!

I don't understand how this does not meet the criteria. I see this in the criteria:

* Bugs which represent severe regressions from the previous release of Ubuntu

Now, this may not be a "severe regression" for everyone. But for many that own a computer with a SIS5513 IDE controller, going from a Feisty install where everything works, to a Gutsy install, where the system can't even boot because the IDE controller cannot recognize the hard drives -- that's a very severe regression! No solution has been presented that allows me to boot my computer with its SiS 5513 IDE controller without recompiling the kernel to include the necessary IDE module.

So, here's what I know in regard to the requested documentation:

1) Impact of the bug: Computers with SiS 5513 IDE controllers have reported everything from complete inability to boot, to very flaky disk access, even locking up the system. One proposed work-around was to force the ide_generic and ide_disk modules -- a "solution" that prevents the use of DMA in CD- / DVD-ROM drives connected to the computer. Thus, an unacceptable solution. That from my personal experience coupled with reports seen on the web forums.

2) I don't know of the bug being addressed in the "development" branch. But I can look back to former stable branches!!!

From the Feisty stock .config file:
CONFIG_BLK_DEV_SIS5513=m

From the Gutsy stock .config file:
# CONFIG_BLK_DEV_SIS5513 is not set

3) The patch: well, just correct the one line in the .config file in the kernel. That's it. Basically -- the module that the kernel folk have written for the SiS 5513 IDE controller should be allowed to be used by those of us with such a controller. Since this is built as a module, I would expect it to only be used by computers with the appropriate hardware. (Note: I tested with the ubuntu .config file, but with the latest 2.6.22 official Linux kernel -- if any Ubuntu-specific changes have been made to this module, there may be problems I am not aware of)

4) Reproducing the bug. I migrated my computer up from a Feisty install, so this *might* be a problem specific to migration, rather than new installs, but I doubt it. First: get a computer with a SiS 5513 IDE controller, as the error is specific to this hardware. Then try a) a fresh install with Gutsy, and b) a fresh install with Feisty, followed by a migration to Gutsy. I would expect both to cause problems. (I remember trying the new install. My computer crashed during the install, though I don't recall the specifics of where) If you're able to reboot successfully off of such a computer, and hdparm reports that DMA is settable for any CD- / DVD drives on the IDE controller, then something else is the problem. But I think you're going to find issues.

5) Regression potential: If you have a computer with a SiS 5513 IDE controller, migrating to Gutsy can bring your computer to a complete halt. If you can *somehow* tweak your computer to boot, even with this controller, than you'll probably find that the drives are not all working as they optimally should. This should only affect the users of this particular style of IDE controller, but the probalem for computers us...

mfajer, original poster of this thread:
http://ubuntuforums.org/showthread.php?t=586894
has an Intel 82801 EB/ER IDE controller, but his lsmod report from Etch shows that this also uses the SIS5513 module. So, it appears that the Intel 82801 EB/ER driver is also broken by dropping this module.

One line of changes in the .config file -- which would only impact those that actually utilize the module -- could sure fix a lot of problems. Here's a partial listing of what appear to be issues directly related to this missing module:

From the forums:
http://ubuntuforums.org/showthread.php?t=586894

Bug 123789, Kernel 2.6.22-x will not boot on laptop

Bug 131945, Bad recognition of DD by ata_piix

Bug 133666, please patch linux-source-2.6.22 to fix ata_piix bug

Bug 152820, Upgrade to linux-image-2.6.22-14 kernel renders my PC unbootable

Bug 155702, pata_sis fails to enable udma100, which causes a long (~1min) delay at boot. Patch available!

Bug 163753, Strange ATA HD configuration

Bug 164722, kernel 2.6.22 crash on toshiba m30x-127

Major regression issue still in Gutsy for all SiS 5513 controllers, and some Intel 82801 controllers. For Gutsy, we are still requesting that the old patch be put back into the kernel -- that is: restore the availability of the sis5513 module.

If there is something we are attempting to fix by removing the sis5513 module -- we failed. And the current 2.6.24 kernels in Hardy do not fix the issue either. Please review what the sis5513 module is doing because I have recompiled the 2.6.24 kerne for my Hardy kernell -- and the module works fine for those of us with SiS 5513 controllers. But removing the module makes such a system unbootable.

Cheers!

Sorry -- I neglected to change the 'Baltix' default.

Hi David,

If you can, please do not task kernel bugs against linux-source-2.6.24. Beginning with the Hardy release, the kernel source package naming convention has changed from linux-source-2.6.24 to just linux. Closing task against linux-source-2.6.24. Thanks.

I'm marking this as a duplicate of 133666. Thanks.

Please note: this is related to upgrade. Please also see bug 152820. There are issues with the upgrade that do not exist with a fresh install. I had to make additional updates to /etc/fstab. These updates should have been identified in the upgrade notes, but they were not.

So, after the changes I indicated in 152820, I am now working -- with the stock Gutsy kernels.

Cheers!