Ubuntu10.10: kernel 2.6.35-22 cannot find intern hard drives.

[Trigger]

I made an upgrade from Ubuntu 10.04 to Ubuntu 10.10. The upgrade went well, no error or such, but when time went to reboot, it stucks just after Grub.
I can still boot with the older kernels (2.6.32-25).

The error message with the recent kernel (2.6.35-22) is:
"udevd-work[173]: `/sbin/modprobe -bv pci:<lots of letters and digits>` unespected exit with status 0x0009
Gave up waiting for root device. Common problem:
-Boot args (cat /proc/cmdline)
-Check root delay
-Check root
-Missing modules (cat /proc/modules; ls /dev)
ALERT! /dev/disk/by-uuid/<disk ID> does not exist. "

In /dev (with the newest kernel), no sda or sdb is present, and the "by-uuid" folder is empty.

Following this link : http://lkml.org/lkml/2010/6/16/323 I found a temporary solution : by inserting pci=nocrs in the kernel line in grub, I can boot with the kernel 2.6.35-22 (from where I post the bug report).

Feel free to ask me more questions or to request more information if that may help you to find the source and a way to correct this.

Thank you very much for all your work.

ProblemType: Bug
DistroRelease: Ubuntu 10.10
Package: udisks 1.0.1+git20100614-3
ProcVersionSignature: Ubuntu 2.6.35-22.34-generic-pae 2.6.35.4
Uname: Linux 2.6.35-22-generic-pae i686
NonfreeKernelModules: nvidia
Architecture: i386
Date: Tue Oct 12 14:46:52 2010
MachineType: Dell Inc. Precision WorkStation T3500
ProcCmdLine: BOOT_IMAGE=/boot/vmlinuz-2.6.35-22-generic-pae pci=nocrs root=UUID=1ea4de6c-1e4d-4701-ac4f-0d6eebc83b63 ro splash quiet
ProcEnviron:
 PATH=(custom, no user)
 LANG=fr_FR.UTF-8
 SHELL=/bin/bash
SourcePackage: udisks
Symptom: storage
Title: Internal hard disk partition cannot be mounted manually
dmi.bios.date: 04/12/2010
dmi.bios.vendor: Dell Inc.
dmi.bios.version: A07
dmi.board.name: 0XPDFK
dmi.board.vendor: Dell Inc.
dmi.board.version: A01
dmi.chassis.type: 7
dmi.chassis.vendor: Dell Inc.
dmi.modalias: dmi:bvnDellInc.:bvrA07:bd04/12/2010:svnDellInc.:pnPrecisionWorkStationT3500:pvr:rvnDellInc.:rn0XPDFK:rvrA01:cvnDellInc.:ct7:cvr:
dmi.product.name: Precision WorkStation T3500
dmi.sys.vendor: Dell Inc.

[hpeck (hp-eck-dv)]

I had a similar effect after upgrading kubuntu from 10.04 to 10.10.
After the first reboot I found, that the kernel reported my former /dev/sda as /dev/sdf and /dev/sdb as /dev/sde. Because I had all filesystems exept /home mounted via UUID, I got the system running but without /home so I could fix it easily by changeing the entry in fstab to mounting by UUID also for the /home filesystem.

To me it seems there was introduced a bug into the kernel changeing the startvalue for enumerating the attached disks, at least for SATA.

[Trigger (triggerds)]

Hi Hpeck,

I don't think our bugs are related: in both cases, the system cannot find the disks, but in your case it is because the name have changed from one version to the other, in my case all my disks are already referenced by their uuid, so no such problem. For me, the system cannot see any disk at all, not even a /dev/sdf or /dev/sde.

I think you should open another bug to signal this one.

[Avi Carmi (avi-carmi)]

Hi all,

I have the same identical problem as Trigger (and Matt, see: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/653238).

Same DELL Precision T3500

there is a new A08 BIOS dated 9/17/10 available from Dell, which I'll install as soon as I am done posting.

http://support.dell.com/support/downloads/download.aspx?fileid=416864

however... found this too in https://bugzilla.kernel.org/show_bug.cgi?id=16228

> FYI: I just updated my T3500 to the latest BIOS A08 from Dell. It still doesn't
> boot 2.6.34 without the proposed patches. So still no BIOS fix...

the above also have list of patches, but I'll have to recall how to apply patches and compile the kernel, last time I've done it was years ago RHEL 5 or 6, so spoiled now by Ubuntu

found this yet in another thread: https://bugs.edge.launchpad.net/ubuntu/+source/linux/+bug/647043/comments/14

which has a link to a patched test kernel, with the same list of patches:

> I had Lisa do a series of tests. We subsequently discovered a series of patches that have been submitted upstream but are a bit > invasive (ie. they won't land in upstream 2.6.36, and will likely have to wait for 2.6.37).
>
> http://marc.info/?l=linux-kernel&m=128476278029918&w=2
> Patch : https://patchwork.kernel.org/patch/189182/
> Patch : https://patchwork.kernel.org/patch/189232/
> Patch : https://patchwork.kernel.org/patch/189242/
> Patch : https://patchwork.kernel.org/patch/189252/
>
> I've subsequently built a test kernel with the above patches applied. Lisa has confirmed this fixes the issue and the
> USB ports work with this test kernel. The test kernel can be found at the following:
>
> http://people.canonical.com/~ogasawara/lp647043/i386/
>
> I'll have to discuss with the Ubuntu Kernel SRU team if they will qualify for a Stable Release Update to Maverick.

-avi

[Avi Carmi (avi-carmi)]

the A08 BIOS update did not resolve the problem.

 pci=nocrs does work with the Ubuntu kernel.

did not yet try the test patched kernel (got to get some real work done...)

-avi

[Stan Ionascu (stan-io)]

Same problem, as the reported describes, same DELL Precision T3500.

[Stan Ionascu (stan-io)]

But the system boots with pci=nocrs.

[kikoune (kikoune)]

Same error here with a DELL Precision T3500. Can boot Ubuntu Maverick 10.10 with option pci=nocrs and 2.6.35 kernel or with previous 2.6.32 kernel or

[ScArcher2 (scott-archer-gmail)]

I'm seeing the same problem with my Dell Precision T3500. (Also a co-worker with the same machine has the same problem).

Are there any side effects to the pci=nocrs workaround?

[Trigger (triggerds)]

Hi All,

I am not so sure that this bug is a duplicate of bug #647043 "Dell Studio 1536 Unable to detect USB ports (Maverick)" because
all my USB devices (keyboard, USB key) were recognized, detected and fully functional. The only devices not detected are the internal hard drive, not an USB one.

If Wawrzyniec Niewodniczański could argue to explain why he thinks that these bugs are in fact the same it would be great.

Thanks.

[Wawrzyniec Niewodniczański (wawrzek)]

This comment https://bugs.launchpad.net/ubuntu/+source/linux/+bug/658560/comments/7 confirm that patch from bug #647043 resolves your issue, therefore, resolving that bug will address your problems as well.

I can agree that using term 'duplicate' is not 100% right in this situation, the better name for relation would be 'depends on' or 'related to', but I don't think such options exist in Launchpad.

BTW. I marked 'duplicate' more bugs based on https://bugs.launchpad.net/ubuntu/+source/linux/+bug/663182/comments/9. IMHO it is easier to follow 1 bug entry with patch rather than 6. So maybe there should be new bug entry titled 'Kernel regression causing trouble for some Dell boxes' with list of the boxes and issues they have.

[Bjorn Helgaas (bjorn-helgaas)]

This is a duplicate of bug #653238, not of bug #647043.

Bug #653238 was marked a duplicate of bug #647043, but that is incorrect.