aic7xxx fails to load

Bug #41061 reported by Topper
110
Affects Status Importance Assigned to Milestone
linux-source-2.6.15 (Ubuntu)
Fix Released
High
Unassigned

Bug Description

I'm using Dapper 32bits on amd64x2.
My Boot drive is Fujitsu 10k U160 SCSI on Adaptec AHA19160.

After the last auto-upgrade from kernel 2.6.15-20-k7 to 2.6.15-21-k7 my system won't boot passed "Uncompressing linux image".
If I wait a few minutes it dies with a message like "can't find /dev/sda1" which is my boot drive.
Booting back on 2.6.15-20-k7 works just fine.
I Found another unlucky person on ubuntuforums with likely the same issue :
http://ubuntuforums.org/showthread.php?t=164873

Revision history for this message
Richard Samson (richard) wrote :

Same error on 2.6.15-21-686, aic7xxx seems broken.

I saw error on aic7xxx module loading, i will try to get them.

Revision history for this message
Petter Holmström (petter-holmstrom) wrote :

I have the exact same problem. Switching back to the old kernel "solved" the problem.

Revision history for this message
Fuctinger (fu--) wrote :

I got the same problem. System hangs at Mounting root file system. Old kernel works fine though.

Revision history for this message
Paul Otto (potto007) wrote :

I experience this problem as well. I am on an ASUS Z63a notebook with SATA connected to a Samsung MP0603H 60gb harddrive. It opens the Ubuntu splash correctly, but during the mounting root portion (the second line of the splash) it hangs for a long time, before I finally give up and reboot.

No problems when returning to 2.6.15-20-386. I haven't yet tried the 686 version.

Revision history for this message
Wilbert (wilbert-scarlet) wrote :

I've got a simular problem too. I boot from IDE but with the latest kernel I miss all my scsi drives.
I've got kernel: 2.6.15-21-386
Scsi: AHA-2940U2/U2W card with seagate drives
driver (under 2.6.15-20-386) aic7xxx

Revision history for this message
Richard Kleeman (kleeman) wrote :

Same problem here with the aic79xx driver. This may be due to the backporting of the new 2.6.17-rc aic79xx (and aic7xxx???) drivers. There have been some very substantial changes recently in both the aic79xx and aic7xxx drivers. I find however that 2.6.17-rc1/2 kernel works well so there may be some kind of source tree problem

See here https://launchpad.net/bugs/14394 last entry.

Topper (t-harley)
Changed in linux-source-2.6.15:
status: Unconfirmed → Confirmed
Revision history for this message
Wilbert (wilbert-scarlet) wrote :

I've got a simular problem too. I boot from IDE but with the latest kernel I miss all my scsi drives.
I've got kernel: 2.6.15-21-386
Scsi: AHA-2940U2/U2W card with seagate drives
driver (under 2.6.15-20-386) aic7xxx

Revision history for this message
benkillin (bennyhur) wrote :

I have the same problem after upgrading from kernel 2.6.15-20-k7 to 2.6.15-21-k7. I have an adaptec SCSI controller with segate drives. Seems no matter what I try, nothing will work except booting to 2.6.15-20-k7.

Revision history for this message
bert_3262 (bert-rijsdijk-casema) wrote :

kernel 2.6.15-21-686 has that problem to. Looking in the dmesg output is see the following error:

aic7xxx: Unknown symbol spi_populate_sync_msg
aic7xxx: Unknown symbol spi_populate_width_msg
aic7xxx: Unknown symbol spi_populate_ppr_msg

And when i compile the latest kernel-source, 2.6.15.6,-ubuntu1 it fails with error:
drivers/built-in.o: In function `ahc_construct_sdtr':aic7xxx_core.c:(.text+0x61381): undefined reference to `spi_populate_sync_msg'
drivers/built-in.o: In function `ahc_construct_wdtr':aic7xxx_core.c:(.text+0x61406): undefined reference to `spi_populate_width_msg'
drivers/built-in.o: In function `ahc_construct_ppr':aic7xxx_core.c:(.text+0x6149f): undefined reference to `spi_populate_ppr_msg'

Revision history for this message
Tim Breuer (breuer) wrote :

Same "Unknown symbol" while trying to install on a scsi harddisk. Problem is 2.6.15-20 has been removed some days ago. Can you add them again?

Matt Zimmerman (mdz)
Changed in linux-source-2.6.15:
status: Confirmed → Unconfirmed
Matt Zimmerman (mdz)
Changed in linux-source-2.6.15:
status: Unconfirmed → Confirmed
Revision history for this message
Stian Jordet (stian-web) wrote :

Indeed, 2.6.15-21-686 doesn't boot for me neither. Integrated dual aic7899.

Revision history for this message
Joerg Beyer (j-beyer) wrote :

I also had the unknown symbols (as reported before)

aic7xxx: Unknown symbol spi_populate_sync_msg
aic7xxx: Unknown symbol spi_populate_width_msg
aic7xxx: Unknown symbol spi_populate_ppr_msg

with the daily snapshot from yesterday. I wanted to install on a scsii disk - obviousliy it fails. I will try the older flight6 cd now.

Revision history for this message
Joerg Beyer (j-beyer) wrote :

any idea where i get (the 686 smp version of) linux-image 2.6.15-20 - it seems to be removed?

thanks
joerg

Revision history for this message
shadowcatcher (shadowcatcher) wrote :

I have the same probleme.
I'm upgrading my kernel to 2.6.17-rc3, SCSI drive is recognised but i couldn't mount the hd.
when i'm doing a df command, message say that's hd is in use with "udev" so i couldn't mount him.
I lose a fat partition too (same that scsi/ntfs):
"you couldn't mount, /dev/hdb1 (or /dev/sda1) already mount".

Udev problem now?

Revision history for this message
Joerg Beyer (j-beyer) wrote : Re: [Bug 41061] Re: aic7xxx fails to load

shadowcatcher schrieb:
> I have the same probleme.
> I'm upgrading my kernel to 2.6.17-rc3, SCSI drive is recognised but i couldn't mount the hd.

2.6.17 from kernel.org? I cant find a newer ubuntu package than 2.6.15.

is this an upstream bug or specific to ubuntu?

 Joerg

> when i'm doing a df command, message say that's hd is in use with "udev" so i couldn't mount him.
> I lose a fat partition too (same that scsi/ntfs):
> "you couldn't mount, /dev/hdb1 (or /dev/sda1) already mount".
>
> Udev problem now?
>

Revision history for this message
Richard Kleeman (kleeman) wrote :

2.6.17 hasn't been released yet. There is a rc3 which you get by patching a 2.6.16 version. 2.6.17-rc3 works on dapper for me really well despite my use of aic79xx. The 21 Ubuntu kernel does not.

Revision history for this message
Timo Aaltonen (tjaalton) wrote :

2.6.15-22 is affected as well. -19 is the latest I have that works.

Changed in linux-source-2.6.15:
status: Confirmed → In Progress
Changed in linux-source-2.6.15:
status: In Progress → Fix Committed
Revision history for this message
XavierB. (xbourvellec) wrote :

I'm also affected by this problem ... Too bad for my girlfriend's scanner
( My kernel is actually 2.6.15-21 btw )

Revision history for this message
Manfred Usselmann (usselmann-m) wrote :

The problem is still there with the new released 2.6.15-22 kernel.
Having only SCSI disks I still need to stay with 2.6.15-20 to be able to boot.

Revision history for this message
Thomas Hofer (th) wrote :

Ben Collins said on IRC that the fix had been commited for 2.6.15-22.34; I believe the current version is still 2.6.15-22.33.

So please do a retest when 2.6.15-22.34 arrives.

Revision history for this message
lyon (r-steiger) wrote :

2.6.15-22.33 is still the latest out. AIC7890 doesn't work on my SMP machine, so using 2.6.15-20-386 without SMP support :(

Revision history for this message
Thomas Hofer (th) wrote :

Seems like the 22.34 package has been created a few hours ago: https://launchpad.net/people/ben-collins/+packages

But I wonder why there are only amd64 and powerpc packages for 22.34 in the pool: ftp://ftp.ubuntu.com/ubuntu/pool/main/l/linux-source-2.6.15

Hmmm, hopefully the remaining packages will follow soon, maybe they're still building...

Revision history for this message
Joerg Beyer (j-beyer) wrote :

OK, just updated to 22.34 and that works for me (on 686-SMP)

thanks for fixing it.

Revision history for this message
Joerg Beyer (j-beyer) wrote :

I had the bug before, it works for me now.

Changed in linux-source-2.6.15:
status: Fix Committed → Confirmed
Revision history for this message
Richard Kleeman (kleeman) wrote :

I can confirm also that 22.34 now works for me as well with the aic79xx driver.

Revision history for this message
Thomas Hofer (th) wrote :

Joerg, you changed the bug status from "fix committed" to "confirmed". I guess you wanted to confirm that the bug is resolved now, but I think this has the exactly the opposite effect.

Revision history for this message
Joerg Beyer (j-beyer) wrote :

Thomas Hofer schrieb:
> Joerg, you changed the bug status from "fix committed" to "confirmed". I
> guess you wanted to confirm that the bug is resolved now, but I think
> this has the exactly the opposite effect.
>
oh excuse me - you are right - what can I do about it?

 sorry
 Joerg

Revision history for this message
Fuctinger (fu--) wrote :

Still doesn't work for me with 22.34. Have to use old kernel to boot.

Revision history for this message
Bruno Sabin (brunosabin) wrote :

I confirm that 22.34 fixed the issue on installation
--
bruno

Revision history for this message
Andy Brook (javahollic) wrote :

Im using AMD64, I updated from flight7 to the .22.34 and I can confirm the SCSI boot issue goes away!

Now just to fix nvida GLcore...

Changed in linux-source-2.6.15:
status: Confirmed → Fix Released
Revision history for this message
Derek Harding (dha) wrote :

For some reason for me using an HP tc2110 with scsi RAID, first the scsi_mod and aic7xxx modules are not loaded by default. So I loaded scsi_mod, fine, then aic7xxx and it gives thos unknown symbol errors. So I load aic7xxx_old and it loads fine and I can see my disks as /dev/sda.

I use French for the language and keyboard but the same issue arises with UK variations.

Going back into GUI for the install, it loads as far as the partitioner then fails with a whole load of python errors which crash the installer.

How can I do a manual install instead of GUI?

--
Regards, Derek

Revision history for this message
Derek Harding (dha) wrote :

OK - so I've found the install iso (instead of the live iso) however, the bug still remains.
:)

Revision history for this message
Andy Brook (javahollic) wrote :

Get a copy of Beta2 that didnt have this problem, a later upgrade came down that removed scsi support from the kernel. Booting into this earlier release will get you up and running, doing an 'apt-get update' and 'apt-get dist-upgrade' will bypass the problem kernel.

Check the date of your ISO, either get an older one or a later one!

Revision history for this message
Derek Harding (dha) wrote :

OK - the latest overnight build works OK (though I did pass the scsi and aic7 modules via grub to be sure) and it is now installing apparently happily.

Having no other hard disks, I couldn't make the install so apt-get update/dist-upgrade presumably wouldn't have helped.

However, can't beat the Ubuntu community for getting things sorted - thanks. :)

Revision history for this message
Fuctinger (fu--) wrote :

i got the same problem even after updating to the new kernel. Could you write
a short how-to for passing the scsi and aic7 modules via grub? Would
really appreciate it, so i can use the latest kernel, too.

Thanks in advance!

Revision history for this message
Timo Aaltonen (tjaalton) wrote :

this still is broken for some users (at least for me and Fuctinger)

Changed in linux-source-2.6.15:
status: Fix Released → Confirmed
Revision history for this message
Fuctinger (fu--) wrote :

i got the same problem even after updating to the new kernel. Could you write
a short how-to for passing the scsi and aic7 modules via grub? Would
really appreciate it, so i can use the latest kernel, too.

Thanks in advance!

Revision history for this message
Derek Harding (dha) wrote :

To pass the module from the grub command line, simply put in the module name - e.g. I add:

scsi_mod aic7xxx

I also delete quiet and change splash to nosplash so that I can see what is happening.
:)

Revision history for this message
Fuctinger (fu--) wrote :

i got the same problem even after updating to the new kernel. Could you write
a short how-to for passing the scsi and aic7 modules via grub? Would
really appreciate it, so i can use the latest kernel, too.

Thanks in advance!

Revision history for this message
Derek Harding (dha) wrote :

To pass the module from the grub command line, simply put in the module name - e.g. I add:

scsi_mod aic7xxx

I also delete quiet and change splash to nosplash so that I can see what is happening.
:)

Revision history for this message
Fuctinger (fu--) wrote :

Seems to be no solution for me. Still doesn't boot with the grub parameters. Hoping for new kernel rls....

Revision history for this message
Fuctinger (fu--) wrote :

Kernel 2.6.15-23 doesn't produce relief either. Pls correct this issue or help me with that.

Revision history for this message
Andy Brook (javahollic) wrote :

for bug history, the latest 2.6.15-23 kernel still works on AMD64 (just rebooted)

Revision history for this message
Paul Otto (potto007) wrote :

I updated to 2.6.15-23 last night, and still no luck. Still hangs waiting for the root partition.... To elaborate on my system a slight bit, my controller is the Intel ICH6 Sata controller.

I am running two custom kernels (since I can't get headers for the Ubuntu kernels 19 or 20) - 2.6.16ck9 and 2.6.17rc3... both boot and work fine.

Hope that helps a little.

-Paul

Revision history for this message
Kiz (ian-stuart) wrote :

Attempting to use Dapper-live CDs to do some prep-work before installing - Beta 2 & flight 7 both fail (flight 7 appears to older than beta 2.... kernel wise)

Revision history for this message
Timo Aaltonen (tjaalton) wrote :

I just installed an identical machine to the one that won't boot the latest kernel, so maybe it is initramfs-tools that failed to build a complete image..

Revision history for this message
Richard Kleeman (kleeman) wrote :

This problem also occured for the aic79xx driver but was fixed in the release 2.6.15 kernel. It reappeared in the Edgy 2.6.17 kernel. See bug # 50684.

Changed in linux-source-2.6.15:
status: Confirmed → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.