kernel 2.6.20-11,13 doesn't boot with sata-hd

Please attach the output of "lspci -vv" and "lspci -vvn"

Thanks

SATA is causing interrupts to be disabled on IRQ 16 (the same IRQ that your ethernet and USB are on).

Not sure why it is doing that. It looks very specific to your machine.

Can you try checking in your BIOS settings to see if your SATA controller can be set to a different mode (AHCI, Legacy, Native, etc) and see if changing that fixes the problem? Sometimes things like this can really be a BIOS bug.

I am affected by the same bug. I use by no means new or exotic hardware. It is an ASUS P4P800SE Mainboard with an Intel 865PE Chipset. I have no problems booting kernel 2.6.19-7 and i never had this kind of problem with any previous ubuntu release. So i am suspecting a kernel bug really. Something changed with 2.6.20. I will check out my bios settings though.

BIOS Onboard IDE Operating Mode is set to "S-ATA", because i use only SATA harddrives. With 2.6.20-x i get something like

irq18: nobody cared
disabling irq 18
...
hde: dma_timer_expiry dma status == 0x24
hde: DMA interrupt recovery
hde: lost interrupt
...
hde: dma_timer_expiry dma status == 0x24
hde: DMA interrupt recovery
hde: lost interrupt

in an endless loop during bootup. I attached a screenshot with these log messages because i cannot find them in /var/log/messages.

If i set BIOS "IDE Operate Mode" to SATA and PATA i get the same, if i set it to PATA i get something similar but with "hda" instead of "hde".

Strange thing is, my primary SATA drive has always been "/dev/sda". Now it seems to be recognized as "/dev/hde" and with "ls /dev" after booting from a herd3 cd, i do not see ANY sdx device.

In my opinion the importance of this bug should be set to "high" or even "critical", because i cannot even boot feisty with 2.6.20-x because of it right now. I do not know how many systems are affected though, but i do not consider my system especially "non-standard".

Let me know, if i can provide you with any further information.

Hey,
good to here that i'm not the only one having problems with this.

It still doesn't work with 20-6. I'm using the latest bios update from the Asus homepage and I tried playing with the settings, but it didn't change anything. I need the sata+pata mode because i use both.

btw, I agree that 'medium' is a modest importance rating for this.

I managed to solve the problem by setting 'Configure S-ATA as RAID' in the bios to 'Yes'. No idea why this works, because I have only on sata-drive and no raid.

The working IDE Configuration is:
 - Onboard IDE Operate Mode [Enhanced Mode]
 - Enhanced Mode Support On [S-ATA]
 - Configure S-ATA as RAID [Yes]
 - Serial ATA BOOTROM [Disabled]

kernel: generic 2.6.20-6.11

i can confirm the bug.

i have an intel 865pe chipset and sata and ide drives

I also agree this bug is a showstopper: I have very common hardware setup namely a dell dimension 4600, a very well sold computer.

This occurs for me on an Asus P4P800, but only with a second SATA disk connected. The first on its own is fine. The two disks are exactly the same model.

I get Donald's error.

Oh, and for my issue, I can append "irqpoll" to the kernel command line and my system boots, but does some things slowly (including bootup, but you can see what it's doing by removing the "quiet" and "splash" options).

Same problem here:

Edgy Eft is running fine on hda1.

I did a "sudo update-manager -d" to upgrade my second instance of edgy, that is situated on hda2, to feisty and it gives me the same errors as discribed by Curtis in this forum thread:
http://ubuntuforums.org/showthread.php?t=370910

My drives:
hda: ide from Samsung
hdb: ide from Maxtor
hdc: cdrw
hdd: combo
hde: sata from WD
hdf: sata from Samsung

In fact, I get the errors for hde and hdf.

The motherboard is an ASRock 775i65GV. The connectors for my 6 drives are all provided by the motherboard.

I succeeded in booting by adding irqpoll to the boot options of the 2.6.20-9 kernel. (as suggested in above)

I was asked in the thread mentioned above to add the following two bug reports:

1) output of sudo lspci -vv

And here the second report; it contains the output of
sudo lspci -vvn

I hope that the reports can be useful...

I managed to get the kubuntu herd 5 live cd to work on my Asus p4p800 deluxe. Here is the solution that I entered in the wiki at this location:
https://wiki.ubuntu.com/Asus_P4P800

You need to use F6 at the boot screen to add these command line options: noapic nolapic irqpoll All three of them. Then it works.

I will add output of lspci -vv as separate file.

Hope this helps everybody who experiences these problems.

John van Spaandonk

output of lcpci -vv

Forgot to mention that I really use the lastest BIOS, version 19
All BIOS settings are at default (enhanced mode, SATA only)
and putting it in SATA legacy mode did not solve the problem.
And here is the output of lcpci -vvn

Ok, i've just tried the installer of herd5, without the "noapic nolapic irqpoll" options the live CD doesn't boot correctly. Now I've tried to use "noapic nolapic irqpoll" and the installer seems work. I'm waiting the install process end... :-)

I just tried to boot from harddisk with the 3 options:
noapic nolapic irqpoll

At first view, I don't see any difference with booting with only the irqpoll option.

In fact, in both cases, there are no disk icons on the desktop; nor is there a trash icon on the panel. (May this is due to the fact, that my feisty is an upgrade of a not very clean edgy.)

Here some further indications that might help:

The feisty system is an upgrade from edgy. In edgy, there was kernel 2.6.17-11 that worked.

In feisty, there also is a 2.6.17-11 kernel. (I don't know whether it is the same or if it has been replaced during the upgrade)

I would like to let you know, that feisty does not start when I tell it to use the 2.6.17-11 kernel. (I did not add the irqpoll option) Unfortunately, I am not able to see the messages at startup because the computer is headless. Are these messages logged? If so, where can I find them?

And now to the main purpose of this comment: The fact, that feisty does not start with the kernel that worked under edgy (I am assuming that it is the same) makes me wonder whether it really is merely a kernel issue.

I have a normal Asus P4P800 motherboard, with an SATA HD containing Edgy and a second IDE HD on which I installed Feisty Herd 5.

When only running the IDE drive, there is no problem. When turning on the SATA drive as well, I get the "lost interrupt" and "dma_timer_expiry" errors on bootup.

After adding the options noapic, nolapic and irqpoll the system boots without problems. I have attached the output of sudo lspci -vv.

Is it normal that there are several different refference to my motherboard? E.g. P5P800-MX, P4P800-SE? I have the standard P4P800...

Could you maybe try an other development of a Linux distribution using kernel 2.6.20 or higher? For example, Fedora 7 Test 2 (https://www.redhat.com/archives/fedora-devel-list/2007-March/msg00025.html) which ships with kernel 2.6.21rc1.

If the problem also appears there, it would be worth reporting a bug upstream at http://bugzilla.kernel.org

Extremely frustrating that it's come to this, but I'm not placing fault on anyone.

I'm not about to install Fedora, but I will try with a vanilla kernel.

It would be sufficient to try if the installation cd boots afaik ;-)

They seem to have a LiveCD:
http://torrent.fedoraproject.org/torrents/f7-test2-livecd-i386.torrent

Unfortunately, my ubuntu machine is headless and I don't know how to test the LiveCD on it.

I'm using ArchLinux with the kernel 2.6.20 (vanilla) and I've no problem.

> I'm using ArchLinux with the kernel 2.6.20 (vanilla) and I've no problem.

:(

Martin,

:( indeed...

I downloaded the fedora LiveCD and it booted. During boot I got some I/O Buffer errors, see attachment, (on device sr0, don't know what that is on fedora), but that could have been the scratches on my CD ;)

Once in Gnome, I could mount my SATA and IDE drive without a problem.

Let me know if you need anything that could help you along.

Don't have any expectations, I am no kernel developer ;-)

Though, it is always good for the Ubuntu kernel developers to know if it is a problem with their kernel only. (And they are reading this ( I hope ))..

Martin

This bug /seems/ to be related: https://launchpad.net/bugs/71040

Hi all,

i had no problem with the current Fedora 7 test 2 LiveCD.
I have two SATA harddrives (and two CDROMs) and both SATA
harddrives were recognized properly (i could see them and
their partitions with ls /dev as "sda" and "sdb")
Further i could mount the partitions on the SATA harddrives.
So yes, it seems this problem is specific to ubuntu.

Why is this bug not assigned to someone? Could someone
please confirm if this bug is still tracked by development /
a bug assignment team?

Donald

The "Ubuntu Kernel Team" is subscribed to this bug report, that's sufficient :)

I wonder whether anybody, that has this problem with the sata drives, was able to find a kernel that runs properly under feisty (even if it is an older kernel)?

Or does the problem occur with every kernel that is available at the moment?

The bug description says this:

> With 2.6.19-7 everything works as expected.

The Edgy kernel 2.6.17 should work, also.

You are right: I copied the kernel 2.6.19-7 from Edgy to Feisty and tried again to boot from it in Feisty. This time it works. I don't know why it did not work the last time I tried; maybe that I did some mistake.

This is still affecting kernel 2.6.20-9. Updated bug title and description.

Could somebody please attach a dmesg from the last working kernel?

I can attach one from a booted system with the irqpoll option.

Sorry, from the last kernel that worked without irqpoll.

Matthew, i attached a dmesg from todays i.e. 07. March 2007 feisty booted with the 2.6.19-7 kernel.
I cannot boot from any of ubuntus 2.6.20-x kernels.

Booted 2.6.20-9 WITH the "irqpoll" option and got the following message:

'irq 16: nobody cared (try booting with the "irqpoll" option)'

I'll try compiling the 2.6.19-7 kernel to see what it does tomorrow (or can I get a package somewhere?)

I'm having the same problem too (booting an IDE disk, but the separate SATA disk is complaining, same as comment 8 above https://bugs.launchpad.net/ubuntu/+source/linux-source-2.6.20/+bug/78288/comments/8 .

No issues previous from Warty through Edgy

I think this issue has been mentioned on the lkml list too:
http://www.gossamer-threads.com/lists/linux/kernel/741745

Can someone confirm/deny if this is the same? I'm not overly familiar with this stuff.

I've just compiled the 2.6.19-7 kernel using the standard Herd5 kernel .config file. Booted without problems, I'll attach the dmesg and lspci output.

Compiled 2.6.21-rc3 from kernel.org, booted and gparted recognized my SATA disk, dmesg:

Hi,

With 'noapic nolapic irqpoll' option the feisty H5 install CD works ONLY when SATA enhenced mode is choosen on BIOS.
If that can help someone...

Does this also happen with kernel 2.6.20-10?

Yes.

Same thing for 2.6.20-10 here as well...

The problem also continues to occur on my computer, when using the kernel 2.6.20-10.

I am also having this problem--if I unplug my SATA (used as storage), I get a normal boot into Feisty--pluged in , the system hangs at loading drives--I'm at work, so can't try anything-----will try this evening.

I can also confirm that adding "irqpoll" to the kernel line "allows" my system to boot. I even get the SATA drive shown in Nautilus. I am using a ASUS P5P800SE motherboard (I believe that it uses the same controlers as the P4P800XX boards). If anyone wants info from me----just ask.

What about kernel 2.6.20-11-generic?

Still the same for me.

Does not boot here either with kernel 2.6.20-11.

Unfortunately, as the computer is headless, I don't see the messages during startup.

Same here:

 uname -a:
 Linux 2.6.20-11-generic #2 SMP Thu Mar 15 08:03:07 UTC 2007 i686 GNU/Linux

 My chipset is an "Intel Corporation 82875P"

 Thanks

The problem does not appear any more on my computer with kernel 2.6.20-12-generic #2

(i386)

wohooo, can anyone else confirm that 2.6.20-12 fixes the problem and close this bug report then please?

2.6.20-12 fixes my duplicate issue #91820. Thanks!

Upgrade to linux-image 2.6.20.12.8 fixed the problem for me as well! Great work!

Fixed for me too , thanks!

Yes, its fixed, thank you very much!

I wonder whether there still is some problem, because my /var/log/syslog is full of these messages:

03/21/2007 02:57:02 PM UbuntuDesktop kernel [ 2298.815372] res 51/20:03:00:00:20/00:00:00:00:00/a0 Emask 0x3 (HSM violation)

03/21/2007 02:57:02 PM UbuntuDesktop kernel [ 2298.815399] ata2: soft resetting port

03/21/2007 02:57:02 PM UbuntuDesktop kernel [ 2299.462261] ata2.00: configured for MWDMA2

03/21/2007 02:57:02 PM UbuntuDesktop kernel [ 2299.626221] ata2.01: configured for MWDMA2

03/21/2007 02:57:02 PM UbuntuDesktop kernel [ 2299.626242] ata2: EH complete

03/21/2007 02:57:02 PM UbuntuDesktop kernel [ 2299.626981] ata2: DRQ=1 with device error, dev_stat 0x59

03/21/2007 02:57:02 PM UbuntuDesktop kernel [ 2299.627012] ata2.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x2 frozen

03/21/2007 02:57:02 PM UbuntuDesktop kernel [ 2299.627021] ata2.00: cmd a0/00:00:00:00:20/00:00:00:00:00/a0 tag 0 cdb 0x0 data 0

03/21/2007 02:57:02 PM UbuntuDesktop kernel [ 2299.627022] res 51/20:03:00:00:20/00:00:00:00:00/a0 Emask 0x3 (HSM violation)

03/21/2007 02:57:02 PM UbuntuDesktop kernel [ 2299.627049] ata2: soft resetting port

03/21/2007 02:57:03 PM UbuntuDesktop kernel [ 2300.273887] ata2.00: configured for MWDMA2

03/21/2007 02:57:03 PM UbuntuDesktop kernel [ 2300.437844] ata2.01: configured for MWDMA2

03/21/2007 02:57:03 PM UbuntuDesktop kernel [ 2300.437866] ata2: EH complete

03/21/2007 02:57:03 PM UbuntuDesktop kernel [ 2300.438500] ata2: DRQ=1 with device error, dev_stat 0x59

03/21/2007 02:57:03 PM UbuntuDesktop kernel [ 2300.438531] ata2.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x2 frozen

03/21/2007 02:57:03 PM UbuntuDesktop kernel [ 2300.438539] ata2.00: cmd a0/00:00:00:00:20/00:00:00:00:00/a0 tag 0 cdb 0x0 data 0

03/21/2007 02:57:03 PM UbuntuDesktop kernel [ 2300.438541] res 51/20:03:00:00:20/00:00:00:00:00/a0 Emask 0x3 (HSM violation)

03/21/2007 02:57:03 PM UbuntuDesktop kernel [ 2300.438567] ata2: soft resetting port

03/21/2007 02:57:04 PM UbuntuDesktop kernel [ 2301.085547] ata2.00: configured for MWDMA2

03/21/2007 02:57:04 PM UbuntuDesktop kernel [ 2301.249501] ata2.01: configured for MWDMA2

03/21/2007 02:57:04 PM UbuntuDesktop kernel [ 2301.249531] ata2: EH complete

03/21/2007 02:57:04 PM UbuntuDesktop kernel [ 2301.251836] ata2: DRQ=1 with device error, dev_stat 0x59

03/21/2007 02:57:04 PM UbuntuDesktop kernel [ 2301.251865] ata2.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x2 frozen

03/21/2007 02:57:04 PM UbuntuDesktop kernel [ 2301.251873] ata2.00: cmd a0/00:00:00:00:20/00:00:00:00:00/a0 tag 0 cdb 0x0 data 0

03/21/2007 02:57:04 PM UbuntuDesktop kernel [ 2301.251875] res 51/20:03:00:00:20/00:00:00:00:00/a0 Emask 0x3 (HSM violation)

03/21/2007 02:57:04 PM UbuntuDesktop kernel [ 2301.251901] ata2: soft resetting port

03/21/2007 02:57:05 PM UbuntuDesktop kernel [ 2301.901144] ata2.00: configured for MWDMA2

03/21/2007 02:57:05 PM UbuntuDesktop kernel [ 2302.065100] ata2.01: configured for MWDMA2

03/21/2007 02:57:05 PM UbuntuDesktop kernel [ 2302.065121] ata2: EH complete

03/21/2007 02:57:05 PM UbuntuDesktop kernel [ 2302.065840] ata2: DRQ=1 with device error, d...

On Wed, Mar 21, 2007 at 02:02:00PM -0000, frafu wrote:
> I wonder whether there still is some problem, because my /var/log/syslog
> is full of these messages:
>
> 03/21/2007 02:57:02 PM UbuntuDesktop kernel [ 2298.815372]
> res 51/20:03:00:00:20/00:00:00:00:00/a0 Emask 0x3 (HSM violation)
>
> 03/21/2007 02:57:02 PM UbuntuDesktop kernel [ 2298.815399] ata2:
> soft resetting port
>
> 03/21/2007 02:57:02 PM UbuntuDesktop kernel [ 2299.462261] ata2.00:
> configured for MWDMA2
>
> 03/21/2007 02:57:02 PM UbuntuDesktop kernel [ 2299.626221] ata2.01:
> configured for MWDMA2
>
> 03/21/2007 02:57:02 PM UbuntuDesktop kernel [ 2299.626242] ata2: EH
> complete
>
> 03/21/2007 02:57:02 PM UbuntuDesktop kernel [ 2299.626981] ata2:
> DRQ=1 with device error, dev_stat 0x59

[snip]

I am not seeing messages like these on my machine. I don't think that this is
related.

-Forest

I suspected that they were related because the messages do not appear when I boot with the 2.6.17-11 kernel.

But they might also be related to the cdroms, because they appear in the drive list with the new 2.6.20-12 kernel, but they are not present in the list with the 2.6.17-11 kernel.

Today the update manager download a new 2.6.20-12 kernel. The computer still goes on writing the messages, that I showed in my comment from the 21/03/2007, in the kernel and system log.

To avoid filling up my system with logs, I continue to use kernel 2.6.17-11.

On Thu, Mar 22, 2007 at 02:05:38PM -0000, frafu wrote:
> Today the update manager download a new 2.6.20-12 kernel. The computer
> still goes on writing the messages, that I showed in my comment from the
> 21/03/2007, in the kernel and system log.
>
> To avoid filling up my system with logs, I continue to use kernel
> 2.6.17-11.

File a new bug for this problem. It is not the same bug. Your issue will get
more appropriate attention if it is represented more clearly.

-Forest

I seem to have a "new" non-bootable problem. First, let me say that the first 2.6.20-12 worked wonders with my system. For the first time I was able to boot "normally" & without "irqpoll" in the kernel line.

I have just upgraded to the "new" 2.6.20-12 Kernel (downloaded last night Pacific time) & a new showstoping problem has happened. As soon as I boot into the new kernel I get "CRC ERROR ---- System Halted"

I have just rebooted with the old 2.6.20-11 kernel & have the logs from the prior boot. What is needed?

> I seem to have a "new" non-bootable problem.

Please report this in a seperate bug. Thanks!

On my Asus P4P800 Deluxe:

2.6.20-12-generic works. Linux wcaspian 2.6.20-12-generic #2 SMP Wed Mar 21 20:55:46 UTC 2007 i686 GNU/Linux
2.6.20-13-generic doesn't work.

Adding pollirq allows 13 to boot but lots of errors in dmesg and general system sluggishness.

Linux wcaspian 2.6.20-13-generic #2 SMP Sun Mar 25 00:21:25 UTC 2007 i686 GNU/Linux

There was an ugly error that scrolled past that suggested a crash in a module but I couldn't capture it as system wouldn't boot. Adding irqpoll allows system to boot but slow and weird errors on desktop including non existant audio CD being mounted.

Linux wcaspian 2.6.20-13-generic #2 SMP Sun Mar 25 00:21:25 UTC 2007 i686 GNU/Linux

With irqpoll on boot line.

I also have high CPU utilisation which I suspect it is due to thrashing on seeks per dmesg output.

Is it possible that this bug affects IDE drives scsi emulated by kernel?
wile writing dvds with "growisofs -Z /dev/scd0=kubuntu-6.10-dvd-i386.iso" at certain point I obtain:

[ 2618.968000] ata1.01: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x2 frozen
[ 2618.968000] ata1.01: cmd a0/01:00:00:00:00/00:00:00:00:00/b0 tag 0 cdb 0x2a data 32768 out

And the process fails, (I tryed with 5 DVD+R and 1 DVD+RW). But if I blank a dvd with "growisofs -Z /dev/scd0=/dev/zero"

Everything goes fine. I attach my lspci -vvn just in case. I already reported this in bug #91210 list, but wanted to ask if there is a connection.

If this is not the place for posting this, sorry guys. I'm new reporting bugs :-P

From remotely supporting an edgy->feisty upgrade, I conclude that this bug reappeared in 2.6.20-16; Using irqpoll to boot the system and then installing 2.6.20-15 resolved the issue.

See also bug #81164 for similar reports.

I'm using a P5LD2-SE Motherboard (Intel ICH7 southbridge chipset) with a 160gb sata drive. The only way that my system boot to begin ubuntu installation is pressing F6 and add the fallow command lines: (Please remove the splash option for more details)

noapic nolapic irqpoll break=top

Then a prompt comes up, just write:

modprobe piix [hit enter]
exit [hit enter]

And the system boots ! If I remove any of those steps the system do not boot... Maybe it may help someone.

Thanks for all eforts and sorry for my bad Enghish =)

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.