Kernel module ipmi_si fails to load

[Asif]

After upgrading the kernel from 2.6.32-36 to 2.6.32-37 no more ipmi_si module available.

$ sudo modprobe ipmi_si
FATAL: Error inserting ipmi_si (/lib/modules/2.6.32-37-generic/kernel/drivers/char/ipmi/ipmi_si.ko): No such device

So I cannot run ipmitool now

$ sudo ipmitool sdr
Could not open device at /dev/ipmi0 or /dev/ipmi/0 or /dev/ipmidev/0: No such file or directory
Get Device ID command failed
Unable to open SDR for reading
---
Architecture: amd64
DistroRelease: Ubuntu 10.04
Package: ipmitool 1.8.11-1ubuntu1
PackageArchitecture: amd64
ProcVersionSignature: Ubuntu 2.6.32-37.81-generic 2.6.32.49+drm33.21
Tags: lucid
Uname: Linux 2.6.32-37-generic x86_64
UserGroups:

[Asif (vadud3)]

previous kernel has this module
~$ ls /lib/modules/2.6.32-36-generic/kernel/drivers/char/ipmi/ipmi_si.ko
/lib/modules/2.6.32-36-generic/kernel/drivers/char/ipmi/ipmi_si.ko

[Asif (vadud3)]

Looks like the module is in the new kernel as well. But still failing to modprobe

~# modprobe -f ipmi_si
FATAL: Error inserting ipmi_si (/lib/modules/2.6.32-37-generic/kernel/drivers/char/ipmi/ipmi_si.ko): Invalid module format

~# ls -al /lib/modules/2.6.32-37-generic/kernel/drivers/char/ipmi/ipmi_si.ko
-rw-r--r-- 1 root root 83024 2011-12-02 21:22 /lib/modules/2.6.32-37-generic/kernel/drivers/char/ipmi/ipmi_si.ko

[Peter Petrakis (peter-petrakis)]

We need a full apport collect to see the fall out from loading the module.

apport-collect -p linux 916200

Thanks.

[Asif (vadud3)]

Can't run apport-collect -p linux 916200, since w3m is not so visible on this remote system through ssh.
Is there any other way I can upload that data? may be apport-bug --save=/tmp/bugfile linux.

But I noticed another lucid 64bit lts server has the similar kernel's ipmi_si module
and loaded fine. So 2.6.32-37-server's ipmi_si loads fine on another system.

But 2.6.32-37-generic's ipmi_si fails to load in this system.
The md5sum are different for the ipmi_si.ko

[Stefan Bader (smb)]

They are of course different. I just checked with a resent 2.6.2-37.81 generic i386 kernel. And here the module loads just fine. Note that you had to load ipmi_devintf always to make ipmitool work. Please check dmesg after the failed load as that normally shows more detailed errors.

[Peter Petrakis (peter-petrakis)]

@Asif

Get the apport data any way you can. I merely offered a popular manner to
accomplish it. Also, please don't change bug states without justification. The
point of Incomplete is part of the bug triage process where a developer has
posed a question etc and is waiting for that request to be satisfied. It doesn't
lower the priority of your issue, in fact it draws more attention to it.

Peter

[Asif (vadud3)]

On Mon, Jan 16, 2012 at 4:01 AM, Stefan Bader
<email address hidden> wrote:
> They are of course different. I just checked with a resent 2.6.2-37.81
> generic i386 kernel. And here the module loads just fine. Note that you
> had to load ipmi_devintf always to make ipmitool work. Please check
> dmesg after the failed load as that normally shows more detailed errors.

# lsmod | grep ipmi
ipmi_devintf 8143 0
ipmi_msghandler 37083 1 ipmi_devintf

# modprobe ipmi_devintf
# modprobe ipmi_si
FATAL: Error inserting ipmi_si
(/lib/modules/2.6.32-37-generic/kernel/drivers/char/ipmi/ipmi_si.ko):
No such device

[379940.240689] IPMI System Interface driver.
[379940.240694] ipmi_si: Trying SMBIOS-specified kcs state machine at
i/o address 0xca2, slave address 0x0, irq 0
[379940.240701] ipmi_si: Interface detection failed
[379940.550664] ipmi_si: Unable to find any System Interface(s)

>
> --
> You received this bug notification because you are subscribed to the bug
> report.
> https://bugs.launchpad.net/bugs/916200
>
> Title:
>  missing kernel module ipmi_si module
>
> To manage notifications about this bug go to:
> https://bugs.launchpad.net/ubuntu/+source/ipmitool/+bug/916200/+subscriptions

--
Asif Iqbal
PGP Key: 0xE62693C5 KeyServer: pgp.mit.edu
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?

[Asif (vadud3)]

apport information

[Asif (vadud3)]

apport information

[Asif (vadud3)]

Here is the output of apport-bug attached.

[Stefan Bader (smb)]

Weirdly seems to skip some probing state.

@Asif, could you please do the following: boot back into the -36 kernel version (which had ipmi working). Not sure whether it can simply be done via iKVM but otherwise you could modify the set default="0" line in /boot/grub/grub.cfg to ="2" and reboot.
Then verify ipmi is up and post the dmesg from that boot. Thanks.

[dann frazier (dannf)]

fyi, it looks like this could occur if a misbehaving bmc's status register contained 0xff. That's another reason to check to see if the older kernel works. If the previously working kernel also shows this problem, I'd suggest removing power from the system to see if you can reset the BMC.

[Stefan Bader (smb)]

True, on my supermicro this would also be possible through the web interface,

[Asif (vadud3)]

On Wed, Jan 18, 2012 at 11:31 AM, dann frazier
<email address hidden> wrote:
> fyi, it looks like this could occur if a misbehaving bmc's status
> register contained 0xff. That's another reason to check to see if the
> older kernel works. If the previously working kernel also shows this
> problem, I'd suggest removing power from the system to see if you can
> reset the BMC.

This was working on -36 kernel and stop working right after the kernel upgrade.

I will try to schedule a downtime to reboot back to older kernel and verify
and collect the dmesg as Stefan suggested.

>
> --
> You received this bug notification because you are subscribed to the bug
> report.
> https://bugs.launchpad.net/bugs/916200
>
> Title:
>  missing kernel module ipmi_si module
>
> To manage notifications about this bug go to:
> https://bugs.launchpad.net/ubuntu/+source/ipmitool/+bug/916200/+subscriptions

--
Asif Iqbal
PGP Key: 0xE62693C5 KeyServer: pgp.mit.edu
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?

[Samantha Jian-Pielak (samantha-jian)]

This issue occurs on 12.04.2, too. It's run on an OCP (Open Compute Project) v2 system.
$ lsmod |grep ipmi
ipmi_devintf 17708 0
ipmi_msghandler 45979 1 ipmi_devintf
$ sudo modprobe ipmi_si
FATAL: Error inserting ipmi_si (/lib/modules/3.5.0-25-generic/kernel/drivers/char/ipmi/ipmi_si.ko): No such device

[62158.420904] IPMI System Interface driver.
[62158.420989] ipmi_si: Adding default-specified kcs state machine
[62158.420995] ipmi_si: Trying default-specified kcs state machine at i/o address 0xca2, slave address 0x0, irq 0
[62158.421003] ipmi_si: Interface detection failed
[62158.422671] ipmi_si: Adding default-specified smic state machine
[62158.422679] ipmi_si: Trying default-specified smic state machine at i/o address 0xca9, slave address 0x0, irq 0
[62158.422686] ipmi_si: Interface detection failed
[62158.446639] ipmi_si: Adding default-specified bt state machine
[62158.446647] ipmi_si: Trying default-specified bt state machine at i/o address 0xe4, slave address 0x0, irq 0
[62158.446656] ipmi_si: Interface detection failed
[62158.446804] ipmi_si: Unable to find any System Interface(s)

$ modinfo ipmi_si
filename: /lib/modules/3.5.0-25-generic/kernel/drivers/char/ipmi/ipmi_si.ko
...

I will upload the apport info shortly.

[Stefan Bader (smb)]

Samantha, could you load this module ever before on that HW?

[Samantha Jian-Pielak (samantha-jian)]

@stefan,
I just realized that this is a new hw, and will need a new driver to get it to work. Please ignore the info I provided.

[Raja Sriram (grssriram008)]

@Samantha , Can you provide the driver details ?

[Jörg Frings-Fürst (jff-de)]

Hello,

The version 1.8.11 is not longer supported. The current version is 4.0.17.

I close this bug

To close I set the status to "Invalid". If the bug still occurs please change the status to 'Confirmed'.

Jörg

[Nicolas Krzywinski (nsk7even)]

This is just an information for those that come to here due to matching search engine keywords, it is only partially related to this bug, so all others please ignore this.

If your system does not power off at all (including hang at reboot) you may have fallen into this trap:
- install ipmitool
- recognise that your system does not meet the requirements for ipmi_si (matching errors as described above)
- remove ipmitool

Now ipmi_si is still present in the system and you can see the correspondig startup messages in system.log and it seems to be able to prevent poweroff.

Solution: PURGE ipmitool (and maybe openipmi as well - at least there was "remaining configuration" present in my system for this package as well and therefore I purged it additionally).

After that, my system was able to reboot again and was able to completely poweroff on its own:
CPU: Clarkdale i3 540
Board: ASUS P7H55D-M Evo
OS: Mythbuntu 12.04.4