linux-image

[X3]

This HAPPENS WITH ANY KARMIC KERNEL EVEN 2.6.31.14
ProblemType: Bug
Well this is really silly from a stable supposedly release that has been tested for months. AND has been reported FIXED on Bug #296478 I'm here to tell you its not fixed, does not work, period.
uname -r is 2.6.31-15-generic ____ KARMIC KOALA 9.10
I have tried this with default files as system is installed with minimal karmic install I get error FATAL: Error inserting coretemp (/lib/modules/2.6.31-15-generic/kernel/drivers/hwmon/coretemp.ko): Invalid module format and I have tried it with the description below. FATAL ERRORS EXIST.
dmesg outputs this [ 4563.803572] coretemp: no symbol version for module_layout WTH??
I have this issue and further more on jaunty kernel 2.6.28-16 modprobe coretemp worked perfectly after I replaced sensors-detect in /usr/sbin/ with this file http://www.lm-sensors.org/svn/lm-sensors/trunk/prog/detect/sensors-detect chmod 755 this file and added a pre-patched coretemp.ko from http://mabene.icomedias.com/coretemp.ko and place it in its path which varies on kernel name but is in my instance /lib/modules/2.6.31-15-generic/kernel/drivers/hwmon/ obviously I renamed the existing files by doing e.g. filename.extension-orig. JUST FOR TESTING PURPOSES.
This is my sensors-detect output The modules load during that so it has to work but in no obvious way.
http://pastebin.com/f2ae9c41b
LM-SENSORS is Version: 1:3.1.1-4~ppa0
sensors -s outputs No sensors found!
Make sure you loaded all the kernel drivers you need.
Try sensors-detect to find out which these are.
modprobe fails with errors as described.
doing /etc/init.d/lm-sensors start or service lm-sensors start outputs * Setting sensors limits [fail] should say ok
when I do # modprobe coretemp or # modprobe -v coretemp I get the same silly error FATAL: Error inserting coretemp (/lib/modules/2.6.31-15-generic/kernel/drivers/hwmon/coretemp.ko): Invalid module format.
This worked fine in Jaunty kernels with no issues in EITHER CASE and all my CPU temps were posted when doing # sensors
Doing sensors-detect is able to see/load modules which are coretemp, ic2-nforce2, ic2-dev I cannot load any of them using modprobe ic2-nforce2, ic2-dev it says FATAL: Module ic2_dev not found.
makes anyone wonder if bug reports are addressed or even read in some cases. silly bugs are allowed to travel down for years without being fixed in some cases.
ok I get it its all free and ppl do it in their spare time they have families/jobs/live like most of us do well I do anywho... just dont get it why this is reported and not fixed. or when ppl just add it as a duplicate of e.g. Bug #296478 and so sometimes even having dozens of 'duplicate' bugs that when u read have nothing in relation in the majority of cases. As a obvious result it gets ignored. So much for support.
The only kernel which fixes this is the 2.6.32rc7 which I have downloaded from mainline for testing. though the system is only accessible via ssh and stays at blinking cursor on machine.

[X3 (x3lectric)]

What no one cares about this at all? Why bother reporting it?

[Andy Whitcroft (apw)]

@X3 -- we had close to 3000 bugs filed in a 1 week period following release, regardless of the number of people assigned to the task you cannot have an instantaneous response to any one bug in that pile. We are not always able to get to bugs in a timely fashion but that does not mean we do not care. Putting sensible titles on bugs and having all the apport data attached, helps to get things to the top of the heap, duplicates do indeed make a bug more likely to be found.

[X3 (x3lectric)]

Well I have supplied the apport file and all the information I found pertinent. I realise that whoever is dealing with this cant jump at any 1 report, I get it ppl have all sorts of things in live to do.

My comment regarding the no one cares part, really came from the OLDER bug I linked to on my post where it was reported fixed and indeed its not if you check down that original bug report to a way older Ubuntu Distro.

I understand also that have two instances probing the same bus can break things & potentially damage hardware and that from 2.6.29 has sorta changed in regards to that so this does not happen. However it does not explain why a older kernel fixes it like the 2.6.32rc7. Nor does it explain why froa very long time this was reported on the other report and no official reply was given.

I apologise if it sounded ungrateful at sme stage, I hope this gets looked at and the necessary patches applied so that everything works as expected.

This has forced me to stop doinbg Karmic based custom minimal installs for the issues posted even before adding any sensors monitoring sw it already presents conflicts with SMbus access. I can only blame it on testers not reporting it properly though in this instace its a silent on boot comment which is apparently not logged.

Heres hoping I can get Karmic posted on my plentifull requests for custom systems atm Im no longer running Karmic besides on VM.

[Andy Whitcroft (apw)]

This error is rather peculiar as it implies that the module file is corrupt on this system:

  FATAL: Error inserting coretemp (/lib/modules/2.6.31-15-generic/kernel/drivers/hwmon/coretemp.ko): Invalid module format.

Cirtainly I am able to insert this module on my karmic systems:

  apw@dm$ lsmod | grep coretemp
  coretemp 7072 0

Could you attach the errant coretemp.ko file itself so we can compare it to those it should match.

[X3 (x3lectric)]

its worth mentioning that Im using a minimal install on a Asrock ION 330 HT and theres many issues with that system and karmic minimal install. as I install Karmic minimal Im greeted straight away with conflict with ic2 smbus conflicts before I get terminal prompt to log in. Maybe this is platform specific which would make things worst but facts are what I have reported is real.

pls find attached the file requested neither files work.

on a side note I have also found way too many bugs associated with minimal installs so many that I don't have time to document and report them all.

[X3 (x3lectric)]

ah also pls note Jaunty works 200% ok with same install and I get to ic2 smbus conflicts nor anything else described here.

Im also aware that the IO ports being access by both acpi and lm sensors is not good however my findings are that 2.6.32 kernel fix all the issues described here

[Thomas Novin (thomasn80)]

The original coretemp.ko is almost 1/5 the size of the custom one, what's so special with the custom module?

[Thomas Novin (thomasn80)]

If I run strings(1) on both these files they are from different kernel versions and report different strings.

$ strings coretemp.ko | grep vermagic
vermagic=2.6.28.9 SMP mod_unload modversions 586
$ strings coretemp.ko-orig | grep vermagic
vermagic=2.6.28-17-generic SMP mod_unload modversions 586

Should these be loadable on a 2.6.31 kernel?

[X3 (x3lectric)]

well honestly like I said karmic was such a nightmare that it is practically impossible to run on this machine I mentioned the coretemp.ko since I cannot run karmic and its too much of a headache to even try fixing it.

as for the size discrepancy, one is plain vanilla as it comes with install which doesnt support my hardware the other is a specific version which supports more hardware in this case my machine.

The -orig is the original file of course 19kb and the other is 95kb. No idea why the files would report different strings or why they do.

they both should load just fine on the karmic kernel when I compiled the 2.6.32 from kernel from scratch these files as they are both load no issues (all systems go) as for the 2.6.31 kernel like I said no way in hell will it load neither file nor will it resolve other issues with the system like the ic2-smbus conflicts...

Really unless I create a custom minimal install disk with the 2.6.32 kernel and install that I see no other solution until this kernel is official and a proper minimal install can be used installing the .32 kernel instead of the offending ones which are now whats on offer labelled stable.

[Brad Figg (brad-figg)]

This bug was filed against a series that is no longer supported and so is being marked as Won't Fix. If this issue still exists in a supported series, please file a new bug.

This change has been made by an automated script, maintained by the Ubuntu Kernel Team.