Created attachment 517818 [details]
Screenshot of kernel panic during boot process
Description of problem:
I recently performed a yum update to update my system to kernel-PAE 2.6.40-4.
The yum update appeared to run just fine,but upon rebooting I receive the following error message:
/bin/sh: error while loading shared libraries: libgcc_s.so.1: cannot open
shared object file: No such file or directory
[ 5.229739] Kernel panic - not syncing: Attempted to kill init!
[ 5.229806] Pid: 1, comm: init Not tainted 2.6.40-4.fc15.x86_64 #1
... call trace ...
Version-Release number of selected component (if applicable):
How reproducible: Always
Steps to Reproduce:
1. Yum update to new kernel
3. Observe you are unable to boot.
Actual results: System does not boot but stops with a kernel panic error message.
Expected results: Normal boot.
See the attached image.
Maybe it's not working for me because I use some older hardware?
I have a regular 3.0 GHz Intel Pentium 4 processor on a Intel D865GLC
I erased and reinstalled this kernel twice but i still get kernel panics during
I also used yumdownloader to download the kernel rpm amd did a "yum
localinstall" but it doesn't help either.
the older kernels boot fine.
BTW, in relation to BUG #729526 (https://bugzilla.redhat.com/show_bug.cgi?id=729526), while erasing kernel-2.6.40-4.fc15.x86_64, yum also removed kmod-nvidia-2.6.40-4.fc15.x86_64 which was not reinstalled. I do have the akmod-nvidia package installed.
There may be an issue with the 2.6.40 kernel and adding modules to it.
> Created attachment 517818 [details]
> Screenshot of kernel panic during boot process
> Description of problem:
> I recently performed a yum update to update my system to kernel-PAE 2.6.40-4.
> The yum update appeared to run just fine,but upon rebooting I receive the
> following error message:
> /bin/sh: error while loading shared libraries: libgcc_s.so.1: cannot open
> shared object file: No such file or directory
Something is going wrong in the creation of your initramfs, and dracut is exiting which panics the kernel (pid 1 can't ever exit). If you look in /boot, do the initramfs-<version>.img files all seem of comparable size (around 14MB) or is the bad one really small, etc?
I checked that.
They all seem of comparable size.
Sorry for my bad english.
I have the same problem.
I have a regular 2.6 GHz Intel Pentium 4 processor on a Intel D865GBF
After upgrade to kernel-PAE-2.6.40-4.fc15.i686 system does not boot.
Just updated to kernel 22.214.171.124-5 in the hope that it would solve this problem. But unfortunaly, it didn't. The same problem is still there!
Is someone working on it yet?
(In reply to comment #5)
> Just updated to kernel 126.96.36.199-5 in the hope that it would solve this problem.
> But unfortunaly, it didn't. The same problem is still there!
> Is someone working on it yet?
There isn't much information to go on in the bug. We'll need to see more of the boot process in order to determine what might be happening. Do you have a serial port on this machine that could be used as the console to capture the boot logs?
Yes, the system has a serial port.
How do I have to do that?
(In reply to comment #7)
> Yes, the system has a serial port.
> How do I have to do that?
Juast adding rdinitdebug to the kernel boot options might be enough to see what's happening.
I was able to capture boot process via ttyS0.
Please see the attached test1.txt, test2.txt and test3.txt.
Created attachment 523152 [details]
Created attachment 523153 [details]
Created attachment 523158 [details]
test 3 (with rdinitdebug kernel parameter)
I provided logs.
Is somebody working on this issue now?
When is this going to be solved?
I can't install newer kernels now.
Can you post the boot log of a working kernel, as well as providing the output of acpidump from that kernel?
Also, have you run memtest86+ on this machine to eliminate bad memory?
Created attachment 526628 [details]
output of acpidump
Created attachment 526629 [details]
boot.log of working kernel
Created attachment 526630 [details]
/var/log/messages of working kernel
(In reply to comment #15)
> Can you post the boot log of a working kernel, as well as providing the output
> of acpidump from that kernel?
> Also, have you run memtest86+ on this machine to eliminate bad memory?
I have created the log files as you requested. Memtest did not produce errors.
I've tried to reproduce this on some of the P4 machines we have here, but none of them hit this issue. Bug 727865 also shows similar errors for similar machines, so at the moment we suspect it's isolated to either a few particular models of Pentium 4 machines, or a particular BIOS version for them, or similar.
We know 2.6.38 works and 2.6.40 (3.0) does not. It would be helpful to know if 2.6.39 worked. Would you be willing to install this older kernel and see if it boots?
Until we can either recreate here or find someone with an afflicted machine that is willing to do a git bisect, this is going to be somewhat hard to narrow down.
> We know 2.6.38 works and 2.6.40 (3.0) does not. It would be helpful to know if
> 2.6.39 worked. Would you be willing to install this older kernel and see if it
This is a F16 kernel, my distr is F15 (current). It seems to me it is impossible to install F16 kernel on F15, is it?
(In reply to comment #21)
> > We know 2.6.38 works and 2.6.40 (3.0) does not. It would be helpful to know if
> > 2.6.39 worked. Would you be willing to install this older kernel and see if it
> > boots?
> > http://koji.fedoraproject.org/koji/buildinfo?buildID=244663
> This is a F16 kernel, my distr is F15 (current). It seems to me it is
> impossible to install F16 kernel on F15, is it?
No. Just download it and use:
yum localinstall <path to rpm>
Okay, I downloaded and installed kernel 2.6.39 per as in yor comment #21.
It gives the same result, it crashes on boot.
So far, the kernels up to 2.6.38 work, up to 188.8.131.52-35 to be exactly.
(In reply to comment #23)
> Okay, I downloaded and installed kernel 2.6.39 per as in yor comment #21.
> It gives the same result, it crashes on boot.
And it crashes with the same backtrace as in comment #12?
> So far, the kernels up to 2.6.38 work, up to 184.108.40.206-35 to be exactly.
If it's the same crash, that's somewhat good news. At least we've narrowed it down a bit.
If you could bear with us and test a few more kernels, we might be able to narrow this down even more. I've listed the 2.6.39 builds below in increasing order of the -rc release. If you install and boot and note if it's good or bad that would be helpful.
The first 2.6.39 kernel release (rc0 - 236996) in the links from comment #25 doesn't already work, so, do I have to continue testing them?
Well, that's it for me guys.
I give it up.
I have only P4 machines running here and they all have problems with latest kernels so I am going to change to some other distro or perhaps say goodbye to linux.
To bad though. I always was convinced linux being far more stable then Microsoft's Windows but it seems it's the other way around now these days.
Can somebody please close this bug for me?
Bye bye RedHat/Fedora, Had some fun for more then 15 years.
(In reply to comment #27)
> Well, that's it for me guys.
> I give it up.
> I have only P4 machines running here and they all have problems with latest
> kernels so I am going to change to some other distro or perhaps say goodbye to
That's somewhat unfortunate.
I would honestly be very surprised if any distro running a 2.6.39 or newer kernel worked on that machine. We really aren't carrying any patches in Fedora that would impact this.
If you do find a distro that works using a 3.x kernel, it would be greatly appreciated if you let us know.
As a someone who encountered an issue with my Fedora 15 machine and was successfully helped by the Redhat team, I say thank you.
To the original poster, I would say that the P4 is quickly (in computer time) becoming a "dusty deck" system given that the last P4 was shipped in 2008 and the architecture was first introduced in 2000.
The ISA is still supported in newer chips which after 12 years are probably at least 6x more powerful. You could potentially consolidate your P4 systems into a single newer Intel multicore processor based system and save on SWaP (size, weight and power) costs. This cost savings could provide reinvested into future upgrades or other pressing needs.
the last line should read:
"This cost savings could be reinvested into future upgrades or other pressing needs."
LOL: On the Fedora project homepage:
400 MHz Pentium Pro or better processor (comparable AMD, Cyrix, and VIA processors also acceptable)
At least 768 MB memory (RAM), 1 GB recommended for best performance
10 GB disk space / storage
Well I think the Fedora folks will need to update this information.
I think as with most FOSS projects ... caveat emptor or should it be caveat venditor? (either way, let the user beware).
I give this one more try.
With kernel parameter maxcpus=0 every kernel boots fine.
PS: A new machine is ordered already.
I only want this bug to be kept open because it must be solved for other P4 users as well.
I have now a P4 (Intel D865PERL + 3.2 GHz processor) especially running for this bug only.
Please continue the investagation,
I am not Fedora user but I am VERY interested in that this bug is going to be fixed.
I have definitely the same problem: Intel P4 3GHz, Intel D865PERL.
My bug report in Launchpad is here:
I have tested almost all main series after 2.6.38 and none of them work for me.
I am also compiling kernels by myself to find the revision that broke the kernel.
By now it is sure that the revision is between 2.6.38 and 2.6.39rc1.
Feel free to contact me it it is needed.
(In reply to comment #35)
> I am not Fedora user but I am VERY interested in that this bug is going to be
> I have definitely the same problem: Intel P4 3GHz, Intel D865PERL.
> My bug report in Launchpad is here:
> I have tested almost all main series after 2.6.38 and none of them work for me.
> I am also compiling kernels by myself to find the revision that broke the
> By now it is sure that the revision is between 2.6.38 and 2.6.39rc1.
Excellent. A git bisect is probably the best way forward on this bug. If you find the particular commit that causes the issues, please let us know.
Matthew Garrett pointed me at a patch for a regression in ACPI yesterday. I've started a scratch build with this patch applied. Could those with an impacted machine please try this kernel when it finishes building and let us know the results?
My apologies, I pasted the wrong link to the scratch build. This is the one that should be tested:
I installed the kernel-PAE, kernel-headers en perf from this koji build.
This kernel does not boot but hangs with the follwing messages:
swapper: segfault at 9ba00 ip (null) sp f4499ca0 error 4
BUG: unable to handle kernel no vm86_info: BAD
note: swapper exited with preempt_count 1
Hmmm, and onother try gives complete different error.
With kernel parameter maxcpus=1 everything boots fine and there is no problem at all with this kernel.
Matthew pointed me to:
That seems like a rather plausible fix for this.
If somebody's able to apply the patch in an new kernel' I'd be happy to test this. I still am keeping a separate older P4 machine alive for this purpose.
This scratch build should have the patch mentioned above:
Testing when it completes would be much appreciated.
(In reply to comment #41)
> Matthew pointed me to:
> That seems like a rather plausible fix for this.
I can confirm, this patch fixed the problem for me.
kernel-3.2.8-3.fc16 has been submitted as an update for Fedora 16.
I confirm too, kernel-PAE-220.127.116.11-1.1.fc15.i686.rpm installed and working good with both cpu's enabled.
* should fix your issue,
* was pushed to the Fedora 16 testing repository,
* should be available at your local mirror within two days.
Update it with:
# su -c 'yum update --enablerepo=updates-testing kernel-3.2.8-3.fc16'
as soon as you are able to.
Please go to the following url:
then log in and leave karma (feedback).
kernel-3.2.9-1.fc16 has been submitted as an update for Fedora 16.
(In reply to comment #43)
> This scratch build should have the patch mentioned above:
> Testing when it completes would be much appreciated.
It is running fine here now, thank you.
kernel-3.2.9-1.fc16 has been pushed to the Fedora 16 stable repository. If problems still persist, please make note of it in this bug report.