nash receiving SIGSEGV on kernel update

[draeath]

Nash is blowing up whenever I try to install kernel-2.6.26.5-45.fc9.x86_64. I'm new to Fedora (and RPM in general), this came up while updating after initial install. This text comes up if I try from the console, and it will hang until I interrupt it. I've left it running for 20 minutes, and all the spawned processes were sleeping the whole while.

I really have no idea what to do with this. I've managed to figure out how to reinstall a package (which was needlessly complicated IMO) as such:

sudo rpm -e --nodeps kernel-2.6.26.5.x86_64
sudo yum update kernel

When I try that, this is what I end up getting. It hangs until CTRL-C is hit at the end, in which case yum/rpm SAY that it was installed, but there is no initrd in /boot (and who knows what else is missing/wrong)

[draeath@epicenter ~]$ sudo rpm -e --nodeps kernel-2.6.26.5.x86_64
[draeath@epicenter ~]$ sudo yum update kernel
Loaded plugins: refresh-packagekit
Setting up Update Process
Resolving Dependencies
--> Running transaction check
---> Package kernel.x86_64 0:2.6.26.5-45.fc9 set to be installed
--> Finished Dependency Resolution

Dependencies Resolved

================================================================================
 Package       Arch          Version                Repository            Size 
================================================================================
Installing:
 kernel        x86_64        2.6.26.5-45.fc9        updates-newkey         19 M

Transaction Summary
================================================================================
Install      1 Package(s)         
Update       0 Package(s)         
Remove       0 Package(s)         

Total download size: 19 M
Is this ok [y/N]: y
Downloading Packages:
kernel-2.6.26.5-45.fc9.x86_64.rpm   |  19 MB     00:31     
Running rpm_check_debug
Running Transaction Test
Finished Transaction Test
Transaction Test Succeeded
Running Transaction
  Installing     : kernel                                            [1/1] 
nash received SIGSEGV!  Backtrace (21):
/sbin/nash[0x40d093]
/lib64/libc.so.6[0x7f53c98be2a0]
/lib64/libc.so.6(realloc+0xe2)[0x7f53c9909352]
/usr/lib64/libnash.so.6.0.52[0x621eee]
/usr/lib64/libnash.so.6.0.52[0x621f95]
/usr/lib64/libnash.so.6.0.52(nash_list_add+0x2f)[0x6227fe]
/usr/lib64/libnash.so.6.0.52[0x628873]
/usr/lib64/libnash.so.6.0.52(nash_dev_tree_process_bdev+0x11c)[0x628acb]
/usr/lib64/libnash.so.6.0.52[0x628cfb]
/usr/lib64/libnash.so.6.0.52[0x628e0d]
/usr/lib64/libnash.so.6.0.52(nash_dev_tree_add_sysfs_dir+0x3e)[0x628c76]
/usr/lib64/libnash.so.6.0.52(nash_dev_tree_populate_from_sysfs+0x1c)[0x628ea5]
/usr/lib64/libnash.so.6.0.52(nashBdevIterNewPoll+0x7d)[0x62462a]
/usr/lib64/libnash.so.6.0.52[0x624989]
/usr/lib64/libnash.so.6.0.52(nashFindFsByUUID+0x24)[0x624b24]
/usr/lib64/libnash.so.6.0.52(nashAGetPathBySpec+0x73)[0x624c60]
/sbin/nash[0x4088a1]
/sbin/nash[0x40cf49]
/sbin/nash[0x40d576]
/lib64/libc.so.6(__libc_start_main+0xfa)[0x7f53c98aa32a]
/sbin/nash[0x404509]
^Cerror: %post(kernel-2.6.26.5-45.fc9.x86_64) scriptlet failed, signal 2

Installed: kernel.x86_64 0:2.6.26.5-45.fc9
Complete!
[draeath@epicenter ~]$ ls /boot
config-2.6.25-14.fc9.x86_64    efi   initrd-2.6.25-14.fc9.x86_64.img  System.map-2.6.25-14.fc9.x86_64    vmlinuz-2.6.25-14.fc9.x86_64
config-2.6.26.5-45.fc9.x86_64  grub  lost+found                       System.map-2.6.26.5-45.fc9.x86_64  vmlinuz-2.6.26.5-45.fc9.x86_64

Note the 2.6.26 initrd- is not present.

Help?

[Nokia]

And who told you rpm -e --nodeps is the proper way to uninstall the kernel ? Try

yum erase

instead.

[JEO]

I recommend that you should report the error by filing a bug report at Fedora bugzilla against the "nash" package.

[axet]

that is buffer overflow bug, do you have usb connected devices?

[draeath]

And who told you rpm -e --nodeps is the proper way to uninstall the kernel ? Try

yum erase

instead.

Google, because I didn't find that anywhere when I looked, and it didn't show in a search.

that is buffer overflow bug, do you have usb connected devices?

Yes, a whole bunch, only two (mouse/keyboard) would be required during boot...

I'll try removing the others. If it still fails then, I'll report it as a bug.

[axet]

try detach usb flash drivers / firewire harddrives...

[draeath]

That did it!

Found (and added my resolution details to) bug 443332

[atrofast]

Greetings,

I'm having the exact same problem on Fedora 8 updating to kernel 2.6.26.6-49 but I can unplug down to just my keyboard (a Dell branded one) and it still segfaults. Are there any other fixes to this problem?
Thanks in advance

Alex

[axet]

apply those patch on mkinitrd source:

diff -ru mkinitrd-6.0.52/bdevid/scsi.c mkinitrd-6.0.52.new/bdevid/scsi.c
--- mkinitrd-6.0.52/bdevid/scsi.c 2008-04-25 19:24:00.000000000 +0400
+++ mkinitrd-6.0.52.new/bdevid/scsi.c 2008-10-05 02:48:35.000000000 +0400
@@ -379,7 +379,7 @@
if (priv->unique_id_83)
return 0;

- if (!(priv->unique_id_83 = calloc(page_83[3] + 1, sizeof (char))))
+ if (!(priv->unique_id_83 = calloc(2 * (page_83[3] + 1), sizeof (char))))
return 1;

#if 0



https://bugzilla.redhat.com/show_bug.cgi?id=440661

[atrofast]

apply those patch on mkinitrd source:

diff -ru mkinitrd-6.0.52/bdevid/scsi.c mkinitrd-6.0.52.new/bdevid/scsi.c
--- mkinitrd-6.0.52/bdevid/scsi.c 2008-04-25 19:24:00.000000000 +0400
+++ mkinitrd-6.0.52.new/bdevid/scsi.c 2008-10-05 02:48:35.000000000 +0400
@@ -379,7 +379,7 @@
if (priv->unique_id_83)
return 0;

- if (!(priv->unique_id_83 = calloc(page_83[3] + 1, sizeof (char))))
+ if (!(priv->unique_id_83 = calloc(2 * (page_83[3] + 1), sizeof (char))))
return 1;

#if 0



https://bugzilla.redhat.com/show_bug.cgi?id=440661

Thank you axet, that did the trick.

Alex

[Thetargos]

I'm getting the same errors, only in my case detaching USB devices does not do the trick (maybe because the onboard sound device is USB based?), I'll try disabling it from the BIOS and try again (though somehow I doubt it will do any good)

[Aaron222]

axet's patch works in my case. Thank you man.

Linux Archive