Description of problem: First of all, I'm not using RHEL5 but CentOS5. People under the IRC #centos channel suggested me to file a bug directly to RedHat due the severity of the problem. So that's why I'm here. The problem was encountered by a mistake trying to link a loop device to a file using the mount command instead losetup, and then after the expected error message from the mount command, I changed mount by losetup and just after press Enter the system freezes. NOTE: maybe the mount and losetup steps need to be completed two times to generate the kernel panic. I've reproduced the problem under different machines using the same up to date CentOS 5 system with the same results. Version-Release number of selected component (if applicable): mount --version says: mount (util-linux 2.13-pre7) How reproducible: Always, maybe need two passes instead only one in some machines. Steps to Reproduce: The following are the steps to reproduce the problem: (other steps may reproduce the problem too, but to make sure use theses ones) 1) Download the floppy boot image from here: http://archive.download.redhat.com/pub/redhat/linux/7.3/en/os/i386/images/boot.img 2) Convert it to a cdrom boot image using the following command: mkisofs -R -b boot.img boot.img > boot-cdrom.iso 3) mount /dev/loop0 boot-cdrom.iso (it will display a normal error message here) 4) losetup /dev/loop0 boot-cdrom.iso It must panic here, otherwise do a "losetup -d /dev/loop0" and go to step 3. Actual results: Kernel panic - not syncing: Fatal exception Expected results: The command should work finely. Additional info:
I could reproduce the crash with 2.6.18-8.1.8.el5, but not with a current development build. The issue looks similar to bug 236880. Possibly fixed already. Jordi, can you test a recent kernel development build from http://people.redhat.com/dzickus/el5/ ?
I have just upgraded my CentOS 5 to the very last kernel 2.6.18-8.1.10.el5 and the kernel panic still persist. Remember that it may need two passes before see the panic (steps 3 and 4)
2.6.18-8.1.10.el5 does not have the fix for bug 236880. Can you please test a newer kernel from http://people.redhat.com/dzickus/el5/ ? E.g. version 49.el5. Thanks.
Ok, I have tested it with the kernel 2.6.18-49.el5 and it works finely. As you said the problem was probably fixed. I tried a lot of passes even I've executed a shell-script with the steps 3 and 4 inside a while [ 1 ] loop during several minutes, and the panic was not seen. I think you can close this issue. Thanks.
*** This bug has been marked as a duplicate of 236880 ***