Created attachment 317542 [details] core dump from gfs2_fsck, gzipped Description of problem: After adding extended attributes to a file system, gfs2_fsck seg faults while checking it. Core was generated by `gfs2_fsck /dev/tankmorph/tankmorph0 -yq'. Program terminated with signal 11, Segmentation fault. #0 0x0804ed86 in pass1c () (gdb) bt #0 0x0804ed86 in pass1c () #1 0x08049cbb in main () #2 0x00578dec in __libc_start_main () from /lib/libc.so.6 #3 0x08048fb1 in _start () Version-Release number of selected component (if applicable): gfs2-utils-0.1.47-1.el5 How reproducible: 100% Steps to Reproduce: 1. add extended attributes to a file system 2. umount 3. fsck Actual results: Expected results: gfs2_fsck should not segfault. Additional info:
Created attachment 317605 [details] Patch to fix the problem This was a simple case of needing to use the "_safe" version of the osi_list_foreach function. That's because there are places where entries are being deleted off the list as I'm running the list. This patch fixes the problem, and I tested it on the file system that had the problem, from system tank-01.
The patch was pushed to the master, STABLE2 and RHEL5 branches of the cluster git tree for inclusion into 5.3. It was tested on the failing file system on tank-01. Changing status to Modified.
Verified with gfs2-utils-0.1.49-1.el5.
An advisory has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on therefore solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHBA-2009-0087.html