Distribution: Fedora Core 3

Hardware Environment: 7200RPM IDE hard disk, i386 architecture

Software Environment: XFS partition is loop mounted (i.e. it's a file stored on
another partition)

Problem Description: I have a corrupted XFS filesystem that gives me the
following error when I try to mount it:

Filesystem "loop0": XFS internal error xlog_valid_rec_header(2) at line 3516 of
file fs/xfs/xfs_log_recover.c.  Caller 0xf9452e98

This is followed by a scary-looking traceback. I assume from the appearance of
the traceback that this is something I should report as a bug. I will attach
output from 2.6.11-rc2-mm1.

Steps to reproduce:
1. Attempt to mount a particular corrupted XFS filesystem. (I would attach it
here or put it on the web or something, but it may contain sensitive information
and it's also 4GB. It gzips down to 1.8GB but that's still large relative to ISP
bandwidth bytecaps and that sort of thing. :(

Is there an XFS equivalent to e2image, or anything of the sort, that I could use
to provide more information?