I've been using XFS filesystem for long time and somewhat track the
developement and try to stay up to date with the changes in 2.4 CVS
I just built a recent kernel from 03-22-2002 cvs on Slackware Linux 8.0
with gcc 2.95.3 that Slackware ships.I took these steps:
update the linux-2.4 tree
generate diff against kernel.org 2.4.18 source (diff -Nur --exclude=CVS
/usr/src/linux-2.4.18 linux > ...)
patch JFS 0.1.16 code first (patch -p1 option)
patch XFS diff from 03-22-2002 (patch -p1 option) all in linux directory
Went to compile kernel as usual and there was no problems, all compiled
fine.XFS is build right into kernel and JFS as module but i dont think
this should matter.
Upon reboot to test new kernel and after my AHA-2940U2W adapter is
scanned i get a message that XFS has bad magic number on /dev/sda2
and it's primary superblock has been corrupted and it just halts
there...i was able to use SYSRQ to reboot the system.
Tried to boot with my previous working kernels (2.4.18 and 2.4.5 both
with XFS) and got the same errors that superblock is corrupted and has
I should note that there was never any oops'es or crashes on my box and
hard disk seems to be fine too.
I was able to boot to my other Linux LFS system that is on another
partition and from there run xfs_repair on the bad partition.All of the
xfs utilities are the latest version as of 03-22-2002.Xfs_repair was
able to recover the primary superblock from it's secondary copy and
repair my partition without any errors.
My question is to what could cause such sudden corrruption of superblock
and if it has to do anything with recent changes that went into CVS?
Is there anything i should do to verify that the filesystem is indeed
I would appreciate any help or suggestions on this matter.Thank you for
such a great filesystem and your hard work to port it to Linux.
( http://www.linuxdiscussions.org )
Description: This is a digitally signed message part