Messages in this thread | | | Date | Wed, 17 Nov 1999 23:54:39 +0100 (CET) | From | Stelian Pop <> | Subject | Re: `filetype' ext2 feature causes dump to SEGV |
| |
On Wed, 17 Nov 1999, Theodore Y. Ts'o wrote:
> It's actually a bug in dump. It's not checking the superblock > compatibility flags, so it didn't notice that it didn't know how to > handle filesystems with the filetype feature. When it tried, it got > itself into a lot of trouble. :-( > > [...] > > See below for the patch to dump-0.4b9/dump/traverse.c.
This was a quick patch !!!
Of course you are absolutelly right, I will integrate the patch in the upcoming release (0.4b10, planned for this week-end).
> Yeah, there really ought to be a debugging flag which turns off all of > the forks. It was a pain to have to manually comment out the fork in > startnewtape() in order to be able to debug this under gdb. Stelian?
Actually this is a little bit tricky because dump uses fork in order to keep checkpoints on the tape level (by duplicating the context). I could add such a debug mode but it would be impossible to recover on tape errors, and maybe it would be impossible to make multi-tape backups (need to look at the code a bit more). But such a special-debug mode can make sense in order to investigate the filesystem access errors, so will probably add it anyway...
By the way, Ted: I plan to reimplement the ext2fs access code in dump (actually the entire traverse.c code), in order to make dump more filesystem independent (to be able to add in the future some support for other filesystems, such as ext3, BSD, etc). I was wondering why dump uses direct disc access (open and read directly the direct, indirect and doubly indirect blocks), instead of using the associated e2fsprogs library functions. Did Rémy done this because the functions didn't exist at this time or was this for performance reasons ?
Thanks.
Stelian. -- /\ / \ Stelian Pop / DS \ Email: pop@cybercable.fr \____/
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |