Messages in this thread | | | Date | Sat, 10 Apr 1999 09:45:48 -0400 (EDT) | From | Alexander Viro <> | Subject | Re: [CFT] rename patch + FAT stuff |
| |
OK, here is a new version of FAT patch. Changes: fixed a couple of races in fs/fat/*.c fixed a bug in fs/buffer.c that gave massive fs corruption in case of umount immediately followed by mount for all FAT-derived filesystems.
Now it passes most of Gordon's tests for VFAT. The rest is some problem with internationalization (didn't look into it yet) and yet another case of old problem with handling aliases. Old code passed the test, but gave dcache corruption if LongDirName was a directory, you had longdi~1/foo opened and moved LongDirName. The whole aliases thing requires fixing and it's the next one on my hitlist. So there... VFAT mostly works. I still consider it as alpha-quality code. Don't try it on your living filesystem. This version still doesn't work with UMSDOS (second item on the hitlist). It is not suitable for NFS exporting, especially with unfsd (knfsd will give quite a few stale fhandles in low-memory situation, but other than that it may work; said that, combined suckitude of NFS and FAT makes exporting FAT over NFS rather strange idea, unless you are *deep* into sadomasochism). Handling of fat_get_entry() is still not optimal (ahem), but at least it works. To be cleaned up after the aliases stuff and initial UMSDOS integration.
Patch lives on
ftp.math.psu.edu/pub/viro/fat-patch-8.gz (incremental over rename-11) ftp.math.psu.edu/pub/viro/rename+fat-patch-8.gz (combined). No changes in rename patch. Cheers, Al
- 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/
| |