Messages in this thread | | | Date | Tue, 16 Feb 1999 14:08:41 -0500 (EST) | From | Alexander Viro <> | Subject | [CFT] New version of rename patch |
| |
Hi. New version of rename patch (this time against -pre4) is out - I've posted it to fsdevel and had put it on ftp://math.psu.edu/pub/viro/rename-patch-7 (thing had grown too large for l-k - ~56K, mostly *removing* dead code ;-)
News: * In case we are doing a cross-directory rename and the source is a directory we have to change .. in it. Thus we need write permissions. Check added. Kudos to James Griffiths. * If source and target are links to the same inode no permissions matter - rename() should do nothing and return 0. Fixed. * In ext2, ufs, smbfs, coda and nfs lots of excessive name length checks trimmed. It's enough to do it once in ->lookup() and return -ENAMETOOLONG if name is bad. Then no other method will receive such dentry. * Related to previous: mknod() on ext2 and ufs would return -ENAMETOOLONG in some cases when -EIO should be returned. Fixed as the side effect of trimmage above. * Dead code removed from {ext2,ufs}_link() - check for target being IMMUTABLE/APPEND-ONLY is already done in VFS, so there is no need in duplication. * Bug introduced by msdos rename() cleanup fixed - rename("foo",".foo") might blow up. Badly.
Please, try it out. 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/
| |