Weekly Edition
Return to the Kernel page
| ||||||||||||||||
|
| ||||||||||||||||
New EA/ACL kernel patches available (0.8.56 release candidate 2)
Hello, I am happy to announce that Sun has no issues with the NFS ACL protocol implementation. A letter confirming that is on the way to SuSE, but even before it arrives we are quite safe with a confirmation from Brent Callaghan. Brent, thank you very much for going through all the trouble for us. Concerning the next release of kernel patches, the following problems were found in 0.8.56 rc1. Version 0.8.56 rc2 fixes these issues, and is found at <http://acl.bestbits.at/pre/2.4.20/>. * Stephen C. Tweedie found a locking order problem leading to a deadlock, most easily trigered over NFS. (Reported by Stephen C. Tweedie <sct@redhat.com>). * The fs_struct structure was not exported, so the kernel nfsd could not be built as a module. (Reported by Eric Jarman <ehj38230@cmsu2.cmsu.edu>). * On Alpha, the number of entries in the system call table was not adjusted to account for the xattr system calls. (Reported by Dennis Henriksen <opus@osrl.net>). * Due to a bug in the meta block cache (fs/mbcache.c), extended attribute block sharing was effectively disabled. * Due to an inverted flags test, ACLs could only be set over NFS with the `-o noacl' flag set. (But then, they wouldn't be checked!) * The NFS ACL protocol has been improved; the listxattr and removexattr inode operations that were so far missing have been added. Thanks go all who have helped to improve the patches. Regards, Andreas. ------------------------------------------------------------------ Andreas Gruenbacher SuSE Linux AG mailto:agruen@suse.de Deutschherrnstr. 15-19 http://www.suse.de/ D-90429 Nuernberg, Germany _______________________________________________ acl-devel mailing list acl-devel@bestbits.at http://acl.bestbits.at/mailman/listinfo/acl-devel |
|
|||||||||||||||