On Wed, Nov 04, 2009 at 06:47:14AM +0000, Edward Tomasz Napierala wrote: > Author: trasz > Date: Wed Nov 4 06:47:14 2009 > New Revision: 198873 > URL: http://svn.freebsd.org/changeset/base/198873 > > Log: > While VAPPEND without VWRITE makes sense for VOP_ACCESSX(9) (e.g. to check > for the permission to create subdirectory (ACE4_ADD_SUBDIRECTORY)), it > doesn't > really make sense for VOP_ACCESS(9). Also, many VOP_ACCESS(9) > implementations > don't expect that. Make sure we don't confuse them. > > Modified: > head/sys/kern/vfs_default.c > > Modified: head/sys/kern/vfs_default.c > ============================================================================== > --- head/sys/kern/vfs_default.c Wed Nov 4 04:41:03 2009 > (r198872) > +++ head/sys/kern/vfs_default.c Wed Nov 4 06:47:14 2009 > (r198873) > @@ -353,6 +353,14 @@ vop_stdaccessx(struct vop_accessx_args * > if (accmode == 0) > return (0); > > + /* > + * Many VOP_APPEND implementations don't expect VAPPEND without VWRITE > + * being set, e.g. they check whether the filesystem is read-only only > + * when VWRITE is set. Make sure we don't confuse them.
But O_APPEND alone is legal and doesn't provide write access. This is fine for file systems to only check for read-only in VWRITE case. If file system you're talking about is ZFS it was simply a bug in ZFS. > + */ > + if (accmode & VAPPEND) > + accmode |= VWRITE; > return (VOP_ACCESS(ap->a_vp, accmode, ap->a_cred, ap->a_td)); -- Pawel Jakub Dawidek http://www.wheel.pl p...@freebsd.org http://www.FreeBSD.org FreeBSD committer Am I Evil? Yes, I Am!
pgpj5wPjogBvO.pgp
Description: PGP signature
