Messages in this thread | | | Subject | Re: [patch 01/15] security: pass path to inode_create | From | Tetsuo Handa <> | Date | Sat, 31 May 2008 19:48:22 +0900 |
| |
Christoph Hellwig wrote: > On Thu, May 29, 2008 at 03:49:04PM +0200, Miklos Szeredi wrote: > > From: Miklos Szeredi <mszeredi@suse.cz> > > > > In the inode_create() security operation and related functions pass > > the path (vfsmount + dentry) to the parent directory instead of the > > inode. AppArmor will need this. > > So you're once again switching vfs_ to a pass a vfsmount argument, this > time hidden under struct path. It's really hard to grasp a "no" > sometimes, isn't it? :) > The vfs-cleanups git tree is developed for passing "vfsmount" argument so that r/o bind mounts shall not fail to check "vfsmount" by reconstructing vfs_*() callers and marking vfs_*() as "static".
If this vfs-cleanups git tree has no problem regarding "vfsmount" argument, I think there is no advantage of adding security_path_*() hooks inside the path_*() functions since everybody calls the vfs_*() functions via path_*() functions.
Passing "vfsmount" to vfs_*() functions is better for AppArmor and TOMOYO and auditing purpose (we can obtain the absolute pathname compared to the relative pathname from the mount point) than adding security_path_*(), isn't it?
Are you worrying that there might be a case where the "vfsmount" argument passed to path_*() functions is invalid or NULL?
| |