Messages in this thread | | | Date | Mon, 20 Oct 2008 17:23:22 -0400 | From | Shaya Potter <> | Subject | Re: [TOMOYO #11 (linux-next) 01/11] Introduce new LSM hooks where vfsmount is available. |
| |
crispin@crispincowan.com wrote: > Quoting Shaya Potter <spotter@cs.columbia.edu>: >> I know I'm late to the game in this, but as I recently asked about this >> and didn't get an answer, I'll re-ask my approach. >> >> Why can't you do this >> >> in lookup() >> >> - resolve rules (not for single process, but for all processes) for >> said path and tag dentry (seem to already have a hook) >> >> in permission() > > Because it doesn't work > http://kerneltrap.org/mailarchive/linux-fsdevel/2007/6/8/319446 > > Quick summary: The difference between the pathname model and the label > model is dynamism. The accessi really is determined by the pathname to > the file that you used to access the file. If you try to compute access > based on attributes tagged onto the file, then you have to re-compute > those attributes every time someone renames a file.
ok. simple question then so why not just recompute them every every rename? are rename's that common relative to all other file operations where we care?
invalidate the dentry on rename, it will force it to go back through lookup() instead of being found in the cache and shouldn't it implicitly recalculate it?
I don't see why one can't maintain the dynamism with a focus just on lookup and permission (minus the multiple names to the same object issues which is a security problem waiting to happen anyways)
I could very well be missing something, I'm fully ready to eat crow. I just felt it should be asked.
| |