Messages in this thread | | | Subject | Re: silent semantic changes with reiser4 | From | Christophe Saout <> | Date | Wed, 25 Aug 2004 23:00:00 +0200 |
| |
Am Mittwoch, den 25.08.2004, 21:42 +0100 schrieb :
> > > For one thing _I_ didn't decide about xattrs anyway. And I still > > > haven't seen a design from you on -fsdevel how you try to solve the > > > problems with files as directories. > > > > Hey, files-as-directories are one of my pet things, so I have to side with > > Hans on this one. I think it just makes sense. A hell of a lot more sense > > than xattrs, anyway, since it allows scripts etc standard tools to touch > > the attributes. > > > > It's the UNIX way. > > Not if you allow link(2) on them.
That doesn't make sense anyway. (actually, I tried what happens and the result was an Oops ;))
It should be completely forbidden to link into a meta-directory or out of such a directory. You could think of those meta-directory as a sysfs for that inode. Of course it's not an own filesystem and that means that there need to be a lot of security precautions in the VFS layer. Where something like that belongs anyway, if done correctly.
> And not if you design and market your > stuff as a general-purpose backdoor into kernel. Note how *EVERY* *DAMN* > *OPERATION* is made possible to override by "plugins". Which is the reason > for deadlocks in question, BTW.
What do you mean? If you tell that file that you want it to be compressed or encrypted or modify some attributes (like ACLs) this isn't necessarily a backdoor.
> Don't fool yourself - that's what Hans is selling. Target market: ISV. > Marketed product: a set of hooks, the wider the better, no matter how > little sense it makes. The reason for doing that outside of core kernel: > bypassing any review and being able to control the product being sold (see > above).
Yes, I don't think it was a good idea either. Probably someone should remove these features and make it a "normal" filesystem. The people who need it now can turn it on again and a real solution could be worked out in Linux 2.7.
I wouldn't use it on a public server anyway now because I'm not convinced some malicious guy could find a way to exploit that. What if you changed into a meta directory using ftp and some manage to break things? This might be very dangerous.
I personally think that the idea of doing something like this (I'm not speaking of the current implementation which I think is really bad) is the right way to go in the long term.
[unhandled content-type:application/pgp-signature] | |