Re: [PATCH] add lock hook to prevent race

Previous message: Crispin Cowan: "Re: [PATCH] permission hook in filemap_nopage"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

chrisat_private

* Chris Wright (chrisat_private) wrote:
> * Antony Edwards (aedwardat_private) wrote:
> > 
> > I think changing the kernel interface is a much cleaner
> > solution (better to prevent the race than detect it),
> > and would be happy to provide a patch to do it. I just
> > didn't want to change any base kernel interfaces in
> > my original patch for acceptance.
> 
> I've sent the patch below to Matthew Wilcox, the maintainer for this kernel
> code.  He had no problem with the changes, and it seems likely it's
> unintrusive enough to make it into 2.4 as well.
> 
> I still like the idea of funneling both locking paths through the same
> LSM hook.  The 2.5 code will change such that the filesystem's ->lock
> method will be called on all file locking code paths.

attached are the patches that implement fcntl style locking mediation.
these patches (relative to the patch in the last email from this thread)
are not much different from antony's.  the first patch changes the core
kernel functionality (where i preferred delaying the check until after
the DAC checking).  the second patch simply updates the modules.

thanks,
-chris