Re: DoS with POSIX file locks?

[Miklos Szeredi]

> Concrete breakage.  Something like:
> 
> clone(CLONE_FILES)
>   /* in child */
>   lock
>   execve
>   lock
> 
> w/out the kludge[1], the lock fails.  I should have a test program about
> that I wrote to test this, although it was originally triggered via some
> LTP or LSB type of test (don't recall which).
> 
> thanks,
> -chris
> 
> [1] happy to see it go.

We all agree on this then.

I'm just little paranoid about a real-world app (LTP/LSB don't matter)
relying on the current semantics.

But maybe there's no other way to find out, than to remove
steal_locks() and see if anybody complains.

> i concur with Trond, there's no sane way to get rid of it w/out
> formalizing CLONE_FILES and locks on exec

Probably there is.  It would involve allocating a separate
lock-owner-ID stored in files_struct but separate from it.  But it's
more complicated than simply not propagating locks on exec in the
CLONE_FILES case.

Miklos
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/