Re: [patch 7/8] fdmap v2 - implement sys_socket2

[Davide Libenzi]

On Sat, 9 Jun 2007, Davide Libenzi wrote:

> On Sat, 9 Jun 2007, Al Viro wrote:
> 
> > That's simply not true.  On the current kernel nothing stops you from e.g.
> > picking a random number and using F_DUPFD.  Voila - there's your randomized
> > descriptor.  Portable to earlier kernels.
> > 
> > Moreover, nonsense^H^H^Hq_fd() can be implemented in userland just fine
> > if we allow F_DUPFD to arbitrary number - just pass it a random one *or*
> > base chosen like davedel is doing (constant + 20bit random chosen at start
> > time).
> 
> That does not work. Or better, it works but it forces *huge* fdtables to 
> be created. To do that, you need "start" to be over NR_FILE, and this 
> creates an fdtable bigger than NR_FILE ((pointer-size + a-few-bits) * 
> NR_FILE) propagated down to every app you fork.

Keep in mind also, that quite a few places in the code, walk through the 
whole fdtable memory. So it is not only a problem of wasted RAM.


- Davide


-
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/