Re: still nfs problems [Was: Linux 2.6.37-rc8]

[Trond Myklebust]

On Thu, 2010-12-30 at 22:17 -0500, George Spelvin wrote: 
> > Uncached_readdir is not really a problem. The real problem is
> > filesystems that generate "infinite directories" by producing looping
> > combinations of cookies.
> > 
> > IOW: I've seen servers that generate cookies in a sequence of a form
> > vaguely resembling
> > 
> > 1 2 3 4 5 6 7 8 9 10 11 12 3...
> > 
> > (with possibly a thousand or so entries between the first and second
> > copy of '3')
> > 
> > The kernel won't loop forever with something like that (because
> > eventually filldir() will declare it is out of buffer space), but
> > userland has a halting problem: it needs to detect that every
> > sys_getdents() call it is making is generating another copy of the
> > sequence associated with '4 5 6 7 8 9 10 11 12 3'...
> 
> Huh?  This is not only an easy problem, it's a well-known problem.
> http://en.wikipedia.org/wiki/Cycle_detection
> 
> 	Here's Brent's algorithm:
> 
> 	n = 0;
> 	saved_cookie = <invalid>
> 	For each cookie {
> 		if (n && cookie == saved_cookie)
> 			die("Loop detected!");
> 		if (++n is a power of 2)
> 			saved_cookie = cookie;
> 	}
> 
> You can tweak the performance with other exponentially-growing
> functions, saving k > 1 old cookies for comparison, etc., but the
> above will work very well.


...and your point would be that an exponentially increasing addition to
the existing number of tests is an acceptable tradeoff in a situation
where the >99.999999999999999% case is that of sane servers with no
looping? I don't think so...

Trond

-- 
Trond Myklebust
Linux NFS client maintainer

NetApp
Trond.Myklebust@xxxxxxxxxx
www.netapp.com

--
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/