Re: [PATCH] sunrpc - fixup userspace buffer possible overrun v2

[Cyrill Gorcunov <gorcunov@gmail.com>]

[Ingo Oeser - Sun, Aug 31, 2008 at 04:09:10PM +0200]
| Hi Cyrill,
| 
| On Sunday 31 August 2008, Cyrill Gorcunov wrote:
| > The conclusion is that proc_do_xprt doesn't check for userside buffer
| > size indeed so fix. Also set lenp to number of bytes were really written.
| 
| Why not use  simple_read_from_buffer() for the read case and
| keep the -EINVAL for the write case.

Ah, thanks Ingo - good idea. Btw does libfs.c depends
on anything?

| 
| > Reported-by: Vegard Nossum <vegard.nossum@gmail.com>
| > Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
| > CC: David Wagner <daw@cs.berkeley.edu>
| > ---
| > 
| > Please review.
| > 
| > Index: linux-2.6.git/net/sunrpc/sysctl.c
| > ===================================================================
| > --- linux-2.6.git.orig/net/sunrpc/sysctl.c	2008-08-31 13:43:46.000000000 +0400
| > +++ linux-2.6.git/net/sunrpc/sysctl.c	2008-08-31 13:58:14.000000000 +0400
| > @@ -60,23 +60,26 @@ static int proc_do_xprt(ctl_table *table
| >  			void __user *buffer, size_t *lenp, loff_t *ppos)
| >  {
| >  	char tmpbuf[256];
| > -	int len;
| > +	size_t len;
| > +
| + ssize_t ret;
| >  	if ((*ppos && !write) || !*lenp) {
| >  		*lenp = 0;
| >  		return 0;
| >  	}
| > +
| >  	if (write)
| >  		return -EINVAL;
| 
|         len = svc_print_xprts(tmpbuf, sizeof(tmpbuf));
| 	ret = simple_read_from_buffer(buffer, ppos, tmpbuf, len);
| 	if (ret >= 0) {
| 		*lenp = ret;
| 		ret = 0;
| 	}
| 
| 	return ret;
|    }
| 
| 
| Best Regards
| 
| Ingo Oeser
| 
		- Cyrill -