[IPV4] Always set hoplimit

[IPV4] Always set hoplimit

[Herbert Xu]

[-- Attachment #1: Type: text/plain, Size: 455 bytes --]

Hi Dave:

Is there any reason why local/broadcast packets should have a ttl of zero?
This is the result of the HOPLIMIT change as those packets doesn't get
their HOPLIMIT set.  This patch sets it for them.

Am I missing something?
-- 
Debian GNU/Linux 3.0 is out! ( http://www.debian.org/ )
Email:  Herbert Xu ~{PmV>HI~} <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt

[-- Attachment #2: p --]
[-- Type: text/plain, Size: 1120 bytes --]

Index: kernel-source-2.4/net/ipv4/route.c
===================================================================
RCS file: /home/gondolin/herbert/src/CVS/debian/kernel-source-2.4/net/ipv4/route.c,v
retrieving revision 1.6
diff -u -r1.6 route.c
--- kernel-source-2.4/net/ipv4/route.c	3 Sep 2003 10:27:21 -0000	1.6
+++ kernel-source-2.4/net/ipv4/route.c	13 Nov 2003 10:58:04 -0000
@@ -1275,9 +1275,6 @@
 			rt->rt_gateway = FIB_RES_GW(*res);
 		memcpy(rt->u.dst.metrics, fi->fib_metrics,
 		       sizeof(rt->u.dst.metrics));
-		if (rt->u.dst.metrics[RTAX_HOPLIMIT-1] == 0)
-			rt->u.dst.metrics[RTAX_HOPLIMIT-1] =
-				sysctl_ip_default_ttl;
 		if (fi->fib_mtu == 0) {
 			rt->u.dst.metrics[RTAX_MTU-1] = rt->u.dst.dev->mtu;
 			if (rt->u.dst.metrics[RTAX_LOCK-1] & (1 << RTAX_MTU) &&
@@ -1291,6 +1288,8 @@
 	} else
 		rt->u.dst.metrics[RTAX_MTU-1]= rt->u.dst.dev->mtu;
 
+	if (rt->u.dst.metrics[RTAX_HOPLIMIT-1] == 0)
+		rt->u.dst.metrics[RTAX_HOPLIMIT-1] = sysctl_ip_default_ttl;
 	if (rt->u.dst.metrics[RTAX_MTU-1] > IP_MAX_MTU)
 		rt->u.dst.metrics[RTAX_MTU-1] = IP_MAX_MTU;
 	if (rt->u.dst.metrics[RTAX_ADVMSS-1] == 0)

Re: [IPV4] Always set hoplimit

[David S. Miller]

On Thu, 13 Nov 2003 22:04:38 +1100
Herbert Xu <herbert@gondor.apana.org.au> wrote:

> Is there any reason why local/broadcast packets should have a ttl of zero?
> This is the result of the HOPLIMIT change as those packets doesn't get
> their HOPLIMIT set.  This patch sets it for them.
> 
> Am I missing something?

I don't believe I changed this behavior intentionally.  Let me
look into the revision history to verify that.  I'm likely going
to apply your fix tonight, thanks.

Although it is an interesting question because both broadcasts and
locally destined packets have no reason to make it past the local
subnet and thus have their TTL/HOPLIMIT decremented or tested.  I
therefore cannot imagine a case where the current behavior would cause
a problem.

Did you run into an actual problem that led you to this issue?  If
so, what was it?  Or was this just determined by pure code review?

Re: [IPV4] Always set hoplimit

[Herbert Xu]

On Tue, Nov 18, 2003 at 04:59:09PM -0800, David S. Miller wrote:
>
> Did you run into an actual problem that led you to this issue?  If
> so, what was it?  Or was this just determined by pure code review?

Someone reported this problem to me after a picky firewall complained
about getting broadcast packets with ttl == 0.

Cheers,
-- 
Debian GNU/Linux 3.0 is out! ( http://www.debian.org/ )
Email:  Herbert Xu ~{PmV>HI~} <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt

Re: [IPV4] Always set hoplimit

[David S. Miller]

On Wed, 19 Nov 2003 12:08:27 +1100
Herbert Xu <herbert@gondor.apana.org.au> wrote:

> On Tue, Nov 18, 2003 at 04:59:09PM -0800, David S. Miller wrote:
> >
> > Did you run into an actual problem that led you to this issue?  If
> > so, what was it?  Or was this just determined by pure code review?
> 
> Someone reported this problem to me after a picky firewall complained
> about getting broadcast packets with ttl == 0.

Ok, that doubly-justifies the fix.

Thanks again Herbert.