[PATCH net-next v2 1/1] net: ipv4: Refine the ipv4_default_advmss

[PATCH net-next v2 1/1] net: ipv4: Refine the ipv4_default_advmss

[gfree.wind]

From: Gao Feng <fgao@ikuai8.com>

1. Don't get the metric RTAX_ADVMSS of dst.
There are two reasons.
1) Its caller dst_metric_advmss has already invoke dst_metric_advmss
before invoke default_advmss.
2) The ipv4_default_advmss is used to get the default mss, it should
not try to get the metric like ip6_default_advmss.

2. Use sizeof(tcphdr)+sizeof(iphdr) instead of literal 40.

3. Define one new macro IPV4_MAX_PMTU instead of 65535 according to
RFC 2675, section 5.1.

Signed-off-by: Gao Feng <fgao@ikuai8.com>
---
 v2: Use min instead of unnecessary min_t, per Joe
 v1: initial version

 include/net/ip.h |  2 ++
 net/ipv4/route.c | 11 ++++-------
 2 files changed, 6 insertions(+), 7 deletions(-)

diff --git a/include/net/ip.h b/include/net/ip.h
index bf264a8..821cedc 100644
--- a/include/net/ip.h
+++ b/include/net/ip.h
@@ -33,6 +33,8 @@
 #include <net/flow.h>
 #include <net/flow_dissector.h>
 
+#define IPV4_MAX_PMTU		65535U		/* RFC 2675, Section 5.1 */
+
 struct sock;
 
 struct inet_skb_parm {
diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index 5e1e605..2d414b0 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -1250,14 +1250,11 @@ static void set_class_tag(struct rtable *rt, u32 tag)
 
 static unsigned int ipv4_default_advmss(const struct dst_entry *dst)
 {
-	unsigned int advmss = dst_metric_raw(dst, RTAX_ADVMSS);
+	unsigned int header_size = sizeof(struct tcphdr) + sizeof(struct iphdr);
+	unsigned int advmss = max_t(unsigned int, dst->dev->mtu - header_size,
+				    ip_rt_min_advmss);
 
-	if (advmss == 0) {
-		advmss = max_t(unsigned int, dst->dev->mtu - 40,
-			       ip_rt_min_advmss);
-		if (advmss > 65535 - 40)
-			advmss = 65535 - 40;
-	}
+	advmss = min(advmss, IPV4_MAX_PMTU - header_size);
 	return advmss;
 }
 
-- 
1.9.1

Re: [PATCH net-next v2 1/1] net: ipv4: Refine the ipv4_default_advmss

[Joe Perches]

On Wed, 2017-04-12 at 10:32 +0800, gfree.wind@foxmail.com wrote:
> diff --git a/net/ipv4/route.c b/net/ipv4/route.c

more trivia:

> @@ -1250,14 +1250,11 @@ static void set_class_tag(struct rtable *rt, u32 tag)
>  
>  static unsigned int ipv4_default_advmss(const struct dst_entry *dst)
>  {
> -	unsigned int advmss = dst_metric_raw(dst, RTAX_ADVMSS);
> +	unsigned int header_size = sizeof(struct tcphdr) + sizeof(struct iphdr);
> +	unsigned int advmss = max_t(unsigned int, dst->dev->mtu - header_size,
> +				    ip_rt_min_advmss);
>  
> -	if (advmss == 0) {
> -		advmss = max_t(unsigned int, dst->dev->mtu - 40,
> -			       ip_rt_min_advmss);
> -		if (advmss > 65535 - 40)
> -			advmss = 65535 - 40;
> -	}
> +	advmss = min(advmss, IPV4_MAX_PMTU - header_size);
>  	return advmss;
>  }

This would probably be simpler to read as:

	return min(advmss, IPV4_MAX_PMTU - header_size);

though it's almost certain that the compiler emits identical
object code.

RE: [PATCH net-next v2 1/1] net: ipv4: Refine the ipv4_default_advmss

[Gao Feng]

> -----Original Message-----
> From: netdev-owner@vger.kernel.org [mailto:netdev-owner@vger.kernel.org]
> On Behalf Of Joe Perches
> Sent: Wednesday, April 12, 2017 11:57 AM
> To: gfree.wind@foxmail.com; davem@davemloft.net; kuznet@ms2.inr.ac.ru;
> jmorris@namei.org; netdev@vger.kernel.org
> Cc: Gao Feng <fgao@ikuai8.com>
> Subject: Re: [PATCH net-next v2 1/1] net: ipv4: Refine the
ipv4_default_advmss
> 
> On Wed, 2017-04-12 at 10:32 +0800, gfree.wind@foxmail.com wrote:
> > diff --git a/net/ipv4/route.c b/net/ipv4/route.c
> 
> more trivia:
> 
> > @@ -1250,14 +1250,11 @@ static void set_class_tag(struct rtable *rt,
> > u32 tag)
> >
> >  static unsigned int ipv4_default_advmss(const struct dst_entry *dst)
> > {
> > -	unsigned int advmss = dst_metric_raw(dst, RTAX_ADVMSS);
> > +	unsigned int header_size = sizeof(struct tcphdr) + sizeof(struct
iphdr);
> > +	unsigned int advmss = max_t(unsigned int, dst->dev->mtu -
header_size,
> > +				    ip_rt_min_advmss);
> >
> > -	if (advmss == 0) {
> > -		advmss = max_t(unsigned int, dst->dev->mtu - 40,
> > -			       ip_rt_min_advmss);
> > -		if (advmss > 65535 - 40)
> > -			advmss = 65535 - 40;
> > -	}
> > +	advmss = min(advmss, IPV4_MAX_PMTU - header_size);
> >  	return advmss;
> >  }
> 
> This would probably be simpler to read as:
> 
> 	return min(advmss, IPV4_MAX_PMTU - header_size);
> 
> though it's almost certain that the compiler emits identical object code.

Yes. I thought about writing it with the way you mentioned, but gave up.
But i could follow you this time.

Best Regards
Feng