Herbert Xu | 1 Jul 2004 01:01
Picon
Picon

Re: Check connect address in NETLINK

On Wed, Jun 30, 2004 at 03:36:06PM -0700, David S. Miller wrote:
> 
> Why don't you combine the two "ERR_PTR(-ECONNREFUSED)" tests
> into one test like:
> 
> 	if ((nlk->pid == 0 && !nlk->data_ready) ||
> 	    (sock->sk_state == NELTINK_CONNECTED &&
> 	     nlk->dst_pid != nlk_sk(ssk)->pid)) {
> 		sock_put(sock);
> 		return ERR_PTR(-ECONNREFUSED);
> 	}
> 
> so we don't have two copies of the "sock_put(); return ERR_PTR()"
> thing emitted by the compiler?

Well at least under i386, gcc (3.3.4) is smart enough to merge these
common exit paths.

But yes we could merge them.  What about the following incremental
patch?
--

-- 
Visit Openswan at http://www.openswan.org/
Email: Herbert Xu ~{PmV>HI~} <herbert <at> gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
--- work-2.6/net/netlink/af_netlink.c.orig	2004-07-01 08:56:55.000000000 +1000
+++ work-2.6/net/netlink/af_netlink.c	2004-07-01 08:59:14.000000000 +1000
 <at>  <at>  -425,21 +425,23  <at>  <at> 
(Continue reading)

Rusty Russell | 1 Jul 2004 03:12
Picon
Gravatar

Re: A problem with a large static arp table

On Thu, 2004-07-01 at 01:14, Roman Zagustin wrote:
> Hello!

Hi,

	Not really my area; pinging netdev to see if they can help.

> I have a router with >800 static ip/mac records.
> It does not work :(
> With small table it is fine.
> The kernel is 2.6 (tried with different patches and versions
>  2.4 - the same)
> 
> Just for example(real):
>  
> root <at> stargate:/proc/sys/net/ipv4/neigh/default# arp -n|grep 172.16.3.230
> 172.16.3.230             ether   00:50:04:6C:50:9E   CM                    eth3
> root <at> stargate:/proc/sys/net/ipv4/neigh/default#
> root <at> stargate:/proc/sys/net/ipv4/neigh/default# tcpdump -i eth3 host 172.16.3.230  -e
> tcpdump: listening on eth3
> 15:38:10.660144 0:6:29:ee:c3:d8 0:2:b3:b4:34:e2 ip 448: 172.16.3.230.1060 >
medium.cs.microlink.lt.http: P 1268207798:1268208192(394) ack 3537529644 win 64240 (DF)
> 15:38:10.674698 0:6:29:ee:c3:d8 0:2:b3:b4:34:e2 ip 453: 172.16.3.230.1077 >
medium.cs.microlink.lt.http: P 1279539300:1279539699(399) ack 2683450794 win 63095 (DF)
> 
> There are no errors/warnings in the log files about neighbour table.
> 
> >From the server I can't ping the 172.16.3.230, but it can ping the
> server. In general, server do not check the ip/mac table at all
> when forwarding received packets.
(Continue reading)

Jeff Garzik | 1 Jul 2004 05:19
Picon
Favicon

Re: [PATCH 1/6 Linux-2.6.7-bk13] prism54 cleanup functions

applied patches 1-6

Jeff Garzik | 1 Jul 2004 05:21
Picon
Favicon

Re: [PATCH 5 2.6.7-bk6] pcnet32: Add HomePNA parameter for 79C978.

applied patches #4-7

Jeff Garzik | 1 Jul 2004 05:23
Picon
Favicon

Re: [PATCH] natsemi 1: switch to netdev_priv()

patches 1 and 2 look OK, but don't apply cleanly against mainline.  I'll 
try against -netdev in a few days

Jeff Garzik | 1 Jul 2004 05:24
Picon
Favicon

Re: [PATCH] (1/3) skfp - cleanup is_XXX functions

applied

Jeff Garzik | 1 Jul 2004 05:24
Picon
Favicon

Re: [PATCH] (2/3) skfp -- sparse __user annotation

applied

Jeff Garzik | 1 Jul 2004 05:26
Picon
Favicon

Re: [PATCH] [sparse] dgrs -- cleanup warnings

bounced:  OK but patch(1) rejected.  please rediff and resubmit.

Jeff Garzik | 1 Jul 2004 05:27
Picon
Favicon

Re: [PATCH] [sparse] get rid of warnings about #if DEBUG

applied

Jeff Garzik | 1 Jul 2004 05:27
Picon
Favicon

Re: [PATCH] get rid of __OPTIMIZE__ requirement in net drivers

applied


Gmane