netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: pravin b shelar <pravins@calsoftinc.com>
To: Herbert Xu <herbert@gondor.apana.org.au>
Cc: christoph@lameter.com, netdev@oss.sgi.com, davem@davemloft.net
Subject: Re: Fwd: atomic_dec_and_test for child dst needed in dst_destroy?
Date: Thu, 07 Apr 2005 18:00:16 +0530	[thread overview]
Message-ID: <425527D8.4030408@calsoftinc.com> (raw)
In-Reply-To: <E1DJV6r-0002Nu-00@gondolin.me.apana.org.au>

[-- Attachment #1: Type: text/plain, Size: 637 bytes --]

Herbert Xu wrote:

>pravin b shelar <pravins@calsoftinc.com> wrote:
>  
>
>>This patch  will remove the need for atomic_dec_and_test for this 
>>particular case.
>>Now we can break down the atomic_dec_and_test to atomic_dec & atomic_read.
>>    
>>
>
>Sorry I overlooked your patch.  Otherwise we could have solved this
>one day earlier :)
>  
>
ah ... I was just waiting :)

>  
>
>>Please comment.
>>    
>>
>
>There is just one problem.  We need an extra barrier before or after
>the goto since we no longer have the implicit barrier given by
>atomic_dec_and_test.
>  
>
I am attaching updated patch with barrier.

Regards,
Pravin.



[-- Attachment #2: dst-atomic_dec_and_test-removed.patch --]
[-- Type: text/x-patch, Size: 1176 bytes --]

This patch removes atomic_dec_and_test from dst_destroy, used to test child dst __refcnt.

Signed-off by: Pravin B. Shelar <pravins@calsoftinc.com>

Index: linux-2.6.11-dsta/net/core/dst.c
===================================================================
--- linux-2.6.11-dsta.orig/net/core/dst.c	2005-04-06 00:18:32.000000000 -0700
+++ linux-2.6.11-dsta/net/core/dst.c	2005-04-07 04:53:28.000000000 -0700
@@ -198,17 +198,21 @@
 
 	dst = child;
 	if (dst) {
-		if (atomic_dec_and_test(&dst->__refcnt)) {
-			/* We were real parent of this dst, so kill child. */
-			if (dst->flags&DST_NOHASH)
-				goto again;
-		} else {
-			/* Child is still referenced, return it for freeing. */
-			if (dst->flags&DST_NOHASH)
-				return dst;
-			/* Child is still in his hash table */
+		if (dst->flags&DST_NOHASH) {
+			atomic_dec(&dst->__refcnt);
+			smp_mb__after_atomic_dec();
+
+			if (atomic_read(&dst->__refcnt))
+				/*-- Child is still referenced, return it for freeing. */
+				return dst;	
+
+			/*We were real parent of this dst, so kill child. */
+			goto again;
 		}
+		else /* Child is still in his hash table */
+			atomic_dec(&dst->__refcnt);
 	}
+
 	return NULL;
 }
 

      reply	other threads:[~2005-04-07 12:30 UTC|newest]

Thread overview: 27+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2005-04-05 18:55 atomic_dec_and_test for child dst needed in dst_destroy? Christoph Lameter
2005-04-05 19:34 ` David S. Miller
2005-04-05 19:47   ` Christoph Lameter
2005-04-05 19:50     ` David S. Miller
2005-04-05 19:58       ` Christoph Lameter
2005-04-05 20:12         ` David S. Miller
2005-04-05 21:45 ` Herbert Xu
2005-04-05 21:48   ` David S. Miller
2005-04-05 22:14   ` Christoph Lameter
2005-04-06  2:19     ` Herbert Xu
2005-04-06  3:19       ` Christoph Lameter
2005-04-06  8:32         ` Herbert Xu
2005-04-06 18:17           ` David S. Miller
2005-04-06 18:48             ` Christoph Lameter
2005-04-07 11:07             ` Herbert Xu
2005-04-07 16:00               ` Christoph Lameter
2005-04-07 21:25                 ` Herbert Xu
2005-04-07 22:30                   ` Christoph Lameter
2005-04-07 23:07                     ` Herbert Xu
2005-04-08  5:45                   ` Christoph Lameter
2005-04-08  5:48                     ` Herbert Xu
2005-04-08 15:05                       ` Christoph Lameter
2005-04-08 21:45                         ` Herbert Xu
2005-04-09 15:28                           ` Christoph Lameter
     [not found]   ` <b82a8917050406002339f732ca@mail.gmail.com>
2005-04-06  8:53     ` Fwd: " pravin b shelar
2005-04-07 11:23       ` Herbert Xu
2005-04-07 12:30         ` pravin b shelar [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=425527D8.4030408@calsoftinc.com \
    --to=pravins@calsoftinc.com \
    --cc=christoph@lameter.com \
    --cc=davem@davemloft.net \
    --cc=herbert@gondor.apana.org.au \
    --cc=netdev@oss.sgi.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).