On Mon, Mar 29, 2010 at 08:57:48PM -0400, J. Bruce Fields wrote:
> Hm, OK, so it does look like tcp_close() can sleep, so we are wrong to
> be calling svc_xprt_put() while holding sv_lock.
>
> The commit ab1b18f "sunrpc: remove unnecessary svc_xprt_put" gets rid of
> one svc_xprt_put(), and the remaining final svc_xprt_put() could easily
> be delayed till after we drop the lock.
So, perhaps we want the following.
--b.
commit 788e69e548cc8d127b90f0de1f7b7e983d1d587a
Author: J. Bruce Fields <bfields@xxxxxxxxxxxxxx>
Date: Mon Mar 29 21:02:31 2010 -0400
svcrpc: don't hold sv_lock over svc_xprt_put()
svc_xprt_put() can call tcp_close(), which can sleep, so we shouldn't be
holding this lock.
In fact, only the xpt_list removal and the sv_tmpcnt decrement should
need the sv_lock here.
Reported-by: Mi Jinlong <mijinlong@xxxxxxxxxxxxxx>
Signed-off-by: J. Bruce Fields <bfields@xxxxxxxxxxxxxx>
diff --git a/net/sunrpc/svc_xprt.c b/net/sunrpc/svc_xprt.c
index 8f0f1fb..c334f54 100644
--- a/net/sunrpc/svc_xprt.c
+++ b/net/sunrpc/svc_xprt.c
@@ -892,12 +892,12 @@ void svc_delete_xprt(struct svc_xprt *xprt)
*/
if (test_bit(XPT_TEMP, &xprt->xpt_flags))
serv->sv_tmpcnt--;
+ spin_unlock_bh(&serv->sv_lock);
while ((dr = svc_deferred_dequeue(xprt)) != NULL)
kfree(dr);
svc_xprt_put(xprt);
- spin_unlock_bh(&serv->sv_lock);
}
void svc_close_xprt(struct svc_xprt *xprt)
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
