linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* dev_get_by_name without dev_put
@ 2001-07-04 13:49 fabrizio.gennari
  2001-07-04 14:06 ` [PATCH] " Arnaldo Carvalho de Melo
  2001-07-10  0:36 ` David S. Miller
  0 siblings, 2 replies; 3+ messages in thread
From: fabrizio.gennari @ 2001-07-04 13:49 UTC (permalink / raw)
  To: linux-kernel

Every time dev_get_by_name is called, and it has returned a valid struct net_device*, dev_put should be called afterwards, because otherwise the machine hangs when the device is unregistered (since dev->refcnt > 1). However, it seems that some drivers do
not call dev_put after dev_get_by_name: for example, drivers/net/pppoe.c at line 573 and net/core/dv.c at line 168. Am I wrong?
---------------------------------------------------------
Fabrizio Gennari          tel. +39 039 203 7816
Philips Research Monza    fax. +39 039 203 7800
via G. Casati 23          fabrizio.gennari@philips.com
20052 Monza (MI) Italy    http://www.research.philips.com


^ permalink raw reply	[flat|nested] 3+ messages in thread

* [PATCH] Re: dev_get_by_name without dev_put
  2001-07-04 13:49 dev_get_by_name without dev_put fabrizio.gennari
@ 2001-07-04 14:06 ` Arnaldo Carvalho de Melo
  2001-07-10  0:36 ` David S. Miller
  1 sibling, 0 replies; 3+ messages in thread
From: Arnaldo Carvalho de Melo @ 2001-07-04 14:06 UTC (permalink / raw)
  To: fabrizio.gennari; +Cc: linux-kernel

Em Wed, Jul 04, 2001 at 03:49:46PM +0200, fabrizio.gennari@philips.com escreveu:
> Every time dev_get_by_name is called, and it has returned a valid struct net_device*, dev_put should be called afterwards, because otherwise the machine hangs when the device is unregistered (since dev->refcnt > 1). However, it seems that some drivers do
> not call dev_put after dev_get_by_name: for example, drivers/net/pppoe.c at line 573 and net/core/dv.c at line 168. Am I wrong?

For pppoe (and one has to look at the set_item and see if delete_item is
needed).

- Arnaldo

--- linux-2.4.5-ac24/drivers/net/pppoe.c	Tue Jul  3 18:45:41 2001
+++ linux-2.4.5-ac24.acme/drivers/net/pppoe.c	Wed Jul  4 11:00:47 2001
@@ -567,17 +567,17 @@
 		if (!dev)
 			goto end;
 
+		po->pppoe_dev = dev;
+
 		if( ! (dev->flags & IFF_UP) )
-			goto end;
+			goto err_put;
 		memcpy(&po->pppoe_pa,
 		       &sp->sa_addr.pppoe,
 		       sizeof(struct pppoe_addr));
 
 		error = set_item(po);
 		if (error < 0)
-			goto end;
-
-		po->pppoe_dev = dev;
+			goto err_put;
 
 		po->chan.hdrlen = (sizeof(struct pppoe_hdr) +
 				   dev->hard_header_len);
@@ -586,6 +586,8 @@
 		po->chan.ops = &pppoe_chan_ops;
 
 		error = ppp_register_channel(&po->chan);
+		if (error)
+			goto err_put;
 
 		sk->state = PPPOX_CONNECTED;
 	}
@@ -595,6 +597,10 @@
  end:
 	release_sock(sk);
 	return error;
+err_put:
+	dev_put(po->pppoe_dev);
+	po->pppoe_dev = NULL;
+	goto end;
 }
 
- Arnaldo

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH] Re: dev_get_by_name without dev_put
  2001-07-04 13:49 dev_get_by_name without dev_put fabrizio.gennari
  2001-07-04 14:06 ` [PATCH] " Arnaldo Carvalho de Melo
@ 2001-07-10  0:36 ` David S. Miller
  1 sibling, 0 replies; 3+ messages in thread
From: David S. Miller @ 2001-07-10  0:36 UTC (permalink / raw)
  To: Arnaldo Carvalho de Melo; +Cc: fabrizio.gennari, linux-kernel


Arnaldo Carvalho de Melo writes:
 > For pppoe (and one has to look at the set_item and see if delete_item is
 > needed).

Applied, sorry for taking so long.

Later,
David S. Miller
davem@redhat.com

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2001-07-10  0:36 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2001-07-04 13:49 dev_get_by_name without dev_put fabrizio.gennari
2001-07-04 14:06 ` [PATCH] " Arnaldo Carvalho de Melo
2001-07-10  0:36 ` David S. Miller

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).