From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: Re: [GIT]: Networking Date: Thu, 02 Jul 2009 12:51:17 +0200 Message-ID: <4A4C9125.8020705@gmail.com> References: <20090630.213927.180401151.davem@davemloft.net> <20090702075724.GA10608@elte.hu> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: David Miller , torvalds@linux-foundation.org, akpm@linux-foundation.org, netdev@vger.kernel.org, linux-kernel@vger.kernel.org To: Ingo Molnar Return-path: Received: from gw1.cosmosbay.com ([212.99.114.194]:50577 "EHLO gw1.cosmosbay.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752375AbZGBKvc (ORCPT ); Thu, 2 Jul 2009 06:51:32 -0400 In-Reply-To: <20090702075724.GA10608@elte.hu> Sender: netdev-owner@vger.kernel.org List-ID: Ingo Molnar a =E9crit : >> The following changes since commit 52989765629e7d182b4f146050ebba0ab= f2cb0b7: >> Linus Torvalds (1): >> Merge git://git.kernel.org/.../davem/net-2.6 >> >> are available in the git repository at: >> >> master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6.git mast= er >=20 > Hm, something in this lot quickly wrecked networking here - see the=20 > tx timeout dump below. It starts with: >=20 > [ 351.004596] WARNING: at net/sched/sch_generic.c:246 dev_watchdog+0= x10b/0x19c() > [ 351.011815] Hardware name: System Product Name > [ 351.016220] NETDEV WATCHDOG: eth0 (forcedeth): transmit queue 0 ti= med out >=20 > Config attached. Unfortunately i've got no time to do bisection=20 > today. forcedeth might have a problem, in its netif_wake_queue() logic, but I could not see why a recent patch could make this problem visible now. CPU0/1: AMD Athlon(tm) 64 X2 Dual Core Processor 3800+ stepping 02 is not a new cpu either :) forcedeth uses an internal tx_stop without appropriate barrier. Could you try following patch ? (random guess as I dont have much time right now) Thank you diff --git a/drivers/net/forcedeth.c b/drivers/net/forcedeth.c index 1094d29..dc6bbde 100644 --- a/drivers/net/forcedeth.c +++ b/drivers/net/forcedeth.c @@ -2165,7 +2165,7 @@ static int nv_start_xmit(struct sk_buff *skb, str= uct net_device *dev) empty_slots =3D nv_get_empty_tx_slots(np); if (unlikely(empty_slots <=3D entries)) { netif_stop_queue(dev); - np->tx_stop =3D 1; + set_mb(np->tx_stop, 1); spin_unlock_irqrestore(&np->lock, flags); return NETDEV_TX_BUSY; } @@ -2286,7 +2286,7 @@ static int nv_start_xmit_optimized(struct sk_buff= *skb, struct net_device *dev) empty_slots =3D nv_get_empty_tx_slots(np); if (unlikely(empty_slots <=3D entries)) { netif_stop_queue(dev); - np->tx_stop =3D 1; + set_mb(np->tx_stop, 1); spin_unlock_irqrestore(&np->lock, flags); return NETDEV_TX_BUSY; } @@ -2564,7 +2564,7 @@ static void nv_tx_timeout(struct net_device *dev) else status =3D readl(base + NvRegIrqStatus) & NVREG_IRQSTAT_MASK; =20 - printk(KERN_INFO "%s: Got tx_timeout. irq: %08x\n", dev->name, status= ); + printk(KERN_INFO "%s: Got tx_timeout. irq: %08x tx_stop=3D%d\n", dev-= >name, status, np->tx_stop); =20 { int i;