netdev - [PATCH 1/2] fec: Fix inconsistent lock state

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Date:	Wed, 15 May 2013 14:06:26 -0300
From:	Fabio Estevam <fabio.estevam@...escale.com>
To:	<davem@...emloft.net>
CC:	<shawn.guo@...aro.org>, <Frank.Li@...escale.com>,
	<kernel@...gutronix.de>, <eric.dumazet@...il.com>,
	<netdev@...r.kernel.org>,
	Fabio Estevam <fabio.estevam@...escale.com>
Subject: [PATCH 1/2] fec: Fix inconsistent lock state

fec_restart() runs in softirq context and we should use the 
netif_tx_lock_bh/netif_tx_unlock_bh() variants to avoid the following warning
that happens since commit 54309fa6 ("net: fec: fix kernel oops when plug/unplug 
cable many times"):

[    9.753168] =================================
[    9.757540] [ INFO: inconsistent lock state ]
[    9.761921] 3.10.0-rc1-next-20130514 #13 Not tainted
[    9.766897] ---------------------------------
[    9.771264] inconsistent {SOFTIRQ-ON-W} -> {IN-SOFTIRQ-W} usage.
[    9.777288] swapper/0 [HC0[0]:SC1[3]:HE1:SE0] takes:
[    9.782261]  (_xmit_ETHER#2){+.?...}, at: [<c03c24a4>] sch_direct_xmit+0xa0/0x2d4
[    9.789879] {SOFTIRQ-ON-W} state was registered at:
[    9.794769]   [<c0059c60>] __lock_acquire+0x528/0x1bc0
[    9.799953]   [<c005b838>] lock_acquire+0xa0/0x108
[    9.804780]   [<c0441320>] _raw_spin_lock+0x28/0x38
[    9.809702]   [<c02f9fc8>] fec_restart+0x5d0/0x664
[    9.814542]   [<c02fa738>] fec_enet_adjust_link+0xa8/0xc0
[    9.819978]   [<c02f7a28>] phy_state_machine+0x2fc/0x370
[    9.825323]   [<c0035ee0>] process_one_work+0x1c0/0x4a0
[    9.830589]   [<c0036594>] worker_thread+0x138/0x394
[    9.835587]   [<c003c620>] kthread+0xa4/0xb0
[    9.839890]   [<c000e820>] ret_from_fork+0x14/0x34
[    9.844728] irq event stamp: 185984
[    9.848226] hardirqs last  enabled at (185984): [<c00232b0>] local_bh_enable_ip+0x84/0xf0
[    9.856450] hardirqs last disabled at (185983): [<c0023270>] local_bh_enable_ip+0x44/0xf0
[    9.864667] softirqs last  enabled at (185966): [<c0023470>] irq_enter+0x64/0x68
[    9.872099] softirqs last disabled at (185967): [<c0023510>] irq_exit+0x9c/0xd8
[    9.879440] 
[    9.879440] other info that might help us debug this:
[    9.885981]  Possible unsafe locking scenario:
[    9.885981] 
[    9.891912]        CPU0
[    9.894364]        ----
[    9.896814]   lock(_xmit_ETHER#2);
[    9.900259]   <Interrupt>
[    9.902884]     lock(_xmit_ETHER#2);
[    9.906500] 
[    9.906500]  *** DEADLOCK ***

Reported-by: Shawn Guo <shawn.guo@...aro.org>
Suggested-by: Eric Dumazet <eric.dumazet@...il.com>
Signed-off-by: Fabio Estevam <fabio.estevam@...escale.com>
---
 drivers/net/ethernet/freescale/fec_main.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/freescale/fec_main.c b/drivers/net/ethernet/freescale/fec_main.c
index ca9825c..658fbc1 100644
--- a/drivers/net/ethernet/freescale/fec_main.c
+++ b/drivers/net/ethernet/freescale/fec_main.c
@@ -451,7 +451,7 @@ fec_restart(struct net_device *ndev, int duplex)
 		netif_device_detach(ndev);
 		napi_disable(&fep->napi);
 		netif_stop_queue(ndev);
-		netif_tx_lock(ndev);
+		netif_tx_lock_bh(ndev);
 	}
 
 	/* Whack a reset.  We should wait for this. */
@@ -619,7 +619,7 @@ fec_restart(struct net_device *ndev, int duplex)
 		netif_device_attach(ndev);
 		napi_enable(&fep->napi);
 		netif_wake_queue(ndev);
-		netif_tx_unlock(ndev);
+		netif_tx_unlock_bh(ndev);
 	}
 }
 
-- 
1.8.1.2


--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html