[PATCH v2] ipv6: addrconf: Fix recursive spin lock call

[PATCH v2] ipv6: addrconf: Fix recursive spin lock call

[subashab]

A rcu stall with the following backtrace was seen on a system with
forwarding, optimistic_dad and use_optimistic set. To reproduce,
set these flags and allow ipv6 autoconf.

This occurs because the device write_lock is acquired while already
holding the read_lock. Back trace below -

INFO: rcu_preempt self-detected stall on CPU { 1}  (t=2100 jiffies
 g=3992 c=3991 q=4471)
<6> Task dump for CPU 1:
<2> kworker/1:0     R  running task    12168    15   2 0x00000002
<2> Workqueue: ipv6_addrconf addrconf_dad_work
<6> Call trace:
<2> [<ffffffc000084da8>] el1_irq+0x68/0xdc
<2> [<ffffffc000cc4e0c>] _raw_write_lock_bh+0x20/0x30
<2> [<ffffffc000bc5dd8>] __ipv6_dev_ac_inc+0x64/0x1b4
<2> [<ffffffc000bcbd2c>] addrconf_join_anycast+0x9c/0xc4
<2> [<ffffffc000bcf9f0>] __ipv6_ifa_notify+0x160/0x29c
<2> [<ffffffc000bcfb7c>] ipv6_ifa_notify+0x50/0x70
<2> [<ffffffc000bd035c>] addrconf_dad_work+0x314/0x334
<2> [<ffffffc0000b64c8>] process_one_work+0x244/0x3fc
<2> [<ffffffc0000b7324>] worker_thread+0x2f8/0x418
<2> [<ffffffc0000bb40c>] kthread+0xe0/0xec

v2: do addrconf_dad_kick inside read lock and then acquire write
lock for ipv6_ifa_notify as suggested by Eric

Fixes: 7fd2561e4ebdd ("net: ipv6: Add a sysctl to make optimistic
addresses useful candidates")

Cc: Eric Dumazet <edumazet@google.com>
Cc: Erik Kline <ek@google.com>
Cc: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: Subash Abhinov Kasiviswanathan <subashab@codeaurora.org>
---
 net/ipv6/addrconf.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c
index 38eedde..9efd9ff 100644
--- a/net/ipv6/addrconf.c
+++ b/net/ipv6/addrconf.c
@@ -3538,6 +3538,7 @@ static void addrconf_dad_begin(struct inet6_ifaddr *ifp)
 {
 	struct inet6_dev *idev = ifp->idev;
 	struct net_device *dev = idev->dev;
+	bool notify = false;

 	addrconf_join_solict(dev, &ifp->addr);

@@ -3583,7 +3584,7 @@ static void addrconf_dad_begin(struct inet6_ifaddr *ifp)
 			/* Because optimistic nodes can use this address,
 			 * notify listeners. If DAD fails, RTM_DELADDR is sent.
 			 */
-			ipv6_ifa_notify(RTM_NEWADDR, ifp);
+			notify = true;
 		}
 	}

@@ -3591,6 +3592,8 @@ static void addrconf_dad_begin(struct inet6_ifaddr *ifp)
 out:
 	spin_unlock(&ifp->lock);
 	read_unlock_bh(&idev->lock);
+	if (notify)
+		ipv6_ifa_notify(RTM_NEWADDR, ifp);
 }

 static void addrconf_dad_start(struct inet6_ifaddr *ifp)
-- 

Re: [PATCH v2] ipv6: addrconf: Fix recursive spin lock call

[Hannes Frederic Sowa]

On 02.02.2016 03:11, subashab@codeaurora.org wrote:
> A rcu stall with the following backtrace was seen on a system with
> forwarding, optimistic_dad and use_optimistic set. To reproduce,
> set these flags and allow ipv6 autoconf.
>
> This occurs because the device write_lock is acquired while already
> holding the read_lock. Back trace below -
>
> INFO: rcu_preempt self-detected stall on CPU { 1}  (t=2100 jiffies
>   g=3992 c=3991 q=4471)
> <6> Task dump for CPU 1:
> <2> kworker/1:0     R  running task    12168    15   2 0x00000002
> <2> Workqueue: ipv6_addrconf addrconf_dad_work
> <6> Call trace:
> <2> [<ffffffc000084da8>] el1_irq+0x68/0xdc
> <2> [<ffffffc000cc4e0c>] _raw_write_lock_bh+0x20/0x30
> <2> [<ffffffc000bc5dd8>] __ipv6_dev_ac_inc+0x64/0x1b4
> <2> [<ffffffc000bcbd2c>] addrconf_join_anycast+0x9c/0xc4
> <2> [<ffffffc000bcf9f0>] __ipv6_ifa_notify+0x160/0x29c
> <2> [<ffffffc000bcfb7c>] ipv6_ifa_notify+0x50/0x70
> <2> [<ffffffc000bd035c>] addrconf_dad_work+0x314/0x334
> <2> [<ffffffc0000b64c8>] process_one_work+0x244/0x3fc
> <2> [<ffffffc0000b7324>] worker_thread+0x2f8/0x418
> <2> [<ffffffc0000bb40c>] kthread+0xe0/0xec
>
> v2: do addrconf_dad_kick inside read lock and then acquire write
> lock for ipv6_ifa_notify as suggested by Eric
>
> Fixes: 7fd2561e4ebdd ("net: ipv6: Add a sysctl to make optimistic
> addresses useful candidates")
>
> Cc: Eric Dumazet <edumazet@google.com>
> Cc: Erik Kline <ek@google.com>
> Cc: Hannes Frederic Sowa <hannes@stressinduktion.org>
> Signed-off-by: Subash Abhinov Kasiviswanathan <subashab@codeaurora.org>

Acked-by: Hannes Frederic Sowa <hannes@stressinduktion.org>

Definitely a stable candidate, thanks!

Re: [PATCH v2] ipv6: addrconf: Fix recursive spin lock call

[Eric Dumazet]

On Tue, 2016-02-02 at 02:11 +0000, subashab@codeaurora.org wrote:
> A rcu stall with the following backtrace was seen on a system with
> forwarding, optimistic_dad and use_optimistic set. To reproduce,
> set these flags and allow ipv6 autoconf.
> 
> This occurs because the device write_lock is acquired while already
> holding the read_lock. Back trace below -

Acked-by: Eric Dumazet <edumazet@google.com>

Re: [PATCH v2] ipv6: addrconf: Fix recursive spin lock call

[David Miller]

From: subashab@codeaurora.org
Date: Tue, 2 Feb 2016 02:11:10 -0000

> A rcu stall with the following backtrace was seen on a system with
> forwarding, optimistic_dad and use_optimistic set. To reproduce,
> set these flags and allow ipv6 autoconf.
> 
> This occurs because the device write_lock is acquired while already
> holding the read_lock. Back trace below -
 ...
> v2: do addrconf_dad_kick inside read lock and then acquire write
> lock for ipv6_ifa_notify as suggested by Eric
> 
> Fixes: 7fd2561e4ebdd ("net: ipv6: Add a sysctl to make optimistic
> addresses useful candidates")
> 
> Cc: Eric Dumazet <edumazet@google.com>
> Cc: Erik Kline <ek@google.com>
> Cc: Hannes Frederic Sowa <hannes@stressinduktion.org>
> Signed-off-by: Subash Abhinov Kasiviswanathan <subashab@codeaurora.org>

Applied and queued up for -stable, thanks.