From: Michel Machado <hidden> Date: 2012-02-21 21:27:51
When the fixed race condition happens:
1. While function neigh_periodic_work scans the neighbor hash table
pointed by field tbl->nht, it unlocks and locks tbl->lock between
buckets in order to call cond_resched.
2. Assume that function neigh_periodic_work calls cond_resched, that is,
the lock tbl->lock is available, and function neigh_hash_grow runs.
3. Once function neigh_hash_grow finishes, and RCU calls
neigh_hash_free_rcu, the original struct neigh_hash_table that function
neigh_periodic_work was using doesn't exist anymore.
4. Once back at neigh_periodic_work, whenever the old struct
neigh_hash_table is accessed, things can go badly.
Signed-off-by: Michel Machado <redacted>
CC: "David S. Miller" <davem@davemloft.net>
CC: Eric Dumazet <redacted>
---
@@ -826,6 +826,8 @@ next_elt:write_unlock_bh(&tbl->lock);cond_resched();write_lock_bh(&tbl->lock);+nht=rcu_dereference_protected(tbl->nht,+lockdep_is_held(&tbl->lock));}/* Cycle through all hash buckets every base_reachable_time/2 ticks.*ARPentrytimeoutsrangefrom1/2base_reachable_timeto3/2
From: Eric Dumazet <hidden> Date: 2012-02-21 21:25:27
Le mardi 21 février 2012 à 16:04 -0500, Michel Machado a écrit :
quoted hunk
When the fixed race condition happens:
1. While function neigh_periodic_work scans the neighbor hash table
pointed by field tbl->nht, it unlocks and locks tbl->lock between
buckets in order to call cond_resched.
2. Assume that function neigh_periodic_work calls cond_resched, that is,
the lock tbl->lock is available, and function neigh_hash_grow runs.
3. Once function neigh_hash_grow finishes, and RCU calls
neigh_hash_free_rcu, the original struct neigh_hash_table that function
neigh_periodic_work was using doesn't exist anymore.
4. Once back at neigh_periodic_work, whenever the old struct
neigh_hash_table is accessed, things can go badly.
Signed-off-by: Michel Machado <redacted>
CC: "David S. Miller" <davem@davemloft.net>
CC: Eric Dumazet <redacted>
---
@@ -826,6 +826,8 @@ next_elt:write_unlock_bh(&tbl->lock);cond_resched();write_lock_bh(&tbl->lock);+nht=rcu_dereference_protected(tbl->nht,+lockdep_is_held(&tbl->lock));}/* Cycle through all hash buckets every base_reachable_time/2 ticks.*ARPentrytimeoutsrangefrom1/2base_reachable_timeto3/2
Nice catch !
Bug introduced in 2.6.37 in commit d6bf781712a (net neigh: RCU
conversion of neigh hash table)
Acked-by: Eric Dumazet <redacted>
From: David Miller <davem@davemloft.net> Date: 2012-02-21 21:29:12
From: Eric Dumazet <redacted>
Date: Tue, 21 Feb 2012 22:25:22 +0100
Le mardi 21 février 2012 à 16:04 -0500, Michel Machado a écrit :
quoted
When the fixed race condition happens:
1. While function neigh_periodic_work scans the neighbor hash table
pointed by field tbl->nht, it unlocks and locks tbl->lock between
buckets in order to call cond_resched.
2. Assume that function neigh_periodic_work calls cond_resched, that is,
the lock tbl->lock is available, and function neigh_hash_grow runs.
3. Once function neigh_hash_grow finishes, and RCU calls
neigh_hash_free_rcu, the original struct neigh_hash_table that function
neigh_periodic_work was using doesn't exist anymore.
4. Once back at neigh_periodic_work, whenever the old struct
neigh_hash_table is accessed, things can go badly.
Signed-off-by: Michel Machado <redacted>
...
Nice catch !
Bug introduced in 2.6.37 in commit d6bf781712a (net neigh: RCU
conversion of neigh hash table)
Acked-by: Eric Dumazet <redacted>