rxrpc, afs: Fix peer hash locking vs RCU callback
Subsystems (3)
David Howells
maintainerPATCH 1/1 rxrpc, afs: Fix peer hash locking vs RCU callback
In its address list, afs now retains pointers to and refs on one or more rxrpc_peer objects. The address list is freed under RCU and at this time, it puts the refs on those peers.
Now, when an rxrpc_peer object runs out of refs, it gets removed from the peer hash table and, for that, rxrpc has to take a spinlock. However, it is now being called from afs's RCU cleanup, which takes place in BH context - but it is just taking an ordinary spinlock.
The put may also be called from non-BH context, and so there exists the possibility of deadlock if the BH-based RCU cleanup happens whilst the hash spinlock is held. This led to the attached lockdep complaint.
Fix this by changing spinlocks of rxnet->peer_hash_lock back to BH-disabling locks.
================================
WARNING: inconsistent lock state
6.13.0-rc5-build2+ #1223 Tainted: G E
--------------------------------
inconsistent {SOFTIRQ-ON-W} -> {IN-SOFTIRQ-W} usage.
swapper/1/0 [HC0[0]:SC1[1]:HE1:SE0] takes:
ffff88810babe228 (&rxnet->peer_hash_lock){+.?.}-{3:3}, at: rxrpc_put_peer+0xcb/0x180
{SOFTIRQ-ON-W} state was registered at:
mark_usage+0x164/0x180
__lock_acquire+0x544/0x990
lock_acquire.part.0+0x103/0x280
_raw_spin_lock+0x2f/0x40
rxrpc_peer_keepalive_worker+0x144/0x440
process_one_work+0x486/0x7c0
process_scheduled_works+0x73/0x90
worker_thread+0x1c8/0x2a0
kthread+0x19b/0x1b0
ret_from_fork+0x24/0x40
ret_from_fork_asm+0x1a/0x30
irq event stamp: 972402
hardirqs last enabled at (972402): [<ffffffff8244360e>] _raw_spin_unlock_irqrestore+0x2e/0x50
hardirqs last disabled at (972401): [<ffffffff82443328>] _raw_spin_lock_irqsave+0x18/0x60
softirqs last enabled at (972300): [<ffffffff810ffbbe>] handle_softirqs+0x3ee/0x430
softirqs last disabled at (972313): [<ffffffff810ffc54>] __irq_exit_rcu+0x44/0x110
other info that might help us debug this:
Possible unsafe locking scenario:
CPU0
----
lock(&rxnet->peer_hash_lock);
<Interrupt>
lock(&rxnet->peer_hash_lock);
*** DEADLOCK ***
1 lock held by swapper/1/0:
#0: ffffffff83576be0 (rcu_callback){....}-{0:0}, at: rcu_lock_acquire+0x7/0x30
stack backtrace:
CPU: 1 UID: 0 PID: 0 Comm: swapper/1 Tainted: G E 6.13.0-rc5-build2+ #1223
Tainted: [E]=UNSIGNED_MODULE
Hardware name: ASUS All Series/H97-PLUS, BIOS 2306 10/09/2014
Call Trace:
<IRQ>
dump_stack_lvl+0x57/0x80
print_usage_bug.part.0+0x227/0x240
valid_state+0x53/0x70
mark_lock_irq+0xa5/0x2f0
mark_lock+0xf7/0x170
mark_usage+0xe1/0x180
__lock_acquire+0x544/0x990
lock_acquire.part.0+0x103/0x280
_raw_spin_lock+0x2f/0x40
rxrpc_put_peer+0xcb/0x180
afs_free_addrlist+0x46/0x90 [kafs]
rcu_do_batch+0x2d2/0x640
rcu_core+0x2f7/0x350
handle_softirqs+0x1ee/0x430
__irq_exit_rcu+0x44/0x110
irq_exit_rcu+0xa/0x30
sysvec_apic_timer_interrupt+0x7f/0xa0
</IRQ>--- net/rxrpc/peer_event.c | 16 ++++++++-------- net/rxrpc/peer_object.c | 12 ++++++------ 2 files changed, 14 insertions(+), 14 deletions(-)
net/rxrpc/peer_event.c +8 -8
@@ -246,7 +246,7 @@ static void rxrpc_peer_keepalive_dispatch(struct rxrpc_net *rxnet,bool use;int slot; -spin_lock (&rxnet->peer_hash_lock); +spin_lock_bh (&rxnet->peer_hash_lock);while (!list_empty (collector)) { peer = list_entry(collector->next,
@@ -257,7 +257,7 @@ static void rxrpc_peer_keepalive_dispatch(struct rxrpc_net *rxnet,continue ; use =__rxrpc_use_local (peer->local, rxrpc_local_use_peer_keepalive); -spin_unlock (&rxnet->peer_hash_lock); +spin_unlock_bh (&rxnet->peer_hash_lock);if (use) { keepalive_at = peer->last_tx_at +RXRPC_KEEPALIVE_TIME ;
@@ -277,17 +277,17 @@ static void rxrpc_peer_keepalive_dispatch(struct rxrpc_net *rxnet,*/ slot += cursor; slot &= mask; -spin_lock (&rxnet->peer_hash_lock); +spin_lock_bh (&rxnet->peer_hash_lock);list_add_tail (&peer->keepalive_link, &rxnet->peer_keepalive[slot & mask]); -spin_unlock (&rxnet->peer_hash_lock); +spin_unlock_bh (&rxnet->peer_hash_lock);rxrpc_unuse_local (peer->local, rxrpc_local_unuse_peer_keepalive); }rxrpc_put_peer (peer, rxrpc_peer_put_keepalive); -spin_lock (&rxnet->peer_hash_lock); +spin_lock_bh (&rxnet->peer_hash_lock); } -spin_unlock (&rxnet->peer_hash_lock); +spin_unlock_bh (&rxnet->peer_hash_lock); }/*
@@ -317,7 +317,7 @@ void rxrpc_peer_keepalive_worker(struct work_struct *work)* second; the bucket at cursor + 1 goes at now + 1s and so * on... */ -spin_lock (&rxnet->peer_hash_lock); +spin_lock_bh (&rxnet->peer_hash_lock);list_splice_init (&rxnet->peer_keepalive_new, &collector); stop = cursor +ARRAY_SIZE (rxnet->peer_keepalive);
@@ -329,7 +329,7 @@ void rxrpc_peer_keepalive_worker(struct work_struct *work) } base = now; -spin_unlock (&rxnet->peer_hash_lock); +spin_unlock_bh (&rxnet->peer_hash_lock); rxnet->peer_keepalive_base = base; rxnet->peer_keepalive_cursor = cursor;
net/rxrpc/peer_object.c +6 -6
@@ -325,10 +325,10 @@ void rxrpc_new_incoming_peer(struct rxrpc_local *local, struct rxrpc_peer *peer) hash_key =rxrpc_peer_hash_key (local, &peer->srx);rxrpc_init_peer (local, peer, hash_key); -spin_lock (&rxnet->peer_hash_lock); +spin_lock_bh (&rxnet->peer_hash_lock);hash_add_rcu (rxnet->peer_hash, &peer->hash_link, hash_key);list_add_tail (&peer->keepalive_link, &rxnet->peer_keepalive_new); -spin_unlock (&rxnet->peer_hash_lock); +spin_unlock_bh (&rxnet->peer_hash_lock); }/*
@@ -360,7 +360,7 @@ struct rxrpc_peer *rxrpc_lookup_peer(struct rxrpc_local *local,return NULL ; } -spin_lock (&rxnet->peer_hash_lock); +spin_lock_bh (&rxnet->peer_hash_lock);/* Need to check that we aren't racing with someone else */ peer =__rxrpc_lookup_peer_rcu (local, srx, hash_key);
@@ -373,7 +373,7 @@ struct rxrpc_peer *rxrpc_lookup_peer(struct rxrpc_local *local, &rxnet->peer_keepalive_new); } -spin_unlock (&rxnet->peer_hash_lock); +spin_unlock_bh (&rxnet->peer_hash_lock);if (peer)rxrpc_free_peer (candidate);
@@ -423,10 +423,10 @@ static void __rxrpc_put_peer(struct rxrpc_peer *peer)ASSERT (hlist_empty (&peer->error_targets)); -spin_lock (&rxnet->peer_hash_lock); +spin_lock_bh (&rxnet->peer_hash_lock);hash_del_rcu (&peer->hash_link);list_del_init (&peer->keepalive_link); -spin_unlock (&rxnet->peer_hash_lock); +spin_unlock_bh (&rxnet->peer_hash_lock);rxrpc_free_peer (peer); }
This patch was applied to netdev/net.git (main) by Jakub Kicinski <[email protected]>:
Hello:
This patch was applied to netdev/net.git (main)
by Jakub Kicinski <[email protected]>:
Quoted 10 lines · On Thu, 23 Jan 2025 08:59:12 +0000 you wrote:
In its address list, afs now retains pointers to and refs on one or more
rxrpc_peerobjects. The address list is freed under RCU and at this time, it puts the refs on those peers.Now, when an
rxrpc_peerobject runs out of refs, it gets removed from the peer hash table and, for that, rxrpc has to take a spinlock. However, it is now being called from afs's RCU cleanup, which takes place in BH context - but it is just taking an ordinary spinlock.[...]
Here is the summary with links:
- [net] rxrpc, afs: Fix peer hash locking vs RCU callback
https://git.kernel.org/netdev/net/c/79d458c13056
You are awesome, thank you!
Signature · 2 lines
Deet-doot-dot, I am a bot. https://korg.docs.kernel.org/patchwork/pwbot.html