Skip to contentSkip to conversation

locking/lockdep: Use local_irq_save() with call_rcu()

PATCH In review Waiman Long 2 participants 3 messages Dec 22, 2020 – Jan 4, 2021 Locking primitives

lore
Share & tools
lore

Waiman Long

author reviewer
#0 lore raw

PATCH 1/1 locking/lockdep: Use local_irq_save() with call_rcu()

The following lockdep splat was hit:

 [  560.638354] WARNING: CPU: 79 PID: 27458 at kernel/rcu/tree_plugin.h:1749 call_rcu+0x6dc/0xf00
    :
 [  560.647761] RIP: 0010:call_rcu+0x6dc/0xf00
 [  560.647763] Code: 0f 8f 29 04 00 00 e8 93 da 1c 00 48 8b 3c 24 57 9d 0f 1f 44 00 00 e9 19 fa ff ff 65 8b 05 38 83 c4 49 85 c0 0f 84 cd fb ff ff <0f> 0b e9 c6 fb ff ff e8 b8 45 51 00 4c 89 f2 48 b8 00 00 00 00 00
 [  560.647764] RSP: 0018:ff11001050097b58 EFLAGS: 00010002
 [  560.647766] RAX: 0000000000000001 RBX: ffffffffbb1f3360 RCX: 0000000000000001
 [  560.647766] RDX: 0000000000000000 RSI: 0000000000000000 RDI: ffffffffb99bac9c
 [  560.647767] RBP: 1fe220020a012f73 R08: 000000010004005c R09: dffffc0000000000
 [  560.647768] R10: dffffc0000000000 R11: 0000000000000003 R12: ff1100105b7f70e1
 [  560.647769] R13: ffffffffb635d8a0 R14: ff1100105b7f72d8 R15: ff1100105b7f7040
 [  560.647770] FS:  00007fd9b3437080(0000) GS:ff1100105b600000(0000) knlGS:0000000000000000
 [  560.647771] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
 [  560.647772] CR2: 00007fd9b30112bc CR3: 000000105e898006 CR4: 0000000000761ee0
 [  560.647773] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
 [  560.647773] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
 [  560.647774] PKRU: 55555554
 [  560.647774] Call Trace:
 [  560.647778]  ? invoke_rcu_core+0x180/0x180
 [  560.647782]  ? __is_module_percpu_address+0xed/0x440
 [  560.647787]  lockdep_unregister_key+0x2ab/0x5b0
 [  560.647791]  destroy_workqueue+0x40b/0x610
 [  560.647862]  xlog_dealloc_log+0x216/0x2b0 [xfs]
    :

This splat is caused by the fact that lockdep_unregister_key() uses raw_local_irq_save() which doesn't update the hardirqs_enabled percpu flag. The call_rcu() function, however, will call lockdep_assert_irqs_disabled() to check the hardirqs_enabled flag which remained set in this case.

Fix this problem by using local_irq_save()/local_irq_restore() pairs whenever call_rcu() is being called.

I think raw_local_irq_save() function can be used if no external function is being called except maybe printk() as it means another lockdep problem exists.

Fixes a0b0fd53e1e67 ("locking/lockdep: Free lock classes that are no longer in use")
---
 kernel/locking/lockdep.c | 20 ++++++++++++--------
 1 file changed, 12 insertions(+), 8 deletions(-)
kernel/locking/lockdep.c +12 -8
index c1418b47f625..2a37af77ede6 100644
@@ -5884,7 +5884,7 @@ static void free_zapped_rcu(struct rcu_head *ch)
 	if (WARN_ON_ONCE(ch != &delayed_free.rcu_head))
 		return;
 
-	raw_local_irq_save(flags);
+	local_irq_save(flags);
 	lockdep_lock();
 
 	/* closed head */
@@ -5898,7 +5898,7 @@ static void free_zapped_rcu(struct rcu_head *ch)
 	call_rcu_zapped(delayed_free.pf + delayed_free.index);
 
 	lockdep_unlock();
-	raw_local_irq_restore(flags);
+	local_irq_restore(flags);
 }
 
 /*
@@ -5941,13 +5941,13 @@ static void lockdep_free_key_range_reg(void *start, unsigned long size)
 
 	init_data_structures_once();
 
-	raw_local_irq_save(flags);
+	local_irq_save(flags);
 	lockdep_lock();
 	pf = get_pending_free();
 	__lockdep_free_key_range(pf, start, size);
 	call_rcu_zapped(pf);
 	lockdep_unlock();
-	raw_local_irq_restore(flags);
+	local_irq_restore(flags);
 
 	/*
 	 * Wait for any possible iterators from look_up_lock_class() to pass
@@ -6043,7 +6043,7 @@ static void lockdep_reset_lock_reg(struct lockdep_map *lock)
 	unsigned long flags;
 	int locked;
 
-	raw_local_irq_save(flags);
+	local_irq_save(flags);
 	locked = graph_lock();
 	if (!locked)
 		goto out_irq;
@@ -6054,7 +6054,7 @@ static void lockdep_reset_lock_reg(struct lockdep_map *lock)
 
 	graph_unlock();
 out_irq:
-	raw_local_irq_restore(flags);
+	local_irq_restore(flags);
 }
 
 /*
@@ -6098,7 +6098,11 @@ void lockdep_unregister_key(struct lock_class_key *key)
 	if (WARN_ON_ONCE(static_obj(key)))
 		return;
 
-	raw_local_irq_save(flags);
+	/*
+	 * local_irq_save() should be used as call_rcu() will check
+	 * hardirqs_enabled state.
+	 */
+	local_irq_save(flags);
 	if (!graph_lock())
 		goto out_irq;
 
@@ -6115,7 +6119,7 @@ void lockdep_unregister_key(struct lock_class_key *key)
 	call_rcu_zapped(pf);
 	graph_unlock();
 out_irq:
-	raw_local_irq_restore(flags);
+	local_irq_restore(flags);
 
 	/* Wait until is_dynamic_key() has finished accessing k->hash_entry. */
 	synchronize_rcu();
Signature · 1 line
2.18.1
maintainer
#1 lore raw
Quoted 26 lines · Waiman Long

The following lockdep splat was hit:

 [  560.638354] WARNING: CPU: 79 PID: 27458 at kernel/rcu/tree_plugin.h:1749 call_rcu+0x6dc/0xf00
    :
 [  560.647761] RIP: 0010:call_rcu+0x6dc/0xf00
 [  560.647763] Code: 0f 8f 29 04 00 00 e8 93 da 1c 00 48 8b 3c 24 57 9d 0f 1f 44 00 00 e9 19 fa ff ff 65 8b 05 38 83 c4 49 85 c0 0f 84 cd fb ff ff <0f> 0b e9 c6 fb ff ff e8 b8 45 51 00 4c 89 f2 48 b8 00 00 00 00 00
 [  560.647764] RSP: 0018:ff11001050097b58 EFLAGS: 00010002
 [  560.647766] RAX: 0000000000000001 RBX: ffffffffbb1f3360 RCX: 0000000000000001
 [  560.647766] RDX: 0000000000000000 RSI: 0000000000000000 RDI: ffffffffb99bac9c
 [  560.647767] RBP: 1fe220020a012f73 R08: 000000010004005c R09: dffffc0000000000
 [  560.647768] R10: dffffc0000000000 R11: 0000000000000003 R12: ff1100105b7f70e1
 [  560.647769] R13: ffffffffb635d8a0 R14: ff1100105b7f72d8 R15: ff1100105b7f7040
 [  560.647770] FS:  00007fd9b3437080(0000) GS:ff1100105b600000(0000) knlGS:0000000000000000
 [  560.647771] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
 [  560.647772] CR2: 00007fd9b30112bc CR3: 000000105e898006 CR4: 0000000000761ee0
 [  560.647773] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
 [  560.647773] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
 [  560.647774] PKRU: 55555554
 [  560.647774] Call Trace:
 [  560.647778]  ? invoke_rcu_core+0x180/0x180
 [  560.647782]  ? __is_module_percpu_address+0xed/0x440
 [  560.647787]  lockdep_unregister_key+0x2ab/0x5b0
 [  560.647791]  destroy_workqueue+0x40b/0x610
 [  560.647862]  xlog_dealloc_log+0x216/0x2b0 [xfs]
    :
Quoted 8 lines

This splat is caused by the fact that lockdep_unregister_key() uses raw_local_irq_save() which doesn't update the hardirqs_enabled percpu flag. The call_rcu() function, however, will call lockdep_assert_irqs_disabled() to check the hardirqs_enabled flag which remained set in this case.

Fix this problem by using local_irq_save()/local_irq_restore() pairs whenever call_rcu() is being called.

I'm not sure I much like all this,.. :/

Quoted 3 lines

I think raw_local_irq_save() function can be used if no external function is being called except maybe printk() as it means another lockdep problem exists.

The reason lockdep is using raw_local_irq_save() is to avoid calling into itself again, notably local_irq_restore() will end up in mark_held_locks().

Quoted 1 line
Fixes a0b0fd53e1e67 ("locking/lockdep: Free lock classes that are no longer in use")

Seems dubious, as the lockdep_assert_irqs_disabled() that triggered was added after that patch.

I'm thinking another solution would be to increment the lockdep recursion count before calling RCU, because then we'll fail __lockdep_enabled and the assertion gets killed. Hmm?

author reviewer
#2 lore raw
Quoted 44 lines · Peter Zijlstra
Quoted 26 lines · Waiman Long

The following lockdep splat was hit:

  [  560.638354] WARNING: CPU: 79 PID: 27458 at kernel/rcu/tree_plugin.h:1749 call_rcu+0x6dc/0xf00
     :
  [  560.647761] RIP: 0010:call_rcu+0x6dc/0xf00
  [  560.647763] Code: 0f 8f 29 04 00 00 e8 93 da 1c 00 48 8b 3c 24 57 9d 0f 1f 44 00 00 e9 19 fa ff ff 65 8b 05 38 83 c4 49 85 c0 0f 84 cd fb ff ff <0f> 0b e9 c6 fb ff ff e8 b8 45 51 00 4c 89 f2 48 b8 00 00 00 00 00
  [  560.647764] RSP: 0018:ff11001050097b58 EFLAGS: 00010002
  [  560.647766] RAX: 0000000000000001 RBX: ffffffffbb1f3360 RCX: 0000000000000001
  [  560.647766] RDX: 0000000000000000 RSI: 0000000000000000 RDI: ffffffffb99bac9c
  [  560.647767] RBP: 1fe220020a012f73 R08: 000000010004005c R09: dffffc0000000000
  [  560.647768] R10: dffffc0000000000 R11: 0000000000000003 R12: ff1100105b7f70e1
  [  560.647769] R13: ffffffffb635d8a0 R14: ff1100105b7f72d8 R15: ff1100105b7f7040
  [  560.647770] FS:  00007fd9b3437080(0000) GS:ff1100105b600000(0000) knlGS:0000000000000000
  [  560.647771] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
  [  560.647772] CR2: 00007fd9b30112bc CR3: 000000105e898006 CR4: 0000000000761ee0
  [  560.647773] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
  [  560.647773] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
  [  560.647774] PKRU: 55555554
  [  560.647774] Call Trace:
  [  560.647778]  ? invoke_rcu_core+0x180/0x180
  [  560.647782]  ? __is_module_percpu_address+0xed/0x440
  [  560.647787]  lockdep_unregister_key+0x2ab/0x5b0
  [  560.647791]  destroy_workqueue+0x40b/0x610
  [  560.647862]  xlog_dealloc_log+0x216/0x2b0 [xfs]
     :
Quoted 8 lines

This splat is caused by the fact that lockdep_unregister_key() uses raw_local_irq_save() which doesn't update the hardirqs_enabled percpu flag. The call_rcu() function, however, will call lockdep_assert_irqs_disabled() to check the hardirqs_enabled flag which remained set in this case.

Fix this problem by using local_irq_save()/local_irq_restore() pairs whenever call_rcu() is being called.

I'm not sure I much like all this,.. :/

Quoted 3 lines

I think raw_local_irq_save() function can be used if no external function is being called except maybe printk() as it means another lockdep problem exists.

The reason lockdep is using raw_local_irq_save() is to avoid calling into itself again, notably local_irq_restore() will end up in mark_held_locks().

Quoted 8 lines
Quoted 1 line
Fixes a0b0fd53e1e67 ("locking/lockdep: Free lock classes that are no longer in use")

Seems dubious, as the lockdep_assert_irqs_disabled() that triggered was added after that patch.

I'm thinking another solution would be to increment the lockdep recursion count before calling RCU, because then we'll fail __lockdep_enabled and the assertion gets killed. Hmm?

Yes, you are right. Incrementing the lockdep recursion count should fix the issue. I was missing the commit 43be4388e94b ("lockdep: Put graph lock/unlock under lock_recursion protection"). This commit will properly increment the percpu lockdep recursion count and disable the call_rcu warning. So please ignore this patch.

Thanks,
Longman