locking/lockdep: Support dynamic lockdep keys
Subsystems (2)
cover locking/lockdep: Support dynamic lockdep keys
Hi Ingo,
As you may know some false positive lockdep reports are the result of the requirement to associate the same static lockdep key with all instances of a locking object. Recently I encountered a lockdep false positive for which I found no elegant way to suppress it other than by modifying the lockdep implementation. Hence this series with two patches: one that modifies lockdep and another patch that shows the false positive I and others ran into and that also shows how to suppress that false positive.
Please keep in mind that I'm not a lockdep expert and hence that I'm not sure whether the approach of patch 1/2 is correct.
Thanks,
Bart.
Bart Van Assche (2):
locking/lockdep: Add support for dynamic depmaps and keys kernel/workqueue: Use dynamic lockdep keys for workqueues
include/linux/lockdep.h | 2 ++ include/linux/workqueue.h | 28 ++++------------------- kernel/locking/lockdep.c | 16 ++++++++++--- kernel/workqueue.c | 48 ++++++++++++++++++++++++++++++++------- 4 files changed, 59 insertions(+), 35 deletions(-)
Signature · 1 line
2.19.1.930.g4563a0d9d0-goog
PATCH 1/2 locking/lockdep: Add support for dynamic depmaps and keys
The lock validator forces to categorize multiple instances of a lock object as the same lock class because it requires that struct lockdep_map and struct lock_class_key instances are static objects. This can result in false positive lockdep reports that are hard to suppress in an elegant way. Hence add support for allocating instances of these objects dynamically.
--- include/linux/lockdep.h | 2 ++ kernel/locking/lockdep.c | 16 +++++++++++++--- 2 files changed, 15 insertions(+), 3 deletions(-)
include/linux/lockdep.h +2 -0
@@ -56,6 +56,7 @@ struct lockdep_subclass_key {struct lock_class_key {struct lockdep_subclass_key subkeys[MAX_LOCKDEP_SUBCLASSES ]; +bool dynamic:1 ; };extern struct lock_class_key __lockdep_no_validate__;
@@ -153,6 +154,7 @@ struct lockdep_map {int cpu;unsigned long ip;#endif + bool dynamic:1; }; static inline void lockdep_copy_map(struct lockdep_map *to,
kernel/locking/lockdep.c +13 -3
@@ -624,6 +624,16 @@ static int static_obj(void *obj) }#endif +static bool is_valid_lockdep_map (struct lockdep_map *lock) +{ +return static_obj (lock) || lock->dynamic; +} + +static bool is_valid_key (struct lock_class_key *key) +{ +return static_obj (key) || key->dynamic; +} +/* * To make lock name printouts unique, we calculate a unique * class->name_version generation counter:
@@ -716,7 +726,7 @@ static bool assign_lock_key(struct lockdep_map *lock) lock->key = (void *)can_addr; else if (__is_module_percpu_address(addr, &can_addr)) lock->key = (void *)can_addr; - else if (static_obj(lock)) + else if (is_valid_lockdep_map(lock)) lock->key = (void *)lock; else {/* Debug-check: all keys must be persistent! */
@@ -752,7 +762,7 @@ register_lock_class(struct lockdep_map *lock, unsigned int subclass, int force)if (!lock->key) {if (!assign_lock_key (lock))return NULL ; - }else if (!static_obj (lock->key)) { + }else if (!is_valid_key (lock->key)) {return NULL ; }
@@ -3118,7 +3128,7 @@ static void __lockdep_init_map(struct lockdep_map *lock, const char *name,/* * Sanity check, the lock-class key must be persistent: */ -if (!static_obj (key)) { +if (!is_valid_key (key)) {printk ("BUG: key %px not in .data!\n" , key);/* * What it says above ^^^^^, I suggest you read it.
Signature · 1 line
2.19.1.930.g4563a0d9d0-goog
Peter Zijlstra
maintainerRe: [PATCH 1/2] locking/lockdep: Add support for dynamic depmaps and keys
Quoted 5 lines · Bart Van Assche
The lock validator forces to categorize multiple instances of a lock object as the same lock class because it requires that struct
lockdep_mapand structlock_class_keyinstances are static objects. This can result in false positive lockdep reports that are hard to suppress in an elegant way. Hence add support for allocating instances of these objects dynamically.
Yeah, I think not. You completely fail to explain how what you propose is correct.
The thing is; we rely on static objects because they provide persistence. Their address will never be re-used.
Dynamic objects do not provide this same guarantee. And when you re-use the key address for something else, you'll mix the chains and things come unstuck.
Re: [PATCH 1/2] locking/lockdep: Add support for dynamic depmaps and keys
Quoted 10 lines · Peter Zijlstra
Quoted 5 lines · Bart Van Assche
The lock validator forces to categorize multiple instances of a lock object as the same lock class because it requires that struct
lockdep_mapand structlock_class_keyinstances are static objects. This can result in false positive lockdep reports that are hard to suppress in an elegant way. Hence add support for allocating instances of these objects dynamically.Yeah, I think not. You completely fail to explain how what you propose is correct.
Quoted 6 lines
The thing is;we rely on static objects because they provide persistence. Their address will never be re-used.
Dynamic objects do not provide this same guarantee. And when you re-use the key address for something else, you'll mix the chains and things come unstuck.
Hi Peter,
Thanks for having taken a look. I will rework this patch taking your feedback into account and will post the new version when it's ready.
Bart.
PATCH 2/2 kernel/workqueue: Use dynamic lockdep keys for workqueues
In reply to Bart Van AsscheCommit 87915adc3f0a ("workqueue: re-add lockdep dependencies for flushing") improved deadlock checking in the workqueue implementation. Unfortunately that patch also introduced a few false positive lockdep complaints. This patch suppresses these false positives by allocating the workqueue mutex lockdep key dynamically. An example of a false positive lockdep complaint suppressed by this report can be found below. The root cause of the lockdep complaint shown below is that the direct I/O code can call alloc_workqueue() from inside a work item created by another alloc_workqueue() call and that both workqueues share the same lockdep key. This patch avoids that that lockdep complaint is triggered by allocating the work queue lockdep keys dynamically. In other words, this patch guarantees that a unique lockdep key is associated with each work queue mutex.
======================================================
WARNING: possible circular locking dependency detected
4.19.0-dbg+ #1 Not tainted
------------------------------------------------------
fio/4129 is trying to acquire lock:00000000a01cfe1a ((wq_completion)"dio/%s"sb->s_id){+.+.}, at: flush_workqueue+0xd0/0x970
but task is already holding lock:00000000a0acecf9 (&sb->s_type->i_mutex_key#14){+.+.}, at: ext4_file_write_iter+0x154/0x710
which lock already depends on the new lock.
the existing dependency chain (in reverse order) is:
-> #2 (&sb->s_type->i_mutex_key#14){+.+.}:
down_write+0x3d/0x80
__generic_file_fsync+0x77/0xf0
ext4_sync_file+0x3c9/0x780
vfs_fsync_range+0x66/0x100
dio_complete+0x2f5/0x360
dio_aio_complete_work+0x1c/0x20
process_one_work+0x481/0x9f0
worker_thread+0x63/0x5a0
kthread+0x1cf/0x1f0
ret_from_fork+0x24/0x30-> #1 ((work_completion)(&dio->complete_work)){+.+.}:
process_one_work+0x447/0x9f0
worker_thread+0x63/0x5a0
kthread+0x1cf/0x1f0
ret_from_fork+0x24/0x30-> #0 ((wq_completion)"dio/%s"sb->s_id){+.+.}:
lock_acquire+0xc5/0x200
flush_workqueue+0xf3/0x970
drain_workqueue+0xec/0x220
destroy_workqueue+0x23/0x350
sb_init_dio_done_wq+0x6a/0x80
do_blockdev_direct_IO+0x1f33/0x4be0
__blockdev_direct_IO+0x79/0x86
ext4_direct_IO+0x5df/0xbb0
generic_file_direct_write+0x119/0x220
__generic_file_write_iter+0x131/0x2d0
ext4_file_write_iter+0x3fa/0x710
aio_write+0x235/0x330
io_submit_one+0x510/0xeb0
__x64_sys_io_submit+0x122/0x340
do_syscall_64+0x71/0x220
entry_SYSCALL_64_after_hwframe+0x49/0xbeother info that might help us debug this:
Chain exists of: (wq_completion)"dio/%s"sb->s_id --> (work_completion)(&dio->complete_work) --> &sb->s_type->i_mutex_key#14 Possible unsafe locking scenario:
CPU0 CPU1
---- ----
lock(&sb->s_type->i_mutex_key#14);
lock((work_completion)(&dio->complete_work));
lock(&sb->s_type->i_mutex_key#14);lock((wq_completion)"dio/%s"sb->s_id);
*** DEADLOCK ***
1 lock held by fio/4129:
#0: 00000000a0acecf9 (&sb->s_type->i_mutex_key#14){+.+.}, at: ext4_file_write_iter+0x154/0x710
stack backtrace: CPU: 3 PID: 4129 Comm: fio Not tainted 4.19.0-dbg+ #1 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.10.2-1 04/01/2014 Call Trace: dump_stack+0x86/0xc5 print_circular_bug.isra.32+0x20a/0x218 __lock_acquire+0x1c68/0x1cf0 lock_acquire+0xc5/0x200 flush_workqueue+0xf3/0x970 drain_workqueue+0xec/0x220 destroy_workqueue+0x23/0x350 sb_init_dio_done_wq+0x6a/0x80 do_blockdev_direct_IO+0x1f33/0x4be0 __blockdev_direct_IO+0x79/0x86 ext4_direct_IO+0x5df/0xbb0 generic_file_direct_write+0x119/0x220 __generic_file_write_iter+0x131/0x2d0 ext4_file_write_iter+0x3fa/0x710 aio_write+0x235/0x330 io_submit_one+0x510/0xeb0 __x64_sys_io_submit+0x122/0x340 do_syscall_64+0x71/0x220 entry_SYSCALL_64_after_hwframe+0x49/0xbe
--- include/linux/workqueue.h | 28 ++++------------------- kernel/workqueue.c | 48 ++++++++++++++++++++++++++++++++------- 2 files changed, 44 insertions(+), 32 deletions(-)
include/linux/workqueue.h +4 -24
@@ -390,43 +390,23 @@ extern struct workqueue_struct *system_freezable_wq;extern struct workqueue_struct *system_power_efficient_wq;extern struct workqueue_struct *system_freezable_power_efficient_wq; -extern struct workqueue_struct * -__alloc_workqueue_key (const char *fmt,unsigned int flags,int max_active, -struct lock_class_key *key,const char *lock_name, ...)__printf (1 ,6 ); -/** * alloc_workqueue - allocate a workqueue * @fmt: printf format for the name of the workqueue * @flags: WQ_* flags * @max_active: max in-flight work items, 0 for default -* @args...: args for @fmt +* remaining args: args for @fmt * * Allocate a workqueue with the specified parameters. For detailed * information on WQ_* flags, please refer to * Documentation/core-api/workqueue.rst. * -* The __lock_name macro dance is to guarantee that single lock_class_key -* doesn't end up with different namesm, which isn't allowed by lockdep. -* * RETURNS: * Pointer to the allocated workqueue on success, %NULL on failure. */ -#ifdef CONFIG_LOCKDEP -#define alloc_workqueue(fmt, flags, max_active, args...) \ -({ \ - static struct lock_class_key __key; \ - const char *__lock_name; \ - \ - __lock_name = "(wq_completion)"#fmt#args; \ - \ - __alloc_workqueue_key((fmt), (flags), (max_active), \ - &__key, __lock_name, ##args); \ -}) -#else -#define alloc_workqueue(fmt, flags, max_active, args...) \ - __alloc_workqueue_key((fmt), (flags), (max_active), \ - NULL, NULL, ##args) -#endif +struct workqueue_struct *alloc_workqueue (const char *fmt, +unsigned int flags, +int max_active, ...);/** * alloc_ordered_workqueue - allocate an ordered workqueue
kernel/workqueue.c +40 -8
89 lines not on this page · raw message
Signature · 1 line
2.19.1.930.g4563a0d9d0-goog