Lockdep warning on boot with 2.6.31-rc5-rt1.1
Peter,
I'm getting this warning from lockdep when booting on my T60.
The two addresses reported (0xffffffff812664a2 and 0xffffffff812664ae) actually bracket one call to mutex_lock() in driver_attach() so I'm not sure what the complaint is.
Clark
=============================================
[ INFO: possible recursive locking detected ]
2.6.31-rc5-rt1.1 #37
---------------------------------------------
swapper/1 is trying to acquire lock:
(&dev->mutex){+.+...}, at: [<ffffffff812664ae>]__driver_attach+0x48/0x81
but task is already holding lock:
(&dev->mutex){+.+...}, at: [<ffffffff812664a2>]__driver_attach+0x3c/0x81
other info that might help us debug this:
1 lock held by swapper/1:
#0: (&dev->mutex){+.+...}, at: [<ffffffff812664a2>]
__driver_attach+0x3c/0x81 stack backtrace: Pid: 1, comm: swapper Not tainted 2.6.31-rc5-rt1.1 #37 Call Trace: [<ffffffff81071e38>] __lock_acquire+0x14ae/0x153c [<ffffffff812664ae>] ? __driver_attach+0x48/0x81 [<ffffffff81071fc3>] lock_acquire+0xfd/0x129 [<ffffffff812664ae>] ? __driver_attach+0x48/0x81 [<ffffffff8137b667>] _mutex_lock+0x31/0x40 [<ffffffff812664ae>] ? __driver_attach+0x48/0x81 [<ffffffff812664ae>] __driver_attach+0x48/0x81 [<ffffffff81266466>] ? __driver_attach+0x0/0x81 [<ffffffff81265b9c>] bus_for_each_dev+0x59/0x8e [<ffffffff81266245>] driver_attach+0x1e/0x20 [<ffffffff81265465>] bus_add_driver+0x13f/0x288 [<ffffffff812667b3>] driver_register+0x9d/0x10e [<ffffffff81207e26>] acpi_bus_register_driver+0x43/0x45 [<ffffffff815f5fdf>] acpi_ec_init+0x37/0x55 [<ffffffff815f5d47>] acpi_init+0x224/0x265 [<ffffffff815f5b23>] ? acpi_init+0x0/0x265 [<ffffffff81009080>] do_one_initcall+0x75/0x18a [<ffffffff8106dd0a>] ? put_lock_stats+0xe/0x27 [<ffffffff8137b3d1>] ? rt_spin_unlock+0x23/0x6d [<ffffffff8106dcc8>] ? get_lock_stats+0x16/0x4a [<ffffffff8106dcc8>] ? get_lock_stats+0x16/0x4a [<ffffffff81145616>] ? proc_register+0x18c/0x1a2 [<ffffffff8137e6fe>] ? sub_preempt_count+0x35/0x49 [<ffffffff8106dd0a>] ? put_lock_stats+0xe/0x27 [<ffffffff8106de1c>] ? lock_release_holdtime+0xf9/0xfe [<ffffffff8137b3d1>] ? rt_spin_unlock+0x23/0x6d [<ffffffff81145616>] ? proc_register+0x18c/0x1a2 [<ffffffff8114574a>] ? create_proc_entry+0x79/0x91 [<ffffffff8109c069>] ? register_irq_proc+0xb3/0xcf [<ffffffff81140000>] ? pde_users_dec+0x20/0x45 [<ffffffff815d370a>] kernel_init+0x17e/0x28b [<ffffffff8100d15a>] child_rip+0xa/0x20 [<ffffffff8100ca94>] ? restore_args+0x0/0x30 [<ffffffff815d358c>] ? kernel_init+0x0/0x28b [<ffffffff8100d150>] ? child_rip+0x0/0x20 ---------------------------
| preempt count: 00000000 ]
| 0-level deep critical section nesting:
----------------------------------------
Quoted 18 lines · Clark Williams
Peter,
I'm getting this warning from lockdep when booting on my T60.
The two addresses reported (
0xffffffff812664a2and0xffffffff812664ae) actually bracket one call tomutex_lock()indriver_attach()so I'm not sure what the complaint is.Clark
=============================================
[ INFO: possible recursive locking detected ]
2.6.31-rc5-rt1.1 #37
---------------------------------------------
swapper/1 is trying to acquire lock:
(&dev->mutex){+.+...}, at: [<ffffffff812664ae>]__driver_attach+0x48/0x81
Quoted 8 lines
but task is already holding lock:
(&dev->mutex){+.+...}, at: [<ffffffff812664a2>]__driver_attach+0x3c/0x81other info that might help us debug this:
1 lock held by swapper/1:
#0: (&dev->mutex){+.+...}, at: [<ffffffff812664a2>]__driver_attach+0x3c/0x81
Oh, that's tglx who's gone wild with sem->mutex conversions.
It used to be that _all_ dev->sem instances were taken on suspend or something like that, I think that got fixed a long while back.
I'd have to look at what the current locking requirements for dev->sem are.
I remember talking to Alan on several occasions about this, and I just went over some of the old emails, but I must say the precise requirements stay hidden from me. Also, I'm not sure these emails are still representative of the current state.
Quoted 10 lines · Peter Zijlstra
Quoted 7 lines · Clark Williams
Peter,
I'm getting this warning from lockdep when booting on my T60.
The two addresses reported (
0xffffffff812664a2and0xffffffff812664ae) actually bracket one call tomutex_lock()indriver_attach()so I'm not sure what the complaint is.Oh, that's tglx who's gone wild with
sem->mutexconversions.
Is this code available somewhere?
Quoted 5 lines
It used to be that _all_
dev->seminstances were taken on suspend or something like that, I think that got fixed a long while back.I'd have to look at what the current locking requirements for
dev->semare.
It is supposed to be locked whenever the driver core invokes a probe, remove, or PM-related callback. Under some circumstances, the parent's semaphore is supposed to be locked as well. Individual subsystems may have their own requirements in addition to these.
The ordering requirement is: Don't try to acquire a device's lock if you already hold the lock for a non-ancestor device. More generally (if more obscurely): If you already hold device A's lock, then don't try to acquire the lock for device B unless you already hold the lock for A & B's most recent common ancestor.
Quoted 4 lines
I remember talking to Alan on several occasions about this, and I just went over some of the old emails, but I must say the precise requirements stay hidden from me. Also, I'm not sure these emails are still representative of the current state.
I think they are, pretty much. The real problem, of course, is that lockdep doesn't understand tree-structured lock orderings. Hence it isn't practical to convert dev->sem into a mutex.
Alan Stern
Quoted 7 lines · Alan Stern
Quoted 5 lines · Peter Zijlstra
Quoted 4 lines · Clark Williams
Peter,
I'm getting this warning from lockdep when booting on my T60.
Quoted 7 lines
Quoted 5 lines
Quoted 3 lines
The two addresses reported (
0xffffffff812664a2and0xffffffff812664ae) actually bracket one call tomutex_lock()indriver_attach()so I'm not sure what the complaint is.Oh, that's tglx who's gone wild with
sem->mutexconversions.Is this code available somewhere?
Its in the -rt tree, but this patch was posted to lkml at:
http://lkml.org/lkml/2009/7/26/36
The -rt tree can be found in various places, but while tglx is out celebrating his holidays the latest can be found through:
http://lkml.org/lkml/2009/8/5/406
Quoted 17 lines
Quoted 5 lines
It used to be that _all_
dev->seminstances were taken on suspend or something like that, I think that got fixed a long while back.I'd have to look at what the current locking requirements for
dev->semare.It is supposed to be locked whenever the driver core invokes a probe, remove, or PM-related callback. Under some circumstances, the parent's semaphore is supposed to be locked as well. Individual subsystems may have their own requirements in addition to these.
The ordering requirement is: Don't try to acquire a device's lock if you already hold the lock for a non-ancestor device. More generally (if more obscurely): If you already hold device A's lock, then don't try to acquire the lock for device B unless you already hold the lock for A & B's most recent common ancestor.
Quoted 8 lines
Quoted 4 lines
I remember talking to Alan on several occasions about this, and I just went over some of the old emails, but I must say the precise requirements stay hidden from me. Also, I'm not sure these emails are still representative of the current state.
I think they are, pretty much. The real problem, of course, is that lockdep doesn't understand tree-structured lock orderings. Hence it isn't practical to convert
dev->seminto a mutex.
Right, well it would if we'd make every instance a class, but since classes should reside in static storage this is far from trivial.
If we'd be able to find a mapping such that we can use a limited number of these classes to represent the needed structure then we're good.
I think I proposed adding a class to each driver or something, but then you countered that a single driver could register itself at conflicting places in the device tree.
Still it might be worth to try that and see where we'll end up and possibly fix up a few drivers to be more intelligent.
/me ponders
Nested busses would be interesting though, suppose we assign a class to a USB bus driver, and we chain USB hubs, you'd get a nesting of similar classes and that'd upset lockdep again :/
The other proposal was creating a fixed list of classes and register each device at a class corresponding to its depth in the tree. I can't remember what was wrong with that, but I seem to have been persuaded that that was hard too.
Quoted 7 lines · Peter Zijlstra
Quoted 3 lines
I think they are, pretty much. The real problem, of course, is that lockdep doesn't understand tree-structured lock orderings. Hence it isn't practical to convert
dev->seminto a mutex.Right, well it would if we'd make every instance a class, but since classes should reside in static storage this is far from trivial.
Quoted 6 lines
If we'd be able to find a mapping such that we can use a limited number of these classes to represent the needed structure then we're good.
I think I proposed adding a class to each driver or something, but then you countered that a single driver could register itself at conflicting places in the device tree.
Also there can be devices in the tree that don't have any driver. And there can be devices that have the same driver as their children.
Quoted 8 lines
Still it might be worth to try that and see where we'll end up and possibly fix up a few drivers to be more intelligent.
/me ponders
Nested busses would be interesting though, suppose we assign a class to a USB bus driver, and we chain USB hubs, you'd get a nesting of similar classes and that'd upset lockdep again :/
Yep.
Quoted 4 lines
The other proposal was creating a fixed list of classes and register each device at a class corresponding to its depth in the tree. I can't remember what was wrong with that, but I seem to have been persuaded that that was hard too.
It probably would work for the most part. However a possible scenario involves first locking a parent and then locking all its children. (I don't know if this ever happens anywhere, but it might.) This can't cause a deadlock but it would run into trouble with depth-based classes.
Alan Stern
Quoted 7 lines · Alan Stern
Quoted 4 lines · Peter Zijlstra
The other proposal was creating a fixed list of classes and register each device at a class corresponding to its depth in the tree. I can't remember what was wrong with that, but I seem to have been persuaded that that was hard too.
Quoted 5 lines
It probably would work for the most part. However a possible scenario involves first locking a parent and then locking all its children. (I don't know if this ever happens anywhere, but it might.) This can't cause a deadlock but it would run into trouble with depth-based classes.
If you know which parent is locked, we can solve that with mutex_lock_nest_lock() [ doesn't currently exist, but is analogous to spin_lock_nest_lock() ] and together with
http://lkml.org/lkml/2009/7/23/222 that would allow you to lock up to 2048 children.
Would something like that work?
Quoted 14 lines · Peter Zijlstra
Quoted 7 lines · Alan Stern
Quoted 4 lines · Peter Zijlstra
The other proposal was creating a fixed list of classes and register each device at a class corresponding to its depth in the tree. I can't remember what was wrong with that, but I seem to have been persuaded that that was hard too.
Quoted 5 lines
It probably would work for the most part. However a possible scenario involves first locking a parent and then locking all its children. (I don't know if this ever happens anywhere, but it might.) This can't cause a deadlock but it would run into trouble with depth-based classes.
Quoted 5 lines
If you know which parent is locked, we can solve that with
mutex_lock_nest_lock()[ doesn't currently exist, but is analogous tospin_lock_nest_lock()] and together with
http://lkml.org/lkml/2009/7/23/222 that would allow you to lock up to 2048 children.
Not only do I know not which parent is locked, I don't even know if this ever happens anywhere at all! My point was purely theoretical.
Quoted 1 line
Would something like that work?
Perhaps -- I don't understand what spin_lock_nest_lock() is supposed to do.
Alan Stern
Quoted 33 lines · Peter Zijlstra
Quoted 18 lines · Clark Williams
Peter,
I'm getting this warning from lockdep when booting on my T60.
The two addresses reported (
0xffffffff812664a2and0xffffffff812664ae) actually bracket one call tomutex_lock()indriver_attach()so I'm not sure what the complaint is.Clark
=============================================
[ INFO: possible recursive locking detected ]
2.6.31-rc5-rt1.1 #37
---------------------------------------------
swapper/1 is trying to acquire lock:
(&dev->mutex){+.+...}, at: [<ffffffff812664ae>]__driver_attach+0x48/0x81Quoted 8 lines
but task is already holding lock:
(&dev->mutex){+.+...}, at: [<ffffffff812664a2>]__driver_attach+0x3c/0x81other info that might help us debug this:
1 lock held by swapper/1:
#0: (&dev->mutex){+.+...}, at: [<ffffffff812664a2>]__driver_attach+0x3c/0x81Oh, that's tglx who's gone wild with
sem->mutexconversions.It used to be that _all_
dev->seminstances were taken on suspend or something like that, I think that got fixed a long while back.
Quoted 7 lines
I'd have to look at what the current locking requirements for
dev->semare.I remember talking to Alan on several occasions about this, and I just went over some of the old emails, but I must say the precise requirements stay hidden from me. Also, I'm not sure these emails are still representative of the current state.
I think you means this thread:
http://lkml.org/lkml/2008/4/17/305
Quoted 7 lines
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Signature · 6 lines
Regards dave -- To unsubscribe from this list: send the line "unsubscribe linux-rt-users" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
2009/8/7 Peter Zijlstra <[email protected]>:
Quoted 23 lines · Clark Williams
Quoted 22 lines
Peter,
I'm getting this warning from lockdep when booting on my T60.
The two addresses reported (
0xffffffff812664a2and0xffffffff812664ae) actually bracket one call tomutex_lock()indriver_attach()so I'm not sure what the complaint is.Clark
=============================================
[ INFO: possible recursive locking detected ]
2.6.31-rc5-rt1.1 #37
---------------------------------------------
swapper/1 is trying to acquire lock:
(&dev->mutex){+.+...}, at: [<ffffffff812664ae>]__driver_attach+0x48/0x81but task is already holding lock:
(&dev->mutex){+.+...}, at: [<ffffffff812664a2>]__driver_attach+0x3c/0x81
Quoted 6 lines
Quoted 4 lines
other info that might help us debug this:
1 lock held by swapper/1:
#0: (&dev->mutex){+.+...}, at: [<ffffffff812664a2>]__driver_attach+0x3c/0x81Oh, that's tglx who's gone wild with
sem->mutexconversions.
Maybe we can introduce some mutex interfaces which bypass lockdep validation temporarily to allow driver core to convert to mutex from sema if lockdep can't validate tree-structured lock orderings.
Thanks.
--
Lei Ming
--
To unsubscribe from this list: send the line "unsubscribe linux-rt-users" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
2009/8/8 Alan Stern <[email protected]>:
Quoted 13 lines · Peter Zijlstra
Quoted 6 lines
It used to be that _all_
dev->seminstances were taken on suspend or something like that, I think that got fixed a long while back.I'd have to look at what the current locking requirements for
dev->semare.It is supposed to be locked whenever the driver core invokes a probe, remove, or PM-related callback. Under some circumstances, the parent's semaphore is supposed to be locked as well. Individual subsystems may have their own requirements in addition to these.
Quoted 6 lines
The ordering requirement is: Don't try to acquire a device's lock if you already hold the lock for a non-ancestor device. More generally (if more obscurely): If you already hold device A's lock, then don't try to acquire the lock for device B unless you already hold the lock for A & B's most recent common ancestor.
It seems that the following case is very common, and A and B have no common ancestor, but we can hold device A and B's lock at the same time, can't we?
Thanks.
device A comes in one bus:
device_add()
->bus_attach_device()
->device_attach():drivers/base/dd.c /*holding device A's lock*/
->...drv->probe() /*sleep here some time*/then device B comes in another bus:
device_add()
->bus_attach_device()
->device_attach():drivers/base/dd.c /*holding device B's lock*/
->...drv->probe() /*sleep here some time*/Signature · 1 line
Lei Ming
Quoted 7 lines · Clark Williams
Peter,
I'm getting this warning from lockdep when booting on my T60.
The two addresses reported (
0xffffffff812664a2and0xffffffff812664ae) actually bracket one call tomutex_lock()indriver_attach()so I'm not sure what the complaint is.
I'm getting a different lockdep warning when booting on my T400 using 2.6.31-rc5; not sure if it's related or not....
In any case, it's screwing up the ability for lockdep to find any other problems.
[ 0.297775] INFO: trying to register non-static key. [ 0.297775] the code is fine but needs lockdep annotation. [ 0.297775] turning off the locking correctness validator. [ 0.297775] Pid: 1, comm: swapper Not tainted 2.6.31-rc5-00256-gf124845 #4 [ 0.297775] Call Trace: [ 0.297775] [<c0511a2f>] ? printk+0x14/0x1d [ 0.297775] [<c016de1f>] register_lock_class+0x5a/0x2a1 [ 0.297775] [<c016e932>] ? mark_lock+0x1e/0x1e4 [ 0.297775] [<c016f673>] __lock_acquire+0x9c/0xb1e [ 0.297775] [<c01be910>] ? mod_zone_page_state+0x9f/0xaf [ 0.297775] [<c016ed98>] ? trace_hardirqs_on_caller+0x103/0x124 [ 0.297775] [<c016e932>] ? mark_lock+0x1e/0x1e4 [ 0.297775] [<c016eb3b>] ? mark_held_locks+0x43/0x5b [ 0.297775] [<c01d490f>] ? kmem_cache_alloc+0xaf/0x127 [ 0.297775] [<c016ed98>] ? trace_hardirqs_on_caller+0x103/0x124 [ 0.297775] [<c0170189>] lock_acquire+0x94/0xb7 [ 0.297775] [<c045d6bd>] ? alloc_netdev_mq+0x105/0x1cc [ 0.297775] [<c0513f8d>] _spin_lock_bh+0x28/0x58 [ 0.297775] [<c045d6bd>] ? alloc_netdev_mq+0x105/0x1cc [ 0.297775] [<c045d6bd>] alloc_netdev_mq+0x105/0x1cc [ 0.297775] [<c03f7bbf>] ? loopback_setup+0x0/0x79 [ 0.297775] [<c03f7c6f>] loopback_net_init+0x25/0x68 [ 0.297782] [<c0457317>] register_pernet_operations+0x2f/0xa1 [ 0.297832] [<c0512fcd>] ? mutex_lock_nested+0x33/0x3b [ 0.297891] [<c0457435>] register_pernet_device+0x24/0x4c [ 0.297951] [<c0796a09>] net_dev_init+0x101/0x150 [ 0.298010] [<c0796908>] ? net_dev_init+0x0/0x150 [ 0.298069] [<c010115c>] do_one_initcall+0x6a/0x177 [ 0.298127] [<c016e932>] ? mark_lock+0x1e/0x1e4 [ 0.298185] [<c016e932>] ? mark_lock+0x1e/0x1e4 [ 0.298244] [<c01b371a>] ? get_page_from_freelist+0x28f/0x3be [ 0.298304] [<c016ed98>] ? trace_hardirqs_on_caller+0x103/0x124 [ 0.298364] [<c016edc4>] ? trace_hardirqs_on+0xb/0xd [ 0.298423] [<c016e932>] ? mark_lock+0x1e/0x1e4 [ 0.298482] [<c016e932>] ? mark_lock+0x1e/0x1e4 [ 0.298540] [<c016edc4>] ? trace_hardirqs_on+0xb/0xd [ 0.298600] [<c03104ad>] ? ida_get_new_above+0x157/0x171 [ 0.298660] [<c0213468>] ? proc_register+0x14b/0x15c [ 0.298719] [<c011e5f6>] ? sched_clock+0x8/0xb [ 0.298777] [<c016d9b1>] ? lock_release_holdtime+0x30/0x131 [ 0.298837] [<c0213468>] ? proc_register+0x14b/0x15c [ 0.298896] [<c0513dfb>] ? _spin_unlock+0x22/0x25 [ 0.298954] [<c0213468>] ? proc_register+0x14b/0x15c [ 0.299013] [<c021359b>] ? create_proc_entry+0x80/0x96 [ 0.299073] [<c0191064>] ? register_irq_proc+0x91/0xad [ 0.299132] [<c01910d8>] ? init_irq_proc+0x58/0x65 [ 0.299191] [<c0768301>] kernel_init+0x131/0x182 [ 0.299249] [<c07681d0>] ? kernel_init+0x0/0x182
- Ted
Quoted 55 lines · Theodore Tso
Quoted 7 lines · Clark Williams
Peter,
I'm getting this warning from lockdep when booting on my T60.
The two addresses reported (
0xffffffff812664a2and0xffffffff812664ae) actually bracket one call tomutex_lock()indriver_attach()so I'm not sure what the complaint is.I'm getting a different lockdep warning when booting on my T400 using 2.6.31-rc5; not sure if it's related or not....
In any case, it's screwing up the ability for lockdep to find any other problems.
[ 0.297775] INFO: trying to register non-static key.
[ 0.297775] the code is fine but needs lockdep annotation.
[ 0.297775] turning off the locking correctness validator.
[ 0.297775] Pid: 1, comm: swapper Not tainted 2.6.31-rc5-00256-gf124845 #4 [ 0.297775] Call Trace:
[ 0.297775] [<c0511a2f>] ?printk+0x14/0x1d
[ 0.297775] [<c016de1f>]register_lock_class+0x5a/0x2a1
[ 0.297775] [<c016e932>] ?mark_lock+0x1e/0x1e4
[ 0.297775] [<c016f673>]__lock_acquire+0x9c/0xb1e
[ 0.297775] [<c01be910>] ?mod_zone_page_state+0x9f/0xaf
[ 0.297775] [<c016ed98>] ?trace_hardirqs_on_caller+0x103/0x124[ 0.297775] [<c016e932>] ?mark_lock+0x1e/0x1e4
[ 0.297775] [<c016eb3b>] ?mark_held_locks+0x43/0x5b
[ 0.297775] [<c01d490f>] ?kmem_cache_alloc+0xaf/0x127
[ 0.297775] [<c016ed98>] ?trace_hardirqs_on_caller+0x103/0x124[ 0.297775] [<c0170189>]lock_acquire+0x94/0xb7
[ 0.297775] [<c045d6bd>] ?alloc_netdev_mq+0x105/0x1cc
[ 0.297775] [<c0513f8d>]_spin_lock_bh+0x28/0x58
[ 0.297775] [<c045d6bd>] ?alloc_netdev_mq+0x105/0x1cc
[ 0.297775] [<c045d6bd>]alloc_netdev_mq+0x105/0x1cc
[ 0.297775] [<c03f7bbf>] ?loopback_setup+0x0/0x79
[ 0.297775] [<c03f7c6f>]loopback_net_init+0x25/0x68
[ 0.297782] [<c0457317>]register_pernet_operations+0x2f/0xa1[ 0.297832] [<c0512fcd>] ?mutex_lock_nested+0x33/0x3b
[ 0.297891] [<c0457435>]register_pernet_device+0x24/0x4c
[ 0.297951] [<c0796a09>]net_dev_init+0x101/0x150
[ 0.298010] [<c0796908>] ?net_dev_init+0x0/0x150
[ 0.298069] [<c010115c>]do_one_initcall+0x6a/0x177
[ 0.298127] [<c016e932>] ?mark_lock+0x1e/0x1e4
[ 0.298185] [<c016e932>] ?mark_lock+0x1e/0x1e4
[ 0.298244] [<c01b371a>] ?get_page_from_freelist+0x28f/0x3be[ 0.298304] [<c016ed98>] ?trace_hardirqs_on_caller+0x103/0x124[ 0.298364] [<c016edc4>] ?trace_hardirqs_on+0xb/0xd
[ 0.298423] [<c016e932>] ?mark_lock+0x1e/0x1e4
[ 0.298482] [<c016e932>] ?mark_lock+0x1e/0x1e4
[ 0.298540] [<c016edc4>] ?trace_hardirqs_on+0xb/0xd
[ 0.298600] [<c03104ad>] ?ida_get_new_above+0x157/0x171
[ 0.298660] [<c0213468>] ?proc_register+0x14b/0x15c
[ 0.298719] [<c011e5f6>] ?sched_clock+0x8/0xb
[ 0.298777] [<c016d9b1>] ?lock_release_holdtime+0x30/0x131
Quoted 9 lines
[ 0.298837] [<c0213468>] ?
proc_register+0x14b/0x15c[ 0.298896] [<c0513dfb>] ?_spin_unlock+0x22/0x25[ 0.298954] [<c0213468>] ?proc_register+0x14b/0x15c[ 0.299013] [<c021359b>] ?create_proc_entry+0x80/0x96[ 0.299073] [<c0191064>] ?register_irq_proc+0x91/0xad[ 0.299132] [<c01910d8>] ?init_irq_proc+0x58/0x65[ 0.299191] [<c0768301>]kernel_init+0x131/0x182[ 0.299249] [<c07681d0>] ?kernel_init+0x0/0x182
It's a different problem, for this issue please see:
http://lkml.org/lkml/2009/8/5/49
http://lkml.org/lkml/2009/8/5/51
Signature · 6 lines
Regards dave -- To unsubscribe from this list: send the line "unsubscribe linux-rt-users" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
Quoted 11 lines · Ming Lei
Quoted 6 lines
The ordering requirement is: Don't try to acquire a device's lock if you already hold the lock for a non-ancestor device. More generally (if more obscurely): If you already hold device A's lock, then don't try to acquire the lock for device B unless you already hold the lock for A & B's most recent common ancestor.
It seems that the following case is very common, and A and B have no common ancestor, but we can hold device A and B's lock at the same time, can't we?
Quoted 7 lines
Thanks.
device A comes in one bus:
device_add() ->bus_attach_device() ->device_attach():drivers/base/dd.c /*holding device A's lock*/ ->...drv->probe() /*sleep here some time*/
So right now thread 1 is sleeping.
Quoted 1 line
then device B comes in another bus:
So all this must happen in a different thread, thread 2:
Quoted 4 lines
device_add()
->bus_attach_device()
->device_attach():drivers/base/dd.c /*holding device B's lock*/
->...drv->probe() /*sleep here some time*/At this point, thread 1 holds A's lock and thread 2 holds B's lock. Neither thread holds both locks.
Alan Stern