lockdep warning in pcmcia_socket_dev_suspend

[Jonathan Nieder]

Hi pcmcia maintainers,

In recent kernels, I started getting a lockdep warning in the kernel
log on suspend.  I see no other symptoms (except that my wireless card
suspends and resumes properly now --- thanks!), but this message is
perfectly reproducible.  I am pretty sure it wasn’t there in v2.6.31.

Grasping at straws, I tried reverting commit 0c570cdeb8 (PM / yenta:
Fix cardbus suspend/resume regression).  That does make the message
go away (and the wireless card suspend problems come back).

This log fragment is for a kernel built from Linus’s commit
v2.6.32-rc5-11-g2fdc246, plus three small local patches to the
Makefile.

Perhaps this comment from drivers/base/power/main.c is relevant:

 * Since device_pm_add() may be called with a device semaphore held,
 * we must never try to acquire a device semaphore while holding
 * dpm_list_mutex.

But I am way out of my depth here.

Thoughts?
Jonathan

PM: Syncing filesystems ... done.
Freezing user space processes ... (elapsed 0.00 seconds) done.
Freezing remaining freezable tasks ... (elapsed 0.00 seconds) done.
Suspending console(s) (use no_console_suspend to debug)
ath5k 0000:02:00.0: PCI INT A disabled
sd 0:0:0:0: [sda] Synchronizing SCSI cache
sd 0:0:0:0: [sda] Stopping disk
parport_pc 00:0f: disabled
Maestro3 0000:00:08.0: PCI INT A disabled
ACPI handle has no context!
uhci_hcd 0000:00:07.2: PCI INT D disabled

=======================================================
[ INFO: possible circular locking dependency detected ]
2.6.32-rc5-00014-g68c786f #13
-------------------------------------------------------
s2ram/2242 is trying to acquire lock:
 (&socket->skt_mutex){+.+.+.}, at: [<ccaa2ce7>] pcmcia_socket_dev_suspend+0x3f/0x8e [pcmcia_core]

but task is already holding lock:
 (pcmcia_socket_list_rwsem){++++..}, at: [<ccaa2cc8>] pcmcia_socket_dev_suspend+0x20/0x8e [pcmcia_core]

which lock already depends on the new lock.


the existing dependency chain (in reverse order) is:

-> #2 (pcmcia_socket_list_rwsem){++++..}:
       [<c015a99a>] __lock_acquire+0xa02/0xb71
       [<c015abbb>] lock_acquire+0xb2/0xdc
       [<c0396dfc>] down_read+0x29/0x4d
       [<ccaa2cc8>] pcmcia_socket_dev_suspend+0x20/0x8e [pcmcia_core]
       [<ccba8951>] yenta_dev_suspend_noirq+0x29/0x9d [yenta_socket]
       [<c02757b1>] pci_pm_suspend_noirq+0x66/0x10b
       [<c02e0319>] pm_noirq_op+0x54/0xdf
       [<c02e0bdd>] dpm_suspend_noirq+0x46/0xb9
       [<c0166155>] suspend_devices_and_enter+0x84/0x149
       [<c01662ca>] enter_state+0xb0/0xd9
       [<c0165a39>] state_store+0xa1/0xc7
       [<c0260f77>] kobj_attr_store+0x27/0x45
       [<c0211a43>] sysfs_write_file+0xd3/0x10f
       [<c01c802e>] vfs_write+0x98/0x106
       [<c01c8176>] sys_write+0x50/0x84
       [<c0102d8b>] sysenter_do_call+0x12/0x32

-> #1 (dpm_list_mtx){+.+...}:
       [<c015a99a>] __lock_acquire+0xa02/0xb71
       [<c015abbb>] lock_acquire+0xb2/0xdc
       [<c039679b>] __mutex_lock_common+0x4a/0x387
       [<c0396b61>] mutex_lock_nested+0x21/0x35
       [<c02e0ccc>] device_pm_add+0x22/0xac
       [<c02d99b8>] device_add+0x2be/0x43a
       [<c026fa25>] pci_bus_add_device+0x1d/0x50
       [<c026fa86>] pci_bus_add_devices+0x2e/0xfc
       [<ccaa6159>] cb_alloc+0xaf/0xd4 [pcmcia_core]
       [<ccaa2874>] socket_insert+0xbc/0xf5 [pcmcia_core]
       [<ccaa31b3>] pccardd+0x177/0x214 [pcmcia_core]
       [<c0147e67>] kthread+0x69/0x6e
       [<c010372f>] kernel_thread_helper+0x7/0x58

-> #0 (&socket->skt_mutex){+.+.+.}:
       [<c015a89d>] __lock_acquire+0x905/0xb71
       [<c015abbb>] lock_acquire+0xb2/0xdc
       [<c039679b>] __mutex_lock_common+0x4a/0x387
       [<c0396b61>] mutex_lock_nested+0x21/0x35
       [<ccaa2ce7>] pcmcia_socket_dev_suspend+0x3f/0x8e [pcmcia_core]
       [<ccba8951>] yenta_dev_suspend_noirq+0x29/0x9d [yenta_socket]
       [<c02757b1>] pci_pm_suspend_noirq+0x66/0x10b
       [<c02e0319>] pm_noirq_op+0x54/0xdf
       [<c02e0bdd>] dpm_suspend_noirq+0x46/0xb9
       [<c0166155>] suspend_devices_and_enter+0x84/0x149
       [<c01662ca>] enter_state+0xb0/0xd9
       [<c0165a39>] state_store+0xa1/0xc7
       [<c0260f77>] kobj_attr_store+0x27/0x45
       [<c0211a43>] sysfs_write_file+0xd3/0x10f
       [<c01c802e>] vfs_write+0x98/0x106
       [<c01c8176>] sys_write+0x50/0x84
       [<c0102d8b>] sysenter_do_call+0x12/0x32

other info that might help us debug this:

4 locks held by s2ram/2242:
 #0:  (&buffer->mutex){+.+.+.}, at: [<c02119a8>] sysfs_write_file+0x38/0x10f
 #1:  (pm_mutex){+.+.+.}, at: [<c016624b>] enter_state+0x31/0xd9
 #2:  (dpm_list_mtx){+.+...}, at: [<c02e0bbe>] dpm_suspend_noirq+0x27/0xb9
 #3:  (pcmcia_socket_list_rwsem){++++..}, at: [<ccaa2cc8>] pcmcia_socket_dev_suspend+0x20/0x8e [pcmcia_core]

stack backtrace:
Pid: 2242, comm: s2ram Not tainted 2.6.32-rc5-00014-g68c786f #13
Call Trace:
 [<c0395658>] ? printk+0x1d/0x35
 [<c0159c14>] print_circular_bug+0x9b/0xb8
 [<c015a89d>] __lock_acquire+0x905/0xb71
 [<ccaa2ce7>] ? pcmcia_socket_dev_suspend+0x3f/0x8e [pcmcia_core]
 [<c015abbb>] lock_acquire+0xb2/0xdc
 [<ccaa2ce7>] ? pcmcia_socket_dev_suspend+0x3f/0x8e [pcmcia_core]
 [<c039679b>] __mutex_lock_common+0x4a/0x387
 [<ccaa2ce7>] ? pcmcia_socket_dev_suspend+0x3f/0x8e [pcmcia_core]
 [<ccaa2ce7>] ? pcmcia_socket_dev_suspend+0x3f/0x8e [pcmcia_core]
 [<c0396b61>] mutex_lock_nested+0x21/0x35
 [<ccaa2ce7>] ? pcmcia_socket_dev_suspend+0x3f/0x8e [pcmcia_core]
 [<ccaa2ce7>] pcmcia_socket_dev_suspend+0x3f/0x8e [pcmcia_core]
 [<ccba8951>] yenta_dev_suspend_noirq+0x29/0x9d [yenta_socket]
 [<c02757b1>] pci_pm_suspend_noirq+0x66/0x10b
 [<c02e0319>] pm_noirq_op+0x54/0xdf
 [<c02e0bdd>] dpm_suspend_noirq+0x46/0xb9
 [<c0166155>] suspend_devices_and_enter+0x84/0x149
 [<c01662ca>] enter_state+0xb0/0xd9
 [<c0165a39>] state_store+0xa1/0xc7
 [<c0165998>] ? state_store+0x0/0xc7
 [<c0260f77>] kobj_attr_store+0x27/0x45
 [<c0211a43>] sysfs_write_file+0xd3/0x10f
 [<c0211970>] ? sysfs_write_file+0x0/0x10f
 [<c01c802e>] vfs_write+0x98/0x106
 [<c01c8176>] sys_write+0x50/0x84
 [<c0102d8b>] sysenter_do_call+0x12/0x32
ACPI: Preparing to enter system sleep state S3