drm/mtrr: possible deadlock

[Dou Liyang]

Hi Dmitry,

At 08/15/2017 03:18 PM, Dmitry Vyukov wrote:
> Hello,
>
> I am getting the following deadlock report while booting linux-next on
> 91dfed74eabcdae9378131546c446442c29bf769 in qemu. Config is attached.
>
>
> WARNING: possible recursive locking detected
> 4.13.0-rc4-next-20170811 #2 Not tainted
> --------------------------------------------
> swapper/0/1 is trying to acquire lock:
>  (cpu_hotplug_lock.rw_sem){++++}, at: [<ffffffff816a19fc>]
> stop_machine+0x1c/0x40 kernel/stop_machine.c:596
>
> but task is already holding lock:
>  (cpu_hotplug_lock.rw_sem){++++}, at: [<ffffffff812aad6d>]
> get_online_cpus include/linux/cpu.h:126 [inline]
>  (cpu_hotplug_lock.rw_sem){++++}, at: [<ffffffff812aad6d>]
> mtrr_add_page+0x1bd/0xe80 arch/x86/kernel/cpu/mtrr/main.c:328
>

I guess you may want this.

https://lkml.org/lkml/2017/8/14/7

and the solution by Thomas

https://lkml.org/lkml/2017/8/14/64

Thanks,
	dou.

> other info that might help us debug this:
>  Possible unsafe locking scenario:
>
>        CPU0
>        ----
>   lock(cpu_hotplug_lock.rw_sem);
>   lock(cpu_hotplug_lock.rw_sem);
>
>  *** DEADLOCK ***
>
>  May be due to missing lock nesting notation
>
> 5 locks held by swapper/0/1:
>  #0:  (&dev->mutex){....}, at: [<ffffffff834aa073>] device_lock
> include/linux/device.h:1081 [inline]
>  #0:  (&dev->mutex){....}, at: [<ffffffff834aa073>]
> __driver_attach+0xd3/0x1c0 drivers/base/dd.c:788
>  #1:  (&dev->mutex){....}, at: [<ffffffff834aa086>] device_lock
> include/linux/device.h:1081 [inline]
>  #1:  (&dev->mutex){....}, at: [<ffffffff834aa086>]
> __driver_attach+0xe6/0x1c0 drivers/base/dd.c:789
>  #2:  (drm_global_mutex){+.+.}, at: [<ffffffff8285e40d>]
> drm_dev_register+0x4d/0x660 drivers/gpu/drm/drm_drv.c:780
>  #3:  (cpu_hotplug_lock.rw_sem){++++}, at: [<ffffffff812aad6d>]
> get_online_cpus include/linux/cpu.h:126 [inline]
>  #3:  (cpu_hotplug_lock.rw_sem){++++}, at: [<ffffffff812aad6d>]
> mtrr_add_page+0x1bd/0xe80 arch/x86/kernel/cpu/mtrr/main.c:328
>  #4:  (mtrr_mutex){+.+.}, at: [<ffffffff812aad7b>]
> mtrr_add_page+0x1cb/0xe80 arch/x86/kernel/cpu/mtrr/main.c:331
>
> stack backtrace:
> CPU: 0 PID: 1 Comm: swapper/0 Not tainted 4.13.0-rc4-next-20170811 #2
> Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
> Call Trace:
>  __dump_stack lib/dump_stack.c:16 [inline]
>  dump_stack+0x194/0x257 lib/dump_stack.c:52
>  print_deadlock_bug kernel/locking/lockdep.c:1797 [inline]
>  check_deadlock kernel/locking/lockdep.c:1844 [inline]
>  validate_chain kernel/locking/lockdep.c:2453 [inline]
>  __lock_acquire+0xed5/0x3bc0 kernel/locking/lockdep.c:3498
>  lock_acquire+0x1d5/0x580 kernel/locking/lockdep.c:4002
>  percpu_down_read_preempt_disable include/linux/percpu-rwsem.h:35 [inline]
>  percpu_down_read include/linux/percpu-rwsem.h:58 [inline]
>  cpus_read_lock+0x42/0x90 kernel/cpu.c:218
>  stop_machine+0x1c/0x40 kernel/stop_machine.c:596
>  set_mtrr arch/x86/kernel/cpu/mtrr/main.c:237 [inline]
>  mtrr_add_page+0x6b7/0xe80 arch/x86/kernel/cpu/mtrr/main.c:373
>  mtrr_add+0x97/0xb0 arch/x86/kernel/cpu/mtrr/main.c:448
>  arch_phys_wc_add+0x55/0x90 arch/x86/kernel/cpu/mtrr/main.c:562
>  cirrus_mm_init+0x43f/0x6a0 drivers/gpu/drm/cirrus/cirrus_ttm.c:274
>  cirrus_driver_load+0xcd/0x250 drivers/gpu/drm/cirrus/cirrus_main.c:182
>  drm_dev_register+0x34f/0x660 drivers/gpu/drm/drm_drv.c:801
>  drm_get_pci_dev+0x1ee/0x600 drivers/gpu/drm/drm_pci.c:262
>  cirrus_pci_probe+0x198/0x210 drivers/gpu/drm/cirrus/cirrus_drv.c:75
>  local_pci_probe+0xdc/0x190 drivers/pci/pci-driver.c:307
>  __pci_device_probe drivers/pci/pci-driver.c:361 [inline]
>  pci_device_probe+0x5a3/0x6a0 drivers/pci/pci-driver.c:426
>  really_probe drivers/base/dd.c:413 [inline]
>  driver_probe_device+0x63c/0xa20 drivers/base/dd.c:557
>  __driver_attach+0x181/0x1c0 drivers/base/dd.c:791
>  bus_for_each_dev+0x154/0x1e0 drivers/base/bus.c:313
>  driver_attach+0x3d/0x50 drivers/base/dd.c:810
>  bus_add_driver+0x48f/0x660 drivers/base/bus.c:669
>  driver_register+0x1bf/0x3c0 drivers/base/driver.c:168
>  __pci_register_driver+0x1d2/0x2c0 drivers/pci/pci-driver.c:1313
>  cirrus_init+0x52/0x5e drivers/gpu/drm/cirrus/cirrus_drv.c:168
>  do_one_initcall+0x9e/0x330 init/main.c:824
>  do_initcall_level init/main.c:890 [inline]
>  do_initcalls init/main.c:898 [inline]
>  do_basic_setup init/main.c:916 [inline]
>  kernel_init_freeable+0x46e/0x526 init/main.c:1066
>  kernel_init+0x13/0x172 init/main.c:991
>  ret_from_fork+0x2a/0x40 arch/x86/entry/entry_64.S:431
>
>