On Thu, 04 Aug 2016, Daniel Vetter <daniel.vetter@xxxxxxxx> wrote:
> drm_connector_register_all requires a few too many locks because our
> connector_list locking is busted. Add another FIXME+hack to work
> around this. This should address the below lockdep splat:
>
> ======================================================
> [ INFO: possible circular locking dependency detected ]
> 4.7.0-rc5+ #524 Tainted: G O
> -------------------------------------------------------
> kworker/u8:0/6 is trying to acquire lock:
> (&dev->mode_config.mutex){+.+.+.}, at: [<ffffffff815afde0>] drm_modeset_lock_all+0x40/0x120
>
> but task is already holding lock:
> ((fb_notifier_list).rwsem){++++.+}, at: [<ffffffff810ac195>] __blocking_notifier_call_chain+0x35/0x70
>
> which lock already depends on the new lock.
>
> the existing dependency chain (in reverse order) is:
>
> -> #1 ((fb_notifier_list).rwsem){++++.+}:
> [<ffffffff810df611>] lock_acquire+0xb1/0x200
> [<ffffffff819a55b4>] down_write+0x44/0x80
> [<ffffffff810abf91>] blocking_notifier_chain_register+0x21/0xb0
> [<ffffffff814c7448>] fb_register_client+0x18/0x20
> [<ffffffff814c6c86>] backlight_device_register+0x136/0x260
> [<ffffffffa0127eb2>] intel_backlight_device_register+0xa2/0x160 [i915]
> [<ffffffffa00f46be>] intel_connector_register+0xe/0x10 [i915]
> [<ffffffffa0112bfb>] intel_dp_connector_register+0x1b/0x80 [i915]
> [<ffffffff8159dfea>] drm_connector_register+0x4a/0x80
> [<ffffffff8159fe44>] drm_connector_register_all+0x64/0xf0
> [<ffffffff815a2a64>] drm_modeset_register_all+0x174/0x1c0
> [<ffffffff81599b72>] drm_dev_register+0xc2/0xd0
> [<ffffffffa00621d7>] i915_driver_load+0x1547/0x2200 [i915]
> [<ffffffffa006d80f>] i915_pci_probe+0x4f/0x70 [i915]
> [<ffffffff814a2135>] local_pci_probe+0x45/0xa0
> [<ffffffff814a349b>] pci_device_probe+0xdb/0x130
> [<ffffffff815c07e3>] driver_probe_device+0x223/0x440
> [<ffffffff815c0ad5>] __driver_attach+0xd5/0x100
> [<ffffffff815be386>] bus_for_each_dev+0x66/0xa0
> [<ffffffff815c002e>] driver_attach+0x1e/0x20
> [<ffffffff815bf9be>] bus_add_driver+0x1ee/0x280
> [<ffffffff815c1810>] driver_register+0x60/0xe0
> [<ffffffff814a1a10>] __pci_register_driver+0x60/0x70
> [<ffffffffa01a905b>] i915_init+0x5b/0x62 [i915]
> [<ffffffff8100042d>] do_one_initcall+0x3d/0x150
> [<ffffffff811a935b>] do_init_module+0x5f/0x1d9
> [<ffffffff81124416>] load_module+0x20e6/0x27e0
> [<ffffffff81124d63>] SYSC_finit_module+0xc3/0xf0
> [<ffffffff81124dae>] SyS_finit_module+0xe/0x10
> [<ffffffff819a83a9>] entry_SYSCALL_64_fastpath+0x1c/0xac
>
> -> #0 (&dev->mode_config.mutex){+.+.+.}:
> [<ffffffff810df0ac>] __lock_acquire+0x10fc/0x1260
> [<ffffffff810df611>] lock_acquire+0xb1/0x200
> [<ffffffff819a3097>] mutex_lock_nested+0x67/0x3c0
> [<ffffffff815afde0>] drm_modeset_lock_all+0x40/0x120
> [<ffffffff8158f79b>] drm_fb_helper_restore_fbdev_mode_unlocked+0x2b/0x80
> [<ffffffff8158f81d>] drm_fb_helper_set_par+0x2d/0x50
> [<ffffffffa0105f7a>] intel_fbdev_set_par+0x1a/0x60 [i915]
> [<ffffffff814c13c6>] fbcon_init+0x586/0x610
> [<ffffffff8154d16a>] visual_init+0xca/0x130
> [<ffffffff8154e611>] do_bind_con_driver+0x1c1/0x3a0
> [<ffffffff8154eaf6>] do_take_over_console+0x116/0x180
> [<ffffffff814bd3a7>] do_fbcon_takeover+0x57/0xb0
> [<ffffffff814c1e48>] fbcon_event_notify+0x658/0x750
> [<ffffffff810abcae>] notifier_call_chain+0x3e/0xb0
> [<ffffffff810ac1ad>] __blocking_notifier_call_chain+0x4d/0x70
> [<ffffffff810ac1e6>] blocking_notifier_call_chain+0x16/0x20
> [<ffffffff814c748b>] fb_notifier_call_chain+0x1b/0x20
> [<ffffffff814c86b1>] register_framebuffer+0x251/0x330
> [<ffffffff8158fa9f>] drm_fb_helper_initial_config+0x25f/0x3f0
> [<ffffffffa0106b48>] intel_fbdev_initial_config+0x18/0x30 [i915]
> [<ffffffff810adfd8>] async_run_entry_fn+0x48/0x150
> [<ffffffff810a3947>] process_one_work+0x1e7/0x750
> [<ffffffff810a3efb>] worker_thread+0x4b/0x4f0
> [<ffffffff810aad4f>] kthread+0xef/0x110
> [<ffffffff819a85ef>] ret_from_fork+0x1f/0x40
>
> other info that might help us debug this:
>
> Possible unsafe locking scenario:
>
> CPU0 CPU1
> ---- ----
> lock((fb_notifier_list).rwsem);
> lock(&dev->mode_config.mutex);
> lock((fb_notifier_list).rwsem);
> lock(&dev->mode_config.mutex);
>
> *** DEADLOCK ***
>
> 6 locks held by kworker/u8:0/6:
> #0: ("events_unbound"){.+.+.+}, at: [<ffffffff810a38c9>] process_one_work+0x169/0x750
> #1: ((&entry->work)){+.+.+.}, at: [<ffffffff810a38c9>] process_one_work+0x169/0x750
> #2: (registration_lock){+.+.+.}, at: [<ffffffff814c8487>] register_framebuffer+0x27/0x330
> #3: (console_lock){+.+.+.}, at: [<ffffffff814c86ce>] register_framebuffer+0x26e/0x330
> #4: (&fb_info->lock){+.+.+.}, at: [<ffffffff814c78dd>] lock_fb_info+0x1d/0x40
> #5: ((fb_notifier_list).rwsem){++++.+}, at: [<ffffffff810ac195>] __blocking_notifier_call_chain+0x35/0x70
>
> stack backtrace:
> CPU: 2 PID: 6 Comm: kworker/u8:0 Tainted: G O 4.7.0-rc5+ #524
> Hardware name: Intel Corp. Broxton P/NOTEBOOK, BIOS APLKRVPA.X64.0138.B33.1606250842 06/25/2016
> Workqueue: events_unbound async_run_entry_fn
> 0000000000000000 ffff8800758577f0 ffffffff814507a5 ffffffff828b9900
> ffffffff828b9900 ffff880075857830 ffffffff810dc6fa ffff880075857880
> ffff88007584d688 0000000000000005 0000000000000006 ffff88007584d6b0
> Call Trace:
> [<ffffffff814507a5>] dump_stack+0x67/0x92
> [<ffffffff810dc6fa>] print_circular_bug+0x1aa/0x200
> [<ffffffff810df0ac>] __lock_acquire+0x10fc/0x1260
> [<ffffffff810df611>] lock_acquire+0xb1/0x200
> [<ffffffff815afde0>] ? drm_modeset_lock_all+0x40/0x120
> [<ffffffff815afde0>] ? drm_modeset_lock_all+0x40/0x120
> [<ffffffff819a3097>] mutex_lock_nested+0x67/0x3c0
> [<ffffffff815afde0>] ? drm_modeset_lock_all+0x40/0x120
> [<ffffffff810fa85f>] ? rcu_read_lock_sched_held+0x7f/0x90
> [<ffffffff81208218>] ? kmem_cache_alloc_trace+0x248/0x2b0
> [<ffffffff815afdc5>] ? drm_modeset_lock_all+0x25/0x120
> [<ffffffff815afde0>] drm_modeset_lock_all+0x40/0x120
> [<ffffffff8158f79b>] drm_fb_helper_restore_fbdev_mode_unlocked+0x2b/0x80
> [<ffffffff8158f81d>] drm_fb_helper_set_par+0x2d/0x50
> [<ffffffffa0105f7a>] intel_fbdev_set_par+0x1a/0x60 [i915]
> [<ffffffff814c13c6>] fbcon_init+0x586/0x610
> [<ffffffff8154d16a>] visual_init+0xca/0x130
> [<ffffffff8154e611>] do_bind_con_driver+0x1c1/0x3a0
> [<ffffffff8154eaf6>] do_take_over_console+0x116/0x180
> [<ffffffff814bd3a7>] do_fbcon_takeover+0x57/0xb0
> [<ffffffff814c1e48>] fbcon_event_notify+0x658/0x750
> [<ffffffff810abcae>] notifier_call_chain+0x3e/0xb0
> [<ffffffff810ac1ad>] __blocking_notifier_call_chain+0x4d/0x70
> [<ffffffff810ac1e6>] blocking_notifier_call_chain+0x16/0x20
> [<ffffffff814c748b>] fb_notifier_call_chain+0x1b/0x20
> [<ffffffff814c86b1>] register_framebuffer+0x251/0x330
> [<ffffffff815b7e8d>] ? vga_switcheroo_client_fb_set+0x5d/0x70
> [<ffffffff8158fa9f>] drm_fb_helper_initial_config+0x25f/0x3f0
> [<ffffffffa0106b48>] intel_fbdev_initial_config+0x18/0x30 [i915]
> [<ffffffff810adfd8>] async_run_entry_fn+0x48/0x150
> [<ffffffff810a3947>] process_one_work+0x1e7/0x750
> [<ffffffff810a38c9>] ? process_one_work+0x169/0x750
> [<ffffffff810a3efb>] worker_thread+0x4b/0x4f0
> [<ffffffff810a3eb0>] ? process_one_work+0x750/0x750
> [<ffffffff810aad4f>] kthread+0xef/0x110
> [<ffffffff819a85ef>] ret_from_fork+0x1f/0x40
> [<ffffffff810aac60>] ? kthread_stop+0x2e0/0x2e0
>
> Cc: Imre Deak <imre.deak@xxxxxxxxx>
> Cc: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx>
> Signed-off-by: Daniel Vetter <daniel.vetter@xxxxxxxxx>
Fixes: ?
Cc: drm-intel-fixes@xxxxxxxxxxxxxxxxxxxxx
Cc: a bunch of people who reported the problem upstream?
BR,
Jani.
> ---
> drivers/gpu/drm/drm_connector.c | 8 +++-----
> 1 file changed, 3 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/gpu/drm/drm_connector.c b/drivers/gpu/drm/drm_connector.c
> index ef921fa09a84..d9104d8b3c6b 100644
> --- a/drivers/gpu/drm/drm_connector.c
> +++ b/drivers/gpu/drm/drm_connector.c
> @@ -401,16 +401,14 @@ int drm_connector_register_all(struct drm_device *dev)
> struct drm_connector *connector;
> int ret;
>
> - mutex_lock(&dev->mode_config.mutex);
> -
> - drm_for_each_connector(connector, dev) {
> + /* FIXME: taking the mode config mutex ends up in a clash with
> + * fbcon/backlight registration */
> + list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
> ret = drm_connector_register(connector);
> if (ret)
> goto err;
> }
>
> - mutex_unlock(&dev->mode_config.mutex);
> -
> return 0;
>
> err:
--
Jani Nikula, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
