From: Ivan Lipski <[email protected]> [Why&How] An interrupt racing DM teardown can reach amdgpu_dm_irq_schedule_work() after the IRQ workqueue is destroyed, passing a NULL wq to queue_work() and crashing in __queue_work(). Bail out early when adev->dm.irq_wq is NULL so a registered handler is simply not queued.
Reviewed-by: Alex Hung <[email protected]> Signed-off-by: Ivan Lipski <[email protected]> --- drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c index 9be63996b0627..560ff697ff950 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c @@ -618,6 +618,9 @@ STATIC_IFN_KUNIT void amdgpu_dm_irq_schedule_work(struct amdgpu_device *adev, DM_IRQ_TABLE_LOCK(adev, irq_table_flags); + if (!adev->dm.irq_wq) + goto out_unlock; + if (list_empty(handler_list)) goto out_unlock; -- 2.43.0
