]> git.ipfire.org Git - thirdparty/kernel/stable.git/commit
drm/amdkfd: KFD release_work possible circular locking
authorPhilip Yang <Philip.Yang@amd.com>
Tue, 18 Feb 2025 01:08:29 +0000 (20:08 -0500)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Wed, 4 Jun 2025 12:40:10 +0000 (14:40 +0200)
commitc8cc14eeb231b6153d407f4f10dbb1bf802cb820
tree710b6ddd9299650b72981f245ef63bce465768f1
parent1964a698cb6bd5b6566ac56c50f7450fb36ee58c
drm/amdkfd: KFD release_work possible circular locking

[ Upstream commit 1b9366c601039d60546794c63fbb83ce8e53b978 ]

If waiting for gpu reset done in KFD release_work, thers is WARNING:
possible circular locking dependency detected

  #2  kfd_create_process
        kfd_process_mutex
          flush kfd release work

  #1  kfd release work
        wait for amdgpu reset work

  #0  amdgpu_device_gpu_reset
        kgd2kfd_pre_reset
          kfd_process_mutex

  Possible unsafe locking scenario:

        CPU0                    CPU1
        ----                    ----
   lock((work_completion)(&p->release_work));
                  lock((wq_completion)kfd_process_wq);
                  lock((work_completion)(&p->release_work));
   lock((wq_completion)amdgpu-reset-dev);

To fix this, KFD create process move flush release work outside
kfd_process_mutex.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
drivers/gpu/drm/amd/amdkfd/kfd_process.c