drm/amdgpu/gfx8: reset compute ring wptr on the GPU on resume

Commit 42cdf6f687da ("drm/amdgpu/gfx8: always restore kcq MQDs") made the
ring pointer always to be reset on resume from suspend. This caused compute
rings to fail since the reset was done without also resetting it for the
firmware. Reset wptr on the GPU to avoid a disconnect between the driver
and firmware wptr.

Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/3911
Fixes: 42cdf6f687da ("drm/amdgpu/gfx8: always restore kcq MQDs")
Signed-off-by: Eeli Haapalainen <eeli.haapalainen@protonmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit 2becafc319db3d96205320f31cc0de4ee5a93747)
Cc: stable@vger.kernel.org

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
index 5ee2237d8ee8f11a51a9a02c97237dcf26eca2fc..bc983ecf3d99a9fb5e5a355aa76c37bb092bedf0 100644 (file)
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
@@ -4640,6 +4640,7 @@ static int gfx_v8_0_kcq_init_queue(struct amdgpu_ring *ring)
                        memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(struct vi_mqd_allocation));
                /* reset ring buffer */
                ring->wptr = 0;
+               atomic64_set((atomic64_t *)ring->wptr_cpu_addr, 0);
                amdgpu_ring_clear_ring(ring);
        }
        return 0;