drm/amdgpu: clean wptr on wb when gpu recovery
authorYintian Tao <yttao@amd.com>
Fri, 28 Feb 2020 06:24:42 +0000 (14:24 +0800)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Wed, 25 Mar 2020 07:25:49 +0000 (08:25 +0100)
[ Upstream commit 2ab7e274b86739f4ceed5d94b6879f2d07b2802f ]

The TDR will be randomly failed due to compute ring
test failure. If the compute ring wptr & 0x7ff(ring_buf_mask)
is 0x100 then after map mqd the compute ring rptr will be
synced with 0x100. And the ring test packet size is also 0x100.
Then after invocation of amdgpu_ring_commit, the cp will not
really handle the packet on the ring buffer because rptr is equal to wptr.

Signed-off-by: Yintian Tao <yttao@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Monk Liu <Monk.Liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c

index 2816d03..14417ce 100644 (file)
@@ -3555,6 +3555,7 @@ static int gfx_v10_0_kcq_init_queue(struct amdgpu_ring *ring)
 
                /* reset ring buffer */
                ring->wptr = 0;
+               atomic64_set((atomic64_t *)&adev->wb.wb[ring->wptr_offs], 0);
                amdgpu_ring_clear_ring(ring);
        } else {
                amdgpu_ring_clear_ring(ring);
index d85e1e5..40034ef 100644 (file)
@@ -3756,6 +3756,7 @@ static int gfx_v9_0_kcq_init_queue(struct amdgpu_ring *ring)
 
                /* reset ring buffer */
                ring->wptr = 0;
+               atomic64_set((atomic64_t *)&adev->wb.wb[ring->wptr_offs], 0);
                amdgpu_ring_clear_ring(ring);
        } else {
                amdgpu_ring_clear_ring(ring);