0 waits forever. We don't actually use the HW semaphore anymore, but if someone uses the packet, set a time out value so we eventually time out and avoid a potential queue or GPU reset.
Signed-off-by: Alex Deucher <[email protected]> --- drivers/gpu/drm/amd/amdgpu/si_dma.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/amdgpu/si_dma.c b/drivers/gpu/drm/amd/amdgpu/si_dma.c index 3e58feb2d5e4f..7cf9dfbf11196 100644 --- a/drivers/gpu/drm/amd/amdgpu/si_dma.c +++ b/drivers/gpu/drm/amd/amdgpu/si_dma.c @@ -149,7 +149,7 @@ static int si_dma_start(struct amdgpu_device *adev) ring = &adev->sdma.instance[i].ring; WREG32(mmDMA_SEM_INCOMPLETE_TIMER_CNTL + sdma_offsets[i], 0); - WREG32(mmDMA_SEM_WAIT_FAIL_TIMER_CNTL + sdma_offsets[i], 0); + WREG32(mmDMA_SEM_WAIT_FAIL_TIMER_CNTL + sdma_offsets[i], 10); /* Set ring buffer size in dwords */ rb_bufsz = order_base_2(ring->ring_size / 4); -- 2.53.0
