Re: [PATCH 112/207] drm/amdgpu: fix SDMA hdp flush engine conflict

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Am 01.06.20 um 20:19 schrieb Alex Deucher:
From: Likun Gao <Likun.Gao@xxxxxxx>

Each of HDP flush engine should be used by one ring, correct allocate of
hdp flush engine to SDMA ring.
Correct me value of each SDMA ring, as it was cleared when init microcode.

Signed-off-by: Likun Gao <Likun.Gao@xxxxxxx>
Signed-off-by: Alex Deucher <alexander.deucher@xxxxxxx>

Reviewed-by: Christian König <christian.koenig@xxxxxxx>

---
  drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c | 6 ++----
  1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c
index 52206050adb9..f072cef28b60 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c
@@ -391,10 +391,7 @@ static void sdma_v5_2_ring_emit_hdp_flush(struct amdgpu_ring *ring)
  	u32 ref_and_mask = 0;
  	const struct nbio_hdp_flush_reg *nbio_hf_reg = adev->nbio.hdp_flush_reg;
- if (ring->me == 0)
-		ref_and_mask = nbio_hf_reg->ref_and_mask_sdma0;
-	else
-		ref_and_mask = nbio_hf_reg->ref_and_mask_sdma1;
+	ref_and_mask = nbio_hf_reg->ref_and_mask_sdma0 << ring->me;
amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_POLL_REGMEM) |
  			  SDMA_PKT_POLL_REGMEM_HEADER_HDP_FLUSH(1) |
@@ -1224,6 +1221,7 @@ static int sdma_v5_2_sw_init(void *handle)
  		ring = &adev->sdma.instance[i].ring;
  		ring->ring_obj = NULL;
  		ring->use_doorbell = true;
+		ring->me = i;
DRM_INFO("use_doorbell being set to: [%s]\n",
  				ring->use_doorbell?"true":"false");

_______________________________________________
amd-gfx mailing list
amd-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/amd-gfx




[Index of Archives]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux