WREG32(sdma_v4_0_get_reg_offset(i, mmSDMA0_GFX_RB_BASE_HI), ring->gpu_addr >> 40);
ring->wptr = 0;
- WREG32(sdma_v4_0_get_reg_offset(i, mmSDMA0_GFX_RB_WPTR), lower_32_bits(ring->wptr) << 2);
- WREG32(sdma_v4_0_get_reg_offset(i, mmSDMA0_GFX_RB_WPTR_HI), upper_32_bits(ring->wptr) << 2);
+
+ /* before programing wptr to a less value, need set minor_ptr_update first */
+ WREG32(sdma_v4_0_get_reg_offset(i, mmSDMA0_GFX_MINOR_PTR_UPDATE), 1);
+
+ if (!amdgpu_sriov_vf(adev)) { /* only bare-metal use register write for wptr */
+ WREG32(sdma_v4_0_get_reg_offset(i, mmSDMA0_GFX_RB_WPTR), lower_32_bits(ring->wptr) << 2);
+ WREG32(sdma_v4_0_get_reg_offset(i, mmSDMA0_GFX_RB_WPTR_HI), upper_32_bits(ring->wptr) << 2);
+ }
doorbell = RREG32(sdma_v4_0_get_reg_offset(i, mmSDMA0_GFX_DOORBELL));
doorbell_offset = RREG32(sdma_v4_0_get_reg_offset(i, mmSDMA0_GFX_DOORBELL_OFFSET));
WREG32(sdma_v4_0_get_reg_offset(i, mmSDMA0_GFX_DOORBELL_OFFSET), doorbell_offset);
nbio_v6_1_sdma_doorbell_range(adev, i, ring->use_doorbell, ring->doorbell_index);
+ if (amdgpu_sriov_vf(adev))
+ sdma_v4_0_ring_set_wptr(ring);
+
+ /* set minor_ptr_update to 0 after wptr programed */
+ WREG32(sdma_v4_0_get_reg_offset(i, mmSDMA0_GFX_MINOR_PTR_UPDATE), 0);
+
/* set utc l1 enable flag always to 1 */
temp = RREG32(sdma_v4_0_get_reg_offset(i, mmSDMA0_CNTL));
temp = REG_SET_FIELD(temp, SDMA0_CNTL, UTC_L1_ENABLE, 1);
WREG32(sdma_v4_0_get_reg_offset(i, mmSDMA0_CNTL), temp);
- /* unhalt engine */
- temp = RREG32(sdma_v4_0_get_reg_offset(i, mmSDMA0_F32_CNTL));
- temp = REG_SET_FIELD(temp, SDMA0_F32_CNTL, HALT, 0);
- WREG32(sdma_v4_0_get_reg_offset(i, mmSDMA0_F32_CNTL), temp);
+ if (!amdgpu_sriov_vf(adev)) {
+ /* unhalt engine */
+ temp = RREG32(sdma_v4_0_get_reg_offset(i, mmSDMA0_F32_CNTL));
+ temp = REG_SET_FIELD(temp, SDMA0_F32_CNTL, HALT, 0);
+ WREG32(sdma_v4_0_get_reg_offset(i, mmSDMA0_F32_CNTL), temp);
+ }
/* enable DMA RB */
rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_ENABLE, 1);
ring->ready = true;
+ if (amdgpu_sriov_vf(adev)) { /* bare-metal sequence doesn't need below to lines */
+ sdma_v4_0_ctx_switch_enable(adev, true);
+ sdma_v4_0_enable(adev, true);
+ }
+
r = amdgpu_ring_test_ring(ring);
if (r) {
ring->ready = false;
(adev->sdma.instance[i].fw->data +
le32_to_cpu(hdr->header.ucode_array_offset_bytes));
- sdma_v4_0_print_ucode_regs(adev);
-
WREG32(sdma_v4_0_get_reg_offset(i, mmSDMA0_UCODE_ADDR), 0);
*/
static int sdma_v4_0_start(struct amdgpu_device *adev)
{
- int r;
+ int r = 0;
if (amdgpu_sriov_vf(adev)) {
- /* disable RB and halt engine */
+ sdma_v4_0_ctx_switch_enable(adev, false);
sdma_v4_0_enable(adev, false);
/* set RB registers */