diff options
Diffstat (limited to 'drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c')
| -rw-r--r-- | drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c | 723 |
1 files changed, 488 insertions, 235 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c index 42f4bd250def..a6e612b4a892 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c @@ -60,10 +60,61 @@ MODULE_FIRMWARE("amdgpu/sdma_5_2_7.bin"); #define SDMA0_HYP_DEC_REG_END 0x5893 #define SDMA1_HYP_DEC_REG_OFFSET 0x20 +static const struct amdgpu_hwip_reg_entry sdma_reg_list_5_2[] = { + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_STATUS_REG), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_STATUS1_REG), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_STATUS2_REG), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_STATUS3_REG), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_UCODE_CHECKSUM), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RB_RPTR_FETCH_HI), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RB_RPTR_FETCH), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_UTCL1_RD_STATUS), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_UTCL1_WR_STATUS), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_UTCL1_RD_XNACK0), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_UTCL1_RD_XNACK1), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_UTCL1_WR_XNACK0), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_UTCL1_WR_XNACK1), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_RB_CNTL), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_RB_RPTR), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_RB_RPTR_HI), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_RB_WPTR), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_RB_WPTR_HI), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_IB_OFFSET), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_IB_BASE_LO), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_IB_BASE_HI), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_IB_CNTL), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_IB_RPTR), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_IB_SUB_REMAIN), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_GFX_DUMMY_REG), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_RB_CNTL), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_RB_RPTR), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_RB_RPTR_HI), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_RB_WPTR), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_RB_WPTR_HI), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_IB_OFFSET), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_IB_BASE_LO), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_IB_BASE_HI), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_PAGE_DUMMY_REG), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_RB_CNTL), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_RB_RPTR), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_RB_RPTR_HI), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_RB_WPTR), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_RB_WPTR_HI), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_IB_OFFSET), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_IB_BASE_LO), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_IB_BASE_HI), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_RLC0_DUMMY_REG), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_INT_STATUS), + SOC15_REG_ENTRY_STR(GC, 0, mmSDMA0_VM_CNTL), + SOC15_REG_ENTRY_STR(GC, 0, mmGRBM_STATUS2) +}; + static void sdma_v5_2_set_ring_funcs(struct amdgpu_device *adev); static void sdma_v5_2_set_buffer_funcs(struct amdgpu_device *adev); static void sdma_v5_2_set_vm_pte_funcs(struct amdgpu_device *adev); static void sdma_v5_2_set_irq_funcs(struct amdgpu_device *adev); +static int sdma_v5_2_stop_queue(struct amdgpu_ring *ring); +static int sdma_v5_2_restore_queue(struct amdgpu_ring *ring); static u32 sdma_v5_2_get_reg_offset(struct amdgpu_device *adev, u32 instance, u32 internal_offset) { @@ -176,6 +227,16 @@ static void sdma_v5_2_ring_set_wptr(struct amdgpu_ring *ring) DRM_DEBUG("calling WDOORBELL64(0x%08x, 0x%016llx)\n", ring->doorbell_index, ring->wptr << 2); WDOORBELL64(ring->doorbell_index, ring->wptr << 2); + if (amdgpu_ip_version(adev, SDMA0_HWIP, 0) == IP_VERSION(5, 2, 1)) { + /* SDMA seems to miss doorbells sometimes when powergating kicks in. + * Updating the wptr directly will wake it. This is only safe because + * we disallow gfxoff in begin_use() and then allow it again in end_use(). + */ + WREG32(sdma_v5_2_get_reg_offset(adev, ring->me, mmSDMA0_GFX_RB_WPTR), + lower_32_bits(ring->wptr << 2)); + WREG32(sdma_v5_2_get_reg_offset(adev, ring->me, mmSDMA0_GFX_RB_WPTR_HI), + upper_32_bits(ring->wptr << 2)); + } } else { DRM_DEBUG("Not using doorbell -- " "mmSDMA%i_GFX_RB_WPTR == 0x%08x " @@ -280,17 +341,21 @@ 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; - 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) | - SDMA_PKT_POLL_REGMEM_HEADER_FUNC(3)); /* == */ - amdgpu_ring_write(ring, (adev->nbio.funcs->get_hdp_flush_done_offset(adev)) << 2); - amdgpu_ring_write(ring, (adev->nbio.funcs->get_hdp_flush_req_offset(adev)) << 2); - amdgpu_ring_write(ring, ref_and_mask); /* reference */ - amdgpu_ring_write(ring, ref_and_mask); /* mask */ - amdgpu_ring_write(ring, SDMA_PKT_POLL_REGMEM_DW5_RETRY_COUNT(0xfff) | - SDMA_PKT_POLL_REGMEM_DW5_INTERVAL(10)); /* retry count, poll interval */ + if (ring->me > 1) { + amdgpu_asic_flush_hdp(adev, ring); + } else { + 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) | + SDMA_PKT_POLL_REGMEM_HEADER_FUNC(3)); /* == */ + amdgpu_ring_write(ring, (adev->nbio.funcs->get_hdp_flush_done_offset(adev)) << 2); + amdgpu_ring_write(ring, (adev->nbio.funcs->get_hdp_flush_req_offset(adev)) << 2); + amdgpu_ring_write(ring, ref_and_mask); /* reference */ + amdgpu_ring_write(ring, ref_and_mask); /* mask */ + amdgpu_ring_write(ring, SDMA_PKT_POLL_REGMEM_DW5_RETRY_COUNT(0xfff) | + SDMA_PKT_POLL_REGMEM_DW5_INTERVAL(10)); /* retry count, poll interval */ + } } /** @@ -331,11 +396,9 @@ static void sdma_v5_2_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 se } if ((flags & AMDGPU_FENCE_FLAG_INT)) { - uint32_t ctx = ring->is_mes_queue ? - (ring->hw_queue_id | AMDGPU_FENCE_MES_QUEUE_FLAG) : 0; /* generate an interrupt */ amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_TRAP)); - amdgpu_ring_write(ring, SDMA_PKT_TRAP_INT_CONTEXT_INT_CONTEXT(ctx)); + amdgpu_ring_write(ring, SDMA_PKT_TRAP_INT_CONTEXT_INT_CONTEXT(0)); } } @@ -344,15 +407,15 @@ static void sdma_v5_2_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 se * sdma_v5_2_gfx_stop - stop the gfx async dma engines * * @adev: amdgpu_device pointer - * + * @inst_mask: mask of dma engine instances to be disabled * Stop the gfx async dma ring buffers. */ -static void sdma_v5_2_gfx_stop(struct amdgpu_device *adev) +static void sdma_v5_2_gfx_stop(struct amdgpu_device *adev, uint32_t inst_mask) { u32 rb_cntl, ib_cntl; int i; - for (i = 0; i < adev->sdma.num_instances; i++) { + for_each_inst(i, inst_mask) { rb_cntl = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL)); rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_ENABLE, 0); WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL), rb_cntl); @@ -443,9 +506,11 @@ static void sdma_v5_2_enable(struct amdgpu_device *adev, bool enable) { u32 f32_cntl; int i; + uint32_t inst_mask; + inst_mask = GENMASK(adev->sdma.num_instances - 1, 0); if (!enable) { - sdma_v5_2_gfx_stop(adev); + sdma_v5_2_gfx_stop(adev, inst_mask); sdma_v5_2_rlc_stop(adev); } @@ -459,14 +524,17 @@ static void sdma_v5_2_enable(struct amdgpu_device *adev, bool enable) } /** - * sdma_v5_2_gfx_resume - setup and start the async dma engines + * sdma_v5_2_gfx_resume_instance - start/restart a certain sdma engine * * @adev: amdgpu_device pointer + * @i: instance + * @restore: used to restore wptr when restart * - * Set up the gfx DMA ring buffers and enable them. - * Returns 0 for success, error for failure. + * Set up the gfx DMA ring buffers and enable them. On restart, we will restore wptr and rptr. + * Return 0 for success. */ -static int sdma_v5_2_gfx_resume(struct amdgpu_device *adev) + +static int sdma_v5_2_gfx_resume_instance(struct amdgpu_device *adev, int i, bool restore) { struct amdgpu_ring *ring; u32 rb_cntl, ib_cntl; @@ -476,139 +544,161 @@ static int sdma_v5_2_gfx_resume(struct amdgpu_device *adev) u32 temp; u32 wptr_poll_cntl; u64 wptr_gpu_addr; - int i, r; - for (i = 0; i < adev->sdma.num_instances; i++) { - ring = &adev->sdma.instance[i].ring; + ring = &adev->sdma.instance[i].ring; - if (!amdgpu_sriov_vf(adev)) - WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_SEM_WAIT_FAIL_TIMER_CNTL), 0); + if (!amdgpu_sriov_vf(adev)) + WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_SEM_WAIT_FAIL_TIMER_CNTL), 0); - /* Set ring buffer size in dwords */ - rb_bufsz = order_base_2(ring->ring_size / 4); - rb_cntl = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL)); - rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_SIZE, rb_bufsz); + /* Set ring buffer size in dwords */ + rb_bufsz = order_base_2(ring->ring_size / 4); + rb_cntl = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL)); + rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_SIZE, rb_bufsz); #ifdef __BIG_ENDIAN - rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_SWAP_ENABLE, 1); - rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, - RPTR_WRITEBACK_SWAP_ENABLE, 1); + rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_SWAP_ENABLE, 1); + rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, + RPTR_WRITEBACK_SWAP_ENABLE, 1); #endif - WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL), rb_cntl); - - /* Initialize the ring buffer's read and write pointers */ + WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL), rb_cntl); + + /* Initialize the ring buffer's read and write pointers */ + if (restore) { + WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_RPTR), lower_32_bits(ring->wptr << 2)); + WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_RPTR_HI), upper_32_bits(ring->wptr << 2)); + WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR), lower_32_bits(ring->wptr << 2)); + WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_HI), upper_32_bits(ring->wptr << 2)); + } else { WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_RPTR), 0); WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_RPTR_HI), 0); WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR), 0); WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_HI), 0); + } - /* setup the wptr shadow polling */ - wptr_gpu_addr = ring->wptr_gpu_addr; - WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_POLL_ADDR_LO), - lower_32_bits(wptr_gpu_addr)); - WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_POLL_ADDR_HI), - upper_32_bits(wptr_gpu_addr)); - wptr_poll_cntl = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, - mmSDMA0_GFX_RB_WPTR_POLL_CNTL)); - wptr_poll_cntl = REG_SET_FIELD(wptr_poll_cntl, - SDMA0_GFX_RB_WPTR_POLL_CNTL, - F32_POLL_ENABLE, 1); - WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_POLL_CNTL), - wptr_poll_cntl); - - /* set the wb address whether it's enabled or not */ - WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_RPTR_ADDR_HI), - upper_32_bits(ring->rptr_gpu_addr) & 0xFFFFFFFF); - WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_RPTR_ADDR_LO), - lower_32_bits(ring->rptr_gpu_addr) & 0xFFFFFFFC); - - rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RPTR_WRITEBACK_ENABLE, 1); - - WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_BASE), ring->gpu_addr >> 8); - WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_BASE_HI), ring->gpu_addr >> 40); - + /* setup the wptr shadow polling */ + wptr_gpu_addr = ring->wptr_gpu_addr; + WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_POLL_ADDR_LO), + lower_32_bits(wptr_gpu_addr)); + WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_POLL_ADDR_HI), + upper_32_bits(wptr_gpu_addr)); + wptr_poll_cntl = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, + mmSDMA0_GFX_RB_WPTR_POLL_CNTL)); + wptr_poll_cntl = REG_SET_FIELD(wptr_poll_cntl, + SDMA0_GFX_RB_WPTR_POLL_CNTL, + F32_POLL_ENABLE, 1); + WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_POLL_CNTL), + wptr_poll_cntl); + + /* set the wb address whether it's enabled or not */ + WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_RPTR_ADDR_HI), + upper_32_bits(ring->rptr_gpu_addr) & 0xFFFFFFFF); + WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_RPTR_ADDR_LO), + lower_32_bits(ring->rptr_gpu_addr) & 0xFFFFFFFC); + + rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RPTR_WRITEBACK_ENABLE, 1); + + WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_BASE), ring->gpu_addr >> 8); + WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_BASE_HI), ring->gpu_addr >> 40); + + if (!restore) ring->wptr = 0; - /* before programing wptr to a less value, need set minor_ptr_update first */ - WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_MINOR_PTR_UPDATE), 1); + /* before programing wptr to a less value, need set minor_ptr_update first */ + WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_MINOR_PTR_UPDATE), 1); - if (!amdgpu_sriov_vf(adev)) { /* only bare-metal use register write for wptr */ - WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR), lower_32_bits(ring->wptr << 2)); - WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_HI), upper_32_bits(ring->wptr << 2)); - } + if (!amdgpu_sriov_vf(adev)) { /* only bare-metal use register write for wptr */ + WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR), lower_32_bits(ring->wptr << 2)); + WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_WPTR_HI), upper_32_bits(ring->wptr << 2)); + } - doorbell = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_DOORBELL)); - doorbell_offset = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_DOORBELL_OFFSET)); + doorbell = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_DOORBELL)); + doorbell_offset = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_DOORBELL_OFFSET)); - if (ring->use_doorbell) { - doorbell = REG_SET_FIELD(doorbell, SDMA0_GFX_DOORBELL, ENABLE, 1); - doorbell_offset = REG_SET_FIELD(doorbell_offset, SDMA0_GFX_DOORBELL_OFFSET, - OFFSET, ring->doorbell_index); - } else { - doorbell = REG_SET_FIELD(doorbell, SDMA0_GFX_DOORBELL, ENABLE, 0); - } - WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_DOORBELL), doorbell); - WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_DOORBELL_OFFSET), doorbell_offset); + if (ring->use_doorbell) { + doorbell = REG_SET_FIELD(doorbell, SDMA0_GFX_DOORBELL, ENABLE, 1); + doorbell_offset = REG_SET_FIELD(doorbell_offset, SDMA0_GFX_DOORBELL_OFFSET, + OFFSET, ring->doorbell_index); + } else { + doorbell = REG_SET_FIELD(doorbell, SDMA0_GFX_DOORBELL, ENABLE, 0); + } + WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_DOORBELL), doorbell); + WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_DOORBELL_OFFSET), doorbell_offset); - adev->nbio.funcs->sdma_doorbell_range(adev, i, ring->use_doorbell, - ring->doorbell_index, - adev->doorbell_index.sdma_doorbell_range); + adev->nbio.funcs->sdma_doorbell_range(adev, i, ring->use_doorbell, + ring->doorbell_index, + adev->doorbell_index.sdma_doorbell_range); - if (amdgpu_sriov_vf(adev)) - sdma_v5_2_ring_set_wptr(ring); + if (amdgpu_sriov_vf(adev)) + sdma_v5_2_ring_set_wptr(ring); - /* set minor_ptr_update to 0 after wptr programed */ + /* set minor_ptr_update to 0 after wptr programed */ - WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_MINOR_PTR_UPDATE), 0); + WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_MINOR_PTR_UPDATE), 0); - /* SRIOV VF has no control of any of registers below */ - if (!amdgpu_sriov_vf(adev)) { - /* set utc l1 enable flag always to 1 */ - temp = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_CNTL)); - temp = REG_SET_FIELD(temp, SDMA0_CNTL, UTC_L1_ENABLE, 1); - - /* enable MCBP */ - temp = REG_SET_FIELD(temp, SDMA0_CNTL, MIDCMD_PREEMPT_ENABLE, 1); - WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_CNTL), temp); - - /* Set up RESP_MODE to non-copy addresses */ - temp = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_UTCL1_CNTL)); - temp = REG_SET_FIELD(temp, SDMA0_UTCL1_CNTL, RESP_MODE, 3); - temp = REG_SET_FIELD(temp, SDMA0_UTCL1_CNTL, REDO_DELAY, 9); - WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_UTCL1_CNTL), temp); - - /* program default cache read and write policy */ - temp = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_UTCL1_PAGE)); - /* clean read policy and write policy bits */ - temp &= 0xFF0FFF; - temp |= ((CACHE_READ_POLICY_L2__DEFAULT << 12) | - (CACHE_WRITE_POLICY_L2__DEFAULT << 14) | - SDMA0_UTCL1_PAGE__LLC_NOALLOC_MASK); - WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_UTCL1_PAGE), temp); - - /* unhalt engine */ - temp = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_F32_CNTL)); - temp = REG_SET_FIELD(temp, SDMA0_F32_CNTL, HALT, 0); - WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_F32_CNTL), temp); - } + /* SRIOV VF has no control of any of registers below */ + if (!amdgpu_sriov_vf(adev)) { + /* set utc l1 enable flag always to 1 */ + temp = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_CNTL)); + temp = REG_SET_FIELD(temp, SDMA0_CNTL, UTC_L1_ENABLE, 1); + + /* enable MCBP */ + temp = REG_SET_FIELD(temp, SDMA0_CNTL, MIDCMD_PREEMPT_ENABLE, 1); + WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_CNTL), temp); + + /* Set up RESP_MODE to non-copy addresses */ + temp = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_UTCL1_CNTL)); + temp = REG_SET_FIELD(temp, SDMA0_UTCL1_CNTL, RESP_MODE, 3); + temp = REG_SET_FIELD(temp, SDMA0_UTCL1_CNTL, REDO_DELAY, 9); + WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_UTCL1_CNTL), temp); + + /* program default cache read and write policy */ + temp = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_UTCL1_PAGE)); + /* clean read policy and write policy bits */ + temp &= 0xFF0FFF; + temp |= ((CACHE_READ_POLICY_L2__DEFAULT << 12) | + (CACHE_WRITE_POLICY_L2__DEFAULT << 14) | + SDMA0_UTCL1_PAGE__LLC_NOALLOC_MASK); + WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_UTCL1_PAGE), temp); + + /* unhalt engine */ + temp = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_F32_CNTL)); + temp = REG_SET_FIELD(temp, SDMA0_F32_CNTL, HALT, 0); + WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_F32_CNTL), temp); + } - /* enable DMA RB */ - rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_ENABLE, 1); - WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL), rb_cntl); + /* enable DMA RB */ + rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_GFX_RB_CNTL, RB_ENABLE, 1); + WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_RB_CNTL), rb_cntl); - ib_cntl = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_IB_CNTL)); - ib_cntl = REG_SET_FIELD(ib_cntl, SDMA0_GFX_IB_CNTL, IB_ENABLE, 1); + ib_cntl = RREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_IB_CNTL)); + ib_cntl = REG_SET_FIELD(ib_cntl, SDMA0_GFX_IB_CNTL, IB_ENABLE, 1); #ifdef __BIG_ENDIAN - ib_cntl = REG_SET_FIELD(ib_cntl, SDMA0_GFX_IB_CNTL, IB_SWAP_ENABLE, 1); + ib_cntl = REG_SET_FIELD(ib_cntl, SDMA0_GFX_IB_CNTL, IB_SWAP_ENABLE, 1); #endif - /* enable DMA IBs */ - WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_IB_CNTL), ib_cntl); + /* enable DMA IBs */ + WREG32_SOC15_IP(GC, sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_GFX_IB_CNTL), ib_cntl); - if (amdgpu_sriov_vf(adev)) { /* bare-metal sequence doesn't need below to lines */ - sdma_v5_2_ctx_switch_enable(adev, true); - sdma_v5_2_enable(adev, true); - } + if (amdgpu_sriov_vf(adev)) { /* bare-metal sequence doesn't need below to lines */ + sdma_v5_2_ctx_switch_enable(adev, true); + sdma_v5_2_enable(adev, true); + } + + return amdgpu_ring_test_helper(ring); +} - r = amdgpu_ring_test_helper(ring); +/** + * sdma_v5_2_gfx_resume - setup and start the async dma engines + * + * @adev: amdgpu_device pointer + * + * Set up the gfx DMA ring buffers and enable them. + * Returns 0 for success, error for failure. + */ +static int sdma_v5_2_gfx_resume(struct amdgpu_device *adev) +{ + int i, r; + + for (i = 0; i < adev->sdma.num_instances; i++) { + r = sdma_v5_2_gfx_resume_instance(adev, i, false); if (r) return r; } @@ -673,37 +763,49 @@ static int sdma_v5_2_load_microcode(struct amdgpu_device *adev) return 0; } -static int sdma_v5_2_soft_reset(void *handle) +static int sdma_v5_2_soft_reset_engine(struct amdgpu_device *adev, u32 instance_id) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; u32 grbm_soft_reset; u32 tmp; - int i; - for (i = 0; i < adev->sdma.num_instances; i++) { - grbm_soft_reset = REG_SET_FIELD(0, - GRBM_SOFT_RESET, SOFT_RESET_SDMA0, - 1); - grbm_soft_reset <<= i; + grbm_soft_reset = REG_SET_FIELD(0, + GRBM_SOFT_RESET, SOFT_RESET_SDMA0, + 1); + grbm_soft_reset <<= instance_id; - tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET); - tmp |= grbm_soft_reset; - DRM_DEBUG("GRBM_SOFT_RESET=0x%08X\n", tmp); - WREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET, tmp); - tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET); + tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET); + tmp |= grbm_soft_reset; + DRM_DEBUG("GRBM_SOFT_RESET=0x%08X\n", tmp); + WREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET, tmp); + tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET); - udelay(50); + udelay(50); - tmp &= ~grbm_soft_reset; - WREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET, tmp); - tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET); + tmp &= ~grbm_soft_reset; + WREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET, tmp); + tmp = RREG32_SOC15(GC, 0, mmGRBM_SOFT_RESET); + return 0; +} + +static int sdma_v5_2_soft_reset(struct amdgpu_ip_block *ip_block) +{ + struct amdgpu_device *adev = ip_block->adev; + int i; + for (i = 0; i < adev->sdma.num_instances; i++) { + sdma_v5_2_soft_reset_engine(adev, i); udelay(50); } return 0; } +static const struct amdgpu_sdma_funcs sdma_v5_2_sdma_funcs = { + .stop_kernel_queue = &sdma_v5_2_stop_queue, + .start_kernel_queue = &sdma_v5_2_restore_queue, + .soft_reset_kernel_queue = &sdma_v5_2_soft_reset_engine, +}; + /** * sdma_v5_2_start - setup and start the async dma engines * @@ -715,6 +817,7 @@ static int sdma_v5_2_soft_reset(void *handle) static int sdma_v5_2_start(struct amdgpu_device *adev) { int r = 0; + struct amdgpu_ip_block *ip_block; if (amdgpu_sriov_vf(adev)) { sdma_v5_2_ctx_switch_enable(adev, false); @@ -735,7 +838,11 @@ static int sdma_v5_2_start(struct amdgpu_device *adev) msleep(1000); } - sdma_v5_2_soft_reset(adev); + ip_block = amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_SDMA); + if (!ip_block) + return -EINVAL; + + sdma_v5_2_soft_reset(ip_block); /* unhalt the MEs */ sdma_v5_2_enable(adev, true); /* enable sdma ring preemption */ @@ -810,28 +917,18 @@ static int sdma_v5_2_ring_test_ring(struct amdgpu_ring *ring) int r; u32 tmp; u64 gpu_addr; - volatile uint32_t *cpu_ptr = NULL; tmp = 0xCAFEDEAD; - if (ring->is_mes_queue) { - uint32_t offset = 0; - offset = amdgpu_mes_ctx_get_offs(ring, - AMDGPU_MES_CTX_PADDING_OFFS); - gpu_addr = amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset); - cpu_ptr = amdgpu_mes_ctx_get_offs_cpu_addr(ring, offset); - *cpu_ptr = tmp; - } else { - r = amdgpu_device_wb_get(adev, &index); - if (r) { - dev_err(adev->dev, "(%d) failed to allocate wb slot\n", r); - return r; - } - - gpu_addr = adev->wb.gpu_addr + (index * 4); - adev->wb.wb[index] = cpu_to_le32(tmp); + r = amdgpu_device_wb_get(adev, &index); + if (r) { + dev_err(adev->dev, "(%d) failed to allocate wb slot\n", r); + return r; } + gpu_addr = adev->wb.gpu_addr + (index * 4); + adev->wb.wb[index] = cpu_to_le32(tmp); + r = amdgpu_ring_alloc(ring, 20); if (r) { DRM_ERROR("amdgpu: dma failed to lock ring %d (%d).\n", ring->idx, r); @@ -848,10 +945,7 @@ static int sdma_v5_2_ring_test_ring(struct amdgpu_ring *ring) amdgpu_ring_commit(ring); for (i = 0; i < adev->usec_timeout; i++) { - if (ring->is_mes_queue) - tmp = le32_to_cpu(*cpu_ptr); - else - tmp = le32_to_cpu(adev->wb.wb[index]); + tmp = le32_to_cpu(adev->wb.wb[index]); if (tmp == 0xDEADBEEF) break; if (amdgpu_emu_mode == 1) @@ -863,8 +957,7 @@ static int sdma_v5_2_ring_test_ring(struct amdgpu_ring *ring) if (i >= adev->usec_timeout) r = -ETIMEDOUT; - if (!ring->is_mes_queue) - amdgpu_device_wb_free(adev, index); + amdgpu_device_wb_free(adev, index); return r; } @@ -887,37 +980,23 @@ static int sdma_v5_2_ring_test_ib(struct amdgpu_ring *ring, long timeout) long r; u32 tmp = 0; u64 gpu_addr; - volatile uint32_t *cpu_ptr = NULL; tmp = 0xCAFEDEAD; memset(&ib, 0, sizeof(ib)); - if (ring->is_mes_queue) { - uint32_t offset = 0; - offset = amdgpu_mes_ctx_get_offs(ring, AMDGPU_MES_CTX_IB_OFFS); - ib.gpu_addr = amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset); - ib.ptr = (void *)amdgpu_mes_ctx_get_offs_cpu_addr(ring, offset); - - offset = amdgpu_mes_ctx_get_offs(ring, - AMDGPU_MES_CTX_PADDING_OFFS); - gpu_addr = amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset); - cpu_ptr = amdgpu_mes_ctx_get_offs_cpu_addr(ring, offset); - *cpu_ptr = tmp; - } else { - r = amdgpu_device_wb_get(adev, &index); - if (r) { - dev_err(adev->dev, "(%ld) failed to allocate wb slot\n", r); - return r; - } + r = amdgpu_device_wb_get(adev, &index); + if (r) { + dev_err(adev->dev, "(%ld) failed to allocate wb slot\n", r); + return r; + } - gpu_addr = adev->wb.gpu_addr + (index * 4); - adev->wb.wb[index] = cpu_to_le32(tmp); + gpu_addr = adev->wb.gpu_addr + (index * 4); + adev->wb.wb[index] = cpu_to_le32(tmp); - r = amdgpu_ib_get(adev, NULL, 256, AMDGPU_IB_POOL_DIRECT, &ib); - if (r) { - DRM_ERROR("amdgpu: failed to get ib (%ld).\n", r); - goto err0; - } + r = amdgpu_ib_get(adev, NULL, 256, AMDGPU_IB_POOL_DIRECT, &ib); + if (r) { + DRM_ERROR("amdgpu: failed to get ib (%ld).\n", r); + goto err0; } ib.ptr[0] = SDMA_PKT_HEADER_OP(SDMA_OP_WRITE) | @@ -945,10 +1024,7 @@ static int sdma_v5_2_ring_test_ib(struct amdgpu_ring *ring, long timeout) goto err1; } - if (ring->is_mes_queue) - tmp = le32_to_cpu(*cpu_ptr); - else - tmp = le32_to_cpu(adev->wb.wb[index]); + tmp = le32_to_cpu(adev->wb.wb[index]); if (tmp == 0xDEADBEEF) r = 0; @@ -956,11 +1032,10 @@ static int sdma_v5_2_ring_test_ib(struct amdgpu_ring *ring, long timeout) r = -EINVAL; err1: - amdgpu_ib_free(adev, &ib, NULL); + amdgpu_ib_free(&ib, NULL); dma_fence_put(f); err0: - if (!ring->is_mes_queue) - amdgpu_device_wb_free(adev, index); + amdgpu_device_wb_free(adev, index); return r; } @@ -1116,7 +1191,28 @@ static void sdma_v5_2_ring_emit_pipeline_sync(struct amdgpu_ring *ring) static void sdma_v5_2_ring_emit_vm_flush(struct amdgpu_ring *ring, unsigned vmid, uint64_t pd_addr) { - amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr); + struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->vm_hub]; + uint32_t req = hub->vmhub_funcs->get_invalidate_req(vmid, 0); + + /* Update the PD address for this VMID. */ + amdgpu_ring_emit_wreg(ring, hub->ctx0_ptb_addr_lo32 + + (hub->ctx_addr_distance * vmid), + lower_32_bits(pd_addr)); + amdgpu_ring_emit_wreg(ring, hub->ctx0_ptb_addr_hi32 + + (hub->ctx_addr_distance * vmid), + upper_32_bits(pd_addr)); + + /* Trigger invalidation. */ + amdgpu_ring_write(ring, + SDMA_PKT_VM_INVALIDATION_HEADER_OP(SDMA_OP_POLL_REGMEM) | + SDMA_PKT_VM_INVALIDATION_HEADER_SUB_OP(SDMA_SUBOP_VM_INVALIDATION) | + SDMA_PKT_VM_INVALIDATION_HEADER_GFX_ENG_ID(ring->vm_inv_eng) | + SDMA_PKT_VM_INVALIDATION_HEADER_MM_ENG_ID(0x1f)); + amdgpu_ring_write(ring, req); + amdgpu_ring_write(ring, 0xFFFFFFFF); + amdgpu_ring_write(ring, + SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_INVALIDATEACK(1 << vmid) | + SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_ADDRESSRANGEHI(0x1F)); } static void sdma_v5_2_ring_emit_wreg(struct amdgpu_ring *ring, @@ -1152,9 +1248,9 @@ static void sdma_v5_2_ring_emit_reg_write_reg_wait(struct amdgpu_ring *ring, amdgpu_ring_emit_reg_wait(ring, reg1, mask, mask); } -static int sdma_v5_2_early_init(void *handle) +static int sdma_v5_2_early_init(struct amdgpu_ip_block *ip_block) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; int r; r = amdgpu_sdma_init_microcode(adev, 0, true); @@ -1204,11 +1300,13 @@ static unsigned sdma_v5_2_seq_to_trap_id(int seq_num) return -EINVAL; } -static int sdma_v5_2_sw_init(void *handle) +static int sdma_v5_2_sw_init(struct amdgpu_ip_block *ip_block) { struct amdgpu_ring *ring; int r, i; - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; + uint32_t reg_count = ARRAY_SIZE(sdma_reg_list_5_2); + uint32_t *ptr; /* SDMA trap event */ for (i = 0; i < adev->sdma.num_instances; i++) { @@ -1220,6 +1318,7 @@ static int sdma_v5_2_sw_init(void *handle) } for (i = 0; i < adev->sdma.num_instances; i++) { + adev->sdma.instance[i].funcs = &sdma_v5_2_sdma_funcs; ring = &adev->sdma.instance[i].ring; ring->ring_obj = NULL; ring->use_doorbell = true; @@ -1240,32 +1339,64 @@ static int sdma_v5_2_sw_init(void *handle) return r; } + adev->sdma.supported_reset = + amdgpu_get_soft_full_reset_mask(&adev->sdma.instance[0].ring); + switch (amdgpu_ip_version(adev, SDMA0_HWIP, 0)) { + case IP_VERSION(5, 2, 0): + case IP_VERSION(5, 2, 2): + case IP_VERSION(5, 2, 3): + case IP_VERSION(5, 2, 4): + if (adev->sdma.instance[0].fw_version >= 76) + adev->sdma.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE; + break; + case IP_VERSION(5, 2, 5): + if (adev->sdma.instance[0].fw_version >= 34) + adev->sdma.supported_reset |= AMDGPU_RESET_TYPE_PER_QUEUE; + break; + default: + break; + } + + /* Allocate memory for SDMA IP Dump buffer */ + ptr = kcalloc(adev->sdma.num_instances * reg_count, sizeof(uint32_t), GFP_KERNEL); + if (ptr) + adev->sdma.ip_dump = ptr; + else + DRM_ERROR("Failed to allocated memory for SDMA IP Dump\n"); + + r = amdgpu_sdma_sysfs_reset_mask_init(adev); + if (r) + return r; + return r; } -static int sdma_v5_2_sw_fini(void *handle) +static int sdma_v5_2_sw_fini(struct amdgpu_ip_block *ip_block) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; int i; for (i = 0; i < adev->sdma.num_instances; i++) amdgpu_ring_fini(&adev->sdma.instance[i].ring); + amdgpu_sdma_sysfs_reset_mask_fini(adev); amdgpu_sdma_destroy_inst_ctx(adev, true); + kfree(adev->sdma.ip_dump); + return 0; } -static int sdma_v5_2_hw_init(void *handle) +static int sdma_v5_2_hw_init(struct amdgpu_ip_block *ip_block) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; return sdma_v5_2_start(adev); } -static int sdma_v5_2_hw_fini(void *handle) +static int sdma_v5_2_hw_fini(struct amdgpu_ip_block *ip_block) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; if (amdgpu_sriov_vf(adev)) return 0; @@ -1276,23 +1407,19 @@ static int sdma_v5_2_hw_fini(void *handle) return 0; } -static int sdma_v5_2_suspend(void *handle) +static int sdma_v5_2_suspend(struct amdgpu_ip_block *ip_block) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - - return sdma_v5_2_hw_fini(adev); + return sdma_v5_2_hw_fini(ip_block); } -static int sdma_v5_2_resume(void *handle) +static int sdma_v5_2_resume(struct amdgpu_ip_block *ip_block) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - - return sdma_v5_2_hw_init(adev); + return sdma_v5_2_hw_init(ip_block); } -static bool sdma_v5_2_is_idle(void *handle) +static bool sdma_v5_2_is_idle(struct amdgpu_ip_block *ip_block) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; u32 i; for (i = 0; i < adev->sdma.num_instances; i++) { @@ -1305,11 +1432,11 @@ static bool sdma_v5_2_is_idle(void *handle) return true; } -static int sdma_v5_2_wait_for_idle(void *handle) +static int sdma_v5_2_wait_for_idle(struct amdgpu_ip_block *ip_block) { unsigned i; u32 sdma0, sdma1, sdma2, sdma3; - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; for (i = 0; i < adev->usec_timeout; i++) { sdma0 = RREG32(sdma_v5_2_get_reg_offset(adev, 0, mmSDMA0_STATUS_REG)); @@ -1324,6 +1451,84 @@ static int sdma_v5_2_wait_for_idle(void *handle) return -ETIMEDOUT; } +static int sdma_v5_2_reset_queue(struct amdgpu_ring *ring, unsigned int vmid) +{ + struct amdgpu_device *adev = ring->adev; + u32 inst_id = ring->me; + + return amdgpu_sdma_reset_engine(adev, inst_id); +} + +static int sdma_v5_2_stop_queue(struct amdgpu_ring *ring) +{ + u32 f32_cntl, freeze, cntl, stat1_reg; + struct amdgpu_device *adev = ring->adev; + int i, j, r = 0; + + if (amdgpu_sriov_vf(adev)) + return -EINVAL; + + i = ring->me; + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); + + /* stop queue */ + sdma_v5_2_gfx_stop(adev, 1 << i); + + /*engine stop SDMA1_F32_CNTL.HALT to 1 and SDMAx_FREEZE freeze bit to 1 */ + freeze = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_FREEZE)); + freeze = REG_SET_FIELD(freeze, SDMA0_FREEZE, FREEZE, 1); + WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_FREEZE), freeze); + + for (j = 0; j < adev->usec_timeout; j++) { + freeze = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_FREEZE)); + + if (REG_GET_FIELD(freeze, SDMA0_FREEZE, FROZEN) & 1) + break; + udelay(1); + } + + + if (j == adev->usec_timeout) { + stat1_reg = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_STATUS1_REG)); + if ((stat1_reg & 0x3FF) != 0x3FF) { + DRM_ERROR("cannot soft reset as sdma not idle\n"); + r = -ETIMEDOUT; + goto err0; + } + } + + f32_cntl = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_F32_CNTL)); + f32_cntl = REG_SET_FIELD(f32_cntl, SDMA0_F32_CNTL, HALT, 1); + WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_F32_CNTL), f32_cntl); + + cntl = RREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_CNTL)); + cntl = REG_SET_FIELD(cntl, SDMA0_CNTL, UTC_L1_ENABLE, 0); + WREG32(sdma_v5_2_get_reg_offset(adev, i, mmSDMA0_CNTL), cntl); + +err0: + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); + return r; +} + +static int sdma_v5_2_restore_queue(struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = ring->adev; + u32 inst_id = ring->me; + u32 freeze; + int r; + + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); + /* unfreeze and unhalt */ + freeze = RREG32(sdma_v5_2_get_reg_offset(adev, inst_id, mmSDMA0_FREEZE)); + freeze = REG_SET_FIELD(freeze, SDMA0_FREEZE, FREEZE, 0); + WREG32(sdma_v5_2_get_reg_offset(adev, inst_id, mmSDMA0_FREEZE), freeze); + + r = sdma_v5_2_gfx_resume_instance(adev, inst_id, true); + + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); + return r; +} + static int sdma_v5_2_ring_preempt_ib(struct amdgpu_ring *ring) { int i, r = 0; @@ -1577,10 +1782,10 @@ static void sdma_v5_2_update_medium_grain_light_sleep(struct amdgpu_device *adev } } -static int sdma_v5_2_set_clockgating_state(void *handle, +static int sdma_v5_2_set_clockgating_state(struct amdgpu_ip_block *ip_block, enum amd_clockgating_state state) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; if (amdgpu_sriov_vf(adev)) return 0; @@ -1606,15 +1811,15 @@ static int sdma_v5_2_set_clockgating_state(void *handle, return 0; } -static int sdma_v5_2_set_powergating_state(void *handle, +static int sdma_v5_2_set_powergating_state(struct amdgpu_ip_block *ip_block, enum amd_powergating_state state) { return 0; } -static void sdma_v5_2_get_clockgating_state(void *handle, u64 *flags) +static void sdma_v5_2_get_clockgating_state(struct amdgpu_ip_block *ip_block, u64 *flags) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_device *adev = ip_block->adev; int data; if (amdgpu_sriov_vf(adev)) @@ -1642,6 +1847,10 @@ static void sdma_v5_2_ring_begin_use(struct amdgpu_ring *ring) * but it shouldn't hurt for other parts since * this GFXOFF will be disallowed anyway when SDMA is * active, this just makes it explicit. + * sdma_v5_2_ring_set_wptr() takes advantage of this + * to update the wptr because sometimes SDMA seems to miss + * doorbells when entering PG. If you remove this, update + * sdma_v5_2_ring_set_wptr() as well! */ amdgpu_gfx_off_ctrl(adev, false); } @@ -1657,10 +1866,51 @@ static void sdma_v5_2_ring_end_use(struct amdgpu_ring *ring) amdgpu_gfx_off_ctrl(adev, true); } -const struct amd_ip_funcs sdma_v5_2_ip_funcs = { +static void sdma_v5_2_print_ip_state(struct amdgpu_ip_block *ip_block, struct drm_printer *p) +{ + struct amdgpu_device *adev = ip_block->adev; + int i, j; + uint32_t reg_count = ARRAY_SIZE(sdma_reg_list_5_2); + uint32_t instance_offset; + + if (!adev->sdma.ip_dump) + return; + + drm_printf(p, "num_instances:%d\n", adev->sdma.num_instances); + for (i = 0; i < adev->sdma.num_instances; i++) { + instance_offset = i * reg_count; + drm_printf(p, "\nInstance:%d\n", i); + + for (j = 0; j < reg_count; j++) + drm_printf(p, "%-50s \t 0x%08x\n", sdma_reg_list_5_2[j].reg_name, + adev->sdma.ip_dump[instance_offset + j]); + } +} + +static void sdma_v5_2_dump_ip_state(struct amdgpu_ip_block *ip_block) +{ + struct amdgpu_device *adev = ip_block->adev; + int i, j; + uint32_t instance_offset; + uint32_t reg_count = ARRAY_SIZE(sdma_reg_list_5_2); + + if (!adev->sdma.ip_dump) + return; + + amdgpu_gfx_off_ctrl(adev, false); + for (i = 0; i < adev->sdma.num_instances; i++) { + instance_offset = i * reg_count; + for (j = 0; j < reg_count; j++) + adev->sdma.ip_dump[instance_offset + j] = + RREG32(sdma_v5_2_get_reg_offset(adev, i, + sdma_reg_list_5_2[j].reg_offset)); + } + amdgpu_gfx_off_ctrl(adev, true); +} + +static const struct amd_ip_funcs sdma_v5_2_ip_funcs = { .name = "sdma_v5_2", .early_init = sdma_v5_2_early_init, - .late_init = NULL, .sw_init = sdma_v5_2_sw_init, .sw_fini = sdma_v5_2_sw_fini, .hw_init = sdma_v5_2_hw_init, @@ -1673,6 +1923,8 @@ const struct amd_ip_funcs sdma_v5_2_ip_funcs = { .set_clockgating_state = sdma_v5_2_set_clockgating_state, .set_powergating_state = sdma_v5_2_set_powergating_state, .get_clockgating_state = sdma_v5_2_get_clockgating_state, + .dump_ip_state = sdma_v5_2_dump_ip_state, + .print_ip_state = sdma_v5_2_print_ip_state, }; static const struct amdgpu_ring_funcs sdma_v5_2_ring_funcs = { @@ -1711,6 +1963,7 @@ static const struct amdgpu_ring_funcs sdma_v5_2_ring_funcs = { .emit_reg_write_reg_wait = sdma_v5_2_ring_emit_reg_write_reg_wait, .init_cond_exec = sdma_v5_2_ring_init_cond_exec, .preempt_ib = sdma_v5_2_ring_preempt_ib, + .reset = sdma_v5_2_reset_queue, }; static void sdma_v5_2_set_ring_funcs(struct amdgpu_device *adev) @@ -1747,7 +2000,7 @@ static void sdma_v5_2_set_irq_funcs(struct amdgpu_device *adev) * @src_offset: src GPU address * @dst_offset: dst GPU address * @byte_count: number of bytes to xfer - * @tmz: if a secure copy should be used + * @copy_flags: copy flags for the buffers * * Copy GPU buffers using the DMA engine. * Used by the amdgpu ttm implementation to move pages if @@ -1757,11 +2010,11 @@ static void sdma_v5_2_emit_copy_buffer(struct amdgpu_ib *ib, uint64_t src_offset, uint64_t dst_offset, uint32_t byte_count, - bool tmz) + uint32_t copy_flags) { ib->ptr[ib->length_dw++] = SDMA_PKT_HEADER_OP(SDMA_OP_COPY) | SDMA_PKT_HEADER_SUB_OP(SDMA_SUBOP_COPY_LINEAR) | - SDMA_PKT_COPY_LINEAR_HEADER_TMZ(tmz ? 1 : 0); + SDMA_PKT_COPY_LINEAR_HEADER_TMZ((copy_flags & AMDGPU_COPY_FLAGS_TMZ) ? 1 : 0); ib->ptr[ib->length_dw++] = byte_count - 1; ib->ptr[ib->length_dw++] = 0; /* src/dst endian swap */ ib->ptr[ib->length_dw++] = lower_32_bits(src_offset); |