[igt-dev] [PATCH i-g-t] tests/amdgpu: add vce decoder test
vitaly.prosyak at amd.com
vitaly.prosyak at amd.com
Thu Jul 6 01:11:42 UTC 2023
From: Vitaly Prosyak <vitaly.prosyak at amd.com>
Ported and refactored drmlib vce_tests.c and the following is done:
1. Remove everywhere global variables.
2. Reuse common functions from amd_mmd_shared.h
3. Properly formatted code to meet igt guidelines.
Signed-off-by: Vitaly Prosyak <vitaly.prosyak at amd.com>
Acked-by: Leo Liu <leo.liu at amd.com>
---
tests/amdgpu/amd_vce_dec.c | 602 +++++++++++++++++++++++++++++++++++++
tests/amdgpu/meson.build | 1 +
2 files changed, 603 insertions(+)
create mode 100644 tests/amdgpu/amd_vce_dec.c
diff --git a/tests/amdgpu/amd_vce_dec.c b/tests/amdgpu/amd_vce_dec.c
new file mode 100644
index 000000000..364534955
--- /dev/null
+++ b/tests/amdgpu/amd_vce_dec.c
@@ -0,0 +1,602 @@
+// SPDX-License-Identifier: MIT
+// Copyright 2023 Advanced Micro Devices, Inc.
+// Copyright 2014 Advanced Micro Devices, Inc.
+
+#include "lib/amdgpu/amd_mmd_shared.h"
+#include "lib/amdgpu/amd_mmd_vce_ib.h"
+
+#define FW_53_0_03 ((53 << 24) | (0 << 16) | (03 << 8))
+
+struct amdgpu_vce_encode {
+ unsigned int width;
+ unsigned int height;
+ struct amdgpu_mmd_bo vbuf;
+ struct amdgpu_mmd_bo bs[2];
+ struct amdgpu_mmd_bo fb[2];
+ struct amdgpu_mmd_bo cpb;
+ unsigned int ib_len;
+ bool two_instance;
+ struct amdgpu_mmd_bo mvrefbuf;
+ struct amdgpu_mmd_bo mvb;
+ unsigned int mvbuf_size;
+};
+
+
+static bool
+is_vce_tests_enable(amdgpu_device_handle device_handle, uint32_t family_id,
+ uint32_t chip_id, uint32_t chip_rev, bool *is_mv_supported)
+{
+ uint64_t ids_flags;
+ uint32_t version, feature;
+ int r;
+ struct amdgpu_gpu_info gpu_info = {0};
+
+ r = amdgpu_query_gpu_info(device_handle, &gpu_info);
+ igt_assert_eq(r, 0);
+ ids_flags = gpu_info.ids_flags;
+
+ amdgpu_query_firmware_version(device_handle, AMDGPU_INFO_FW_VCE, 0,
+ 0, &version, &feature);
+
+ if (family_id >= AMDGPU_FAMILY_RV || family_id == AMDGPU_FAMILY_SI ||
+ is_gfx_pipe_removed(family_id, chip_id, chip_rev)) {
+ igt_info("\n\nThe ASIC NOT support VCE, tests are disabled\n");
+ return false;
+ }
+
+ if (!(chip_id == (chip_rev + 0x3C) || /* FIJI */
+ chip_id == (chip_rev + 0x50) || /* Polaris 10*/
+ chip_id == (chip_rev + 0x5A) || /* Polaris 11*/
+ chip_id == (chip_rev + 0x64) || /* Polaris 12*/
+ (family_id >= AMDGPU_FAMILY_AI && !ids_flags))) /* dGPU > Polaris */
+ igt_info("\n\nThe ASIC NOT support VCE MV, tests are disabled\n");
+ else if (version < FW_53_0_03)
+ igt_info("\n\nThe ASIC FW version NOT support VCE MV, tests are disabled\n");
+ else
+ *is_mv_supported = true;
+
+ return true;
+}
+
+static void
+amdgpu_cs_vce_create(amdgpu_device_handle device_handle,
+ struct amdgpu_vce_encode *enc, struct mmd_context *context, bool is_mv_supported)
+{
+ unsigned int align = (context->family_id >= AMDGPU_FAMILY_AI) ? 256 : 16;
+ int len, r;
+
+ enc->width = vce_create[6];
+ enc->height = vce_create[7];
+
+ context->num_resources = 0;
+ alloc_resource(device_handle, &enc->fb[0], 4096, AMDGPU_GEM_DOMAIN_GTT);
+ context->resources[context->num_resources++] = enc->fb[0].handle;
+ context->resources[context->num_resources++] = context->ib_handle;
+
+ len = 0;
+ memcpy(context->ib_cpu, vce_session, sizeof(vce_session));
+ len += sizeof(vce_session) / 4;
+ memcpy((context->ib_cpu + len), vce_taskinfo, sizeof(vce_taskinfo));
+ len += sizeof(vce_taskinfo) / 4;
+ memcpy((context->ib_cpu + len), vce_create, sizeof(vce_create));
+ context->ib_cpu[len + 8] = ALIGN(enc->width, align);
+ context->ib_cpu[len + 9] = ALIGN(enc->width, align);
+ if (is_mv_supported == true) {/* disableTwoInstance */
+ if (context->family_id >= AMDGPU_FAMILY_AI)
+ context->ib_cpu[len + 11] = 0x01000001;
+ else
+ context->ib_cpu[len + 11] = 0x01000201;
+ }
+ len += sizeof(vce_create) / 4;
+ memcpy((context->ib_cpu + len), vce_feedback, sizeof(vce_feedback));
+ context->ib_cpu[len + 2] = enc->fb[0].addr >> 32;
+ context->ib_cpu[len + 3] = enc->fb[0].addr;
+ len += sizeof(vce_feedback) / 4;
+
+ r = submit(device_handle, context, len, AMDGPU_HW_IP_VCE);
+ igt_assert_eq(r, 0);
+
+ free_resource(&enc->fb[0]);
+}
+
+static void
+amdgpu_cs_vce_config(amdgpu_device_handle device_handle,
+ struct mmd_context *context, bool is_mv_supported)
+{
+ int len = 0, r;
+
+ memcpy((context->ib_cpu + len), vce_session, sizeof(vce_session));
+ len += sizeof(vce_session) / 4;
+ memcpy((context->ib_cpu + len), vce_taskinfo, sizeof(vce_taskinfo));
+ context->ib_cpu[len + 3] = 2;
+ context->ib_cpu[len + 6] = 0xffffffff;
+ len += sizeof(vce_taskinfo) / 4;
+ memcpy((context->ib_cpu + len), vce_rate_ctrl, sizeof(vce_rate_ctrl));
+ len += sizeof(vce_rate_ctrl) / 4;
+ memcpy((context->ib_cpu + len), vce_config_ext, sizeof(vce_config_ext));
+ len += sizeof(vce_config_ext) / 4;
+ memcpy((context->ib_cpu + len), vce_motion_est, sizeof(vce_motion_est));
+ len += sizeof(vce_motion_est) / 4;
+ memcpy((context->ib_cpu + len), vce_rdo, sizeof(vce_rdo));
+ len += sizeof(vce_rdo) / 4;
+ memcpy((context->ib_cpu + len), vce_pic_ctrl, sizeof(vce_pic_ctrl));
+ if (is_mv_supported == true)
+ context->ib_cpu[len + 27] = 0x00000001; /* encSliceMode */
+ len += sizeof(vce_pic_ctrl) / 4;
+
+ r = submit(device_handle, context, len, AMDGPU_HW_IP_VCE);
+ igt_assert_eq(r, 0);
+}
+
+static void amdgpu_cs_vce_encode_idr(amdgpu_device_handle device_handle,
+ struct mmd_context *context, struct amdgpu_vce_encode *enc)
+{
+
+ uint64_t luma_offset, chroma_offset;
+ unsigned int align = (context->family_id >= AMDGPU_FAMILY_AI) ? 256 : 16;
+ unsigned int luma_size = ALIGN(enc->width, align) * ALIGN(enc->height, 16);
+ int len = 0, i, r;
+
+ luma_offset = enc->vbuf.addr;
+ chroma_offset = luma_offset + luma_size;
+
+ memcpy((context->ib_cpu + len), vce_session, sizeof(vce_session));
+ len += sizeof(vce_session) / 4;
+ memcpy((context->ib_cpu + len), vce_taskinfo, sizeof(vce_taskinfo));
+ len += sizeof(vce_taskinfo) / 4;
+ memcpy((context->ib_cpu + len), vce_bs_buffer, sizeof(vce_bs_buffer));
+ context->ib_cpu[len + 2] = enc->bs[0].addr >> 32;
+ context->ib_cpu[len + 3] = enc->bs[0].addr;
+ len += sizeof(vce_bs_buffer) / 4;
+ memcpy((context->ib_cpu + len), vce_context_buffer, sizeof(vce_context_buffer));
+ context->ib_cpu[len + 2] = enc->cpb.addr >> 32;
+ context->ib_cpu[len + 3] = enc->cpb.addr;
+ len += sizeof(vce_context_buffer) / 4;
+ memcpy((context->ib_cpu + len), vce_aux_buffer, sizeof(vce_aux_buffer));
+ for (i = 0; i < 8; ++i)
+ context->ib_cpu[len + 2 + i] = luma_size * 1.5 * (i + 2);
+ for (i = 0; i < 8; ++i)
+ context->ib_cpu[len + 10 + i] = luma_size * 1.5;
+ len += sizeof(vce_aux_buffer) / 4;
+ memcpy((context->ib_cpu + len), vce_feedback, sizeof(vce_feedback));
+ context->ib_cpu[len + 2] = enc->fb[0].addr >> 32;
+ context->ib_cpu[len + 3] = enc->fb[0].addr;
+ len += sizeof(vce_feedback) / 4;
+ memcpy((context->ib_cpu + len), vce_encode, sizeof(vce_encode));
+ context->ib_cpu[len + 9] = luma_offset >> 32;
+ context->ib_cpu[len + 10] = luma_offset;
+ context->ib_cpu[len + 11] = chroma_offset >> 32;
+ context->ib_cpu[len + 12] = chroma_offset;
+ context->ib_cpu[len + 14] = ALIGN(enc->width, align);
+ context->ib_cpu[len + 15] = ALIGN(enc->width, align);
+ context->ib_cpu[len + 73] = luma_size * 1.5;
+ context->ib_cpu[len + 74] = luma_size * 2.5;
+ len += sizeof(vce_encode) / 4;
+ enc->ib_len = len;
+ if (!enc->two_instance) {
+ r = submit(device_handle, context, len, AMDGPU_HW_IP_VCE);
+ igt_assert_eq(r, 0);
+ }
+}
+
+static void amdgpu_cs_vce_encode_p(amdgpu_device_handle device_handle,
+ struct mmd_context *context, struct amdgpu_vce_encode *enc)
+{
+ uint64_t luma_offset, chroma_offset;
+ int len, i, r;
+ unsigned int align = (context->family_id >= AMDGPU_FAMILY_AI) ? 256 : 16;
+ unsigned int luma_size = ALIGN(enc->width, align) * ALIGN(enc->height, 16);
+
+ len = (enc->two_instance) ? enc->ib_len : 0;
+ luma_offset = enc->vbuf.addr;
+ chroma_offset = luma_offset + luma_size;
+
+ if (!enc->two_instance) {
+ memcpy((context->ib_cpu + len), vce_session, sizeof(vce_session));
+ len += sizeof(vce_session) / 4;
+ }
+ memcpy((context->ib_cpu + len), vce_taskinfo, sizeof(vce_taskinfo));
+ len += sizeof(vce_taskinfo) / 4;
+ memcpy((context->ib_cpu + len), vce_bs_buffer, sizeof(vce_bs_buffer));
+ context->ib_cpu[len + 2] = enc->bs[1].addr >> 32;
+ context->ib_cpu[len + 3] = enc->bs[1].addr;
+ len += sizeof(vce_bs_buffer) / 4;
+ memcpy((context->ib_cpu + len), vce_context_buffer, sizeof(vce_context_buffer));
+ context->ib_cpu[len + 2] = enc->cpb.addr >> 32;
+ context->ib_cpu[len + 3] = enc->cpb.addr;
+ len += sizeof(vce_context_buffer) / 4;
+ memcpy((context->ib_cpu + len), vce_aux_buffer, sizeof(vce_aux_buffer));
+ for (i = 0; i < 8; ++i)
+ context->ib_cpu[len + 2 + i] = luma_size * 1.5 * (i + 2);
+ for (i = 0; i < 8; ++i)
+ context->ib_cpu[len + 10 + i] = luma_size * 1.5;
+ len += sizeof(vce_aux_buffer) / 4;
+ memcpy((context->ib_cpu + len), vce_feedback, sizeof(vce_feedback));
+ context->ib_cpu[len + 2] = enc->fb[1].addr >> 32;
+ context->ib_cpu[len + 3] = enc->fb[1].addr;
+ len += sizeof(vce_feedback) / 4;
+ memcpy((context->ib_cpu + len), vce_encode, sizeof(vce_encode));
+ context->ib_cpu[len + 2] = 0;
+ context->ib_cpu[len + 9] = luma_offset >> 32;
+ context->ib_cpu[len + 10] = luma_offset;
+ context->ib_cpu[len + 11] = chroma_offset >> 32;
+ context->ib_cpu[len + 12] = chroma_offset;
+ context->ib_cpu[len + 14] = ALIGN(enc->width, align);
+ context->ib_cpu[len + 15] = ALIGN(enc->width, align);
+ context->ib_cpu[len + 18] = 0;
+ context->ib_cpu[len + 19] = 0;
+ context->ib_cpu[len + 56] = 3;
+ context->ib_cpu[len + 57] = 0;
+ context->ib_cpu[len + 58] = 0;
+ context->ib_cpu[len + 59] = luma_size * 1.5;
+ context->ib_cpu[len + 60] = luma_size * 2.5;
+ context->ib_cpu[len + 73] = 0;
+ context->ib_cpu[len + 74] = luma_size;
+ context->ib_cpu[len + 81] = 1;
+ context->ib_cpu[len + 82] = 1;
+ len += sizeof(vce_encode) / 4;
+
+ r = submit(device_handle, context, len, AMDGPU_HW_IP_VCE);
+ igt_assert_eq(r, 0);
+}
+
+static void check_result(struct amdgpu_vce_encode *enc)
+{
+ uint64_t sum;
+ uint32_t s[2] = {180325, 15946};
+ uint32_t *ptr, size;
+ int i, j, r;
+
+ for (i = 0; i < 2; ++i) {
+ r = amdgpu_bo_cpu_map(enc->fb[i].handle, (void **)&enc->fb[i].ptr);
+ igt_assert_eq(r, 0);
+ ptr = (uint32_t *)enc->fb[i].ptr;
+ size = ptr[4] - ptr[9];
+ r = amdgpu_bo_cpu_unmap(enc->fb[i].handle);
+ igt_assert_eq(r, 0);
+ r = amdgpu_bo_cpu_map(enc->bs[i].handle, (void **)&enc->bs[i].ptr);
+ igt_assert_eq(r, 0);
+ for (j = 0, sum = 0; j < size; ++j)
+ sum += enc->bs[i].ptr[j];
+ igt_assert_eq(sum, s[i]);
+ r = amdgpu_bo_cpu_unmap(enc->bs[i].handle);
+ igt_assert_eq(r, 0);
+ }
+}
+
+static void
+amdgpu_cs_vce_encode(amdgpu_device_handle device_handle, struct mmd_context *context,
+ struct amdgpu_vce_encode *enc, bool is_mv_supported)
+{
+ uint32_t vbuf_size, bs_size = 0x154000, cpb_size;
+ unsigned int align = (context->family_id >= AMDGPU_FAMILY_AI) ? 256 : 16;
+ int i, r;
+
+ vbuf_size = ALIGN(enc->width, align) * ALIGN(enc->height, 16) * 1.5;
+ cpb_size = vbuf_size * 10;
+ context->num_resources = 0;
+ alloc_resource(device_handle, &enc->fb[0], 4096, AMDGPU_GEM_DOMAIN_GTT);
+ context->resources[context->num_resources++] = enc->fb[0].handle;
+ alloc_resource(device_handle, &enc->fb[1], 4096, AMDGPU_GEM_DOMAIN_GTT);
+ context->resources[context->num_resources++] = enc->fb[1].handle;
+ alloc_resource(device_handle, &enc->bs[0], bs_size, AMDGPU_GEM_DOMAIN_GTT);
+ context->resources[context->num_resources++] = enc->bs[0].handle;
+ alloc_resource(device_handle, &enc->bs[1], bs_size, AMDGPU_GEM_DOMAIN_GTT);
+ context->resources[context->num_resources++] = enc->bs[1].handle;
+ alloc_resource(device_handle, &enc->vbuf, vbuf_size, AMDGPU_GEM_DOMAIN_VRAM);
+ context->resources[context->num_resources++] = enc->vbuf.handle;
+ alloc_resource(device_handle, &enc->cpb, cpb_size, AMDGPU_GEM_DOMAIN_VRAM);
+ context->resources[context->num_resources++] = enc->cpb.handle;
+ context->resources[context->num_resources++] = context->ib_handle;
+
+ r = amdgpu_bo_cpu_map(enc->vbuf.handle, (void **)&enc->vbuf.ptr);
+ igt_assert_eq(r, 0);
+
+ memset(enc->vbuf.ptr, 0, vbuf_size);
+ for (i = 0; i < enc->height; ++i) {
+ memcpy(enc->vbuf.ptr, (frame + i * enc->width), enc->width);
+ enc->vbuf.ptr += ALIGN(enc->width, align);
+ }
+ for (i = 0; i < enc->height / 2; ++i) {
+ memcpy(enc->vbuf.ptr, ((frame + enc->height * enc->width) +
+ i * enc->width), enc->width);
+ enc->vbuf.ptr += ALIGN(enc->width, align);
+ }
+
+ r = amdgpu_bo_cpu_unmap(enc->vbuf.handle);
+ igt_assert_eq(r, 0);
+
+ amdgpu_cs_vce_config(device_handle, context, is_mv_supported);
+
+ if (context->family_id >= AMDGPU_FAMILY_VI) {
+ vce_taskinfo[3] = 3;
+ amdgpu_cs_vce_encode_idr(device_handle, context, enc);
+ amdgpu_cs_vce_encode_p(device_handle, context, enc);
+ check_result(enc);
+
+ /* two pipes */
+ vce_encode[16] = 0;
+ amdgpu_cs_vce_encode_idr(device_handle, context, enc);
+ amdgpu_cs_vce_encode_p(device_handle, context, enc);
+ check_result(enc);
+
+ /* two instances */
+ if (context->vce_harvest_config == 0) {
+ enc->two_instance = true;
+ vce_taskinfo[2] = 0x83;
+ vce_taskinfo[4] = 1;
+ amdgpu_cs_vce_encode_idr(device_handle, context, enc);
+ vce_taskinfo[2] = 0xffffffff;
+ vce_taskinfo[4] = 2;
+ amdgpu_cs_vce_encode_p(device_handle, context, enc);
+ check_result(enc);
+ }
+ } else {
+ vce_taskinfo[3] = 3;
+ vce_encode[16] = 0;
+ amdgpu_cs_vce_encode_idr(device_handle, context, enc);
+ amdgpu_cs_vce_encode_p(device_handle, context, enc);
+ check_result(enc);
+ }
+
+ free_resource(&enc->fb[0]);
+ free_resource(&enc->fb[1]);
+ free_resource(&enc->bs[0]);
+ free_resource(&enc->bs[1]);
+ free_resource(&enc->vbuf);
+ free_resource(&enc->cpb);
+}
+
+static void amdgpu_cs_vce_mv(amdgpu_device_handle device_handle,
+ struct mmd_context *context, struct amdgpu_vce_encode *enc)
+{
+ uint64_t luma_offset, chroma_offset;
+ uint64_t mv_ref_luma_offset;
+ unsigned int align = (context->family_id >= AMDGPU_FAMILY_AI) ? 256 : 16;
+ unsigned int luma_size = ALIGN(enc->width, align) * ALIGN(enc->height, 16);
+ int len = 0, i, r;
+
+ luma_offset = enc->vbuf.addr;
+ chroma_offset = luma_offset + luma_size;
+ mv_ref_luma_offset = enc->mvrefbuf.addr;
+
+ memcpy((context->ib_cpu + len), vce_session, sizeof(vce_session));
+ len += sizeof(vce_session) / 4;
+ memcpy((context->ib_cpu + len), vce_taskinfo, sizeof(vce_taskinfo));
+ len += sizeof(vce_taskinfo) / 4;
+ memcpy((context->ib_cpu + len), vce_bs_buffer, sizeof(vce_bs_buffer));
+ context->ib_cpu[len + 2] = enc->bs[0].addr >> 32;
+ context->ib_cpu[len + 3] = enc->bs[0].addr;
+ len += sizeof(vce_bs_buffer) / 4;
+ memcpy((context->ib_cpu + len), vce_context_buffer, sizeof(vce_context_buffer));
+ context->ib_cpu[len + 2] = enc->cpb.addr >> 32;
+ context->ib_cpu[len + 3] = enc->cpb.addr;
+ len += sizeof(vce_context_buffer) / 4;
+ memcpy((context->ib_cpu + len), vce_aux_buffer, sizeof(vce_aux_buffer));
+ for (i = 0; i < 8; ++i)
+ context->ib_cpu[len + 2 + i] = luma_size * 1.5 * (i + 2);
+ for (i = 0; i < 8; ++i)
+ context->ib_cpu[len + 10 + i] = luma_size * 1.5;
+ len += sizeof(vce_aux_buffer) / 4;
+ memcpy((context->ib_cpu + len), vce_feedback, sizeof(vce_feedback));
+ context->ib_cpu[len + 2] = enc->fb[0].addr >> 32;
+ context->ib_cpu[len + 3] = enc->fb[0].addr;
+ len += sizeof(vce_feedback) / 4;
+ memcpy((context->ib_cpu + len), vce_mv_buffer, sizeof(vce_mv_buffer));
+ context->ib_cpu[len + 2] = mv_ref_luma_offset >> 32;
+ context->ib_cpu[len + 3] = mv_ref_luma_offset;
+ context->ib_cpu[len + 4] = ALIGN(enc->width, align);
+ context->ib_cpu[len + 5] = ALIGN(enc->width, align);
+ context->ib_cpu[len + 6] = luma_size;
+ context->ib_cpu[len + 7] = enc->mvb.addr >> 32;
+ context->ib_cpu[len + 8] = enc->mvb.addr;
+ len += sizeof(vce_mv_buffer) / 4;
+ memcpy((context->ib_cpu + len), vce_encode, sizeof(vce_encode));
+ context->ib_cpu[len + 2] = 0;
+ context->ib_cpu[len + 3] = 0;
+ context->ib_cpu[len + 4] = 0x154000;
+ context->ib_cpu[len + 9] = luma_offset >> 32;
+ context->ib_cpu[len + 10] = luma_offset;
+ context->ib_cpu[len + 11] = chroma_offset >> 32;
+ context->ib_cpu[len + 12] = chroma_offset;
+ context->ib_cpu[len + 13] = ALIGN(enc->height, 16);
+ context->ib_cpu[len + 14] = ALIGN(enc->width, align);
+ context->ib_cpu[len + 15] = ALIGN(enc->width, align);
+ /* encDisableMBOffloading-encDisableTwoPipeMode-encInputPicArrayMode-encInputPicAddrMode */
+ context->ib_cpu[len + 16] = 0x01010000;
+ context->ib_cpu[len + 18] = 0; /* encPicType */
+ context->ib_cpu[len + 19] = 0; /* encIdrFlag */
+ context->ib_cpu[len + 20] = 0; /* encIdrPicId */
+ context->ib_cpu[len + 21] = 0; /* encMGSKeyPic */
+ context->ib_cpu[len + 22] = 0; /* encReferenceFlag */
+ context->ib_cpu[len + 23] = 0; /* encTemporalLayerIndex */
+ context->ib_cpu[len + 55] = 0; /* pictureStructure */
+ context->ib_cpu[len + 56] = 0; /* encPicType -ref[0] */
+ context->ib_cpu[len + 61] = 0; /* pictureStructure */
+ context->ib_cpu[len + 62] = 0; /* encPicType -ref[1] */
+ context->ib_cpu[len + 67] = 0; /* pictureStructure */
+ context->ib_cpu[len + 68] = 0; /* encPicType -ref1 */
+ context->ib_cpu[len + 81] = 1; /* frameNumber */
+ context->ib_cpu[len + 82] = 2; /* pictureOrderCount */
+ context->ib_cpu[len + 83] = 0xffffffff; /* numIPicRemainInRCGOP */
+ context->ib_cpu[len + 84] = 0xffffffff; /* numPPicRemainInRCGOP */
+ context->ib_cpu[len + 85] = 0xffffffff; /* numBPicRemainInRCGOP */
+ context->ib_cpu[len + 86] = 0xffffffff; /* numIRPicRemainInRCGOP */
+ context->ib_cpu[len + 87] = 0; /* remainedIntraRefreshPictures */
+ len += sizeof(vce_encode) / 4;
+
+ enc->ib_len = len;
+ r = submit(device_handle, context, len, AMDGPU_HW_IP_VCE);
+ igt_assert_eq(r, 0);
+}
+
+static void check_mv_result(struct amdgpu_vce_encode *enc)
+{
+ uint64_t sum;
+ uint32_t s = 140790;
+ int j, r;
+
+ r = amdgpu_bo_cpu_map(enc->fb[0].handle, (void **)&enc->fb[0].ptr);
+ igt_assert_eq(r, 0);
+ r = amdgpu_bo_cpu_unmap(enc->fb[0].handle);
+ igt_assert_eq(r, 0);
+ r = amdgpu_bo_cpu_map(enc->mvb.handle, (void **)&enc->mvb.ptr);
+ igt_assert_eq(r, 0);
+ for (j = 0, sum = 0; j < enc->mvbuf_size; ++j)
+ sum += enc->mvb.ptr[j];
+ igt_assert_eq(sum, s);
+ r = amdgpu_bo_cpu_unmap(enc->mvb.handle);
+ igt_assert_eq(r, 0);
+}
+
+static void
+amdgpu_cs_vce_encode_mv(amdgpu_device_handle device_handle, struct mmd_context *context,
+ struct amdgpu_vce_encode *enc, bool is_mv_supported)
+{
+ uint32_t vbuf_size, bs_size = 0x154000, cpb_size;
+ unsigned int align = (context->family_id >= AMDGPU_FAMILY_AI) ? 256 : 16;
+ int i, r;
+
+ vbuf_size = ALIGN(enc->width, align) * ALIGN(enc->height, 16) * 1.5;
+ enc->mvbuf_size = ALIGN(enc->width, 16) * ALIGN(enc->height, 16) / 8;
+ cpb_size = vbuf_size * 10;
+ context->num_resources = 0;
+ alloc_resource(device_handle, &enc->fb[0], 4096, AMDGPU_GEM_DOMAIN_GTT);
+ context->resources[context->num_resources++] = enc->fb[0].handle;
+ alloc_resource(device_handle, &enc->bs[0], bs_size, AMDGPU_GEM_DOMAIN_GTT);
+ context->resources[context->num_resources++] = enc->bs[0].handle;
+ alloc_resource(device_handle, &enc->mvb, enc->mvbuf_size, AMDGPU_GEM_DOMAIN_GTT);
+ context->resources[context->num_resources++] = enc->mvb.handle;
+ alloc_resource(device_handle, &enc->vbuf, vbuf_size, AMDGPU_GEM_DOMAIN_VRAM);
+ context->resources[context->num_resources++] = enc->vbuf.handle;
+ alloc_resource(device_handle, &enc->mvrefbuf, vbuf_size, AMDGPU_GEM_DOMAIN_VRAM);
+ context->resources[context->num_resources++] = enc->mvrefbuf.handle;
+ alloc_resource(device_handle, &enc->cpb, cpb_size, AMDGPU_GEM_DOMAIN_VRAM);
+ context->resources[context->num_resources++] = enc->cpb.handle;
+ context->resources[context->num_resources++] = context->ib_handle;
+
+ r = amdgpu_bo_cpu_map(enc->vbuf.handle, (void **)&enc->vbuf.ptr);
+ igt_assert_eq(r, 0);
+
+ memset(enc->vbuf.ptr, 0, vbuf_size);
+ for (i = 0; i < enc->height; ++i) {
+ memcpy(enc->vbuf.ptr, (frame + i * enc->width), enc->width);
+ enc->vbuf.ptr += ALIGN(enc->width, align);
+ }
+ for (i = 0; i < enc->height / 2; ++i) {
+ memcpy(enc->vbuf.ptr, ((frame + enc->height * enc->width) + i * enc->width), enc->width);
+ enc->vbuf.ptr += ALIGN(enc->width, align);
+ }
+
+ r = amdgpu_bo_cpu_unmap(enc->vbuf.handle);
+ igt_assert_eq(r, 0);
+
+ r = amdgpu_bo_cpu_map(enc->mvrefbuf.handle, (void **)&enc->mvrefbuf.ptr);
+ igt_assert_eq(r, 0);
+
+ memset(enc->mvrefbuf.ptr, 0, vbuf_size);
+ for (i = 0; i < enc->height; ++i) {
+ memcpy(enc->mvrefbuf.ptr, (frame + (enc->height - i - 1) * enc->width), enc->width);
+ enc->mvrefbuf.ptr += ALIGN(enc->width, align);
+ }
+ for (i = 0; i < enc->height / 2; ++i) {
+ memcpy(enc->mvrefbuf.ptr,
+ ((frame + enc->height * enc->width) + (enc->height / 2 - i - 1) * enc->width), enc->width);
+ enc->mvrefbuf.ptr += ALIGN(enc->width, align);
+ }
+
+ r = amdgpu_bo_cpu_unmap(enc->mvrefbuf.handle);
+ igt_assert_eq(r, 0);
+
+ amdgpu_cs_vce_config(device_handle, context, is_mv_supported);
+
+ vce_taskinfo[3] = 3;
+ amdgpu_cs_vce_mv(device_handle, context, enc);
+ check_mv_result(enc);
+
+ free_resource(&enc->fb[0]);
+ free_resource(&enc->bs[0]);
+ free_resource(&enc->vbuf);
+ free_resource(&enc->cpb);
+ free_resource(&enc->mvrefbuf);
+ free_resource(&enc->mvb);
+}
+
+static void
+amdgpu_cs_vce_destroy(amdgpu_device_handle device_handle, struct mmd_context *context,
+ struct amdgpu_vce_encode *enc)
+{
+ int len, r;
+
+ context->num_resources = 0;
+ alloc_resource(device_handle, &enc->fb[0], 4096, AMDGPU_GEM_DOMAIN_GTT);
+ context->resources[context->num_resources++] = enc->fb[0].handle;
+ context->resources[context->num_resources++] = context->ib_handle;
+
+ len = 0;
+ memcpy(context->ib_cpu, vce_session, sizeof(vce_session));
+ len += sizeof(vce_session) / 4;
+ memcpy((context->ib_cpu + len), vce_taskinfo, sizeof(vce_taskinfo));
+ context->ib_cpu[len + 3] = 1;
+ len += sizeof(vce_taskinfo) / 4;
+ memcpy((context->ib_cpu + len), vce_feedback, sizeof(vce_feedback));
+ context->ib_cpu[len + 2] = enc->fb[0].addr >> 32;
+ context->ib_cpu[len + 3] = enc->fb[0].addr;
+ len += sizeof(vce_feedback) / 4;
+ memcpy((context->ib_cpu + len), vce_destroy, sizeof(vce_destroy));
+ len += sizeof(vce_destroy) / 4;
+
+ r = submit(device_handle, context, len, AMDGPU_HW_IP_VCE);
+ igt_assert_eq(r, 0);
+
+ free_resource(&enc->fb[0]);
+}
+
+igt_main
+{
+ amdgpu_device_handle device;
+ struct mmd_context context = {};
+ struct amdgpu_vce_encode enc = {};
+ int fd = -1;
+ bool is_mv_supported = false;
+
+ igt_fixture {
+ uint32_t major, minor;
+ int err;
+
+ fd = drm_open_driver(DRIVER_AMDGPU);
+ err = amdgpu_device_initialize(fd, &major, &minor, &device);
+ igt_require(err == 0);
+ igt_info("Initialized amdgpu, driver version %d.%d\n",
+ major, minor);
+ err = mmd_context_init(device, &context);
+ igt_require(err == 0);
+ igt_skip_on(!is_vce_tests_enable(device, context.family_id, context.chip_id,
+ context.chip_rev, &is_mv_supported));
+ }
+ igt_describe("Test whether vce enc is created");
+ igt_subtest("amdgpu_cs_vce_create")
+ amdgpu_cs_vce_create(device, &enc, &context, is_mv_supported);
+
+ igt_describe("Test whether vce enc encodes");
+ igt_subtest("amdgpu_cs_vce_encode")
+ amdgpu_cs_vce_encode(device, &context, &enc, is_mv_supported);
+
+ if (is_mv_supported) {
+ igt_describe("Test whether vce enc encodes mv");
+ igt_subtest("amdgpu_cs_vce_encode_mv")
+ amdgpu_cs_vce_encode_mv(device, &context, &enc, is_mv_supported);
+ }
+
+ igt_describe("Test whether vce enc is destroyed");
+ igt_subtest("amdgpu_cs_vce_destroy")
+ amdgpu_cs_vce_destroy(device, &context, &enc);
+
+
+ igt_fixture {
+ mmd_context_clean(device, &context);
+ amdgpu_device_deinitialize(device);
+ drm_close_driver(fd);
+ }
+
+}
diff --git a/tests/amdgpu/meson.build b/tests/amdgpu/meson.build
index 576333aab..e6e6eaabf 100644
--- a/tests/amdgpu/meson.build
+++ b/tests/amdgpu/meson.build
@@ -26,6 +26,7 @@ if libdrm_amdgpu.found()
'amd_psr',
'amd_uvd_dec',
'amd_uvd_enc',
+ 'amd_vce_dec',
'amd_vrr_range',
]
amdgpu_deps += libdrm_amdgpu
--
2.25.1
More information about the igt-dev
mailing list