diff options
Diffstat (limited to 'gen6_vme.c')
| -rw-r--r-- | gen6_vme.c | 740 |
1 files changed, 740 insertions, 0 deletions
diff --git a/gen6_vme.c b/gen6_vme.c new file mode 100644 index 0000000..ccf26bc --- /dev/null +++ b/gen6_vme.c @@ -0,0 +1,740 @@ +/* + * Copyright © 2009 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL PRECISION INSIGHT AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: + * Zhou Chang <chang.zhou@intel.com> + * + */ + +#include <stdio.h> +#include <string.h> +#include <assert.h> + +#include <va/va_backend.h> + +#include "intel_batchbuffer.h" +#include "intel_driver.h" + +#include "i965_defines.h" +#include "i965_drv_video.h" +#include "gen6_vme.h" + +#define VME_INTRA_SHADER 0 +#define VME_INTER_SHADER 1 + +#define CURBE_ALLOCATION_SIZE 37 /* in 256-bit */ +#define CURBE_TOTAL_DATA_LENGTH (4 * 32) /* in byte, it should be less than or equal to CURBE_ALLOCATION_SIZE * 32 */ +#define CURBE_URB_ENTRY_LENGTH 4 /* in 256-bit, it should be less than or equal to CURBE_TOTAL_DATA_LENGTH / 32 */ + +static uint32_t gen6_vme_intra_frame[][4] = { +#include "shaders/vme/intra_frame.g6b" + {0,0,0,0} +}; + +static uint32_t gen6_vme_inter_frame[][4] = { +#include "shaders/vme/inter_frame.g6b" + {0,0,0,0} +}; + +static struct media_kernel gen6_vme_kernels[] = { + { + "VME Intra Frame", + VME_INTRA_SHADER, /*index*/ + gen6_vme_intra_frame, + sizeof(gen6_vme_intra_frame), + NULL + }, + { + "VME inter Frame", + VME_INTER_SHADER, + gen6_vme_inter_frame, + sizeof(gen6_vme_inter_frame), + NULL + } +}; + +#define GEN6_VME_KERNEL_NUMBER ARRAY_ELEMS(gen6_vme_kernels) + +static void +gen6_vme_set_common_surface_tiling(struct i965_surface_state *ss, unsigned int tiling) +{ + switch (tiling) { + case I915_TILING_NONE: + ss->ss3.tiled_surface = 0; + ss->ss3.tile_walk = 0; + break; + case I915_TILING_X: + ss->ss3.tiled_surface = 1; + ss->ss3.tile_walk = I965_TILEWALK_XMAJOR; + break; + case I915_TILING_Y: + ss->ss3.tiled_surface = 1; + ss->ss3.tile_walk = I965_TILEWALK_YMAJOR; + break; + } +} + +static void +gen6_vme_set_source_surface_tiling(struct i965_surface_state2 *ss, unsigned int tiling) +{ + switch (tiling) { + case I915_TILING_NONE: + ss->ss2.tiled_surface = 0; + ss->ss2.tile_walk = 0; + break; + case I915_TILING_X: + ss->ss2.tiled_surface = 1; + ss->ss2.tile_walk = I965_TILEWALK_XMAJOR; + break; + case I915_TILING_Y: + ss->ss2.tiled_surface = 1; + ss->ss2.tile_walk = I965_TILEWALK_YMAJOR; + break; + } +} + +/* only used for VME source surface state */ +static void gen6_vme_source_surface_state(VADriverContextP ctx, + int index, + struct object_surface *obj_surface) +{ + struct i965_driver_data *i965 = i965_driver_data(ctx); + struct gen6_media_state *media_state = &i965->gen6_media_state; + struct i965_surface_state2 *ss; + dri_bo *bo; + int w, h, w_pitch, h_pitch; + unsigned int tiling, swizzle; + + assert(obj_surface->bo); + dri_bo_get_tiling(obj_surface->bo, &tiling, &swizzle); + + w = obj_surface->orig_width; + h = obj_surface->orig_height; + w_pitch = obj_surface->width; + h_pitch = obj_surface->height; + + bo = dri_bo_alloc(i965->intel.bufmgr, + "VME surface state", + sizeof(struct i965_surface_state2), + 0x1000); + assert(bo); + dri_bo_map(bo, 1); + assert(bo->virtual); + ss = bo->virtual; + memset(ss, 0, sizeof(*ss)); + + ss->ss0.surface_base_address = obj_surface->bo->offset; + + ss->ss1.cbcr_pixel_offset_v_direction = 2; + ss->ss1.width = w - 1; + ss->ss1.height = h - 1; + + ss->ss2.surface_format = MFX_SURFACE_PLANAR_420_8; + ss->ss2.interleave_chroma = 1; + ss->ss2.pitch = w_pitch - 1; + ss->ss2.half_pitch_for_chroma = 0; + + gen6_vme_set_source_surface_tiling(ss, tiling); + + /* UV offset for interleave mode */ + ss->ss3.x_offset_for_cb = 0; + ss->ss3.y_offset_for_cb = h_pitch; + + dri_bo_unmap(bo); + + dri_bo_emit_reloc(bo, + I915_GEM_DOMAIN_RENDER, 0, + 0, + offsetof(struct i965_surface_state2, ss0), + obj_surface->bo); + + assert(index < MAX_MEDIA_SURFACES_GEN6); + media_state->surface_state[index].bo = bo; +} + +static void +gen6_vme_media_source_surface_state(VADriverContextP ctx, + int index, + struct object_surface *obj_surface) +{ + struct i965_driver_data *i965 = i965_driver_data(ctx); + struct gen6_media_state *media_state = &i965->gen6_media_state; + struct i965_surface_state *ss; + dri_bo *bo; + int w, h, w_pitch; + unsigned int tiling, swizzle; + + w = obj_surface->orig_width; + h = obj_surface->orig_height; + w_pitch = obj_surface->width; + + /* Y plane */ + dri_bo_get_tiling(obj_surface->bo, &tiling, &swizzle); + bo = dri_bo_alloc(i965->intel.bufmgr, + "surface state", + sizeof(struct i965_surface_state), + 0x1000); + assert(bo); + + dri_bo_map(bo, True); + assert(bo->virtual); + ss = bo->virtual; + memset(ss, 0, sizeof(*ss)); + ss->ss0.surface_type = I965_SURFACE_2D; + ss->ss0.surface_format = I965_SURFACEFORMAT_R8_UNORM; + ss->ss1.base_addr = obj_surface->bo->offset; + ss->ss2.width = w / 4 - 1; + ss->ss2.height = h - 1; + ss->ss3.pitch = w_pitch - 1; + gen6_vme_set_common_surface_tiling(ss, tiling); + dri_bo_emit_reloc(bo, + I915_GEM_DOMAIN_RENDER, + 0, + 0, + offsetof(struct i965_surface_state, ss1), + obj_surface->bo); + dri_bo_unmap(bo); + + assert(index < MAX_MEDIA_SURFACES_GEN6); + media_state->surface_state[index].bo = bo; +} + +static VAStatus +gen6_vme_output_buffer_setup(VADriverContextP ctx, + VAContextID context, + struct mfc_encode_state *encode_state, + int index) + +{ + struct i965_driver_data *i965 = i965_driver_data(ctx); + struct gen6_media_state *media_state = &i965->gen6_media_state; + struct object_context *obj_context = CONTEXT(context); + struct i965_surface_state *ss; + dri_bo *bo; + int width_in_mbs = ALIGN(obj_context->picture_width, 16) / 16; + int height_in_mbs = ALIGN(obj_context->picture_height, 16) / 16; + int num_entries; + VAEncSliceParameterBuffer *pSliceParameter = (VAEncSliceParameterBuffer *)encode_state->slice_params[0]->buffer; + int is_intra = pSliceParameter->slice_flags.bits.is_intra; + + if ( is_intra ) { + media_state->vme_output.num_blocks = width_in_mbs * height_in_mbs; + } else { + media_state->vme_output.num_blocks = width_in_mbs * height_in_mbs * 4; + } + media_state->vme_output.size_block = 16; /* an OWORD */ + media_state->vme_output.pitch = ALIGN(media_state->vme_output.size_block, 16); + bo = dri_bo_alloc(i965->intel.bufmgr, + "VME output buffer", + media_state->vme_output.num_blocks * media_state->vme_output.pitch, + 0x1000); + assert(bo); + media_state->vme_output.bo = bo; + + bo = dri_bo_alloc(i965->intel.bufmgr, + "VME output buffer state", + sizeof(struct i965_surface_state), + 0x1000); + assert(bo); + dri_bo_map(bo, 1); + assert(bo->virtual); + ss = bo->virtual; + memset(ss, 0, sizeof(*ss)); + + /* always use 16 bytes as pitch on Sandy Bridge */ + num_entries = media_state->vme_output.num_blocks * media_state->vme_output.pitch / 16; + ss->ss0.render_cache_read_mode = 1; + ss->ss0.surface_type = I965_SURFACE_BUFFER; + ss->ss1.base_addr = media_state->vme_output.bo->offset; + ss->ss2.width = ((num_entries - 1) & 0x7f); + ss->ss2.height = (((num_entries - 1) >> 7) & 0x1fff); + ss->ss3.depth = (((num_entries - 1) >> 20) & 0x7f); + ss->ss3.pitch = media_state->vme_output.pitch - 1; + dri_bo_emit_reloc(bo, + I915_GEM_DOMAIN_RENDER, I915_GEM_DOMAIN_RENDER, + 0, + offsetof(struct i965_surface_state, ss1), + media_state->vme_output.bo); + + dri_bo_unmap(bo); + + assert(index < MAX_MEDIA_SURFACES_GEN6); + media_state->surface_state[index].bo = bo; + return VA_STATUS_SUCCESS; +} + +static VAStatus gen6_vme_surface_setup(VADriverContextP ctx, + VAContextID context, + struct mfc_encode_state *encode_state, + int is_intra) +{ + struct i965_driver_data *i965 = i965_driver_data(ctx); + struct gen6_media_state *media_state = &i965->gen6_media_state; + struct object_surface *obj_surface; + unsigned int *binding_table; + dri_bo *bo = media_state->binding_table.bo; + int i; + VAEncPictureParameterBufferH264 *pPicParameter = (VAEncPictureParameterBufferH264 *)encode_state->pic_param->buffer; + + /*Setup surfaces state*/ + /* current picture for encoding */ + obj_surface = SURFACE(encode_state->current_render_target); + assert(obj_surface); + gen6_vme_source_surface_state(ctx, 0, obj_surface); + gen6_vme_media_source_surface_state(ctx, 4, obj_surface); + + if ( ! is_intra ) { + /* reference 0 */ + obj_surface = SURFACE(pPicParameter->reference_picture); + assert(obj_surface); + gen6_vme_source_surface_state(ctx, 1, obj_surface); + /* reference 1, FIXME: */ + // obj_surface = SURFACE(pPicParameter->reference_picture); + // assert(obj_surface); + //gen6_vme_source_surface_state(ctx, 2, obj_surface); + } + + /* VME output */ + gen6_vme_output_buffer_setup(ctx, context, encode_state, 3); + + /*Building binding table*/ + dri_bo_map(bo, 1); + assert(bo->virtual); + binding_table = bo->virtual; + memset(binding_table, 0, bo->size); + + for (i = 0; i < MAX_MEDIA_SURFACES_GEN6; i++) { + if (media_state->surface_state[i].bo) { + binding_table[i] = media_state->surface_state[i].bo->offset; + dri_bo_emit_reloc(bo, + I915_GEM_DOMAIN_INSTRUCTION, 0, + 0, + i * sizeof(*binding_table), + media_state->surface_state[i].bo); + } + } + + dri_bo_unmap(media_state->binding_table.bo); + + return VA_STATUS_SUCCESS; +} + +static VAStatus gen6_vme_interface_setup(VADriverContextP ctx, + VAContextID context, + struct mfc_encode_state *encode_state) +{ + struct i965_driver_data *i965 = i965_driver_data(ctx); + struct gen6_media_state *media_state = &i965->gen6_media_state; + struct gen6_interface_descriptor_data *desc; + int i; + dri_bo *bo; + + bo = media_state->idrt.bo; + dri_bo_map(bo, 1); + assert(bo->virtual); + desc = bo->virtual; + + for (i = 0; i < GEN6_VME_KERNEL_NUMBER; i++) { + struct media_kernel *kernel; + kernel = &gen6_vme_kernels[i]; + assert(sizeof(*desc) == 32); + /*Setup the descritor table*/ + memset(desc, 0, sizeof(*desc)); + desc->desc0.kernel_start_pointer = (kernel->bo->offset >> 6); + desc->desc2.sampler_count = 1; /* FIXME: */ + desc->desc2.sampler_state_pointer = (media_state->vme_state.bo->offset >> 5); + desc->desc3.binding_table_entry_count = 1; /* FIXME: */ + desc->desc3.binding_table_pointer = (media_state->binding_table.bo->offset >> 5); + desc->desc4.constant_urb_entry_read_offset = 0; + desc->desc4.constant_urb_entry_read_length = CURBE_URB_ENTRY_LENGTH; + + /*kernel start*/ + dri_bo_emit_reloc(bo, + I915_GEM_DOMAIN_INSTRUCTION, 0, + 0, + i * sizeof(*desc) + offsetof(struct gen6_interface_descriptor_data, desc0), + kernel->bo); + /*Sampler State(VME state pointer)*/ + dri_bo_emit_reloc(bo, + I915_GEM_DOMAIN_INSTRUCTION, 0, + (1 << 2), // + i * sizeof(*desc) + offsetof(struct gen6_interface_descriptor_data, desc2), + media_state->vme_state.bo); + /*binding table*/ + dri_bo_emit_reloc(bo, + I915_GEM_DOMAIN_INSTRUCTION, 0, + 4, //One Entry + i * sizeof(*desc) + offsetof(struct gen6_interface_descriptor_data, desc3), + media_state->binding_table.bo); + desc++; + } + dri_bo_unmap(bo); + + return VA_STATUS_SUCCESS; +} + +static VAStatus gen6_vme_constant_setup(VADriverContextP ctx, + VAContextID context, + struct mfc_encode_state *encode_state) +{ + struct i965_driver_data *i965 = i965_driver_data(ctx); + struct gen6_media_state *media_state = &i965->gen6_media_state; + unsigned char *constant_buffer; + + dri_bo_map(media_state->curbe.bo, 1); + assert(media_state->curbe.bo->virtual); + constant_buffer = media_state->curbe.bo->virtual; + + /*TODO copy buffer into CURB*/ + + dri_bo_unmap( media_state->curbe.bo); + + return VA_STATUS_SUCCESS; +} + +static VAStatus gen6_vme_vme_state_setup(VADriverContextP ctx, VAContextID context, struct mfc_encode_state *encode_state, int is_intra) +{ + struct i965_driver_data *i965 = i965_driver_data(ctx); + struct gen6_media_state *media_state = &i965->gen6_media_state; + unsigned int *vme_state_message; + int i; + + //building VME state message + dri_bo_map(media_state->vme_state.bo, 1); + assert(media_state->vme_state.bo->virtual); + vme_state_message = (unsigned int *)media_state->vme_state.bo->virtual; + + for(i = 0;i < 32; i++) { + vme_state_message[i] = 0x11; + } + vme_state_message[16] = 0x42424242; //cost function LUT set 0 for Intra + + dri_bo_unmap( media_state->vme_state.bo); + return VA_STATUS_SUCCESS; +} + +static void gen6_vme_pipeline_select(VADriverContextP ctx) +{ + BEGIN_BATCH(ctx, 1); + OUT_BATCH(ctx, CMD_PIPELINE_SELECT | PIPELINE_SELECT_MEDIA); + ADVANCE_BATCH(ctx); +} + +static void gen6_vme_state_base_address(VADriverContextP ctx) +{ + BEGIN_BATCH(ctx, 10); + + OUT_BATCH(ctx, CMD_STATE_BASE_ADDRESS | 8); + + OUT_BATCH(ctx, 0 | BASE_ADDRESS_MODIFY); //General State Base Address + OUT_BATCH(ctx, 0 | BASE_ADDRESS_MODIFY); //Surface State Base Address + OUT_BATCH(ctx, 0 | BASE_ADDRESS_MODIFY); //Dynamic State Base Address + OUT_BATCH(ctx, 0 | BASE_ADDRESS_MODIFY); //Indirect Object Base Address + OUT_BATCH(ctx, 0 | BASE_ADDRESS_MODIFY); //Instruction Base Address + + OUT_BATCH(ctx, 0xFFFFF000 | BASE_ADDRESS_MODIFY); //General State Access Upper Bound + OUT_BATCH(ctx, 0xFFFFF000 | BASE_ADDRESS_MODIFY); //Dynamic State Access Upper Bound + OUT_BATCH(ctx, 0xFFFFF000 | BASE_ADDRESS_MODIFY); //Indirect Object Access Upper Bound + OUT_BATCH(ctx, 0xFFFFF000 | BASE_ADDRESS_MODIFY); //Instruction Access Upper Bound + + /* + OUT_BATCH(ctx, 0 | BASE_ADDRESS_MODIFY); //LLC Coherent Base Address + OUT_BATCH(ctx, 0xFFFFF000 | BASE_ADDRESS_MODIFY ); //LLC Coherent Upper Bound + */ + + ADVANCE_BATCH(ctx); +} + +static void gen6_vme_vfe_state(VADriverContextP ctx) +{ + struct i965_driver_data *i965 = i965_driver_data(ctx); + struct gen6_media_state *media_state = &i965->gen6_media_state; + + BEGIN_BATCH(ctx, 8); + + OUT_BATCH(ctx, CMD_MEDIA_VFE_STATE | 6); /*Gen6 CMD_MEDIA_STATE_POINTERS = CMD_MEDIA_STATE */ + OUT_BATCH(ctx, 0); /*Scratch Space Base Pointer and Space*/ + OUT_BATCH(ctx, (media_state->vfe_state.max_num_threads << 16) + | (media_state->vfe_state.num_urb_entries << 8) + | (media_state->vfe_state.gpgpu_mode << 2) ); /*Maximum Number of Threads , Number of URB Entries, MEDIA Mode*/ + OUT_BATCH(ctx, 0); /*Debug: Object ID*/ + OUT_BATCH(ctx, (media_state->vfe_state.urb_entry_size << 16) + | media_state->vfe_state.curbe_allocation_size); /*URB Entry Allocation Size , CURBE Allocation Size*/ + OUT_BATCH(ctx, 0); /*Disable Scoreboard*/ + OUT_BATCH(ctx, 0); /*Disable Scoreboard*/ + OUT_BATCH(ctx, 0); /*Disable Scoreboard*/ + + ADVANCE_BATCH(ctx); + +} + +static void gen6_vme_curbe_load(VADriverContextP ctx) +{ + struct i965_driver_data *i965 = i965_driver_data(ctx); + struct gen6_media_state *media_state = &i965->gen6_media_state; + + BEGIN_BATCH(ctx, 4); + + OUT_BATCH(ctx, CMD_MEDIA_CURBE_LOAD | 2); + OUT_BATCH(ctx, 0); + + OUT_BATCH(ctx, CURBE_TOTAL_DATA_LENGTH); + OUT_RELOC(ctx, media_state->curbe.bo, I915_GEM_DOMAIN_INSTRUCTION, 0, 0); + + ADVANCE_BATCH(ctx); +} + +static void gen6_vme_idrt(VADriverContextP ctx) +{ + struct i965_driver_data *i965 = i965_driver_data(ctx); + struct gen6_media_state *media_state = &i965->gen6_media_state; + + BEGIN_BATCH(ctx, 4); + + OUT_BATCH(ctx, CMD_MEDIA_INTERFACE_LOAD | 2); + OUT_BATCH(ctx, 0); + OUT_BATCH(ctx, GEN6_VME_KERNEL_NUMBER * sizeof(struct gen6_interface_descriptor_data)); + OUT_RELOC(ctx, media_state->idrt.bo, I915_GEM_DOMAIN_INSTRUCTION, 0, 0); + + ADVANCE_BATCH(ctx); +} + +static int gen6_vme_media_object(VADriverContextP ctx, + VAContextID context, + struct mfc_encode_state *encode_state, + int mb_x, int mb_y, + int kernel) +{ + struct i965_driver_data *i965 = i965_driver_data(ctx); + struct object_surface *obj_surface = SURFACE( encode_state->current_render_target); + int mb_width = ALIGN(obj_surface->orig_width, 16) / 16; + int len_in_dowrds = 6 + 1; + + BEGIN_BATCH(ctx, len_in_dowrds); + + OUT_BATCH(ctx, CMD_MEDIA_OBJECT | (len_in_dowrds - 2)); + OUT_BATCH(ctx, kernel); /*Interface Descriptor Offset*/ + OUT_BATCH(ctx, 0); + OUT_BATCH(ctx, 0); + OUT_BATCH(ctx, 0); + OUT_BATCH(ctx, 0); + + /*inline data */ + OUT_BATCH(ctx, mb_width << 16 | mb_y << 8 | mb_x); /*M0.0 Refrence0 X,Y, not used in Intra*/ + ADVANCE_BATCH(ctx); + + return len_in_dowrds * 4; +} + +static void gen6_vme_media_init(VADriverContextP ctx) +{ + int i; + struct i965_driver_data *i965 = i965_driver_data(ctx); + struct gen6_media_state *media_state = &i965->gen6_media_state; + dri_bo *bo; + + /* constant buffer */ + dri_bo_unreference(media_state->curbe.bo); + bo = dri_bo_alloc(i965->intel.bufmgr, + "Buffer", + CURBE_TOTAL_DATA_LENGTH, 64); + assert(bo); + media_state->curbe.bo = bo; + + /* surface state */ + for (i = 0; i < MAX_MEDIA_SURFACES_GEN6; i++) { + dri_bo_unreference(media_state->surface_state[i].bo); + media_state->surface_state[i].bo = NULL; + } + + /* binding table */ + dri_bo_unreference(media_state->binding_table.bo); + bo = dri_bo_alloc(i965->intel.bufmgr, + "Buffer", + MAX_MEDIA_SURFACES_GEN6 * sizeof(unsigned int), 32); + assert(bo); + media_state->binding_table.bo = bo; + + /* interface descriptor remapping table */ + dri_bo_unreference(media_state->idrt.bo); + bo = dri_bo_alloc(i965->intel.bufmgr, + "Buffer", + MAX_INTERFACE_DESC_GEN6 * sizeof(struct gen6_interface_descriptor_data), 16); + assert(bo); + media_state->idrt.bo = bo; + + /* VME output buffer */ + dri_bo_unreference(media_state->vme_output.bo); + media_state->vme_output.bo = NULL; + + /* VME state */ + dri_bo_unreference(media_state->vme_state.bo); + bo = dri_bo_alloc(i965->intel.bufmgr, + "Buffer", + 1024*16, 64); + assert(bo); + media_state->vme_state.bo = bo; + + media_state->vfe_state.max_num_threads = 60 - 1; + media_state->vfe_state.num_urb_entries = 16; + media_state->vfe_state.gpgpu_mode = 0; + media_state->vfe_state.urb_entry_size = 59 - 1; + media_state->vfe_state.curbe_allocation_size = CURBE_ALLOCATION_SIZE - 1; +} + +static void gen6_vme_pipeline_programing(VADriverContextP ctx, + VAContextID context, + struct mfc_encode_state *encode_state) +{ + struct i965_driver_data *i965 = i965_driver_data(ctx); + struct object_context *obj_context = CONTEXT(context); + int width_in_mbs = (obj_context->picture_width + 15) / 16; + int height_in_mbs = (obj_context->picture_height + 15) / 16; + int x, y; + int emit_new_state = 1, object_len_in_bytes; + VAEncSliceParameterBuffer *pSliceParameter = (VAEncSliceParameterBuffer *)encode_state->slice_params[0]->buffer; + int is_intra = pSliceParameter->slice_flags.bits.is_intra; + + intel_batchbuffer_start_atomic(ctx, 0x1000); + + for(y = 0; y < height_in_mbs; y++){ + for(x = 0; x < width_in_mbs; x++){ + + if (emit_new_state) { + /*Step1: MI_FLUSH/PIPE_CONTROL*/ + BEGIN_BATCH(ctx, 4); + OUT_BATCH(ctx, CMD_PIPE_CONTROL | 0x02); + OUT_BATCH(ctx, 0); + OUT_BATCH(ctx, 0); + OUT_BATCH(ctx, 0); + ADVANCE_BATCH(ctx); + + /*Step2: State command PIPELINE_SELECT*/ + gen6_vme_pipeline_select(ctx); + + /*Step3: State commands configuring pipeline states*/ + gen6_vme_state_base_address(ctx); + gen6_vme_vfe_state(ctx); + gen6_vme_curbe_load(ctx); + gen6_vme_idrt(ctx); + + emit_new_state = 0; + } + + /*Step4: Primitive commands*/ + object_len_in_bytes = gen6_vme_media_object(ctx, context, encode_state, x, y, is_intra ? VME_INTRA_SHADER : VME_INTER_SHADER); + + if (intel_batchbuffer_check_free_space(ctx, object_len_in_bytes) == 0) { + intel_batchbuffer_end_atomic(ctx); + intel_batchbuffer_flush(ctx); + emit_new_state = 1; + intel_batchbuffer_start_atomic(ctx, 0x1000); + } + } + } + + intel_batchbuffer_end_atomic(ctx); +} + +static VAStatus gen6_vme_prepare(VADriverContextP ctx, + VAContextID context, + struct mfc_encode_state *encode_state) +{ + VAStatus vaStatus = VA_STATUS_SUCCESS; + VAEncSliceParameterBuffer *pSliceParameter = (VAEncSliceParameterBuffer *)encode_state->slice_params[0]->buffer; + int is_intra = pSliceParameter->slice_flags.bits.is_intra; + + /*Setup all the memory object*/ + gen6_vme_surface_setup(ctx, context, encode_state, is_intra); + gen6_vme_interface_setup(ctx, context, encode_state); + gen6_vme_constant_setup(ctx, context, encode_state); + gen6_vme_vme_state_setup(ctx, context, encode_state, is_intra); + + /*Programing media pipeline*/ + gen6_vme_pipeline_programing(ctx, context, encode_state); + + return vaStatus; +} + +static VAStatus gen6_vme_run(VADriverContextP ctx, + VAContextID context, + struct mfc_encode_state *encode_state) +{ + intel_batchbuffer_flush(ctx); + + return VA_STATUS_SUCCESS; +} + +static VAStatus gen6_vme_stop(VADriverContextP ctx, + VAContextID context, + struct mfc_encode_state *encode_state) +{ + return VA_STATUS_SUCCESS; +} + +VAStatus gen6_vme_media_pipeline(VADriverContextP ctx, + VAContextID context, + struct mfc_encode_state *encode_state) +{ + gen6_vme_media_init(ctx); + gen6_vme_prepare(ctx, context, encode_state); + gen6_vme_run(ctx, context, encode_state); + gen6_vme_stop(ctx, context, encode_state); + + return VA_STATUS_SUCCESS; +} + +Bool gen6_vme_init(VADriverContextP ctx) +{ + struct i965_driver_data *i965 = i965_driver_data(ctx); + int i; + + for (i = 0; i < GEN6_VME_KERNEL_NUMBER; i++) { + /*Load kernel into GPU memory*/ + struct media_kernel *kernel = &gen6_vme_kernels[i]; + + kernel->bo = dri_bo_alloc(i965->intel.bufmgr, + kernel->name, + kernel->size, + 0x1000); + assert(kernel->bo); + dri_bo_subdata(kernel->bo, 0, kernel->size, kernel->bin); + } + + return True; +} + +Bool gen6_vme_terminate(VADriverContextP ctx) +{ + int i; + + for (i = 0; i < GEN6_VME_KERNEL_NUMBER; i++) { + /*Load kernel into GPU memory*/ + struct media_kernel *kernel = &gen6_vme_kernels[i]; + + dri_bo_unreference(kernel->bo); + kernel->bo = NULL; + } + + return True; +} |