/* * 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, sublicense, * 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 NONINFRINGEMENT. IN NO EVENT SHALL * THE AUTHORS OR COPYRIGHT HOLDERS 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. * * Author: * Zou Nan hai */ #include "xvmc_vld.h" #include "i965_hwmc.h" #include "i810_reg.h" #include "brw_defines.h" #include "brw_structs.h" #ifndef ALIGN #define ALIGN(m,n) (((m) + (n) - 1) & ~((n) - 1)) #endif #define BATCH_STRUCT(x) intelBatchbufferData(&x, sizeof(x), 0) #define VLD_MAX_SLICE_SIZE (32 * 1024) #define CS_SIZE 30 #define URB_SIZE 384 /* idct table */ #define C0 23170 #define C1 22725 #define C2 21407 #define C3 19266 #define C4 16383 #define C5 12873 #define C6 8867 #define C7 4520 const uint32_t idct_table[] = { C4, C1, C2, C3, C4, C5, C6, C7, //g5 C4, C1, C2, C3, C4, C5, C6, C7, C4, C3, C6, -C7, -C4, -C1, -C2, -C5, C4, C3, C6, -C7, -C4, -C1, -C2, -C5, C4, C5, -C6, -C1, -C4, C7, C2, C3, C4, C5, -C6, -C1, -C4, C7, C2, C3, C4, C7, -C2, -C5, C4, C3, -C6, -C1, C4, C7, -C2, -C5, C4, C3, -C6, -C1, C4, -C7, -C2, C5, C4, -C3, -C6, C1, C4, -C7, -C2, C5, C4, -C3, -C6, C1, C4, -C5, -C6, C1, -C4, -C7, C2, -C3, C4, -C5, -C6, C1, -C4, -C7, C2, -C3, C4, -C3, C6, C7, -C4, C1, -C2, C5, C4, -C3, C6, C7, -C4, C1, -C2, C5, C4, -C1, C2, -C3, C4, -C5, C6, -C7, C4, -C1, C2, -C3, C4, -C5, C6, -C7 //g20 }; #undef C0 #undef C1 #undef C2 #undef C3 #undef C4 #undef C5 #undef C6 #undef C7 #define INTERFACE_NUM 8 enum interface { FRAME_INTRA = 0, FRAME_FRAME_PRED_FORWARD, FRAME_FRAME_PRED_BACKWARD, FRAME_FRAME_PRED_BIDIRECT, FRAME_FIELD_PRED_FORWARD, FRAME_FIELD_PRED_BACKWARD, FRAME_FIELD_PRED_BIDIRECT, LIB_INTERFACE }; /*kernels for vld mode*/ static uint32_t lib_kernel[][4] = { #include "shader/vld/lib.g4b" }; static uint32_t ipicture_kernel[][4] = { #include "shader/vld/ipicture.g4b" }; static uint32_t frame_forward_kernel[][4] = { #include "shader/vld/frame_forward.g4b" }; static uint32_t frame_backward_kernel[][4] = { #include "shader/vld/frame_backward.g4b" }; static uint32_t frame_f_b_kernel[][4] = { #include "shader/vld/frame_f_b.g4b" }; static uint32_t field_forward_kernel[][4] = { #include "shader/vld/field_forward.g4b" }; static uint32_t field_backward_kernel[][4] = { #include "shader/vld/field_backward.g4b" }; static uint32_t field_f_b_kernel[][4] = { #include "shader/vld/field_f_b.g4b" }; /* on IGDNG */ static uint32_t lib_kernel_gen5[][4] = { #include "shader/vld/lib.g4b.gen5" }; static uint32_t ipicture_kernel_gen5[][4] = { #include "shader/vld/ipicture.g4b.gen5" }; static uint32_t frame_forward_kernel_gen5[][4] = { #include "shader/vld/frame_forward.g4b.gen5" }; static uint32_t frame_backward_kernel_gen5[][4] = { #include "shader/vld/frame_backward.g4b.gen5" }; static uint32_t frame_f_b_kernel_gen5[][4] = { #include "shader/vld/frame_f_b.g4b.gen5" }; static uint32_t field_forward_kernel_gen5[][4] = { #include "shader/vld/field_forward.g4b.gen5" }; static uint32_t field_backward_kernel_gen5[][4] = { #include "shader/vld/field_backward.g4b.gen5" }; static uint32_t field_f_b_kernel_gen5[][4] = { #include "shader/vld/field_f_b.g4b.gen5" }; /*kernels for mc mode*/ static uint32_t lib_kernel_idct[][4] = { #include "shader/mc/lib_igd.g4b" }; static uint32_t ipicture_kernel_idct[][4] = { #include "shader/mc/ipicture_igd.g4b" }; static uint32_t frame_forward_kernel_idct[][4] = { #include "shader/mc/frame_forward_igd.g4b" }; static uint32_t frame_backward_kernel_idct[][4] = { #include "shader/mc/frame_backward_igd.g4b" }; static uint32_t frame_f_b_kernel_idct[][4] = { #include "shader/mc/frame_f_b_igd.g4b" }; static uint32_t field_forward_kernel_idct[][4] = { #include "shader/mc/field_forward_igd.g4b" }; static uint32_t field_backward_kernel_idct[][4] = { #include "shader/mc/field_backward_igd.g4b" }; static uint32_t field_f_b_kernel_idct[][4] = { #include "shader/mc/field_f_b_igd.g4b" }; /* on IGDNG */ static uint32_t lib_kernel_idct_gen5[][4] = { #include "shader/mc/lib_igd.g4b.gen5" }; static uint32_t ipicture_kernel_idct_gen5[][4] = { #include "shader/mc/ipicture_igd.g4b.gen5" }; static uint32_t frame_forward_kernel_idct_gen5[][4] = { #include "shader/mc/frame_forward_igd.g4b.gen5" }; static uint32_t frame_backward_kernel_idct_gen5[][4] = { #include "shader/mc/frame_backward_igd.g4b.gen5" }; static uint32_t frame_f_b_kernel_idct_gen5[][4] = { #include "shader/mc/frame_f_b_igd.g4b.gen5" }; static uint32_t field_forward_kernel_idct_gen5[][4] = { #include "shader/mc/field_forward_igd.g4b.gen5" }; static uint32_t field_backward_kernel_idct_gen5[][4] = { #include "shader/mc/field_backward_igd.g4b.gen5" }; static uint32_t field_f_b_kernel_idct_gen5[][4] = { #include "shader/mc/field_f_b_igd.g4b.gen5" }; struct media_kernel { uint32_t(*bin)[4]; int size; }; static struct media_kernel media_kernels[] = { /*kernels for vld mode */ {ipicture_kernel, sizeof(ipicture_kernel)} , {frame_forward_kernel, sizeof(frame_forward_kernel)} , {frame_backward_kernel, sizeof(frame_backward_kernel)} , {frame_f_b_kernel, sizeof(frame_f_b_kernel)} , {field_forward_kernel, sizeof(field_forward_kernel)} , {field_backward_kernel, sizeof(field_backward_kernel)} , {field_f_b_kernel, sizeof(field_f_b_kernel)} , {lib_kernel, sizeof(lib_kernel)} , /*kernels for mc mode */ {ipicture_kernel_idct, sizeof(ipicture_kernel_idct)} , {frame_forward_kernel_idct, sizeof(frame_forward_kernel_idct)} , {frame_backward_kernel_idct, sizeof(frame_backward_kernel_idct)} , {frame_f_b_kernel_idct, sizeof(frame_f_b_kernel_idct)} , {field_forward_kernel_idct, sizeof(field_forward_kernel_idct)} , {field_backward_kernel_idct, sizeof(field_backward_kernel_idct)} , {field_f_b_kernel_idct, sizeof(field_f_b_kernel_idct)} , {lib_kernel_idct, sizeof(lib_kernel_idct)} }; static struct media_kernel media_gen5_kernels[] = { /*kernels for vld mode */ {ipicture_kernel_gen5, sizeof(ipicture_kernel_gen5)} , {frame_forward_kernel_gen5, sizeof(frame_forward_kernel_gen5)} , {frame_backward_kernel_gen5, sizeof(frame_backward_kernel_gen5)} , {frame_f_b_kernel_gen5, sizeof(frame_f_b_kernel_gen5)} , {field_forward_kernel_gen5, sizeof(field_forward_kernel_gen5)} , {field_backward_kernel_gen5, sizeof(field_backward_kernel_gen5)} , {field_f_b_kernel_gen5, sizeof(field_f_b_kernel_gen5)} , {lib_kernel_gen5, sizeof(lib_kernel_gen5)} , /*kernels for mc mode */ {ipicture_kernel_idct_gen5, sizeof(ipicture_kernel_idct_gen5)} , {frame_forward_kernel_idct_gen5, sizeof(frame_forward_kernel_idct_gen5)} , {frame_backward_kernel_idct_gen5, sizeof(frame_backward_kernel_idct_gen5)} , {frame_f_b_kernel_idct_gen5, sizeof(frame_f_b_kernel_idct_gen5)} , {field_forward_kernel_idct_gen5, sizeof(field_forward_kernel_idct_gen5)} , {field_backward_kernel_idct_gen5, sizeof(field_backward_kernel_idct_gen5)} , {field_f_b_kernel_idct_gen5, sizeof(field_f_b_kernel_idct_gen5)} , {lib_kernel_idct_gen5, sizeof(lib_kernel_idct_gen5)} }; #define MEDIA_KERNEL_NUM (sizeof(media_kernels)/sizeof(media_kernels[0])) struct media_kernel_obj { dri_bo *bo; }; struct interface_descriptor_obj { dri_bo *bo; struct media_kernel_obj kernels[MEDIA_KERNEL_NUM]; }; struct vfe_state_obj { dri_bo *bo; struct interface_descriptor_obj interface; }; struct vld_state_obj { dri_bo *bo; }; struct surface_obj { dri_bo *bo; }; struct surface_state_obj { struct surface_obj surface; dri_bo *bo; }; struct binding_table_obj { dri_bo *bo; struct surface_state_obj surface_states[I965_MAX_SURFACES]; }; struct slice_data_obj { dri_bo *bo; }; struct mb_data_obj { dri_bo *bo; }; struct cs_state_obj { dri_bo *bo; }; static struct media_state { struct vfe_state_obj vfe_state; struct vld_state_obj vld_state; struct binding_table_obj binding_table; struct cs_state_obj cs_object; struct slice_data_obj slice_data; struct mb_data_obj mb_data; } media_state; /* XvMCQMatrix * 2 + idct_table + 8 * kernel offset pointer */ #define CS_OBJECT_SIZE (32*20 + sizeof(unsigned int) * 8) static int free_object(struct media_state *s) { int i; #define FREE_ONE_BO(bo) \ if (bo) \ drm_intel_bo_unreference(bo) FREE_ONE_BO(s->vfe_state.bo); FREE_ONE_BO(s->vfe_state.interface.bo); for (i = 0; i < MEDIA_KERNEL_NUM; i++) FREE_ONE_BO(s->vfe_state.interface.kernels[i].bo); FREE_ONE_BO(s->binding_table.bo); for (i = 0; i < I965_MAX_SURFACES; i++) FREE_ONE_BO(s->binding_table.surface_states[i].bo); FREE_ONE_BO(s->slice_data.bo); FREE_ONE_BO(s->mb_data.bo); FREE_ONE_BO(s->cs_object.bo); FREE_ONE_BO(s->vld_state.bo); } static int alloc_object(struct media_state *s) { int i; for (i = 0; i < I965_MAX_SURFACES; i++) { s->binding_table.surface_states[i].bo = drm_intel_bo_alloc(xvmc_driver->bufmgr, "surface_state", sizeof(struct brw_surface_state), 0x1000); if (!s->binding_table.surface_states[i].bo) goto out; } return 0; out: free_object(s); return BadAlloc; } static void flush() { #define FLUSH_STATE_CACHE 1 struct brw_mi_flush f; memset(&f, 0, sizeof(f)); f.opcode = CMD_MI_FLUSH; f.flags = (1 << FLUSH_STATE_CACHE); BATCH_STRUCT(f); } static Status vfe_state(int vfe_mode) { struct brw_vfe_state tmp, *vfe_state = &tmp; memset(vfe_state, 0, sizeof(*vfe_state)); if (vfe_mode == VFE_VLD_MODE) { vfe_state->vfe0.extend_vfe_state_present = 1; } else { vfe_state->vfe0.extend_vfe_state_present = 0; } vfe_state->vfe1.vfe_mode = vfe_mode; vfe_state->vfe1.num_urb_entries = 1; vfe_state->vfe1.children_present = 0; vfe_state->vfe1.urb_entry_alloc_size = 2; vfe_state->vfe1.max_threads = 31; vfe_state->vfe2.interface_descriptor_base = media_state.vfe_state.interface.bo->offset >> 4; if (media_state.vfe_state.bo) drm_intel_bo_unreference(media_state.vfe_state.bo); media_state.vfe_state.bo = drm_intel_bo_alloc(xvmc_driver->bufmgr, "vfe state", sizeof(struct brw_vfe_state), 0x1000); if (!media_state.vfe_state.bo) return BadAlloc; drm_intel_bo_subdata(media_state.vfe_state.bo, 0, sizeof(tmp), &tmp); drm_intel_bo_emit_reloc(media_state.vfe_state.bo, offsetof(struct brw_vfe_state, vfe2), media_state.vfe_state.interface.bo, 0, I915_GEM_DOMAIN_INSTRUCTION, 0); return Success; } static Status interface_descriptor() { int i; struct brw_interface_descriptor tmp, *desc = &tmp; if (media_state.vfe_state.interface.bo) drm_intel_bo_unreference(media_state.vfe_state.interface.bo); media_state.vfe_state.interface.bo = drm_intel_bo_alloc(xvmc_driver->bufmgr, "interfaces", MEDIA_KERNEL_NUM * sizeof(struct brw_interface_descriptor), 0x1000); if (!media_state.vfe_state.interface.bo) return BadAlloc; for (i = 0; i < MEDIA_KERNEL_NUM; i++) { memset(desc, 0, sizeof(*desc)); desc->desc0.grf_reg_blocks = 15; desc->desc0.kernel_start_pointer = media_state.vfe_state.interface.kernels[i].bo->offset >> 6; desc->desc1.const_urb_entry_read_offset = 0; desc->desc1.const_urb_entry_read_len = 30; desc->desc3.binding_table_entry_count = I965_MAX_SURFACES - 1; desc->desc3.binding_table_pointer = media_state.binding_table.bo->offset >> 5; drm_intel_bo_subdata(media_state.vfe_state.interface.bo, i * sizeof(tmp), sizeof(tmp), desc); drm_intel_bo_emit_reloc(media_state.vfe_state.interface.bo, i * sizeof(*desc) + offsetof(struct brw_interface_descriptor, desc0), media_state.vfe_state. interface.kernels[i].bo, desc->desc0.grf_reg_blocks, I915_GEM_DOMAIN_INSTRUCTION, 0); drm_intel_bo_emit_reloc(media_state.vfe_state.interface.bo, i * sizeof(*desc) + offsetof(struct brw_interface_descriptor, desc3), media_state.binding_table.bo, desc->desc3.binding_table_entry_count, I915_GEM_DOMAIN_INSTRUCTION, 0); } return Success; } static int setup_media_kernels(struct i965_xvmc_context *i965_ctx) { int i; assert(MEDIA_KERNEL_NUM == sizeof(media_gen5_kernels) / sizeof(media_gen5_kernels[0])); for (i = 0; i < MEDIA_KERNEL_NUM; i++) { if (i965_ctx->is_igdng) media_state.vfe_state.interface.kernels[i].bo = drm_intel_bo_alloc(xvmc_driver->bufmgr, "kernel", media_gen5_kernels[i].size, 0x1000); else media_state.vfe_state.interface.kernels[i].bo = drm_intel_bo_alloc(xvmc_driver->bufmgr, "kernels", media_kernels[i].size, 0x1000); if (!media_state.vfe_state.interface.kernels[i].bo) goto out; } for (i = 0; i < MEDIA_KERNEL_NUM; i++) { dri_bo *bo = media_state.vfe_state.interface.kernels[i].bo; if (i965_ctx->is_igdng) drm_intel_bo_subdata(bo, 0, media_gen5_kernels[i].size, media_gen5_kernels[i].bin); else drm_intel_bo_subdata(bo, 0, media_kernels[i].size, media_kernels[i].bin); } return 0; out: free_object(&media_state); return BadAlloc; } static Status binding_tables() { unsigned int table[I965_MAX_SURFACES]; int i; if (media_state.binding_table.bo) drm_intel_bo_unreference(media_state.binding_table.bo); media_state.binding_table.bo = drm_intel_bo_alloc(xvmc_driver->bufmgr, "binding_table", I965_MAX_SURFACES * 4, 0x1000); if (!media_state.binding_table.bo) return BadAlloc; for (i = 0; i < I965_MAX_SURFACES; i++) { table[i] = media_state.binding_table.surface_states[i].bo->offset; drm_intel_bo_emit_reloc(media_state.binding_table.bo, i * sizeof(unsigned int), media_state. binding_table.surface_states[i].bo, 0, I915_GEM_DOMAIN_INSTRUCTION, 0); } drm_intel_bo_subdata(media_state.binding_table.bo, 0, sizeof(table), table); return Success; } static Status cs_init(int interface_offset) { char buf[CS_OBJECT_SIZE]; unsigned int *lib_reloc; int i; if (media_state.cs_object.bo) drm_intel_bo_unreference(media_state.cs_object.bo); media_state.cs_object.bo = drm_intel_bo_alloc(xvmc_driver->bufmgr, "cs object", CS_OBJECT_SIZE, 64); if (!media_state.cs_object.bo) return BadAlloc; memcpy(buf + 32 * 4, idct_table, sizeof(idct_table)); /* idct lib reloction */ lib_reloc = (unsigned int *)(buf + 32 * 20); for (i = 0; i < 8; i++) lib_reloc[i] = media_state.vfe_state.interface.kernels[LIB_INTERFACE + interface_offset].bo-> offset; drm_intel_bo_subdata(media_state.cs_object.bo, 32 * 4, 32 * 16 + 8 * sizeof(unsigned int), buf + 32 * 4); for (i = 0; i < 8; i++) drm_intel_bo_emit_reloc(media_state.cs_object.bo, 32 * 20 + sizeof(unsigned int) * i, media_state.vfe_state. interface.kernels[LIB_INTERFACE + interface_offset].bo, 0, I915_GEM_DOMAIN_INSTRUCTION, 0); return Success; } static Status create_context(Display * display, XvMCContext * context, int priv_count, CARD32 * priv_data) { struct i965_xvmc_context *i965_ctx; i965_ctx = (struct i965_xvmc_context *)priv_data; context->privData = priv_data; if (alloc_object(&media_state)) return BadAlloc; if (setup_media_kernels(i965_ctx)) return BadAlloc; return Success; } static Status destroy_context(Display * display, XvMCContext * context) { struct i965_xvmc_context *i965_ctx; i965_ctx = context->privData; Xfree(i965_ctx); return Success; } #define STRIDE(w) (w) #define SIZE_YUV420(w, h) (h * (STRIDE(w) + STRIDE(w >> 1))) static Status create_surface(Display * display, XvMCContext * context, XvMCSurface * surface, int priv_count, CARD32 * priv_data) { struct i965_xvmc_surface *priv_surface = (struct i965_xvmc_surface *)priv_data; size_t size = SIZE_YUV420(priv_surface->w, priv_surface->h); surface->privData = priv_data; priv_surface->bo = drm_intel_bo_alloc(xvmc_driver->bufmgr, "surface", size, 0x1000); return Success; } static Status destroy_surface(Display * display, XvMCSurface * surface) { struct i965_xvmc_surface *priv_surface = surface->privData; XSync(display, False); drm_intel_bo_unreference(priv_surface->bo); return Success; } static Status load_qmatrix(Display * display, XvMCContext * context, const XvMCQMatrix * qmx) { Status ret; ret = cs_init(0); if (ret != Success) return ret; drm_intel_bo_subdata(media_state.cs_object.bo, 0, 64, qmx->intra_quantiser_matrix); drm_intel_bo_subdata(media_state.cs_object.bo, 64, 64, qmx->non_intra_quantiser_matrix); return Success; } static Status get_surface_status(Display * display, XvMCSurface * surface, int *status) { *status = 0; return Success; } static Status vld_state(const XvMCMpegControl * control) { struct brw_vld_state tmp, *vld = &tmp; if (media_state.vld_state.bo) drm_intel_bo_unreference(media_state.vld_state.bo); media_state.vld_state.bo = drm_intel_bo_alloc(xvmc_driver->bufmgr, "vld state", sizeof(struct brw_vld_state), 64); if (!media_state.vld_state.bo) return BadAlloc; memset(vld, 0, sizeof(*vld)); vld->vld0.f_code_0_0 = control->FHMV_range + 1; vld->vld0.f_code_0_1 = control->FVMV_range + 1; vld->vld0.f_code_1_0 = control->BHMV_range + 1; vld->vld0.f_code_1_1 = control->BVMV_range + 1; vld->vld0.intra_dc_precision = control->intra_dc_precision; vld->vld0.picture_structure = control->picture_structure; vld->vld0.top_field_first = !!(control->flags & XVMC_TOP_FIELD_FIRST); vld->vld0.frame_predict_frame_dct = !!(control->flags & XVMC_PRED_DCT_FRAME); vld->vld0.concealment_motion_vector = !!(control->flags & XVMC_CONCEALMENT_MOTION_VECTORS); vld->vld0.quantizer_scale_type = !!(control->flags & XVMC_Q_SCALE_TYPE); vld->vld0.intra_vlc_format = !!(control->flags & XVMC_INTRA_VLC_FORMAT); vld->vld0.scan_order = !!(control->flags & XVMC_ALTERNATE_SCAN); vld->vld1.picture_coding_type = control->picture_coding_type; vld->desc_remap_table0.index_0 = FRAME_INTRA; vld->desc_remap_table0.index_1 = FRAME_FRAME_PRED_FORWARD; vld->desc_remap_table0.index_2 = FRAME_FIELD_PRED_FORWARD; vld->desc_remap_table0.index_3 = FRAME_FIELD_PRED_BIDIRECT; /* dual prime */ vld->desc_remap_table0.index_4 = FRAME_FRAME_PRED_BACKWARD; vld->desc_remap_table0.index_5 = FRAME_FIELD_PRED_BACKWARD; vld->desc_remap_table0.index_6 = FRAME_FRAME_PRED_BIDIRECT; vld->desc_remap_table0.index_7 = FRAME_FIELD_PRED_BIDIRECT; vld->desc_remap_table1.index_8 = FRAME_INTRA; vld->desc_remap_table1.index_9 = FRAME_FRAME_PRED_FORWARD; vld->desc_remap_table1.index_10 = FRAME_FIELD_PRED_FORWARD; vld->desc_remap_table1.index_11 = FRAME_FIELD_PRED_BIDIRECT; vld->desc_remap_table1.index_12 = FRAME_FRAME_PRED_BACKWARD; vld->desc_remap_table1.index_13 = FRAME_FIELD_PRED_BACKWARD; vld->desc_remap_table1.index_14 = FRAME_FRAME_PRED_BIDIRECT; vld->desc_remap_table1.index_15 = FRAME_FIELD_PRED_BIDIRECT; drm_intel_bo_subdata(media_state.vld_state.bo, 0, sizeof(tmp), vld); return Success; } static Status setup_media_surface(int index, dri_bo * bo, unsigned long offset, int w, int h, Bool write) { struct brw_surface_state tmp, *ss = &tmp; memset(ss, 0, sizeof(*ss)); ss->ss0.surface_type = BRW_SURFACE_2D; ss->ss0.surface_format = BRW_SURFACEFORMAT_R8_SINT; ss->ss1.base_addr = offset + bo->offset; ss->ss2.width = w - 1; ss->ss2.height = h - 1; ss->ss3.pitch = w - 1; if (media_state.binding_table.surface_states[index].bo) drm_intel_bo_unreference(media_state. binding_table.surface_states[index]. bo); media_state.binding_table.surface_states[index].bo = drm_intel_bo_alloc(xvmc_driver->bufmgr, "surface_state", sizeof(struct brw_surface_state), 0x1000); if (!media_state.binding_table.surface_states[index].bo) return BadAlloc; drm_intel_bo_subdata(media_state.binding_table.surface_states[index].bo, 0, sizeof(*ss), ss); drm_intel_bo_emit_reloc(media_state.binding_table. surface_states[index].bo, offsetof(struct brw_surface_state, ss1), bo, offset, I915_GEM_DOMAIN_RENDER, write ? I915_GEM_DOMAIN_RENDER : 0); return Success; } static Status setup_surface(struct i965_xvmc_surface *target, struct i965_xvmc_surface *past, struct i965_xvmc_surface *future, int w, int h) { Status ret; ret = setup_media_surface(0, target->bo, 0, w, h, TRUE); if (ret != Success) return ret; ret = setup_media_surface(1, target->bo, w * h, w / 2, h / 2, TRUE); if (ret != Success) return ret; ret = setup_media_surface(2, target->bo, w * h + w * h / 4, w / 2, h / 2, TRUE); if (ret != Success) return ret; if (past) { ret = setup_media_surface(4, past->bo, 0, w, h, FALSE); if (ret != Success) return ret; ret = setup_media_surface(5, past->bo, w * h, w / 2, h / 2, FALSE); if (ret != Success) return ret; ret = setup_media_surface(6, past->bo, w * h + w * h / 4, w / 2, h / 2, FALSE); if (ret != Success) return ret; } if (future) { ret = setup_media_surface(7, future->bo, 0, w, h, FALSE); if (ret != Success) return ret; ret = setup_media_surface(8, future->bo, w * h, w / 2, h / 2, FALSE); if (ret != Success) return ret; ret = setup_media_surface(9, future->bo, w * h + w * h / 4, w / 2, h / 2, FALSE); if (ret != Success) return ret; } return Success; } static Status begin_surface(Display * display, XvMCContext * context, XvMCSurface * target, XvMCSurface * past, XvMCSurface * future, const XvMCMpegControl * control) { struct i965_xvmc_contex *i965_ctx; struct i965_xvmc_surface *priv_target, *priv_past, *priv_future; intel_xvmc_context_ptr intel_ctx; Status ret; intel_ctx = intel_xvmc_find_context(context->context_id); priv_target = target->privData; priv_past = past ? past->privData : NULL; priv_future = future ? future->privData : NULL; ret = vld_state(control); if (ret != Success) return ret; ret = setup_surface(priv_target, priv_past, priv_future, context->width, context->height); if (ret != Success) return ret; ret = binding_tables(); if (ret != Success) return ret; ret = interface_descriptor(); if (ret != Success) return ret; ret = vfe_state(VFE_VLD_MODE); if (ret != Success) return ret; LOCK_HARDWARE(intel_ctx->hw_context); flush(); UNLOCK_HARDWARE(intel_ctx->hw_context); return Success; } static Status put_slice(Display * display, XvMCContext * context, unsigned char *slice, int nbytes) { return Success; } static void state_base_address(struct i965_xvmc_context *i965_ctx) { BATCH_LOCALS; if (i965_ctx->is_igdng) { BEGIN_BATCH(8); OUT_BATCH(BRW_STATE_BASE_ADDRESS | 6); OUT_BATCH(0 | BASE_ADDRESS_MODIFY); OUT_BATCH(0 | BASE_ADDRESS_MODIFY); OUT_BATCH(0 | BASE_ADDRESS_MODIFY); OUT_BATCH(0 | BASE_ADDRESS_MODIFY); OUT_BATCH(0 | BASE_ADDRESS_MODIFY); OUT_BATCH(0 | BASE_ADDRESS_MODIFY); OUT_BATCH(0 | BASE_ADDRESS_MODIFY); ADVANCE_BATCH(); } else { BEGIN_BATCH(6); OUT_BATCH(BRW_STATE_BASE_ADDRESS | 4); OUT_BATCH(0 | BASE_ADDRESS_MODIFY); OUT_BATCH(0 | BASE_ADDRESS_MODIFY); OUT_BATCH(0 | BASE_ADDRESS_MODIFY); OUT_BATCH(0 | BASE_ADDRESS_MODIFY); OUT_BATCH(0 | BASE_ADDRESS_MODIFY); ADVANCE_BATCH(); } } static void pipeline_select() { BATCH_LOCALS; BEGIN_BATCH(1); OUT_BATCH(NEW_PIPELINE_SELECT | PIPELINE_SELECT_MEDIA); ADVANCE_BATCH(); } static void media_state_pointers(int vfe_mode) { BATCH_LOCALS; BEGIN_BATCH(3); OUT_BATCH(BRW_MEDIA_STATE_POINTERS | 1); if (vfe_mode == VFE_VLD_MODE) OUT_RELOC(media_state.vld_state.bo, I915_GEM_DOMAIN_INSTRUCTION, 0, 1); else OUT_BATCH(0); OUT_RELOC(media_state.vfe_state.bo, I915_GEM_DOMAIN_INSTRUCTION, 0, 0); ADVANCE_BATCH(); } static void align_urb_fence() { BATCH_LOCALS; int i, offset_to_next_cacheline; unsigned long batch_offset; BEGIN_BATCH(3); batch_offset = (void *)batch_ptr - xvmc_driver->alloc.ptr; offset_to_next_cacheline = ALIGN(batch_offset, 64) - batch_offset; if (offset_to_next_cacheline <= 12 && offset_to_next_cacheline != 0) { for (i = 0; i < offset_to_next_cacheline / 4; i++) OUT_BATCH(0); ADVANCE_BATCH(); } } static void urb_layout() { BATCH_LOCALS; align_urb_fence(); BEGIN_BATCH(3); OUT_BATCH(BRW_URB_FENCE | UF0_VFE_REALLOC | UF0_CS_REALLOC | UF0_SF_REALLOC | UF0_CLIP_REALLOC | UF0_GS_REALLOC | UF0_VS_REALLOC | 1); OUT_BATCH((0 << UF1_CLIP_FENCE_SHIFT) | (0 << UF1_GS_FENCE_SHIFT) | (0 << UF1_VS_FENCE_SHIFT)); OUT_BATCH((0 << UF2_CS_FENCE_SHIFT) | (0 << UF2_SF_FENCE_SHIFT) | ((URB_SIZE - CS_SIZE - 1) << UF2_VFE_FENCE_SHIFT) | /* VFE_SIZE */ ((URB_SIZE) << UF2_CS_FENCE_SHIFT)); /* CS_SIZE */ ADVANCE_BATCH(); } static void cs_urb_layout() { BATCH_LOCALS; BEGIN_BATCH(2); OUT_BATCH(BRW_CS_URB_STATE | 0); OUT_BATCH((CS_SIZE << 4) | /* URB Entry Allocation Size */ (1 << 0)); /* Number of URB Entries */ ADVANCE_BATCH(); } static void cs_buffer() { BATCH_LOCALS; BEGIN_BATCH(2); OUT_BATCH(BRW_CONSTANT_BUFFER | 0 | (1 << 8)); OUT_RELOC(media_state.cs_object.bo, I915_GEM_DOMAIN_INSTRUCTION, 0, CS_SIZE); ADVANCE_BATCH(); } /* kick media object to gpu in idct mode*/ static void send_media_object(XvMCMacroBlock * mb, dri_bo * bo, uint32_t offset, enum interface interface) { BATCH_LOCALS; BEGIN_BATCH(13); OUT_BATCH(BRW_MEDIA_OBJECT | 11); OUT_BATCH(interface); OUT_BATCH(6 * 128); OUT_RELOC(bo, I915_GEM_DOMAIN_INSTRUCTION, 0, offset); OUT_BATCH(mb->x << 4); OUT_BATCH(mb->y << 4); OUT_RELOC(bo, I915_GEM_DOMAIN_INSTRUCTION, 0, offset); OUT_BATCH_SHORT(mb->coded_block_pattern); OUT_BATCH_SHORT(mb->PMV[0][0][0]); OUT_BATCH_SHORT(mb->PMV[0][0][1]); OUT_BATCH_SHORT(mb->PMV[0][1][0]); OUT_BATCH_SHORT(mb->PMV[0][1][1]); OUT_BATCH_SHORT(mb->PMV[1][0][0]); OUT_BATCH_SHORT(mb->PMV[1][0][1]); OUT_BATCH_SHORT(mb->PMV[1][1][0]); OUT_BATCH_SHORT(mb->PMV[1][1][1]); OUT_BATCH_CHAR(mb->dct_type); OUT_BATCH_CHAR(mb->motion_vertical_field_select); OUT_BATCH(0xffffffff); ADVANCE_BATCH(); } /* kick media object to gpu in vld mode*/ static void vld_send_media_object(dri_bo * bo, int slice_len, int mb_h_pos, int mb_v_pos, int mb_bit_offset, int mb_count, int q_scale_code) { BATCH_LOCALS; BEGIN_BATCH(6); OUT_BATCH(BRW_MEDIA_OBJECT | 4); OUT_BATCH(0); OUT_BATCH(slice_len); OUT_RELOC(bo, I915_GEM_DOMAIN_INSTRUCTION, 0, 0); OUT_BATCH((mb_h_pos << 24) | (mb_v_pos << 16) | (mb_count << 8) | (mb_bit_offset)); OUT_BATCH(q_scale_code << 24); ADVANCE_BATCH(); } static Status put_slice2(Display * display, XvMCContext * context, unsigned char *slice, int nbytes, int sliceCode) { unsigned int bit_buf; intel_xvmc_context_ptr intel_ctx; struct i965_xvmc_context *i965_ctx; int q_scale_code, mb_row; i965_ctx = (struct i965_xvmc_context *)context->privData; mb_row = *(slice - 1) - 1; bit_buf = (slice[0] << 24) | (slice[1] << 16) | (slice[2] << 8) | (slice[3]); q_scale_code = bit_buf >> 27; if (media_state.slice_data.bo) { drm_intel_gem_bo_unmap_gtt(media_state.slice_data.bo); drm_intel_bo_unreference(media_state.slice_data.bo); } media_state.slice_data.bo = drm_intel_bo_alloc(xvmc_driver->bufmgr, "slice data", VLD_MAX_SLICE_SIZE, 64); if (!media_state.slice_data.bo) return BadAlloc; drm_intel_gem_bo_map_gtt(media_state.slice_data.bo); memcpy(media_state.slice_data.bo->virtual, slice, nbytes); intel_ctx = intel_xvmc_find_context(context->context_id); LOCK_HARDWARE(intel_ctx->hw_context); state_base_address(i965_ctx); pipeline_select(); media_state_pointers(VFE_VLD_MODE); urb_layout(); cs_urb_layout(); cs_buffer(); vld_send_media_object(media_state.slice_data.bo, nbytes, 0, mb_row, 6, 127, q_scale_code); intelFlushBatch(TRUE); UNLOCK_HARDWARE(intel_ctx->hw_context); return Success; } static Status put_surface(Display * display, XvMCSurface * surface, Drawable draw, short srcx, short srcy, unsigned short srcw, unsigned short srch, short destx, short desty, unsigned short destw, unsigned short desth, int flags, struct intel_xvmc_command *data) { struct i965_xvmc_surface *private_surface = surface->privData; uint32_t handle; drm_intel_bo_flink(private_surface->bo, &handle); data->handle = handle; return Success; } static Status render_surface(Display * display, XvMCContext * context, unsigned int picture_structure, XvMCSurface * target_surface, XvMCSurface * past_surface, XvMCSurface * future_surface, unsigned int flags, unsigned int num_macroblocks, unsigned int first_macroblock, XvMCMacroBlockArray * macroblock_array, XvMCBlockArray * blocks) { struct i965_xvmc_surface *priv_target, *priv_past, *priv_future; intel_xvmc_context_ptr intel_ctx; XvMCMacroBlock *mb; Status ret; unsigned short *block_ptr; int i, j; int block_offset = 0; struct i965_xvmc_context *i965_ctx; intel_ctx = intel_xvmc_find_context(context->context_id); if (!intel_ctx) { XVMC_ERR("Can't find intel xvmc context\n"); return BadValue; } i965_ctx = (struct i965_xvmc_context *)context->privData; priv_target = target_surface->privData; priv_past = past_surface ? past_surface->privData : NULL; priv_future = future_surface ? future_surface->privData : NULL; ret = setup_surface(priv_target, priv_past, priv_future, context->width, context->height); if (ret != Success) return ret; ret = binding_tables(); if (ret != Success) return ret; ret = interface_descriptor(); if (ret != Success) return ret; ret = cs_init(INTERFACE_NUM); if (ret != Success) return ret; ret = vfe_state(VFE_GENERIC_MODE); if (ret != Success) return ret; if (media_state.mb_data.bo) { drm_intel_gem_bo_unmap_gtt(media_state.mb_data.bo); drm_intel_bo_unreference(media_state.mb_data.bo); } unsigned int block_num = (((context->width + 15) >> 4) * ((context->height + 15) >> 4)); unsigned int surface_size = (64 * sizeof(short) * 6 * block_num); media_state.mb_data.bo = drm_intel_bo_alloc(xvmc_driver->bufmgr, "macroblock data", surface_size, 64); if (!media_state.mb_data.bo) return BadAlloc; drm_intel_gem_bo_map_gtt(media_state.mb_data.bo); block_ptr = media_state.mb_data.bo->virtual; unsigned short *mb_block_ptr; for (i = first_macroblock; i < num_macroblocks + first_macroblock; i++) { mb = ¯oblock_array->macro_blocks[i]; mb_block_ptr = &blocks->blocks[(mb->index << 6)]; if (mb->coded_block_pattern & 0x20) { for (j = 0; j < 8; j++) memcpy(block_ptr + 16 * j, mb_block_ptr + 8 * j, 16); mb_block_ptr += 64; } if (mb->coded_block_pattern & 0x10) { for (j = 0; j < 8; j++) memcpy(block_ptr + 16 * j + 8, mb_block_ptr + 8 * j, 16); mb_block_ptr += 64; } block_ptr += 2 * 64; if (mb->coded_block_pattern & 0x08) { for (j = 0; j < 8; j++) memcpy(block_ptr + 16 * j, mb_block_ptr + 8 * j, 16); mb_block_ptr += 64; } if (mb->coded_block_pattern & 0x04) { for (j = 0; j < 8; j++) memcpy(block_ptr + 16 * j + 8, mb_block_ptr + 8 * j, 16); mb_block_ptr += 64; } block_ptr += 2 * 64; if (mb->coded_block_pattern & 0x2) { memcpy(block_ptr, mb_block_ptr, 128); mb_block_ptr += 64; } block_ptr += 64; if (mb->coded_block_pattern & 0x1) memcpy(block_ptr, mb_block_ptr, 128); block_ptr += 64; } LOCK_HARDWARE(intel_ctx->hw_context); state_base_address(i965_ctx); flush(); pipeline_select(); urb_layout(); media_state_pointers(VFE_GENERIC_MODE); cs_urb_layout(); cs_buffer(); for (i = first_macroblock; i < num_macroblocks + first_macroblock; i++, block_offset += 128 * 6) { mb = ¯oblock_array->macro_blocks[i]; if (mb->macroblock_type & XVMC_MB_TYPE_INTRA) { send_media_object(mb, media_state.mb_data.bo, block_offset, FRAME_INTRA + INTERFACE_NUM); } else { if (((mb->motion_type & 3) == XVMC_PREDICTION_FRAME)) { if ((mb->macroblock_type & XVMC_MB_TYPE_MOTION_FORWARD)) { if ((mb->macroblock_type & XVMC_MB_TYPE_MOTION_BACKWARD)) { send_media_object(mb, media_state.mb_data. bo, block_offset, FRAME_FRAME_PRED_BIDIRECT + INTERFACE_NUM); } else { send_media_object(mb, media_state.mb_data. bo, block_offset, FRAME_FRAME_PRED_FORWARD + INTERFACE_NUM); } } else if ((mb->macroblock_type & XVMC_MB_TYPE_MOTION_BACKWARD)) { send_media_object(mb, media_state. mb_data.bo, block_offset, FRAME_FRAME_PRED_BACKWARD + INTERFACE_NUM); } } else if ((mb->motion_type & 3) == XVMC_PREDICTION_FIELD) { if ((mb->macroblock_type & XVMC_MB_TYPE_MOTION_FORWARD)) { if (((mb->macroblock_type & XVMC_MB_TYPE_MOTION_BACKWARD))) { send_media_object(mb, media_state.mb_data. bo, block_offset, FRAME_FIELD_PRED_BIDIRECT + INTERFACE_NUM); } else { send_media_object(mb, media_state.mb_data. bo, block_offset, FRAME_FIELD_PRED_FORWARD + INTERFACE_NUM); } } else if ((mb->macroblock_type & XVMC_MB_TYPE_MOTION_BACKWARD)) { send_media_object(mb, media_state. mb_data.bo, block_offset, FRAME_FIELD_PRED_BACKWARD + INTERFACE_NUM); } } else { send_media_object(mb, media_state.mb_data.bo, block_offset, FRAME_FIELD_PRED_BIDIRECT + INTERFACE_NUM); /*dual prime */ } } } intelFlushBatch(TRUE); UNLOCK_HARDWARE(intel_ctx->hw_context); return Success; } struct _intel_xvmc_driver xvmc_vld_driver = { .type = XVMC_I965_MPEG2_VLD, .create_context = create_context, .destroy_context = destroy_context, .create_surface = create_surface, .destroy_surface = destroy_surface, .load_qmatrix = load_qmatrix, .get_surface_status = get_surface_status, .begin_surface = begin_surface, .render_surface = render_surface, .put_surface = put_surface, .put_slice = put_slice, .put_slice2 = put_slice2 };