/* * Copyright (C) 2017 Rob Clark * Copyright © 2018 Google, Inc. * * 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. * * Authors: * Rob Clark */ #include "util/u_dump.h" #include "freedreno_blitter.h" #include "freedreno_fence.h" #include "freedreno_resource.h" #include "fd6_blitter.h" #include "fd6_format.h" #include "fd6_emit.h" /* Make sure none of the requested dimensions extend beyond the size of the * resource. Not entirely sure why this happens, but sometimes it does, and * w/ 2d blt doesn't have wrap modes like a sampler, so force those cases * back to u_blitter */ static bool ok_dims(const struct pipe_resource *r, const struct pipe_box *b, int lvl) { int last_layer = r->target == PIPE_TEXTURE_3D ? u_minify(r->depth0, lvl) : r->array_size; return (b->x >= 0) && (b->x + b->width <= u_minify(r->width0, lvl)) && (b->y >= 0) && (b->y + b->height <= u_minify(r->height0, lvl)) && (b->z >= 0) && (b->z + b->depth <= last_layer); } static bool ok_format(enum pipe_format pfmt) { enum a6xx_color_fmt fmt = fd6_pipe2color(pfmt); if (fmt == ~0) return false; if (fd6_ifmt(fmt) == 0) return false; return true; } #define DEBUG_BLIT_FALLBACK 0 #define fail_if(cond) \ do { \ if (cond) { \ if (DEBUG_BLIT_FALLBACK) { \ fprintf(stderr, "falling back: %s for blit:\n", #cond); \ util_dump_blit_info(stderr, info); \ fprintf(stderr, "\nsrc: "); \ util_dump_resource(stderr, info->src.resource); \ fprintf(stderr, "\ndst: "); \ util_dump_resource(stderr, info->dst.resource); \ fprintf(stderr, "\n"); \ } \ return false; \ } \ } while (0) static bool can_do_blit(const struct pipe_blit_info *info) { /* I think we can do scaling, but not in z dimension since that would * require blending.. */ fail_if(info->dst.box.depth != info->src.box.depth); /* We can blit if both or neither formats are compressed formats... */ fail_if(util_format_is_compressed(info->src.format) != util_format_is_compressed(info->src.format)); /* Fail if unsupported format: */ fail_if(!ok_format(info->src.format)); fail_if(!ok_format(info->dst.format)); /* ... but only if they're the same compression format. */ fail_if(util_format_is_compressed(info->src.format) && info->src.format != info->dst.format); /* hw ignores {SRC,DST}_INFO.COLOR_SWAP if {SRC,DST}_INFO.TILE_MODE * is set (not linear). We can kind of get around that when tiling/ * untiling by setting both src and dst COLOR_SWAP=WZYX, but that * means the formats must match: */ fail_if((fd_resource(info->dst.resource)->tile_mode || fd_resource(info->src.resource)->tile_mode) && info->dst.format != info->src.format); /* src box can be inverted, which we don't support.. dst box cannot: */ fail_if((info->src.box.width < 0) || (info->src.box.height < 0)); fail_if(!ok_dims(info->src.resource, &info->src.box, info->src.level)); fail_if(!ok_dims(info->dst.resource, &info->dst.box, info->dst.level)); debug_assert(info->dst.box.width >= 0); debug_assert(info->dst.box.height >= 0); debug_assert(info->dst.box.depth >= 0); fail_if(info->dst.resource->nr_samples + info->src.resource->nr_samples > 2); fail_if(info->window_rectangle_include); fail_if(info->render_condition_enable); fail_if(info->alpha_blend); fail_if(info->mask != util_format_get_mask(info->src.format)); fail_if(info->mask != util_format_get_mask(info->dst.format)); return true; } static void emit_setup(struct fd_ringbuffer *ring) { OUT_PKT7(ring, CP_EVENT_WRITE, 1); OUT_RING(ring, PC_CCU_INVALIDATE_COLOR); OUT_PKT7(ring, CP_EVENT_WRITE, 1); OUT_RING(ring, LRZ_FLUSH); OUT_PKT7(ring, CP_SKIP_IB2_ENABLE_GLOBAL, 1); OUT_RING(ring, 0x0); OUT_WFI5(ring); OUT_PKT4(ring, REG_A6XX_RB_CCU_CNTL, 1); OUT_RING(ring, 0x10000000); } static uint32_t blit_control(enum a6xx_color_fmt fmt) { unsigned blit_cntl = 0xf00000; blit_cntl |= A6XX_RB_2D_BLIT_CNTL_COLOR_FORMAT(fmt); blit_cntl |= A6XX_RB_2D_BLIT_CNTL_IFMT(fd6_ifmt(fmt)); return blit_cntl; } /* buffers need to be handled specially since x/width can exceed the bounds * supported by hw.. if necessary decompose into (potentially) two 2D blits */ static void emit_blit_buffer(struct fd_ringbuffer *ring, const struct pipe_blit_info *info) { const struct pipe_box *sbox = &info->src.box; const struct pipe_box *dbox = &info->dst.box; struct fd_resource *src, *dst; unsigned sshift, dshift; if (DEBUG_BLIT_FALLBACK) { fprintf(stderr, "buffer blit: "); util_dump_blit_info(stderr, info); fprintf(stderr, "\ndst resource: "); util_dump_resource(stderr, info->dst.resource); fprintf(stderr, "\nsrc resource: "); util_dump_resource(stderr, info->src.resource); fprintf(stderr, "\n"); } src = fd_resource(info->src.resource); dst = fd_resource(info->dst.resource); debug_assert(src->cpp == 1); debug_assert(dst->cpp == 1); debug_assert(info->src.resource->format == info->dst.resource->format); debug_assert((sbox->y == 0) && (sbox->height == 1)); debug_assert((dbox->y == 0) && (dbox->height == 1)); debug_assert((sbox->z == 0) && (sbox->depth == 1)); debug_assert((dbox->z == 0) && (dbox->depth == 1)); debug_assert(sbox->width == dbox->width); debug_assert(info->src.level == 0); debug_assert(info->dst.level == 0); /* * Buffers can have dimensions bigger than max width, remap into * multiple 1d blits to fit within max dimension * * Note that blob uses .ARRAY_PITCH=128 for blitting buffers, which * seems to prevent overfetch related faults. Not quite sure what * the deal is there. * * Low 6 bits of SRC/DST addresses need to be zero (ie. address * aligned to 64) so we need to shift src/dst x1/x2 to make up the * difference. On top of already splitting up the blit so width * isn't > 16k. * * We perhaps could do a bit better, if src and dst are aligned but * in the worst case this means we have to split the copy up into * 16k (0x4000) minus 64 (0x40). */ sshift = sbox->x & 0x3f; dshift = dbox->x & 0x3f; OUT_PKT7(ring, CP_SET_MARKER, 1); OUT_RING(ring, A2XX_CP_SET_MARKER_0_MODE(RM6_BLIT2DSCALE)); uint32_t blit_cntl = blit_control(RB6_R8_UNORM) | 0x20000000; OUT_PKT4(ring, REG_A6XX_RB_2D_BLIT_CNTL, 1); OUT_RING(ring, blit_cntl); OUT_PKT4(ring, REG_A6XX_GRAS_2D_BLIT_CNTL, 1); OUT_RING(ring, blit_cntl); for (unsigned off = 0; off < sbox->width; off += (0x4000 - 0x40)) { unsigned soff, doff, w, p; soff = (sbox->x + off) & ~0x3f; doff = (dbox->x + off) & ~0x3f; w = MIN2(sbox->width - off, (0x4000 - 0x40)); p = align(w, 64); debug_assert((soff + w) <= fd_bo_size(src->bo)); debug_assert((doff + w) <= fd_bo_size(dst->bo)); /* * Emit source: */ OUT_PKT4(ring, REG_A6XX_SP_PS_2D_SRC_INFO, 13); OUT_RING(ring, A6XX_SP_PS_2D_SRC_INFO_COLOR_FORMAT(RB6_R8_UNORM) | A6XX_SP_PS_2D_SRC_INFO_TILE_MODE(TILE6_LINEAR) | A6XX_SP_PS_2D_SRC_INFO_COLOR_SWAP(WZYX) | 0x500000); OUT_RING(ring, A6XX_SP_PS_2D_SRC_SIZE_WIDTH(sshift + w) | A6XX_SP_PS_2D_SRC_SIZE_HEIGHT(1)); /* SP_PS_2D_SRC_SIZE */ OUT_RELOC(ring, src->bo, soff, 0, 0); /* SP_PS_2D_SRC_LO/HI */ OUT_RING(ring, A6XX_SP_PS_2D_SRC_PITCH_PITCH(p)); OUT_RING(ring, 0x00000000); OUT_RING(ring, 0x00000000); OUT_RING(ring, 0x00000000); OUT_RING(ring, 0x00000000); OUT_RING(ring, 0x00000000); OUT_RING(ring, 0x00000000); OUT_RING(ring, 0x00000000); OUT_RING(ring, 0x00000000); /* * Emit destination: */ OUT_PKT4(ring, REG_A6XX_RB_2D_DST_INFO, 9); OUT_RING(ring, A6XX_RB_2D_DST_INFO_COLOR_FORMAT(RB6_R8_UNORM) | A6XX_RB_2D_DST_INFO_TILE_MODE(TILE6_LINEAR) | A6XX_RB_2D_DST_INFO_COLOR_SWAP(WZYX)); OUT_RELOC(ring, dst->bo, doff, 0, 0); /* RB_2D_DST_LO/HI */ OUT_RING(ring, A6XX_RB_2D_DST_SIZE_PITCH(p)); OUT_RING(ring, 0x00000000); OUT_RING(ring, 0x00000000); OUT_RING(ring, 0x00000000); OUT_RING(ring, 0x00000000); OUT_RING(ring, 0x00000000); /* * Blit command: */ OUT_PKT4(ring, REG_A6XX_GRAS_2D_SRC_TL_X, 4); OUT_RING(ring, A6XX_GRAS_2D_SRC_TL_X_X(sshift)); OUT_RING(ring, A6XX_GRAS_2D_SRC_BR_X_X(sshift + w - 1)); OUT_RING(ring, A6XX_GRAS_2D_SRC_TL_Y_Y(0)); OUT_RING(ring, A6XX_GRAS_2D_SRC_BR_Y_Y(0)); OUT_PKT4(ring, REG_A6XX_GRAS_2D_DST_TL, 2); OUT_RING(ring, A6XX_GRAS_2D_DST_TL_X(dshift) | A6XX_GRAS_2D_DST_TL_Y(0)); OUT_RING(ring, A6XX_GRAS_2D_DST_BR_X(dshift + w - 1) | A6XX_GRAS_2D_DST_BR_Y(0)); OUT_PKT7(ring, CP_EVENT_WRITE, 1); OUT_RING(ring, 0x3f); OUT_WFI5(ring); OUT_PKT4(ring, REG_A6XX_RB_UNKNOWN_8C01, 1); OUT_RING(ring, 0); OUT_PKT4(ring, REG_A6XX_SP_2D_SRC_FORMAT, 1); OUT_RING(ring, 0xf180); OUT_PKT4(ring, REG_A6XX_RB_UNKNOWN_8E04, 1); OUT_RING(ring, 0x01000000); OUT_PKT7(ring, CP_BLIT, 1); OUT_RING(ring, CP_BLIT_0_OP(BLIT_OP_SCALE)); OUT_WFI5(ring); OUT_PKT4(ring, REG_A6XX_RB_UNKNOWN_8E04, 1); OUT_RING(ring, 0); } } static void emit_blit_texture(struct fd_ringbuffer *ring, const struct pipe_blit_info *info) { const struct pipe_box *sbox = &info->src.box; const struct pipe_box *dbox = &info->dst.box; struct fd_resource *src, *dst; struct fd_resource_slice *sslice, *dslice; enum a6xx_color_fmt sfmt, dfmt; enum a6xx_tile_mode stile, dtile; enum a3xx_color_swap sswap, dswap; unsigned spitch, dpitch; unsigned sx1, sy1, sx2, sy2; unsigned dx1, dy1, dx2, dy2; if (DEBUG_BLIT_FALLBACK) { fprintf(stderr, "texture blit: "); util_dump_blit_info(stderr, info); fprintf(stderr, "\ndst resource: "); util_dump_resource(stderr, info->dst.resource); fprintf(stderr, "\nsrc resource: "); util_dump_resource(stderr, info->src.resource); fprintf(stderr, "\n"); } src = fd_resource(info->src.resource); dst = fd_resource(info->dst.resource); sslice = fd_resource_slice(src, info->src.level); dslice = fd_resource_slice(dst, info->dst.level); sfmt = fd6_pipe2color(info->src.format); dfmt = fd6_pipe2color(info->dst.format); int blocksize = util_format_get_blocksize(info->src.format); int blockwidth = util_format_get_blockwidth(info->src.format); int blockheight = util_format_get_blockheight(info->src.format); int nelements; stile = fd_resource_level_linear(info->src.resource, info->src.level) ? TILE6_LINEAR : src->tile_mode; dtile = fd_resource_level_linear(info->dst.resource, info->dst.level) ? TILE6_LINEAR : dst->tile_mode; sswap = fd6_pipe2swap(info->src.format); dswap = fd6_pipe2swap(info->dst.format); if (util_format_is_compressed(info->src.format)) { debug_assert(info->src.format == info->dst.format); sfmt = dfmt = RB6_R8_UNORM; nelements = blocksize; } else { debug_assert(!util_format_is_compressed(info->dst.format)); nelements = 1; } spitch = DIV_ROUND_UP(sslice->pitch, blockwidth) * src->cpp; dpitch = DIV_ROUND_UP(dslice->pitch, blockwidth) * dst->cpp; sx1 = sbox->x / blockwidth * nelements; sy1 = sbox->y / blockheight; sx2 = DIV_ROUND_UP(sbox->x + sbox->width, blockwidth) * nelements - 1; sy2 = DIV_ROUND_UP(sbox->y + sbox->height, blockheight) - 1; dx1 = dbox->x / blockwidth * nelements; dy1 = dbox->y / blockheight; dx2 = DIV_ROUND_UP(dbox->x + dbox->width, blockwidth) * nelements - 1; dy2 = DIV_ROUND_UP(dbox->y + dbox->height, blockheight) - 1; uint32_t width = DIV_ROUND_UP(u_minify(src->base.width0, info->src.level), blockwidth) * nelements; uint32_t height = DIV_ROUND_UP(u_minify(src->base.height0, info->src.level), blockheight); /* if dtile, then dswap ignored by hw, and likewise if stile then sswap * ignored by hw.. but in this case we have already rejected the blit * if src and dst formats differ, so juse use WZYX for both src and * dst swap mode (so we don't change component order) */ if (stile || dtile) { debug_assert(info->src.format == info->dst.format); sswap = dswap = WZYX; } OUT_PKT7(ring, CP_SET_MARKER, 1); OUT_RING(ring, A2XX_CP_SET_MARKER_0_MODE(RM6_BLIT2DSCALE)); uint32_t blit_cntl = blit_control(dfmt); if (dtile != stile) blit_cntl |= 0x20000000; if (info->scissor_enable) { OUT_PKT4(ring, REG_A6XX_GRAS_RESOLVE_CNTL_1, 2); OUT_RING(ring, A6XX_GRAS_RESOLVE_CNTL_1_X(info->scissor.minx) | A6XX_GRAS_RESOLVE_CNTL_1_Y(info->scissor.miny)); OUT_RING(ring, A6XX_GRAS_RESOLVE_CNTL_1_X(info->scissor.maxx - 1) | A6XX_GRAS_RESOLVE_CNTL_1_Y(info->scissor.maxy - 1)); blit_cntl |= A6XX_RB_2D_BLIT_CNTL_SCISSOR; } OUT_PKT4(ring, REG_A6XX_RB_2D_BLIT_CNTL, 1); OUT_RING(ring, blit_cntl); OUT_PKT4(ring, REG_A6XX_GRAS_2D_BLIT_CNTL, 1); OUT_RING(ring, blit_cntl); for (unsigned i = 0; i < info->dst.box.depth; i++) { unsigned soff = fd_resource_offset(src, info->src.level, sbox->z + i); unsigned doff = fd_resource_offset(dst, info->dst.level, dbox->z + i); /* * Emit source: */ uint32_t filter = 0; if (info->filter == PIPE_TEX_FILTER_LINEAR) filter = A6XX_SP_PS_2D_SRC_INFO_FILTER; OUT_PKT4(ring, REG_A6XX_SP_PS_2D_SRC_INFO, 13); OUT_RING(ring, A6XX_SP_PS_2D_SRC_INFO_COLOR_FORMAT(sfmt) | A6XX_SP_PS_2D_SRC_INFO_TILE_MODE(stile) | A6XX_SP_PS_2D_SRC_INFO_COLOR_SWAP(sswap) | 0x500000 | filter); OUT_RING(ring, A6XX_SP_PS_2D_SRC_SIZE_WIDTH(width) | A6XX_SP_PS_2D_SRC_SIZE_HEIGHT(height)); /* SP_PS_2D_SRC_SIZE */ OUT_RELOC(ring, src->bo, soff, 0, 0); /* SP_PS_2D_SRC_LO/HI */ OUT_RING(ring, A6XX_SP_PS_2D_SRC_PITCH_PITCH(spitch)); OUT_RING(ring, 0x00000000); OUT_RING(ring, 0x00000000); OUT_RING(ring, 0x00000000); OUT_RING(ring, 0x00000000); OUT_RING(ring, 0x00000000); OUT_RING(ring, 0x00000000); OUT_RING(ring, 0x00000000); OUT_RING(ring, 0x00000000); /* * Emit destination: */ OUT_PKT4(ring, REG_A6XX_RB_2D_DST_INFO, 9); OUT_RING(ring, A6XX_RB_2D_DST_INFO_COLOR_FORMAT(dfmt) | A6XX_RB_2D_DST_INFO_TILE_MODE(dtile) | A6XX_RB_2D_DST_INFO_COLOR_SWAP(dswap)); OUT_RELOC(ring, dst->bo, doff, 0, 0); /* RB_2D_DST_LO/HI */ OUT_RING(ring, A6XX_RB_2D_DST_SIZE_PITCH(dpitch)); OUT_RING(ring, 0x00000000); OUT_RING(ring, 0x00000000); OUT_RING(ring, 0x00000000); OUT_RING(ring, 0x00000000); OUT_RING(ring, 0x00000000); /* * Blit command: */ OUT_PKT4(ring, REG_A6XX_GRAS_2D_SRC_TL_X, 4); OUT_RING(ring, A6XX_GRAS_2D_SRC_TL_X_X(sx1)); OUT_RING(ring, A6XX_GRAS_2D_SRC_BR_X_X(sx2)); OUT_RING(ring, A6XX_GRAS_2D_SRC_TL_Y_Y(sy1)); OUT_RING(ring, A6XX_GRAS_2D_SRC_BR_Y_Y(sy2)); OUT_PKT4(ring, REG_A6XX_GRAS_2D_DST_TL, 2); OUT_RING(ring, A6XX_GRAS_2D_DST_TL_X(dx1) | A6XX_GRAS_2D_DST_TL_Y(dy1)); OUT_RING(ring, A6XX_GRAS_2D_DST_BR_X(dx2) | A6XX_GRAS_2D_DST_BR_Y(dy2)); OUT_PKT7(ring, CP_EVENT_WRITE, 1); OUT_RING(ring, 0x3f); OUT_WFI5(ring); OUT_PKT4(ring, REG_A6XX_RB_UNKNOWN_8C01, 1); OUT_RING(ring, 0); OUT_PKT4(ring, REG_A6XX_SP_2D_SRC_FORMAT, 1); OUT_RING(ring, A6XX_SP_2D_SRC_FORMAT_COLOR_FORMAT(sfmt) | COND(util_format_is_pure_sint(info->src.format), A6XX_SP_2D_SRC_FORMAT_SINT) | COND(util_format_is_pure_uint(info->src.format), A6XX_SP_2D_SRC_FORMAT_UINT) | COND(util_format_is_snorm(info->src.format), A6XX_SP_2D_SRC_FORMAT_SINT | A6XX_SP_2D_SRC_FORMAT_NORM) | COND(util_format_is_unorm(info->src.format), // TODO sometimes blob uses UINT+NORM but dEQP seems unhappy about that // A6XX_SP_2D_SRC_FORMAT_UINT | A6XX_SP_2D_SRC_FORMAT_NORM) | 0xf000); OUT_PKT4(ring, REG_A6XX_RB_UNKNOWN_8E04, 1); OUT_RING(ring, 0x01000000); OUT_PKT7(ring, CP_BLIT, 1); OUT_RING(ring, CP_BLIT_0_OP(BLIT_OP_SCALE)); OUT_WFI5(ring); OUT_PKT4(ring, REG_A6XX_RB_UNKNOWN_8E04, 1); OUT_RING(ring, 0); } } static void emit_blit(struct fd_context *ctx, const struct pipe_blit_info *info) { struct fd_batch *batch; fd_fence_ref(ctx->base.screen, &ctx->last_fence, NULL); batch = fd_bc_alloc_batch(&ctx->screen->batch_cache, ctx, true); fd6_emit_restore(batch, batch->draw); fd6_emit_lrz_flush(batch->draw); mtx_lock(&ctx->screen->lock); fd_batch_resource_used(batch, fd_resource(info->src.resource), false); fd_batch_resource_used(batch, fd_resource(info->dst.resource), true); mtx_unlock(&ctx->screen->lock); emit_setup(batch->draw); if ((info->src.resource->target == PIPE_BUFFER) && (info->dst.resource->target == PIPE_BUFFER)) { assert(fd_resource(info->src.resource)->tile_mode == TILE6_LINEAR); assert(fd_resource(info->dst.resource)->tile_mode == TILE6_LINEAR); emit_blit_buffer(batch->draw, info); } else { /* I don't *think* we need to handle blits between buffer <-> !buffer */ debug_assert(info->src.resource->target != PIPE_BUFFER); debug_assert(info->dst.resource->target != PIPE_BUFFER); emit_blit_texture(batch->draw, info); } fd6_event_write(batch, batch->draw, 0x1d, true); fd6_event_write(batch, batch->draw, FACENESS_FLUSH, true); fd6_event_write(batch, batch->draw, CACHE_FLUSH_TS, true); fd_resource(info->dst.resource)->valid = true; batch->needs_flush = true; fd_batch_flush(batch, false, false); fd_batch_reference(&batch, NULL); } static bool fd6_blit(struct fd_context *ctx, const struct pipe_blit_info *info) { if (!can_do_blit(info)) { return false; } emit_blit(ctx, info); return true; } void fd6_blitter_init(struct pipe_context *pctx) { if (fd_mesa_debug & FD_DBG_NOBLIT) return; fd_context(pctx)->blit = fd6_blit; } unsigned fd6_tile_mode(const struct pipe_resource *tmpl) { /* basically just has to be a format we can blit, so uploads/downloads * via linear staging buffer works: */ return TILE6_3; }