/* * Copyright 2018-2019 Alyssa Rosenzweig * * 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. * */ #include "pan_bo.h" #include "pan_context.h" #include "pan_util.h" #include "util/format/u_format.h" static struct mali_sfbd_format panfrost_sfbd_format(struct pipe_surface *surf) { /* Explode details on the format */ const struct util_format_description *desc = util_format_description(surf->format); /* The swizzle for rendering is inverted from texturing */ unsigned char swizzle[4]; panfrost_invert_swizzle(desc->swizzle, swizzle); struct mali_sfbd_format fmt = { .unk1 = 0x1, .swizzle = panfrost_translate_swizzle_4(swizzle), .nr_channels = MALI_POSITIVE(desc->nr_channels), .unk2 = 0x4, .unk3 = 0xb, }; if (desc->colorspace == UTIL_FORMAT_COLORSPACE_SRGB) fmt.unk2 |= MALI_SFBD_FORMAT_SRGB; /* sRGB handled as a dedicated flag */ enum pipe_format linearized = util_format_linear(surf->format); /* If RGB, we're good to go */ if (util_format_is_unorm8(desc)) return fmt; switch (linearized) { case PIPE_FORMAT_B5G6R5_UNORM: fmt.unk1 = 0x5; fmt.nr_channels = MALI_POSITIVE(2); fmt.unk2 = 0x5; break; case PIPE_FORMAT_A4B4G4R4_UNORM: case PIPE_FORMAT_B4G4R4A4_UNORM: fmt.unk1 = 0x4; fmt.nr_channels = MALI_POSITIVE(1); fmt.unk2 = 0x5; break; default: unreachable("Invalid format rendering"); } return fmt; } static void panfrost_sfbd_clear( struct panfrost_batch *batch, struct mali_single_framebuffer *sfbd) { if (batch->clear & PIPE_CLEAR_COLOR) { sfbd->clear_color_1 = batch->clear_color[0][0]; sfbd->clear_color_2 = batch->clear_color[0][1]; sfbd->clear_color_3 = batch->clear_color[0][2]; sfbd->clear_color_4 = batch->clear_color[0][3]; } if (batch->clear & PIPE_CLEAR_DEPTH) { sfbd->clear_depth_1 = batch->clear_depth; sfbd->clear_depth_2 = batch->clear_depth; sfbd->clear_depth_3 = batch->clear_depth; sfbd->clear_depth_4 = batch->clear_depth; } if (batch->clear & PIPE_CLEAR_STENCIL) { sfbd->clear_stencil = batch->clear_stencil; } /* Set flags based on what has been cleared, for the SFBD case */ /* XXX: What do these flags mean? */ int clear_flags = 0x101100; if (!(batch->clear & ~(PIPE_CLEAR_COLOR | PIPE_CLEAR_DEPTH | PIPE_CLEAR_STENCIL))) { /* On a tiler like this, it's fastest to clear all three buffers at once */ clear_flags |= MALI_CLEAR_FAST; } else { clear_flags |= MALI_CLEAR_SLOW; if (batch->clear & PIPE_CLEAR_STENCIL) clear_flags |= MALI_CLEAR_SLOW_STENCIL; } sfbd->clear_flags = clear_flags; } static void panfrost_sfbd_set_cbuf( struct mali_single_framebuffer *fb, struct pipe_surface *surf) { struct panfrost_resource *rsrc = pan_resource(surf->texture); unsigned level = surf->u.tex.level; unsigned first_layer = surf->u.tex.first_layer; assert(surf->u.tex.last_layer == first_layer); signed stride = rsrc->slices[level].stride; mali_ptr base = panfrost_get_texture_address(rsrc, level, first_layer); fb->format = panfrost_sfbd_format(surf); fb->framebuffer = base; fb->stride = stride; if (rsrc->layout == MALI_TEXTURE_LINEAR) fb->format.block = MALI_BLOCK_LINEAR; else if (rsrc->layout == MALI_TEXTURE_TILED) { fb->format.block = MALI_BLOCK_TILED; fb->stride *= 16; } else { fprintf(stderr, "Invalid render layout\n"); assert(0); } } static void panfrost_sfbd_set_zsbuf( struct mali_single_framebuffer *fb, struct pipe_surface *surf) { struct panfrost_resource *rsrc = pan_resource(surf->texture); struct panfrost_context *ctx = pan_context(surf->context); unsigned level = surf->u.tex.level; assert(surf->u.tex.first_layer == 0); if (rsrc->layout != MALI_TEXTURE_TILED) unreachable("Invalid render layout."); fb->depth_buffer = rsrc->bo->gpu + rsrc->slices[level].offset; fb->depth_stride = rsrc->slices[level].stride; /* No stencil? Job done. */ if (!ctx->depth_stencil || !ctx->depth_stencil->stencil[0].enabled) return; if (panfrost_is_z24s8_variant(surf->format)) { /* Stencil data is interleaved with depth */ fb->stencil_buffer = fb->depth_buffer; fb->stencil_stride = fb->depth_stride; } else if (surf->format == PIPE_FORMAT_Z32_UNORM || surf->format == PIPE_FORMAT_Z32_FLOAT) { /* No stencil, nothing to do */ } else if (surf->format == PIPE_FORMAT_Z32_FLOAT_S8X24_UINT) { /* Stencil data in separate buffer */ struct panfrost_resource *stencil = rsrc->separate_stencil; struct panfrost_slice stencil_slice = stencil->slices[level]; fb->stencil_buffer = stencil->bo->gpu + stencil_slice.offset; fb->stencil_stride = stencil_slice.stride; } else unreachable("Unsupported depth/stencil format."); } static struct mali_single_framebuffer panfrost_emit_sfbd(struct panfrost_batch *batch, unsigned vertex_count) { struct panfrost_context *ctx = batch->ctx; struct pipe_context *gallium = (struct pipe_context *) ctx; struct panfrost_screen *screen = pan_screen(gallium->screen); unsigned width = batch->key.width; unsigned height = batch->key.height; /* TODO: Why do we need to make the stack bigger than other platforms? */ unsigned shift = panfrost_get_stack_shift(MAX2(batch->stack_size, 512)); /* TODO: where do we specify the shift? */ struct mali_single_framebuffer framebuffer = { .width = MALI_POSITIVE(width), .height = MALI_POSITIVE(height), .shared_memory = { .shared_workgroup_count = ~0, .scratchpad = panfrost_batch_get_scratchpad(batch, shift, screen->thread_tls_alloc, screen->core_count)->gpu, }, .format = { .unk3 = 0x3, }, .clear_flags = 0x1000, .tiler = panfrost_emit_midg_tiler(batch, vertex_count), }; return framebuffer; } void panfrost_attach_sfbd(struct panfrost_batch *batch, unsigned vertex_count) { struct mali_single_framebuffer sfbd = panfrost_emit_sfbd(batch, vertex_count); memcpy(batch->framebuffer.cpu, &sfbd, sizeof(sfbd)); } /* Creates an SFBD for the FRAGMENT section of the bound framebuffer */ mali_ptr panfrost_sfbd_fragment(struct panfrost_batch *batch, bool has_draws) { struct mali_single_framebuffer fb = panfrost_emit_sfbd(batch, has_draws); panfrost_sfbd_clear(batch, &fb); /* SFBD does not support MRT natively; sanity check */ assert(batch->key.nr_cbufs <= 1); if (batch->key.nr_cbufs) { struct pipe_surface *surf = batch->key.cbufs[0]; struct panfrost_resource *rsrc = pan_resource(surf->texture); struct panfrost_bo *bo = rsrc->bo; panfrost_sfbd_set_cbuf(&fb, surf); if (rsrc->checksummed) { unsigned level = surf->u.tex.level; struct panfrost_slice *slice = &rsrc->slices[level]; fb.checksum_stride = slice->checksum_stride; fb.checksum = bo->gpu + slice->checksum_offset; } } if (batch->key.zsbuf) panfrost_sfbd_set_zsbuf(&fb, batch->key.zsbuf); if (batch->requirements & PAN_REQ_MSAA) { fb.format.unk1 |= MALI_SFBD_FORMAT_MSAA_A; fb.format.unk2 |= MALI_SFBD_FORMAT_MSAA_B; } return panfrost_upload_transient(batch, &fb, sizeof(fb)); }