--- /dev/null
+/*
+ * Copyright (C) 2010 Advanced Micro Devices, Inc.
+ *
+ * All Rights Reserved.
+ *
+ * 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 COPYRIGHT OWNER(S) 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.
+ *
+ */
+
+#include "radeon_common.h"
+#include "radeon_context.h"
+#include "radeon_blit.h"
+
+static inline uint32_t cmdpacket0(struct radeon_screen *rscrn,
+ int reg, int count)
+{
+ if (count)
+ return CP_PACKET0(reg, count - 1);
+ return CP_PACKET2;
+}
+
+static inline void emit_vtx_state(struct r100_context *r100)
+{
+ BATCH_LOCALS(&r100->radeon);
+
+ BEGIN_BATCH(8);
+ if (r100->radeon.radeonScreen->chip_flags & RADEON_CHIPSET_TCL) {
+ OUT_BATCH_REGVAL(RADEON_SE_CNTL_STATUS, 0);
+ } else {
+ OUT_BATCH_REGVAL(RADEON_SE_CNTL_STATUS, RADEON_TCL_BYPASS);
+
+ }
+ OUT_BATCH_REGVAL(RADEON_SE_COORD_FMT, (RADEON_VTX_XY_PRE_MULT_1_OVER_W0 |
+ RADEON_TEX1_W_ROUTING_USE_W0));
+ OUT_BATCH_REGVAL(RADEON_SE_VTX_FMT, RADEON_SE_VTX_FMT_XY | RADEON_SE_VTX_FMT_ST0);
+ OUT_BATCH_REGVAL(RADEON_SE_CNTL, (RADEON_DIFFUSE_SHADE_GOURAUD |
+ RADEON_BFACE_SOLID |
+ RADEON_FFACE_SOLID |
+ RADEON_VTX_PIX_CENTER_OGL |
+ RADEON_ROUND_MODE_ROUND |
+ RADEON_ROUND_PREC_4TH_PIX));
+ END_BATCH();
+}
+
+static void inline emit_tx_setup(struct r100_context *r100,
+ gl_format mesa_format,
+ struct radeon_bo *bo,
+ intptr_t offset,
+ unsigned width,
+ unsigned height,
+ unsigned pitch)
+{
+ uint32_t txformat = RADEON_TXFORMAT_NON_POWER2;
+ BATCH_LOCALS(&r100->radeon);
+
+ assert(width <= 2047);
+ assert(height <= 2047);
+ assert(offset % 32 == 0);
+
+ /* XXX others? BE/LE? */
+ switch (mesa_format) {
+ case MESA_FORMAT_ARGB8888:
+ txformat |= RADEON_TXFORMAT_ARGB8888 | RADEON_TXFORMAT_ALPHA_IN_MAP;
+ break;
+ case MESA_FORMAT_XRGB8888:
+ txformat |= RADEON_TXFORMAT_ARGB8888;
+ break;
+ case MESA_FORMAT_RGB565:
+ txformat |= RADEON_TXFORMAT_RGB565;
+ break;
+ case MESA_FORMAT_ARGB1555:
+ txformat |= RADEON_TXFORMAT_ARGB1555 | RADEON_TXFORMAT_ALPHA_IN_MAP;
+ break;
+ case MESA_FORMAT_A8:
+ txformat |= RADEON_TXFORMAT_I8 | RADEON_TXFORMAT_ALPHA_IN_MAP;
+ break;
+ default:
+ break;
+ }
+
+ BEGIN_BATCH(18);
+ OUT_BATCH_REGVAL(RADEON_PP_CNTL, RADEON_TEX_0_ENABLE | RADEON_TEX_BLEND_0_ENABLE);
+ OUT_BATCH_REGVAL(RADEON_PP_TXCBLEND_0, (RADEON_COLOR_ARG_A_ZERO |
+ RADEON_COLOR_ARG_B_ZERO |
+ RADEON_COLOR_ARG_C_T0_COLOR |
+ RADEON_BLEND_CTL_ADD |
+ RADEON_CLAMP_TX));
+ OUT_BATCH_REGVAL(RADEON_PP_TXABLEND_0, (RADEON_ALPHA_ARG_A_ZERO |
+ RADEON_ALPHA_ARG_B_ZERO |
+ RADEON_ALPHA_ARG_C_T0_ALPHA |
+ RADEON_BLEND_CTL_ADD |
+ RADEON_CLAMP_TX));
+ OUT_BATCH_REGVAL(RADEON_PP_TXFILTER_0, (RADEON_CLAMP_S_CLAMP_LAST |
+ RADEON_CLAMP_T_CLAMP_LAST |
+ RADEON_MAG_FILTER_NEAREST |
+ RADEON_MIN_FILTER_NEAREST));
+ OUT_BATCH_REGVAL(RADEON_PP_TXFORMAT_0, txformat);
+ OUT_BATCH_REGVAL(RADEON_PP_TEX_SIZE_0, ((width - 1) |
+ ((height - 1) << RADEON_TEX_VSIZE_SHIFT)));
+ OUT_BATCH_REGVAL(RADEON_PP_TEX_PITCH_0, pitch - 32);
+
+ OUT_BATCH_REGSEQ(RADEON_PP_TXOFFSET_0, 1);
+ OUT_BATCH_RELOC(0, bo, 0, RADEON_GEM_DOMAIN_GTT|RADEON_GEM_DOMAIN_VRAM, 0, 0);
+
+ END_BATCH();
+}
+
+static inline void emit_cb_setup(struct r100_context *r100,
+ struct radeon_bo *bo,
+ intptr_t offset,
+ gl_format mesa_format,
+ unsigned pitch,
+ unsigned width,
+ unsigned height)
+{
+ uint32_t dst_pitch = pitch;
+ uint32_t dst_format = 0;
+ BATCH_LOCALS(&r100->radeon);
+
+ /* XXX others? BE/LE? */
+ switch (mesa_format) {
+ case MESA_FORMAT_ARGB8888:
+ case MESA_FORMAT_XRGB8888:
+ dst_format = RADEON_COLOR_FORMAT_ARGB8888;
+ break;
+ case MESA_FORMAT_RGB565:
+ dst_format = RADEON_COLOR_FORMAT_RGB565;
+ break;
+ case MESA_FORMAT_ARGB1555:
+ dst_format = RADEON_COLOR_FORMAT_ARGB1555;
+ break;
+ case MESA_FORMAT_A8:
+ dst_format = RADEON_COLOR_FORMAT_RGB8;
+ break;
+ default:
+ break;
+ }
+
+ BEGIN_BATCH_NO_AUTOSTATE(18);
+ OUT_BATCH_REGVAL(RADEON_RE_TOP_LEFT, 0);
+ OUT_BATCH_REGVAL(RADEON_RE_WIDTH_HEIGHT, ((width << RADEON_RE_WIDTH_SHIFT) |
+ (height << RADEON_RE_HEIGHT_SHIFT)));
+ OUT_BATCH_REGVAL(RADEON_RB3D_PLANEMASK, 0xffffffff);
+ OUT_BATCH_REGVAL(RADEON_RB3D_BLENDCNTL, RADEON_SRC_BLEND_GL_ONE | RADEON_DST_BLEND_GL_ZERO);
+ OUT_BATCH_REGVAL(RADEON_RB3D_CNTL, dst_format);
+
+ OUT_BATCH_REGSEQ(RADEON_RB3D_COLOROFFSET, 1);
+ OUT_BATCH_RELOC(0, bo, 0, 0, RADEON_GEM_DOMAIN_GTT|RADEON_GEM_DOMAIN_VRAM, 0);
+ OUT_BATCH_REGSEQ(RADEON_RB3D_COLORPITCH, 1);
+ OUT_BATCH_RELOC(dst_pitch, bo, dst_pitch, 0, RADEON_GEM_DOMAIN_GTT|RADEON_GEM_DOMAIN_VRAM, 0);
+
+ END_BATCH();
+}
+
+static GLboolean validate_buffers(struct r100_context *r100,
+ struct radeon_bo *src_bo,
+ struct radeon_bo *dst_bo)
+{
+ int ret;
+ radeon_cs_space_add_persistent_bo(r100->radeon.cmdbuf.cs,
+ src_bo, RADEON_GEM_DOMAIN_VRAM, 0);
+
+ radeon_cs_space_add_persistent_bo(r100->radeon.cmdbuf.cs,
+ dst_bo, 0, RADEON_GEM_DOMAIN_VRAM);
+
+ ret = radeon_cs_space_check_with_bo(r100->radeon.cmdbuf.cs,
+ first_elem(&r100->radeon.dma.reserved)->bo,
+ RADEON_GEM_DOMAIN_GTT, 0);
+ if (ret)
+ return GL_FALSE;
+
+ return GL_TRUE;
+}
+
+/**
+ * Calculate texcoords for given image region.
+ * Output values are [minx, maxx, miny, maxy]
+ */
+static inline void calc_tex_coords(float img_width, float img_height,
+ float x, float y,
+ float reg_width, float reg_height,
+ unsigned flip_y, float *buf)
+{
+ buf[0] = x / img_width;
+ buf[1] = buf[0] + reg_width / img_width;
+ buf[2] = y / img_height;
+ buf[3] = buf[2] + reg_height / img_height;
+ if (flip_y)
+ {
+ float tmp = buf[2];
+ buf[2] = 1.0 - buf[3];
+ buf[3] = 1.0 - tmp;
+ }
+}
+
+static inline void emit_draw_packet(struct r100_context *r100,
+ unsigned src_width, unsigned src_height,
+ unsigned src_x_offset, unsigned src_y_offset,
+ unsigned dst_x_offset, unsigned dst_y_offset,
+ unsigned reg_width, unsigned reg_height,
+ unsigned flip_y)
+{
+ float texcoords[4];
+ float verts[12];
+ BATCH_LOCALS(&r100->radeon);
+
+ calc_tex_coords(src_width, src_height,
+ src_x_offset, src_y_offset,
+ reg_width, reg_height,
+ flip_y, texcoords);
+
+ verts[0] = dst_x_offset;
+ verts[1] = dst_y_offset + reg_height;
+ verts[2] = texcoords[0];
+ verts[3] = texcoords[2];
+
+ verts[4] = dst_x_offset + reg_width;
+ verts[5] = dst_y_offset + reg_height;
+ verts[6] = texcoords[1];
+ verts[7] = texcoords[2];
+
+ verts[8] = dst_x_offset + reg_width;
+ verts[9] = dst_y_offset;
+ verts[10] = texcoords[1];
+ verts[11] = texcoords[3];
+
+ BEGIN_BATCH(15);
+ OUT_BATCH(RADEON_CP_PACKET3_3D_DRAW_IMMD | (13 << 16));
+ OUT_BATCH(RADEON_CP_VC_FRMT_XY | RADEON_CP_VC_FRMT_ST0);
+ OUT_BATCH(RADEON_CP_VC_CNTL_PRIM_WALK_RING |
+ RADEON_CP_VC_CNTL_PRIM_TYPE_RECT_LIST |
+ RADEON_CP_VC_CNTL_MAOS_ENABLE |
+ RADEON_CP_VC_CNTL_VTX_FMT_RADEON_MODE |
+ (3 << 16));
+ OUT_BATCH_TABLE(verts, 12);
+ END_BATCH();
+}
+
+/**
+ * Copy a region of [@a width x @a height] pixels from source buffer
+ * to destination buffer.
+ * @param[in] r100 r100 context
+ * @param[in] src_bo source radeon buffer object
+ * @param[in] src_offset offset of the source image in the @a src_bo
+ * @param[in] src_mesaformat source image format
+ * @param[in] src_pitch aligned source image width
+ * @param[in] src_width source image width
+ * @param[in] src_height source image height
+ * @param[in] src_x_offset x offset in the source image
+ * @param[in] src_y_offset y offset in the source image
+ * @param[in] dst_bo destination radeon buffer object
+ * @param[in] dst_offset offset of the destination image in the @a dst_bo
+ * @param[in] dst_mesaformat destination image format
+ * @param[in] dst_pitch aligned destination image width
+ * @param[in] dst_width destination image width
+ * @param[in] dst_height destination image height
+ * @param[in] dst_x_offset x offset in the destination image
+ * @param[in] dst_y_offset y offset in the destination image
+ * @param[in] width region width
+ * @param[in] height region height
+ * @param[in] flip_y set if y coords of the source image need to be flipped
+ */
+GLboolean r100_blit(struct r100_context *r100,
+ struct radeon_bo *src_bo,
+ intptr_t src_offset,
+ gl_format src_mesaformat,
+ unsigned src_pitch,
+ unsigned src_width,
+ unsigned src_height,
+ unsigned src_x_offset,
+ unsigned src_y_offset,
+ struct radeon_bo *dst_bo,
+ intptr_t dst_offset,
+ gl_format dst_mesaformat,
+ unsigned dst_pitch,
+ unsigned dst_width,
+ unsigned dst_height,
+ unsigned dst_x_offset,
+ unsigned dst_y_offset,
+ unsigned reg_width,
+ unsigned reg_height,
+ unsigned flip_y)
+{
+ if (_mesa_get_format_bits(src_mesaformat, GL_DEPTH_BITS) > 0)
+ return GL_FALSE;
+
+ /* Make sure that colorbuffer has even width - hw limitation */
+ if (dst_pitch % 2 > 0)
+ ++dst_pitch;
+
+ /* Rendering to small buffer doesn't work.
+ * Looks like a hw limitation.
+ */
+ if (dst_pitch < 32)
+ return GL_FALSE;
+
+ /* Need to clamp the region size to make sure
+ * we don't read outside of the source buffer
+ * or write outside of the destination buffer.
+ */
+ if (reg_width + src_x_offset > src_width)
+ reg_width = src_width - src_x_offset;
+ if (reg_height + src_y_offset > src_height)
+ reg_height = src_height - src_y_offset;
+ if (reg_width + dst_x_offset > dst_width)
+ reg_width = dst_width - dst_x_offset;
+ if (reg_height + dst_y_offset > dst_height)
+ reg_height = dst_height - dst_y_offset;
+
+ if (src_bo == dst_bo) {
+ return GL_FALSE;
+ }
+
+ if (0) {
+ fprintf(stderr, "src: size [%d x %d], pitch %d, "
+ "offset [%d x %d], format %s, bo %p\n",
+ src_width, src_height, src_pitch,
+ src_x_offset, src_y_offset,
+ _mesa_get_format_name(src_mesaformat),
+ src_bo);
+ fprintf(stderr, "dst: pitch %d, offset[%d x %d], format %s, bo %p\n",
+ dst_pitch, dst_x_offset, dst_y_offset,
+ _mesa_get_format_name(dst_mesaformat), dst_bo);
+ fprintf(stderr, "region: %d x %d\n", reg_width, reg_height);
+ }
+
+ /* Flush is needed to make sure that source buffer has correct data */
+ radeonFlush(r100->radeon.glCtx);
+
+ rcommonEnsureCmdBufSpace(&r100->radeon, 59, __FUNCTION__);
+
+ if (!validate_buffers(r100, src_bo, dst_bo))
+ return GL_FALSE;
+
+ /* 8 */
+ emit_vtx_state(r100);
+ /* 18 */
+ emit_tx_setup(r100, src_mesaformat, src_bo, src_offset, src_width, src_height, src_pitch);
+ /* 18 */
+ emit_cb_setup(r100, dst_bo, dst_offset, dst_mesaformat, dst_pitch, dst_width, dst_height);
+ /* 15 */
+ emit_draw_packet(r100, src_width, src_height,
+ src_x_offset, src_y_offset,
+ dst_x_offset, dst_y_offset,
+ reg_width, reg_height,
+ flip_y);
+
+ radeonFlush(r100->radeon.glCtx);
+
+ return GL_TRUE;
+}
--- /dev/null
+/*
+ * Copyright (C) 2009 Maciej Cencora <m.cencora@gmail.com>
+ *
+ * All Rights Reserved.
+ *
+ * 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 COPYRIGHT OWNER(S) 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.
+ *
+ */
+
+#include "radeon_common.h"
+#include "radeon_context.h"
+
+#include "main/image.h"
+#include "main/teximage.h"
+#include "main/texstate.h"
+#include "drivers/common/meta.h"
+
+#include "radeon_mipmap_tree.h"
+#include "radeon_blit.h"
+#include <main/debug.h>
+
+// TODO:
+// need to pass correct pitch for small dst textures!
+static GLboolean
+do_copy_texsubimage(GLcontext *ctx,
+ GLenum target, GLint level,
+ struct radeon_tex_obj *tobj,
+ radeon_texture_image *timg,
+ GLint dstx, GLint dsty,
+ GLint x, GLint y,
+ GLsizei width, GLsizei height)
+{
+ struct r100_context *r100 = R100_CONTEXT(ctx);
+ struct radeon_renderbuffer *rrb;
+
+ if (_mesa_get_format_bits(timg->base.TexFormat, GL_DEPTH_BITS) > 0) {
+ rrb = radeon_get_depthbuffer(&r100->radeon);
+ } else {
+ rrb = radeon_get_colorbuffer(&r100->radeon);
+ }
+
+ if (!timg->mt) {
+ radeon_validate_texture_miptree(ctx, &tobj->base);
+ }
+
+ assert(rrb && rrb->bo);
+ assert(timg->mt->bo);
+ assert(timg->base.Width >= dstx + width);
+ assert(timg->base.Height >= dsty + height);
+
+ intptr_t src_offset = rrb->draw_offset;
+ intptr_t dst_offset = radeon_miptree_image_offset(timg->mt, _mesa_tex_target_to_face(target), level);
+
+ if (src_offset % 32 || dst_offset % 32) {
+ return GL_FALSE;
+ }
+
+ if (0) {
+ fprintf(stderr, "%s: copying to face %d, level %d\n",
+ __FUNCTION__, _mesa_tex_target_to_face(target), level);
+ fprintf(stderr, "to: x %d, y %d, offset %d\n", dstx, dsty, (uint32_t) dst_offset);
+ fprintf(stderr, "from (%dx%d) width %d, height %d, offset %d, pitch %d\n",
+ x, y, rrb->base.Width, rrb->base.Height, (uint32_t) src_offset, rrb->pitch/rrb->cpp);
+ fprintf(stderr, "src size %d, dst size %d\n", rrb->bo->size, timg->mt->bo->size);
+
+ }
+
+ /* blit from src buffer to texture */
+ return r100_blit(r100, rrb->bo, src_offset, rrb->base.Format, rrb->pitch,
+ rrb->base.Width, rrb->base.Height, x, y,
+ timg->mt->bo, dst_offset, timg->base.TexFormat,
+ timg->base.Width, timg->base.Width, timg->base.Height,
+ dstx, dsty, width, height, 1);
+}
+
+static void
+r100CopyTexImage2D(GLcontext *ctx, GLenum target, GLint level,
+ GLenum internalFormat,
+ GLint x, GLint y, GLsizei width, GLsizei height,
+ GLint border)
+{
+ struct gl_texture_unit *texUnit = _mesa_get_current_tex_unit(ctx);
+ struct gl_texture_object *texObj =
+ _mesa_select_tex_object(ctx, texUnit, target);
+ struct gl_texture_image *texImage =
+ _mesa_select_tex_image(ctx, texObj, target, level);
+ int srcx, srcy, dstx, dsty;
+
+ if (border)
+ goto fail;
+
+ /* Setup or redefine the texture object, mipmap tree and texture
+ * image. Don't populate yet.
+ */
+ ctx->Driver.TexImage2D(ctx, target, level, internalFormat,
+ width, height, border,
+ GL_RGBA, GL_UNSIGNED_BYTE, NULL,
+ &ctx->DefaultPacking, texObj, texImage);
+
+ srcx = x;
+ srcy = y;
+ dstx = 0;
+ dsty = 0;
+ if (!_mesa_clip_copytexsubimage(ctx,
+ &dstx, &dsty,
+ &srcx, &srcy,
+ &width, &height)) {
+ return;
+ }
+
+ if (!do_copy_texsubimage(ctx, target, level,
+ radeon_tex_obj(texObj), (radeon_texture_image *)texImage,
+ 0, 0, x, y, width, height)) {
+ goto fail;
+ }
+
+ return;
+
+fail:
+ _mesa_meta_CopyTexImage2D(ctx, target, level, internalFormat, x, y,
+ width, height, border);
+}
+
+static void
+r100CopyTexSubImage2D(GLcontext *ctx, GLenum target, GLint level,
+ GLint xoffset, GLint yoffset,
+ GLint x, GLint y,
+ GLsizei width, GLsizei height)
+{
+ struct gl_texture_unit *texUnit = _mesa_get_current_tex_unit(ctx);
+ struct gl_texture_object *texObj = _mesa_select_tex_object(ctx, texUnit, target);
+ struct gl_texture_image *texImage = _mesa_select_tex_image(ctx, texObj, target, level);
+
+ if (!do_copy_texsubimage(ctx, target, level,
+ radeon_tex_obj(texObj), (radeon_texture_image *)texImage,
+ xoffset, yoffset, x, y, width, height)) {
+
+ //DEBUG_FALLBACKS
+
+ _mesa_meta_CopyTexSubImage2D(ctx, target, level,
+ xoffset, yoffset, x, y, width, height);
+ }
+}
+
+
+void r100_init_texcopy_functions(struct dd_function_table *table)
+{
+ table->CopyTexImage2D = r100CopyTexImage2D;
+ table->CopyTexSubImage2D = r100CopyTexSubImage2D;
+}
projects / mesa.git / commitdiff