diff options
Diffstat (limited to 'src')
53 files changed, 618 insertions, 279 deletions
diff --git a/src/gallium/auxiliary/gallivm/lp_bld_alpha.h b/src/gallium/auxiliary/gallivm/lp_bld_alpha.h index 634575670d..fe3cedcc48 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_alpha.h +++ b/src/gallium/auxiliary/gallivm/lp_bld_alpha.h @@ -35,7 +35,7 @@ #define LP_BLD_ALPHA_H -#include <llvm-c/Core.h> +#include "os/os_llvm.h" struct pipe_alpha_state; struct lp_type; diff --git a/src/gallium/auxiliary/gallivm/lp_bld_arit.c b/src/gallium/auxiliary/gallivm/lp_bld_arit.c index f55d2b6d15..90c84c6a4e 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_arit.c +++ b/src/gallium/auxiliary/gallivm/lp_bld_arit.c @@ -232,6 +232,37 @@ lp_build_add(struct lp_build_context *bld, } +/** Return the sum of the elements of a */ +LLVMValueRef +lp_build_sum_vector(struct lp_build_context *bld, + LLVMValueRef a) +{ + const struct lp_type type = bld->type; + LLVMValueRef index, res; + int i; + + if (a == bld->zero) + return bld->zero; + if (a == bld->undef) + return bld->undef; + assert(type.length > 1); + + assert(!bld->type.norm); + + index = LLVMConstInt(LLVMInt32Type(), 0, 0); + res = LLVMBuildExtractElement(bld->builder, a, index, ""); + + for (i = 1; i < type.length; i++) { + index = LLVMConstInt(LLVMInt32Type(), i, 0); + res = LLVMBuildAdd(bld->builder, res, + LLVMBuildExtractElement(bld->builder, a, index, ""), + ""); + } + + return res; +} + + /** * Generate a - b */ diff --git a/src/gallium/auxiliary/gallivm/lp_bld_arit.h b/src/gallium/auxiliary/gallivm/lp_bld_arit.h index 55385e3a66..7a10fe1220 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_arit.h +++ b/src/gallium/auxiliary/gallivm/lp_bld_arit.h @@ -37,7 +37,7 @@ #define LP_BLD_ARIT_H -#include <llvm-c/Core.h> +#include "os/os_llvm.h" struct lp_type; @@ -57,6 +57,10 @@ lp_build_add(struct lp_build_context *bld, LLVMValueRef b); LLVMValueRef +lp_build_sum_vector(struct lp_build_context *bld, + LLVMValueRef a); + +LLVMValueRef lp_build_sub(struct lp_build_context *bld, LLVMValueRef a, LLVMValueRef b); diff --git a/src/gallium/auxiliary/gallivm/lp_bld_blend.h b/src/gallium/auxiliary/gallivm/lp_bld_blend.h index da272e549f..5a9e1c1fb2 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_blend.h +++ b/src/gallium/auxiliary/gallivm/lp_bld_blend.h @@ -40,7 +40,7 @@ * for a standalone example. */ -#include <llvm-c/Core.h> +#include "os/os_llvm.h" #include "pipe/p_format.h" diff --git a/src/gallium/auxiliary/gallivm/lp_bld_const.h b/src/gallium/auxiliary/gallivm/lp_bld_const.h index cb8e1c7b00..4078636103 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_const.h +++ b/src/gallium/auxiliary/gallivm/lp_bld_const.h @@ -37,7 +37,7 @@ #define LP_BLD_CONST_H -#include <llvm-c/Core.h> +#include "os/os_llvm.h" #include <pipe/p_compiler.h> diff --git a/src/gallium/auxiliary/gallivm/lp_bld_conv.h b/src/gallium/auxiliary/gallivm/lp_bld_conv.h index 948e68fae4..78e8155ff7 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_conv.h +++ b/src/gallium/auxiliary/gallivm/lp_bld_conv.h @@ -37,7 +37,7 @@ #define LP_BLD_CONV_H -#include <llvm-c/Core.h> +#include "os/os_llvm.h" struct lp_type; diff --git a/src/gallium/auxiliary/gallivm/lp_bld_debug.h b/src/gallium/auxiliary/gallivm/lp_bld_debug.h index 583e6132b4..441ad94786 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_debug.h +++ b/src/gallium/auxiliary/gallivm/lp_bld_debug.h @@ -30,7 +30,7 @@ #define LP_BLD_DEBUG_H -#include <llvm-c/Core.h> +#include "os/os_llvm.h" #include "pipe/p_compiler.h" #include "util/u_string.h" diff --git a/src/gallium/auxiliary/gallivm/lp_bld_depth.h b/src/gallium/auxiliary/gallivm/lp_bld_depth.h index 79d6981bb5..8be80024ae 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_depth.h +++ b/src/gallium/auxiliary/gallivm/lp_bld_depth.h @@ -36,7 +36,7 @@ #define LP_BLD_DEPTH_H -#include <llvm-c/Core.h> +#include "os/os_llvm.h" struct pipe_depth_state; diff --git a/src/gallium/auxiliary/gallivm/lp_bld_flow.h b/src/gallium/auxiliary/gallivm/lp_bld_flow.h index 8bb22543ee..e158836549 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_flow.h +++ b/src/gallium/auxiliary/gallivm/lp_bld_flow.h @@ -35,7 +35,7 @@ #define LP_BLD_FLOW_H -#include <llvm-c/Core.h> +#include "os/os_llvm.h" struct lp_type; diff --git a/src/gallium/auxiliary/gallivm/lp_bld_format.h b/src/gallium/auxiliary/gallivm/lp_bld_format.h index 970bee379f..8972c0dc17 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_format.h +++ b/src/gallium/auxiliary/gallivm/lp_bld_format.h @@ -34,7 +34,7 @@ * Pixel format helpers. */ -#include <llvm-c/Core.h> +#include "os/os_llvm.h" #include "pipe/p_format.h" diff --git a/src/gallium/auxiliary/gallivm/lp_bld_interp.h b/src/gallium/auxiliary/gallivm/lp_bld_interp.h index ca958cdf34..177b5e943e 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_interp.h +++ b/src/gallium/auxiliary/gallivm/lp_bld_interp.h @@ -41,7 +41,7 @@ #define LP_BLD_INTERP_H -#include <llvm-c/Core.h> +#include "os/os_llvm.h" #include "tgsi/tgsi_exec.h" diff --git a/src/gallium/auxiliary/gallivm/lp_bld_intr.h b/src/gallium/auxiliary/gallivm/lp_bld_intr.h index f813f27074..7d5506c733 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_intr.h +++ b/src/gallium/auxiliary/gallivm/lp_bld_intr.h @@ -37,7 +37,7 @@ #define LP_BLD_INTR_H -#include <llvm-c/Core.h> +#include "os/os_llvm.h" /** diff --git a/src/gallium/auxiliary/gallivm/lp_bld_logic.c b/src/gallium/auxiliary/gallivm/lp_bld_logic.c index fa65895367..f3df3dd138 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_logic.c +++ b/src/gallium/auxiliary/gallivm/lp_bld_logic.c @@ -74,13 +74,11 @@ lp_build_compare(LLVMBuilderRef builder, LLVMValueRef a, LLVMValueRef b) { - LLVMTypeRef vec_type = lp_build_vec_type(type); LLVMTypeRef int_vec_type = lp_build_int_vec_type(type); LLVMValueRef zeros = LLVMConstNull(int_vec_type); LLVMValueRef ones = LLVMConstAllOnes(int_vec_type); LLVMValueRef cond; LLVMValueRef res; - unsigned i; assert(func >= PIPE_FUNC_NEVER); assert(func <= PIPE_FUNC_ALWAYS); @@ -94,11 +92,12 @@ lp_build_compare(LLVMBuilderRef builder, /* XXX: It is not clear if we should use the ordered or unordered operators */ -#if !defined(HAVE_LLVM) || HAVE_LLVM < 0x0207 +#if HAVE_LLVM < 0x0207 #if defined(PIPE_ARCH_X86) || defined(PIPE_ARCH_X86_64) if(type.width * type.length == 128) { if(type.floating && util_cpu_caps.has_sse) { /* float[4] comparison */ + LLVMTypeRef vec_type = lp_build_vec_type(type); LLVMValueRef args[3]; unsigned cc; boolean swap; @@ -168,6 +167,7 @@ lp_build_compare(LLVMBuilderRef builder, const char *pcmpgt; LLVMValueRef args[2]; LLVMValueRef res; + LLVMTypeRef vec_type = lp_build_vec_type(type); switch (type.width) { case 8: @@ -259,11 +259,15 @@ lp_build_compare(LLVMBuilderRef builder, cond = LLVMBuildFCmp(builder, op, a, b, ""); res = LLVMBuildSExt(builder, cond, int_vec_type, ""); #else - res = LLVMGetUndef(int_vec_type); if (type.length == 1) { - res = LLVMBuildFCmp(builder, op, a, b, ""); + cond = LLVMBuildFCmp(builder, op, a, b, ""); + res = LLVMBuildSExt(builder, cond, int_vec_type, ""); } else { + unsigned i; + + res = LLVMGetUndef(int_vec_type); + debug_printf("%s: warning: using slow element-wise float" " vector comparison\n", __FUNCTION__); for (i = 0; i < type.length; ++i) { @@ -311,11 +315,15 @@ lp_build_compare(LLVMBuilderRef builder, cond = LLVMBuildICmp(builder, op, a, b, ""); res = LLVMBuildSExt(builder, cond, int_vec_type, ""); #else - res = LLVMGetUndef(int_vec_type); if (type.length == 1) { - res = LLVMBuildICmp(builder, op, a, b, ""); + cond = LLVMBuildICmp(builder, op, a, b, ""); + res = LLVMBuildSExt(builder, cond, int_vec_type, ""); } else { + unsigned i; + + res = LLVMGetUndef(int_vec_type); + debug_printf("%s: warning: using slow element-wise int" " vector comparison\n", __FUNCTION__); @@ -357,6 +365,8 @@ lp_build_cmp(struct lp_build_context *bld, /** * Return mask ? a : b; + * + * mask is a bitwise mask, composed of 0 or ~0 for each element. */ LLVMValueRef lp_build_select(struct lp_build_context *bld, @@ -371,6 +381,7 @@ lp_build_select(struct lp_build_context *bld, return a; if (type.length == 1) { + mask = LLVMBuildTrunc(bld->builder, mask, LLVMInt1Type(), ""); res = LLVMBuildSelect(bld->builder, mask, a, b, ""); } else { diff --git a/src/gallium/auxiliary/gallivm/lp_bld_logic.h b/src/gallium/auxiliary/gallivm/lp_bld_logic.h index a399ebf39e..b54ec13b70 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_logic.h +++ b/src/gallium/auxiliary/gallivm/lp_bld_logic.h @@ -37,7 +37,7 @@ #define LP_BLD_LOGIC_H -#include <llvm-c/Core.h> +#include "os/os_llvm.h" #include "pipe/p_defines.h" /* For PIPE_FUNC_xxx */ diff --git a/src/gallium/auxiliary/gallivm/lp_bld_pack.c b/src/gallium/auxiliary/gallivm/lp_bld_pack.c index 4c61d10749..23398f41f9 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_pack.c +++ b/src/gallium/auxiliary/gallivm/lp_bld_pack.c @@ -256,7 +256,7 @@ lp_build_pack2(LLVMBuilderRef builder, LLVMValueRef lo, LLVMValueRef hi) { -#if !(HAVE_LLVM >= 0x0207) +#if HAVE_LLVM < 0x0207 LLVMTypeRef src_vec_type = lp_build_vec_type(src_type); #endif LLVMTypeRef dst_vec_type = lp_build_vec_type(dst_type); diff --git a/src/gallium/auxiliary/gallivm/lp_bld_pack.h b/src/gallium/auxiliary/gallivm/lp_bld_pack.h index fb2a34984a..346a17d580 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_pack.h +++ b/src/gallium/auxiliary/gallivm/lp_bld_pack.h @@ -37,7 +37,7 @@ #define LP_BLD_PACK_H -#include <llvm-c/Core.h> +#include "os/os_llvm.h" struct lp_type; diff --git a/src/gallium/auxiliary/gallivm/lp_bld_sample.h b/src/gallium/auxiliary/gallivm/lp_bld_sample.h index 5b8da5dbf2..7f08bfaac1 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_sample.h +++ b/src/gallium/auxiliary/gallivm/lp_bld_sample.h @@ -36,7 +36,7 @@ #define LP_BLD_SAMPLE_H -#include <llvm-c/Core.h> +#include "os/os_llvm.h" struct pipe_texture; struct pipe_sampler_state; @@ -113,9 +113,9 @@ struct lp_sampler_dynamic_state unsigned unit); LLVMValueRef - (*stride)( struct lp_sampler_dynamic_state *state, - LLVMBuilderRef builder, - unsigned unit); + (*row_stride)( struct lp_sampler_dynamic_state *state, + LLVMBuilderRef builder, + unsigned unit); LLVMValueRef (*data_ptr)( struct lp_sampler_dynamic_state *state, diff --git a/src/gallium/auxiliary/gallivm/lp_bld_sample_soa.c b/src/gallium/auxiliary/gallivm/lp_bld_sample_soa.c index dadae61115..a410688350 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_sample_soa.c +++ b/src/gallium/auxiliary/gallivm/lp_bld_sample_soa.c @@ -48,6 +48,7 @@ #include "lp_bld_logic.h" #include "lp_bld_swizzle.h" #include "lp_bld_pack.h" +#include "lp_bld_flow.h" #include "lp_bld_format.h" #include "lp_bld_sample.h" @@ -138,6 +139,34 @@ lp_build_get_const_mipmap_level(struct lp_build_sample_context *bld, } +/** + * Dereference stride_array[mipmap_level] array to get a stride. + * Return stride as a vector. + */ +static LLVMValueRef +lp_build_get_level_stride_vec(struct lp_build_sample_context *bld, + LLVMValueRef stride_array, LLVMValueRef level) +{ + LLVMValueRef indexes[2], stride; + indexes[0] = LLVMConstInt(LLVMInt32Type(), 0, 0); + indexes[1] = level; + stride = LLVMBuildGEP(bld->builder, stride_array, indexes, 2, ""); + stride = LLVMBuildLoad(bld->builder, stride, ""); + stride = lp_build_broadcast_scalar(&bld->int_coord_bld, stride); + return stride; +} + + +/** Dereference stride_array[0] array to get a stride (as vector). */ +static LLVMValueRef +lp_build_get_const_level_stride_vec(struct lp_build_sample_context *bld, + LLVMValueRef stride_array, int level) +{ + LLVMValueRef lvl = LLVMConstInt(LLVMInt32Type(), level, 0); + return lp_build_get_level_stride_vec(bld, stride_array, lvl); +} + + static int texture_dims(enum pipe_texture_target tex) { @@ -1007,7 +1036,7 @@ lp_build_sample_image_nearest(struct lp_build_sample_context *bld, } } else { - y = NULL; + y = z = NULL; } /* @@ -1075,6 +1104,7 @@ lp_build_sample_image_linear(struct lp_build_sample_context *bld, } else { y0 = y1 = t_fpart = NULL; + z0 = z1 = r_fpart = NULL; } /* @@ -1092,9 +1122,11 @@ lp_build_sample_image_linear(struct lp_build_sample_context *bld, if (dims == 1) { /* Interpolate two samples from 1D image to produce one color */ - colors_out[chan] = lp_build_lerp(&bld->texel_bld, s_fpart, - neighbors[0][0][chan], - neighbors[0][1][chan]); + for (chan = 0; chan < 4; chan++) { + colors_out[chan] = lp_build_lerp(&bld->texel_bld, s_fpart, + neighbors[0][0][chan], + neighbors[0][1][chan]); + } } else { /* 2D/3D texture */ @@ -1169,6 +1201,93 @@ lp_build_sample_image_linear(struct lp_build_sample_context *bld, } +/** + * Generate code to do cube face selection and per-face texcoords. + */ +static void +lp_build_cube_lookup(struct lp_build_sample_context *bld, + LLVMValueRef s, + LLVMValueRef t, + LLVMValueRef r, + LLVMValueRef *face, + LLVMValueRef *face_s, + LLVMValueRef *face_t) +{ +#if 0 + struct lp_build_context *float_bld = &bld->float_bld; + LLVMValueRef rx, ry, rz; + LLVMValueRef arx, ary, arz; + LLVMValueRef sc, tc, ma; + LLVMValueRef c25 = LLVMConstReal(LLVMFloatType(), 0.25); + LLVMValueRef arx_ge_ary, arx_ge_arz; + LLVMValueRef ary_ge_arx, ary_ge_arz; + LLVMValueRef arx_ge_ary_arz, ary_ge_arx_arz; + LLVMValueRef rx_pos, ry_pos, rz_pos; + + assert(bld->coord_bld.type.length == 4); + + /* + * Use the average of the four pixel's texcoords to choose the face. + */ + rx = lp_build_mul(float_bld, c25, + lp_build_sum_vector(&bld->coord_bld, s)); + ry = lp_build_mul(float_bld, c25, + lp_build_sum_vector(&bld->coord_bld, t)); + rz = lp_build_mul(float_bld, c25, + lp_build_sum_vector(&bld->coord_bld, r)); + + arx = lp_build_abs(float_bld, rx); + ary = lp_build_abs(float_bld, ry); + arz = lp_build_abs(float_bld, rz); + + /* + * Compare sign/magnitude of rx,ry,rz to determine face + */ + arx_ge_ary = LLVMBuildFCmp(bld->builder, LLVMRealUGE, arx, ary, ""); + arx_ge_arz = LLVMBuildFCmp(bld->builder, LLVMRealUGE, arx, arz, ""); + ary_ge_arx = LLVMBuildFCmp(bld->builder, LLVMRealUGE, ary, arx, ""); + ary_ge_arz = LLVMBuildFCmp(bld->builder, LLVMRealUGE, ary, arz, ""); + + arx_ge_ary_arz = LLVMBuildAnd(bld->builder, arx_ge_ary, arx_ge_arz, ""); + ary_ge_arx_arz = LLVMBuildAnd(bld->builder, ary_ge_arx, ary_ge_arz, ""); + + rx_pos = LLVMBuildFCmp(bld->builder, LLVMRealUGE, rx, float_bld->zero, ""); + ry_pos = LLVMBuildFCmp(bld->builder, LLVMRealUGE, ry, float_bld->zero, ""); + rz_pos = LLVMBuildFCmp(bld->builder, LLVMRealUGE, rz, float_bld->zero, ""); + + { + struct lp_build_flow_context *flow_ctx; + struct lp_build_if_state if_ctx, if2_ctx; + + flow_ctx = lp_build_flow_create(bld->builder); + + lp_build_if(&if_ctx, flow_ctx, bld->builder, arx_ge_ary_arz); + { +#if 0 + lp_build_if(&if2_ctx, flow_ctx, bld->builder, rx_pos); + { + /* Positive X face */ + } + lp_build_else(&if2_ctx); + { + /* Negative X face */ + } + lp_build_endif(&if2_ctx); +#endif + } + lp_build_else(&if_ctx); + { + + + } + lp_build_endif(&if_ctx); + + lp_build_flow_destroy(flow_ctx); + } +#endif +} + + /** * General texture sampling codegen. @@ -1187,7 +1306,7 @@ lp_build_sample_general(struct lp_build_sample_context *bld, LLVMValueRef width_vec, LLVMValueRef height_vec, LLVMValueRef depth_vec, - LLVMValueRef row_stride_vec, + LLVMValueRef row_stride_array, LLVMValueRef img_stride_vec, LLVMValueRef data_array, LLVMValueRef *colors_out) @@ -1198,10 +1317,10 @@ lp_build_sample_general(struct lp_build_sample_context *bld, const int dims = texture_dims(bld->static_state->target); LLVMValueRef lod, lod_fpart; LLVMValueRef ilevel0, ilevel1, ilevel0_vec, ilevel1_vec; - LLVMValueRef width0_vec, height0_vec, depth0_vec; - LLVMValueRef width1_vec, height1_vec, depth1_vec; - LLVMValueRef row_stride0_vec, row_stride1_vec; - LLVMValueRef img_stride0_vec, img_stride1_vec; + LLVMValueRef width0_vec = NULL, height0_vec = NULL, depth0_vec = NULL; + LLVMValueRef width1_vec = NULL, height1_vec = NULL, depth1_vec = NULL; + LLVMValueRef row_stride0_vec = NULL, row_stride1_vec = NULL; + LLVMValueRef img_stride0_vec = NULL, img_stride1_vec = NULL; LLVMValueRef data_ptr0, data_ptr1; int chan; @@ -1245,9 +1364,12 @@ lp_build_sample_general(struct lp_build_sample_context *bld, width0_vec = lp_build_minify(bld, width_vec, ilevel0_vec); if (dims >= 2) { height0_vec = lp_build_minify(bld, height_vec, ilevel0_vec); - row_stride0_vec = lp_build_minify(bld, row_stride_vec, ilevel0_vec); + row_stride0_vec = lp_build_get_level_stride_vec(bld, row_stride_array, + ilevel0); if (dims == 3) { depth0_vec = lp_build_minify(bld, depth_vec, ilevel0_vec); + img_stride0_vec = lp_build_mul(&bld->int_coord_bld, + row_stride0_vec, height0_vec); } } if (mip_filter == PIPE_TEX_MIPFILTER_LINEAR) { @@ -1255,9 +1377,12 @@ lp_build_sample_general(struct lp_build_sample_context *bld, width1_vec = lp_build_minify(bld, width_vec, ilevel1_vec); if (dims >= 2) { height1_vec = lp_build_minify(bld, height_vec, ilevel1_vec); - row_stride1_vec = lp_build_minify(bld, row_stride_vec, ilevel1_vec); + row_stride1_vec = lp_build_get_level_stride_vec(bld, row_stride_array, + ilevel1); if (dims == 3) { depth1_vec = lp_build_minify(bld, depth_vec, ilevel1_vec); + img_stride1_vec = lp_build_mul(&bld->int_coord_bld, + row_stride1_vec, height1_vec); } } } @@ -1266,7 +1391,8 @@ lp_build_sample_general(struct lp_build_sample_context *bld, * Choose cube face, recompute texcoords. */ if (bld->static_state->target == PIPE_TEXTURE_CUBE) { - + LLVMValueRef face, face_s, face_t; + lp_build_cube_lookup(bld, s, t, r, &face, &face_s, &face_t); } /* @@ -1377,7 +1503,7 @@ lp_build_sample_2d_linear_aos(struct lp_build_sample_context *bld, LLVMValueRef t, LLVMValueRef width, LLVMValueRef height, - LLVMValueRef stride, + LLVMValueRef stride_array, LLVMValueRef data_array, LLVMValueRef *texel) { @@ -1394,6 +1520,7 @@ lp_build_sample_2d_linear_aos(struct lp_build_sample_context *bld, LLVMValueRef neighbors_hi[2][2]; LLVMValueRef packed, packed_lo, packed_hi; LLVMValueRef unswizzled[4]; + LLVMValueRef stride; lp_build_context_init(&i32, builder, lp_type_int_vec(32)); lp_build_context_init(&h16, builder, lp_type_ufixed(16)); @@ -1501,6 +1628,8 @@ lp_build_sample_2d_linear_aos(struct lp_build_sample_context *bld, t_fpart_hi = LLVMBuildShuffleVector(builder, t_fpart, h16.undef, shuffle_hi, ""); } + stride = lp_build_get_const_level_stride_vec(bld, stride_array, 0); + /* * Fetch the pixels as 4 x 32bit (rgba order might differ): * @@ -1625,7 +1754,7 @@ lp_build_sample_soa(LLVMBuilderRef builder, LLVMValueRef width, width_vec; LLVMValueRef height, height_vec; LLVMValueRef depth, depth_vec; - LLVMValueRef stride, stride_vec; + LLVMValueRef stride_array; LLVMValueRef data_array; LLVMValueRef s; LLVMValueRef t; @@ -1661,7 +1790,7 @@ lp_build_sample_soa(LLVMBuilderRef builder, width = dynamic_state->width(dynamic_state, builder, unit); height = dynamic_state->height(dynamic_state, builder, unit); depth = dynamic_state->depth(dynamic_state, builder, unit); - stride = dynamic_state->stride(dynamic_state, builder, unit); + stride_array = dynamic_state->row_stride(dynamic_state, builder, unit); data_array = dynamic_state->data_ptr(dynamic_state, builder, unit); /* Note that data_array is an array[level] of pointers to texture images */ @@ -1672,9 +1801,9 @@ lp_build_sample_soa(LLVMBuilderRef builder, width_vec = lp_build_broadcast_scalar(&bld.uint_coord_bld, width); height_vec = lp_build_broadcast_scalar(&bld.uint_coord_bld, height); depth_vec = lp_build_broadcast_scalar(&bld.uint_coord_bld, depth); - stride_vec = lp_build_broadcast_scalar(&bld.uint_coord_bld, stride); if (lp_format_is_rgba8(bld.format_desc) && + static_state->target == PIPE_TEXTURE_2D && static_state->min_img_filter == PIPE_TEX_FILTER_LINEAR && static_state->mag_img_filter == PIPE_TEX_FILTER_LINEAR && static_state->min_mip_filter == PIPE_TEX_MIPFILTER_NONE && @@ -1682,13 +1811,13 @@ lp_build_sample_soa(LLVMBuilderRef builder, is_simple_wrap_mode(static_state->wrap_t)) { /* special case */ lp_build_sample_2d_linear_aos(&bld, s, t, width_vec, height_vec, - stride_vec, data_array, texel); + stride_array, data_array, texel); } else { lp_build_sample_general(&bld, unit, s, t, r, width, height, depth, width_vec, height_vec, depth_vec, - stride_vec, NULL, data_array, + stride_array, NULL, data_array, texel); } diff --git a/src/gallium/auxiliary/gallivm/lp_bld_struct.h b/src/gallium/auxiliary/gallivm/lp_bld_struct.h index 740392f561..34478c10f5 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_struct.h +++ b/src/gallium/auxiliary/gallivm/lp_bld_struct.h @@ -37,7 +37,7 @@ #define LP_BLD_STRUCT_H -#include <llvm-c/Core.h> +#include "os/os_llvm.h" #include <llvm-c/Target.h> #include "util/u_debug.h" diff --git a/src/gallium/auxiliary/gallivm/lp_bld_swizzle.h b/src/gallium/auxiliary/gallivm/lp_bld_swizzle.h index b9472127a6..57b5cc079f 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_swizzle.h +++ b/src/gallium/auxiliary/gallivm/lp_bld_swizzle.h @@ -37,7 +37,7 @@ #define LP_BLD_SWIZZLE_H -#include <llvm-c/Core.h> +#include "os/os_llvm.h" struct lp_type; diff --git a/src/gallium/auxiliary/gallivm/lp_bld_tgsi.h b/src/gallium/auxiliary/gallivm/lp_bld_tgsi.h index eddb7a83fa..0f2f8a65b1 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_tgsi.h +++ b/src/gallium/auxiliary/gallivm/lp_bld_tgsi.h @@ -35,7 +35,7 @@ #ifndef LP_BLD_TGSI_H #define LP_BLD_TGSI_H -#include <llvm-c/Core.h> +#include "os/os_llvm.h" struct tgsi_token; diff --git a/src/gallium/auxiliary/gallivm/lp_bld_tgsi_soa.c b/src/gallium/auxiliary/gallivm/lp_bld_tgsi_soa.c index b3a0fe7d9b..5ec59d636c 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_tgsi_soa.c +++ b/src/gallium/auxiliary/gallivm/lp_bld_tgsi_soa.c @@ -41,6 +41,7 @@ #include "util/u_debug.h" #include "util/u_math.h" #include "util/u_memory.h" +#include "tgsi/tgsi_dump.h" #include "tgsi/tgsi_info.h" #include "tgsi/tgsi_parse.h" #include "tgsi/tgsi_util.h" diff --git a/src/gallium/auxiliary/gallivm/lp_bld_type.c b/src/gallium/auxiliary/gallivm/lp_bld_type.c index c327ba045a..796af88caa 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_type.c +++ b/src/gallium/auxiliary/gallivm/lp_bld_type.c @@ -58,7 +58,10 @@ LLVMTypeRef lp_build_vec_type(struct lp_type type) { LLVMTypeRef elem_type = lp_build_elem_type(type); - return LLVMVectorType(elem_type, type.length); + if (type.length == 1) + return elem_type; + else + return LLVMVectorType(elem_type, type.length); } @@ -115,6 +118,9 @@ lp_check_vec_type(struct lp_type type, LLVMTypeRef vec_type) if(!vec_type) return FALSE; + if (type.length == 1) + return lp_check_elem_type(type, vec_type); + if(LLVMGetTypeKind(vec_type) != LLVMVectorTypeKind) return FALSE; @@ -153,7 +159,10 @@ LLVMTypeRef lp_build_int_vec_type(struct lp_type type) { LLVMTypeRef elem_type = lp_build_int_elem_type(type); - return LLVMVectorType(elem_type, type.length); + if (type.length == 1) + return elem_type; + else + return LLVMVectorType(elem_type, type.length); } diff --git a/src/gallium/auxiliary/gallivm/lp_bld_type.h b/src/gallium/auxiliary/gallivm/lp_bld_type.h index 4daa904e63..5b351476ac 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_type.h +++ b/src/gallium/auxiliary/gallivm/lp_bld_type.h @@ -37,7 +37,7 @@ #define LP_BLD_TYPE_H -#include <llvm-c/Core.h> +#include "os/os_llvm.h" #include <pipe/p_compiler.h> diff --git a/src/gallium/auxiliary/os/os_llvm.h b/src/gallium/auxiliary/os/os_llvm.h new file mode 100644 index 0000000000..d5edfbfe92 --- /dev/null +++ b/src/gallium/auxiliary/os/os_llvm.h @@ -0,0 +1,47 @@ +/************************************************************************** + * + * Copyright 2010 VMware, 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, 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 VMWARE 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. + * + **************************************************************************/ + +/** + * @file + * Wrapper for LLVM header file #includes. + */ + + +#ifndef OS_LLVM_H +#define OS_LLVM_H + + +#include <llvm-c/Core.h> + + +/** Set version to 0 if missing to avoid #ifdef HAVE_LLVM everywhere */ +#ifndef HAVE_LLVM +#define HAVE_LLVM 0x0 +#endif + + +#endif /* OS_LLVM_H */ diff --git a/src/gallium/auxiliary/util/u_blitter.c b/src/gallium/auxiliary/util/u_blitter.c index 33d09085f0..36d582491f 100644 --- a/src/gallium/auxiliary/util/u_blitter.c +++ b/src/gallium/auxiliary/util/u_blitter.c @@ -106,9 +106,9 @@ struct blitter_context_priv struct blitter_context *util_blitter_create(struct pipe_context *pipe) { struct blitter_context_priv *ctx; - struct pipe_blend_state blend = { 0 }; - struct pipe_depth_stencil_alpha_state dsa = { { 0 } }; - struct pipe_rasterizer_state rs_state = { 0 }; + struct pipe_blend_state blend; + struct pipe_depth_stencil_alpha_state dsa; + struct pipe_rasterizer_state rs_state; struct pipe_sampler_state *sampler_state; struct pipe_vertex_element velem[2]; unsigned i; @@ -131,12 +131,14 @@ struct blitter_context *util_blitter_create(struct pipe_context *pipe) ctx->blitter.saved_num_sampler_states = ~0; /* blend state objects */ + memset(&blend, 0, sizeof(blend)); ctx->blend_keep_color = pipe->create_blend_state(pipe, &blend); blend.rt[0].colormask = PIPE_MASK_RGBA; ctx->blend_write_color = pipe->create_blend_state(pipe, &blend); /* depth stencil alpha state objects */ + memset(&dsa, 0, sizeof(dsa)); ctx->dsa_keep_depth_stencil = pipe->create_depth_stencil_alpha_state(pipe, &dsa); diff --git a/src/gallium/drivers/llvmpipe/lp_buffer.c b/src/gallium/drivers/llvmpipe/lp_buffer.c index dab20cb639..6e0f37393e 100644 --- a/src/gallium/drivers/llvmpipe/lp_buffer.c +++ b/src/gallium/drivers/llvmpipe/lp_buffer.c @@ -33,7 +33,6 @@ #include "lp_screen.h" #include "lp_buffer.h" -#include "state_tracker/sw_winsys.h" static void * llvmpipe_buffer_map(struct pipe_screen *screen, diff --git a/src/gallium/drivers/llvmpipe/lp_jit.c b/src/gallium/drivers/llvmpipe/lp_jit.c index 08c8f93794..5887613120 100644 --- a/src/gallium/drivers/llvmpipe/lp_jit.c +++ b/src/gallium/drivers/llvmpipe/lp_jit.c @@ -57,7 +57,8 @@ lp_jit_init_globals(struct llvmpipe_screen *screen) elem_types[LP_JIT_TEXTURE_HEIGHT] = LLVMInt32Type(); elem_types[LP_JIT_TEXTURE_DEPTH] = LLVMInt32Type(); elem_types[LP_JIT_TEXTURE_LAST_LEVEL] = LLVMInt32Type(); - elem_types[LP_JIT_TEXTURE_STRIDE] = LLVMInt32Type(); + elem_types[LP_JIT_TEXTURE_ROW_STRIDE] = + LLVMArrayType(LLVMInt32Type(), LP_MAX_TEXTURE_2D_LEVELS); elem_types[LP_JIT_TEXTURE_DATA] = LLVMArrayType(LLVMPointerType(LLVMInt8Type(), 0), LP_MAX_TEXTURE_2D_LEVELS); @@ -76,9 +77,9 @@ lp_jit_init_globals(struct llvmpipe_screen *screen) LP_CHECK_MEMBER_OFFSET(struct lp_jit_texture, last_level, screen->target, texture_type, LP_JIT_TEXTURE_LAST_LEVEL); - LP_CHECK_MEMBER_OFFSET(struct lp_jit_texture, stride, + LP_CHECK_MEMBER_OFFSET(struct lp_jit_texture, row_stride, screen->target, texture_type, - LP_JIT_TEXTURE_STRIDE); + LP_JIT_TEXTURE_ROW_STRIDE); LP_CHECK_MEMBER_OFFSET(struct lp_jit_texture, data, screen->target, texture_type, LP_JIT_TEXTURE_DATA); diff --git a/src/gallium/drivers/llvmpipe/lp_jit.h b/src/gallium/drivers/llvmpipe/lp_jit.h index 5cc7a12c03..13167ae3bf 100644 --- a/src/gallium/drivers/llvmpipe/lp_jit.h +++ b/src/gallium/drivers/llvmpipe/lp_jit.h @@ -51,7 +51,7 @@ struct lp_jit_texture uint32_t height; uint32_t depth; uint32_t last_level; - uint32_t stride; + uint32_t row_stride[LP_MAX_TEXTURE_2D_LEVELS]; const void *data[LP_MAX_TEXTURE_2D_LEVELS]; }; @@ -61,7 +61,7 @@ enum { LP_JIT_TEXTURE_HEIGHT, LP_JIT_TEXTURE_DEPTH, LP_JIT_TEXTURE_LAST_LEVEL, - LP_JIT_TEXTURE_STRIDE, + LP_JIT_TEXTURE_ROW_STRIDE, LP_JIT_TEXTURE_DATA }; diff --git a/src/gallium/drivers/llvmpipe/lp_screen.h b/src/gallium/drivers/llvmpipe/lp_screen.h index f4e62cbf08..d977f98cfa 100644 --- a/src/gallium/drivers/llvmpipe/lp_screen.h +++ b/src/gallium/drivers/llvmpipe/lp_screen.h @@ -34,9 +34,7 @@ #ifndef LP_SCREEN_H #define LP_SCREEN_H -#include <llvm-c/Core.h> -#include <llvm-c/Analysis.h> -#include <llvm-c/Target.h> +#include "os/os_llvm.h" #include <llvm-c/ExecutionEngine.h> #include "pipe/p_screen.h" diff --git a/src/gallium/drivers/llvmpipe/lp_setup.c b/src/gallium/drivers/llvmpipe/lp_setup.c index d6d37c4809..c870f89d01 100644 --- a/src/gallium/drivers/llvmpipe/lp_setup.c +++ b/src/gallium/drivers/llvmpipe/lp_setup.c @@ -471,13 +471,13 @@ lp_setup_set_sampler_textures( struct setup_context *setup, jit_tex->height = tex->height0; jit_tex->depth = tex->depth0; jit_tex->last_level = tex->last_level; - jit_tex->stride = lp_tex->stride[0]; if (!lp_tex->dt) { /* regular texture - setup array of mipmap level pointers */ int j; for (j = 0; j <= tex->last_level; j++) { jit_tex->data[j] = (ubyte *) lp_tex->data + lp_tex->level_offset[j]; + jit_tex->row_stride[j] = lp_tex->stride[j]; } } else { @@ -490,6 +490,7 @@ lp_setup_set_sampler_textures( struct setup_context *setup, struct sw_winsys *winsys = screen->winsys; jit_tex->data[0] = winsys->displaytarget_map(winsys, lp_tex->dt, PIPE_BUFFER_USAGE_CPU_READ); + jit_tex->row_stride[0] = lp_tex->stride[0]; assert(jit_tex->data[0]); } diff --git a/src/gallium/drivers/llvmpipe/lp_state.h b/src/gallium/drivers/llvmpipe/lp_state.h index 6dbdc195bf..a5a1a72074 100644 --- a/src/gallium/drivers/llvmpipe/lp_state.h +++ b/src/gallium/drivers/llvmpipe/lp_state.h @@ -31,7 +31,7 @@ #ifndef LP_STATE_H #define LP_STATE_H -#include <llvm-c/Core.h> +#include "os/os_llvm.h" #include "pipe/p_state.h" #include "tgsi/tgsi_scan.h" diff --git a/src/gallium/drivers/llvmpipe/lp_state_fs.c b/src/gallium/drivers/llvmpipe/lp_state_fs.c index c4b79dd415..9a8de0edfd 100644 --- a/src/gallium/drivers/llvmpipe/lp_state_fs.c +++ b/src/gallium/drivers/llvmpipe/lp_state_fs.c @@ -40,7 +40,7 @@ * - depth/stencil test (stencil TBI) * - blending * - * This file has only the glue to assembly the fragment pipeline. The actual + * This file has only the glue to assemble the fragment pipeline. The actual * plumbing of converting Gallium state into LLVM IR is done elsewhere, in the * lp_bld_*.[ch] files, and in a complete generic and reusable way. Here we * muster the LLVM JIT execution engine to create a function that follows an @@ -95,6 +95,9 @@ #include "lp_tex_sample.h" +#include <llvm-c/Analysis.h> + + static const unsigned char quad_offset_x[4] = {0, 1, 0, 1}; static const unsigned char quad_offset_y[4] = {0, 0, 1, 1}; diff --git a/src/gallium/drivers/llvmpipe/lp_test.h b/src/gallium/drivers/llvmpipe/lp_test.h index a9b99945f9..1df9897898 100644 --- a/src/gallium/drivers/llvmpipe/lp_test.h +++ b/src/gallium/drivers/llvmpipe/lp_test.h @@ -41,7 +41,7 @@ #include <stdio.h> #include <float.h> -#include <llvm-c/Core.h> +#include "os/os_llvm.h" #include <llvm-c/Analysis.h> #include <llvm-c/ExecutionEngine.h> #include <llvm-c/Target.h> diff --git a/src/gallium/drivers/llvmpipe/lp_test_format.c b/src/gallium/drivers/llvmpipe/lp_test_format.c index d05157991b..2c4d7fb6e1 100644 --- a/src/gallium/drivers/llvmpipe/lp_test_format.c +++ b/src/gallium/drivers/llvmpipe/lp_test_format.c @@ -29,7 +29,7 @@ #include <stdlib.h> #include <stdio.h> -#include <llvm-c/Core.h> +#include "os/os_llvm.h" #include <llvm-c/Analysis.h> #include <llvm-c/ExecutionEngine.h> #include <llvm-c/Target.h> diff --git a/src/gallium/drivers/llvmpipe/lp_tex_sample.h b/src/gallium/drivers/llvmpipe/lp_tex_sample.h index cb59a94464..799df182b6 100644 --- a/src/gallium/drivers/llvmpipe/lp_tex_sample.h +++ b/src/gallium/drivers/llvmpipe/lp_tex_sample.h @@ -29,7 +29,7 @@ #define LP_TEX_SAMPLE_H -#include <llvm-c/Core.h> +#include "os/os_llvm.h" struct lp_sampler_static_state; diff --git a/src/gallium/drivers/llvmpipe/lp_tex_sample_llvm.c b/src/gallium/drivers/llvmpipe/lp_tex_sample_llvm.c index 5a3cf37d6d..662508af61 100644 --- a/src/gallium/drivers/llvmpipe/lp_tex_sample_llvm.c +++ b/src/gallium/drivers/llvmpipe/lp_tex_sample_llvm.c @@ -147,7 +147,7 @@ LP_LLVM_TEXTURE_MEMBER(width, LP_JIT_TEXTURE_WIDTH, TRUE) LP_LLVM_TEXTURE_MEMBER(height, LP_JIT_TEXTURE_HEIGHT, TRUE) LP_LLVM_TEXTURE_MEMBER(depth, LP_JIT_TEXTURE_DEPTH, TRUE) LP_LLVM_TEXTURE_MEMBER(last_level, LP_JIT_TEXTURE_LAST_LEVEL, TRUE) -LP_LLVM_TEXTURE_MEMBER(stride, LP_JIT_TEXTURE_STRIDE, TRUE) +LP_LLVM_TEXTURE_MEMBER(row_stride, LP_JIT_TEXTURE_ROW_STRIDE, FALSE) LP_LLVM_TEXTURE_MEMBER(data_ptr, LP_JIT_TEXTURE_DATA, FALSE) @@ -204,7 +204,7 @@ lp_llvm_sampler_soa_create(const struct lp_sampler_static_state *static_state, sampler->dynamic_state.base.height = lp_llvm_texture_height; sampler->dynamic_state.base.depth = lp_llvm_texture_depth; sampler->dynamic_state.base.last_level = lp_llvm_texture_last_level; - sampler->dynamic_state.base.stride = lp_llvm_texture_stride; + sampler->dynamic_state.base.row_stride = lp_llvm_texture_row_stride; sampler->dynamic_state.base.data_ptr = lp_llvm_texture_data_ptr; sampler->dynamic_state.static_state = static_state; sampler->dynamic_state.context_ptr = context_ptr; diff --git a/src/gallium/drivers/softpipe/Makefile b/src/gallium/drivers/softpipe/Makefile index 1c6e4ae076..239655d628 100644 --- a/src/gallium/drivers/softpipe/Makefile +++ b/src/gallium/drivers/softpipe/Makefile @@ -8,6 +8,7 @@ C_SOURCES = \ sp_fs_sse.c \ sp_buffer.c \ sp_clear.c \ + sp_fence.c \ sp_flush.c \ sp_query.c \ sp_context.c \ diff --git a/src/gallium/drivers/softpipe/SConscript b/src/gallium/drivers/softpipe/SConscript index 27ab00b036..9949a53adf 100644 --- a/src/gallium/drivers/softpipe/SConscript +++ b/src/gallium/drivers/softpipe/SConscript @@ -11,6 +11,7 @@ softpipe = env.ConvenienceLibrary( 'sp_clear.c', 'sp_context.c', 'sp_draw_arrays.c', + 'sp_fence.c', 'sp_flush.c', 'sp_prim_vbuf.c', 'sp_setup.c', diff --git a/src/gallium/drivers/softpipe/sp_fence.c b/src/gallium/drivers/softpipe/sp_fence.c new file mode 100644 index 0000000000..66c5214113 --- /dev/null +++ b/src/gallium/drivers/softpipe/sp_fence.c @@ -0,0 +1,70 @@ +/************************************************************************** + * + * Copyright 2010 VMware, 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, 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 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 + * THE COPYRIGHT HOLDERS, AUTHORS 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. + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + **************************************************************************/ + + +#include "pipe/p_screen.h" +#include "util/u_debug.h" +#include "sp_fence.h" + + +static void +softpipe_fence_reference(struct pipe_screen *screen, + struct pipe_fence_handle **ptr, + struct pipe_fence_handle *fence) +{ + assert(!*ptr); + assert(!fence); +} + + +static int +softpipe_fence_signalled(struct pipe_screen *screen, + struct pipe_fence_handle *fence, + unsigned flags) +{ + assert(!fence); + return 0; +} + + +static int +softpipe_fence_finish(struct pipe_screen *screen, + struct pipe_fence_handle *fence, + unsigned flags) +{ + assert(!fence); + return 0; +} + + +void +softpipe_init_screen_fence_funcs(struct pipe_screen *screen) +{ + screen->fence_reference = softpipe_fence_reference; + screen->fence_finish = softpipe_fence_finish; + screen->fence_signalled = softpipe_fence_signalled; +} diff --git a/src/gallium/drivers/softpipe/sp_fence.h b/src/gallium/drivers/softpipe/sp_fence.h new file mode 100644 index 0000000000..39c33243bd --- /dev/null +++ b/src/gallium/drivers/softpipe/sp_fence.h @@ -0,0 +1,40 @@ +/************************************************************************** + * + * Copyright 2010 VMware, 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, 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 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 + * THE COPYRIGHT HOLDERS, AUTHORS 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. + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + **************************************************************************/ + + +#ifndef SP_FENCE_H_ +#define SP_FENCE_H_ + + +struct pipe_screen; + + +void +softpipe_init_screen_fence_funcs(struct pipe_screen *screen); + + +#endif /* SP_FENCE_H_ */ diff --git a/src/gallium/drivers/softpipe/sp_screen.c b/src/gallium/drivers/softpipe/sp_screen.c index e0a2ef604e..d62bfa3d63 100644 --- a/src/gallium/drivers/softpipe/sp_screen.c +++ b/src/gallium/drivers/softpipe/sp_screen.c @@ -36,6 +36,7 @@ #include "sp_screen.h" #include "sp_context.h" #include "sp_buffer.h" +#include "sp_fence.h" #include "sp_public.h" @@ -239,6 +240,7 @@ softpipe_create_screen(struct sw_winsys *winsys) softpipe_init_screen_texture_funcs(&screen->base); softpipe_init_screen_buffer_funcs(&screen->base); + softpipe_init_screen_fence_funcs(&screen->base); return &screen->base; } diff --git a/src/gallium/drivers/softpipe/sp_tex_sample.c b/src/gallium/drivers/softpipe/sp_tex_sample.c index ef7ccf4189..fa9e19b282 100644 --- a/src/gallium/drivers/softpipe/sp_tex_sample.c +++ b/src/gallium/drivers/softpipe/sp_tex_sample.c @@ -1614,7 +1614,6 @@ sample_cube(struct tgsi_sampler *tgsi_sampler, struct sp_sampler_varient *samp = sp_sampler_varient(tgsi_sampler); unsigned j; float ssss[4], tttt[4]; - unsigned face; /* major axis @@ -1628,7 +1627,8 @@ sample_cube(struct tgsi_sampler *tgsi_sampler, -rz TEXTURE_CUBE_MAP_NEGATIVE_Z_EXT -rx -ry rz */ - /* First choose the cube face. + /* Choose the cube face and compute new s/t coords for the 2D face. + * * Use the same cube face for all four pixels in the quad. * * This isn't ideal, but if we want to use a different cube face @@ -1647,85 +1647,37 @@ sample_cube(struct tgsi_sampler *tgsi_sampler, const float arx = fabsf(rx), ary = fabsf(ry), arz = fabsf(rz); if (arx >= ary && arx >= arz) { - if (rx >= 0.0F) { - face = PIPE_TEX_FACE_POS_X; - } - else { - face = PIPE_TEX_FACE_NEG_X; + float sign = (rx >= 0.0F) ? 1.0F : -1.0F; + uint face = (rx >= 0.0F) ? PIPE_TEX_FACE_POS_X : PIPE_TEX_FACE_NEG_X; + for (j = 0; j < QUAD_SIZE; j++) { + const float ima = -0.5F / fabsf(s[j]); + ssss[j] = sign * p[j] * ima + 0.5F; + tttt[j] = t[j] * ima + 0.5F; + samp->faces[j] = face; } } else if (ary >= arx && ary >= arz) { - if (ry >= 0.0F) { - face = PIPE_TEX_FACE_POS_Y; - } - else { - face = PIPE_TEX_FACE_NEG_Y; + float sign = (ry >= 0.0F) ? 1.0F : -1.0F; + uint face = (ry >= 0.0F) ? PIPE_TEX_FACE_POS_Y : PIPE_TEX_FACE_NEG_Y; + for (j = 0; j < QUAD_SIZE; j++) { + const float ima = -0.5F / fabsf(t[j]); + ssss[j] = -s[j] * ima + 0.5F; + tttt[j] = sign * -p[j] * ima + 0.5F; + samp->faces[j] = face; } } else { - if (rz > 0.0F) { - face = PIPE_TEX_FACE_POS_Z; - } - else { - face = PIPE_TEX_FACE_NEG_Z; + float sign = (rz >= 0.0F) ? 1.0F : -1.0F; + uint face = (rz >= 0.0F) ? PIPE_TEX_FACE_POS_Z : PIPE_TEX_FACE_NEG_Z; + for (j = 0; j < QUAD_SIZE; j++) { + const float ima = -0.5 / fabsf(p[j]); + ssss[j] = sign * -s[j] * ima + 0.5F; + tttt[j] = t[j] * ima + 0.5F; + samp->faces[j] = face; } } } - /* Now compute the 2D _face_ texture coords from the - * 3D _cube_ texture coords. - */ - for (j = 0; j < QUAD_SIZE; j++) { - const float rx = s[j], ry = t[j], rz = p[j]; - const float arx = fabsf(rx), ary = fabsf(ry), arz = fabsf(rz); - float sc, tc, ma; - - switch (face) { - case PIPE_TEX_FACE_POS_X: - sc = -rz; - tc = -ry; - ma = arx; - break; - case PIPE_TEX_FACE_NEG_X: - sc = rz; - tc = -ry; - ma = arx; - break; - case PIPE_TEX_FACE_POS_Y: - sc = rx; - tc = rz; - ma = ary; - break; - case PIPE_TEX_FACE_NEG_Y: - sc = rx; - tc = -rz; - ma = ary; - break; - case PIPE_TEX_FACE_POS_Z: - sc = rx; - tc = -ry; - ma = arz; - break; - case PIPE_TEX_FACE_NEG_Z: - sc = -rx; - tc = -ry; - ma = arz; - break; - default: - assert(0 && "bad cube face"); - sc = 0.0F; - tc = 0.0F; - ma = 0.0F; - } - - { - const float ima = 1.0 / ma; - ssss[j] = ( sc * ima + 1.0F ) * 0.5F; - tttt[j] = ( tc * ima + 1.0F ) * 0.5F; - samp->faces[j] = face; - } - } - /* In our little pipeline, the compare stage is next. If compare * is not active, this will point somewhere deeper into the * pipeline, eg. to mip_filter or even img_filter. diff --git a/src/gallium/drivers/svga/svgadump/svga_shader_dump.c b/src/gallium/drivers/svga/svgadump/svga_shader_dump.c index 705ca29e8f..4ee1bf2c35 100644 --- a/src/gallium/drivers/svga/svgadump/svga_shader_dump.c +++ b/src/gallium/drivers/svga/svgadump/svga_shader_dump.c @@ -360,7 +360,9 @@ dump_dstreg(struct sh_dstreg dstreg, union { struct sh_reg reg; struct sh_dstreg dstreg; - } u = { { 0 } }; + } u; + + memset(&u, 0, sizeof(u)); assert( (dstreg.modifier & (SVGA3DDSTMOD_SATURATE | SVGA3DDSTMOD_PARTIALPRECISION)) == dstreg.modifier ); diff --git a/src/gallium/state_trackers/dri/dri_extensions.c b/src/gallium/state_trackers/dri/dri_extensions.c index 1259813a41..800677a2d1 100644 --- a/src/gallium/state_trackers/dri/dri_extensions.c +++ b/src/gallium/state_trackers/dri/dri_extensions.c @@ -33,110 +33,14 @@ #include "dri_context.h" #include "state_tracker/st_context.h" -#define need_GL_ARB_map_buffer_range -#define need_GL_ARB_multisample -#define need_GL_ARB_occlusion_query -#define need_GL_ARB_point_parameters -#define need_GL_ARB_provoking_vertex -#define need_GL_ARB_shader_objects -#define need_GL_ARB_texture_compression -#define need_GL_ARB_vertex_array_object -#define need_GL_ARB_vertex_buffer_object -#define need_GL_ARB_vertex_program -#define need_GL_ARB_vertex_shader -#define need_GL_ARB_window_pos -#define need_GL_EXT_blend_color -#define need_GL_EXT_blend_equation_separate -#define need_GL_EXT_blend_func_separate -#define need_GL_EXT_blend_minmax -#define need_GL_EXT_cull_vertex -#define need_GL_EXT_draw_buffers2 -#define need_GL_EXT_fog_coord -#define need_GL_EXT_framebuffer_object -#define need_GL_EXT_multi_draw_arrays -#define need_GL_EXT_provoking_vertex -#define need_GL_EXT_secondary_color -#define need_GL_EXT_stencil_two_side -#define need_GL_APPLE_vertex_array_object -#define need_GL_NV_vertex_program -#define need_GL_VERSION_2_0 -#define need_GL_VERSION_2_1 -#include "main/remap_helper.h" #include "utils.h" -/** - * Extension strings exported by the driver. - */ -static const struct dri_extension card_extensions[] = { - {"GL_ARB_fragment_shader", NULL}, - {"GL_ARB_map_buffer_range", GL_ARB_map_buffer_range_functions}, - {"GL_ARB_multisample", GL_ARB_multisample_functions}, - {"GL_ARB_multitexture", NULL}, - {"GL_ARB_occlusion_query", GL_ARB_occlusion_query_functions}, - {"GL_ARB_pixel_buffer_object", NULL}, - {"GL_ARB_provoking_vertex", GL_ARB_provoking_vertex_functions}, - {"GL_ARB_point_parameters", GL_ARB_point_parameters_functions}, - {"GL_ARB_shading_language_100", GL_VERSION_2_0_functions }, - {"GL_ARB_shading_language_120", GL_VERSION_2_1_functions }, - {"GL_ARB_shader_objects", GL_ARB_shader_objects_functions}, - {"GL_ARB_texture_border_clamp", NULL}, - {"GL_ARB_texture_compression", GL_ARB_texture_compression_functions}, - {"GL_ARB_texture_cube_map", NULL}, - {"GL_ARB_texture_env_add", NULL}, - {"GL_ARB_texture_env_combine", NULL}, - {"GL_ARB_texture_env_dot3", NULL}, - {"GL_ARB_texture_mirrored_repeat", NULL}, - {"GL_ARB_texture_non_power_of_two", NULL}, - {"GL_ARB_texture_rectangle", NULL}, - {"GL_ARB_vertex_array_object", GL_ARB_vertex_array_object_functions}, - {"GL_ARB_vertex_buffer_object", GL_ARB_vertex_buffer_object_functions}, - {"GL_ARB_vertex_shader", GL_ARB_vertex_shader_functions}, - {"GL_ARB_vertex_program", GL_ARB_vertex_program_functions}, - {"GL_ARB_window_pos", GL_ARB_window_pos_functions}, - {"GL_EXT_blend_color", GL_EXT_blend_color_functions}, - {"GL_EXT_blend_equation_separate", GL_EXT_blend_equation_separate_functions}, - {"GL_EXT_blend_func_separate", GL_EXT_blend_func_separate_functions}, - {"GL_EXT_blend_minmax", GL_EXT_blend_minmax_functions}, - {"GL_EXT_blend_subtract", NULL}, - {"GL_EXT_cull_vertex", GL_EXT_cull_vertex_functions}, - {"GL_EXT_draw_buffers2", GL_EXT_draw_buffers2_functions}, - {"GL_EXT_fog_coord", GL_EXT_fog_coord_functions}, - {"GL_EXT_framebuffer_object", GL_EXT_framebuffer_object_functions}, - {"GL_EXT_multi_draw_arrays", GL_EXT_multi_draw_arrays_functions}, - {"GL_EXT_packed_depth_stencil", NULL}, - {"GL_EXT_pixel_buffer_object", NULL}, - {"GL_EXT_provoking_vertex", GL_EXT_provoking_vertex_functions}, - {"GL_EXT_secondary_color", GL_EXT_secondary_color_functions}, - {"GL_EXT_stencil_two_side", GL_EXT_stencil_two_side_functions}, - {"GL_EXT_stencil_wrap", NULL}, - {"GL_EXT_texture_edge_clamp", NULL}, - {"GL_EXT_texture_env_combine", NULL}, - {"GL_EXT_texture_env_dot3", NULL}, - {"GL_EXT_texture_filter_anisotropic", NULL}, - {"GL_EXT_texture_lod_bias", NULL}, - {"GL_3DFX_texture_compression_FXT1", NULL}, - {"GL_APPLE_client_storage", NULL}, - {"GL_APPLE_vertex_array_object", GL_APPLE_vertex_array_object_functions}, - {"GL_MESA_pack_invert", NULL}, - {"GL_MESA_ycbcr_texture", NULL}, - {"GL_NV_blend_square", NULL}, - {"GL_NV_vertex_program", GL_NV_vertex_program_functions}, - {"GL_NV_vertex_program1_1", NULL}, - {"GL_SGIS_generate_mipmap", NULL}, - {NULL, NULL} -}; - void dri_init_extensions(struct dri_context *ctx) { - /* The card_extensions list should be pruned according to the - * capabilities of the pipe_screen. This is actually something - * that can/should be done inside st_create_context(). - * XXX Not pruning is very bogus. Always all these extensions above - * will be advertized, regardless what st_init_extensions - * (which depends on the pipe cap bits) does. - */ - driInitExtensions(ctx->st->ctx, card_extensions, GL_TRUE); + /* New extensions should be added in mesa/state_tracker/st_extensions.c + * and not in this file. */ + driInitExtensions(ctx->st->ctx, NULL, GL_FALSE); } /* vim: set sw=3 ts=8 sts=3 expandtab: */ diff --git a/src/gallium/state_trackers/xorg/xorg_crtc.c b/src/gallium/state_trackers/xorg/xorg_crtc.c index 7268a79a23..f60e9abb76 100644 --- a/src/gallium/state_trackers/xorg/xorg_crtc.c +++ b/src/gallium/state_trackers/xorg/xorg_crtc.c @@ -39,6 +39,7 @@ #include <xf86.h> #include <xf86i2c.h> #include <xf86Crtc.h> +#include <cursorstr.h> #include "xorg_tracker.h" #include "xf86Modes.h" diff --git a/src/mesa/drivers/dri/r200/r200_blit.c b/src/mesa/drivers/dri/r200/r200_blit.c index 55e8d32512..e187fc0f61 100644 --- a/src/mesa/drivers/dri/r200/r200_blit.c +++ b/src/mesa/drivers/dri/r200/r200_blit.c @@ -50,6 +50,9 @@ unsigned r200_check_blit(gl_format mesa_format) case MESA_FORMAT_A8: case MESA_FORMAT_L8: case MESA_FORMAT_I8: + /* swizzled */ + case MESA_FORMAT_RGBA8888: + case MESA_FORMAT_RGBA8888_REV: break; default: return 0; @@ -88,7 +91,8 @@ static inline void emit_vtx_state(struct r200_context *r200) } static void inline emit_tx_setup(struct r200_context *r200, - gl_format mesa_format, + gl_format src_mesa_format, + gl_format dst_mesa_format, struct radeon_bo *bo, intptr_t offset, unsigned width, @@ -103,7 +107,7 @@ static void inline emit_tx_setup(struct r200_context *r200, assert(offset % 32 == 0); /* XXX others? BE/LE? */ - switch (mesa_format) { + switch (src_mesa_format) { case MESA_FORMAT_ARGB8888: txformat |= R200_TXFORMAT_ARGB8888 | R200_TXFORMAT_ALPHA_IN_MAP; break; @@ -139,20 +143,130 @@ static void inline emit_tx_setup(struct r200_context *r200, break; } - BEGIN_BATCH(28); - OUT_BATCH_REGVAL(RADEON_PP_CNTL, RADEON_TEX_0_ENABLE | RADEON_TEX_BLEND_0_ENABLE); + switch (dst_mesa_format) { + case MESA_FORMAT_ARGB8888: + case MESA_FORMAT_XRGB8888: + case MESA_FORMAT_RGB565: + case MESA_FORMAT_ARGB4444: + case MESA_FORMAT_ARGB1555: + case MESA_FORMAT_A8: + case MESA_FORMAT_L8: + case MESA_FORMAT_I8: + default: + /* no swizzle required */ + BEGIN_BATCH(10); + OUT_BATCH_REGVAL(RADEON_PP_CNTL, (RADEON_TEX_0_ENABLE | + RADEON_TEX_BLEND_0_ENABLE)); + OUT_BATCH_REGVAL(R200_PP_TXCBLEND_0, (R200_TXC_ARG_A_ZERO | + R200_TXC_ARG_B_ZERO | + R200_TXC_ARG_C_R0_COLOR | + R200_TXC_OP_MADD)); + OUT_BATCH_REGVAL(R200_PP_TXCBLEND2_0, (R200_TXC_CLAMP_0_1 | + R200_TXC_OUTPUT_REG_R0)); + OUT_BATCH_REGVAL(R200_PP_TXABLEND_0, (R200_TXA_ARG_A_ZERO | + R200_TXA_ARG_B_ZERO | + R200_TXA_ARG_C_R0_ALPHA | + R200_TXA_OP_MADD)); + OUT_BATCH_REGVAL(R200_PP_TXABLEND2_0, (R200_TXA_CLAMP_0_1 | + R200_TXA_OUTPUT_REG_R0)); + END_BATCH(); + break; + case MESA_FORMAT_RGBA8888: + BEGIN_BATCH(10); + OUT_BATCH_REGVAL(RADEON_PP_CNTL, (RADEON_TEX_0_ENABLE | + RADEON_TEX_BLEND_0_ENABLE)); + OUT_BATCH_REGVAL(R200_PP_TXCBLEND_0, (R200_TXC_ARG_A_ZERO | + R200_TXC_ARG_B_ZERO | + R200_TXC_ARG_C_R0_COLOR | + R200_TXC_OP_MADD)); + OUT_BATCH_REGVAL(R200_PP_TXCBLEND2_0, (R200_TXC_CLAMP_0_1 | + R200_TXC_OUTPUT_ROTATE_GBA | + R200_TXC_OUTPUT_REG_R0)); + OUT_BATCH_REGVAL(R200_PP_TXABLEND_0, (R200_TXA_ARG_A_ZERO | + R200_TXA_ARG_B_ZERO | + R200_TXA_ARG_C_R0_ALPHA | + R200_TXA_OP_MADD)); + OUT_BATCH_REGVAL(R200_PP_TXABLEND2_0, (R200_TXA_CLAMP_0_1 | + (R200_TXA_REPL_RED << R200_TXA_REPL_ARG_C_SHIFT) | + R200_TXA_OUTPUT_REG_R0)); + END_BATCH(); + break; + case MESA_FORMAT_RGBA8888_REV: + BEGIN_BATCH(34); + OUT_BATCH_REGVAL(RADEON_PP_CNTL, (RADEON_TEX_0_ENABLE | + RADEON_TEX_BLEND_0_ENABLE | + RADEON_TEX_BLEND_1_ENABLE | + RADEON_TEX_BLEND_2_ENABLE | + RADEON_TEX_BLEND_3_ENABLE)); + /* r1.r = r0.b */ + OUT_BATCH_REGVAL(R200_PP_TXCBLEND_0, (R200_TXC_ARG_A_ZERO | + R200_TXC_ARG_B_ZERO | + R200_TXC_ARG_C_R0_COLOR | + R200_TXC_OP_MADD)); + OUT_BATCH_REGVAL(R200_PP_TXCBLEND2_0, (R200_TXC_CLAMP_0_1 | + R200_TXC_OUTPUT_MASK_R | + (R200_TXC_REPL_BLUE << R200_TXC_REPL_ARG_C_SHIFT) | + R200_TXC_OUTPUT_REG_R1)); + /* r1.a = r0.a */ + OUT_BATCH_REGVAL(R200_PP_TXABLEND_0, (R200_TXA_ARG_A_ZERO | + R200_TXA_ARG_B_ZERO | + R200_TXA_ARG_C_R0_ALPHA | + R200_TXA_OP_MADD)); + OUT_BATCH_REGVAL(R200_PP_TXABLEND2_0, (R200_TXA_CLAMP_0_1 | + R200_TXA_OUTPUT_REG_R1)); + /* r1.g = r0.g */ + OUT_BATCH_REGVAL(R200_PP_TXCBLEND_1, (R200_TXC_ARG_A_ZERO | + R200_TXC_ARG_B_ZERO | + R200_TXC_ARG_C_R0_COLOR | + R200_TXC_OP_MADD)); + OUT_BATCH_REGVAL(R200_PP_TXCBLEND2_1, (R200_TXC_CLAMP_0_1 | + R200_TXC_OUTPUT_MASK_G | + (R200_TXC_REPL_GREEN << R200_TXC_REPL_ARG_C_SHIFT) | + R200_TXC_OUTPUT_REG_R1)); + /* r1.a = r0.a */ + OUT_BATCH_REGVAL(R200_PP_TXABLEND_1, (R200_TXA_ARG_A_ZERO | + R200_TXA_ARG_B_ZERO | + R200_TXA_ARG_C_R0_ALPHA | + R200_TXA_OP_MADD)); + OUT_BATCH_REGVAL(R200_PP_TXABLEND2_1, (R200_TXA_CLAMP_0_1 | + R200_TXA_OUTPUT_REG_R1)); + /* r1.b = r0.r */ + OUT_BATCH_REGVAL(R200_PP_TXCBLEND_2, (R200_TXC_ARG_A_ZERO | + R200_TXC_ARG_B_ZERO | + R200_TXC_ARG_C_R0_COLOR | + R200_TXC_OP_MADD)); + OUT_BATCH_REGVAL(R200_PP_TXCBLEND2_2, (R200_TXC_CLAMP_0_1 | + R200_TXC_OUTPUT_MASK_B | + (R200_TXC_REPL_RED << R200_TXC_REPL_ARG_C_SHIFT) | + R200_TXC_OUTPUT_REG_R1)); + /* r1.a = r0.a */ + OUT_BATCH_REGVAL(R200_PP_TXABLEND_2, (R200_TXA_ARG_A_ZERO | + R200_TXA_ARG_B_ZERO | + R200_TXA_ARG_C_R0_ALPHA | + R200_TXA_OP_MADD)); + OUT_BATCH_REGVAL(R200_PP_TXABLEND2_2, (R200_TXA_CLAMP_0_1 | + R200_TXA_OUTPUT_REG_R1)); + /* r0.rgb = r1.rgb */ + OUT_BATCH_REGVAL(R200_PP_TXCBLEND_3, (R200_TXC_ARG_A_ZERO | + R200_TXC_ARG_B_ZERO | + R200_TXC_ARG_C_R1_COLOR | + R200_TXC_OP_MADD)); + OUT_BATCH_REGVAL(R200_PP_TXCBLEND2_3, (R200_TXC_CLAMP_0_1 | + R200_TXC_OUTPUT_REG_R0)); + /* r0.a = r1.a */ + OUT_BATCH_REGVAL(R200_PP_TXABLEND_3, (R200_TXA_ARG_A_ZERO | + R200_TXA_ARG_B_ZERO | + R200_TXA_ARG_C_R1_ALPHA | + R200_TXA_OP_MADD)); + OUT_BATCH_REGVAL(R200_PP_TXABLEND2_3, (R200_TXA_CLAMP_0_1 | + R200_TXA_OUTPUT_REG_R0)); + END_BATCH(); + break; + } + + BEGIN_BATCH(18); OUT_BATCH_REGVAL(R200_PP_CNTL_X, 0); OUT_BATCH_REGVAL(R200_PP_TXMULTI_CTL_0, 0); - OUT_BATCH_REGVAL(R200_PP_TXCBLEND_0, (R200_TXC_ARG_A_ZERO | - R200_TXC_ARG_B_ZERO | - R200_TXC_ARG_C_R0_COLOR | - R200_TXC_OP_MADD)); - OUT_BATCH_REGVAL(R200_PP_TXCBLEND2_0, R200_TXC_CLAMP_0_1 | R200_TXC_OUTPUT_REG_R0); - OUT_BATCH_REGVAL(R200_PP_TXABLEND_0, (R200_TXA_ARG_A_ZERO | - R200_TXA_ARG_B_ZERO | - R200_TXA_ARG_C_R0_ALPHA | - R200_TXA_OP_MADD)); - OUT_BATCH_REGVAL(R200_PP_TXABLEND2_0, R200_TXA_CLAMP_0_1 | R200_TXA_OUTPUT_REG_R0); OUT_BATCH_REGVAL(R200_PP_TXFILTER_0, (R200_CLAMP_S_CLAMP_LAST | R200_CLAMP_T_CLAMP_LAST | R200_MAG_FILTER_NEAREST | @@ -161,7 +275,7 @@ static void inline emit_tx_setup(struct r200_context *r200, OUT_BATCH_REGVAL(R200_PP_TXFORMAT_X_0, 0); OUT_BATCH_REGVAL(R200_PP_TXSIZE_0, ((width - 1) | ((height - 1) << RADEON_TEX_VSIZE_SHIFT))); - OUT_BATCH_REGVAL(R200_PP_TXPITCH_0, pitch * _mesa_get_format_bytes(mesa_format) - 32); + OUT_BATCH_REGVAL(R200_PP_TXPITCH_0, pitch * _mesa_get_format_bytes(src_mesa_format) - 32); OUT_BATCH_REGSEQ(R200_PP_TXOFFSET_0, 1); OUT_BATCH_RELOC(0, bo, 0, RADEON_GEM_DOMAIN_GTT|RADEON_GEM_DOMAIN_VRAM, 0, 0); @@ -185,6 +299,8 @@ static inline void emit_cb_setup(struct r200_context *r200, switch (mesa_format) { case MESA_FORMAT_ARGB8888: case MESA_FORMAT_XRGB8888: + case MESA_FORMAT_RGBA8888: + case MESA_FORMAT_RGBA8888_REV: dst_format = RADEON_COLOR_FORMAT_ARGB8888; break; case MESA_FORMAT_RGB565: @@ -401,15 +517,15 @@ unsigned r200_blit(GLcontext *ctx, /* Flush is needed to make sure that source buffer has correct data */ radeonFlush(r200->radeon.glCtx); - rcommonEnsureCmdBufSpace(&r200->radeon, 78, __FUNCTION__); + rcommonEnsureCmdBufSpace(&r200->radeon, 102, __FUNCTION__); if (!validate_buffers(r200, src_bo, dst_bo)) return GL_FALSE; /* 14 */ emit_vtx_state(r200); - /* 28 */ - emit_tx_setup(r200, src_mesaformat, src_bo, src_offset, src_width, src_height, src_pitch); + /* 52 */ + emit_tx_setup(r200, src_mesaformat, dst_mesaformat, src_bo, src_offset, src_width, src_height, src_pitch); /* 22 */ emit_cb_setup(r200, dst_bo, dst_offset, dst_mesaformat, dst_pitch, dst_width, dst_height); /* 14 */ diff --git a/src/mesa/drivers/dri/r200/r200_reg.h b/src/mesa/drivers/dri/r200/r200_reg.h index 59115212ce..2e52c7c13f 100644 --- a/src/mesa/drivers/dri/r200/r200_reg.h +++ b/src/mesa/drivers/dri/r200/r200_reg.h @@ -1265,6 +1265,10 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. #define R200_TXC_OUTPUT_MASK_G (5 << 20) #define R200_TXC_OUTPUT_MASK_B (6 << 20) #define R200_TXC_OUTPUT_MASK_NONE (7 << 20) +#define R200_TXC_OUTPUT_ROTATE_RGB (0 << 24) +#define R200_TXC_OUTPUT_ROTATE_ARG (1 << 24) +#define R200_TXC_OUTPUT_ROTATE_GBA (2 << 24) +#define R200_TXC_OUTPUT_ROTATE_RGA (3 << 24) #define R200_TXC_REPL_NORMAL 0 #define R200_TXC_REPL_RED 1 #define R200_TXC_REPL_GREEN 2 diff --git a/src/mesa/drivers/windows/gdi/wmesa.c b/src/mesa/drivers/windows/gdi/wmesa.c index 91ddc3615a..22b0c46b4f 100644 --- a/src/mesa/drivers/windows/gdi/wmesa.c +++ b/src/mesa/drivers/windows/gdi/wmesa.c @@ -248,16 +248,6 @@ static void wmesa_flush(GLcontext *ctx) */ /* - * Set the color index used to clear the color buffer. - */ -static void clear_index(GLcontext *ctx, GLuint index) -{ - WMesaContext pwc = wmesa_context(ctx); - /* Note that indexed mode is not supported yet */ - pwc->clearColorRef = RGB(0,0,0); -} - -/* * Set the color used to clear the color buffer. */ static void clear_color(GLcontext *ctx, const GLfloat color[4]) @@ -482,7 +472,7 @@ static void write_rgba_span_front(const GLcontext *ctx, }; } BGRA; BGRA *bgra, c; - int i; + GLuint i; if (n < 16) { // the value 16 is just guessed y=FLIP(y); @@ -827,9 +817,9 @@ static void read_rgba_span_32(const GLcontext *ctx, lpdw = ((LPDWORD)(pwfb->pbPixels + pwfb->ScanWidth * y)) + x; for (i=0; i<n; i++) { pixel = lpdw[i]; - rgba[i][RCOMP] = (pixel & 0x00ff0000) >> 16; - rgba[i][GCOMP] = (pixel & 0x0000ff00) >> 8; - rgba[i][BCOMP] = (pixel & 0x000000ff); + rgba[i][RCOMP] = (GLubyte)((pixel & 0x00ff0000) >> 16); + rgba[i][GCOMP] = (GLubyte)((pixel & 0x0000ff00) >> 8); + rgba[i][BCOMP] = (GLubyte)(pixel & 0x000000ff); rgba[i][ACOMP] = 255; } } @@ -851,9 +841,9 @@ static void read_rgba_pixels_32(const GLcontext *ctx, GLint y2 = FLIP(y[i]); lpdw = ((LPDWORD)(pwfb->pbPixels + pwfb->ScanWidth * y2)) + x[i]; pixel = *lpdw; - rgba[i][RCOMP] = (pixel & 0x00ff0000) >> 16; - rgba[i][GCOMP] = (pixel & 0x0000ff00) >> 8; - rgba[i][BCOMP] = (pixel & 0x000000ff); + rgba[i][RCOMP] = (GLubyte)((pixel & 0x00ff0000) >> 16); + rgba[i][GCOMP] = (GLubyte)((pixel & 0x0000ff00) >> 8); + rgba[i][BCOMP] = (GLubyte)(pixel & 0x000000ff); rgba[i][ACOMP] = 255; } } @@ -1271,7 +1261,7 @@ wmesa_renderbuffer_storage(GLcontext *ctx, * on if we're drawing to the front or back color buffer. */ void wmesa_set_renderbuffer_funcs(struct gl_renderbuffer *rb, int pixelformat, - BYTE cColorBits, int double_buffer) + int cColorBits, int double_buffer) { if (double_buffer) { /* back buffer */ @@ -1483,7 +1473,6 @@ WMesaContext WMesaCreateContext(HDC hDC, functions.GetBufferSize = wmesa_get_buffer_size; functions.Flush = wmesa_flush; functions.Clear = clear; - functions.ClearIndex = clear_index; functions.ClearColor = clear_color; functions.ResizeBuffers = wmesa_resize_buffers; functions.Viewport = wmesa_viewport; diff --git a/src/mesa/drivers/windows/gdi/wmesadef.h b/src/mesa/drivers/windows/gdi/wmesadef.h index 83a42e6082..1c0e245111 100644 --- a/src/mesa/drivers/windows/gdi/wmesadef.h +++ b/src/mesa/drivers/windows/gdi/wmesadef.h @@ -27,7 +27,7 @@ struct wmesa_framebuffer HDC hDC; int pixelformat; GLuint ScanWidth; - BYTE cColorBits; + int cColorBits; /* back buffer DIB fields */ HDC dib_hDC; BITMAPINFO bmi; diff --git a/src/mesa/es/state_tracker/st_cb_drawtex.c b/src/mesa/es/state_tracker/st_cb_drawtex.c index 0a5cba9d92..f75f4861a2 100644 --- a/src/mesa/es/state_tracker/st_cb_drawtex.c +++ b/src/mesa/es/state_tracker/st_cb_drawtex.c @@ -120,8 +120,11 @@ st_DrawTex(GLcontext *ctx, GLfloat x, GLfloat y, GLfloat z, GLboolean emitColor; uint semantic_names[2 + MAX_TEXTURE_UNITS]; uint semantic_indexes[2 + MAX_TEXTURE_UNITS]; + struct pipe_vertex_element velements[2 + MAX_TEXTURE_UNITS]; GLbitfield inputs = VERT_BIT_POS; + st_validate_state(st); + /* determine if we need vertex color */ if (ctx->FragmentProgram._Current->Base.InputsRead & FRAG_BIT_COL0) emitColor = GL_TRUE; @@ -232,6 +235,7 @@ st_DrawTex(GLcontext *ctx, GLfloat x, GLfloat y, GLfloat z, cso_save_viewport(cso); cso_save_vertex_shader(cso); + cso_save_vertex_elements(cso); { void *vs = lookup_shader(pipe, numAttribs, @@ -239,6 +243,14 @@ st_DrawTex(GLcontext *ctx, GLfloat x, GLfloat y, GLfloat z, cso_set_vertex_shader_handle(cso, vs); } + for (i = 0; i < numAttribs; i++) { + velements[i].src_offset = i * 4 * sizeof(float); + velements[i].instance_divisor = 0; + velements[i].vertex_buffer_index = 0; + velements[i].src_format = PIPE_FORMAT_R32G32B32A32_FLOAT; + } + cso_set_vertex_elements(cso, numAttribs, velements); + /* viewport state: viewport matching window dims */ { const struct gl_framebuffer *fb = st->ctx->DrawBuffer; @@ -270,6 +282,7 @@ st_DrawTex(GLcontext *ctx, GLfloat x, GLfloat y, GLfloat z, /* restore state */ cso_restore_viewport(cso); cso_restore_vertex_shader(cso); + cso_restore_vertex_elements(cso); } diff --git a/src/mesa/state_tracker/st_extensions.c b/src/mesa/state_tracker/st_extensions.c index bc240e0893..290ee36b0f 100644 --- a/src/mesa/state_tracker/st_extensions.c +++ b/src/mesa/state_tracker/st_extensions.c @@ -172,6 +172,7 @@ void st_init_extensions(struct st_context *st) ctx->Extensions.ARB_vertex_array_object = GL_TRUE; ctx->Extensions.ARB_vertex_buffer_object = GL_TRUE; ctx->Extensions.ARB_vertex_program = GL_TRUE; + ctx->Extensions.ARB_window_pos = GL_TRUE; ctx->Extensions.EXT_blend_color = GL_TRUE; ctx->Extensions.EXT_blend_func_separate = GL_TRUE; @@ -196,10 +197,17 @@ void st_init_extensions(struct st_context *st) ctx->Extensions.APPLE_vertex_array_object = GL_TRUE; + ctx->Extensions.MESA_pack_invert = GL_TRUE; + ctx->Extensions.NV_blend_square = GL_TRUE; ctx->Extensions.NV_texgen_reflection = GL_TRUE; ctx->Extensions.NV_texture_env_combine4 = GL_TRUE; ctx->Extensions.NV_texture_rectangle = GL_TRUE; +#if 0 + /* possibly could support the following two */ + ctx->Extensions.NV_vertex_program = GL_TRUE; + ctx->Extensions.NV_vertex_program1_1 = GL_TRUE; +#endif #if FEATURE_OES_draw_texture ctx->Extensions.OES_draw_texture = GL_TRUE; diff --git a/src/mesa/swrast/s_span.c b/src/mesa/swrast/s_span.c index 29f070686f..687c8eb0bf 100644 --- a/src/mesa/swrast/s_span.c +++ b/src/mesa/swrast/s_span.c @@ -60,6 +60,7 @@ void _swrast_span_default_attribs(GLcontext *ctx, SWspan *span) { + GLchan r, g, b, a; /* Z*/ { const GLfloat depthMax = ctx->DrawBuffer->_DepthMaxF; @@ -80,7 +81,6 @@ _swrast_span_default_attribs(GLcontext *ctx, SWspan *span) span->attrStepY[FRAG_ATTRIB_WPOS][3] = 0.0; /* primary color, or color index */ - GLchan r, g, b, a; UNCLAMPED_FLOAT_TO_CHAN(r, ctx->Current.RasterColor[0]); UNCLAMPED_FLOAT_TO_CHAN(g, ctx->Current.RasterColor[1]); UNCLAMPED_FLOAT_TO_CHAN(b, ctx->Current.RasterColor[2]); |