diff options
Diffstat (limited to 'src/mesa')
47 files changed, 732 insertions, 551 deletions
diff --git a/src/mesa/Makefile.mgw b/src/mesa/Makefile.mgw index ebec5c055a..3b52834bd1 100644 --- a/src/mesa/Makefile.mgw +++ b/src/mesa/Makefile.mgw @@ -1,5 +1,5 @@ # Mesa 3-D graphics library -# Version: 5.1 +# Version: 7.0 # # Copyright (C) 1999-2003 Brian Paul All Rights Reserved. # @@ -26,6 +26,16 @@ # Email : dborca@users.sourceforge.net # Web : http://www.geocities.com/dborca +# MinGW core-gl makefile updated for Mesa 7.0 +# +# updated : by Heromyth, on 2007-7-21 +# Email : zxpmyth@yahoo.com.cn +# Bugs : 1) All the default settings work fine. But the setting X86=1 can't work. +# The others havn't been tested yet. +# 2) The generated DLLs are *not* compatible with the ones built +# with the other compilers like VC8, especially for GLUT. +# 3) Although more tests are needed, it can be used individually! + # # Available options: @@ -52,7 +62,6 @@ # - .PHONY: all clean .INTERMEDIATE: x86/gen_matypes.exe .SUFFIXES: .rc .res @@ -60,6 +69,8 @@ # Set this to the prefix of your build tools, i.e. mingw32- TOOLS_PREFIX = mingw32- + + TOP = ../.. GLIDE ?= $(TOP)/glide3 LIBDIR = $(TOP)/lib @@ -71,11 +82,25 @@ else GL_IMP = libopengl32.a endif -LDLIBS = -lgdi32 +GL_DEF = gl.def + +include $(TOP)/configs/config.mgw +GL_USING_STDCALL ?= 1 + +MESA_LIB = libmesa.a + +LDLIBS = -lgdi32 -luser32 -liberty +LDFLAGS = -Wl,--out-implib=$(LIBDIR)/$(GL_IMP) -Wl,--output-def=$(LIBDIR)/gl.def CC = $(TOOLS_PREFIX)gcc -CFLAGS += -DBUILD_GL32 -D_OPENGL32_ -CFLAGS += $(INCLUDE_DIRS) +CFLAGS += -DBUILD_GL32 -D_OPENGL32_ -D_DLL -DMESA_MINWARN -DNDEBUG -D_USRDLL -DGDI_EXPORTS + +ifeq ($(GL_USING_STDCALL),1) + LDFLAGS += -Wl,--add-stdcall-alias +else + CFLAGS += -DGL_NO_STDCALL +endif + CFLAGS += -DUSE_EXTERNAL_DXTN_LIB=1 ifeq ($(FX),1) CFLAGS += -I$(GLIDE)/include -DFX @@ -104,6 +129,8 @@ endif include sources +CFLAGS += $(INCLUDE_DIRS) + ifeq ($(X86),1) CFLAGS += -DUSE_X86_ASM CFLAGS += -DUSE_MMX_ASM @@ -140,10 +167,9 @@ RESOURCE = $(GL_RES:.rc=.res) .c.o: $(CC) -o $@ $(CFLAGS) -c $< -.S.o: - $(CC) -o $@ $(CFLAGS) -c $< .s.o: $(CC) -o $@ $(CFLAGS) -x assembler-with-cpp -c $< + .rc.res: windres -o $@ -Irc -Ocoff $< @@ -153,9 +179,8 @@ $(LIBDIR): mkdir -p $(LIBDIR) $(LIBDIR)/$(GL_DLL) $(LIBDIR)/$(GL_IMP): $(OBJECTS) $(RESOURCE) - $(TOOLS_PREFIX)dllwrap -o $(LIBDIR)/$(GL_DLL) --output-lib $(LIBDIR)/$(GL_IMP) \ - --target i386-mingw32 --def $(GL_DEF) -Wl,-enable-stdcall-fixup \ - $^ $(LDLIBS) + $(CC) -shared -fPIC -o $(LIBDIR)/$(GL_DLL) $(LDFLAGS) \ + $^ $(LDLIBS) $(X86_OBJECTS): x86/matypes.h @@ -187,17 +212,21 @@ tnl/t_vtx_x86_gcc.o: tnl/t_vtx_x86_gcc.S $(CC) -o $@ $(CFLAGS) -DSTDCALL_API -c $< clean: - -$(call UNLINK,array_cache/*.o) -$(call UNLINK,glapi/*.o) -$(call UNLINK,main/*.o) -$(call UNLINK,math/*.o) + -$(call UNLINK,vbo/*.o) -$(call UNLINK,shader/*.o) + -$(call UNLINK,shader/slang/*.o) + -$(call UNLINK,shader/grammar/*.o) -$(call UNLINK,sparc/*.o) -$(call UNLINK,ppc/*.o) -$(call UNLINK,swrast/*.o) -$(call UNLINK,swrast_setup/*.o) -$(call UNLINK,tnl/*.o) -$(call UNLINK,x86/*.o) + -$(call UNLINK,x86/rtasm/*.o) + -$(call UNLINK,x86-64/*.o) -$(call UNLINK,drivers/common/*.o) -$(call UNLINK,drivers/glide/*.o) -$(call UNLINK,drivers/windows/fx/*.o) diff --git a/src/mesa/drivers/dri/i810/i810state.c b/src/mesa/drivers/dri/i810/i810state.c index 3ad25282d9..e0d5b2b487 100644 --- a/src/mesa/drivers/dri/i810/i810state.c +++ b/src/mesa/drivers/dri/i810/i810state.c @@ -380,7 +380,10 @@ static void i810CullFaceFrontFace(GLcontext *ctx, GLenum unused) static void i810LineWidth( GLcontext *ctx, GLfloat widthf ) { i810ContextPtr imesa = I810_CONTEXT( ctx ); - int width = (int)ctx->Line._Width; + /* AA, non-AA limits are same */ + const int width = (int) CLAMP(ctx->Line.Width, + ctx->Const.MinLineWidth, + ctx->Const.MaxLineWidth); imesa->LcsLineWidth = 0; if (width & 1) imesa->LcsLineWidth |= LCS_LINEWIDTH_1_0; @@ -396,7 +399,10 @@ static void i810LineWidth( GLcontext *ctx, GLfloat widthf ) static void i810PointSize( GLcontext *ctx, GLfloat sz ) { i810ContextPtr imesa = I810_CONTEXT( ctx ); - int size = (int)ctx->Point._Size; + /* AA, non-AA limits are same */ + const int size = (int) CLAMP(ctx->Point.Size, + ctx->Const.MinPointSize, + ctx->Const.MaxPointSize); imesa->LcsPointSize = 0; if (size & 1) imesa->LcsPointSize |= LCS_LINEWIDTH_1_0; diff --git a/src/mesa/drivers/dri/i810/i810tris.c b/src/mesa/drivers/dri/i810/i810tris.c index 3e09427bb9..2c4ee06633 100644 --- a/src/mesa/drivers/dri/i810/i810tris.c +++ b/src/mesa/drivers/dri/i810/i810tris.c @@ -112,7 +112,9 @@ static __inline__ void i810_draw_quad( i810ContextPtr imesa, static __inline__ void i810_draw_point( i810ContextPtr imesa, i810VertexPtr tmp ) { - GLfloat sz = imesa->glCtx->Point._Size * .5; + GLfloat sz = 0.5 * CLAMP(imesa->glCtx->Point.Size, + imesa->glCtx->Const.MinPointSize, + imesa->glCtx->Const.MaxPointSize); int vertsize = imesa->vertex_size; GLuint *vb = i810AllocDmaLow( imesa, 2 * 4 * vertsize ); int j; diff --git a/src/mesa/drivers/dri/i915tex/intel_batchbuffer.h b/src/mesa/drivers/dri/i915tex/intel_batchbuffer.h index 59261f7274..212f130101 100644 --- a/src/mesa/drivers/dri/i915tex/intel_batchbuffer.h +++ b/src/mesa/drivers/dri/i915tex/intel_batchbuffer.h @@ -9,7 +9,7 @@ struct intel_context; #define BATCH_SZ 16384 #define BATCH_RESERVED 16 -#define MAX_RELOCS 400 +#define MAX_RELOCS 4096 #define INTEL_BATCH_NO_CLIPRECTS 0x1 #define INTEL_BATCH_CLIPRECTS 0x2 diff --git a/src/mesa/drivers/dri/i915tex/intel_fbo.c b/src/mesa/drivers/dri/i915tex/intel_fbo.c index 2dd3617f93..a09db46163 100644 --- a/src/mesa/drivers/dri/i915tex/intel_fbo.c +++ b/src/mesa/drivers/dri/i915tex/intel_fbo.c @@ -509,7 +509,7 @@ intel_new_renderbuffer(GLcontext * ctx, GLuint name) */ static void intel_bind_framebuffer(GLcontext * ctx, GLenum target, - struct gl_framebuffer *fb) + struct gl_framebuffer *fb, struct gl_framebuffer *fbread) { if (target == GL_FRAMEBUFFER_EXT || target == GL_DRAW_FRAMEBUFFER_EXT) { intel_draw_buffer(ctx, fb); diff --git a/src/mesa/drivers/dri/i965/brw_sf_state.c b/src/mesa/drivers/dri/i965/brw_sf_state.c index bfac52d765..9a6e5f5f19 100644 --- a/src/mesa/drivers/dri/i965/brw_sf_state.c +++ b/src/mesa/drivers/dri/i965/brw_sf_state.c @@ -173,7 +173,8 @@ static void upload_sf_unit( struct brw_context *brw ) /* _NEW_LINE */ - sf.sf6.line_width = brw->attribs.Line->_Width * (1<<1); + /* XXX use ctx->Const.Min/MaxLineWidth here */ + sf.sf6.line_width = CLAMP(brw->attribs.Line->Width, 1.0, 5.0) * (1<<1); sf.sf6.line_endcap_aa_region_width = 1; if (brw->attribs.Line->SmoothFlag) @@ -183,7 +184,8 @@ static void upload_sf_unit( struct brw_context *brw ) /* _NEW_POINT */ sf.sf6.point_rast_rule = 1; /* opengl conventions */ - sf.sf7.point_size = brw->attribs.Point->_Size * (1<<3); + /* XXX clamp max depends on AA vs. non-AA */ + sf.sf7.point_size = CLAMP(brw->attribs.Point->Size, 1.0, 3.0) * (1<<3); sf.sf7.use_point_size_state = !brw->attribs.Point->_Attenuated; /* might be BRW_NEW_PRIMITIVE if we have to adjust pv for polygons: diff --git a/src/mesa/drivers/dri/mach64/mach64_native_vb.c b/src/mesa/drivers/dri/mach64/mach64_native_vb.c index 75cf0e2ed2..248fa2a9a2 100644 --- a/src/mesa/drivers/dri/mach64/mach64_native_vb.c +++ b/src/mesa/drivers/dri/mach64/mach64_native_vb.c @@ -103,7 +103,7 @@ void TAG(translate_vertex)(GLcontext *ctx, assert( p + 1 - (CARD32 *)src == 10 ); - dst->pointSize = ctx->Point._Size; + dst->pointSize = ctx->Point.Size; } diff --git a/src/mesa/drivers/dri/mach64/mach64_tris.c b/src/mesa/drivers/dri/mach64/mach64_tris.c index 08cc1849a1..369f610442 100644 --- a/src/mesa/drivers/dri/mach64/mach64_tris.c +++ b/src/mesa/drivers/dri/mach64/mach64_tris.c @@ -673,7 +673,10 @@ static __inline void mach64_draw_line( mach64ContextPtr mmesa, #if MACH64_NATIVE_VTXFMT GLcontext *ctx = mmesa->glCtx; const GLuint vertsize = mmesa->vertex_size; - GLint width = (GLint)(mmesa->glCtx->Line._Width * 2.0); /* 2 fractional bits for hardware */ + /* 2 fractional bits for hardware: */ + const int width = (int) (2.0 * CLAMP(mmesa->glCtx->Line.Width, + mmesa->glCtx->Const.MinLineWidth, + mmesa->glCtx->Const.MaxLineWidth)); GLfloat ooa; GLuint *pxy0, *pxy1; GLuint xy0old, xy0, xy1old, xy1; @@ -691,9 +694,6 @@ static __inline void mach64_draw_line( mach64ContextPtr mmesa, mach64_print_vertex( ctx, v1 ); } - if( !width ) - width = 1; /* round to the nearest supported width */ - pxy0 = &v0->ui[xyoffset]; xy0old = *pxy0; xy0 = LE32_IN( &xy0old ); @@ -961,7 +961,10 @@ static __inline void mach64_draw_point( mach64ContextPtr mmesa, #if MACH64_NATIVE_VTXFMT GLcontext *ctx = mmesa->glCtx; const GLuint vertsize = mmesa->vertex_size; - GLint sz = (GLint)(mmesa->glCtx->Point._Size * 2.0); /* 2 fractional bits for hardware */ + /* 2 fractional bits for hardware: */ + GLint sz = (GLint) (2.0 * CLAMP(mmesa->glCtx->Point.Size, + ctx->Const.MinPointSize, + ctx->Const.MaxPointSize)); GLfloat ooa; GLuint *pxy; GLuint xyold, xy; diff --git a/src/mesa/drivers/dri/mga/mgatris.c b/src/mesa/drivers/dri/mga/mgatris.c index 2b7ea05b14..91b413ae76 100644 --- a/src/mesa/drivers/dri/mga/mgatris.c +++ b/src/mesa/drivers/dri/mga/mgatris.c @@ -104,8 +104,10 @@ static void __inline__ mga_draw_quad( mgaContextPtr mmesa, static __inline__ void mga_draw_point( mgaContextPtr mmesa, mgaVertexPtr tmp ) { - GLfloat sz = mmesa->glCtx->Point._Size * .5; - int vertex_size = mmesa->vertex_size; + const GLfloat sz = 0.5 * CLAMP(mmesa->glCtx->Point.Size, + mmesa->glCtx->Const.MinPointSize, + mmesa->glCtx->Const.MaxPointSize); + const int vertex_size = mmesa->vertex_size; GLuint *vb = mgaAllocDmaLow( mmesa, 6 * 4 * vertex_size ); int j; @@ -165,7 +167,9 @@ static __inline__ void mga_draw_line( mgaContextPtr mmesa, GLuint vertex_size = mmesa->vertex_size; GLuint *vb = mgaAllocDmaLow( mmesa, 6 * 4 * vertex_size ); GLfloat dx, dy, ix, iy; - GLfloat width = mmesa->glCtx->Line._Width; + const GLfloat width = CLAMP(mmesa->glCtx->Line.Width, + mmesa->glCtx->Const.MinLineWidth, + mmesa->glCtx->Const.MaxLineWidth); GLint j; #if 0 diff --git a/src/mesa/drivers/dri/nouveau/nouveau_buffers.c b/src/mesa/drivers/dri/nouveau/nouveau_buffers.c index f98d666563..d498f616c9 100644 --- a/src/mesa/drivers/dri/nouveau/nouveau_buffers.c +++ b/src/mesa/drivers/dri/nouveau/nouveau_buffers.c @@ -390,9 +390,12 @@ nouveauNewRenderbuffer(GLcontext *ctx, GLuint name) } static void -nouveauBindFramebuffer(GLcontext *ctx, GLenum target, struct gl_framebuffer *fb) +nouveauBindFramebuffer(GLcontext *ctx, GLenum target, + struct gl_framebuffer *fb, struct gl_framebuffer *fbread) { - nouveau_build_framebuffer(ctx, fb); + if (target == GL_FRAMEBUFFER_EXT || target == GL_DRAW_FRAMEBUFFER_EXT) { + nouveau_build_framebuffer(ctx, fb); + } } static void diff --git a/src/mesa/drivers/dri/r200/r200_state.c b/src/mesa/drivers/dri/r200/r200_state.c index 2115799b9b..1d975ecd57 100644 --- a/src/mesa/drivers/dri/r200/r200_state.c +++ b/src/mesa/drivers/dri/r200/r200_state.c @@ -772,9 +772,11 @@ static void r200LineWidth( GLcontext *ctx, GLfloat widthf ) R200_STATECHANGE( rmesa, set ); /* Line width is stored in U6.4 format. + * Same min/max limits for AA, non-AA lines. */ rmesa->hw.lin.cmd[LIN_SE_LINE_WIDTH] &= ~0xffff; - rmesa->hw.lin.cmd[LIN_SE_LINE_WIDTH] |= (GLuint)(ctx->Line._Width * 16.0); + rmesa->hw.lin.cmd[LIN_SE_LINE_WIDTH] |= (GLuint) + (CLAMP(widthf, ctx->Const.MinLineWidth, ctx->Const.MaxLineWidth) * 16.0); if ( widthf > 1.0 ) { rmesa->hw.set.cmd[SET_SE_CNTL] |= R200_WIDELINE_ENABLE; diff --git a/src/mesa/drivers/dri/r300/r300_context.h b/src/mesa/drivers/dri/r300/r300_context.h index 0349bac9a2..be6909724a 100644 --- a/src/mesa/drivers/dri/r300/r300_context.h +++ b/src/mesa/drivers/dri/r300/r300_context.h @@ -568,7 +568,7 @@ struct r300_vertex_shader_fragment { union { GLuint d[VSF_MAX_FRAGMENT_LENGTH]; float f[VSF_MAX_FRAGMENT_LENGTH]; - struct r300_vertprog_instruction i[VSF_MAX_FRAGMENT_LENGTH / 4]; + GLuint i[VSF_MAX_FRAGMENT_LENGTH]; } body; }; diff --git a/src/mesa/drivers/dri/r300/r300_state.c b/src/mesa/drivers/dri/r300/r300_state.c index 6789efd428..088216c76e 100644 --- a/src/mesa/drivers/dri/r300/r300_state.c +++ b/src/mesa/drivers/dri/r300/r300_state.c @@ -733,8 +733,8 @@ static void r300Fogfv(GLcontext * ctx, GLenum pname, const GLfloat * param) static void r300PointSize(GLcontext * ctx, GLfloat size) { r300ContextPtr r300 = R300_CONTEXT(ctx); - - size = ctx->Point._Size; + /* same size limits for AA, non-AA points */ + size = CLAMP(size, ctx->Const.MinPointSize, ctx->Const.MaxPointSize); R300_STATECHANGE(r300, ps); r300->hw.ps.cmd[R300_PS_POINTSIZE] = @@ -749,8 +749,9 @@ static void r300LineWidth(GLcontext * ctx, GLfloat widthf) { r300ContextPtr r300 = R300_CONTEXT(ctx); - widthf = ctx->Line._Width; - + widthf = CLAMP(widthf, + ctx->Const.MinPointSize, + ctx->Const.MaxPointSize); R300_STATECHANGE(r300, lcntl); r300->hw.lcntl.cmd[1] = R300_LINE_CNT_HO | R300_LINE_CNT_VE | (int)(widthf * 6.0); @@ -1566,30 +1567,31 @@ static void r300SetupDefaultVertexProgram(r300ContextPtr rmesa) for (i = VERT_ATTRIB_POS; i < VERT_ATTRIB_MAX; i++) { if (rmesa->state.sw_tcl_inputs[i] != -1) { - prog->program.body.i[program_end].opcode = EASY_VSF_OP(MUL, o_reg++, ALL, RESULT); - prog->program.body.i[program_end].src[0] = VSF_REG(rmesa->state.sw_tcl_inputs[i]); - prog->program.body.i[program_end].src[1] = VSF_ATTR_UNITY(rmesa->state.sw_tcl_inputs[i]); - prog->program.body.i[program_end].src[2] = VSF_UNITY(rmesa->state.sw_tcl_inputs[i]); - program_end++; + prog->program.body.i[program_end + 0] = EASY_VSF_OP(MUL, o_reg++, ALL, RESULT); + prog->program.body.i[program_end + 1] = VSF_REG(rmesa->state.sw_tcl_inputs[i]); + prog->program.body.i[program_end + 2] = VSF_ATTR_UNITY(rmesa->state.sw_tcl_inputs[i]); + prog->program.body.i[program_end + 3] = VSF_UNITY(rmesa->state.sw_tcl_inputs[i]); + program_end += 4; } } - prog->program.length = program_end * 4; + prog->program.length = program_end; - r300SetupVertexProgramFragment(rmesa, R300_PVS_UPLOAD_PROGRAM, &(prog->program)); + r300SetupVertexProgramFragment(rmesa, R300_PVS_UPLOAD_PROGRAM, + &(prog->program)); inst_count = (prog->program.length / 4) - 1; R300_STATECHANGE(rmesa, pvs); rmesa->hw.pvs.cmd[R300_PVS_CNTL_1] = - (0 << R300_PVS_CNTL_1_PROGRAM_START_SHIFT) | - (inst_count << R300_PVS_CNTL_1_POS_END_SHIFT) | - (inst_count << R300_PVS_CNTL_1_PROGRAM_END_SHIFT); + (0 << R300_PVS_CNTL_1_PROGRAM_START_SHIFT) | + (inst_count << R300_PVS_CNTL_1_POS_END_SHIFT) | + (inst_count << R300_PVS_CNTL_1_PROGRAM_END_SHIFT); rmesa->hw.pvs.cmd[R300_PVS_CNTL_2] = - (0 << R300_PVS_CNTL_2_PARAM_OFFSET_SHIFT) | - (param_count << R300_PVS_CNTL_2_PARAM_COUNT_SHIFT); + (0 << R300_PVS_CNTL_2_PARAM_OFFSET_SHIFT) | + (param_count << R300_PVS_CNTL_2_PARAM_COUNT_SHIFT); rmesa->hw.pvs.cmd[R300_PVS_CNTL_3] = - (inst_count << R300_PVS_CNTL_3_PROGRAM_UNKNOWN_SHIFT) | - (inst_count << R300_PVS_CNTL_3_PROGRAM_UNKNOWN2_SHIFT); + (inst_count << R300_PVS_CNTL_3_PROGRAM_UNKNOWN_SHIFT) | + (inst_count << R300_PVS_CNTL_3_PROGRAM_UNKNOWN2_SHIFT); } static void r300SetupRealVertexProgram(r300ContextPtr rmesa) diff --git a/src/mesa/drivers/dri/r300/r300_vertprog.c b/src/mesa/drivers/dri/r300/r300_vertprog.c index 0fb6110494..7d4e8c9511 100644 --- a/src/mesa/drivers/dri/r300/r300_vertprog.c +++ b/src/mesa/drivers/dri/r300/r300_vertprog.c @@ -31,6 +31,8 @@ USE OR OTHER DEALINGS IN THE SOFTWARE. * \author Aapo Tahkola <aet@rasterburn.org> * * \author Oliver McFadden <z3ro.geek@gmail.com> + * + * For a description of the vertex program instruction set see r300_reg.h. */ #include "glheader.h" @@ -366,20 +368,19 @@ static GLboolean valid_dst(struct r300_vertex_program *vp, * component. */ -static void t_opcode_abs(struct r300_vertex_program *vp, - struct prog_instruction *vpi, - struct r300_vertprog_instruction *o_inst, - struct prog_src_register src[3]) +static GLuint *t_opcode_abs(struct r300_vertex_program *vp, + struct prog_instruction *vpi, GLuint * inst, + struct prog_src_register src[3]) { //MAX RESULT 1.X Y Z W PARAM 0{} {X Y Z W} PARAM 0{X Y Z W } {X Y Z W} neg Xneg Yneg Zneg W - o_inst->opcode = + inst[0] = MAKE_VSF_OP(R300_VPI_OUT_OP_MAX, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = t_src(vp, &src[0]); - o_inst->src[1] = + inst[1] = t_src(vp, &src[0]); + inst[2] = MAKE_VSF_SOURCE(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 0)), t_swizzle(GET_SWZ(src[0].Swizzle, 1)), @@ -389,13 +390,14 @@ static void t_opcode_abs(struct r300_vertex_program *vp, (!src[0]. NegateBase) ? VSF_FLAG_ALL : VSF_FLAG_NONE) | (src[0].RelAddr << 4); - o_inst->src[2] = 0; + inst[3] = 0; + + return inst; } -static void t_opcode_add(struct r300_vertex_program *vp, - struct prog_instruction *vpi, - struct r300_vertprog_instruction *o_inst, - struct prog_src_register src[3]) +static GLuint *t_opcode_add(struct r300_vertex_program *vp, + struct prog_instruction *vpi, GLuint * inst, + struct prog_src_register src[3]) { unsigned long hw_op; @@ -405,53 +407,55 @@ static void t_opcode_add(struct r300_vertex_program *vp, PROGRAM_TEMPORARY) ? R300_VPI_OUT_OP_MAD_2 : R300_VPI_OUT_OP_MAD; - o_inst->opcode = + inst[0] = MAKE_VSF_OP(hw_op, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = ONE_SRC_0; - o_inst->src[1] = t_src(vp, &src[0]); - o_inst->src[2] = t_src(vp, &src[1]); + inst[1] = ONE_SRC_0; + inst[2] = t_src(vp, &src[0]); + inst[3] = t_src(vp, &src[1]); #else - o_inst->opcode = + inst[0] = MAKE_VSF_OP(R300_VPI_OUT_OP_ADD, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = t_src(vp, &src[0]); - o_inst->src[1] = t_src(vp, &src[1]); - o_inst->src[2] = ZERO_SRC_1; + inst[1] = t_src(vp, &src[0]); + inst[2] = t_src(vp, &src[1]); + inst[3] = ZERO_SRC_1; #endif + + return inst; } -static void t_opcode_arl(struct r300_vertex_program *vp, - struct prog_instruction *vpi, - struct r300_vertprog_instruction *o_inst, - struct prog_src_register src[3]) +static GLuint *t_opcode_arl(struct r300_vertex_program *vp, + struct prog_instruction *vpi, GLuint * inst, + struct prog_src_register src[3]) { - o_inst->opcode = + inst[0] = MAKE_VSF_OP(R300_VPI_OUT_OP_ARL, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = t_src(vp, &src[0]); - o_inst->src[1] = ZERO_SRC_0; - o_inst->src[2] = ZERO_SRC_0; + inst[1] = t_src(vp, &src[0]); + inst[2] = ZERO_SRC_0; + inst[3] = ZERO_SRC_0; + + return inst; } -static void t_opcode_dp3(struct r300_vertex_program *vp, - struct prog_instruction *vpi, - struct r300_vertprog_instruction *o_inst, - struct prog_src_register src[3]) +static GLuint *t_opcode_dp3(struct r300_vertex_program *vp, + struct prog_instruction *vpi, GLuint * inst, + struct prog_src_register src[3]) { //DOT RESULT 1.X Y Z W PARAM 0{} {X Y Z ZERO} PARAM 0{} {X Y Z ZERO} - o_inst->opcode = + inst[0] = MAKE_VSF_OP(R300_VPI_OUT_OP_DOT, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = + inst[1] = MAKE_VSF_SOURCE(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 0)), t_swizzle(GET_SWZ(src[0].Swizzle, 1)), @@ -461,7 +465,7 @@ static void t_opcode_dp3(struct r300_vertex_program *vp, NegateBase ? VSF_FLAG_XYZ : VSF_FLAG_NONE) | (src[0].RelAddr << 4); - o_inst->src[1] = + inst[2] = MAKE_VSF_SOURCE(t_src_index(vp, &src[1]), t_swizzle(GET_SWZ(src[1].Swizzle, 0)), t_swizzle(GET_SWZ(src[1].Swizzle, 1)), @@ -471,36 +475,38 @@ static void t_opcode_dp3(struct r300_vertex_program *vp, NegateBase ? VSF_FLAG_XYZ : VSF_FLAG_NONE) | (src[1].RelAddr << 4); - o_inst->src[2] = ZERO_SRC_1; + inst[3] = ZERO_SRC_1; + + return inst; } -static void t_opcode_dp4(struct r300_vertex_program *vp, - struct prog_instruction *vpi, - struct r300_vertprog_instruction *o_inst, - struct prog_src_register src[3]) +static GLuint *t_opcode_dp4(struct r300_vertex_program *vp, + struct prog_instruction *vpi, GLuint * inst, + struct prog_src_register src[3]) { - o_inst->opcode = + inst[0] = MAKE_VSF_OP(R300_VPI_OUT_OP_DOT, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = t_src(vp, &src[0]); - o_inst->src[1] = t_src(vp, &src[1]); - o_inst->src[2] = ZERO_SRC_1; + inst[1] = t_src(vp, &src[0]); + inst[2] = t_src(vp, &src[1]); + inst[3] = ZERO_SRC_1; + + return inst; } -static void t_opcode_dph(struct r300_vertex_program *vp, - struct prog_instruction *vpi, - struct r300_vertprog_instruction *o_inst, - struct prog_src_register src[3]) +static GLuint *t_opcode_dph(struct r300_vertex_program *vp, + struct prog_instruction *vpi, GLuint * inst, + struct prog_src_register src[3]) { //DOT RESULT 1.X Y Z W PARAM 0{} {X Y Z ONE} PARAM 0{} {X Y Z W} - o_inst->opcode = + inst[0] = MAKE_VSF_OP(R300_VPI_OUT_OP_DOT, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = + inst[1] = MAKE_VSF_SOURCE(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 0)), t_swizzle(GET_SWZ(src[0].Swizzle, 1)), @@ -509,81 +515,84 @@ static void t_opcode_dph(struct r300_vertex_program *vp, src[0]. NegateBase ? VSF_FLAG_XYZ : VSF_FLAG_NONE) | (src[0].RelAddr << 4); - o_inst->src[1] = t_src(vp, &src[1]); - o_inst->src[2] = ZERO_SRC_1; + inst[2] = t_src(vp, &src[1]); + inst[3] = ZERO_SRC_1; + + return inst; } -static void t_opcode_dst(struct r300_vertex_program *vp, - struct prog_instruction *vpi, - struct r300_vertprog_instruction *o_inst, - struct prog_src_register src[3]) +static GLuint *t_opcode_dst(struct r300_vertex_program *vp, + struct prog_instruction *vpi, GLuint * inst, + struct prog_src_register src[3]) { - o_inst->opcode = + inst[0] = MAKE_VSF_OP(R300_VPI_OUT_OP_DST, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = t_src(vp, &src[0]); - o_inst->src[1] = t_src(vp, &src[1]); - o_inst->src[2] = ZERO_SRC_1; + inst[1] = t_src(vp, &src[0]); + inst[2] = t_src(vp, &src[1]); + inst[3] = ZERO_SRC_1; + + return inst; } -static void t_opcode_ex2(struct r300_vertex_program *vp, - struct prog_instruction *vpi, - struct r300_vertprog_instruction *o_inst, - struct prog_src_register src[3]) +static GLuint *t_opcode_ex2(struct r300_vertex_program *vp, + struct prog_instruction *vpi, GLuint * inst, + struct prog_src_register src[3]) { - o_inst->opcode = + inst[0] = MAKE_VSF_OP(R300_VPI_OUT_OP_EX2, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = t_src_scalar(vp, &src[0]); - o_inst->src[1] = ZERO_SRC_0; - o_inst->src[2] = ZERO_SRC_0; + inst[1] = t_src_scalar(vp, &src[0]); + inst[2] = ZERO_SRC_0; + inst[3] = ZERO_SRC_0; + + return inst; } -static void t_opcode_exp(struct r300_vertex_program *vp, - struct prog_instruction *vpi, - struct r300_vertprog_instruction *o_inst, - struct prog_src_register src[3]) +static GLuint *t_opcode_exp(struct r300_vertex_program *vp, + struct prog_instruction *vpi, GLuint * inst, + struct prog_src_register src[3]) { - o_inst->opcode = + inst[0] = MAKE_VSF_OP(R300_VPI_OUT_OP_EXP, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = t_src_scalar(vp, &src[0]); - o_inst->src[1] = ZERO_SRC_0; - o_inst->src[2] = ZERO_SRC_0; + inst[1] = t_src_scalar(vp, &src[0]); + inst[2] = ZERO_SRC_0; + inst[3] = ZERO_SRC_0; + + return inst; } -static struct r300_vertprog_instruction *t_opcode_flr(struct r300_vertex_program *vp, - struct prog_instruction *vpi, - struct r300_vertprog_instruction *o_inst, - struct prog_src_register src[3], - int *u_temp_i) +static GLuint *t_opcode_flr(struct r300_vertex_program *vp, + struct prog_instruction *vpi, GLuint * inst, + struct prog_src_register src[3], int *u_temp_i) { /* FRC TMP 0.X Y Z W PARAM 0{} {X Y Z W} ADD RESULT 1.X Y Z W PARAM 0{} {X Y Z W} TMP 0{X Y Z W } {X Y Z W} neg Xneg Yneg Zneg W */ - o_inst->opcode = + inst[0] = MAKE_VSF_OP(R300_VPI_OUT_OP_FRC, *u_temp_i, t_dst_mask(vpi->DstReg.WriteMask), VSF_OUT_CLASS_TMP); - o_inst->src[0] = t_src(vp, &src[0]); - o_inst->src[1] = ZERO_SRC_0; - o_inst->src[2] = ZERO_SRC_0; - o_inst++; + inst[1] = t_src(vp, &src[0]); + inst[2] = ZERO_SRC_0; + inst[3] = ZERO_SRC_0; + inst += 4; - o_inst->opcode = + inst[0] = MAKE_VSF_OP(R300_VPI_OUT_OP_ADD, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = t_src(vp, &src[0]); - o_inst->src[1] = + inst[1] = t_src(vp, &src[0]); + inst[2] = MAKE_VSF_SOURCE(*u_temp_i, VSF_IN_COMPONENT_X, VSF_IN_COMPONENT_Y, VSF_IN_COMPONENT_Z, VSF_IN_COMPONENT_W, VSF_IN_CLASS_TMP, @@ -592,40 +601,40 @@ static struct r300_vertprog_instruction *t_opcode_flr(struct r300_vertex_program NegateBase) ? VSF_FLAG_ALL : VSF_FLAG_NONE /*VSF_FLAG_ALL */ ); - o_inst->src[2] = ZERO_SRC_0; + inst[3] = ZERO_SRC_0; (*u_temp_i)--; - return o_inst; + return inst; } -static void t_opcode_frc(struct r300_vertex_program *vp, - struct prog_instruction *vpi, - struct r300_vertprog_instruction *o_inst, - struct prog_src_register src[3]) +static GLuint *t_opcode_frc(struct r300_vertex_program *vp, + struct prog_instruction *vpi, GLuint * inst, + struct prog_src_register src[3]) { - o_inst->opcode = + inst[0] = MAKE_VSF_OP(R300_VPI_OUT_OP_FRC, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = t_src(vp, &src[0]); - o_inst->src[1] = ZERO_SRC_0; - o_inst->src[2] = ZERO_SRC_0; + inst[1] = t_src(vp, &src[0]); + inst[2] = ZERO_SRC_0; + inst[3] = ZERO_SRC_0; + + return inst; } -static void t_opcode_lg2(struct r300_vertex_program *vp, - struct prog_instruction *vpi, - struct r300_vertprog_instruction *o_inst, - struct prog_src_register src[3]) +static GLuint *t_opcode_lg2(struct r300_vertex_program *vp, + struct prog_instruction *vpi, GLuint * inst, + struct prog_src_register src[3]) { // LG2 RESULT 1.X Y Z W PARAM 0{} {X X X X} - o_inst->opcode = + inst[0] = MAKE_VSF_OP(R300_VPI_OUT_OP_LG2, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = + inst[1] = MAKE_VSF_SOURCE(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 0)), t_swizzle(GET_SWZ(src[0].Swizzle, 0)), @@ -635,70 +644,70 @@ static void t_opcode_lg2(struct r300_vertex_program *vp, src[0]. NegateBase ? VSF_FLAG_ALL : VSF_FLAG_NONE) | (src[0].RelAddr << 4); - o_inst->src[1] = ZERO_SRC_0; - o_inst->src[2] = ZERO_SRC_0; + inst[2] = ZERO_SRC_0; + inst[3] = ZERO_SRC_0; + + return inst; } -static void t_opcode_lit(struct r300_vertex_program *vp, - struct prog_instruction *vpi, - struct r300_vertprog_instruction *o_inst, - struct prog_src_register src[3]) +static GLuint *t_opcode_lit(struct r300_vertex_program *vp, + struct prog_instruction *vpi, GLuint * inst, + struct prog_src_register src[3]) { //LIT TMP 1.Y Z TMP 1{} {X W Z Y} TMP 1{} {Y W Z X} TMP 1{} {Y X Z W} - o_inst->opcode = + inst[0] = MAKE_VSF_OP(R300_VPI_OUT_OP_LIT, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); /* NOTE: Users swizzling might not work. */ - o_inst->src[0] = MAKE_VSF_SOURCE(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 0)), // x - t_swizzle(GET_SWZ(src[0].Swizzle, 3)), // w - VSF_IN_COMPONENT_ZERO, // z - t_swizzle(GET_SWZ(src[0].Swizzle, 1)), // y - t_src_class(src[0].File), - src[0]. - NegateBase ? VSF_FLAG_ALL : - VSF_FLAG_NONE) | (src[0]. - RelAddr << 4); - o_inst->src[1] = MAKE_VSF_SOURCE(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 1)), // y - t_swizzle(GET_SWZ(src[0].Swizzle, 3)), // w - VSF_IN_COMPONENT_ZERO, // z - t_swizzle(GET_SWZ(src[0].Swizzle, 0)), // x - t_src_class(src[0].File), - src[0]. - NegateBase ? VSF_FLAG_ALL : - VSF_FLAG_NONE) | (src[0]. - RelAddr << 4); - o_inst->src[2] = MAKE_VSF_SOURCE(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 1)), // y - t_swizzle(GET_SWZ(src[0].Swizzle, 0)), // x - VSF_IN_COMPONENT_ZERO, // z - t_swizzle(GET_SWZ(src[0].Swizzle, 3)), // w - t_src_class(src[0].File), - src[0]. - NegateBase ? VSF_FLAG_ALL : - VSF_FLAG_NONE) | (src[0]. - RelAddr << 4); + inst[1] = MAKE_VSF_SOURCE(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 0)), // x + t_swizzle(GET_SWZ(src[0].Swizzle, 3)), // w + VSF_IN_COMPONENT_ZERO, // z + t_swizzle(GET_SWZ(src[0].Swizzle, 1)), // y + t_src_class(src[0].File), + src[0]. + NegateBase ? VSF_FLAG_ALL : + VSF_FLAG_NONE) | (src[0].RelAddr << 4); + inst[2] = MAKE_VSF_SOURCE(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 1)), // y + t_swizzle(GET_SWZ(src[0].Swizzle, 3)), // w + VSF_IN_COMPONENT_ZERO, // z + t_swizzle(GET_SWZ(src[0].Swizzle, 0)), // x + t_src_class(src[0].File), + src[0]. + NegateBase ? VSF_FLAG_ALL : + VSF_FLAG_NONE) | (src[0].RelAddr << 4); + inst[3] = MAKE_VSF_SOURCE(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 1)), // y + t_swizzle(GET_SWZ(src[0].Swizzle, 0)), // x + VSF_IN_COMPONENT_ZERO, // z + t_swizzle(GET_SWZ(src[0].Swizzle, 3)), // w + t_src_class(src[0].File), + src[0]. + NegateBase ? VSF_FLAG_ALL : + VSF_FLAG_NONE) | (src[0].RelAddr << 4); + + return inst; } -static void t_opcode_log(struct r300_vertex_program *vp, - struct prog_instruction *vpi, - struct r300_vertprog_instruction *o_inst, - struct prog_src_register src[3]) +static GLuint *t_opcode_log(struct r300_vertex_program *vp, + struct prog_instruction *vpi, GLuint * inst, + struct prog_src_register src[3]) { - o_inst->opcode = + inst[0] = MAKE_VSF_OP(R300_VPI_OUT_OP_LOG, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = t_src_scalar(vp, &src[0]); - o_inst->src[1] = ZERO_SRC_0; - o_inst->src[2] = ZERO_SRC_0; + inst[1] = t_src_scalar(vp, &src[0]); + inst[2] = ZERO_SRC_0; + inst[3] = ZERO_SRC_0; + + return inst; } -static void t_opcode_mad(struct r300_vertex_program *vp, - struct prog_instruction *vpi, - struct r300_vertprog_instruction *o_inst, - struct prog_src_register src[3]) +static GLuint *t_opcode_mad(struct r300_vertex_program *vp, + struct prog_instruction *vpi, GLuint * inst, + struct prog_src_register src[3]) { unsigned long hw_op; @@ -708,80 +717,84 @@ static void t_opcode_mad(struct r300_vertex_program *vp, PROGRAM_TEMPORARY) ? R300_VPI_OUT_OP_MAD_2 : R300_VPI_OUT_OP_MAD; - o_inst->opcode = + inst[0] = MAKE_VSF_OP(hw_op, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = t_src(vp, &src[0]); - o_inst->src[1] = t_src(vp, &src[1]); - o_inst->src[2] = t_src(vp, &src[2]); + inst[1] = t_src(vp, &src[0]); + inst[2] = t_src(vp, &src[1]); + inst[3] = t_src(vp, &src[2]); + + return inst; } -static void t_opcode_max(struct r300_vertex_program *vp, - struct prog_instruction *vpi, - struct r300_vertprog_instruction *o_inst, - struct prog_src_register src[3]) +static GLuint *t_opcode_max(struct r300_vertex_program *vp, + struct prog_instruction *vpi, GLuint * inst, + struct prog_src_register src[3]) { - o_inst->opcode = + inst[0] = MAKE_VSF_OP(R300_VPI_OUT_OP_MAX, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = t_src(vp, &src[0]); - o_inst->src[1] = t_src(vp, &src[1]); - o_inst->src[2] = ZERO_SRC_1; + inst[1] = t_src(vp, &src[0]); + inst[2] = t_src(vp, &src[1]); + inst[3] = ZERO_SRC_1; + + return inst; } -static void t_opcode_min(struct r300_vertex_program *vp, - struct prog_instruction *vpi, - struct r300_vertprog_instruction *o_inst, - struct prog_src_register src[3]) +static GLuint *t_opcode_min(struct r300_vertex_program *vp, + struct prog_instruction *vpi, GLuint * inst, + struct prog_src_register src[3]) { - o_inst->opcode = + inst[0] = MAKE_VSF_OP(R300_VPI_OUT_OP_MIN, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = t_src(vp, &src[0]); - o_inst->src[1] = t_src(vp, &src[1]); - o_inst->src[2] = ZERO_SRC_1; + inst[1] = t_src(vp, &src[0]); + inst[2] = t_src(vp, &src[1]); + inst[3] = ZERO_SRC_1; + + return inst; } -static void t_opcode_mov(struct r300_vertex_program *vp, - struct prog_instruction *vpi, - struct r300_vertprog_instruction *o_inst, - struct prog_src_register src[3]) +static GLuint *t_opcode_mov(struct r300_vertex_program *vp, + struct prog_instruction *vpi, GLuint * inst, + struct prog_src_register src[3]) { //ADD RESULT 1.X Y Z W PARAM 0{} {X Y Z W} PARAM 0{} {ZERO ZERO ZERO ZERO} #if 1 - o_inst->opcode = + inst[0] = MAKE_VSF_OP(R300_VPI_OUT_OP_ADD, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = t_src(vp, &src[0]); - o_inst->src[1] = ZERO_SRC_0; - o_inst->src[2] = ZERO_SRC_0; + inst[1] = t_src(vp, &src[0]); + inst[2] = ZERO_SRC_0; + inst[3] = ZERO_SRC_0; #else hw_op = (src[0].File == PROGRAM_TEMPORARY) ? R300_VPI_OUT_OP_MAD_2 : R300_VPI_OUT_OP_MAD; - o_inst->opcode = + inst[0] = MAKE_VSF_OP(hw_op, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = t_src(vp, &src[0]); - o_inst->src[1] = ONE_SRC_0; - o_inst->src[2] = ZERO_SRC_0; + inst[1] = t_src(vp, &src[0]); + inst[2] = ONE_SRC_0; + inst[3] = ZERO_SRC_0; #endif + + return inst; } -static void t_opcode_mul(struct r300_vertex_program *vp, - struct prog_instruction *vpi, - struct r300_vertprog_instruction *o_inst, - struct prog_src_register src[3]) +static GLuint *t_opcode_mul(struct r300_vertex_program *vp, + struct prog_instruction *vpi, GLuint * inst, + struct prog_src_register src[3]) { unsigned long hw_op; @@ -792,94 +805,100 @@ static void t_opcode_mul(struct r300_vertex_program *vp, PROGRAM_TEMPORARY) ? R300_VPI_OUT_OP_MAD_2 : R300_VPI_OUT_OP_MAD; - o_inst->opcode = + inst[0] = MAKE_VSF_OP(hw_op, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = t_src(vp, &src[0]); - o_inst->src[1] = t_src(vp, &src[1]); + inst[1] = t_src(vp, &src[0]); + inst[2] = t_src(vp, &src[1]); + + inst[3] = ZERO_SRC_1; - o_inst->src[2] = ZERO_SRC_1; + return inst; } -static void t_opcode_pow(struct r300_vertex_program *vp, - struct prog_instruction *vpi, - struct r300_vertprog_instruction *o_inst, - struct prog_src_register src[3]) +static GLuint *t_opcode_pow(struct r300_vertex_program *vp, + struct prog_instruction *vpi, GLuint * inst, + struct prog_src_register src[3]) { - o_inst->opcode = + inst[0] = MAKE_VSF_OP(R300_VPI_OUT_OP_POW, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = t_src_scalar(vp, &src[0]); - o_inst->src[1] = ZERO_SRC_0; - o_inst->src[2] = t_src_scalar(vp, &src[1]); + inst[1] = t_src_scalar(vp, &src[0]); + inst[2] = ZERO_SRC_0; + inst[3] = t_src_scalar(vp, &src[1]); + + return inst; } -static void t_opcode_rcp(struct r300_vertex_program *vp, - struct prog_instruction *vpi, - struct r300_vertprog_instruction *o_inst, - struct prog_src_register src[3]) +static GLuint *t_opcode_rcp(struct r300_vertex_program *vp, + struct prog_instruction *vpi, GLuint * inst, + struct prog_src_register src[3]) { - o_inst->opcode = + inst[0] = MAKE_VSF_OP(R300_VPI_OUT_OP_RCP, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = t_src_scalar(vp, &src[0]); - o_inst->src[1] = ZERO_SRC_0; - o_inst->src[2] = ZERO_SRC_0; + inst[1] = t_src_scalar(vp, &src[0]); + inst[2] = ZERO_SRC_0; + inst[3] = ZERO_SRC_0; + + return inst; } -static void t_opcode_rsq(struct r300_vertex_program *vp, - struct prog_instruction *vpi, - struct r300_vertprog_instruction *o_inst, - struct prog_src_register src[3]) +static GLuint *t_opcode_rsq(struct r300_vertex_program *vp, + struct prog_instruction *vpi, GLuint * inst, + struct prog_src_register src[3]) { - o_inst->opcode = + inst[0] = MAKE_VSF_OP(R300_VPI_OUT_OP_RSQ, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = t_src_scalar(vp, &src[0]); - o_inst->src[1] = ZERO_SRC_0; - o_inst->src[2] = ZERO_SRC_0; + inst[1] = t_src_scalar(vp, &src[0]); + inst[2] = ZERO_SRC_0; + inst[3] = ZERO_SRC_0; + + return inst; } -static void t_opcode_sge(struct r300_vertex_program *vp, - struct prog_instruction *vpi, - struct r300_vertprog_instruction *o_inst, - struct prog_src_register src[3]) +static GLuint *t_opcode_sge(struct r300_vertex_program *vp, + struct prog_instruction *vpi, GLuint * inst, + struct prog_src_register src[3]) { - o_inst->opcode = + inst[0] = MAKE_VSF_OP(R300_VPI_OUT_OP_SGE, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = t_src(vp, &src[0]); - o_inst->src[1] = t_src(vp, &src[1]); - o_inst->src[2] = ZERO_SRC_1; + inst[1] = t_src(vp, &src[0]); + inst[2] = t_src(vp, &src[1]); + inst[3] = ZERO_SRC_1; + + return inst; } -static void t_opcode_slt(struct r300_vertex_program *vp, - struct prog_instruction *vpi, - struct r300_vertprog_instruction *o_inst, - struct prog_src_register src[3]) +static GLuint *t_opcode_slt(struct r300_vertex_program *vp, + struct prog_instruction *vpi, GLuint * inst, + struct prog_src_register src[3]) { - o_inst->opcode = + inst[0] = MAKE_VSF_OP(R300_VPI_OUT_OP_SLT, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = t_src(vp, &src[0]); - o_inst->src[1] = t_src(vp, &src[1]); - o_inst->src[2] = ZERO_SRC_1; + inst[1] = t_src(vp, &src[0]); + inst[2] = t_src(vp, &src[1]); + inst[3] = ZERO_SRC_1; + + return inst; } -static void t_opcode_sub(struct r300_vertex_program *vp, - struct prog_instruction *vpi, - struct r300_vertprog_instruction *o_inst, - struct prog_src_register src[3]) +static GLuint *t_opcode_sub(struct r300_vertex_program *vp, + struct prog_instruction *vpi, GLuint * inst, + struct prog_src_register src[3]) { unsigned long hw_op; @@ -891,13 +910,13 @@ static void t_opcode_sub(struct r300_vertex_program *vp, PROGRAM_TEMPORARY) ? R300_VPI_OUT_OP_MAD_2 : R300_VPI_OUT_OP_MAD; - o_inst->opcode = + inst[0] = MAKE_VSF_OP(hw_op, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = t_src(vp, &src[0]); - o_inst->src[1] = ONE_SRC_0; - o_inst->src[2] = + inst[1] = t_src(vp, &src[0]); + inst[2] = ONE_SRC_0; + inst[3] = MAKE_VSF_SOURCE(t_src_index(vp, &src[1]), t_swizzle(GET_SWZ(src[1].Swizzle, 0)), t_swizzle(GET_SWZ(src[1].Swizzle, 1)), @@ -908,13 +927,13 @@ static void t_opcode_sub(struct r300_vertex_program *vp, NegateBase) ? VSF_FLAG_ALL : VSF_FLAG_NONE) | (src[1].RelAddr << 4); #else - o_inst->opcode = + inst[0] = MAKE_VSF_OP(R300_VPI_OUT_OP_ADD, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = t_src(vp, &src[0]); - o_inst->src[1] = + inst[1] = t_src(vp, &src[0]); + inst[2] = MAKE_VSF_SOURCE(t_src_index(vp, &src[1]), t_swizzle(GET_SWZ(src[1].Swizzle, 0)), t_swizzle(GET_SWZ(src[1].Swizzle, 1)), @@ -924,113 +943,110 @@ static void t_opcode_sub(struct r300_vertex_program *vp, (!src[1]. NegateBase) ? VSF_FLAG_ALL : VSF_FLAG_NONE) | (src[1].RelAddr << 4); - o_inst->src[2] = 0; + inst[3] = 0; #endif + + return inst; } -static void t_opcode_swz(struct r300_vertex_program *vp, - struct prog_instruction *vpi, - struct r300_vertprog_instruction *o_inst, - struct prog_src_register src[3]) +static GLuint *t_opcode_swz(struct r300_vertex_program *vp, + struct prog_instruction *vpi, GLuint * inst, + struct prog_src_register src[3]) { //ADD RESULT 1.X Y Z W PARAM 0{} {X Y Z W} PARAM 0{} {ZERO ZERO ZERO ZERO} #if 1 - o_inst->opcode = + inst[0] = MAKE_VSF_OP(R300_VPI_OUT_OP_ADD, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = t_src(vp, &src[0]); - o_inst->src[1] = ZERO_SRC_0; - o_inst->src[2] = ZERO_SRC_0; + inst[1] = t_src(vp, &src[0]); + inst[2] = ZERO_SRC_0; + inst[3] = ZERO_SRC_0; #else hw_op = (src[0].File == PROGRAM_TEMPORARY) ? R300_VPI_OUT_OP_MAD_2 : R300_VPI_OUT_OP_MAD; - o_inst->opcode = + inst[0] = MAKE_VSF_OP(hw_op, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = t_src(vp, &src[0]); - o_inst->src[1] = ONE_SRC_0; - o_inst->src[2] = ZERO_SRC_0; + inst[1] = t_src(vp, &src[0]); + inst[2] = ONE_SRC_0; + inst[3] = ZERO_SRC_0; #endif + + return inst; } -static struct r300_vertprog_instruction *t_opcode_xpd(struct r300_vertex_program *vp, - struct prog_instruction *vpi, - struct r300_vertprog_instruction *o_inst, - struct prog_src_register src[3], - int *u_temp_i) +static GLuint *t_opcode_xpd(struct r300_vertex_program *vp, + struct prog_instruction *vpi, GLuint * inst, + struct prog_src_register src[3], int *u_temp_i) { /* mul r0, r1.yzxw, r2.zxyw mad r0, -r2.yzxw, r1.zxyw, r0 NOTE: might need MAD_2 */ - o_inst->opcode = + inst[0] = MAKE_VSF_OP(R300_VPI_OUT_OP_MAD, *u_temp_i, t_dst_mask(vpi->DstReg.WriteMask), VSF_OUT_CLASS_TMP); - o_inst->src[0] = MAKE_VSF_SOURCE(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 1)), // y - t_swizzle(GET_SWZ(src[0].Swizzle, 2)), // z - t_swizzle(GET_SWZ(src[0].Swizzle, 0)), // x - t_swizzle(GET_SWZ(src[0].Swizzle, 3)), // w - t_src_class(src[0].File), - src[0]. - NegateBase ? VSF_FLAG_ALL : - VSF_FLAG_NONE) | (src[0]. - RelAddr << 4); + inst[1] = MAKE_VSF_SOURCE(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 1)), // y + t_swizzle(GET_SWZ(src[0].Swizzle, 2)), // z + t_swizzle(GET_SWZ(src[0].Swizzle, 0)), // x + t_swizzle(GET_SWZ(src[0].Swizzle, 3)), // w + t_src_class(src[0].File), + src[0]. + NegateBase ? VSF_FLAG_ALL : + VSF_FLAG_NONE) | (src[0].RelAddr << 4); - o_inst->src[1] = MAKE_VSF_SOURCE(t_src_index(vp, &src[1]), t_swizzle(GET_SWZ(src[1].Swizzle, 2)), // z - t_swizzle(GET_SWZ(src[1].Swizzle, 0)), // x - t_swizzle(GET_SWZ(src[1].Swizzle, 1)), // y - t_swizzle(GET_SWZ(src[1].Swizzle, 3)), // w - t_src_class(src[1].File), - src[1]. - NegateBase ? VSF_FLAG_ALL : - VSF_FLAG_NONE) | (src[1]. - RelAddr << 4); + inst[2] = MAKE_VSF_SOURCE(t_src_index(vp, &src[1]), t_swizzle(GET_SWZ(src[1].Swizzle, 2)), // z + t_swizzle(GET_SWZ(src[1].Swizzle, 0)), // x + t_swizzle(GET_SWZ(src[1].Swizzle, 1)), // y + t_swizzle(GET_SWZ(src[1].Swizzle, 3)), // w + t_src_class(src[1].File), + src[1]. + NegateBase ? VSF_FLAG_ALL : + VSF_FLAG_NONE) | (src[1].RelAddr << 4); - o_inst->src[2] = ZERO_SRC_1; - o_inst++; + inst[3] = ZERO_SRC_1; + inst += 4; (*u_temp_i)--; - o_inst->opcode = + inst[0] = MAKE_VSF_OP(R300_VPI_OUT_OP_MAD, t_dst_index(vp, &vpi->DstReg), t_dst_mask(vpi->DstReg.WriteMask), t_dst_class(vpi->DstReg.File)); - o_inst->src[0] = MAKE_VSF_SOURCE(t_src_index(vp, &src[1]), t_swizzle(GET_SWZ(src[1].Swizzle, 1)), // y - t_swizzle(GET_SWZ(src[1].Swizzle, 2)), // z - t_swizzle(GET_SWZ(src[1].Swizzle, 0)), // x - t_swizzle(GET_SWZ(src[1].Swizzle, 3)), // w - t_src_class(src[1].File), - (!src[1]. - NegateBase) ? VSF_FLAG_ALL : - VSF_FLAG_NONE) | (src[1]. - RelAddr << 4); + inst[1] = MAKE_VSF_SOURCE(t_src_index(vp, &src[1]), t_swizzle(GET_SWZ(src[1].Swizzle, 1)), // y + t_swizzle(GET_SWZ(src[1].Swizzle, 2)), // z + t_swizzle(GET_SWZ(src[1].Swizzle, 0)), // x + t_swizzle(GET_SWZ(src[1].Swizzle, 3)), // w + t_src_class(src[1].File), + (!src[1]. + NegateBase) ? VSF_FLAG_ALL : + VSF_FLAG_NONE) | (src[1].RelAddr << 4); - o_inst->src[1] = MAKE_VSF_SOURCE(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 2)), // z - t_swizzle(GET_SWZ(src[0].Swizzle, 0)), // x - t_swizzle(GET_SWZ(src[0].Swizzle, 1)), // y - t_swizzle(GET_SWZ(src[0].Swizzle, 3)), // w - t_src_class(src[0].File), - src[0]. - NegateBase ? VSF_FLAG_ALL : - VSF_FLAG_NONE) | (src[0]. - RelAddr << 4); + inst[2] = MAKE_VSF_SOURCE(t_src_index(vp, &src[0]), t_swizzle(GET_SWZ(src[0].Swizzle, 2)), // z + t_swizzle(GET_SWZ(src[0].Swizzle, 0)), // x + t_swizzle(GET_SWZ(src[0].Swizzle, 1)), // y + t_swizzle(GET_SWZ(src[0].Swizzle, 3)), // w + t_src_class(src[0].File), + src[0]. + NegateBase ? VSF_FLAG_ALL : + VSF_FLAG_NONE) | (src[0].RelAddr << 4); - o_inst->src[2] = + inst[3] = MAKE_VSF_SOURCE(*u_temp_i + 1, VSF_IN_COMPONENT_X, VSF_IN_COMPONENT_Y, VSF_IN_COMPONENT_Z, VSF_IN_COMPONENT_W, VSF_IN_CLASS_TMP, VSF_FLAG_NONE); - return o_inst; + return inst; } static void t_inputs_outputs(struct r300_vertex_program *vp) @@ -1067,7 +1083,7 @@ static void t_inputs_outputs(struct r300_vertex_program *vp) if (vp->key.OutputsWritten & (1 << VERT_RESULT_BFC0)) { vp->outputs[VERT_RESULT_BFC0] = vp->outputs[VERT_RESULT_COL0] + 2; - cur_reg = vp->outputs[VERT_RESULT_BFC0] + 1; + cur_reg = vp->outputs[VERT_RESULT_BFC0] + 2; } if (vp->key.OutputsWritten & (1 << VERT_RESULT_BFC1)) { @@ -1092,7 +1108,7 @@ static void r300TranslateVertexShader(struct r300_vertex_program *vp, struct prog_instruction *vpi) { int i; - struct r300_vertprog_instruction *o_inst; + GLuint *inst; unsigned long num_operands; /* Initial value should be last tmp reg that hw supports. Strangely enough r300 doesnt mind even though these would be out of range. @@ -1108,8 +1124,8 @@ static void r300TranslateVertexShader(struct r300_vertex_program *vp, t_inputs_outputs(vp); - for (o_inst = vp->program.body.i; vpi->Opcode != OPCODE_END; - vpi++, o_inst++) { + for (inst = vp->program.body.i; vpi->Opcode != OPCODE_END; + vpi++, inst += 4) { FREE_TEMPS(); @@ -1129,12 +1145,12 @@ static void r300TranslateVertexShader(struct r300_vertex_program *vp, if (num_operands == 3) { /* TODO: scalars */ if (CMP_SRCS(src[1], src[2]) || CMP_SRCS(src[0], src[2])) { - o_inst->opcode = + inst[0] = MAKE_VSF_OP(R300_VPI_OUT_OP_ADD, u_temp_i, VSF_FLAG_ALL, VSF_OUT_CLASS_TMP); - o_inst->src[0] = + inst[1] = MAKE_VSF_SOURCE(t_src_index (vp, &src[2]), SWIZZLE_X, SWIZZLE_Y, @@ -1144,9 +1160,9 @@ static void r300TranslateVertexShader(struct r300_vertex_program *vp, VSF_FLAG_NONE) | (src[2].RelAddr << 4); - o_inst->src[1] = ZERO_SRC_2; - o_inst->src[2] = ZERO_SRC_2; - o_inst++; + inst[2] = ZERO_SRC_2; + inst[3] = ZERO_SRC_2; + inst += 4; src[2].File = PROGRAM_TEMPORARY; src[2].Index = u_temp_i; @@ -1157,12 +1173,12 @@ static void r300TranslateVertexShader(struct r300_vertex_program *vp, if (num_operands >= 2) { if (CMP_SRCS(src[1], src[0])) { - o_inst->opcode = + inst[0] = MAKE_VSF_OP(R300_VPI_OUT_OP_ADD, u_temp_i, VSF_FLAG_ALL, VSF_OUT_CLASS_TMP); - o_inst->src[0] = + inst[1] = MAKE_VSF_SOURCE(t_src_index (vp, &src[0]), SWIZZLE_X, SWIZZLE_Y, @@ -1172,9 +1188,9 @@ static void r300TranslateVertexShader(struct r300_vertex_program *vp, VSF_FLAG_NONE) | (src[0].RelAddr << 4); - o_inst->src[1] = ZERO_SRC_0; - o_inst->src[2] = ZERO_SRC_0; - o_inst++; + inst[2] = ZERO_SRC_0; + inst[3] = ZERO_SRC_0; + inst += 4; src[0].File = PROGRAM_TEMPORARY; src[0].Index = u_temp_i; @@ -1185,87 +1201,89 @@ static void r300TranslateVertexShader(struct r300_vertex_program *vp, switch (vpi->Opcode) { case OPCODE_ABS: - t_opcode_abs(vp, vpi, o_inst, src); + inst = t_opcode_abs(vp, vpi, inst, src); break; case OPCODE_ADD: - t_opcode_add(vp, vpi, o_inst, src); + inst = t_opcode_add(vp, vpi, inst, src); break; case OPCODE_ARL: - t_opcode_arl(vp, vpi, o_inst, src); + inst = t_opcode_arl(vp, vpi, inst, src); break; case OPCODE_DP3: - t_opcode_dp3(vp, vpi, o_inst, src); + inst = t_opcode_dp3(vp, vpi, inst, src); break; case OPCODE_DP4: - t_opcode_dp4(vp, vpi, o_inst, src); + inst = t_opcode_dp4(vp, vpi, inst, src); break; case OPCODE_DPH: - t_opcode_dph(vp, vpi, o_inst, src); + inst = t_opcode_dph(vp, vpi, inst, src); break; case OPCODE_DST: - t_opcode_dst(vp, vpi, o_inst, src); + inst = t_opcode_dst(vp, vpi, inst, src); break; case OPCODE_EX2: - t_opcode_ex2(vp, vpi, o_inst, src); + inst = t_opcode_ex2(vp, vpi, inst, src); break; case OPCODE_EXP: - t_opcode_exp(vp, vpi, o_inst, src); + inst = t_opcode_exp(vp, vpi, inst, src); break; case OPCODE_FLR: - /* FIXME */ - o_inst = t_opcode_flr(vp, vpi, o_inst, src, &u_temp_i); + inst = + t_opcode_flr(vp, vpi, inst, src, /* FIXME */ + &u_temp_i); break; case OPCODE_FRC: - t_opcode_frc(vp, vpi, o_inst, src); + inst = t_opcode_frc(vp, vpi, inst, src); break; case OPCODE_LG2: - t_opcode_lg2(vp, vpi, o_inst, src); + inst = t_opcode_lg2(vp, vpi, inst, src); break; case OPCODE_LIT: - t_opcode_lit(vp, vpi, o_inst, src); + inst = t_opcode_lit(vp, vpi, inst, src); break; case OPCODE_LOG: - t_opcode_log(vp, vpi, o_inst, src); + inst = t_opcode_log(vp, vpi, inst, src); break; case OPCODE_MAD: - t_opcode_mad(vp, vpi, o_inst, src); + inst = t_opcode_mad(vp, vpi, inst, src); break; case OPCODE_MAX: - t_opcode_max(vp, vpi, o_inst, src); + inst = t_opcode_max(vp, vpi, inst, src); break; case OPCODE_MIN: - t_opcode_min(vp, vpi, o_inst, src); + inst = t_opcode_min(vp, vpi, inst, src); break; case OPCODE_MOV: - t_opcode_mov(vp, vpi, o_inst, src); + inst = t_opcode_mov(vp, vpi, inst, src); break; case OPCODE_MUL: - t_opcode_mul(vp, vpi, o_inst, src); + inst = t_opcode_mul(vp, vpi, inst, src); break; case OPCODE_POW: - t_opcode_pow(vp, vpi, o_inst, src); + inst = t_opcode_pow(vp, vpi, inst, src); break; case OPCODE_RCP: - t_opcode_rcp(vp, vpi, o_inst, src); + inst = t_opcode_rcp(vp, vpi, inst, src); break; case OPCODE_RSQ: - t_opcode_rsq(vp, vpi, o_inst, src); + inst = t_opcode_rsq(vp, vpi, inst, src); break; case OPCODE_SGE: - t_opcode_sge(vp, vpi, o_inst, src); + inst = t_opcode_sge(vp, vpi, inst, src); break; case OPCODE_SLT: - t_opcode_slt(vp, vpi, o_inst, src); + inst = t_opcode_slt(vp, vpi, inst, src); break; case OPCODE_SUB: - t_opcode_sub(vp, vpi, o_inst, src); + inst = t_opcode_sub(vp, vpi, inst, src); break; case OPCODE_SWZ: - t_opcode_swz(vp, vpi, o_inst, src); + inst = t_opcode_swz(vp, vpi, inst, src); break; case OPCODE_XPD: - /* FIXME */ - o_inst = t_opcode_xpd(vp, vpi, o_inst, src, &u_temp_i); + inst = + t_opcode_xpd(vp, vpi, inst, src, /* FIXME */ + &u_temp_i); break; default: assert(0); @@ -1273,7 +1291,7 @@ static void r300TranslateVertexShader(struct r300_vertex_program *vp, } } - vp->program.length = (o_inst - vp->program.body.i) * 4; + vp->program.length = (inst - vp->program.body.i); if (vp->program.length >= VSF_MAX_FRAGMENT_LENGTH) { vp->program.length = 0; vp->native = GL_FALSE; diff --git a/src/mesa/drivers/dri/r300/r300_vertprog.h b/src/mesa/drivers/dri/r300/r300_vertprog.h index 2d399e243a..3df0eee799 100644 --- a/src/mesa/drivers/dri/r300/r300_vertprog.h +++ b/src/mesa/drivers/dri/r300/r300_vertprog.h @@ -3,11 +3,6 @@ #include "r300_reg.h" -struct r300_vertprog_instruction { - GLuint opcode; - GLuint src[3]; -}; - #define VSF_FLAG_X 1 #define VSF_FLAG_Y 2 #define VSF_FLAG_Z 4 diff --git a/src/mesa/drivers/dri/savage/savagetris.c b/src/mesa/drivers/dri/savage/savagetris.c index 3dd821a4d3..4ce2f60b4f 100644 --- a/src/mesa/drivers/dri/savage/savagetris.c +++ b/src/mesa/drivers/dri/savage/savagetris.c @@ -131,7 +131,9 @@ static __inline__ void savage_draw_point (savageContextPtr imesa, u_int32_t *vb = savageAllocVtxBuf (imesa, 6*vertsize); const GLfloat x = tmp->v.x; const GLfloat y = tmp->v.y; - const GLfloat sz = imesa->glCtx->Point._Size * .5; + const GLfloat sz = 0.5 * CLAMP(imesa->glCtx->Point.Size, + imesa->glCtx->Const.MinPointSize, + imesa->glCtx->Const.MaxPointSize); GLuint j; *(float *)&vb[0] = x - sz; @@ -164,7 +166,9 @@ static __inline__ void savage_draw_line (savageContextPtr imesa, savageVertexPtr v1 ) { GLuint vertsize = imesa->HwVertexSize; u_int32_t *vb = savageAllocVtxBuf (imesa, 6*vertsize); - GLfloat width = imesa->glCtx->Line._Width; + const GLfloat width = CLAMP(imesa->glCtx->Line.Width, + imesa->glCtx->Const.MinLineWidth, + imesa->glCtx->Const.MaxLineWidth); GLfloat dx, dy, ix, iy; GLuint j; @@ -234,7 +238,9 @@ static __inline__ void savage_ptex_line (savageContextPtr imesa, savageVertexPtr v1 ) { GLuint vertsize = imesa->HwVertexSize; u_int32_t *vb = savageAllocVtxBuf (imesa, 6*vertsize); - GLfloat width = imesa->glCtx->Line._Width; + const GLfloat width = CLAMP(imesa->glCtx->Line.Width, + imesa->glCtx->Const.MinLineWidth, + imesa->glCtx->Const.MaxLineWidth); GLfloat dx, dy, ix, iy; savageVertex tmp0, tmp1; GLuint j; @@ -281,7 +287,9 @@ static __inline__ void savage_ptex_point (savageContextPtr imesa, u_int32_t *vb = savageAllocVtxBuf (imesa, 6*vertsize); const GLfloat x = v0->v.x; const GLfloat y = v0->v.y; - const GLfloat sz = imesa->glCtx->Point._Size * .5; + const GLfloat sz = 0.5 * CLAMP(imesa->glCtx->Point.Size, + imesa->glCtx->Const.MinPointSize, + imesa->glCtx->Const.MaxPointSize); savageVertex tmp; GLuint j; diff --git a/src/mesa/drivers/dri/tdfx/tdfx_tris.c b/src/mesa/drivers/dri/tdfx/tdfx_tris.c index 96f9ae27fc..7252a7e7dc 100644 --- a/src/mesa/drivers/dri/tdfx/tdfx_tris.c +++ b/src/mesa/drivers/dri/tdfx/tdfx_tris.c @@ -184,7 +184,7 @@ tdfx_translate_vertex( GLcontext *ctx, const tdfxVertex *src, SWvertex *dst) } } - dst->pointSize = ctx->Point._Size; + dst->pointSize = ctx->Point.Size; } diff --git a/src/mesa/drivers/windows/gdi/wgl.c b/src/mesa/drivers/windows/gdi/wgl.c index fb23d210db..dad3dc1160 100644 --- a/src/mesa/drivers/windows/gdi/wgl.c +++ b/src/mesa/drivers/windows/gdi/wgl.c @@ -33,15 +33,30 @@ /* We're essentially building part of GDI here, so define this so that * we get the right export linkage. */ #ifdef __MINGW32__ -#include <GL/gl.h> + +#include <stdarg.h> +#include <windef.h> +#include <wincon.h> +#include <winbase.h> + +# if defined(BUILD_GL32) +# define WINGDIAPI __declspec(dllexport) +# else +# define __W32API_USE_DLLIMPORT__ +# endif + +#include <wingdi.h> +#include "GL/mesa_wgl.h" #include <stdlib.h> + #else + #define _GDI32_ -#endif #include <windows.h> -#include "glapi.h" +#endif +#include "glapi.h" #include "GL/wmesa.h" /* protos for wmesa* functions */ /* @@ -339,7 +354,7 @@ WINGDIAPI int GLAPIENTRY wglGetPixelFormat(HDC hdc) } WINGDIAPI BOOL GLAPIENTRY wglSetPixelFormat(HDC hdc,int iPixelFormat, - const PIXELFORMATDESCRIPTOR *ppfd) + const PIXELFORMATDESCRIPTOR *ppfd) { (void) hdc; @@ -392,12 +407,12 @@ static BOOL wglUseFontBitmaps_FX(HDC fontDevice, DWORD firstChar, bitDevice = CreateCompatibleDC(fontDevice); - // Swap fore and back colors so the bitmap has the right polarity + /* Swap fore and back colors so the bitmap has the right polarity */ tempColor = GetBkColor(bitDevice); SetBkColor(bitDevice, GetTextColor(bitDevice)); SetTextColor(bitDevice, tempColor); - // Place chars based on base line + /* Place chars based on base line */ VERIFY(SetTextAlign(bitDevice, TA_BASELINE) != GDI_ERROR ? 1 : 0); for(i = 0; i < (int)numChars; i++) { @@ -410,36 +425,36 @@ static BOOL wglUseFontBitmaps_FX(HDC fontDevice, DWORD firstChar, curChar = (char)(i + firstChar); - // Find how high/wide this character is + /* Find how high/wide this character is */ VERIFY(GetTextExtentPoint32(bitDevice, &curChar, 1, &size)); - // Create the output bitmap + /* Create the output bitmap */ charWidth = size.cx; charHeight = size.cy; - // Round up to the next multiple of 32 bits + /* Round up to the next multiple of 32 bits */ bmapWidth = ((charWidth + 31) / 32) * 32; bmapHeight = charHeight; bitObject = CreateCompatibleBitmap(bitDevice, bmapWidth, bmapHeight); - //VERIFY(bitObject); + /* VERIFY(bitObject); */ - // Assign the output bitmap to the device + /* Assign the output bitmap to the device */ origBmap = SelectObject(bitDevice, bitObject); (void) VERIFY(origBmap); VERIFY( PatBlt( bitDevice, 0, 0, bmapWidth, bmapHeight,BLACKNESS ) ); - // Use our source font on the device + /* Use our source font on the device */ VERIFY(SelectObject(bitDevice, GetCurrentObject(fontDevice,OBJ_FONT))); - // Draw the character + /* Draw the character */ VERIFY(TextOut(bitDevice, 0, metric.tmAscent, &curChar, 1)); - // Unselect our bmap object + /* Unselect our bmap object */ VERIFY(SelectObject(bitDevice, origBmap)); - // Convert the display dependant representation to a 1 bit deep DIB + /* Convert the display dependant representation to a 1 bit deep DIB */ numBytes = (bmapWidth * bmapHeight) / 8; bmap = malloc(numBytes); dibInfo->bmiHeader.biWidth = bmapWidth; @@ -447,24 +462,24 @@ static BOOL wglUseFontBitmaps_FX(HDC fontDevice, DWORD firstChar, res = GetDIBits(bitDevice, bitObject, 0, bmapHeight, bmap, dibInfo, DIB_RGB_COLORS); - //VERIFY(res); + /* VERIFY(res); */ - // Create the GL object + /* Create the GL object */ glNewList(i + listBase, GL_COMPILE); glBitmap(bmapWidth, bmapHeight, 0.0, (GLfloat)metric.tmDescent, (GLfloat)charWidth, 0.0, bmap); glEndList(); - // CheckGL(); + /* CheckGL(); */ - // Destroy the bmap object + /* Destroy the bmap object */ DeleteObject(bitObject); - // Deallocate the bitmap data + /* Deallocate the bitmap data */ free(bmap); } - // Destroy the DC + /* Destroy the DC */ VERIFY(DeleteDC(bitDevice)); free(dibInfo); diff --git a/src/mesa/drivers/windows/gdi/wmesa.c b/src/mesa/drivers/windows/gdi/wmesa.c index 2eec188912..5b67439f0f 100644 --- a/src/mesa/drivers/windows/gdi/wmesa.c +++ b/src/mesa/drivers/windows/gdi/wmesa.c @@ -6,6 +6,7 @@ #include "wmesadef.h" #include "colors.h" #include <GL/wmesa.h> +#include <winuser.h> #include "context.h" #include "extensions.h" #include "framebuffer.h" @@ -114,7 +115,7 @@ static void wmSetPixelFormat(WMesaFramebuffer pwfb, HDC hDC) { pwfb->cColorBits = GetDeviceCaps(hDC, BITSPIXEL); - // Only 16 and 32 bit targets are supported now + /* Only 16 and 32 bit targets are supported now */ assert(pwfb->cColorBits == 0 || pwfb->cColorBits == 16 || pwfb->cColorBits == 32); @@ -1171,7 +1172,7 @@ WMesaContext WMesaCreateContext(HDC hDC, /* I do not understand this contributed code */ /* Support memory and device contexts */ if(WindowFromDC(hDC) != NULL) { - c->hDC = GetDC(WindowFromDC(hDC)); // huh ???? + c->hDC = GetDC(WindowFromDC(hDC)); /* huh ???? */ } else { c->hDC = hDC; @@ -1404,6 +1405,7 @@ void WMesaSwapBuffers( HDC hdc ) * table entries. Hopefully, I'll find a better solution. The * dispatch table generation scripts ought to be making these dummy * stubs as well. */ +#if !defined(__MINGW32__) || !defined(GL_NO_STDCALL) void gl_dispatch_stub_543(void){} void gl_dispatch_stub_544(void){} void gl_dispatch_stub_545(void){} @@ -1471,3 +1473,4 @@ void gl_dispatch_stub_769(void){} void gl_dispatch_stub_770(void){} void gl_dispatch_stub_771(void){} +#endif diff --git a/src/mesa/drivers/windows/gdi/wmesadef.h b/src/mesa/drivers/windows/gdi/wmesadef.h index 97b063a8ba..83a42e6082 100644 --- a/src/mesa/drivers/windows/gdi/wmesadef.h +++ b/src/mesa/drivers/windows/gdi/wmesadef.h @@ -1,6 +1,8 @@ #ifndef WMESADEF_H #define WMESADEF_H - +#ifdef __MINGW32__ +#include <windows.h> +#endif #include "context.h" diff --git a/src/mesa/drivers/x11/fakeglx.c b/src/mesa/drivers/x11/fakeglx.c index 1587df66bc..7a170b4d3d 100644 --- a/src/mesa/drivers/x11/fakeglx.c +++ b/src/mesa/drivers/x11/fakeglx.c @@ -296,6 +296,11 @@ save_glx_visual( Display *dpy, XVisualInfo *vinfo, } } + if (stereoFlag) { + /* stereo not supported */ + return NULL; + } + /* Comparing IDs uses less memory but sometimes fails. */ /* XXX revisit this after 3.0 is finished. */ if (_mesa_getenv("MESA_GLX_VISUAL_HACK")) @@ -1079,7 +1084,7 @@ choose_visual( Display *dpy, int screen, const int *list, GLboolean fbConfig ) else { stereo_flag = GL_TRUE; } - return NULL; /* stereo not supported */ + break; case GLX_AUX_BUFFERS: parselist++; numAux = *parselist++; diff --git a/src/mesa/main/attrib.c b/src/mesa/main/attrib.c index 1aa0a02fc7..b422198f92 100644 --- a/src/mesa/main/attrib.c +++ b/src/mesa/main/attrib.c @@ -98,9 +98,13 @@ _mesa_PushAttrib(GLbitfield mask) } if (mask & GL_COLOR_BUFFER_BIT) { + GLuint i; struct gl_colorbuffer_attrib *attr; attr = MALLOC_STRUCT( gl_colorbuffer_attrib ); MEMCPY( attr, &ctx->Color, sizeof(struct gl_colorbuffer_attrib) ); + /* push the Draw FBO's DrawBuffer[] state, not ctx->Color.DrawBuffer[] */ + for (i = 0; i < ctx->Const.MaxDrawBuffers; i ++) + attr->DrawBuffer[i] = ctx->DrawBuffer->ColorDrawBuffer[i]; newnode = new_attrib_node( GL_COLOR_BUFFER_BIT ); newnode->data = attr; newnode->next = head; diff --git a/src/mesa/main/buffers.c b/src/mesa/main/buffers.c index c280f89e1d..0e6ca8ea1c 100644 --- a/src/mesa/main/buffers.c +++ b/src/mesa/main/buffers.c @@ -370,6 +370,14 @@ _mesa_DrawBuffer(GLenum buffer) /* if we get here, there's no error so set new state */ _mesa_drawbuffers(ctx, 1, &buffer, &destMask); + + /* + * Call device driver function. + */ + if (ctx->Driver.DrawBuffers) + ctx->Driver.DrawBuffers(ctx, 1, &buffer); + else if (ctx->Driver.DrawBuffer) + ctx->Driver.DrawBuffer(ctx, buffer); } @@ -435,6 +443,14 @@ _mesa_DrawBuffersARB(GLsizei n, const GLenum *buffers) /* OK, if we get here, there were no errors so set the new state */ _mesa_drawbuffers(ctx, n, buffers, destMask); + + /* + * Call device driver function. + */ + if (ctx->Driver.DrawBuffers) + ctx->Driver.DrawBuffers(ctx, n, buffers); + else if (ctx->Driver.DrawBuffer) + ctx->Driver.DrawBuffer(ctx, buffers[0]); } @@ -463,14 +479,15 @@ set_color_output(GLcontext *ctx, GLuint output, GLenum buffer, /* not really needed, will be set later */ fb->_NumColorDrawBuffers[output] = 0; + if (fb->Name == 0) /* Set traditional state var */ - ctx->Color.DrawBuffer[output] = buffer; + ctx->Color.DrawBuffer[output] = buffer; } /** * Helper routine used by _mesa_DrawBuffer, _mesa_DrawBuffersARB and - * _mesa_PopAttrib to set drawbuffer state. + * other places (window fbo fixup) to set fbo (and the old ctx) fields. * All error checking will have been done prior to calling this function * so nothing should go wrong at this point. * \param ctx current context @@ -509,30 +526,15 @@ _mesa_drawbuffers(GLcontext *ctx, GLuint n, const GLenum *buffers, } ctx->NewState |= _NEW_COLOR; - - /* - * Call device driver function. - */ - if (ctx->Driver.DrawBuffers) - ctx->Driver.DrawBuffers(ctx, n, buffers); - else if (ctx->Driver.DrawBuffer) - ctx->Driver.DrawBuffer(ctx, buffers[0]); } - -/** - * Called by glReadBuffer to set the source renderbuffer for reading pixels. - * \param mode color buffer such as GL_FRONT, GL_BACK, etc. - */ -void GLAPIENTRY -_mesa_ReadBuffer(GLenum buffer) +GLboolean +_mesa_readbuffer_update_fields(GLcontext *ctx, GLenum buffer) { struct gl_framebuffer *fb; GLbitfield supportedMask; GLint srcBuffer; - GET_CURRENT_CONTEXT(ctx); - ASSERT_OUTSIDE_BEGIN_END_AND_FLUSH(ctx); fb = ctx->ReadBuffer; @@ -548,12 +550,12 @@ _mesa_ReadBuffer(GLenum buffer) srcBuffer = read_buffer_enum_to_index(buffer); if (srcBuffer == -1) { _mesa_error(ctx, GL_INVALID_ENUM, "glReadBuffer(buffer=0x%x)", buffer); - return; + return GL_FALSE; } supportedMask = supported_buffer_bitmask(ctx, fb); if (((1 << srcBuffer) & supportedMask) == 0) { _mesa_error(ctx, GL_INVALID_OPERATION, "glReadBuffer(buffer=0x%x)", buffer); - return; + return GL_FALSE; } } @@ -563,6 +565,27 @@ _mesa_ReadBuffer(GLenum buffer) fb->ColorReadBuffer = buffer; fb->_ColorReadBufferIndex = srcBuffer; + return GL_TRUE; +} + + + +/** + * Called by glReadBuffer to set the source renderbuffer for reading pixels. + * \param mode color buffer such as GL_FRONT, GL_BACK, etc. + */ +void GLAPIENTRY +_mesa_ReadBuffer(GLenum buffer) +{ + GET_CURRENT_CONTEXT(ctx); + ASSERT_OUTSIDE_BEGIN_END_AND_FLUSH(ctx); + + if (MESA_VERBOSE & VERBOSE_API) + _mesa_debug(ctx, "glReadBuffer %s\n", _mesa_lookup_enum_by_nr(buffer)); + + if (!_mesa_readbuffer_update_fields(ctx, buffer)) + return; + ctx->NewState |= _NEW_PIXEL; /* diff --git a/src/mesa/main/buffers.h b/src/mesa/main/buffers.h index fcc2152342..208e7af2b9 100644 --- a/src/mesa/main/buffers.h +++ b/src/mesa/main/buffers.h @@ -56,6 +56,9 @@ extern void _mesa_drawbuffers(GLcontext *ctx, GLuint n, const GLenum *buffers, const GLbitfield *destMask); +extern GLboolean +_mesa_readbuffer_update_fields(GLcontext *ctx, GLenum buffer); + extern void GLAPIENTRY _mesa_ReadBuffer( GLenum mode ); diff --git a/src/mesa/main/context.c b/src/mesa/main/context.c index 2ad1badac7..00e4c8328e 100644 --- a/src/mesa/main/context.c +++ b/src/mesa/main/context.c @@ -1496,9 +1496,8 @@ _mesa_make_current( GLcontext *newCtx, GLframebuffer *drawBuffer, if (!newCtx->DrawBuffer || newCtx->DrawBuffer->Name == 0) { _mesa_reference_framebuffer(&newCtx->DrawBuffer, drawBuffer); /* fix up the fb fields - these will end up wrong otherwise - if the DRIdrawable changes, and someone may rely on them. - */ - /* What a mess!?! */ + if the DRIdrawable changes, and everything relies on them. + This is a bit messy (same as needed in _mesa_BindFramebufferEXT) */ int i; GLenum buffers[MAX_DRAW_BUFFERS]; for(i = 0; i < newCtx->Const.MaxDrawBuffers; i++) { @@ -1508,7 +1507,7 @@ _mesa_make_current( GLcontext *newCtx, GLframebuffer *drawBuffer, } if (!newCtx->ReadBuffer || newCtx->ReadBuffer->Name == 0) { _mesa_reference_framebuffer(&newCtx->ReadBuffer, readBuffer); - _mesa_ReadBuffer(newCtx->Pixel.ReadBuffer); + _mesa_readbuffer_update_fields(newCtx, newCtx->Pixel.ReadBuffer); } newCtx->NewState |= _NEW_BUFFERS; diff --git a/src/mesa/main/dd.h b/src/mesa/main/dd.h index 88f33943b3..caa50dd682 100644 --- a/src/mesa/main/dd.h +++ b/src/mesa/main/dd.h @@ -782,7 +782,7 @@ struct dd_function_table { struct gl_framebuffer * (*NewFramebuffer)(GLcontext *ctx, GLuint name); struct gl_renderbuffer * (*NewRenderbuffer)(GLcontext *ctx, GLuint name); void (*BindFramebuffer)(GLcontext *ctx, GLenum target, - struct gl_framebuffer *fb); + struct gl_framebuffer *fb, struct gl_framebuffer *fbread); void (*FramebufferRenderbuffer)(GLcontext *ctx, struct gl_framebuffer *fb, GLenum attachment, diff --git a/src/mesa/main/fbobject.c b/src/mesa/main/fbobject.c index e3bada5ae8..6f7effcce7 100644 --- a/src/mesa/main/fbobject.c +++ b/src/mesa/main/fbobject.c @@ -29,6 +29,7 @@ */ +#include "buffers.h" #include "context.h" #include "fbobject.h" #include "framebuffer.h" @@ -924,7 +925,7 @@ check_end_texture_render(GLcontext *ctx, struct gl_framebuffer *fb) void GLAPIENTRY _mesa_BindFramebufferEXT(GLenum target, GLuint framebuffer) { - struct gl_framebuffer *newFb; + struct gl_framebuffer *newFb, *newFbread; GLboolean bindReadBuf, bindDrawBuf; GET_CURRENT_CONTEXT(ctx); @@ -984,12 +985,14 @@ _mesa_BindFramebufferEXT(GLenum target, GLuint framebuffer) } _mesa_HashInsert(ctx->Shared->FrameBuffers, framebuffer, newFb); } + newFbread = newFb; } else { /* Binding the window system framebuffer (which was originally set * with MakeCurrent). */ newFb = ctx->WinSysDrawBuffer; + newFbread = ctx->WinSysReadBuffer; } ASSERT(newFb); @@ -999,8 +1002,16 @@ _mesa_BindFramebufferEXT(GLenum target, GLuint framebuffer) * XXX check if re-binding same buffer and skip some of this code. */ + /* for window-framebuffers, re-initialize the fbo values, as they + could be wrong (makecurrent with a new drawable while still a fbo + was bound will lead to default init fbo values). + note that therefore the context ReadBuffer/DrawBuffer values are not + valid while fbo's are bound!!! */ if (bindReadBuf) { - _mesa_reference_framebuffer(&ctx->ReadBuffer, newFb); + _mesa_reference_framebuffer(&ctx->ReadBuffer, newFbread); + if (!newFbread->Name) { + _mesa_readbuffer_update_fields(ctx, ctx->Pixel.ReadBuffer); + } } if (bindDrawBuf) { @@ -1008,14 +1019,22 @@ _mesa_BindFramebufferEXT(GLenum target, GLuint framebuffer) check_end_texture_render(ctx, ctx->DrawBuffer); /* check if time to delete this framebuffer */ _mesa_reference_framebuffer(&ctx->DrawBuffer, newFb); - if (newFb->Name != 0) { + if (!newFb->Name) { + GLuint i; + GLenum buffers[MAX_DRAW_BUFFERS]; + for(i = 0; i < ctx->Const.MaxDrawBuffers; i++) { + buffers[i] = ctx->Color.DrawBuffer[i]; + } + _mesa_drawbuffers(ctx, ctx->Const.MaxDrawBuffers, buffers, NULL); + } + else { /* check if newly bound framebuffer has any texture attachments */ check_begin_texture_render(ctx, newFb); } } if (ctx->Driver.BindFramebuffer) { - ctx->Driver.BindFramebuffer(ctx, target, newFb); + ctx->Driver.BindFramebuffer(ctx, target, newFb, newFbread); } } diff --git a/src/mesa/main/framebuffer.c b/src/mesa/main/framebuffer.c index dc10d9ffbc..c9b30d3252 100644 --- a/src/mesa/main/framebuffer.c +++ b/src/mesa/main/framebuffer.c @@ -654,6 +654,27 @@ update_color_read_buffer(GLcontext *ctx, struct gl_framebuffer *fb) } +static void +update_framebuffer(GLcontext *ctx, struct gl_framebuffer *fb) +{ + /* Completeness only matters for user-created framebuffers */ + if (fb->Name != 0) { + /* XXX: EXT_framebuffer_blit: + framebuffer must still be complete wrt read/draw? */ + _mesa_test_framebuffer_completeness(ctx, fb); + _mesa_update_framebuffer_visual(fb); + } + + /* update_color_draw/read_buffers not needed for + read/draw only fb, but shouldn't hurt ??? */ + update_color_draw_buffers(ctx, fb); + update_color_read_buffer(ctx, fb); + _mesa_update_depth_buffer(ctx, fb, BUFFER_DEPTH); + _mesa_update_stencil_buffer(ctx, fb, BUFFER_STENCIL); + + compute_depth_max(fb); +} + /** * Update state related to the current draw/read framebuffers. * Specifically, update these framebuffer fields: @@ -671,19 +692,11 @@ void _mesa_update_framebuffer(GLcontext *ctx) { struct gl_framebuffer *fb = ctx->DrawBuffer; + struct gl_framebuffer *fbread = ctx->ReadBuffer; - /* Completeness only matters for user-created framebuffers */ - if (fb->Name != 0) { - _mesa_test_framebuffer_completeness(ctx, fb); - _mesa_update_framebuffer_visual(fb); - } - - update_color_draw_buffers(ctx, fb); - update_color_read_buffer(ctx, fb); - _mesa_update_depth_buffer(ctx, fb, BUFFER_DEPTH); - _mesa_update_stencil_buffer(ctx, fb, BUFFER_STENCIL); - - compute_depth_max(fb); + update_framebuffer(ctx, fb); + if (fbread != fb) + update_framebuffer(ctx, fbread); } diff --git a/src/mesa/main/glheader.h b/src/mesa/main/glheader.h index 63dd002a41..fd4127558a 100644 --- a/src/mesa/main/glheader.h +++ b/src/mesa/main/glheader.h @@ -92,7 +92,7 @@ #endif #ifdef WGLAPI -#undef WGLAPI +# undef WGLAPI #endif #if !defined(OPENSTEP) && (defined(__WIN32__) && !defined(__CYGWIN__)) && !defined(BUILD_FOR_SNAP) diff --git a/src/mesa/main/imports.c b/src/mesa/main/imports.c index 8a5dfdb4b8..3ae56c8b0b 100644 --- a/src/mesa/main/imports.c +++ b/src/mesa/main/imports.c @@ -575,7 +575,11 @@ _mesa_ffs(int i) * if no bits set. */ int +#ifdef __MINGW32__ +_mesa_ffsll(long val) +#else _mesa_ffsll(long long val) +#endif { #ifdef ffsll return ffsll(val); diff --git a/src/mesa/main/imports.h b/src/mesa/main/imports.h index 9be8014a13..ebdfc452a7 100644 --- a/src/mesa/main/imports.h +++ b/src/mesa/main/imports.h @@ -700,7 +700,11 @@ extern int _mesa_ffs(int i); extern int +#ifdef __MINGW32__ +_mesa_ffsll(long i); +#else _mesa_ffsll(long long i); +#endif extern unsigned int _mesa_bitcount(unsigned int n); diff --git a/src/mesa/main/lines.c b/src/mesa/main/lines.c index dc7195d4eb..0c2dbf915a 100644 --- a/src/mesa/main/lines.c +++ b/src/mesa/main/lines.c @@ -55,9 +55,6 @@ _mesa_LineWidth( GLfloat width ) FLUSH_VERTICES(ctx, _NEW_LINE); ctx->Line.Width = width; - ctx->Line._Width = CLAMP(width, - ctx->Const.MinLineWidth, - ctx->Const.MaxLineWidth); if (ctx->Driver.LineWidth) ctx->Driver.LineWidth(ctx, width); @@ -105,13 +102,12 @@ _mesa_LineStipple( GLint factor, GLushort pattern ) * Initializes __GLcontextRec::Line and line related constants in * __GLcontextRec::Const. */ -void GLAPIENTRY _mesa_init_line( GLcontext * ctx ) +void GLAPIENTRY +_mesa_init_line( GLcontext * ctx ) { - /* Line group */ ctx->Line.SmoothFlag = GL_FALSE; ctx->Line.StippleFlag = GL_FALSE; ctx->Line.Width = 1.0; - ctx->Line._Width = 1.0; ctx->Line.StipplePattern = 0xffff; ctx->Line.StippleFactor = 1; } diff --git a/src/mesa/main/mtypes.h b/src/mesa/main/mtypes.h index c150aae4d7..52448ee04e 100644 --- a/src/mesa/main/mtypes.h +++ b/src/mesa/main/mtypes.h @@ -918,7 +918,6 @@ struct gl_line_attrib GLushort StipplePattern; /**< Stipple pattern */ GLint StippleFactor; /**< Stipple repeat factor */ GLfloat Width; /**< Line width */ - GLfloat _Width; /**< Clamped Line width */ }; @@ -1064,7 +1063,6 @@ struct gl_point_attrib { GLboolean SmoothFlag; /**< True if GL_POINT_SMOOTH is enabled */ GLfloat Size; /**< User-specified point size */ - GLfloat _Size; /**< Size clamped to Const.Min/MaxPointSize */ GLfloat Params[3]; /**< GL_EXT_point_parameters */ GLfloat MinSize, MaxSize; /**< GL_EXT_point_parameters */ GLfloat Threshold; /**< GL_EXT_point_parameters */ @@ -1134,13 +1132,13 @@ struct gl_stencil_attrib * An index for each type of texture object */ /*@{*/ -#define TEXTURE_1D_INDEX 0 -#define TEXTURE_2D_INDEX 1 -#define TEXTURE_3D_INDEX 2 -#define TEXTURE_CUBE_INDEX 3 -#define TEXTURE_RECT_INDEX 4 -#define TEXTURE_1D_ARRAY_INDEX 5 -#define TEXTURE_2D_ARRAY_INDEX 6 +#define TEXTURE_1D_INDEX 0 +#define TEXTURE_2D_INDEX 1 +#define TEXTURE_3D_INDEX 2 +#define TEXTURE_CUBE_INDEX 3 +#define TEXTURE_RECT_INDEX 4 +#define TEXTURE_1D_ARRAY_INDEX 5 +#define TEXTURE_2D_ARRAY_INDEX 6 /*@}*/ /** @@ -1148,13 +1146,13 @@ struct gl_stencil_attrib * Used for Texture.Unit[]._ReallyEnabled flags. */ /*@{*/ -#define TEXTURE_1D_BIT (1 << TEXTURE_1D_INDEX) -#define TEXTURE_2D_BIT (1 << TEXTURE_2D_INDEX) -#define TEXTURE_3D_BIT (1 << TEXTURE_3D_INDEX) -#define TEXTURE_CUBE_BIT (1 << TEXTURE_CUBE_INDEX) -#define TEXTURE_RECT_BIT (1 << TEXTURE_RECT_INDEX) -#define TEXTURE_1D_ARRAY_BIT (1 << TEXTURE_1D_ARRAY_INDEX) -#define TEXTURE_2D_ARRAY_BIT (1 << TEXTURE_2D_ARRAY_INDEX) +#define TEXTURE_1D_BIT (1 << TEXTURE_1D_INDEX) +#define TEXTURE_2D_BIT (1 << TEXTURE_2D_INDEX) +#define TEXTURE_3D_BIT (1 << TEXTURE_3D_INDEX) +#define TEXTURE_CUBE_BIT (1 << TEXTURE_CUBE_INDEX) +#define TEXTURE_RECT_BIT (1 << TEXTURE_RECT_INDEX) +#define TEXTURE_1D_ARRAY_BIT (1 << TEXTURE_1D_ARRAY_INDEX) +#define TEXTURE_2D_ARRAY_BIT (1 << TEXTURE_2D_ARRAY_INDEX) /*@}*/ @@ -1328,8 +1326,6 @@ struct gl_texture_format }; -#define MAX_3D_TEXTURE_SIZE (1 << (MAX_3D_TEXTURE_LEVELS - 1)) - /** * Texture image state. Describes the dimensions of a texture image, * the texel format and pointers to Texel Fetch functions. @@ -1394,7 +1390,7 @@ struct gl_texture_image #define FACE_NEG_Y 3 #define FACE_POS_Z 4 #define FACE_NEG_Z 5 -#define MAX_FACES 6 +#define MAX_FACES 6 /*@}*/ @@ -2194,12 +2190,11 @@ struct gl_shared_state * \todo Improve the granularity of locking. */ /*@{*/ - _glthread_Mutex TexMutex; /**< texobj thread safety */ - GLuint TextureStateStamp; /**< state notification for shared tex */ + _glthread_Mutex TexMutex; /**< texobj thread safety */ + GLuint TextureStateStamp; /**< state notification for shared tex */ /*@}*/ - /** * \name Vertex/fragment programs */ diff --git a/src/mesa/main/points.c b/src/mesa/main/points.c index 0f562420b0..e83db5de78 100644 --- a/src/mesa/main/points.c +++ b/src/mesa/main/points.c @@ -57,10 +57,6 @@ _mesa_PointSize( GLfloat size ) FLUSH_VERTICES(ctx, _NEW_POINT); ctx->Point.Size = size; - /* XXX correct clamp limits? */ - ctx->Point._Size = CLAMP(ctx->Point.Size, - ctx->Point.MinSize, - ctx->Point.MaxSize); if (ctx->Driver.PointSize) ctx->Driver.PointSize(ctx, size); @@ -253,7 +249,6 @@ _mesa_init_point(GLcontext *ctx) ctx->Point.SmoothFlag = GL_FALSE; ctx->Point.Size = 1.0; - ctx->Point._Size = 1.0; ctx->Point.Params[0] = 1.0; ctx->Point.Params[1] = 0.0; ctx->Point.Params[2] = 0.0; diff --git a/src/mesa/main/shaders.c b/src/mesa/main/shaders.c index 58be1f46e5..7bf8808767 100644 --- a/src/mesa/main/shaders.c +++ b/src/mesa/main/shaders.c @@ -83,7 +83,7 @@ _mesa_CreateShader(GLenum type) } -GLhandleARB APIENTRY +GLhandleARB GLAPIENTRY _mesa_CreateShaderObjectARB(GLenum type) { GET_CURRENT_CONTEXT(ctx); @@ -99,7 +99,7 @@ _mesa_CreateProgram(void) } -GLhandleARB APIENTRY +GLhandleARB GLAPIENTRY _mesa_CreateProgramObjectARB(void) { GET_CURRENT_CONTEXT(ctx); @@ -319,7 +319,7 @@ _mesa_GetUniformivARB(GLhandleARB program, GLint location, GLint * params) #if 0 -GLint APIENTRY +GLint GLAPIENTRY _mesa_GetUniformLocation(GLuint program, const GLcharARB *name) { GET_CURRENT_CONTEXT(ctx); @@ -336,7 +336,7 @@ _mesa_GetHandleARB(GLenum pname) } -GLint APIENTRY +GLint GLAPIENTRY _mesa_GetUniformLocationARB(GLhandleARB programObj, const GLcharARB *name) { GET_CURRENT_CONTEXT(ctx); diff --git a/src/mesa/main/state.c b/src/mesa/main/state.c index 66f8ac6408..444f227760 100644 --- a/src/mesa/main/state.c +++ b/src/mesa/main/state.c @@ -1068,7 +1068,7 @@ update_tricaps(GLcontext *ctx, GLbitfield new_state) if (1/*new_state & _NEW_POINT*/) { if (ctx->Point.SmoothFlag) ctx->_TriangleCaps |= DD_POINT_SMOOTH; - if (ctx->Point._Size != 1.0F) + if (ctx->Point.Size != 1.0F) ctx->_TriangleCaps |= DD_POINT_SIZE; if (ctx->Point._Attenuated) ctx->_TriangleCaps |= DD_POINT_ATTEN; @@ -1082,7 +1082,7 @@ update_tricaps(GLcontext *ctx, GLbitfield new_state) ctx->_TriangleCaps |= DD_LINE_SMOOTH; if (ctx->Line.StippleFlag) ctx->_TriangleCaps |= DD_LINE_STIPPLE; - if (ctx->Line._Width != 1.0) + if (ctx->Line.Width != 1.0) ctx->_TriangleCaps |= DD_LINE_WIDTH; } diff --git a/src/mesa/main/texcompress_fxt1.c b/src/mesa/main/texcompress_fxt1.c index 411d51cfcc..b6991f45ed 100644 --- a/src/mesa/main/texcompress_fxt1.c +++ b/src/mesa/main/texcompress_fxt1.c @@ -302,7 +302,12 @@ const struct gl_texture_format _mesa_texformat_rgba_fxt1 = { #define FX64_NATIVE 1 +#ifdef __MINGW32__ +typedef unsigned long Fx64; +#else typedef unsigned long long Fx64; +#endif + #define FX64_MOV32(a, b) a = b #define FX64_OR32(a, b) a |= b diff --git a/src/mesa/shader/slang/slang_codegen.c b/src/mesa/shader/slang/slang_codegen.c index 35e80e0452..8b2bdd74b5 100644 --- a/src/mesa/shader/slang/slang_codegen.c +++ b/src/mesa/shader/slang/slang_codegen.c @@ -536,7 +536,7 @@ new_not(slang_ir_node *n) static slang_ir_node * new_inlined_function_call(slang_ir_node *code, slang_label *name) { - slang_ir_node *n = new_node1(IR_FUNC, code); + slang_ir_node *n = new_node1(IR_CALL, code); assert(name); if (n) n->Label = name; @@ -1202,17 +1202,29 @@ _slang_gen_function_call(slang_assemble_ctx *A, slang_function *fun, /* non-assembly function */ inlined = slang_inline_function_call(A, fun, oper, dest); if (inlined && _slang_find_node_type(inlined, SLANG_OPER_RETURN)) { - /* This inlined function has one or more 'return' statements. + slang_operation *callOper; + /* The function we're calling has one or more 'return' statements. * So, we can't truly inline this function because we need to * implement 'return' with RET (and CAL). + * Nevertheless, we performed "inlining" to make a new instance + * of the function body to deal with static register allocation. + * * XXX check if there's one 'return' and if it's the very last * statement in the function - we can optimize that case. */ assert(inlined->type == SLANG_OPER_BLOCK_NEW_SCOPE || inlined->type == SLANG_OPER_SEQUENCE); - inlined->type = SLANG_OPER_INLINED_CALL; - inlined->fun = fun; - inlined->label = _slang_label_new_unique((char*) fun->header.a_name); + if (_slang_function_has_return_value(fun) && !dest) { + assert(inlined->children[0].type == SLANG_OPER_VARIABLE_DECL); + assert(inlined->children[2].type == SLANG_OPER_IDENTIFIER); + callOper = &inlined->children[1]; + } + else { + callOper = inlined; + } + callOper->type = SLANG_OPER_INLINED_CALL; + callOper->fun = fun; + callOper->label = _slang_label_new_unique((char*) fun->header.a_name); } } @@ -1949,8 +1961,7 @@ static slang_ir_node * _slang_gen_return(slang_assemble_ctx * A, slang_operation *oper) { const GLboolean haveReturnValue - = (oper->num_children == 1 && - oper->children[0].type != SLANG_OPER_VOID); + = (oper->num_children == 1 && oper->children[0].type != SLANG_OPER_VOID); /* error checking */ assert(A->CurFunction); @@ -1960,7 +1971,7 @@ _slang_gen_return(slang_assemble_ctx * A, slang_operation *oper) return NULL; } else if (!haveReturnValue && - A->CurFunction->header.type.specifier.type != SLANG_SPEC_VOID) { + A->CurFunction->header.type.specifier.type != SLANG_SPEC_VOID) { slang_info_log_error(A->log, "return statement requires an expression"); return NULL; } @@ -2236,7 +2247,9 @@ _slang_gen_assignment(slang_assemble_ctx * A, slang_operation *oper) } if (var->type.qualifier == SLANG_QUAL_CONST || var->type.qualifier == SLANG_QUAL_ATTRIBUTE || - var->type.qualifier == SLANG_QUAL_UNIFORM) { + var->type.qualifier == SLANG_QUAL_UNIFORM || + (var->type.qualifier == SLANG_QUAL_VARYING && + A->program->Target == GL_FRAGMENT_PROGRAM_ARB)) { slang_info_log_error(A->log, "illegal assignment to read-only variable '%s'", (char *) oper->children[0].a_id); @@ -2264,10 +2277,11 @@ _slang_gen_assignment(slang_assemble_ctx * A, slang_operation *oper) lhs = _slang_gen_operation(A, &oper->children[0]); if (lhs) { - if (lhs->Store->File != PROGRAM_OUTPUT && - lhs->Store->File != PROGRAM_TEMPORARY && - lhs->Store->File != PROGRAM_VARYING && - lhs->Store->File != PROGRAM_UNDEFINED) { + if (!(lhs->Store->File == PROGRAM_OUTPUT || + lhs->Store->File == PROGRAM_TEMPORARY || + (lhs->Store->File == PROGRAM_VARYING && + A->program->Target == GL_VERTEX_PROGRAM_ARB) || + lhs->Store->File == PROGRAM_UNDEFINED)) { slang_info_log_error(A->log, "illegal assignment to read-only l-value"); return NULL; diff --git a/src/mesa/shader/slang/slang_emit.c b/src/mesa/shader/slang/slang_emit.c index 02c74095a9..fe13f2865c 100644 --- a/src/mesa/shader/slang/slang_emit.c +++ b/src/mesa/shader/slang/slang_emit.c @@ -780,16 +780,18 @@ emit_label(slang_emit_info *emitInfo, const slang_ir_node *n) /** - * Emit code for an inlined function call (subroutine). + * Emit code for a function call. + * Note that for each time a function is called, we emit the function's + * body code again because the set of available registers may be different. */ static struct prog_instruction * -emit_func(slang_emit_info *emitInfo, slang_ir_node *n) +emit_fcall(slang_emit_info *emitInfo, slang_ir_node *n) { struct gl_program *progSave; struct prog_instruction *inst; GLuint subroutineId; - assert(n->Opcode == IR_FUNC); + assert(n->Opcode == IR_CALL); assert(n->Label); /* save/push cur program */ @@ -1687,10 +1689,10 @@ emit(slang_emit_info *emitInfo, slang_ir_node *n) case IR_KILL: return emit_kill(emitInfo); - case IR_FUNC: - /* new variable scope for subroutines/function calls*/ + case IR_CALL: + /* new variable scope for subroutines/function calls */ _slang_push_var_table(emitInfo->vt); - inst = emit_func(emitInfo, n); + inst = emit_fcall(emitInfo, n); _slang_pop_var_table(emitInfo->vt); return inst; @@ -1782,7 +1784,7 @@ _slang_resolve_subroutines(slang_emit_info *emitInfo) emitInfo->NumSubroutines = 0; /* Examine CAL instructions. - * At this point, the BranchTarget field of the CAL instructions is + * At this point, the BranchTarget field of the CAL instruction is * the number/id of the subroutine to call (an index into the * emitInfo->Subroutines list). * Translate that into an actual instruction location now. diff --git a/src/mesa/shader/slang/slang_ir.c b/src/mesa/shader/slang/slang_ir.c index a29f302687..92e8d0345e 100644 --- a/src/mesa/shader/slang/slang_ir.c +++ b/src/mesa/shader/slang/slang_ir.c @@ -311,7 +311,7 @@ _slang_print_ir_tree(const slang_ir_node *n, int indent) printf("RETURN\n"); break; case IR_CALL: - printf("CALL\n"); + printf("CALL %s\n", n->Label->Name); break; case IR_LOOP: diff --git a/src/mesa/shader/slang/slang_ir.h b/src/mesa/shader/slang/slang_ir.h index 69db4b5451..c7c0ddbf9a 100644 --- a/src/mesa/shader/slang/slang_ir.h +++ b/src/mesa/shader/slang/slang_ir.h @@ -62,8 +62,6 @@ typedef enum IR_RETURN, /* return from subroutine */ IR_CALL, /* call subroutine */ - IR_FUNC, /* inlined function code */ - IR_LOOP, /* high-level loop-begin / loop-end */ /* Children[0] = loop body */ /* Children[1] = loop tail code, or NULL */ diff --git a/src/mesa/swrast/s_aalinetemp.h b/src/mesa/swrast/s_aalinetemp.h index 69a1f0cd39..e7911fec3b 100644 --- a/src/mesa/swrast/s_aalinetemp.h +++ b/src/mesa/swrast/s_aalinetemp.h @@ -132,7 +132,9 @@ NAME(line)(GLcontext *ctx, const SWvertex *v0, const SWvertex *v1) line.dx = line.x1 - line.x0; line.dy = line.y1 - line.y0; line.len = SQRTF(line.dx * line.dx + line.dy * line.dy); - line.halfWidth = 0.5F * ctx->Line._Width; + line.halfWidth = 0.5F * CLAMP(ctx->Line.Width, + ctx->Const.MinLineWidthAA, + ctx->Const.MaxLineWidthAA); if (line.len == 0.0 || IS_INF_OR_NAN(line.len)) return; diff --git a/src/mesa/swrast/s_lines.c b/src/mesa/swrast/s_lines.c index 781146e67f..3de438760b 100644 --- a/src/mesa/swrast/s_lines.c +++ b/src/mesa/swrast/s_lines.c @@ -1,6 +1,6 @@ /* * Mesa 3-D graphics library - * Version: 6.5.3 + * Version: 7.1 * * Copyright (C) 1999-2007 Brian Paul All Rights Reserved. * @@ -63,12 +63,13 @@ compute_stipple_mask( GLcontext *ctx, GLuint len, GLubyte mask[] ) static void draw_wide_line( GLcontext *ctx, SWspan *span, GLboolean xMajor ) { - GLint width, start; + const GLint width = (GLint) CLAMP(ctx->Line.Width, + ctx->Const.MinLineWidth, + ctx->Const.MaxLineWidth); + GLint start; ASSERT(span->end < MAX_WIDTH); - width = (GLint) CLAMP( ctx->Line._Width, MIN_LINE_WIDTH, MAX_LINE_WIDTH ); - if (width & 1) start = width / 2; else @@ -143,7 +144,7 @@ draw_wide_line( GLcontext *ctx, SWspan *span, GLboolean xMajor ) span.arrayMask |= SPAN_MASK; \ compute_stipple_mask(ctx, span.end, span.array->mask); \ } \ - if (ctx->Line._Width > 1.0) { \ + if (ctx->Line.Width > 1.0) { \ draw_wide_line(ctx, &span, (GLboolean)(dx > dy)); \ } \ else { \ @@ -161,7 +162,7 @@ draw_wide_line( GLcontext *ctx, SWspan *span, GLboolean xMajor ) span.arrayMask |= SPAN_MASK; \ compute_stipple_mask(ctx, span.end, span.array->mask); \ } \ - if (ctx->Line._Width > 1.0) { \ + if (ctx->Line.Width > 1.0) { \ draw_wide_line(ctx, &span, (GLboolean)(dx > dy)); \ } \ else { \ @@ -180,7 +181,7 @@ draw_wide_line( GLcontext *ctx, SWspan *span, GLboolean xMajor ) span.arrayMask |= SPAN_MASK; \ compute_stipple_mask(ctx, span.end, span.array->mask); \ } \ - if (ctx->Line._Width > 1.0) { \ + if (ctx->Line.Width > 1.0) { \ draw_wide_line(ctx, &span, (GLboolean)(dx > dy)); \ } \ else { \ @@ -274,16 +275,21 @@ _swrast_choose_line( GLcontext *ctx ) USE(general_line); } else if (ctx->Depth.Test - || ctx->Line._Width != 1.0 + || ctx->Line.Width != 1.0 || ctx->Line.StippleFlag) { /* no texture, but Z, fog, width>1, stipple, etc. */ if (rgbmode) +#if CHAN_BITS == 32 + USE(general_line); +#else USE(rgba_line); +#endif else USE(ci_line); } else { ASSERT(!ctx->Depth.Test); + ASSERT(ctx->Line.Width == 1.0); /* simple lines */ if (rgbmode) USE(simple_no_z_rgba_line); diff --git a/src/mesa/swrast/s_points.c b/src/mesa/swrast/s_points.c index 8eba53c807..4768fbea97 100644 --- a/src/mesa/swrast/s_points.c +++ b/src/mesa/swrast/s_points.c @@ -76,7 +76,7 @@ sprite_point(GLcontext *ctx, const SWvertex *vert) } else { /* use constant point size */ - size = ctx->Point._Size; /* already clamped to user range */ + size = ctx->Point.Size; } /* clamp to non-AA implementation limits */ size = CLAMP(size, ctx->Const.MinPointSize, ctx->Const.MaxPointSize); @@ -227,7 +227,7 @@ smooth_point(GLcontext *ctx, const SWvertex *vert) } else { /* use constant point size */ - size = ctx->Point._Size; /* this is already clamped */ + size = ctx->Point.Size; } /* clamp to AA implementation limits */ size = CLAMP(size, ctx->Const.MinPointSizeAA, ctx->Const.MaxPointSizeAA); @@ -361,7 +361,7 @@ large_point(GLcontext *ctx, const SWvertex *vert) } else { /* use constant point size */ - size = ctx->Point._Size; /* already clamped to user range */ + size = ctx->Point.Size; } /* clamp to non-AA implementation limits */ size = CLAMP(size, ctx->Const.MinPointSize, ctx->Const.MaxPointSize); @@ -550,7 +550,7 @@ _swrast_choose_point(GLcontext *ctx) else if (ctx->Point.SmoothFlag) { swrast->Point = smooth_point; } - else if (ctx->Point._Size > 1.0 || + else if (ctx->Point.Size > 1.0 || ctx->Point._Attenuated || ctx->VertexProgram.PointSizeEnabled) { swrast->Point = large_point; diff --git a/src/mesa/swrast/s_span.c b/src/mesa/swrast/s_span.c index f23272c2be..cfc65bee87 100644 --- a/src/mesa/swrast/s_span.c +++ b/src/mesa/swrast/s_span.c @@ -1358,7 +1358,7 @@ _swrast_write_rgba_span( GLcontext *ctx, SWspan *span) #if CHAN_BITS == 32 if ((span->arrayAttribs & FRAG_BIT_COL0) == 0) { - interpolate_int_colors(ctx, span); + interpolate_active_attribs(ctx, span, FRAG_BIT_COL0); } #else if ((span->arrayMask & SPAN_RGBA) == 0) { diff --git a/src/mesa/tnl/t_vertex.c b/src/mesa/tnl/t_vertex.c index 6aae602037..a6728c318f 100644 --- a/src/mesa/tnl/t_vertex.c +++ b/src/mesa/tnl/t_vertex.c @@ -233,7 +233,7 @@ void _tnl_get_attr( GLcontext *ctx, const void *vin, /* If the hardware vertex doesn't have point size then use size from * GLcontext. XXX this will be wrong if drawing attenuated points! */ - dest[0] = ctx->Point._Size; + dest[0] = ctx->Point.Size; } else { _mesa_memcpy( dest, ctx->Current.Attrib[attr], 4*sizeof(GLfloat)); diff --git a/src/mesa/tnl_dd/t_dd_vb.c b/src/mesa/tnl_dd/t_dd_vb.c index ab3bb37631..3cedd90119 100644 --- a/src/mesa/tnl_dd/t_dd_vb.c +++ b/src/mesa/tnl_dd/t_dd_vb.c @@ -184,7 +184,7 @@ void TAG(translate_vertex)(GLcontext *ctx, } } - dst->pointSize = ctx->Point._Size; + dst->pointSize = ctx->Point.Size; } |