X-Git-Url: https://git.libre-soc.org/?a=blobdiff_plain;f=src%2Fmesa%2Fdrivers%2Fdri%2Fr200%2Fr200_tcl.c;h=2ad35d4390647ca8ea62e978bc88bc7080c81a5a;hb=584b84256b07e106cd7295495355eb21226465d7;hp=1090666dd3f40fb31f9db8cfa1826c5dd3738e79;hpb=aaaf38d37215aa1536ff34ec370ecfc04111af22;p=mesa.git diff --git a/src/mesa/drivers/dri/r200/r200_tcl.c b/src/mesa/drivers/dri/r200/r200_tcl.c index 1090666dd3f..2ad35d43906 100644 --- a/src/mesa/drivers/dri/r200/r200_tcl.c +++ b/src/mesa/drivers/dri/r200/r200_tcl.c @@ -1,4 +1,3 @@ -/* $XFree86: xc/lib/GL/mesa/src/drv/r200/r200_tcl.c,v 1.2 2002/12/16 16:18:55 dawes Exp $ */ /* Copyright (C) The Weather Channel, Inc. 2002. All Rights Reserved. @@ -40,7 +39,7 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. #include "colormac.h" #include "light.h" -#include "array_cache/acache.h" +#include "vbo/vbo.h" #include "tnl/tnl.h" #include "tnl/t_pipeline.h" @@ -62,13 +61,16 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. #define HAVE_TRI_STRIPS 1 #define HAVE_TRI_STRIP_1 0 #define HAVE_TRI_FANS 1 -#define HAVE_QUADS 0 /* hw quad verts in wrong order??? */ +#define HAVE_QUADS 1 #define HAVE_QUAD_STRIPS 1 #define HAVE_POLYGONS 1 #define HAVE_ELTS 1 -#define HW_POINTS R200_VF_PRIM_POINTS +#define HW_POINTS ((ctx->Point.PointSprite || \ + ((ctx->_TriangleCaps & (DD_POINT_SIZE | DD_POINT_ATTEN)) && \ + !(ctx->_TriangleCaps & (DD_POINT_SMOOTH)))) ? \ + R200_VF_PRIM_POINT_SPRITES : R200_VF_PRIM_POINTS) #define HW_LINES R200_VF_PRIM_LINES #define HW_LINE_LOOP 0 #define HW_LINE_STRIP R200_VF_PRIM_LINE_STRIP @@ -140,33 +142,55 @@ static GLboolean discrete_prim[0x10] = { static GLushort *r200AllocElts( r200ContextPtr rmesa, GLuint nr ) { - if (rmesa->dma.flush) - rmesa->dma.flush( rmesa ); + if (rmesa->dma.flush == r200FlushElts && + rmesa->store.cmd_used + nr*2 < R200_CMD_BUF_SZ) { - r200EmitAOS( rmesa, - rmesa->tcl.aos_components, - rmesa->tcl.nr_aos_components, 0 ); + GLushort *dest = (GLushort *)(rmesa->store.cmd_buf + + rmesa->store.cmd_used); + + rmesa->store.cmd_used += nr*2; + + return dest; + } + else { + if (rmesa->dma.flush) + rmesa->dma.flush( rmesa ); + + r200EnsureCmdBufSpace( rmesa, AOS_BUFSZ(rmesa->tcl.nr_aos_components) + + rmesa->hw.max_state_size + ELTS_BUFSZ(nr) ); + + r200EmitAOS( rmesa, + rmesa->tcl.aos_components, + rmesa->tcl.nr_aos_components, 0 ); - return r200AllocEltsOpenEnded( rmesa, rmesa->tcl.hw_primitive, nr ); + return r200AllocEltsOpenEnded( rmesa, rmesa->tcl.hw_primitive, nr ); + } } -#define CLOSE_ELTS() R200_NEWPRIM( rmesa ) +#define CLOSE_ELTS() \ +do { \ + if (0) R200_NEWPRIM( rmesa ); \ +} \ +while (0) /* TODO: Try to extend existing primitive if both are identical, * discrete and there are no intervening state changes. (Somewhat * duplicates changes to DrawArrays code) */ -static void EMIT_PRIM( GLcontext *ctx, - GLenum prim, - GLuint hwprim, - GLuint start, - GLuint count) +static void r200EmitPrim( GLcontext *ctx, + GLenum prim, + GLuint hwprim, + GLuint start, + GLuint count) { r200ContextPtr rmesa = R200_CONTEXT( ctx ); r200TclPrimitive( ctx, prim, hwprim ); + r200EnsureCmdBufSpace( rmesa, AOS_BUFSZ(rmesa->tcl.nr_aos_components) + + rmesa->hw.max_state_size + VBUF_BUFSZ ); + r200EmitAOS( rmesa, rmesa->tcl.aos_components, rmesa->tcl.nr_aos_components, @@ -179,7 +203,9 @@ static void EMIT_PRIM( GLcontext *ctx, count - start ); } - +#define EMIT_PRIM(ctx, prim, hwprim, start, count) do { \ + r200EmitPrim( ctx, prim, hwprim, start, count ); \ + (void) rmesa; } while (0) /* Try & join small primitives */ @@ -196,12 +222,15 @@ static void EMIT_PRIM( GLcontext *ctx, #ifdef MESA_BIG_ENDIAN /* We could do without (most of) this ugliness if dest was always 32 bit word aligned... */ -#define EMIT_ELT(dest, offset, x) do { \ +#define EMIT_ELT(dest, offset, x) do { \ int off = offset + ( ( (GLuint)dest & 0x2 ) >> 1 ); \ GLushort *des = (GLushort *)( (GLuint)dest & ~0x2 ); \ - (des)[ off + 1 - 2 * ( off & 1 ) ] = (GLushort)(x); } while (0) + (des)[ off + 1 - 2 * ( off & 1 ) ] = (GLushort)(x); \ + (void)rmesa; } while (0) #else -#define EMIT_ELT(dest, offset, x) (dest)[offset] = (GLushort) (x) +#define EMIT_ELT(dest, offset, x) do { \ + (dest)[offset] = (GLushort) (x); \ + (void)rmesa; } while (0) #endif #define EMIT_TWO_ELTS(dest, offset, x, y) *(GLuint *)((dest)+offset) = ((y)<<16)|(x); @@ -240,12 +269,107 @@ void r200TclPrimitive( GLcontext *ctx, if (newprim != rmesa->tcl.hw_primitive || !discrete_prim[hw_prim&0xf]) { + /* need to disable perspective-correct texturing for point sprites */ + if ((prim & PRIM_MODE_MASK) == GL_POINTS && ctx->Point.PointSprite) { + if (rmesa->hw.set.cmd[SET_RE_CNTL] & R200_PERSPECTIVE_ENABLE) { + R200_STATECHANGE( rmesa, set ); + rmesa->hw.set.cmd[SET_RE_CNTL] &= ~R200_PERSPECTIVE_ENABLE; + } + } + else if (!(rmesa->hw.set.cmd[SET_RE_CNTL] & R200_PERSPECTIVE_ENABLE)) { + R200_STATECHANGE( rmesa, set ); + rmesa->hw.set.cmd[SET_RE_CNTL] |= R200_PERSPECTIVE_ENABLE; + } R200_NEWPRIM( rmesa ); rmesa->tcl.hw_primitive = newprim; } } +/**********************************************************************/ +/* Fog blend factor computation for hw tcl */ +/* same calculation used as in t_vb_fog.c */ +/**********************************************************************/ + +#define FOG_EXP_TABLE_SIZE 256 +#define FOG_MAX (10.0) +#define EXP_FOG_MAX .0006595 +#define FOG_INCR (FOG_MAX/FOG_EXP_TABLE_SIZE) +static GLfloat exp_table[FOG_EXP_TABLE_SIZE]; + +#if 1 +#define NEG_EXP( result, narg ) \ +do { \ + GLfloat f = (GLfloat) (narg * (1.0/FOG_INCR)); \ + GLint k = (GLint) f; \ + if (k > FOG_EXP_TABLE_SIZE-2) \ + result = (GLfloat) EXP_FOG_MAX; \ + else \ + result = exp_table[k] + (f-k)*(exp_table[k+1]-exp_table[k]); \ +} while (0) +#else +#define NEG_EXP( result, narg ) \ +do { \ + result = exp(-narg); \ +} while (0) +#endif + + +/** + * Initialize the exp_table[] lookup table for approximating exp(). + */ +void +r200InitStaticFogData( void ) +{ + GLfloat f = 0.0F; + GLint i = 0; + for ( ; i < FOG_EXP_TABLE_SIZE ; i++, f += FOG_INCR) { + exp_table[i] = (GLfloat) exp(-f); + } +} + + +/** + * Compute per-vertex fog blend factors from fog coordinates by + * evaluating the GL_LINEAR, GL_EXP or GL_EXP2 fog function. + * Fog coordinates are distances from the eye (typically between the + * near and far clip plane distances). + * Note the fog (eye Z) coords may be negative so we use ABS(z) below. + * Fog blend factors are in the range [0,1]. + */ +float +r200ComputeFogBlendFactor( GLcontext *ctx, GLfloat fogcoord ) +{ + GLfloat end = ctx->Fog.End; + GLfloat d, temp; + const GLfloat z = FABSF(fogcoord); + + switch (ctx->Fog.Mode) { + case GL_LINEAR: + if (ctx->Fog.Start == ctx->Fog.End) + d = 1.0F; + else + d = 1.0F / (ctx->Fog.End - ctx->Fog.Start); + temp = (end - z) * d; + return CLAMP(temp, 0.0F, 1.0F); + break; + case GL_EXP: + d = ctx->Fog.Density; + NEG_EXP( temp, d * z ); + return temp; + break; + case GL_EXP2: + d = ctx->Fog.Density*ctx->Fog.Density; + NEG_EXP( temp, d * z * z ); + return temp; + break; + default: + _mesa_problem(ctx, "Bad fog mode in make_fog_coord"); + return 0; + } +} + + /**********************************************************************/ /* Render pipeline stage */ /**********************************************************************/ @@ -260,6 +384,13 @@ static GLboolean r200_run_tcl_render( GLcontext *ctx, TNLcontext *tnl = TNL_CONTEXT(ctx); struct vertex_buffer *VB = &tnl->vb; GLuint i; + GLubyte *vimap_rev; +/* use hw fixed order for simplicity, pos 0, weight 1, normal 2, fog 3, + color0 - color3 4-7, texcoord0 - texcoord5 8-13, pos 1 14. Must not use + more than 12 of those at the same time. */ + GLubyte map_rev_fixed[15] = {255, 255, 255, 255, 255, 255, 255, 255, + 255, 255, 255, 255, 255, 255, 255}; + /* TODO: separate this from the swtnl pipeline */ @@ -272,8 +403,86 @@ static GLboolean r200_run_tcl_render( GLcontext *ctx, if (VB->Count == 0) return GL_FALSE; + /* Validate state: + */ + if (rmesa->NewGLState) + r200ValidateState( ctx ); + + if (!ctx->VertexProgram._Enabled) { + /* NOTE: inputs != tnl->render_inputs - these are the untransformed + * inputs. + */ + map_rev_fixed[0] = VERT_ATTRIB_POS; + /* technically there is no reason we always need VA_COLOR0. In theory + could disable it depending on lighting, color materials, texturing... */ + map_rev_fixed[4] = VERT_ATTRIB_COLOR0; + + if (ctx->Light.Enabled) { + map_rev_fixed[2] = VERT_ATTRIB_NORMAL; + } + + /* this also enables VA_COLOR1 when using separate specular + lighting model, which is unnecessary. + FIXME: OTOH, we're missing the case where a ATI_fragment_shader accesses + the secondary color (if lighting is disabled). The chip seems + misconfigured for that though elsewhere (tcl output, might lock up) */ + if (ctx->_TriangleCaps & DD_SEPARATE_SPECULAR) { + map_rev_fixed[5] = VERT_ATTRIB_COLOR1; + } + + if ( (ctx->Fog.FogCoordinateSource == GL_FOG_COORD) && ctx->Fog.Enabled ) { + map_rev_fixed[3] = VERT_ATTRIB_FOG; + } + + for (i = 0 ; i < ctx->Const.MaxTextureUnits; i++) { + if (ctx->Texture.Unit[i]._ReallyEnabled) { + if (rmesa->TexGenNeedNormals[i]) { + map_rev_fixed[2] = VERT_ATTRIB_NORMAL; + } + map_rev_fixed[8 + i] = VERT_ATTRIB_TEX0 + i; + } + } + vimap_rev = &map_rev_fixed[0]; + } + else { + /* vtx_tcl_output_vtxfmt_0/1 need to match configuration of "fragment + part", since using some vertex interpolator later which is not in + out_vtxfmt0/1 will lock up. It seems to be ok to write in vertex + prog to a not enabled output however, so just don't mess with it. + We only need to change compsel. */ + GLuint out_compsel = 0; + GLuint vp_out = rmesa->curr_vp_hw->mesa_program.Base.OutputsWritten; + + vimap_rev = &rmesa->curr_vp_hw->inputmap_rev[0]; + assert(vp_out & (1 << VERT_RESULT_HPOS)); + out_compsel = R200_OUTPUT_XYZW; + if (vp_out & (1 << VERT_RESULT_COL0)) { + out_compsel |= R200_OUTPUT_COLOR_0; + } + if (vp_out & (1 << VERT_RESULT_COL1)) { + out_compsel |= R200_OUTPUT_COLOR_1; + } + if (vp_out & (1 << VERT_RESULT_FOGC)) { + out_compsel |= R200_OUTPUT_DISCRETE_FOG; + } + if (vp_out & (1 << VERT_RESULT_PSIZ)) { + out_compsel |= R200_OUTPUT_PT_SIZE; + } + for (i = VERT_RESULT_TEX0; i < VERT_RESULT_TEX6; i++) { + if (vp_out & (1 << i)) { + out_compsel |= R200_OUTPUT_TEX_0 << (i - VERT_RESULT_TEX0); + } + } + if (rmesa->hw.vtx.cmd[VTX_TCL_OUTPUT_COMPSEL] != out_compsel) { + R200_STATECHANGE( rmesa, vtx ); + rmesa->hw.vtx.cmd[VTX_TCL_OUTPUT_COMPSEL] = out_compsel; + } + } + + /* Do the actual work: + */ r200ReleaseArrays( ctx, ~0 /* stage->changed_inputs */ ); - r200EmitArrays( ctx, stage->inputs ); + r200EmitArrays( ctx, vimap_rev ); rmesa->tcl.Elts = VB->Elts; @@ -297,84 +506,15 @@ static GLboolean r200_run_tcl_render( GLcontext *ctx, -static void r200_check_tcl_render( GLcontext *ctx, - struct tnl_pipeline_stage *stage ) -{ - r200ContextPtr rmesa = R200_CONTEXT(ctx); - GLuint inputs = VERT_BIT_POS; - GLuint unit; - - /* Validate state: - */ - if (rmesa->NewGLState) - r200ValidateState( ctx ); - - if (ctx->RenderMode == GL_RENDER) { - /* Make all this event-driven: - */ - if (ctx->Light.Enabled) { - inputs |= VERT_BIT_NORMAL; - - if (1 || ctx->Light.ColorMaterialEnabled) { - inputs |= VERT_BIT_COLOR0; - } - } - else { - inputs |= VERT_BIT_COLOR0; - - if (ctx->_TriangleCaps & DD_SEPARATE_SPECULAR) { - inputs |= VERT_BIT_COLOR1; - } - } - - for (unit = 0 ; unit < ctx->Const.MaxTextureUnits; unit++) { - if (ctx->Texture.Unit[unit]._ReallyEnabled) { - if (ctx->Texture.Unit[unit].TexGenEnabled) { - if (rmesa->TexGenNeedNormals[unit]) { - inputs |= VERT_BIT_NORMAL; - } - } else { - inputs |= VERT_BIT_TEX(unit); - } - } - } - - stage->inputs = inputs; - stage->active = 1; - } - else - stage->active = 0; -} - -static void r200_init_tcl_render( GLcontext *ctx, - struct tnl_pipeline_stage *stage ) -{ - stage->check = r200_check_tcl_render; - stage->check( ctx, stage ); -} - -static void dtr( struct tnl_pipeline_stage *stage ) -{ - (void)stage; -} - - /* Initial state for tcl stage. */ const struct tnl_pipeline_stage _r200_tcl_stage = { "r200 render", - (_DD_NEW_SEPARATE_SPECULAR | - _NEW_LIGHT| - _NEW_TEXTURE| - _NEW_FOG| - _NEW_RENDERMODE), /* re-check (new inputs) */ - 0, /* re-run (always runs) */ - GL_TRUE, /* active */ - 0, 0, /* inputs (set in check_render), outputs */ - 0, 0, /* changed_inputs, private */ - dtr, /* destructor */ - r200_init_tcl_render, /* check - initially set to alloc data */ + NULL, /* private */ + NULL, + NULL, + NULL, r200_run_tcl_render /* run */ }; @@ -411,18 +551,9 @@ static void transition_to_swtnl( GLcontext *ctx ) * need to put the card into D3D mode to make it work: */ R200_STATECHANGE( rmesa, vap ); - rmesa->hw.vap.cmd[VAP_SE_VAP_CNTL] &= ~R200_VAP_TCL_ENABLE; - rmesa->hw.vap.cmd[VAP_SE_VAP_CNTL] |= R200_VAP_D3D_TEX_DEFAULT; - - R200_STATECHANGE( rmesa, vte ); - rmesa->hw.vte.cmd[VTE_SE_VTE_CNTL] &= ~R200_VTX_W0_FMT; - - R200_STATECHANGE( rmesa, set ); - rmesa->hw.set.cmd[SET_RE_CNTL] |= (R200_VTX_STQ0_D3D | - R200_VTX_STQ1_D3D); + rmesa->hw.vap.cmd[VAP_SE_VAP_CNTL] &= ~(R200_VAP_TCL_ENABLE|R200_VAP_PROG_VTX_SHADER_ENABLE); } - static void transition_to_hwtnl( GLcontext *ctx ) { r200ContextPtr rmesa = R200_CONTEXT(ctx); @@ -437,7 +568,7 @@ static void transition_to_hwtnl( GLcontext *ctx ) if ( rmesa->dma.flush ) rmesa->dma.flush( rmesa ); - rmesa->dma.flush = 0; + rmesa->dma.flush = NULL; if (rmesa->swtcl.indexed_verts.buf) r200ReleaseDmaRegion( rmesa, &rmesa->swtcl.indexed_verts, @@ -445,18 +576,24 @@ static void transition_to_hwtnl( GLcontext *ctx ) R200_STATECHANGE( rmesa, vap ); rmesa->hw.vap.cmd[VAP_SE_VAP_CNTL] |= R200_VAP_TCL_ENABLE; - rmesa->hw.vap.cmd[VAP_SE_VAP_CNTL] &= ~(R200_VAP_FORCE_W_TO_ONE | - R200_VAP_D3D_TEX_DEFAULT); + rmesa->hw.vap.cmd[VAP_SE_VAP_CNTL] &= ~R200_VAP_FORCE_W_TO_ONE; + + if (ctx->VertexProgram._Enabled) { + rmesa->hw.vap.cmd[VAP_SE_VAP_CNTL] |= R200_VAP_PROG_VTX_SHADER_ENABLE; + } + + if ( ((rmesa->hw.ctx.cmd[CTX_PP_FOG_COLOR] & R200_FOG_USE_MASK) + == R200_FOG_USE_SPEC_ALPHA) && + (ctx->Fog.FogCoordinateSource == GL_FOG_COORD )) { + R200_STATECHANGE( rmesa, ctx ); + rmesa->hw.ctx.cmd[CTX_PP_FOG_COLOR] &= ~R200_FOG_USE_MASK; + rmesa->hw.ctx.cmd[CTX_PP_FOG_COLOR] |= R200_FOG_USE_VTX_FOG; + } R200_STATECHANGE( rmesa, vte ); rmesa->hw.vte.cmd[VTE_SE_VTE_CNTL] &= ~(R200_VTX_XY_FMT|R200_VTX_Z_FMT); rmesa->hw.vte.cmd[VTE_SE_VTE_CNTL] |= R200_VTX_W0_FMT; - R200_STATECHANGE( rmesa, set ); - rmesa->hw.set.cmd[SET_RE_CNTL] &= ~(R200_VTX_STQ0_D3D | - R200_VTX_STQ1_D3D); - - if (R200_DEBUG & DEBUG_FALLBACKS) fprintf(stderr, "R200 end tcl fallback\n"); } @@ -470,7 +607,12 @@ static char *fallbackStrings[] = { "Texgen unit 0", "Texgen unit 1", "Texgen unit 2", - "User disable" + "Texgen unit 3", + "Texgen unit 4", + "Texgen unit 5", + "User disable", + "Bitmap as points", + "Vertex program" };