kolibrios/programs/develop/libraries/TinyGL/asm_fork/vertex.asm
IgorA ea910a143c 1) fix and optimizations
2) add quadric object 'Sphere'

git-svn-id: svn://kolibrios.org@5218 a494cfbc-eb01-0410-851d-a64ba20cac60
2014-12-05 21:12:56 +00:00

694 lines
18 KiB
NASM

align 4
proc glopNormal uses ecx esi edi, context:dword, p:dword
mov esi,[p]
add esi,4
mov edi,[context]
add edi,offs_cont_current_normal
mov ecx,3
rep movsd
mov dword[edi],0.0 ;context.current_normal.W = 0.0
ret
endp
align 4
proc glopTexCoord uses ecx esi edi, context:dword, p:dword
mov esi,[p]
add esi,4
mov edi,[context]
add edi,offs_cont_current_tex_coord
mov ecx,4
rep movsd
ret
endp
align 4
proc glopEdgeFlag uses eax ebx, context:dword, p:dword
mov eax,[context]
mov ebx,[p]
mov ebx,[ebx+4] ;ebx = p[1]
mov dword[eax+offs_cont_current_edge_flag],ebx
ret
endp
align 4
proc glopColor uses eax ecx esi edi, context:dword, p:dword
locals
q rd 7
endl
;current_color[3] = p[1-4]
;longcurrent_color[2] = p[5-7]
mov esi,[p]
add esi,4
mov edi,[context]
add edi,offs_cont_current_color
mov ecx,7
rep movsd
mov eax,[context]
cmp dword[eax+offs_cont_color_material_enabled],0
je @f
mov dword[q],OP_Material
mov ecx,[eax+offs_cont_current_color_material_mode]
mov dword[q+4],ecx
mov ecx,[eax+offs_cont_current_color_material_type]
mov dword[q+8],ecx
mov esi,[p]
add esi,4
mov edi,ebp
sub edi,16 ;edi = &q[3]
mov ecx,4
rep movsd
stdcall glopMaterial, eax,[q]
@@:
ret
endp
align 4
proc gl_eval_viewport uses eax, context:dword
locals
zsize dd ? ;float
endl
mov eax,[context]
add eax,offs_cont_viewport ;eax = (GLViewport*) v
mov dword[zsize],(1 shl (ZB_Z_BITS + ZB_POINT_Z_FRAC_BITS))
fild dword[zsize]
fstp dword[zsize]
fld1
fld1
fadd st1,st0 ;st1 = 2.0
fdiv st0,st1 ;st0 = 0.5
fild dword[eax+offs_vpor_xsize]
fsub st0,st1
fdiv st0,st2
fst dword[eax+offs_vpor_scale+offs_X]
fiadd dword[eax+offs_vpor_xmin]
fstp dword[eax+offs_vpor_trans+offs_X]
fild dword[eax+offs_vpor_ysize]
fsub st0,st1
fdiv st0,st2
fchs
fst dword[eax+offs_vpor_scale+offs_Y]
fchs
fiadd dword[eax+offs_vpor_ymin]
fstp dword[eax+offs_vpor_trans+offs_Y]
fld dword[zsize]
fsub st0,st1
fdiv st0,st2
fchs
fst dword[eax+offs_vpor_scale+offs_Z]
fchs
mov dword[zsize],(1 shl ZB_POINT_Z_FRAC_BITS) / 2
fiadd dword[zsize]
fstp dword[eax+offs_vpor_trans+offs_Z]
if DEBUG ;gl_eval_viewport
stdcall dbg_print,f_ev,txt_nl
add eax,offs_vpor_scale
stdcall gl_print_matrix,eax,1
add eax,8
stdcall gl_print_matrix,eax,1
end if
ret
endp
align 4
proc glopBegin uses eax ebx ecx edx, context:dword, p:dword
locals
tmp M4
endl
; assert(c->in_begin == 0);
mov edx,[context]
mov ebx,[p]
mov ebx,[ebx+4] ;ebx = p[1]
mov [edx+offs_cont_begin_type],ebx
mov dword[edx+offs_cont_in_begin],1
mov dword[edx+offs_cont_vertex_n],0
mov dword[edx+offs_cont_vertex_cnt],0
bt dword[edx+offs_cont_matrix_model_projection_updated],0
jnc .end_mmpu
cmp dword[edx+offs_cont_lighting_enabled],0 ;if(context.lighting_enabled)
je @f
; precompute inverse modelview
mov ebx,ebp
sub ebx,64
stdcall gl_M4_Inv, ebx, edx+offs_cont_matrix_stack_ptr
stdcall gl_M4_Transpose, edx+offs_cont_matrix_model_view_inv, ebx
jmp .end_if_0
@@:
mov ecx,edx
add ecx,offs_cont_matrix_model_projection
; precompute projection matrix
stdcall gl_M4_Mul, ecx,dword[edx+offs_cont_matrix_stack_ptr+4],dword[edx+offs_cont_matrix_stack_ptr]
; test to accelerate computation
mov dword[edx+offs_cont_matrix_model_projection_no_w_transform],0
fldz
fld dword[ecx+12*4]
fcomp st1
fstsw ax
sahf
jne .end_if_0
fld dword[ecx+13*4]
fcomp st1
fstsw ax
sahf
jne .end_if_0
fld dword[ecx+14*4]
fcomp st1
fstsw ax
sahf
jne .end_if_0
mov dword[edx+offs_cont_matrix_model_projection_no_w_transform],1
.end_if_0:
; test if the texture matrix is not Identity
mov eax,edx
add eax,offs_cont_matrix_stack_ptr+8
stdcall gl_M4_IsId,eax
xor eax,1
mov dword[edx+offs_cont_apply_texture_matrix],eax
mov dword[edx+offs_cont_matrix_model_projection_updated],0
.end_mmpu:
; viewport
cmp dword[edx+offs_cont_viewport+offs_vpor_updated],0 ;if (context.viewport.updated)
je @f
stdcall gl_eval_viewport,edx
mov dword[edx+offs_cont_viewport+offs_vpor_updated],0
@@:
; triangle drawing functions
cmp dword[edx+offs_cont_render_mode],GL_SELECT
jne @f
mov dword[edx+offs_cont_draw_triangle_front],gl_draw_triangle_select
mov dword[edx+offs_cont_draw_triangle_back],gl_draw_triangle_select
jmp .end_if_2
@@:
cmp dword[edx+offs_cont_polygon_mode_front],GL_POINT
jne @f
mov dword[edx+offs_cont_draw_triangle_front],gl_draw_triangle_point
jmp .end_if_1
@@:
cmp dword[edx+offs_cont_polygon_mode_front],GL_LINE
jne @f
mov dword[edx+offs_cont_draw_triangle_front],gl_draw_triangle_line
jmp .end_if_1
@@: ;default:
mov dword[edx+offs_cont_draw_triangle_front],gl_draw_triangle_fill
.end_if_1:
cmp dword[edx+offs_cont_polygon_mode_back],GL_POINT
jne @f
mov dword[edx+offs_cont_draw_triangle_back],gl_draw_triangle_point
jmp .end_if_2
@@:
cmp dword[edx+offs_cont_polygon_mode_back],GL_LINE
jne @f
mov dword[edx+offs_cont_draw_triangle_back],gl_draw_triangle_line
jmp .end_if_2
@@: ;default:
mov dword[edx+offs_cont_draw_triangle_back],gl_draw_triangle_fill
.end_if_2:
ret
endp
; coords, tranformation , clip code and projection
; TODO : handle all cases
align 4
proc gl_vertex_transform, context:dword, v:dword
pushad
mov eax,[context]
mov edx,[v]
cmp dword[eax+offs_cont_lighting_enabled],0 ;if (context.lighting_enabled)
je .els_0
; eye coordinates needed for lighting
mov ebx,dword[eax+offs_cont_matrix_stack_ptr]
finit
fld dword[edx+offs_vert_coord+offs_X]
fld dword[edx+offs_vert_coord+offs_Y]
fld dword[edx+offs_vert_coord+offs_Z]
mov ecx,4
.cycle_0:
fld dword[ebx] ;st0 = m[0]
fmul st0,st3 ;st0 *= v.coord.X
fld dword[ebx+4] ;st0 = m[1]
fmul st0,st3 ;st0 *= v.coord.Y
fld dword[ebx+8] ;st0 = m[2]
fmul st0,st3 ;st0 *= v.coord.Z
fadd dword[ebx+12] ;st0 += m[3]
faddp ;st0 += v.coord.Z * m[2]
faddp ;st0 += v.coord.Y * m[1]
fstp dword[edx+offs_vert_ec] ;v.ec.X = v.coord.X * m[0] + v.coord.Y * m[1] + v.coord.Z * m[2] + m[3]
add ebx,16 ;следущая строка матрицы
add edx,4 ;следущая координата вектора
loop .cycle_0
ffree st0
fincstp
ffree st0
fincstp
ffree st0
fincstp
; projection coordinates
mov ebx,dword[eax+offs_cont_matrix_stack_ptr+4]
mov edx,[v]
fld dword[edx+offs_vert_ec+offs_X]
fld dword[edx+offs_vert_ec+offs_Y]
fld dword[edx+offs_vert_ec+offs_Z]
mov ecx,4
.cycle_1:
fld dword[ebx] ;st0 = m[0]
fmul st0,st3 ;st0 *= v.ec.X
fld dword[ebx+4] ;st0 = m[1]
fmul st0,st3 ;st0 *= v.ec.Y
fld dword[ebx+8] ;st0 = m[2]
fmul st0,st3 ;st0 *= v.ec.Z
fadd dword[ebx+12] ;st0 += m[3]
faddp ;st0 += v.ec.Z * m[2]
faddp ;st0 += v.ec.Y * m[1]
fstp dword[edx+offs_vert_pc] ;v.pc.X = v.ec.X * m[0] + v.ec.Y * m[1] + v.ec.Z * m[2] + m[3]
add ebx,16 ;следущая строка матрицы
add edx,4 ;следущая координата вектора
loop .cycle_1
ffree st0
fincstp
ffree st0
fincstp
ffree st0
fincstp
mov ebx,eax
add ebx,offs_cont_matrix_model_view_inv
mov edi,eax
add edi,offs_cont_current_normal
mov edx,[v]
fld dword[edi+offs_X]
fld dword[edi+offs_Y]
fld dword[edi+offs_Z]
mov ecx,3
.cycle_2:
fld dword[ebx] ;st0 = m[0]
fmul st0,st3 ;st0 *= n.X
fld dword[ebx+4] ;st0 = m[1]
fmul st0,st3 ;st0 *= n.Y
fld dword[ebx+8] ;st0 = m[2]
fmul st0,st3 ;st0 *= n.Z
faddp ;st0 += n.Z * m[2]
faddp ;st0 += n.Y * m[1]
fstp dword[edx+offs_vert_normal] ;v.normal.X = n.X * m[0] + n.Y * m[1] + n.Z * m[2]
add ebx,16 ;следущая строка матрицы
add edx,4 ;следущая координата вектора
loop .cycle_2
cmp dword[eax+offs_cont_normalize_enabled],0
je .end_els
mov edx,[v]
add edx,offs_vert_normal
stdcall gl_V3_Norm,edx
jmp .end_els
.els_0:
; no eye coordinates needed, no normal
; NOTE: W = 1 is assumed
mov ebx,eax
add ebx,offs_cont_matrix_model_projection
finit
fld dword[edx+offs_vert_coord+offs_X]
fld dword[edx+offs_vert_coord+offs_Y]
fld dword[edx+offs_vert_coord+offs_Z]
mov esi,edx
add esi,offs_vert_pc
fld dword[ebx] ;st0 = m[0]
fmul st0,st3 ;st0 *= v.coord.X
fld dword[ebx+4] ;st0 = m[1]
fmul st0,st3 ;st0 *= v.coord.Y
fld dword[ebx+8] ;st0 = m[2]
fmul st0,st3 ;st0 *= v.coord.Z
fadd dword[ebx+12] ;st0 += m[3]
faddp ;st0 += v.coord.Z * m[2]
faddp ;st0 += v.coord.Y * m[1]
fstp dword[esi] ;v.pc.X = v.coord.X * m[0] + v.coord.Y * m[1] + v.coord.Z * m[2] + m[3]
fld dword[ebx+16] ;st0 = m[4]
fmul st0,st3 ;st0 *= v.coord.X
fld dword[ebx+20] ;st0 = m[5]
fmul st0,st3 ;st0 *= v.coord.Y
fld dword[ebx+24] ;st0 = m[6]
fmul st0,st3 ;st0 *= v.coord.Z
fadd dword[ebx+28] ;st0 += m[7]
faddp ;st0 += v.coord.Z * m[6]
faddp ;st0 += v.coord.Y * m[5]
fstp dword[esi+4] ;v.pc.X = v.coord.X * m[4] + v.coord.Y * m[5] + v.coord.Z * m[6] + m[7]
fld dword[ebx+32] ;st0 = m[8]
fmul st0,st3 ;st0 *= v.coord.X
fld dword[ebx+36] ;st0 = m[9]
fmul st0,st3 ;st0 *= v.coord.Y
fld dword[ebx+40] ;st0 = m[10]
fmul st0,st3 ;st0 *= v.coord.Z
fadd dword[ebx+44] ;st0 += m[11]
faddp ;st0 += v.coord.Z * m[10]
faddp ;st0 += v.coord.Y * m[9]
fstp dword[esi+8] ;v.pc.X = v.coord.X * m[8] + v.coord.Y * m[9] + v.coord.Z * m[10] + m[11]
cmp dword[eax+offs_cont_matrix_model_projection_no_w_transform],0
je .els_1
;if (context.matrix_model_projection_no_w_transform)
mov ebx,dword[ebx+60] ;ebx = m[15]
mov dword[esi+12],ebx ;v.pc.W = m[15]
jmp .end_els
.els_1:
fld dword[ebx+48] ;st0 = m[12]
fmul st0,st3 ;st0 *= v.coord.X
fld dword[ebx+52] ;st0 = m[13]
fmul st0,st3 ;st0 *= v.coord.Y
fld dword[ebx+56] ;st0 = m[14]
fmul st0,st3 ;st0 *= v.coord.Z
fadd dword[ebx+60] ;st0 += m[15]
faddp ;st0 += v.coord.Z * m[14]
faddp ;st0 += v.coord.Y * m[13]
fstp dword[esi+12] ;v.pc.W = v.coord.X * m[12] + v.coord.Y * m[13] + v.coord.Z * m[14] + m[15]
.end_els:
ffree st0
fincstp
ffree st0
fincstp
ffree st0
fincstp
if DEBUG ;gl_vertex_transform
stdcall dbg_print,f_vt,txt_nl
mov edx,[v]
add edx,offs_vert_pc
stdcall gl_print_matrix,edx,1
end if
mov edx,[v]
stdcall gl_clipcode, dword[edx+offs_vert_pc+offs_X], dword[edx+offs_vert_pc+offs_Y],\
dword[edx+offs_vert_pc+offs_Z], dword[edx+offs_vert_pc+offs_W]
mov dword[edx+offs_vert_clip_code],eax
popad
ret
endp
align 4
proc glopVertex, context:dword, p:dword
locals
;ebx = GLVertex * v
n dd ? ;ebp-4
endl
pushad
mov edx,[context]
; assert(c->in_begin != 0);
mov ecx,[edx+offs_cont_vertex_n]
mov [n],ecx
inc dword[edx+offs_cont_vertex_cnt]
; quick fix to avoid crashes on large polygons
mov ecx,[edx+offs_cont_vertex_max]
cmp dword[n],ecx
jl @f
shl dword[edx+offs_cont_vertex_max],1 ; just double size
imul ecx,2*sizeof.GLVertex
stdcall gl_malloc,ecx
cmp eax,0
jne .no_err
;gl_fatal_error("unable to allocate GLVertex array.\n");
.no_err:
mov edi,eax
mov ebx,eax
mov esi,[edx+offs_cont_vertex]
mov ecx,[n]
imul ecx,(sizeof.GLVertex)/4 ;((...)/4) что-бы использовать movsd вместо movsb
rep movsd
stdcall gl_free,dword[edx+offs_cont_vertex]
mov dword[edx+offs_cont_vertex],ebx
@@:
; new vertex entry
mov ebx,[n]
imul ebx,sizeof.GLVertex
add ebx,[edx+offs_cont_vertex]
inc dword[n]
mov esi,[p]
add esi,4
mov edi,ebx
add edi,offs_vert_coord ;edi = &v.coord
mov ecx,4
rep movsd
stdcall gl_vertex_transform, edx, ebx
; color
cmp dword[edx+offs_cont_lighting_enabled],0
je .els_0
stdcall gl_shade_vertex, edx,ebx
jmp @f
.els_0:
mov eax,[edx+offs_cont_current_color]
mov [ebx+offs_vert_color],eax
mov eax,[edx+offs_cont_current_color+4]
mov [ebx+offs_vert_color+4],eax
mov eax,[edx+offs_cont_current_color+8]
mov [ebx+offs_vert_color+8],eax
@@:
; tex coords
cmp dword[edx+offs_cont_texture_2d_enabled],0
je @f
cmp dword[edx+offs_cont_apply_texture_matrix],0
je .els_1
; gl_M4_MulV4(&v->tex_coord, c->matrix_stack_ptr[2], &c->current_tex_coord);
jmp @f
.els_1:
mov eax,[edx+offs_cont_current_tex_coord]
mov [ebx+offs_vert_tex_coord],eax
@@:
; precompute the mapping to the viewport
cmp dword[ebx+offs_vert_clip_code],0
jne @f
stdcall gl_transform_to_viewport, edx,ebx
@@:
; edge flag
mov eax,[edx+offs_cont_current_edge_flag]
mov dword[ebx+offs_vert_edge_flag],eax ;v.edge_flag = context.current_edge_flag
cmp dword[edx+offs_cont_begin_type],GL_POINTS
jne @f
stdcall gl_draw_point, edx, dword[edx+offs_cont_vertex] ;dword[edx+...] = &context.vertex[0]
mov dword[n],0
jmp .end_f
@@:
cmp dword[edx+offs_cont_begin_type],GL_LINES
jne @f
cmp dword[n],2
jne .end_f
mov eax,[edx+offs_cont_vertex]
push eax
add eax,sizeof.GLVertex
stdcall gl_draw_line, edx, eax
xor eax,eax
mov dword[n],eax
jmp .end_f
@@:
cmp dword[edx+offs_cont_begin_type],GL_LINE_STRIP
je .li_loop
cmp dword[edx+offs_cont_begin_type],GL_LINE_LOOP
jne @f
.li_loop:
cmp dword[n],1
jne .els_2
mov esi,[edx+offs_cont_vertex]
mov edi,esi
add edi,2*sizeof.GLVertex
mov ecx,(sizeof.GLVertex)/4 ;((...)/4) что-бы использовать movsd вместо movsb
rep movsd ;context.vertex[2] = context.vertex[0]
jmp .end_f
.els_2:
cmp dword[n],2
jne .end_f ;else if (n == 2)
mov eax,[edx+offs_cont_vertex]
push eax
mov edi,eax
add eax,sizeof.GLVertex
mov esi,eax
stdcall gl_draw_line, edx, eax
mov ecx,(sizeof.GLVertex)/4 ;((...)/4) что-бы использовать movsd вместо movsb
rep movsd ;context.vertex[0] = context.vertex[1]
mov dword[n],1
jmp .end_f
@@:
cmp dword[edx+offs_cont_begin_type],GL_TRIANGLES
jne @f
cmp dword[n],3
jne .end_f
mov eax,[edx+offs_cont_vertex]
mov [esp-4],eax
add eax,sizeof.GLVertex
mov [esp-8],eax
add eax,sizeof.GLVertex
mov [esp-12],eax
sub esp,12
stdcall gl_draw_triangle, edx
xor eax,eax
mov dword[n],eax
jmp .end_f
@@:
cmp dword[edx+offs_cont_begin_type],GL_TRIANGLE_STRIP
jne @f
; if (c->vertex_cnt >= 3) {
; if (n == 3)
; n = 0;
; /* needed to respect triangle orientation */
; switch(c->vertex_cnt & 1) {
; case 0:
; gl_draw_triangle(c,&c->vertex[2],&c->vertex[1],&c->vertex[0]);
; break;
; default:
; case 1:
; gl_draw_triangle(c,&c->vertex[0],&c->vertex[1],&c->vertex[2]);
; break;
; }
; }
jmp .end_f
@@:
cmp dword[edx+offs_cont_begin_type],GL_TRIANGLE_FAN
jne @f
cmp dword[n],3
jne .end_f
mov eax,[edx+offs_cont_vertex]
mov [esp-4],eax
add eax,sizeof.GLVertex
mov [esp-8],eax
mov edi,eax
add eax,sizeof.GLVertex
mov [esp-12],eax
mov esi,eax
sub esp,12
stdcall gl_draw_triangle, edx
mov ecx,(sizeof.GLVertex)/4 ;((...)/4) что-бы использовать movsd вместо movsb
rep movsd ;context.vertex[1] = context.vertex[2]
mov dword[n],2
jmp .end_f
@@:
cmp dword[edx+offs_cont_begin_type],GL_QUADS
jne @f
cmp dword[n],4
jne .end_f
mov eax,[edx+offs_cont_vertex]
add eax,2*sizeof.GLVertex
mov dword[eax+offs_vert_edge_flag],0 ;context.vertex[2].edge_flag = 0
push eax
sub eax,sizeof.GLVertex
push eax
sub eax,sizeof.GLVertex
stdcall gl_draw_triangle, edx,eax ;v0,v1,v2
mov dword[eax+offs_vert_edge_flag],0 ;context.vertex[0].edge_flag = 0
add eax,2*sizeof.GLVertex
mov dword[eax+offs_vert_edge_flag],1 ;context.vertex[2].edge_flag = 1
add eax,sizeof.GLVertex
push eax
sub eax,sizeof.GLVertex
push eax
sub eax,2*sizeof.GLVertex
stdcall gl_draw_triangle, edx,eax ;v0,v2,v3
xor eax,eax
mov dword[n],eax
jmp .end_f
@@:
cmp dword[edx+offs_cont_begin_type],GL_QUAD_STRIP
jne @f
cmp dword[n],4
jne .end_f
mov eax,[edx+offs_cont_vertex]
mov [esp-4],eax
mov edi,eax
add eax,sizeof.GLVertex
mov [esp-8],eax
add eax,sizeof.GLVertex
mov [esp-12],eax
mov esi,eax
sub esp,12
stdcall gl_draw_triangle, edx ;v0,v1,v2
mov [esp-12],eax
add eax,sizeof.GLVertex
mov [esp-8],eax
sub eax,2*sizeof.GLVertex
mov [esp-4],eax
sub esp,12
stdcall gl_draw_triangle, edx ;v1,v3,v2
mov ecx,(sizeof.GLVertex)/2 ;((...)/2) копируем 2 вершины
rep movsd ;context.vertex[0] = context.vertex[2], context.vertex[1] = context.vertex[3]
mov dword[n],2
jmp .end_f
@@:
cmp dword[edx+offs_cont_begin_type],GL_POLYGON
je .end_f
; default:
; gl_fatal_error("glBegin: type %x not handled\n", c->begin_type);
; }
.end_f:
mov ecx,[n]
mov [edx+offs_cont_vertex_n],ecx
popad
ret
endp
align 4
proc glopEnd uses eax ebx ecx, context:dword, p:dword
mov eax,[context]
; assert(c->in_begin == 1);
cmp dword[eax+offs_cont_begin_type],GL_LINE_LOOP
jne .else_i
cmp dword[eax+offs_cont_vertex_cnt],3
jl .end_i
mov ebx,[eax+offs_cont_vertex]
push ebx
add ebx,2*sizeof.GLVertex
push ebx
stdcall gl_draw_line, eax
jmp .end_i
.else_i:
cmp dword[eax+offs_cont_begin_type],GL_POLYGON
jne .end_i
mov ebx,dword[eax+offs_cont_vertex_cnt]
@@: ;while (ebx >= 3)
cmp ebx,3
jl .end_i
dec ebx
mov ecx,ebx
imul ecx,sizeof.GLVertex
add ecx,[eax+offs_cont_vertex]
push ecx ;ecx = &context.vertex[i]
sub ecx,sizeof.GLVertex
push ecx ;ecx = &context.vertex[i-1]
stdcall gl_draw_triangle, eax,[eax+offs_cont_vertex]
jmp @b
.end_i:
mov dword[eax+offs_cont_in_begin],0
ret
endp