some fixes & optimizations

git-svn-id: svn://kolibrios.org@5353 a494cfbc-eb01-0410-851d-a64ba20cac60

programs/develop/libraries/TinyGL/asm_fork/api.asm

@@ -27,6 +27,13 @@ proc glVertex2f, x:dword, y:dword
 	ret
 endp
 
+align 4
+proc glVertex2fv uses eax, v:dword
+	mov eax,[v]
+	stdcall glVertex4f,[eax],[eax+4],0.0,1.0
+	ret
+endp
+
 align 4
 proc glVertex3f, x:dword, y:dword, z:dword
 	stdcall glVertex4f,[x],[y],[z],1.0
@@ -40,6 +47,13 @@ proc glVertex3fv uses eax, v:dword
 	ret
 endp
 
+align 4
+proc glVertex4fv uses eax, v:dword
+	mov eax,[v]
+	stdcall glVertex4f,[eax],[eax+4],[eax+8],[eax+12]
+	ret
+endp
+
 ; glNormal
 
 align 4
@@ -548,7 +562,7 @@ endl
 	ret
 endp
 
-;/* lightening */
+; lightening
 
 align 4
 proc glMaterialfv uses eax ecx, mode:dword, type:dword, v:dword
@@ -721,7 +735,7 @@ endl
 	ret
 endp
 
-;/* clear */
+; clear
 
 align 4
 proc glClear uses eax, mask:dword
@@ -775,7 +789,7 @@ endl
 	ret
 endp
 
-;/* textures */
+; textures
 
 align 4
 proc glTexImage2D uses ecx edi esi,\
@@ -879,7 +893,7 @@ endl
 	ret
 endp
 
-;/* selection */
+; selection
 
 align 4
 proc glInitNames uses eax
@@ -954,7 +968,7 @@ endl
 	ret
 endp
 
-;/* Special Functions */
+; Special Functions
 
 align 4
 proc glCallList uses eax, list:dword

programs/develop/libraries/TinyGL/asm_fork/arrays.asm

@@ -174,18 +174,23 @@ align 4
 			add esi,[edx+offs_cont_color_array] ;esi = &context.color_array[i]
 			mov edi,ebp
 			sub edi,28 ;edi = &p[1]
-			movsd ;p[1].f = context.color_array[i]
+			mov ebx,[esi+8]
-			movsd
+			mov [edi],ebx   ;p[1].f = context.color_array[i+2]
-			movsd
+			mov ebx,[esi+4]
+			mov [edi+4],ebx ;p[2].f = context.color_array[i+1]
+			mov ebx,[esi]
+			mov [edi+8],ebx ;p[3].f = context.color_array[i]
+			add edi,12
 			cmp dword[size],3
 			jle .e1
+				add esi,12
 				movsd
-				sub edi,20 ;&p[0]
 				jmp .e2
 			.e1:
 				mov dword[edi],1.0 ;если задано 3 параметра, то 4-й ставим по умолчанию 1.0
-				sub edi,16 ;&p[0]
 			.e2:
+			mov edi,ebp
+			sub edi,32 ;edi = &p[0]
 			mov ebx,ebp
 			sub ebx,12 ;ebp-12 = &p[5]
 			push ebx
@@ -256,6 +261,7 @@ align 4
 			.e6:
 				mov dword[edi],0.0 ;если задано 2 параметра, то 3-й ставим по умолчанию 0.0
 				add edi,4
+				jmp .e8 ;и 4-й тоже ставим по умолчанию
 			.e7:
 			cmp dword[size],3
 			jle .e8
@@ -302,7 +308,7 @@ endp
 align 4
 proc glopDrawElements uses eax ebx edx, context:dword, param:dword
 locals
-	p rd 5
+	p rd 8
 endl
 
 	mov edx,[context]
@@ -318,7 +324,7 @@ endl
 	mov eax,[ebx+4]
 	mov [p+4],eax ;p[1].i = param[1].i
 	mov eax,ebp
-	sub eax,20 ;=sizeof(dd)*5
+	sub eax,32 ;=sizeof(dd)*8
 	stdcall glopBegin, edx,eax
 
 ;  for (int ii=0; ii<count; ii++) {
@@ -360,7 +366,7 @@ endl
 ;    }
 ;  }
 	mov eax,ebp
-	sub eax,20 ;=sizeof(dd)*5
+	sub eax,32 ;=sizeof(dd)*8
 	stdcall glopEnd, edx,eax
 	ret
 endp

programs/develop/libraries/TinyGL/asm_fork/clip.asm

@@ -13,6 +13,10 @@ offs_Y equ 4
 offs_Z equ 8
 offs_W equ 12
 
+if DEBUG
+f_ttv db ' gl_transform_to_viewport',0
+end if
+
 align 4
 proc gl_transform_to_viewport uses eax ebx ecx, context:dword,v:dword
 locals
@@ -69,8 +73,8 @@ endl
 	.end_if:
   
 	; texture
-	bt dword[eax+offs_cont_texture_2d_enabled],0
+	cmp dword[eax+offs_cont_texture_2d_enabled],0
-	jnc @f
+	je @f
 		mov dword[point],dword(ZB_POINT_S_MAX - ZB_POINT_S_MIN)
 		fild dword[point]
 		fmul dword[ebx+offs_vert_tex_coord] ;st0 *= v.tex_coord.X
@@ -84,7 +88,7 @@ endl
 		add dword[ebx+offs_vert_zp+offs_zbup_s],ZB_POINT_T_MIN
 	@@:
 if DEBUG ;gl_transform_to_viewport
-push edi
+pushad
 	mov ecx,80
 	mov eax,[ebx+offs_vert_zp]
 	lea edi,[buf_param]
@@ -106,7 +110,7 @@ push edi
 
 	stdcall str_n_cat,edi,txt_nl,2
 	stdcall dbg_print,f_ttv,buf_param
-pop edi
+popad
 end if
 	ret
 endp
@@ -170,11 +174,11 @@ proc interpolate uses eax ebx ecx, q:dword,p0:dword,p1:dword,t:dword
 	fld dword[t]
 
 	; интерполяция по координатам
-	fld dword[ecx+offs_vert_pc+offs_X]
+	fld dword[ecx+offs_vert_pc]
-	fsub dword[ebx+offs_vert_pc+offs_X]
+	fsub dword[ebx+offs_vert_pc]
 	fmul st0,st1
-	fadd dword[ebx+offs_vert_pc+offs_X]
+	fadd dword[ebx+offs_vert_pc]
-	fstp dword[eax+offs_vert_pc+offs_X]
+	fstp dword[eax+offs_vert_pc] ;q.pc.X = p0.pc.X + (p1.pc.X - p0.pc.X) * t
 
 	fld dword[ecx+offs_vert_pc+offs_Y]
 	fsub dword[ebx+offs_vert_pc+offs_Y]
@@ -663,8 +667,8 @@ proc updateTmp uses eax ebx ecx edx, context:dword, q:dword, p0:dword, p1:dword,
 	stdcall gl_clipcode, [ebx+offs_vert_pc+offs_X],[ebx+offs_vert_pc+offs_Y],\
 		[ebx+offs_vert_pc+offs_Z],[ebx+offs_vert_pc+offs_W]
 	mov dword[ebx+offs_vert_clip_code],eax
-	cmp eax,0 ;if (q.clip_code==0)
+	or eax,eax ;if (q.clip_code==0)
-	jne @f
+	jnz @f
 		stdcall gl_transform_to_viewport,[context],ebx
 		mov eax,ebx
 		add eax,offs_vert_zp+offs_zbup_b
@@ -689,7 +693,6 @@ pushad
 	mov ebx,[p0]
 	mov ecx,[p1]
 	mov edx,[p2]
-
 	mov edi,[ebx+offs_vert_clip_code]
 	mov dword[cc],edi
 	mov eax,[ecx+offs_vert_clip_code]
@@ -700,8 +703,9 @@ pushad
 	or edi,eax ;co = cc[0] | cc[1] | cc[2]
 
 	; we handle the non clipped case here to go faster
-	cmp edi,0 ;if (co==0)
+	;or edi,___ - было выше
-	jne .els_0
+	jnz .els_0
+		;if (co==0)
 		mov edi,dword[edx+offs_vert_zp+offs_zbup_x]
 		sub edi,dword[ebx+offs_vert_zp+offs_zbup_x]
 		mov dword[norm],edi
@@ -728,13 +732,12 @@ pushad
 		fstsw ax
 		sahf
 		je .end_f
-		jb @f
+		jbe @f ;jb @f ???
 			inc dword[front] ;front = 0.0 > norm
 		@@:
 		mov edi,[context]
 		mov eax,dword[edi+offs_cont_current_front_face]
 		xor dword[front],eax ;front ^= context.current_front_face
-
 		; back face culling
 		cmp dword[edi+offs_cont_cull_face_enabled],0
 		je .els_1
@@ -765,8 +768,8 @@ pushad
 		;eax = cc[2]
 		and eax,[cc]
 		and eax,[cc+4] ;eax = c_and = cc[0] & cc[1] & cc[2]
-		cmp eax,0 ;if (c_and==0)
+		or eax,eax ;if (c_and==0)
-		jne .end_f
+		jnz .end_f
 			stdcall gl_draw_triangle_clip, [context],ebx,ecx,edx,0
 	.end_f:
 popad
@@ -799,8 +802,8 @@ pushad
 	or edi,eax
 	mov [co],edi ;co = cc[0] | cc[1] | cc[2]
 
-	cmp edi,0 ;if (co == 0)
+	or edi,edi ;if (co == 0)
-	jne .els_0
+	jnz .els_0
 		stdcall gl_draw_triangle, [context],ebx,ecx,edx
 		jmp .end_f
 	.els_0:
@@ -809,8 +812,8 @@ pushad
 		and eax,[cc+4] ;c_and = cc[0] & cc[1] & cc[2]
 
 		; the triangle is completely outside
-		cmp eax,0 ;if (c_and!=0) return
+		or eax,eax ;if (c_and!=0) return
-		jne .end_f
+		jnz .end_f
 
 		; find the next direction to clip
 		.cycle_0: ;while (clip_bit < 6 && (co & (1 << clip_bit)) == 0)
@@ -821,8 +824,8 @@ pushad
 		mov ecx,[clip_bit]
 		shl eax,cl
 		and eax,[co]
-		cmp eax,0
+		or eax,eax
-		jne .cycle_0_end
+		jnz .cycle_0_end
 			inc dword[clip_bit]
 			jmp .cycle_0
 		.cycle_0_end:
@@ -854,13 +857,13 @@ end if
 
 	mov ecx,[p1] ;востанавливаем после shl ___,cl
 
-	cmp eax,0 ;if (co1)
+	or eax,eax ;if (co1)
-	je .els_1
+	jz .els_1
 		; one point outside
 		mov eax,[cc]
 		and eax,[clip_mask]
-		cmp eax,0 ;if (cc[0] & clip_mask)
+		or eax,eax ;if (cc[0] & clip_mask)
-		je .els_2
+		jz .els_2
 			;q[0]=p0 q[1]=p1 q[2]=p2
 			mov [q],ebx
 			mov [q+4],ecx
@@ -869,8 +872,8 @@ end if
 		.els_2:
 		mov eax,[cc+4]
 		and eax,[clip_mask]
-		cmp eax,0 ;else if (cc[1] & clip_mask)
+		or eax,eax ;else if (cc[1] & clip_mask)
-		je .els_3
+		jz .els_3
 			;q[0]=p1 q[1]=p2 q[2]=p0
 			mov [q],ecx
 			mov [q+4],edx
@@ -917,7 +920,7 @@ end if
 		stdcall updateTmp,[context],edi,ebx,edx,eax ;updateTmp(c,&tmp2,q[0],q[2],tt)
 
 		mov eax,[ebx+offs_vert_edge_flag]
-		mov [tmp1+offs_vert_edge_flag],eax ;q[0].edge_flag
+		mov [tmp1.edge_flag],eax ;q[0].edge_flag
 		mov eax,[edx+offs_vert_edge_flag]
 		mov [edge_flag_tmp],eax	;q[2].edge_flag
 		mov dword[edx+offs_vert_edge_flag],0 ;q[2].edge_flag=0
@@ -992,11 +995,9 @@ end if
 		sub edi,offs_vert_pc
 		stdcall updateTmp,[context],edi,[q],[q+8],eax
 
-		mov edx,[q+8]
+		mov dword[tmp1.edge_flag],1
-
+		mov eax,[edx+offs_vert_edge_flag-offs_vert_pc]
-		mov dword[tmp1+offs_vert_edge_flag],1
+		mov dword[tmp2.edge_flag],eax ;tmp2.edge_flag = q[2].edge_flag
-		mov eax,[edx+offs_vert_edge_flag]
-		mov dword[tmp1+offs_vert_edge_flag],eax ;tmp2.edge_flag = q[2].edge_flag
 		mov eax,[clip_bit]
 		inc eax
 		push eax

programs/develop/libraries/TinyGL/asm_fork/examples/test_array0.asm

@@ -192,7 +192,7 @@ err_message_import db 'Error on load import library tinygl.obj',0
 ;--------------------------------------------------
 
 i_end:
-	rb 1024
+	rb 4096
 stacktop:
 cur_dir_path:
 	rb 4096

programs/develop/libraries/TinyGL/asm_fork/matrix.asm

@@ -384,11 +384,13 @@ proc glopScale uses eax ebx ecx, context:dword, p:dword
 	loop @b
 
 if DEBUG ;glopScale
+pushad
 	stdcall dbg_print,txt_gl_scal,txt_nl
 	mov ebx,[eax+offs_cont_matrix_mode]
 	shl ebx,2
 	add ebx,eax
 	stdcall gl_print_matrix,[ebx+offs_cont_matrix_stack_ptr],4
+popad
 end if
 	gl_matrix_update eax,ebx
 	ret
@@ -429,11 +431,13 @@ proc glopTranslate uses eax ebx ecx, context:dword, p:dword
 	fincstp
 
 if DEBUG ;glopTranslate
+pushad
 	stdcall dbg_print,txt_gl_tran,txt_nl
 	mov ebx,[eax+offs_cont_matrix_mode]
 	shl ebx,2
 	add ebx,eax
 	stdcall gl_print_matrix,[ebx+offs_cont_matrix_stack_ptr],4
+popad
 end if
 	gl_matrix_update eax,ebx
 	ret

programs/develop/libraries/TinyGL/asm_fork/tinygl.asm

@@ -12,9 +12,9 @@ include 'zmath.asm'
 include 'clip.asm'
 include 'vertex.asm'
 include 'api.asm'
-include 'list.asm'
+include 'list.asm' ;gl_add_op
 include 'init.asm'
-include 'matrix.asm'
+include 'matrix.asm' ;gl_print_matrix
 include 'texture.asm'
 include 'misc.asm'
 include 'clear.asm'
@@ -36,42 +36,7 @@ include 'glu.asm'
 
 if DEBUG
 include 'info_fun_float.inc'
-end if
 
-; ***
-glVertex2d: ;(double ,double)
-glVertex2fv: ;(float *)
-glVertex2dv: ;(double *)
-glVertex3d: ;(double ,double ,double)
-glVertex3dv: ;(double *)
-glVertex4d: ;(double ,double ,double, double )
-glVertex4fv: ;(float *)
-glVertex4dv: ;(double *)
-glColor3d: ;(double ,double ,double)
-glColor3dv: ;(double *)
-glColor4d: ;(double ,double ,double, double )
-glColor4dv: ;(double *)
-glNormal3d: ;(double ,double ,double)
-glNormal3dv: ;(double *)
-glTexCoord1f: ;(float)
-glTexCoord1d: ;(double)
-glTexCoord1fv: ;(float *)
-glTexCoord1dv: ;(double *)
-glTexCoord2d: ;(double ,double)
-glTexCoord2dv: ;(double *)
-glTexCoord3f: ;(float ,float ,float)
-glTexCoord3d: ;(double ,double ,double)
-glTexCoord3fv: ;(float *)
-glTexCoord3dv: ;(double *)
-glTexCoord4d: ;(double ,double ,double, double )
-glTexCoord4fv: ;(float *)
-glTexCoord4dv: ;(double *)
-glGenTextures: ;(int n, unsigned int *textures)
-glDeleteTextures: ;(int n, const unsigned int *textures)
-glGetIntegerv: ;(int pname,int *params)
-glGetFloatv: ;(int pname, float *v)
-
-if DEBUG
 align 4
 txt_nl db 13,10,0
 txt_sp db ' ',0
@@ -140,6 +105,37 @@ align 4
 	ret	   ;пока в стеке храниться кол-во вызовов то столько раз мы и будем вызываться
 end if
 
+; ***
+glVertex2d: ;(double ,double)
+glVertex2dv: ;(double *)
+glVertex3d: ;(double ,double ,double)
+glVertex3dv: ;(double *)
+glVertex4d: ;(double ,double ,double, double )
+glVertex4dv: ;(double *)
+glColor3d: ;(double ,double ,double)
+glColor3dv: ;(double *)
+glColor4d: ;(double ,double ,double, double )
+glColor4dv: ;(double *)
+glNormal3d: ;(double ,double ,double)
+glNormal3dv: ;(double *)
+glTexCoord1f: ;(float)
+glTexCoord1d: ;(double)
+glTexCoord1fv: ;(float *)
+glTexCoord1dv: ;(double *)
+glTexCoord2d: ;(double ,double)
+glTexCoord2dv: ;(double *)
+glTexCoord3f: ;(float ,float ,float)
+glTexCoord3d: ;(double ,double ,double)
+glTexCoord3fv: ;(float *)
+glTexCoord3dv: ;(double *)
+glTexCoord4d: ;(double ,double ,double, double )
+glTexCoord4fv: ;(float *)
+glTexCoord4dv: ;(double *)
+glGenTextures: ;(int n, unsigned int *textures)
+glDeleteTextures: ;(int n, const unsigned int *textures)
+glGetIntegerv: ;(int pname,int *params)
+glGetFloatv: ;(int pname, float *v)
+
 align 4
 f_kosgl_1 db 'kosglMakeCurrent',0
 err_0 db 'Error while initializing Z buffer',13,10,0
@@ -151,20 +147,13 @@ f_vp db ' glViewport',0
 err_4 db 'error while resizing display',13,10,0
 err_5 db 'size too small',13,10,0
 f_clipcode db ' gl_clipcode',0
-f_ttv db ' gl_transform_to_viewport',0
-f_vt db ' gl_vertex_transform',0
-f_ev db ' gl_eval_viewport',0
 f_zbz db ' ZB_line_z',0
 f_zb db ' ZB_line',0
-f_m4m db 'gl_M4_Mul',0
-f_m4ml db 'gl_M4_MulLeft',0
 f_find_l db 'find_list',0
 f_alloc_l db 'alloc_list',0
 f_is_l db 'glIsList',0
 f_gen_l db 'glGenLists',0
 f_end_l db 'glEndList',0
-;f_fill_trf db 'ZB_fillTriangleFlat',0
-;f_fill_trrgb db 'ZB_fillTriangleSmooth',0
 f_fill_tr db 'ZB_fillTriangle...',0
 f_fill_tr_nl db ' lines',0
 f_fill_tr_nll db ' len',0

programs/develop/libraries/TinyGL/asm_fork/vertex.asm

@@ -1,3 +1,6 @@
+if DEBUG
+f_vt db ' gl_vertex_transform',0
+end if
 
 align 4
 proc glopNormal uses ecx esi edi, context:dword, p:dword
@@ -117,7 +120,7 @@ proc glopBegin uses eax ebx ecx edx, context:dword, p:dword
 locals
 	tmp M4
 endl
-;    assert(c->in_begin == 0);
+;assert(context.in_begin == 0)
 
 	mov edx,[context]
 	mov ebx,[p]
@@ -137,7 +140,7 @@ if DEBUG ;context.matrix_stack_ptr[0]
 end if
 		; precompute inverse modelview
 		mov ebx,ebp
-		sub ebx,64
+		sub ebx,sizeof.M4
 		stdcall gl_M4_Inv, ebx,dword[edx+offs_cont_matrix_stack_ptr]
 if DEBUG ;tmp
 	stdcall dbg_print,txt_sp,txt_nl
@@ -165,18 +168,21 @@ end if
 		fcomp st1
 		fstsw ax
 		sahf
-		jne .end_if_0
+		jne @f
 		fld dword[ecx+13*4]
 		fcomp st1
 		fstsw ax
 		sahf
-		jne .end_if_0
+		jne @f
 		fld dword[ecx+14*4]
 		fcomp st1
 		fstsw ax
 		sahf
-		jne .end_if_0
+		jne @f
 			mov dword[edx+offs_cont_matrix_model_projection_no_w_transform],1
+		@@:
+		ffree st0 ;0.0
+		fincstp
 	.end_if_0:
 
 		; test if the texture matrix is not Identity

programs/develop/libraries/TinyGL/asm_fork/zmath.asm

@@ -3,6 +3,11 @@
 
 ; ******* Gestion des matrices 4x4 ******
 
+if DEBUG
+f_m4m db 'gl_M4_Mul',0
+f_m4ml db 'gl_M4_MulLeft',0
+end if
+
 align 4
 proc gl_M4_Id uses eax ecx edi, a:dword
 	mov edi,[a]
@@ -115,23 +120,21 @@ pushad
 
 	mov edx,[c]
 	mov dword[i],0
+	mov eax,ebp
+	sub eax,sizeof.M4
 	.cycle_0: ;i
-		xor ebx,ebx
+		xor ebx,ebx ;j=0
 		.cycle_1: ;j
-			finit
 			fldz ;sum=0
-			xor ecx,ecx
+			xor ecx,ecx ;k=0
-			mov eax,ebp
-			sub eax,sizeof.M4
 			M4_reg edi,eax,dword[i],0
 			.cycle_2: ;k
 				fld dword[edi]
 				add edi,4
 				M4_reg esi,[b],ecx,ebx
 				fmul dword[esi]
-				fadd st0,st1 ;sum += a[i][k] * b[k][j]
+				faddp ;sum += a[i][k] * b[k][j]
 				inc ecx
-				add eax,4
 				cmp ecx,4
 				jl .cycle_2
 			fstp dword[edx] ;c[i][j] = sum
@@ -142,7 +145,6 @@ pushad
 		inc dword[i]
 		cmp dword[i],4
 		jl .cycle_0
-	finit
 if DEBUG ;gl_M4_MulLeft
 	stdcall dbg_print,f_m4ml,txt_nl
 	stdcall gl_print_matrix,[c],4