some fixes

git-svn-id: svn://kolibrios.org@5256 a494cfbc-eb01-0410-851d-a64ba20cac60
This commit is contained in:
IgorA 2014-12-22 09:21:22 +00:00
parent f8f660cfef
commit 40df1da8cf
8 changed files with 1151 additions and 269 deletions

View File

@ -0,0 +1,230 @@
use32
org 0x0
db 'MENUET01'
dd 0x1
dd start
dd i_end
dd mem,stacktop
dd 0,cur_dir_path
include '../../../../../../programs/proc32.inc'
include '../../../../../../programs/macros.inc'
include '../../../../../../programs/develop/libraries/box_lib/load_lib.mac'
include '../../../../../../programs/dll.inc'
include '../opengl_const.inc'
@use_library
align 4
start:
load_library name_tgl, cur_dir_path, library_path, system_path, \
err_message_found_lib, head_f_l, import_lib_tinygl, err_message_import, head_f_i
cmp eax,-1
jz button.exit
mcall 40,0x27
stdcall [kosglMakeCurrent], 10,10,300,225,ctx1
stdcall [glEnable], GL_DEPTH_TEST
stdcall [gluNewQuadric]
mov [qObj],eax
stdcall [glClearColor], 0.5,0.5,0.5,0.0
stdcall [glShadeModel], GL_SMOOTH
call draw_3d
align 4
red_win:
call draw_window
align 4
still:
mcall 10
cmp al,1
jz red_win
cmp al,2
jz key
cmp al,3
jz button
jmp still
align 4
draw_window:
pushad
mcall 12,1
mov edx,0x33ffffff ;0x73ffffff
mcall 0,(50 shl 16)+330,(30 shl 16)+275,,,caption
stdcall [kosglSwapBuffers]
mcall 12,2
popad
ret
align 4
key:
mcall 2
cmp ah,27 ;Esc
je button.exit
cmp ah,61 ;+
jne @f
fld dword[scale]
fadd dword[delt_sc]
fstp dword[scale]
call draw_3d
stdcall [kosglSwapBuffers]
@@:
cmp ah,45 ;-
jne @f
fld dword[scale]
fsub dword[delt_sc]
fstp dword[scale]
call draw_3d
stdcall [kosglSwapBuffers]
@@:
cmp ah,178 ;Up
jne @f
fld dword[angle_y]
fadd dword[delt_size]
fstp dword[angle_y]
call draw_3d
stdcall [kosglSwapBuffers]
@@:
cmp ah,177 ;Down
jne @f
fld dword[angle_y]
fsub dword[delt_size]
fstp dword[angle_y]
call draw_3d
stdcall [kosglSwapBuffers]
@@:
cmp ah,176 ;Left
jne @f
fld dword[angle_z]
fadd dword[delt_size]
fstp dword[angle_z]
call draw_3d
stdcall [kosglSwapBuffers]
@@:
cmp ah,179 ;Right
jne @f
fld dword[angle_z]
fsub dword[delt_size]
fstp dword[angle_z]
call draw_3d
stdcall [kosglSwapBuffers]
@@:
jmp still
align 4
button:
mcall 17
cmp ah,1
jne still
.exit:
stdcall [gluDeleteQuadric], [qObj]
mcall -1
align 4
caption db 'Test gluSphere, [Esc] - exit, [<-],[->],[Up],[Down] - rotate',0
align 4
ctx1 db 28 dup (0) ;TinyGLContext or KOSGLContext
;sizeof.TinyGLContext = 28
align 4
draw_3d:
stdcall [glClear], GL_COLOR_BUFFER_BIT + GL_DEPTH_BUFFER_BIT ;очистим буфер цвета и глубины
stdcall [glColor3f], 1.0, 1.0, 0.0
stdcall [glPushMatrix]
call SetLight
stdcall [glTranslatef], 0.0,0.0,0.5
stdcall [glScalef], [scale], [scale], [scale]
stdcall [glRotatef], [angle_z],0.0,0.0,1.0
stdcall [glRotatef], [angle_y],0.0,1.0,0.0
stdcall [gluSphere], [qObj], 1.0, 16,16
stdcall [glColor3f], 1.0, 0.0, 0.0
stdcall [glTranslatef], -1.6,0.0,0.0
stdcall [gluSphere], [qObj], 0.55, 8,8
stdcall [glColor3f], 0.0, 0.0, 1.0
stdcall [glTranslatef], 3.2,0.0,0.0
stdcall [gluSphere], [qObj], 0.55, 8,8
stdcall [glPopMatrix]
ret
align 4
SetLight:
stdcall [glLightfv], GL_LIGHT0, GL_POSITION, light_position
stdcall [glLightfv], GL_LIGHT0, GL_SPOT_DIRECTION, light_dir
stdcall [glLightfv], GL_LIGHT0, GL_DIFFUSE, white_light
stdcall [glLightfv], GL_LIGHT0, GL_SPECULAR, white_light
stdcall [glEnable], GL_COLOR_MATERIAL
stdcall [glColorMaterial], GL_FRONT, GL_AMBIENT_AND_DIFFUSE
stdcall [glMaterialfv], GL_FRONT, GL_SPECULAR, mat_specular
stdcall [glMaterialf], GL_FRONT, GL_SHININESS, mat_shininess
stdcall [glLightModelfv], GL_LIGHT_MODEL_AMBIENT, lmodel_ambient
stdcall [glEnable],GL_LIGHTING
stdcall [glEnable],GL_LIGHT0
ret
qObj dd 0
scale dd 0.4
delt_sc dd 0.05
angle_z dd 0.0
angle_y dd 0.0
delt_size dd 3.0
light_position dd 0.0, 0.0, 2.0, 1000.0 ; Расположение источника [0][1][2], чем ближе [3] к 0, тем ярче свет
light_dir dd 0.0,0.0,0.0 ;направление лампы
mat_specular dd 0.3, 0.3, 0.3, 1.0 ; Цвет блика
mat_shininess dd 3.0 ; Размер блика (обратная пропорция)
white_light dd 0.8, 0.8, 0.8, 1.0 ; Цвет и интенсивность освещения, генерируемого источником
lmodel_ambient dd 0.2, 0.2, 0.2, 1.0 ; Параметры фонового освещения
;--------------------------------------------------
align 4
import_lib_tinygl:
macro E_LIB n
{
n dd sz_#n
}
include '../export.inc'
dd 0,0
macro E_LIB n
{
sz_#n db `n,0
}
include '../export.inc'
;--------------------------------------------------
system_path db '/sys/lib/'
name_tgl db 'tinygl.obj',0
err_message_found_lib db 'Sorry I cannot load library tinygl.obj',0
head_f_i:
head_f_l db 'System error',0
err_message_import db 'Error on load import library tinygl.obj',0
;--------------------------------------------------
i_end:
rb 1024
stacktop:
cur_dir_path:
rb 4096
library_path:
rb 4096
mem:

View File

@ -1,51 +1,76 @@
;include 'msghandling.inc'
align 4 align 4
proc glopMaterial, context:dword, p:dword proc glopMaterial uses eax ebx ecx edi esi, context:dword, p:dword
; int mode=p[1].i; ; edi -> GLMaterial *m
; int type=p[2].i; mov eax,[context]
; float *v=&p[3].f; mov ebx,[p]
; int i; mov ecx,[ebx+4] ;ecx = p[1]
; GLMaterial *m;
; if (mode == GL_FRONT_AND_BACK) { cmp ecx,GL_FRONT_AND_BACK ;if (mode == GL_FRONT_AND_BACK)
; p[1].i=GL_FRONT; jne @f
; glopMaterial(c,p); mov dword[ebx+4],GL_FRONT ;p[1].i=GL_FRONT
; mode=GL_BACK; mov edi,ebp
; } add edi,12
; if (mode == GL_FRONT) m=&c->materials[0]; stdcall glopMaterial,eax,edi
; else m=&c->materials[1]; mov ecx,GL_BACK
@@:
mov edi,[eax+offs_cont_materials]
cmp ecx,GL_FRONT ;if (mode == GL_FRONT) m=&context.materials[0]
je @f
add edi,sizeof.GLMaterial ;else m=&context.materials[1]
@@:
; switch(type) { mov ecx,4
; case GL_EMISSION: mov esi,ebx ;esi = &p
; for(i=0;i<4;i++) add esi,12 ;esi = &p[3]
; m->emission.v[i]=v[i]; mov ebx,[ebx+8] ;ebx = p[2]
; break; cmp ebx,GL_EMISSION
; case GL_AMBIENT: jne @f
; for(i=0;i<4;i++) ;add edi,offs_mate_emission ;offs_mate_emission=0
; m->ambient.v[i]=v[i]; rep movsd
; break; jmp .end_f
; case GL_DIFFUSE: @@:
; for(i=0;i<4;i++) cmp ebx,GL_AMBIENT
; m->diffuse.v[i]=v[i]; jne @f
; break; add edi,offs_mate_ambient
; case GL_SPECULAR: rep movsd
; for(i=0;i<4;i++) jmp .end_f
; m->specular.v[i]=v[i]; @@:
; break; cmp ebx,GL_DIFFUSE
; case GL_SHININESS: jne @f
; m->shininess=v[0]; add edi,offs_mate_diffuse
; m->shininess_i = (v[0]/128.0f)*SPECULAR_BUFFER_RESOLUTION; rep movsd
; break; jmp .end_f
; case GL_AMBIENT_AND_DIFFUSE: @@:
; for(i=0;i<4;i++) cmp ebx,GL_SPECULAR
; m->diffuse.v[i]=v[i]; jne @f
; for(i=0;i<4;i++) add edi,offs_mate_specular
; m->ambient.v[i]=v[i]; rep movsd
; break; jmp .end_f
; default: @@:
cmp ebx,GL_SHININESS
jne @f
fld dword[esi]
add edi,offs_mate_shininess
movsd
mov dword[edi],SPECULAR_BUFFER_RESOLUTION
fdiv dword[an180f]
fimul dword[edi]
fistp dword[edi] ;m.shininess_i = (v[0]/128.0f)*SPECULAR_BUFFER_RESOLUTION
jmp .end_f
@@:
cmp ebx,GL_AMBIENT_AND_DIFFUSE
jne @f
add edi,offs_mate_ambient
rep movsd
sub esi,16
;edi = &offs_mate_diffuse
mov ecx,4
rep movsd
jmp .end_f
@@: ;default
; assert(0); ; assert(0);
; } .end_f:
ret ret
endp endp
@ -61,13 +86,15 @@ proc glopColorMaterial uses eax ebx ecx, context:dword, p:dword
endp endp
align 4 align 4
proc glopLight uses eax ebx ecx edx, context:dword, p:dword proc glopLight context:dword, p:dword
locals
pos V4
endl
pushad
mov eax,[context] mov eax,[context]
mov ebx,[p] mov ebx,[p]
mov edx,[ebx+4] ;edx = p[1] mov edx,[ebx+4] ;edx = p[1]
; V4 v;
; assert(edx >= GL_LIGHT0 && edx < GL_LIGHT0+MAX_LIGHTS ); ; assert(edx >= GL_LIGHT0 && edx < GL_LIGHT0+MAX_LIGHTS );
sub edx,GL_LIGHT0 sub edx,GL_LIGHT0
@ -108,26 +135,40 @@ proc glopLight uses eax ebx ecx edx, context:dword, p:dword
@@: @@:
cmp ecx,GL_POSITION cmp ecx,GL_POSITION
jne @f jne @f
; { mov edi,ebx ;ebx = [ebp+12] = [p] = &p[0]
; for(i=0;i<4;i++) v.v[i]=p[3+i].f; add edi,12 ;&p[3]
; V4 pos; mov esi,ebp
; gl_M4_MulV4(&pos,c->matrix_stack_ptr[0],&v); sub esi,16 ;&pos
; stdcall gl_M4_MulV4, esi,dword[eax+offs_cont_matrix_stack_ptr],edi
; l->position=pos; mov edi,edx
; add edi,offs_ligh_position
; if (l->position.v[3] == 0) { mov ecx,4
; l->norm_position.X=pos.X; rep movsd ;l.position=pos
; l->norm_position.Y=pos.Y;
; l->norm_position.Z=pos.Z; fld dword[edi-4] ;if(l.position.v[3] == 0)
; ftst
; gl_V3_Norm(&l->norm_position); fstsw ax
; } sahf
; } jne .end_i
;mov esi,ebp
sub esi,16 ;&pos
mov edi,edx
add edi,offs_ligh_norm_position
mov ecx,3
rep movsd ;l.norm_position=pos[1,2,3]
;mov edi,edx
;add edi,offs_ligh_norm_position
sub edi,12
stdcall gl_V3_Norm,edi ;&l.norm_position
.end_i:
ffree st0
fincstp
jmp .end_f jmp .end_f
@@: @@:
cmp ecx,GL_SPOT_DIRECTION cmp ecx,GL_SPOT_DIRECTION
jne @f jne @f
mov esi,ebx mov esi,ebx ;&p[0]
add esi,12 add esi,12
mov edi,edx mov edi,edx
add edi,offs_ligh_spot_direction add edi,offs_ligh_spot_direction
@ -150,15 +191,25 @@ proc glopLight uses eax ebx ecx edx, context:dword, p:dword
jmp .end_f jmp .end_f
@@: @@:
cmp ecx,GL_SPOT_CUTOFF cmp ecx,GL_SPOT_CUTOFF
jne @f jne .end_spot_c
; { fld dword[ebp+12] ;float a=v.v[0]
; float a=v.v[0];
; assert(a == 180 || (a>=0 && a<=90)); ; assert(a == 180 || (a>=0 && a<=90));
; l->spot_cutoff=a; fst dword[edi+offs_ligh_spot_cutoff] ;l.spot_cutoff=a
; if (a != 180) l->cos_spot_cutoff=cos(a * M_PI / 180.0); fcom dword[an180f] ;if (a != 180)
; } fstsw ax
sahf
jne @f
fldpi
fmulp
fdiv dword[an180f]
fcos
fstp dword[edi+offs_ligh_spot_cutoff] ;l.cos_spot_cutoff=cos(a * M_PI / 180.0)
jmp .end_f jmp .end_f
@@: @@:
ffree st0
fincstp
jmp .end_f
.end_spot_c:
cmp ecx,GL_CONSTANT_ATTENUATION cmp ecx,GL_CONSTANT_ATTENUATION
mov ecx,[ebx+12] mov ecx,[ebx+12]
mov [edi+offs_ligh_attenuation],ecx ;l->attenuation[0]=p[3] mov [edi+offs_ligh_attenuation],ecx ;l->attenuation[0]=p[3]
@ -176,6 +227,7 @@ proc glopLight uses eax ebx ecx edx, context:dword, p:dword
@@: ;default: @@: ;default:
; assert(0); ; assert(0);
.end_f: .end_f:
popad
ret ret
endp endp
@ -184,46 +236,69 @@ align 4
proc glopLightModel uses ebx ecx esi edi, context:dword, p:dword proc glopLightModel uses ebx ecx esi edi, context:dword, p:dword
mov edi,[context] mov edi,[context]
mov ebx,[p] mov ebx,[p]
mov esi,[ebx+8] mov ebx,[ebx+4]
mov esi,[ebp+12] ;&p[0]
add esi,8 ;&p[2]
cmp dword[ebx+4],GL_LIGHT_MODEL_AMBIENT cmp ebx,GL_LIGHT_MODEL_AMBIENT
jne @f jne @f
mov ecx,4 mov ecx,4
mov edi,dword[edi+offs_cont_ambient_light_model] add edi,offs_cont_ambient_light_model
rep movsd ;for(i=0;i<4;i++) context.ambient_light_model.v[i]=v[i] rep movsd ;for(i=0;i<4;i++) context.ambient_light_model.v[i]=v[i]
jmp .end_f jmp .end_f
@@: @@:
cmp dword[ebx+4],GL_LIGHT_MODEL_LOCAL_VIEWER cmp ebx,GL_LIGHT_MODEL_LOCAL_VIEWER
jne @f jne @f
fld dword[esi] ;st0 = p[2].v[0] fld dword[esi] ;st0 = p[2]
fistp dword[edi+offs_cont_local_light_model] fistp dword[edi+offs_cont_local_light_model]
jmp .end_f jmp .end_f
@@: @@:
cmp dword[ebx+4],GL_LIGHT_MODEL_TWO_SIDE cmp ebx,GL_LIGHT_MODEL_TWO_SIDE
jne @f jne @f
fld dword[esi] ;st0 = p[2].v[0] fld dword[esi] ;st0 = p[2]
fistp dword[edi+offs_cont_light_model_two_side] fistp dword[edi+offs_cont_light_model_two_side]
jmp .end_f jmp .end_f
@@: ;default: @@: ;default:
; tgl_warning("glopLightModel: illegal pname: 0x%x\n", dword[ebx+4]); ; tgl_warning("glopLightModel: illegal pname: 0x%x\n", ebx);
; //assert(0); ; //assert(0);
.end_f: .end_f:
ret ret
endp endp
;static inline float clampf(float a,float min,float max) macro clampf a, min, max
;{ {
; if (a<min) return min; local .o_1
; else if (a>max) return max; local .o_2
; else return a; local .end_m
;} fld dword a ;if (a<=0.0)
ftst
fstsw ax
sahf
ja .o_1
ffree st0
fincstp
mov eax,0.0
jmp .end_m ;return 0.0
.o_1:
fld1 ;else if (a>=1.0)
fcompp
fstsw ax
sahf
jb .o_2
mov eax,1.0
jmp .end_m ;return 1.0
.o_2:
mov eax,dword a ;else return a
.end_m:
}
align 4 align 4
proc gl_enable_disable_light uses eax ebx ecx, context:dword, light:dword, v:dword proc gl_enable_disable_light uses eax ebx ecx, context:dword, light:dword, v:dword
mov eax,[context] mov eax,[context]
mov ebx,[light] mov ebx,[light]
imul ebx,sizeof.GLLight imul ebx,sizeof.GLLight
add ebx,[eax+offs_cont_lights] add ebx,eax
add ebx,offs_cont_lights
xor ecx,ecx xor ecx,ecx
cmp dword[ebx+offs_ligh_enabled],0 cmp dword[ebx+offs_ligh_enabled],0
@ -231,8 +306,8 @@ proc gl_enable_disable_light uses eax ebx ecx, context:dword, light:dword, v:dwo
not ecx not ecx
@@: @@:
and ecx,[v] and ecx,[v]
cmp ecx,0 or ecx,ecx
je @f jz @f
;if (v && !l.enabled) ;if (v && !l.enabled)
mov dword[ebx+offs_ligh_enabled],1 mov dword[ebx+offs_ligh_enabled],1
mov ecx,[eax+offs_cont_first_light] mov ecx,[eax+offs_cont_first_light]
@ -247,8 +322,8 @@ proc gl_enable_disable_light uses eax ebx ecx, context:dword, light:dword, v:dwo
not ecx not ecx
@@: @@:
and ecx,[ebx+offs_ligh_enabled] and ecx,[ebx+offs_ligh_enabled]
cmp ecx,0 or ecx,ecx
je .end_f jz .end_f
;else if (!v && l.enabled) ;else if (!v && l.enabled)
mov dword[ebx+offs_ligh_enabled],0 ;l.enabled = 0 mov dword[ebx+offs_ligh_enabled],0 ;l.enabled = 0
mov ecx,[ebx+offs_ligh_next] mov ecx,[ebx+offs_ligh_next]
@ -269,130 +344,388 @@ proc gl_enable_disable_light uses eax ebx ecx, context:dword, light:dword, v:dwo
ret ret
endp endp
align 4
fl_1e_3 dd 1.0e-3
; non optimized lightening model ; non optimized lightening model
align 4 align 4
proc gl_shade_vertex, context:dword, v:dword proc gl_shade_vertex, context:dword, v:dword
locals
R dd ? ;float
G dd ? ;float
B dd ? ;float
A dd ? ;float
s V3
d V3
dist dd ? ;float
tmp dd ? ;float
att dd ? ;float
dot_spot dd ? ;float
lR dd ? ;float
lB dd ? ;float
lG dd ? ;float
twoside dd ? ;int
idx dd ? ;int
n V3 ;ebp-24
vcoord V3 ;ebp-12
endl
pushad
; ebx -> GLLight *l
; ecx -> GLMaterial *m
; esi -> GLVertex *v
mov esi,[v]
mov edx,[context]
mov ecx,[edx+offs_cont_materials] ;ecx(m) = &context.materials[0]
mov eax,[edx+offs_cont_light_model_two_side]
mov [twoside],eax
; float R,G,B,A; add esi,offs_vert_normal
; GLMaterial *m; mov edi,ebp
; GLLight *l; sub edi,24 ;edi = &n
; V3 n,s,d; movsd ;n.X=v.normal.X
; float dist,tmp,att,dot,dot_spot,dot_spec; movsd ;n.Y=v.normal.Y
; int twoside = c->light_model_two_side; movsd ;n.Z=v.normal.Z
mov esi,[v]
; m=&c->materials[0]; fld dword[edx+offs_cont_ambient_light_model]
fmul dword[ecx+offs_mate_ambient]
fadd dword[ecx] ;offs_mate_emission=0
fstp dword[R] ;R=m.emission.v[0]+m.ambient.v[0]*context.ambient_light_model.v[0]
fld dword[edx+offs_cont_ambient_light_model+4]
fmul dword[ecx+offs_mate_ambient+4]
fadd dword[ecx+offs_mate_emission+4]
fstp dword[G]
fld dword[edx+offs_cont_ambient_light_model+8]
fmul dword[ecx+offs_mate_ambient+8]
fadd dword[ecx+offs_mate_emission+8]
fstp dword[B]
clampf [ecx+offs_mate_diffuse+12],0,1
mov [A],eax ;A=clampf(m.diffuse.v[3],0,1)
; n.X=v->normal.X; mov ebx,[edx+offs_cont_first_light]
; n.Y=v->normal.Y; .cycle_0: ;for(l=context.first_light;l!=NULL;l=l.next)
; n.Z=v->normal.Z; or ebx,ebx
jz .cycle_0_end
; R=m->emission.v[0]+m->ambient.v[0]*c->ambient_light_model.v[0]; ; ambient
; G=m->emission.v[1]+m->ambient.v[1]*c->ambient_light_model.v[1]; fld dword[ecx+offs_mate_ambient]
; B=m->emission.v[2]+m->ambient.v[2]*c->ambient_light_model.v[2]; fmul dword[ebx+offs_ligh_ambient]
; A=clampf(m->diffuse.v[3],0,1); fstp dword[lR] ;lR=l.ambient.v[0] * m.ambient.v[0]
fld dword[ecx+offs_mate_ambient+4]
fmul dword[ebx+offs_ligh_ambient+4]
fstp dword[lG] ;lG=l.ambient.v[1] * m.ambient.v[1]
fld dword[ecx+offs_mate_ambient+8]
fmul dword[ebx+offs_ligh_ambient+8]
fstp dword[lB] ;lB=l.ambient.v[2] * m.ambient.v[2]
; for(l=c->first_light;l!=NULL;l=l->next) { cmp dword[ebx+offs_ligh_position+12],0 ;if (l.position.v[3] == 0)
; float lR,lB,lG; jne .els_0
; light at infinity
mov eax,[ebx+offs_ligh_position]
mov [d],eax ;d.X=l.position.v[0]
mov eax,[ebx+offs_ligh_position+4]
mov [d+4],eax ;d.Y=l.position.v[1]
mov eax,[ebx+offs_ligh_position+8]
mov [d+8],eax ;d.Z=l.position.v[2]
mov dword[att],1.0
jmp .els_0_end
.els_0:
; distance attenuation
fld dword[ebx+offs_ligh_position]
fsub dword[esi+offs_vert_ec]
fstp dword[d] ;d.X=l.position.v[0]-v.ec.v[0]
fld dword[ebx+offs_ligh_position+offs_Y]
fsub dword[esi+offs_vert_ec+offs_Y]
fstp dword[d+offs_Y] ;d.Y=l.position.v[1]-v.ec.v[1]
fld dword[ebx+offs_ligh_position+offs_Z]
fsub dword[esi+offs_vert_ec+offs_Z]
fstp dword[d+offs_Z] ;d.Z=l.position.v[2]-v.ec.v[2]
fld dword[d]
fmul st0,st0
fld dword[d+offs_Y]
fmul st0,st0
faddp
fld dword[d+offs_Z]
fmul st0,st0
faddp
fsqrt
fst dword[dist] ;dist=sqrt(d.X^2+d^2+d^2)
fcom dword[fl_1e_3]
fstsw ax
sahf
jbe @f ;if (dist>1.0e-3)
fld1
fdiv st0,st1
fld dword[d]
fmul st0,st1
fstp dword[d]
fld dword[d+offs_Y]
fmul st0,st1
fstp dword[d+offs_Y]
fld dword[d+offs_Z]
fmul st0,st1
fstp dword[d+offs_Z]
ffree st0 ;1.0/dist
fincstp
@@:
fld dword[ebx+offs_ligh_attenuation+8]
fmul st0,st1 ;st0 = dist * l.attenuation[2]
fadd dword[ebx+offs_ligh_attenuation+4]
fmul st0,st1
fadd dword[ebx+offs_ligh_attenuation]
fld1
fdiv st0,st1
fstp dword[att] ;att = 1.0f/(l.attenuation[0]+dist*(l.attenuation[1]+dist*l.attenuation[2]))
ffree st0 ;1.0
fincstp
ffree st0 ;dist
fincstp
.els_0_end:
fld dword[d]
fmul dword[n]
fld dword[d+offs_Y]
fmul dword[n+offs_Y]
faddp
fld dword[d+offs_Z]
fmul dword[n+offs_Z]
faddp ;dot = d.X*n.X+d.Y*n.Y+d.Z*n.Z
cmp dword[twoside],0 ;if (twoside && dot < 0)
je @f
ftst ;if (dot<0)
fstsw ax
sahf
jae @f
fchs ;dot = -dot
@@:
ftst ;if (dot>0)
fstsw ax
sahf
jle .if0_end
; diffuse light
fld dword[ecx+offs_mate_diffuse]
fmul dword[ebx+offs_ligh_diffuse]
fmul st0,st1
fadd dword[lR]
fstp dword[lR] ;lR+=dot * l.diffuse.v[0] * m.diffuse.v[0]
fld dword[ecx+offs_mate_diffuse+4]
fmul dword[ebx+offs_ligh_diffuse+4]
fmul st0,st1
fadd dword[lG]
fstp dword[lG] ;lG+=dot * l.diffuse.v[1] * m.diffuse.v[1]
fld dword[ecx+offs_mate_diffuse+8]
fmul dword[ebx+offs_ligh_diffuse+8]
fmul st0,st1
fadd dword[lB]
fstp dword[lB] ;lB+=dot * l.diffuse.v[2] * m.diffuse.v[2]
ffree st0 ;dot
fincstp
; /* ambient */ ; spot light
; lR=l->ambient.v[0] * m->ambient.v[0]; fld dword[ebx+offs_ligh_spot_cutoff]
; lG=l->ambient.v[1] * m->ambient.v[1]; fcomp dword[an180f] ;if (l.spot_cutoff != 180)
; lB=l->ambient.v[2] * m->ambient.v[2]; fstsw ax
sahf
jne .if1_end
fld dword[ebx+offs_ligh_norm_spot_direction]
fmul dword[d]
fld dword[ebx+offs_ligh_norm_spot_direction+4]
fmul dword[d+offs_Y]
faddp
fld dword[ebx+offs_ligh_norm_spot_direction+8]
fmul dword[d+offs_Z]
faddp
fchs
fst dword[dot_spot]
cmp dword[twoside],0 ;if (twoside && dot_spot < 0)
je @f
ftst ;if (dot_spot<0)
fstsw ax
sahf
jae @f
fchs ;dot_spot = -dot_spot
@@:
fcom dword[ebx+offs_ligh_cos_spot_cutoff] ;if (dot_spot < l.cos_spot_cutoff)
fstsw ax
sahf
jae .els_1
; no contribution
ffree st0 ;dot_spot
fincstp
mov ebx,[ebx+offs_ligh_next]
jmp .cycle_0 ;continue
.els_1:
; TODO: optimize
fld dword[ebx+offs_ligh_spot_exponent]
ftst ;if (l.spot_exponent > 0)
fstsw ax
sahf
jbe @f
fxch st1 ;dot_spot <--> l.spot_exponent
;Вычисляем x^y
;fld y
;fld x
fyl2x ;Стек FPU теперь содержит: st0=z=y*log2(x):
;Теперь считаем 2**z:
fld st0 ;Создаем еще одну копию z
frndint ;Округляем
fsubr st0,st1 ;st1=z, st0=z-trunc(z)
f2xm1 ;st1=z, st0=2**(z-trunc(z))-1
fld1
faddp ;st1=z, st0=2**(z-trunc(z))
fscale ;st1=z, st0=(2**trunc(z))*(2**(z-trunc(z)))=2**t
fxch st1
fstp st ;Результат остается на вершине стека st0
fmul dword[att]
fstp dword[att] ;att=att*pow(dot_spot,l.spot_exponent)
jmp .if1_end
@@:
ffree st0 ;l.spot_exponent
fincstp
ffree st0 ;dot_spot
fincstp
.if1_end:
; if (l->position.v[3] == 0) { ; specular light
; /* light at infinity */ cmp dword[edx+offs_cont_local_light_model],0 ;if (c.local_light_model)
; d.X=l->position.v[0]; je .els_2
; d.Y=l->position.v[1]; mov eax,[esi+offs_vert_ec]
; d.Z=l->position.v[2]; mov [vcoord],eax ;vcoord.X=v.ec.X
; att=1; mov eax,[esi+offs_vert_ec+offs_Y]
; } else { mov [vcoord+offs_Y],eax ;vcoord.Y=v.ec.Y
; /* distance attenuation */ mov eax,[esi+offs_vert_ec+offs_Z]
; d.X=l->position.v[0]-v->ec.v[0]; mov [vcoord+offs_Z],eax ;vcoord.Z=v.ec.Z
; d.Y=l->position.v[1]-v->ec.v[1]; mov eax,ebp
; d.Z=l->position.v[2]-v->ec.v[2]; sub eax,12 ;eax = &vcoord
; dist=sqrt(d.X*d.X+d.Y*d.Y+d.Z*d.Z); stdcall gl_V3_Norm, eax
; if (dist>1E-3) { fld dword[d]
; tmp=1/dist; fsub dword[vcoord]
; d.X*=tmp; fstp dword[s] ;s.X=d.X-vcoord.X
; d.Y*=tmp; fld dword[d+offs_Y]
; d.Z*=tmp; fsub dword[vcoord+offs_Y]
; } fstp dword[s+offs_Y] ;s.Y=d.Y-vcoord.Y
; att=1.0f/(l->attenuation[0]+dist*(l->attenuation[1]+ fld dword[d+offs_Z]
; dist*l->attenuation[2])); fsub dword[vcoord+offs_Z]
; } fstp dword[s+offs_Z] ;s.Z=d.Z-vcoord.Z
; dot=d.X*n.X+d.Y*n.Y+d.Z*n.Z; jmp .els_2_end
; if (twoside && dot < 0) dot = -dot; .els_2:
; if (dot>0) { mov eax,[d]
; /* diffuse light */ mov [s],eax ;s.X=d.X
; lR+=dot * l->diffuse.v[0] * m->diffuse.v[0]; mov eax,[d+offs_Y]
; lG+=dot * l->diffuse.v[1] * m->diffuse.v[1]; mov [s+offs_Y],eax ;s.Y=d.Y
; lB+=dot * l->diffuse.v[2] * m->diffuse.v[2]; fld1
; fadd dword[d+offs_Z]
; /* spot light */ fstp dword[s+offs_Z] ;s.Z=d.Z+1.0
; if (l->spot_cutoff != 180) { .els_2_end:
; dot_spot=-(d.X*l->norm_spot_direction.v[0]+ fld dword[n]
; d.Y*l->norm_spot_direction.v[1]+ fmul st0,st0
; d.Z*l->norm_spot_direction.v[2]); fld dword[n+offs_Y]
; if (twoside && dot_spot < 0) dot_spot = -dot_spot; fmul st0,st0
; if (dot_spot < l->cos_spot_cutoff) { faddp
; /* no contribution */ fld dword[n+offs_Z]
; continue; fmul st0,st0
; } else { faddp ;dot_spec = n.X^2 +n.Y^2 +n.Z^2
; /* TODO: optimize */ cmp dword[twoside],0 ;if (twoside && dot_spec < 0)
; if (l->spot_exponent > 0) { je @f
; att=att*pow(dot_spot,l->spot_exponent); ftst ;if (dot_spec < 0)
; } fstsw ax
; } sahf
; } jae @f
fchs ;dot_spec = -dot_spec
@@:
ftst ;if (dot_spec > 0)
fstsw ax
sahf
jae .if0_end
fld dword[s]
fmul st0,st0
fld dword[s+offs_Y]
fmul st0,st0
faddp
fld dword[s+offs_Z]
fmul st0,st0
faddp
fsqrt
fcom dword[fl_1e_3]
fstsw ax
sahf
jbe @f ;if (tmp > 1.0e-3)
fdiv st1,st0 ;dot_spec /= tmp
@@:
ffree st0 ;tmp
fincstp
; /* specular light */ ; TODO: optimize
; testing specular buffer code
; dot_spec= pow(dot_spec,m.shininess)
stdcall specbuf_get_buffer, edx, dword[ecx+offs_mate_shininess_i], dword[ecx+offs_mate_shininess]
;eax = specbuf
mov dword[idx],SPECULAR_BUFFER_SIZE
fild dword[idx]
fld1
fcomp
fstsw ax
sahf
jae @f
fmul st0,st1
@@:
fistp dword[idx] ;if (dot_spec < 1.0) idx = (int)(dot_spec*SPECULAR_BUFFER_SIZE)
;else idx = SPECULAR_BUFFER_SIZE;
ffree st0 ;dot_spec
fincstp
shl dword[idx],2
add eax,dword[idx]
fld dword[eax+offs_spec_buf] ;dot_spec = specbuf.buf[idx]
fld dword[ebx+offs_ligh_specular]
fmul st0,st1
fmul dword[ecx+offs_mate_specular]
fadd dword[lR]
fstp dword[lR] ;lR+=dot_spec * l.specular.v[0] * m.specular.v[0]
fld dword[ebx+offs_ligh_specular+offs_Y]
fmul st0,st1
fmul dword[ecx+offs_mate_specular+offs_Y]
fadd dword[lG]
fstp dword[lG] ;lG+=dot_spec * l.specular.v[1] * m.specular.v[1]
fld dword[ebx+offs_ligh_specular+offs_Z]
fmul st0,st1
fmul dword[ecx+offs_mate_specular+offs_Z]
fadd dword[lB]
fstp dword[lB] ;lB+=dot_spec * l.specular.v[2] * m.specular.v[2]
ffree st0 ;dot_spec
fincstp
jmp .if2_end
.if0_end:
ffree st0 ;dot [or] dot_spec
fincstp
.if2_end:
; if (c->local_light_model) { fld dword[att]
; V3 vcoord; fld dword[lR]
; vcoord.X=v->ec.X; fmul st0,st1
; vcoord.Y=v->ec.Y; fadd dword[R]
; vcoord.Z=v->ec.Z; fstp dword[R] ;R += att * lR
; gl_V3_Norm(&vcoord); fld dword[lG]
; s.X=d.X-vcoord.X; fmul st0,st1
; s.Y=d.Y-vcoord.X; fadd dword[G]
; s.Z=d.Z-vcoord.X; fstp dword[G] ;G += att * lG
; } else { fld dword[lB]
; s.X=d.X; fmul st0,st1
; s.Y=d.Y; fadd dword[B]
; s.Z=d.Z+1.0; fstp dword[B] ;B += att * lB
; } ffree st0 ;att
; dot_spec=n.X*s.X+n.Y*s.Y+n.Z*s.Z; fincstp
; if (twoside && dot_spec < 0) dot_spec = -dot_spec; mov ebx,[ebx+offs_ligh_next]
; if (dot_spec>0) { jmp .cycle_0
; GLSpecBuf *specbuf; .cycle_0_end:
; int idx;
; tmp=sqrt(s.X*s.X+s.Y*s.Y+s.Z*s.Z);
; if (tmp > 1E-3) {
; dot_spec=dot_spec / tmp;
; }
; /* TODO: optimize */ clampf [R],0,1
; /* testing specular buffer code */ mov [esi+offs_vert_color],eax ;v.color.v[0]=clampf(R,0,1)
; /* dot_spec= pow(dot_spec,m->shininess);*/ clampf [G],0,1
; specbuf = specbuf_get_buffer(c, m->shininess_i, m->shininess); mov [esi+offs_vert_color+4],eax ;v.color.v[1]=clampf(G,0,1)
; idx = (int)(dot_spec*SPECULAR_BUFFER_SIZE); clampf [B],0,1
; if (idx > SPECULAR_BUFFER_SIZE) idx = SPECULAR_BUFFER_SIZE; mov [esi+offs_vert_color+8],eax ;v.color.v[2]=clampf(B,0,1)
; dot_spec = specbuf->buf[idx]; mov eax,[A]
; lR+=dot_spec * l->specular.v[0] * m->specular.v[0]; mov [esi+offs_vert_color+12],eax ;v.color.v[3]=A
; lG+=dot_spec * l->specular.v[1] * m->specular.v[1]; popad
; lB+=dot_spec * l->specular.v[2] * m->specular.v[2];
; }
; }
; R+=att * lR;
; G+=att * lG;
; B+=att * lB;
; }
; v->color.v[0]=clampf(R,0,1);
; v->color.v[1]=clampf(G,0,1);
; v->color.v[2]=clampf(B,0,1);
; v->color.v[3]=A;
ret ret
endp endp

View File

@ -186,7 +186,8 @@ proc glopEnableDisable uses eax ebx ecx, context:dword, p:dword
jl .els_0 jl .els_0
cmp ebx,GL_LIGHT0+MAX_LIGHTS cmp ebx,GL_LIGHT0+MAX_LIGHTS
jge .els_0 ;if (GL_LIGHT0 <= ebx < GL_LIGHT0+MAX_LIGHTS) jge .els_0 ;if (GL_LIGHT0 <= ebx < GL_LIGHT0+MAX_LIGHTS)
stdcall gl_enable_disable_light, eax,ebx-GL_LIGHT0,ecx sub ebx,GL_LIGHT0
stdcall gl_enable_disable_light, eax,ebx,ecx
jmp .end_f jmp .end_f
.els_0: .els_0:
;//fprintf(stderr,"glEnableDisable: 0x%X not supported.\n",code); ;//fprintf(stderr,"glEnableDisable: 0x%X not supported.\n",code);

View File

@ -0,0 +1,111 @@
align 4
proc calc_buf uses ebx ecx, buf:dword, shininess:dword
locals
val dd ? ;float
f_inc dd ? ;float
endl
mov dword[val],0.0f
mov dword[f_inc],SPECULAR_BUFFER_SIZE
fld1
fidiv dword[f_inc]
fstp dword[f_inc] ;f_inc = 1.0f/SPECULAR_BUFFER_SIZE
xor ecx,ecx
align 4
.cycle_0: ;for (i = 0; i <= SPECULAR_BUFFER_SIZE; i++)
cmp ecx,SPECULAR_BUFFER_SIZE
jg @f
;Вычисляем x^y
fld dword[shininess]
fld dword[val]
fyl2x ;Стек FPU теперь содержит: st0=z=y*log2(x):
;Теперь считаем 2**z:
fld st0 ;Создаем еще одну копию z
frndint ;Округляем
fsubr st0,st1 ;st1=z, st0=z-trunc(z)
f2xm1 ;st1=z, st0=2**(z-trunc(z))-1
fld1
faddp ;st1=z, st0=2**(z-trunc(z))
fscale ;st1=z, st0=(2**trunc(z))*(2**(z-trunc(z)))=2**t
fxch st1
fstp st ;Результат остается на вершине стека st0
mov ebx,ecx
shl ebx,2
add ebx,offs_spec_buf
add ebx,[buf]
fstp dword[ebx] ;buf.buf[i] = pow(val, shininess)
fld dword[val]
fadd dword[f_inc]
fstp dword[val] ;val += f_inc
inc ecx
jmp .cycle_0
@@:
ret
endp
align 4
proc specbuf_get_buffer uses ebx ecx edx, context:dword, shininess_i:dword, shininess:dword
locals
found dd ? ;GLSpecBuf *
oldest dd ? ;GLSpecBuf *
endl
mov edx,[context]
mov eax,[edx+offs_cont_specbuf_first]
mov [found],eax
mov [oldest],eax
mov ebx,[shininess_i]
.cycle_0:
or eax,eax ;while (found)
jz @f
cmp [eax+offs_spec_shininess_i],ebx ;while (found.shininess_i != shininess_i)
je @f
mov ecx,[oldest]
mov ecx,[ecx+offs_spec_last_used]
cmp [eax+offs_spec_last_used],ecx ;if (found.last_used < oldest.last_used)
jge .end_0
mov [oldest],eax ;oldest = found
.end_0:
mov eax,[eax+offs_spec_next] ;found = found.next
jmp .cycle_0
@@:
cmp dword[found],0 ;if (found) /* hey, found one! */
je @f
mov eax,[found]
mov ecx,[edx+offs_cont_specbuf_used_counter]
mov [eax+offs_spec_last_used],ecx ;found.last_used = context.specbuf_used_counter
inc dword[edx+offs_cont_specbuf_used_counter]
jmp .end_f ;return found
@@:
cmp dword[oldest],0 ;if (oldest == NULL || context.specbuf_num_buffers < MAX_SPECULAR_BUFFERS)
je @f
cmp dword[edx+offs_cont_specbuf_num_buffers],MAX_SPECULAR_BUFFERS
jl @f
jmp .end_1
@@:
; create new buffer
stdcall gl_malloc, sizeof.GLSpecBuf
;if (!eax) gl_fatal_error("could not allocate specular buffer")
inc dword[edx+offs_cont_specbuf_num_buffers]
mov ecx,[edx+offs_cont_specbuf_first]
mov [eax+offs_spec_next],ecx
mov [edx+offs_cont_specbuf_first],eax
mov ecx,[edx+offs_cont_specbuf_used_counter]
mov [eax+offs_spec_last_used],ecx
inc dword[edx+offs_cont_specbuf_used_counter]
mov [eax+offs_spec_shininess_i],ebx
stdcall calc_buf, eax,dword[shininess]
jmp .end_f
.end_1:
; overwrite the lru buffer
;tgl_trace("overwriting spec buffer :(\n");
mov eax,[oldest]
mov [eax+offs_spec_shininess_i],ebx
mov ecx,[edx+offs_cont_specbuf_used_counter]
mov [eax+offs_spec_last_used],ecx
inc dword[edx+offs_cont_specbuf_used_counter]
stdcall calc_buf, eax,dword[shininess]
.end_f:
ret
endp

View File

@ -29,6 +29,7 @@ include 'ztriangle.asm'
;include 'image_util.asm' ;include 'image_util.asm'
;include 'msghandling.asm' ;include 'msghandling.asm'
include 'arrays.asm' include 'arrays.asm'
include 'specbuf.asm'
include 'kosgl.asm' include 'kosgl.asm'
include 'glu.asm' include 'glu.asm'

View File

@ -137,11 +137,25 @@ endl
cmp dword[edx+offs_cont_lighting_enabled],0 ;if(context.lighting_enabled) cmp dword[edx+offs_cont_lighting_enabled],0 ;if(context.lighting_enabled)
je @f je @f
if DEBUG ;context.matrix_stack_ptr[0]
stdcall gl_print_matrix,dword[edx+offs_cont_matrix_stack_ptr],4
end if
; precompute inverse modelview ; precompute inverse modelview
mov ebx,ebp mov ebx,ebp
sub ebx,64 sub ebx,64
stdcall gl_M4_Inv, ebx, edx+offs_cont_matrix_stack_ptr stdcall gl_M4_Inv, ebx,dword[edx+offs_cont_matrix_stack_ptr]
stdcall gl_M4_Transpose, edx+offs_cont_matrix_model_view_inv, ebx if DEBUG ;tmp
stdcall dbg_print,txt_sp,txt_nl
stdcall gl_print_matrix,ebx,4
end if
push ebx
mov ebx,edx
add ebx,offs_cont_matrix_model_view_inv
stdcall gl_M4_Transpose, ebx
if DEBUG ;context.matrix_model_view_inv
stdcall dbg_print,txt_sp,txt_nl
stdcall gl_print_matrix,ebx,4
end if
jmp .end_if_0 jmp .end_if_0
@@: @@:
mov ecx,edx mov ecx,edx
@ -296,7 +310,7 @@ pushad
add edi,offs_cont_current_normal add edi,offs_cont_current_normal
mov edx,[v] mov edx,[v]
fld dword[edi+offs_X] fld dword[edi] ;edi = &n
fld dword[edi+offs_Y] fld dword[edi+offs_Y]
fld dword[edi+offs_Z] fld dword[edi+offs_Z]
@ -458,7 +472,6 @@ pushad
stdcall gl_vertex_transform, edx, ebx stdcall gl_vertex_transform, edx, ebx
; color ; color
cmp dword[edx+offs_cont_lighting_enabled],0 cmp dword[edx+offs_cont_lighting_enabled],0
je .els_0 je .els_0
stdcall gl_shade_vertex, edx,ebx stdcall gl_shade_vertex, edx,ebx

View File

@ -47,6 +47,11 @@ struct GLSpecBuf
next dd ? ;struct GLSpecBuf* next dd ? ;struct GLSpecBuf*
ends ends
offs_spec_shininess_i equ 0
offs_spec_last_used equ 4
offs_spec_buf equ 8
offs_spec_next equ 8+4*(SPECULAR_BUFFER_SIZE+1)
struct GLLight struct GLLight
ambient V4 ambient V4
diffuse V4 diffuse V4

View File

@ -70,7 +70,6 @@ pushad
.cycle_0: ;i .cycle_0: ;i
xor ebx,ebx xor ebx,ebx
.cycle_1: ;j .cycle_1: ;j
finit
fldz ;sum=0 fldz ;sum=0
xor ecx,ecx xor ecx,ecx
M4_reg edi,[a],eax,0 M4_reg edi,[a],eax,0
@ -79,7 +78,7 @@ pushad
add edi,4 add edi,4
M4_reg esi,[b],ecx,ebx M4_reg esi,[b],ecx,ebx
fmul dword[esi] fmul dword[esi]
fadd st0,st1 ;sum += a[i][k] * b[k][j] faddp ;sum += a[i][k] * b[k][j]
inc ecx inc ecx
cmp ecx,4 cmp ecx,4
jl .cycle_2 jl .cycle_2
@ -91,7 +90,6 @@ pushad
inc eax inc eax
cmp eax,4 cmp eax,4
jl .cycle_0 jl .cycle_0
finit
if DEBUG ;gl_M4_Mul if DEBUG ;gl_M4_Mul
stdcall dbg_print,f_m4m,txt_nl stdcall dbg_print,f_m4m,txt_nl
stdcall gl_print_matrix,[c],4 stdcall gl_print_matrix,[c],4
@ -187,13 +185,42 @@ endp
; a->Z=b->m[2][0]*c->X+b->m[2][1]*c->Y+b->m[2][2]*c->Z; ; a->Z=b->m[2][0]*c->X+b->m[2][1]*c->Y+b->m[2][2]*c->Z;
;} ;}
;void gl_M4_MulV4(V4 *a,M4 *b,V4 *c) align 4
;{ proc gl_M4_MulV4 uses ebx ecx edx, a:dword, b:dword, c:dword ;V4 *a, M4 *b, V4 *c
; a->X=b->m[0][0]*c->X+b->m[0][1]*c->Y+b->m[0][2]*c->Z+b->m[0][3]*c->W; mov ebx,[b]
; a->Y=b->m[1][0]*c->X+b->m[1][1]*c->Y+b->m[1][2]*c->Z+b->m[1][3]*c->W; mov edx,[c]
; a->Z=b->m[2][0]*c->X+b->m[2][1]*c->Y+b->m[2][2]*c->Z+b->m[2][3]*c->W; fld dword[edx]
; a->W=b->m[3][0]*c->X+b->m[3][1]*c->Y+b->m[3][2]*c->Z+b->m[3][3]*c->W; fld dword[edx+4]
;} fld dword[edx+8]
fld dword[edx+12]
mov edx,[a]
mov ecx,4
.cycle_1:
fld dword[ebx] ;st0 = m[_][0]
fmul st0,st4 ;st0 *= c.X
fld dword[ebx+4] ;st0 = m[_][1]
fmul st0,st4 ;st0 *= c.Y
faddp
fld dword[ebx+8] ;st0 = m[_][2]
fmul st0,st3 ;st0 *= c.Z
faddp
fld dword[ebx+12] ;st0 += m[_][3]
fmul st0,st2 ;st0 *= c.Z
faddp
fstp dword[edx] ;a.X = b.m[_][0]*c.X +b.m[_][1]*c.Y +b.m[_][2]*c.Z +b.m[_][3]*c.W
add ebx,16 ;ñëåäóùàÿ ñòðîêà ìàòðèöû
add edx,4 ;ñëåäóùàÿ êîîðäèíàòà âåêòîðà
loop .cycle_1
ffree st0
fincstp
ffree st0
fincstp
ffree st0
fincstp
ffree st0
fincstp
ret
endp
; transposition of a 4x4 matrix ; transposition of a 4x4 matrix
align 4 align 4
@ -258,58 +285,219 @@ endp
; Note : m is destroyed ; Note : m is destroyed
align 4 align 4
proc Matrix_Inv uses ecx, r:dword, m:dword, n:dword ;(float *r,float *m,int n) proc Matrix_Inv uses ebx ecx edx edi esi, r:dword, m:dword, n:dword ;(float *r,float *m,int n)
; int i,j,k,l; locals
; float max,tmp,t; max dd ? ;float
tmp dd ?
endl
; /* identitée dans r */ ; identitée dans r
; for(i=0;i<n*n;i++) r[i]=0; mov eax,0.0
; for(i=0;i<n;i++) r[i*n+i]=1; mov ecx,[n]
imul ecx,ecx
mov edi,[r]
rep stosd ;for(i=0;i<n*n;i++) r[i]=0
mov eax,1.0
xor ebx,ebx
mov edi,[r]
mov ecx,[n]
shl ecx,2
@@: ;for(i=0;i<n;i++)
cmp ebx,[n]
jge .end_0
stosd ;r[i*n+i]=1
add edi,ecx
inc ebx
jmp @b
.end_0:
; for(j=0;j<n;j++) { ; ebx -> n
; ecx -> j
; edx -> k
; edi -> i
; esi -> l
mov ebx,[n]
xor ecx,ecx
.cycle_0: ;for(j=0;j<n;j++)
cmp ecx,ebx
jge .cycle_0_end
; recherche du nombre de plus grand module sur la colonne j
mov eax,ecx
imul eax,ebx
add eax,ecx
shl eax,2
add eax,[m]
mov eax,[eax]
mov [max],eax ;max=m[j*n+j]
mov edx,ecx ;k=j
mov edi,ecx
inc edi
.cycle_1: ;for(i=j+1;i<n;i++)
cmp edi,ebx
jge .cycle_1_end
mov eax,edi
imul eax,ebx
add eax,ecx
shl eax,2
add eax,[m]
fld dword[eax]
fcom dword[max] ;if (fabs(m[i*n+j])>fabs(max))
fstsw ax
sahf
jbe @f
mov edx,edi ;k=i
fst dword[max]
@@:
ffree st0
fincstp
inc edi
jmp .cycle_1
.cycle_1_end:
; /* recherche du nombre de plus grand module sur la colonne j */ ; non intersible matrix
; max=m[j*n+j]; fld dword[max]
; k=j; ftst ;if (max==0)
; for(i=j+1;i<n;i++) fstsw ax
; if (fabs(m[i*n+j])>fabs(max)) { ffree st0
; k=i; fincstp
; max=m[i*n+j]; sahf
; } jne @f
xor eax,eax
inc eax
jmp .end_f ;return 1
@@:
; /* non intersible matrix */ ; permutation des lignes j et k
; if (max==0) return 1; cmp ecx,edx ;if (j!=k)
je .cycle_2_end
xor edi,edi
.cycle_2: ;for(i=0;i<n;i++)
cmp edi,ebx
jge .cycle_2_end
;òóò ïîêà esi != l
mov eax,ecx
imul eax,ebx
add eax,edi
shl eax,2
add eax,[m]
mov esi,[eax]
mov [tmp],esi ;tmp=m[j*n+i]
mov esi,edx
imul esi,ebx
add esi,edi
shl esi,2
add esi,[m]
m2m dword[eax],dword[esi] ;m[j*n+i]=m[k*n+i]
mov eax,[tmp]
mov [esi],eax ;m[k*n+i]=tmp
; /* permutation des lignes j et k */ mov eax,ecx
; if (k!=j) { imul eax,ebx
; for(i=0;i<n;i++) { add eax,edi
; tmp=m[j*n+i]; shl eax,2
; m[j*n+i]=m[k*n+i]; add eax,[r]
; m[k*n+i]=tmp; mov esi,[eax]
; mov [tmp],esi ;tmp=r[j*n+i]
; tmp=r[j*n+i]; mov esi,edx
; r[j*n+i]=r[k*n+i]; imul esi,ebx
; r[k*n+i]=tmp; add esi,edi
; } shl esi,2
; } add esi,[r]
m2m dword[eax],dword[esi] ;r[j*n+i]=r[k*n+i]
mov eax,[tmp]
mov [esi],eax ;r[k*n+i]=tmp
inc edi
jmp .cycle_2
.cycle_2_end:
; /* multiplication de la ligne j par 1/max */ ; multiplication de la ligne j par 1/max
; max=1/max; fld1
; for(i=0;i<n;i++) { fdiv dword[max]
; m[j*n+i]*=max; fst dword[max] ;max=1/max
; r[j*n+i]*=max; xor edi,edi
; } mov eax,ecx
imul eax,ebx
shl eax,2
.cycle_3: ;for(i=0;i<n;i++)
cmp edi,ebx
jge .cycle_3_end
add eax,[m]
fld dword[eax]
fmul st0,st1
fstp dword[eax] ;m[j*n+i]*=max
sub eax,[m]
add eax,[r]
fld dword[eax]
fmul st0,st1
fstp dword[eax] ;r[j*n+i]*=max
sub eax,[r]
add eax,4
inc edi
jmp .cycle_3
.cycle_3_end:
ffree st0 ;max
fincstp
; for(l=0;l<n;l++) if (l!=j) { xor esi,esi
; t=m[l*n+j]; .cycle_4: ;for(l=0;l<n;l++)
; for(i=0;i<n;i++) { cmp esi,ebx
; m[l*n+i]-=m[j*n+i]*t; jge .cycle_4_end
; r[l*n+i]-=r[j*n+i]*t; cmp esi,ecx ;if (l!=j)
; } je .cycle_5_end
; } mov eax,esi
; } imul eax,ebx
add eax,ecx
shl eax,2
add eax,[m]
fld dword[eax] ;t=m[l*n+j]
xor edi,edi
.cycle_5: ;for(i=0;i<n;i++)
cmp edi,ebx
jge .cycle_5_end
mov eax,ecx
imul eax,ebx
add eax,edi
shl eax,2
add eax,[m]
fld dword[eax]
fmul st0,st1
mov eax,esi
imul eax,ebx
add eax,edi
shl eax,2
add eax,[m]
fsub dword[eax]
fchs
fstp dword[eax] ;m[l*n+i]-=m[j*n+i]*t
mov eax,ecx
imul eax,ebx
add eax,edi
shl eax,2
add eax,[r]
fld dword[eax]
fmul st0,st1
mov eax,esi
imul eax,ebx
add eax,edi
shl eax,2
add eax,[r]
fsub dword[eax]
fchs
fstp dword[eax] ;r[l*n+i]-=r[j*n+i]*t
inc edi
jmp .cycle_5
.cycle_5_end:
ffree st0 ;t
fincstp
inc esi
jmp .cycle_4
.cycle_4_end:
inc ecx
jmp .cycle_0
.cycle_0_end:
; return 0; xor eax,eax ;return 0
.end_f:
ret ret
endp endp
@ -409,12 +597,12 @@ align 4
proc gl_V3_Norm uses ebx, a:dword proc gl_V3_Norm uses ebx, a:dword
mov ebx,[a] mov ebx,[a]
fld dword[ebx] fld dword[ebx]
fmul dword[ebx] fmul st0,st0
fld dword[ebx+4] fld dword[ebx+4]
fmul dword[ebx+4] fmul st0,st0
faddp faddp
fld dword[ebx+8] fld dword[ebx+8]
fmul dword[ebx+8] fmul st0,st0
faddp faddp
fsqrt ;st0 = sqrt(a.X^2 +a.Y^2 +a.Z^2) fsqrt ;st0 = sqrt(a.X^2 +a.Y^2 +a.Z^2)
fldz fldz