forked from KolibriOS/kolibrios
330 lines
7.4 KiB
Plaintext
330 lines
7.4 KiB
Plaintext
ROUND2 equ 10
|
|
|
|
stencil_tri:
|
|
; procedure calculate triangle in stencil buffer
|
|
; ----------------in - eax - x1 shl 16 + y1 ----------------------
|
|
; -------------------- ebx - x2 shl 16 + y2 ----------------------
|
|
; -------------------- ecx - x3 shl 16 + y3 ----------------------
|
|
; -------------------- esi - pointer to s-buffer -----------------
|
|
; -------------------- xmm0 - lo -> hi z1, z2, z3 as dword float
|
|
|
|
.x1 equ [ebp-2]
|
|
.y1 equ [ebp-4]
|
|
.x2 equ [ebp-6]
|
|
.y2 equ [ebp-8]
|
|
.x3 equ [ebp-10]
|
|
.y3 equ [ebp-12]
|
|
|
|
.dx12 equ dword[ebp-20]
|
|
.dx13 equ dword[ebp-24]
|
|
.dx23 equ dword[ebp-28]
|
|
.dz12 equ dword[ebp-32]
|
|
.dz13 equ dword[ebp-36]
|
|
.dz23 equ dword[ebp-40]
|
|
.zz2 equ [ebp-44]
|
|
.zz1 equ [ebp-48]
|
|
.z3 equ [ebp-56]
|
|
.z2 equ [ebp-60]
|
|
.z1 equ [ebp-64]
|
|
.s_buff equ [ebp-68]
|
|
|
|
push ebp
|
|
mov ebp,esp
|
|
sub esp,128
|
|
and ebp,0xfffffff0
|
|
.sort2:
|
|
cmp ax,bx
|
|
jle .sort1
|
|
xchg eax,ebx
|
|
shufps xmm0,xmm0,11100001b
|
|
.sort1:
|
|
cmp bx,cx
|
|
jle .sort3
|
|
xchg ebx,ecx
|
|
shufps xmm0,xmm0,11011000b
|
|
jmp .sort2
|
|
.sort3:
|
|
mov .y1,eax ; store triangle coordinates in user friendly variables
|
|
mov .y2,ebx
|
|
mov .y3,ecx
|
|
|
|
; mov edx,100.11
|
|
; movd xmm0,edx
|
|
; shufps xmm0,xmm0,11100000b
|
|
|
|
movaps .z1,xmm0
|
|
; mov dword .z1,edx
|
|
; mov .z2,edx
|
|
; mov .z3,edx
|
|
mov .s_buff,esi
|
|
|
|
mov edx,80008000h ; eax,ebx,ecx are ANDd together into edx which means that
|
|
and edx,ebx ; if *all* of them are negative a sign flag is raised
|
|
and edx,ecx
|
|
and edx,eax
|
|
test edx,80008000h ; Check both X&Y at once
|
|
jne .loop2_end
|
|
|
|
mov bx,.y2 ; calc delta 12
|
|
sub bx,.y1
|
|
jnz .dx12_make
|
|
mov .dx12,0
|
|
mov .dz12,0
|
|
jmp .dx12_done
|
|
.dx12_make:
|
|
mov ax,.x2
|
|
sub ax,.x1
|
|
cwde
|
|
movsx ebx,bx
|
|
shl eax,ROUND2
|
|
cdq
|
|
idiv ebx
|
|
mov .dx12,eax
|
|
|
|
movss xmm1,.z2
|
|
cvtsi2ss xmm2,ebx
|
|
subss xmm1,.z1
|
|
divss xmm1,xmm2
|
|
movss .dz12,xmm1
|
|
; mov .dz12, dword 0.11
|
|
|
|
.dx12_done:
|
|
mov bx,.y3 ; calc delta 13
|
|
sub bx,.y1
|
|
jnz .dx13_make
|
|
mov .dx13,0
|
|
mov .dz13,0
|
|
jmp .dx13_done
|
|
.dx13_make:
|
|
mov ax,.x3
|
|
sub ax,.x1
|
|
cwde
|
|
movsx ebx,bx
|
|
shl eax,ROUND2
|
|
cdq
|
|
idiv ebx
|
|
mov .dx13,eax
|
|
|
|
movss xmm1,.z3
|
|
cvtsi2ss xmm2,ebx
|
|
subss xmm1,.z1
|
|
divss xmm1,xmm2
|
|
movss .dz13,xmm1
|
|
; mov .dz13, dword 0.11
|
|
|
|
.dx13_done:
|
|
mov bx,.y3 ; calc delta 23
|
|
sub bx,.y2
|
|
jnz .dx23_make
|
|
mov .dx23,0
|
|
mov .dz23,0
|
|
jmp .dx23_done
|
|
.dx23_make:
|
|
mov ax,.x3
|
|
sub ax,.x2
|
|
cwde
|
|
movsx ebx,bx
|
|
shl eax,ROUND2
|
|
cdq
|
|
idiv ebx
|
|
mov .dx23,eax
|
|
|
|
movss xmm1,.z3
|
|
cvtsi2ss xmm2,ebx
|
|
subss xmm1,.z2
|
|
divss xmm1,xmm2
|
|
movss .dz23,xmm1
|
|
|
|
.dx23_done:
|
|
mov edx,.z1
|
|
|
|
mov .zz1,edx
|
|
mov .zz2,edx
|
|
movsx eax,word .x1
|
|
shl eax,ROUND2 ; eax - cur x1
|
|
mov ebx,eax ; ebx - cur x2
|
|
|
|
mov cx,.y1
|
|
cmp cx,.y2
|
|
jge .loop1_end
|
|
.loop1:
|
|
|
|
pushad
|
|
sar ebx,ROUND2
|
|
sar eax,ROUND2
|
|
; movq xmm0,.zz1
|
|
movlps xmm0,.zz1
|
|
; mov edx,0.11
|
|
; movd xmm0,edx
|
|
; shufps xmm0,xmm0,11100000b
|
|
mov esi,.s_buff
|
|
|
|
call stencil_line
|
|
|
|
popad
|
|
add eax,.dx13
|
|
add ebx,.dx12
|
|
|
|
movss xmm1,.zz1
|
|
movss xmm2,.zz2
|
|
addss xmm1,.dz13
|
|
addss xmm2,.dz12
|
|
movss .zz1,xmm1
|
|
movss .zz2,xmm2
|
|
|
|
add cx,1
|
|
cmp cx,.y2
|
|
jl .loop1
|
|
|
|
.loop1_end:
|
|
|
|
mov edx,.z2
|
|
mov .zz2,edx
|
|
movsx ebx,word .x2
|
|
shl ebx,ROUND2
|
|
|
|
mov cx,.y2
|
|
cmp cx,.y3
|
|
jge .loop2_end
|
|
.loop2:
|
|
pushad
|
|
|
|
sar ebx,ROUND2
|
|
sar eax,ROUND2
|
|
movlps xmm0,.zz1
|
|
mov esi,.s_buff
|
|
|
|
|
|
call stencil_line
|
|
|
|
popad
|
|
|
|
add eax,.dx13
|
|
add ebx,.dx23
|
|
|
|
movss xmm1,.zz1
|
|
movss xmm2,.zz2
|
|
addss xmm1,.dz13
|
|
addss xmm2,.dz23
|
|
movss .zz1,xmm1
|
|
movss .zz2,xmm2
|
|
|
|
|
|
add cx,1
|
|
cmp cx,.y3
|
|
jl .loop2
|
|
.loop2_end:
|
|
|
|
add esp,128
|
|
pop ebp
|
|
|
|
ret
|
|
|
|
stencil_line:
|
|
;----------------------------------------------------
|
|
;-------------in xmm0 - lo -> hi z1, z2
|
|
;--------------- eax - x1 ---------------------------
|
|
;--------------- ebx - x2 ---------------------------
|
|
;--------------- ecx - y-----------------------------
|
|
;--------------- esi - pointer to s-buffer
|
|
|
|
|
|
.dz equ [ebp-4]
|
|
.z2 equ [ebp-8]
|
|
.z1 equ [ebp-12]
|
|
.x2 equ [ebp-16]
|
|
.x1 equ [ebp-20]
|
|
.s_buf equ [ebp-24]
|
|
; cmp eax,ebx
|
|
; je @f
|
|
; int3
|
|
; @@:
|
|
push ebp
|
|
mov ebp,esp
|
|
sub esp,64
|
|
; cmp eax,0
|
|
; jg @f
|
|
;
|
|
; @@:
|
|
or cx,cx
|
|
jle .l_quit
|
|
|
|
cmp cx,[size_y_var]
|
|
jge .l_quit
|
|
|
|
movzx ecx,cx
|
|
mov .s_buf,esi
|
|
cmp eax,ebx
|
|
je .l_quit
|
|
jl .l_ok
|
|
|
|
xchg eax,ebx
|
|
shufps xmm0,xmm0,11100001b
|
|
.l_ok:
|
|
|
|
cmp ax,[size_x_var]
|
|
jge .l_quit
|
|
cmp bx,0
|
|
jle .l_quit
|
|
|
|
movlps .z1,xmm0
|
|
mov .x1,eax
|
|
mov .x2,ebx
|
|
|
|
sub ebx,eax
|
|
movss xmm0,.z2
|
|
cvtsi2ss xmm1,ebx
|
|
subss xmm0,.z1
|
|
divss xmm0,xmm1
|
|
movss .dz,xmm0
|
|
|
|
movzx edx,word[size_x_var]
|
|
cmp eax,1
|
|
jge @f
|
|
mov eax,.x1
|
|
neg eax
|
|
cvtsi2ss xmm2,eax
|
|
mulss xmm2,.dz
|
|
addss xmm2,.z1
|
|
movss .z1,xmm2
|
|
mov dword .x1,0
|
|
movzx edx,word[size_x_var]
|
|
sub edx,1
|
|
@@:
|
|
cmp .x2,edx
|
|
jl @f
|
|
mov .x2,edx
|
|
|
|
@@:
|
|
; mov eax,.x1
|
|
; cmp .x2,eax
|
|
; je .l_quit
|
|
movzx edx,word[size_x_var]
|
|
mov esi,.s_buf
|
|
mov eax,ecx ; y
|
|
mul edx
|
|
add eax,.x1
|
|
|
|
shl eax,2
|
|
add esi,eax
|
|
|
|
mov ecx,.x2
|
|
sub ecx,.x1
|
|
movss xmm2,.z1 ; cz
|
|
.ccalc:
|
|
movss xmm1,xmm2
|
|
cmpltss xmm1,dword[esi]
|
|
movd eax,xmm1
|
|
cmp eax,-1
|
|
jnz @f
|
|
movss dword[esi],xmm2
|
|
@@:
|
|
add esi,4
|
|
addss xmm2,.dz
|
|
sub ecx,1
|
|
jnz .ccalc
|
|
.l_quit:
|
|
mov esp,ebp
|
|
pop ebp
|
|
ret
|