sms, basic gamegear support

This commit is contained in:
kub 2021-10-13 21:30:54 +02:00
parent 0df7401c02
commit 466fa07953
16 changed files with 451 additions and 214 deletions

View file

@ -1642,27 +1642,42 @@ FinalizeLine555:
add r3, r10, #OFS_EST_HighPal
mov lr, #0xff
ldr r1, [r10, #OFS_EST_HighCol]
ldr r0, [r10, #OFS_EST_DrawLineDest]
add r1, r1, #8
ldrb r12, [r8, #OFS_Pico_video_reg+12]
mov lr, lr, lsl #1
tst r12, #1
movne r2, #320/8 @ len
bne .fl_no32colRGB555
ldr r5, [r10, #OFS_EST_PicoOpt]
mov r2, #256/8
ldr r1, [r10, #OFS_EST_HighCol]
ldr r0, [r10, #OFS_EST_DrawLineDest]
ldr r4, [r5]
tst r4, #0x4000
bne .fl_32scale_RGB555
tst r4, #0x0100
addeq r0, r0, #32*2
ldr r7, [r5, #OFS_PicoIn_AHW-OFS_PicoIn_opt]
ldrb r12,[r8, #OFS_Pico_video_reg+12]
ldr r2, [r8, #OFS_Pico_m_hardware]
add r1, r1, #8
tst r7, #0x10
beq .fl_no20colRGB555
and r7, r2, #0x3
cmp r7, #0x3 @ Game Gear, LCD?
bne .fl_no20colRGB555
mov r2, #160/8 @ len = 160
tst r4, #0x4000 @ EN_SOFTSCALE?
bne .fl_20scale_RGB555 @ scale 160->320
beq .fl_checkborder
.fl_no20colRGB555:
tst r12, #1 @ h32?
movne r2, #320/8 @ len = 320
bne .fl_no32colRGB555
moveq r2, #256/8 @ len = 256
tst r4, #0x4000 @ EN_SOFTSCALE?
bne .fl_32scale_RGB555 @ scale 256->320
.fl_checkborder:
tst r4, #0x0100 @ DIS_32C_BORDER?
rsbeq r4, r2, #320/8 @ pd += (320-len)/2
addeq r0, r0, r4, lsl #3
.fl_no32colRGB555:
#ifdef UNALIGNED_DRAWLINEDEST
@ this is basically for Gizmondo, which has unaligned odd lines in the framebuffer
tst r0, #2
@ -1715,9 +1730,9 @@ FinalizeLine555:
bne .fl_32scale_RGB555u
#endif
ands r5, r5, #0x3
addne pc, pc, r5, lsl #2
b .fl_32scale_nn
and r5, r5, #0x3
add pc, pc, r5, lsl #2
nop
b .fl_32scale_nn
b .fl_32scale_snn
b .fl_32scale_bl2
@ -1866,6 +1881,9 @@ FinalizeLine555:
ldmfd sp!, {r4-r10,pc}
.fl_32scale_bl4:
// TODO this should reflect the bl4 C algorithm, but it doesn't, it's bln.
and r9, r9, r9, lsl #1 @ nuke 2 LSBs to avoid spilling for n/4
.fl_32loop_bl4:
ldr r12, [r1], #4
ldr r7, [r1], #4
@ -1926,10 +1944,147 @@ FinalizeLine555:
subs r2, r2, #1
stmia r0!, {r4,r5,r6,r8,r10}
bne .fl_32scale_bl4
bne .fl_32loop_bl4
ldmfd sp!, {r4-r10,pc}
.fl_20scale_RGB555:
ldr r5, [r5, #OFS_PicoIn_filter-OFS_PicoIn_opt]
mov r9, #0xf700 @ f800 07e0 001f | e000 0780 001c | 3800 01e0 0007
orr r9, r9, #0x00de
#ifdef UNALIGNED_DRAWLINEDEST
tst r0, #2
bne .fl_20scale_RGB555u
#endif
and r5, r5, #0x2
add pc, pc, r5, lsl #1
nop
b .fl_20scale_nn
b .fl_20scale_bl2
.fl_20scale_nn:
ldr r12, [r1], #4
ldr r7, [r1], #4
and r4, lr, r12, lsl #1
ldrh r4, [r3, r4]
and r5, lr, r12, lsr #7
ldrh r5, [r3, r5]
and r6, lr, r12, lsr #15
ldrh r6, [r3, r6]
and r8 ,lr, r12, lsr #23
ldrh r8 ,[r3, r8 ]
orr r4, r4, r4, lsl #16
orr r5, r5, r5, lsl #16
orr r6, r6, r6, lsl #16
orr r8, r8, r8, lsl #16
stmia r0!, {r4,r5,r6,r8}
and r4, lr, r7, lsl #1
ldrh r4, [r3, r4]
and r5, lr, r7, lsr #7
ldrh r5, [r3, r5]
and r6 ,lr, r7, lsr #15
ldrh r6 ,[r3, r6]
and r8, lr, r7, lsr #23
ldrh r8, [r3, r8]
orr r4, r4, r4, lsl #16
orr r5, r5, r5, lsl #16
orr r6, r6, r6, lsl #16
orr r8, r8, r8, lsl #16
stmia r0!, {r4,r5,r6,r8}
subs r2, r2, #1
bne .fl_20scale_nn
ldmfd sp!, {r4-r10,pc}
.fl_20scale_bl2:
ldr r8, [r1]
and r8, lr, r8, lsl #1
ldrh r8, [r3, r8]
and r8, r8, r9
lsl r8, r8, #16
.fl_20loop_bl2:
ldr r12, [r1], #4
ldr r7, [r1], #4
and r4, lr, r12, lsl #1
ldrh r4, [r3, r4]
and r5, lr, r12, lsr #7
ldrh r5, [r3, r5]
and r6, lr, r12, lsr #15
ldrh r6, [r3, r6]
and r4, r4, r9
add r10,r4, r8, lsr #16
mov r10,r10,lsr #1
orr r4, r10,r4, lsl #16 @ (px-1+px0)/2 | px0
and r8 ,lr, r12, lsr #23
ldrh r8 ,[r3, r8]
and r5, r5, r9
add r10,r5, r4, lsr #16
mov r10,r10,lsr #1
orr r5, r10,r5, lsl #16 @ (px0 +px1)/2 | px1
and r6, r6, r9
add r10,r6, r5, lsr #16
mov r10,r10,lsr #1
orr r6, r10,r6, lsl #16 @ (px1 +px2)/2 | px2
and r8, r8, r9
add r10,r8, r6, lsr #16
mov r10,r10,lsr #1
orr r8, r10,r8, lsl #16 @ (px2 +px3)/2 | px3
stmia r0!, {r4,r5,r6,r8}
and r4, lr, r7, lsl #1
ldrh r4, [r3, r4]
and r5, lr, r7, lsr #7
ldrh r5, [r3, r5]
and r6, lr, r7, lsr #15
ldrh r6, [r3, r6]
and r4, r4, r9
add r10,r4, r8, lsr #16
mov r10,r10,lsr #1
orr r4, r10,r4, lsl #16 @ (px-1+px0)/2 | px0
and r8 ,lr, r7, lsr #23
ldrh r8 ,[r3, r8]
and r5, r5, r9
add r10,r5, r4, lsr #16
mov r10,r10,lsr #1
orr r5, r10,r5, lsl #16 @ (px0 +px1)/2 | px1
and r6, r6, r9
add r10,r6, r5, lsr #16
mov r10,r10,lsr #1
orr r6, r10,r6, lsl #16 @ (px1 +px2)/2 | px2
and r8, r8, r9
add r10,r8, r6, lsr #16
mov r10,r10,lsr #1
orr r8, r10,r8, lsl #16 @ (px2 +px3)/2 | px3
subs r2, r2, #1
stmia r0!, {r4,r5,r6,r8}
bne .fl_20loop_bl2
ldmfd sp!, {r4-r10,pc}
#ifdef UNALIGNED_DRAWLINEDEST
@ unaligned versions of loops
@ warning: starts drawing 2bytes before dst