portrait mode asm

diff --git a/gp2x/upscale_aspect.s b/gp2x/upscale_aspect.s
index ed67c63..41bcf0e 100644 (file)
--- a/gp2x/upscale_aspect.s
+++ b/gp2x/upscale_aspect.s
@@ -316,5 +316,264 @@ loop2:
 
     ldmfd   sp!,{r4-r11,pc}
 
+
+.global upscale_aspect_row @ void *dst, void *linesx4, u32 row
+upscale_aspect_row:
+    stmfd   sp!,{r4-r11,lr}
+    mov     lr,     #0x0000001f
+    orr     lr, lr, #0x0000f800 @ for "unpacked" form of
+    orr     lr, lr, #0x07e00000        @ 00000ggg'ggg00000'rrrrr000'000bbbbb
+    mov     r12,    #0x00000001
+    orr     r12,r12,#0x00000800
+    orr     r12,r12,#0x00200000 @ rounding constant
+
+    mov     r8,     #(240/6)   @ cols
+
+    add     r0, r0, #(240*320)*2
+    add     r0, r0, #12*2
+    add     r0, r0, r2, lsl #3
+
+uar_loop:
+    ldr     r10,[r1]
+    ldr     r11,[r1, #240*2*1]
+
+    unpack_lo r4, r10
+    unpack_hi r5, r10
+    unpack_lo r6, r11
+    unpack_hi r7, r11
+
+    ldr     r11,[r1, #240*2*2]
+
+    do_3_5  r4, r6
+    orr     r2, r2, r2, lsr #16
+    mov     r3, r10,    lsl #16
+    mov     r3, r3,     lsr #16
+    orr     r2, r3, r2, lsl #16
+    str     r2, [r0, #-240*2]! @ 0,8
+
+    unpack_lo r10,r11
+    unpack_hi r9, r11
+
+    do_3_5  r10,r6
+    orr     r2, r2, r2, lsl #16
+    mov     r3, r11,    lsl #16
+    orr     r2, r3, r2, lsr #16
+    str     r2, [r0, #4]       @ 16,24
+
+    do_3_5  r4, r5
+    orr     r3, r2, r2, lsl #16
+
+    do_14_7_7_4 r7, r5, r6, r4
+    orr     r2, r2, r2, lsr #16
+    mov     r3, r3,     lsr #16
+    orr     r2, r3, r2, lsl #16
+    str     r2, [r0, #-240*2]! @ 1,9
+
+    ldr     r11,[r1, #4]
+
+    do_14_7_7_4 r7, r6, r9, r10
+    orr     r3, r2, r2, lsl #16
+
+    do_3_5  r10,r9
+    orr     r2, r2, r2, lsr #16
+    mov     r3, r3,     lsr #16
+    orr     r2, r3, r2, lsl #16
+    str     r2, [r0, #4]       @ 17,25
+
+    ldr     r10,[r1, #240*2*1+4]
+
+    unpack_lo r4, r11
+    unpack_lo r6, r10
+
+    do_3_5  r4, r5
+    orr     r3, r2, r2, lsl #16
+
+    do_14_7_7_4 r7, r5, r6, r4
+    orr     r2, r2, r2, lsr #16
+    mov     r3, r3,     lsr #16
+    orr     r2, r3, r2, lsl #16
+    str     r2, [r0, #-240*2]! @ 2,10
+
+    do_3_5  r4, r6
+
+    ldr     r4, [r1, #240*2*2+4]
+
+    orr     r2, r2, r2, lsr #16
+    mov     r3, r11,    lsl #16
+    mov     r3, r3,     lsr #16
+    orr     r2, r3, r2, lsl #16
+    str     r2, [r0, #-240*2]  @ 3,11
+
+    unpack_lo r5, r4
+
+    do_14_7_7_4 r7, r6, r9, r5
+    orr     r3, r2, r2, lsl #16
+
+    do_3_5  r5, r9
+    orr     r2, r2, r2, lsr #16
+    mov     r2, r2,     lsl #16
+    orr     r2, r2, r3, lsr #16
+    str     r2, [r0, #4]       @ 18,26
+
+    do_3_5  r5, r6
+    orr     r2, r2, r2, lsl #16
+    mov     r3, r4,     lsl #16
+    orr     r2, r3, r2, lsr #16
+    str     r2, [r0, #-240*2+4]        @ 19,27
+
+    unpack_hi r5, r11
+    unpack_hi r6, r10
+    unpack_hi r7, r4
+
+    ldr     r10,[r1, #8]
+
+    do_3_5  r5, r6
+    orr     r2, r2, r2, lsr #16
+    mov     r3, r11,    lsr #16
+    orr     r2, r3, r2, lsl #16
+    str     r2, [r0, #-240*2*2]! @ 4,12
+
+    ldr     r11,[r1, #240*2*1+8]
+
+    do_3_5  r7, r6
+    orr     r2, r2, r2, lsl #16
+    mov     r3, r4,     lsr #16
+    mov     r3, r3,     lsl #16
+    orr     r2, r3, r2, lsr #16
+    str     r2, [r0, #4]       @ 20,28
+
+    unpack_lo r4, r10
+    unpack_lo r9, r11
+
+    ldr     r11,[r1, #240*2*2+8]
+
+    do_3_5  r5, r4
+    orr     r3, r2, r2, lsl #16
+
+    do_14_7_7_4 r9, r4, r6, r5
+    orr     r2, r2, r2, lsr #16
+    mov     r2, r2,     lsl #16
+    orr     r2, r2, r3, lsr #16
+    str     r2, [r0, #-240*2]! @ 5,13
+
+    unpack_lo r5, r11
+
+    do_14_7_7_4 r9, r5, r6, r7
+    orr     r3, r2, r2, lsl #16
+
+    do_3_5  r7, r5
+    orr     r2, r2, r2, lsr #16
+    mov     r3, r3,     lsr #16
+    orr     r2, r3, r2, lsl #16
+    str     r2, [r0, #4]       @ 21,29
+
+    ldr     r7, [r1, #240*2*1+8]
+
+    unpack_hi r6, r10
+    unpack_hi r7, r7
+
+    do_3_5  r6, r4
+    orr     r3, r2, r2, lsl #16
+
+    do_14_7_7_4 r9, r4, r7, r6
+    orr     r2, r2, r2, lsr #16
+    mov     r2, r2,     lsl #16
+    orr     r2, r2, r3, lsr #16
+    str     r2, [r0, #-240*2]! @ 6,14
+
+    unpack_hi r4, r11
+
+    do_14_7_7_4 r9, r5, r7, r4
+    orr     r3, r2, r2, lsl #16
+
+    do_3_5  r4, r5
+    orr     r2, r2, r2, lsr #16
+    mov     r3, r3,     lsr #16
+    orr     r2, r3, r2, lsl #16
+    str     r2, [r0, #4]       @ 22,30
+
+    do_3_5  r6, r7
+    orr     r2, r2, r2, lsr #16
+    mov     r3, r10,    lsr #16
+    orr     r2, r3, r2, lsl #16
+    str     r2, [r0, #-240*2]! @ 7,15
+
+    do_3_5  r4, r7
+    orr     r2, r2, r2, lsl #16
+    mov     r3, r11,    lsr #16
+    mov     r3, r3,     lsl #16
+    orr     r2, r3, r2, lsr #16
+    str     r2, [r0, #4]       @ 23,31
+
+    subs    r8, r8, #1
+    add     r1, r1, #12
+    bne     uar_loop
+
+    ldmfd   sp!,{r4-r11,pc}
+
+
+@ bonus function
+
+@ input: r2-r5
+@ output: r7,r8
+@ trash: r6
+.macro rb_line_low
+    mov     r6, r2, lsl #16
+    mov     r7, r3, lsl #16
+    orr     r7, r7, r6, lsr #16
+    mov     r6, r4, lsl #16
+    mov     r8, r5, lsl #16
+    orr     r8, r8, r6, lsr #16
+.endm
+
+.macro rb_line_hi
+    mov     r6, r2, lsr #16
+    mov     r7, r3, lsr #16
+    orr     r7, r6, r7, lsl #16
+    mov     r6, r4, lsr #16
+    mov     r8, r5, lsr #16
+    orr     r8, r6, r8, lsl #16
+.endm
+
+.global do_rotated_blit @ void *dst, void *linesx4, u32 y
+do_rotated_blit:
+    stmfd   sp!,{r4-r8,lr}
+
+    add     r0, r0, #(240*320)*2
+    sub     r0, r0, #(240*40)*2
+    sub     r0, r0, #(240-40+4)*2      @ y starts from 4
+    add     r0, r0, r2, lsl #1
+
+    mov     lr, #240/4
+
+rotated_blit_loop:
+    ldr     r2, [r1, #240*0*2]
+    ldr     r3, [r1, #240*1*2]
+    ldr     r4, [r1, #240*2*2]
+    ldr     r5, [r1, #240*3*2]
+    rb_line_low
+    stmia   r0, {r7,r8}
+    sub     r0, r0, #240*2
+    rb_line_hi
+    stmia   r0, {r7,r8}
+    sub     r0, r0, #240*2
+
+    ldr     r2, [r1, #240*0*2+4]
+    ldr     r3, [r1, #240*1*2+4]
+    ldr     r4, [r1, #240*2*2+4]
+    ldr     r5, [r1, #240*3*2+4]
+    rb_line_low
+    stmia   r0, {r7,r8}
+    sub     r0, r0, #240*2
+    rb_line_hi
+    stmia   r0, {r7,r8}
+    sub     r0, r0, #240*2
+
+    subs    lr, lr, #1
+    add     r1, r1, #8
+    bne     rotated_blit_loop
+
+    ldmfd   sp!,{r4-r8,pc}
+
 @ vim:filetype=armasm