X-Git-Url: https://notaz.gp2x.de/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=Pico%2FDraw.s;h=3a3da814d7753e105d274d853a60e59ec4046ab4;hb=1b0ac8adc97d65c686a48c4375c07907de86be24;hp=6338aabd8bce926fc4e88fe506ff026bbfcbf05e;hpb=2ec14aec8ff1068bf6265f3ae9c502bf58cfc686;p=picodrive.git

diff --git a/Pico/Draw.s b/Pico/Draw.s
index 6338aab..3a3da81 100644
--- a/Pico/Draw.s
+++ b/Pico/Draw.s
@@ -1647,8 +1647,14 @@ FinalizeLineRGB555:
     addeq   r0, r0, #32*2
 
 .fl_no32colRGB555:
-.fl_loopRGB555:
 
+.if UNALIGNED_DRAWLINEDEST
+    @ this is basically for Gizmondo, which has unaligned odd lines in the framebuffer
+    tst     r0, #2
+    bne     .fl_RGB555u
+.endif
+
+.fl_loopRGB555:
     ldr     r12, [r1], #4
     ldr     r7,  [r1], #4
 
@@ -1670,12 +1676,12 @@ FinalizeLineRGB555:
     ldrh    r6, [r3, r6]
     and     r12,lr, r7, lsr #15
     ldrh    r12,[r3, r12]
+    and     r7, lr, r7, lsr #23
+    ldrh    r7, [r3, r7]
     orr     r8, r8, r6, lsl #16
 
-    and     r6, lr, r7, lsr #23
-    ldrh    r6, [r3, r6]             @ 1 cycle interlock here (r6)
     subs    r2, r2, #1
-    orr     r12,r12, r6, lsl #16
+    orr     r12,r12, r7, lsl #16
 
     stmia   r0!, {r4,r5,r8,r12}
     bne     .fl_loopRGB555
@@ -1689,6 +1695,11 @@ FinalizeLineRGB555:
     mov     r9, #0x3900 @ f800 07e0 001f | e000 0780 001c | 3800 01e0 0007
     orr     r9, r9, #0x00e7
 
+.if UNALIGNED_DRAWLINEDEST
+    tst     r0, #2
+    bne     .fl_32scale_RGB555u
+.endif
+
 .fl_loop32scale_RGB555:
     ldr     r12, [r1], #4
     ldr     r7,  [r1], #4
@@ -1745,6 +1756,121 @@ FinalizeLineRGB555:
     ldmfd   sp!, {r4-r8,lr}
     bx      lr
 
+.if UNALIGNED_DRAWLINEDEST
+    @ unaligned versions of loops
+    @ warning: starts drawing 2bytes before dst
+
+.fl_RGB555u:
+    sub     r0, r0, #2              @ initial adjustment
+    mov     r8, #0
+
+.fl_loopRGB555u:
+    ldr     r12, [r1], #4
+    ldr     r7,  [r1], #4
+
+    and     r6, lr, r12,lsl #1
+    ldrh    r6, [r3, r6]
+    and     r5, lr, r12,lsr #7
+    ldrh    r5, [r3, r5]
+    orr     r4, r8, r6, lsl #16
+
+    and     r6, lr, r12,lsr #15
+    ldrh    r6, [r3, r6]
+    and     r8, lr, r12,lsr #23
+    ldrh    r8, [r3, r8]
+    orr     r5, r5, r6, lsl #16
+
+    and     r6, lr, r7, lsl #1
+    ldrh    r6, [r3, r6]
+    and     r12,lr, r7, lsr #7
+    ldrh    r12,[r3, r12]
+    orr     r6, r8, r6, lsl #16
+
+    and     r8, lr, r7, lsr #15
+    ldrh    r8, [r3, r8]
+    and     r7, lr, r7, lsr #23
+
+    subs    r2, r2, #1
+    orr     r12,r12,r8, lsl #16
+    ldrh    r8, [r3, r7]
+
+    stmia   r0!, {r4,r5,r6,r12}
+    bne     .fl_loopRGB555u
+
+    strh    r8, [r0], #2
+
+    ldmfd   sp!, {r4-r8,lr}
+    bx      lr
+
+
+.fl_32scale_RGB555u:
+    sub     r0, r0, #2              @ initial adjustment
+    mov     r4, #0
+
+    @ r9  f800 07e0 001f | e000 0780 001c | 3800 01e0 0007
+.fl_loop32scale_RGB555u:
+    ldr     r12, [r1], #4
+    ldr     r7,  [r1], #4
+
+    and     r6, lr, r12,lsl #1
+    ldrh    r6, [r3, r6]
+    and     r5, lr, r12,lsr #7
+    ldrh    r5, [r3, r5]
+    and     r6, r6, r9, lsl #2
+    orr     r4, r4, r6, lsl #16       @ r4 = pix_d -1, 0
+
+    and     r5, r5, r9, lsl #2
+    sub     r8, r5, r5, lsr #2        @ r8 = 3/4 pix_s 1
+    add     r6, r8, r6, lsr #2        @ r6 = (1/4 pix_s 0) + (3/4 pix_s 1)
+    orr     r5, r6, r5, lsl #15
+
+    and     r6, lr, r12,lsr #15
+    ldrh    r6, [r3, r6]
+    and     r12,lr, r12,lsr #23
+    ldrh    r12,[r3, r12]
+    and     r6, r6, r9, lsl #2
+    add     r5, r5, r6, lsl #15       @ r5 = pix_d 1, 2
+
+    and     r8, lr, r7, lsl #1
+    ldrh    r8, [r3, r8]
+    and     r10,lr, r7, lsr #7
+    ldrh    r10,[r3, r10]
+    and     r12,r12,r9, lsl #2
+    sub     r6, r6, r6, lsr #2        @ r6 = 3/4 pix_s 2
+    add     r6, r6, r12,lsr #2
+    orr     r6, r6, r12,lsl #16       @ r6 = pix_d 3, 4
+
+    and     r8, r8, r9, lsl #2
+    and     r10,r10,r9, lsl #2
+    sub     r12,r10,r10,lsr #2        @ r12 = 3/4 pix_s 5
+    orr     r8, r8, r8, lsl #14
+    add     r8, r8, r12,lsl #16       @ r8 = pix_d 5, 6
+    and     r12,lr, r7, lsr #15
+    ldrh    r12,[r3, r12]
+    and     r7, lr, r7, lsr #23
+    ldrh    r7, [r3, r7]
+    and     r12,r12,r9, lsl #2
+    add     r10,r10,r12
+    mov     r10,r10,    lsr #1
+    sub     r12,r12,r12,lsr #2        @ r12 = 3/4 pix_s 6
+    orr     r10,r10,r12,lsl #16
+    and     r7, r7, r9, lsl #2
+    add     r10,r10,r7, lsl #14       @ r10 = pix_d 7, 8
+
+    subs    r2, r2, #1
+
+    stmia   r0!, {r4,r5,r6,r8,r10}
+    mov     r4, r7
+    bne     .fl_loop32scale_RGB555u
+
+    strh    r4, [r0], #2
+
+    ldmfd   sp!, {r9,r10}
+    ldmfd   sp!, {r4-r8,lr}
+    bx      lr
+
+.endif @ UNALIGNED_DRAWLINEDEST
+
 
 @ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@