.text
+.macro call_2x_pixels type, subp
+function ff_\type\()_pixels16\subp\()_armv6, export=1
+ push {r0-r3, lr}
+ bl ff_\type\()_pixels8\subp\()_armv6
+ pop {r0-r3, lr}
+ add r0, r0, #8
+ add r1, r1, #8
+ b ff_\type\()_pixels8\subp\()_armv6
+.endfunc
+.endm
+
+call_2x_pixels avg
+call_2x_pixels put, _x2
+call_2x_pixels put, _y2
+call_2x_pixels put, _x2_no_rnd
+call_2x_pixels put, _y2_no_rnd
+
+function ff_put_pixels16_armv6, export=1
+ push {r4-r11}
+1:
+ ldr r5, [r1, #4]
+ ldr r6, [r1, #8]
+ ldr r7, [r1, #12]
+ ldr r4, [r1], r2
+ strd r6, r7, [r0, #8]
+ ldr r9, [r1, #4]
+ strd r4, r5, [r0], r2
+ ldr r10, [r1, #8]
+ ldr r11, [r1, #12]
+ ldr r8, [r1], r2
+ strd r10, r11, [r0, #8]
+ subs r3, r3, #2
+ strd r8, r9, [r0], r2
+ bne 1b
+
+ pop {r4-r11}
+ bx lr
+.endfunc
+
+function ff_put_pixels8_armv6, export=1
+ push {r4-r7}
+1:
+ ldr r5, [r1, #4]
+ ldr r4, [r1], r2
+ ldr r7, [r1, #4]
+ strd r4, r5, [r0], r2
+ ldr r6, [r1], r2
+ subs r3, r3, #2
+ strd r6, r7, [r0], r2
+ bne 1b
+
+ pop {r4-r7}
+ bx lr
+.endfunc
+
+function ff_put_pixels8_x2_armv6, export=1
+ push {r4-r11, lr}
+ mov r12, #1
+ orr r12, r12, r12, lsl #8
+ orr r12, r12, r12, lsl #16
+1:
+ ldr r4, [r1]
+ subs r3, r3, #2
+ ldr r5, [r1, #4]
+ ldr r7, [r1, #5]
+ lsr r6, r4, #8
+ ldr r8, [r1, r2]!
+ orr r6, r6, r5, lsl #24
+ ldr r9, [r1, #4]
+ ldr r11, [r1, #5]
+ lsr r10, r8, #8
+ add r1, r1, r2
+ orr r10, r10, r9, lsl #24
+ eor r14, r4, r6
+ uhadd8 r4, r4, r6
+ eor r6, r5, r7
+ uhadd8 r5, r5, r7
+ and r14, r14, r12
+ and r6, r6, r12
+ uadd8 r4, r4, r14
+ eor r14, r8, r10
+ uadd8 r5, r5, r6
+ eor r6, r9, r11
+ uhadd8 r8, r8, r10
+ and r14, r14, r12
+ uhadd8 r9, r9, r11
+ and r6, r6, r12
+ uadd8 r8, r8, r14
+ strd r4, r5, [r0], r2
+ uadd8 r9, r9, r6
+ strd r8, r9, [r0], r2
+ bne 1b
+
+ pop {r4-r11, pc}
+.endfunc
+
+function ff_put_pixels8_y2_armv6, export=1
+ push {r4-r11}
+ mov r12, #1
+ orr r12, r12, r12, lsl #8
+ orr r12, r12, r12, lsl #16
+ ldr r4, [r1]
+ ldr r5, [r1, #4]
+ ldr r6, [r1, r2]!
+ ldr r7, [r1, #4]
+1:
+ subs r3, r3, #2
+ uhadd8 r8, r4, r6
+ eor r10, r4, r6
+ uhadd8 r9, r5, r7
+ eor r11, r5, r7
+ and r10, r10, r12
+ ldr r4, [r1, r2]!
+ uadd8 r8, r8, r10
+ and r11, r11, r12
+ uadd8 r9, r9, r11
+ ldr r5, [r1, #4]
+ uhadd8 r10, r4, r6
+ eor r6, r4, r6
+ uhadd8 r11, r5, r7
+ and r6, r6, r12
+ eor r7, r5, r7
+ uadd8 r10, r10, r6
+ and r7, r7, r12
+ ldr r6, [r1, r2]!
+ uadd8 r11, r11, r7
+ strd r8, r9, [r0], r2
+ ldr r7, [r1, #4]
+ strd r10, r11, [r0], r2
+ bne 1b
+
+ pop {r4-r11}
+ bx lr
+.endfunc
+
+function ff_put_pixels8_x2_no_rnd_armv6, export=1
+ push {r4-r9, lr}
+1:
+ subs r3, r3, #2
+ ldr r4, [r1]
+ ldr r5, [r1, #4]
+ ldr r7, [r1, #5]
+ ldr r8, [r1, r2]!
+ ldr r9, [r1, #4]
+ ldr r14, [r1, #5]
+ add r1, r1, r2
+ lsr r6, r4, #8
+ orr r6, r6, r5, lsl #24
+ lsr r12, r8, #8
+ orr r12, r12, r9, lsl #24
+ uhadd8 r4, r4, r6
+ uhadd8 r5, r5, r7
+ uhadd8 r8, r8, r12
+ uhadd8 r9, r9, r14
+ stm r0, {r4,r5}
+ add r0, r0, r2
+ stm r0, {r8,r9}
+ add r0, r0, r2
+ bne 1b
+
+ pop {r4-r9, pc}
+.endfunc
+
+function ff_put_pixels8_y2_no_rnd_armv6, export=1
+ push {r4-r9, lr}
+ ldr r4, [r1]
+ ldr r5, [r1, #4]
+ ldr r6, [r1, r2]!
+ ldr r7, [r1, #4]
+1:
+ subs r3, r3, #2
+ uhadd8 r8, r4, r6
+ ldr r4, [r1, r2]!
+ uhadd8 r9, r5, r7
+ ldr r5, [r1, #4]
+ uhadd8 r12, r4, r6
+ ldr r6, [r1, r2]!
+ uhadd8 r14, r5, r7
+ ldr r7, [r1, #4]
+ stm r0, {r8,r9}
+ add r0, r0, r2
+ stm r0, {r12,r14}
+ add r0, r0, r2
+ bne 1b
+
+ pop {r4-r9, pc}
+.endfunc
+
+function ff_avg_pixels8_armv6, export=1
+ pld [r1, r2]
+ push {r4-r10, lr}
+ mov lr, #1
+ orr lr, lr, lr, lsl #8
+ orr lr, lr, lr, lsl #16
+ ldrd r4, r5, [r0]
+ ldr r10, [r1, #4]
+ ldr r9, [r1], r2
+ subs r3, r3, #2
+1:
+ pld [r1, r2]
+ eor r8, r4, r9
+ uhadd8 r4, r4, r9
+ eor r12, r5, r10
+ ldrd r6, r7, [r0, r2]
+ uhadd8 r5, r5, r10
+ and r8, r8, lr
+ ldr r10, [r1, #4]
+ and r12, r12, lr
+ uadd8 r4, r4, r8
+ ldr r9, [r1], r2
+ eor r8, r6, r9
+ uadd8 r5, r5, r12
+ pld [r1, r2, lsl #1]
+ eor r12, r7, r10
+ uhadd8 r6, r6, r9
+ strd r4, r5, [r0], r2
+ uhadd8 r7, r7, r10
+ beq 2f
+ and r8, r8, lr
+ ldrd r4, r5, [r0, r2]
+ uadd8 r6, r6, r8
+ ldr r10, [r1, #4]
+ and r12, r12, lr
+ subs r3, r3, #2
+ uadd8 r7, r7, r12
+ ldr r9, [r1], r2
+ strd r6, r7, [r0], r2
+ b 1b
+2:
+ and r8, r8, lr
+ and r12, r12, lr
+ uadd8 r6, r6, r8
+ uadd8 r7, r7, r12
+ strd r6, r7, [r0], r2
+
+ pop {r4-r10, pc}
+.endfunc
+
function ff_add_pixels_clamped_armv6, export=1
push {r4-r8,lr}
mov r3, #8
bgt 1b
pop {r4-r8,pc}
.endfunc
+
+function ff_pix_abs16_armv6, export=1
+ ldr r0, [sp]
+ push {r4-r9, lr}
+ mov r12, #0
+ mov lr, #0
+ ldm r1, {r4-r7}
+ ldr r8, [r2]
+1:
+ ldr r9, [r2, #4]
+ pld [r1, r3]
+ usada8 r12, r4, r8, r12
+ ldr r8, [r2, #8]
+ pld [r2, r3]
+ usada8 lr, r5, r9, lr
+ ldr r9, [r2, #12]
+ usada8 r12, r6, r8, r12
+ subs r0, r0, #1
+ usada8 lr, r7, r9, lr
+ beq 2f
+ add r1, r1, r3
+ ldm r1, {r4-r7}
+ add r2, r2, r3
+ ldr r8, [r2]
+ b 1b
+2:
+ add r0, r12, lr
+ pop {r4-r9, pc}
+.endfunc
+
+function ff_pix_abs16_x2_armv6, export=1
+ ldr r12, [sp]
+ push {r4-r11, lr}
+ mov r0, #0
+ mov lr, #1
+ orr lr, lr, lr, lsl #8
+ orr lr, lr, lr, lsl #16
+1:
+ ldr r8, [r2]
+ ldr r9, [r2, #4]
+ lsr r10, r8, #8
+ ldr r4, [r1]
+ lsr r6, r9, #8
+ orr r10, r10, r9, lsl #24
+ ldr r5, [r2, #8]
+ eor r11, r8, r10
+ uhadd8 r7, r8, r10
+ orr r6, r6, r5, lsl #24
+ and r11, r11, lr
+ uadd8 r7, r7, r11
+ ldr r8, [r1, #4]
+ usada8 r0, r4, r7, r0
+ eor r7, r9, r6
+ lsr r10, r5, #8
+ and r7, r7, lr
+ uhadd8 r4, r9, r6
+ ldr r6, [r2, #12]
+ uadd8 r4, r4, r7
+ pld [r1, r3]
+ orr r10, r10, r6, lsl #24
+ usada8 r0, r8, r4, r0
+ ldr r4, [r1, #8]
+ eor r11, r5, r10
+ ldrb r7, [r2, #16]
+ and r11, r11, lr
+ uhadd8 r8, r5, r10
+ ldr r5, [r1, #12]
+ uadd8 r8, r8, r11
+ pld [r2, r3]
+ lsr r10, r6, #8
+ usada8 r0, r4, r8, r0
+ orr r10, r10, r7, lsl #24
+ subs r12, r12, #1
+ eor r11, r6, r10
+ add r1, r1, r3
+ uhadd8 r9, r6, r10
+ and r11, r11, lr
+ uadd8 r9, r9, r11
+ add r2, r2, r3
+ usada8 r0, r5, r9, r0
+ bgt 1b
+
+ pop {r4-r11, pc}
+.endfunc
+
+.macro usad_y2 p0, p1, p2, p3, n0, n1, n2, n3
+ ldr \n0, [r2]
+ eor \n1, \p0, \n0
+ uhadd8 \p0, \p0, \n0
+ and \n1, \n1, lr
+ ldr \n2, [r1]
+ uadd8 \p0, \p0, \n1
+ ldr \n1, [r2, #4]
+ usada8 r0, \p0, \n2, r0
+ pld [r1, r3]
+ eor \n3, \p1, \n1
+ uhadd8 \p1, \p1, \n1
+ and \n3, \n3, lr
+ ldr \p0, [r1, #4]
+ uadd8 \p1, \p1, \n3
+ ldr \n2, [r2, #8]
+ usada8 r0, \p1, \p0, r0
+ pld [r2, r3]
+ eor \p0, \p2, \n2
+ uhadd8 \p2, \p2, \n2
+ and \p0, \p0, lr
+ ldr \p1, [r1, #8]
+ uadd8 \p2, \p2, \p0
+ ldr \n3, [r2, #12]
+ usada8 r0, \p2, \p1, r0
+ eor \p1, \p3, \n3
+ uhadd8 \p3, \p3, \n3
+ and \p1, \p1, lr
+ ldr \p0, [r1, #12]
+ uadd8 \p3, \p3, \p1
+ add r1, r1, r3
+ usada8 r0, \p3, \p0, r0
+ add r2, r2, r3
+.endm
+
+function ff_pix_abs16_y2_armv6, export=1
+ pld [r1]
+ pld [r2]
+ ldr r12, [sp]
+ push {r4-r11, lr}
+ mov r0, #0
+ mov lr, #1
+ orr lr, lr, lr, lsl #8
+ orr lr, lr, lr, lsl #16
+ ldr r4, [r2]
+ ldr r5, [r2, #4]
+ ldr r6, [r2, #8]
+ ldr r7, [r2, #12]
+ add r2, r2, r3
+1:
+ usad_y2 r4, r5, r6, r7, r8, r9, r10, r11
+ subs r12, r12, #2
+ usad_y2 r8, r9, r10, r11, r4, r5, r6, r7
+ bgt 1b
+
+ pop {r4-r11, pc}
+.endfunc
+
+function ff_pix_abs8_armv6, export=1
+ pld [r2, r3]
+ ldr r12, [sp]
+ push {r4-r9, lr}
+ mov r0, #0
+ mov lr, #0
+ ldrd r4, r5, [r1], r3
+1:
+ subs r12, r12, #2
+ ldr r7, [r2, #4]
+ ldr r6, [r2], r3
+ ldrd r8, r9, [r1], r3
+ usada8 r0, r4, r6, r0
+ pld [r2, r3]
+ usada8 lr, r5, r7, lr
+ ldr r7, [r2, #4]
+ ldr r6, [r2], r3
+ beq 2f
+ ldrd r4, r5, [r1], r3
+ usada8 r0, r8, r6, r0
+ pld [r2, r3]
+ usada8 lr, r9, r7, lr
+ b 1b
+2:
+ usada8 r0, r8, r6, r0
+ usada8 lr, r9, r7, lr
+ add r0, r0, lr
+ pop {r4-r9, pc}
+.endfunc
projects / ffmpeg.git / blobdiff