falkTX
/
FFmpeg
mirror of https://github.com/falkTX/FFmpeg.git
1
0
Fork 0
Code Issues 0 Releases 338 Wiki Activity
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
13133 Commits
32 Branches
436MB
Tree: c1ac39b3df
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from 'c1ac39b3df'
${ noResults }
FFmpeg/libavcodec/armv4l/dsputil_arm_s.S

698 lines
18KB
Raw Blame History 

  1. @

  2. @ ARMv4L optimized DSP utils

  3. @ Copyright (c) 2004 AGAWA Koji <i (AT) atty (DOT) jp>

  4. @

  5. @ This file is part of FFmpeg.

  6. @

  7. @ FFmpeg is free software; you can redistribute it and/or

  8. @ modify it under the terms of the GNU Lesser General Public

  9. @ License as published by the Free Software Foundation; either

  10. @ version 2.1 of the License, or (at your option) any later version.

  11. @

  12. @ FFmpeg is distributed in the hope that it will be useful,

  13. @ but WITHOUT ANY WARRANTY; without even the implied warranty of

  14. @ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU

  15. @ Lesser General Public License for more details.

  16. @

  17. @ You should have received a copy of the GNU Lesser General Public

  18. @ License along with FFmpeg; if not, write to the Free Software

  19. @ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA

  20. @

  21. 

  22. #include "config.h"

  23. 

  24. #ifndef HAVE_PLD

  25. .macro pld reg

  26. .endm

  27. #endif

  28. 

  29. .macro  ADJ_ALIGN_QUADWORD_D shift, Rd0, Rd1, Rd2, Rd3, Rn0, Rn1, Rn2, Rn3, Rn4

  30.         mov \Rd0, \Rn0, lsr #(\shift * 8)

  31.         mov \Rd1, \Rn1, lsr #(\shift * 8)

  32.         mov \Rd2, \Rn2, lsr #(\shift * 8)

  33.         mov \Rd3, \Rn3, lsr #(\shift * 8)

  34.         orr \Rd0, \Rd0, \Rn1, lsl #(32 - \shift * 8)

  35.         orr \Rd1, \Rd1, \Rn2, lsl #(32 - \shift * 8)

  36.         orr \Rd2, \Rd2, \Rn3, lsl #(32 - \shift * 8)

  37.         orr \Rd3, \Rd3, \Rn4, lsl #(32 - \shift * 8)

  38. .endm

  39. .macro  ADJ_ALIGN_DOUBLEWORD shift, R0, R1, R2

  40.         mov \R0, \R0, lsr #(\shift * 8)

  41.         orr \R0, \R0, \R1, lsl #(32 - \shift * 8)

  42.         mov \R1, \R1, lsr #(\shift * 8)

  43.         orr \R1, \R1, \R2, lsl #(32 - \shift * 8)

  44. .endm

  45. .macro  ADJ_ALIGN_DOUBLEWORD_D shift, Rdst0, Rdst1, Rsrc0, Rsrc1, Rsrc2

  46.         mov \Rdst0, \Rsrc0, lsr #(\shift * 8)

  47.         mov \Rdst1, \Rsrc1, lsr #(\shift * 8)

  48.         orr \Rdst0, \Rdst0, \Rsrc1, lsl #(32 - (\shift * 8))

  49.         orr \Rdst1, \Rdst1, \Rsrc2, lsl #(32 - (\shift * 8))

  50. .endm

  51. 

  52. .macro  RND_AVG32 Rd0, Rd1, Rn0, Rn1, Rm0, Rm1, Rmask

  53.         @ Rd = (Rn | Rm) - (((Rn ^ Rm) & ~0x01010101) >> 1)

  54.         @ Rmask = 0xFEFEFEFE

  55.         @ Rn = destroy

  56.         eor \Rd0, \Rn0, \Rm0

  57.         eor \Rd1, \Rn1, \Rm1

  58.         orr \Rn0, \Rn0, \Rm0

  59.         orr \Rn1, \Rn1, \Rm1

  60.         and \Rd0, \Rd0, \Rmask

  61.         and \Rd1, \Rd1, \Rmask

  62.         sub \Rd0, \Rn0, \Rd0, lsr #1

  63.         sub \Rd1, \Rn1, \Rd1, lsr #1

  64. .endm

  65. 

  66. .macro  NO_RND_AVG32 Rd0, Rd1, Rn0, Rn1, Rm0, Rm1, Rmask

  67.         @ Rd = (Rn & Rm) - (((Rn ^ Rm) & ~0x01010101) >> 1)

  68.         @ Rmask = 0xFEFEFEFE

  69.         @ Rn = destroy

  70.         eor \Rd0, \Rn0, \Rm0

  71.         eor \Rd1, \Rn1, \Rm1

  72.         and \Rn0, \Rn0, \Rm0

  73.         and \Rn1, \Rn1, \Rm1

  74.         and \Rd0, \Rd0, \Rmask

  75.         and \Rd1, \Rd1, \Rmask

  76.         add \Rd0, \Rn0, \Rd0, lsr #1

  77.         add \Rd1, \Rn1, \Rd1, lsr #1

  78. .endm

  79. 

  80. @ ----------------------------------------------------------------

  81.         .align 8

  82.         .global put_pixels16_arm

  83. put_pixels16_arm:

  84.         @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h)

  85.         @ block = word aligned, pixles = unaligned

  86.         pld [r1]

  87.         stmfd sp!, {r4-r11, lr} @ R14 is also called LR

  88.         adr r5, 5f

  89.         ands r4, r1, #3

  90.         bic r1, r1, #3

  91.         add r5, r5, r4, lsl #2

  92.         ldrne pc, [r5]

  93. 1:

  94.         ldmia r1, {r4-r7}

  95.         add r1, r1, r2

  96.         stmia r0, {r4-r7}

  97.         pld [r1]

  98.         subs r3, r3, #1

  99.         add r0, r0, r2

  100.         bne 1b

  101.         ldmfd sp!, {r4-r11, pc}

  102.         .align 8

  103. 2:

  104.         ldmia r1, {r4-r8}

  105.         add r1, r1, r2

  106.         ADJ_ALIGN_QUADWORD_D 1, r9, r10, r11, r12, r4, r5, r6, r7, r8

  107.         pld [r1]

  108.         subs r3, r3, #1

  109.         stmia r0, {r9-r12}

  110.         add r0, r0, r2

  111.         bne 2b

  112.         ldmfd sp!, {r4-r11, pc}

  113.         .align 8

  114. 3:

  115.         ldmia r1, {r4-r8}

  116.         add r1, r1, r2

  117.         ADJ_ALIGN_QUADWORD_D 2, r9, r10, r11, r12, r4, r5, r6, r7, r8

  118.         pld [r1]

  119.         subs r3, r3, #1

  120.         stmia r0, {r9-r12}

  121.         add r0, r0, r2

  122.         bne 3b

  123.         ldmfd sp!, {r4-r11, pc}

  124.         .align 8

  125. 4:

  126.         ldmia r1, {r4-r8}

  127.         add r1, r1, r2

  128.         ADJ_ALIGN_QUADWORD_D 3, r9, r10, r11, r12, r4, r5, r6, r7, r8

  129.         pld [r1]

  130.         subs r3, r3, #1

  131.         stmia r0, {r9-r12}

  132.         add r0, r0, r2

  133.         bne 4b

  134.         ldmfd sp!, {r4-r11,pc}

  135.         .align 8

  136. 5:

  137.         .word 1b

  138.         .word 2b

  139.         .word 3b

  140.         .word 4b

  141. 

  142. @ ----------------------------------------------------------------

  143.         .align 8

  144.         .global put_pixels8_arm

  145. put_pixels8_arm:

  146.         @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h)

  147.         @ block = word aligned, pixles = unaligned

  148.         pld [r1]

  149.         stmfd sp!, {r4-r5,lr} @ R14 is also called LR

  150.         adr r5, 5f

  151.         ands r4, r1, #3

  152.         bic r1, r1, #3

  153.         add r5, r5, r4, lsl #2

  154.         ldrne pc, [r5]

  155. 1:

  156.         ldmia r1, {r4-r5}

  157.         add r1, r1, r2

  158.         subs r3, r3, #1

  159.         pld [r1]

  160.         stmia r0, {r4-r5}

  161.         add r0, r0, r2

  162.         bne 1b

  163.         ldmfd sp!, {r4-r5,pc}

  164.         .align 8

  165. 2:

  166.         ldmia r1, {r4-r5, r12}

  167.         add r1, r1, r2

  168.         ADJ_ALIGN_DOUBLEWORD 1, r4, r5, r12

  169.         pld [r1]

  170.         subs r3, r3, #1

  171.         stmia r0, {r4-r5}

  172.         add r0, r0, r2

  173.         bne 2b

  174.         ldmfd sp!, {r4-r5,pc}

  175.         .align 8

  176. 3:

  177.         ldmia r1, {r4-r5, r12}

  178.         add r1, r1, r2

  179.         ADJ_ALIGN_DOUBLEWORD 2, r4, r5, r12

  180.         pld [r1]

  181.         subs r3, r3, #1

  182.         stmia r0, {r4-r5}

  183.         add r0, r0, r2

  184.         bne 3b

  185.         ldmfd sp!, {r4-r5,pc}

  186.         .align 8

  187. 4:

  188.         ldmia r1, {r4-r5, r12}

  189.         add r1, r1, r2

  190.         ADJ_ALIGN_DOUBLEWORD 3, r4, r5, r12

  191.         pld [r1]

  192.         subs r3, r3, #1

  193.         stmia r0, {r4-r5}

  194.         add r0, r0, r2

  195.         bne 4b

  196.         ldmfd sp!, {r4-r5,pc}

  197.         .align 8

  198. 5:

  199.         .word 1b

  200.         .word 2b

  201.         .word 3b

  202.         .word 4b

  203. 

  204. @ ----------------------------------------------------------------

  205.         .align 8

  206.         .global put_pixels8_x2_arm

  207. put_pixels8_x2_arm:

  208.         @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h)

  209.         @ block = word aligned, pixles = unaligned

  210.         pld [r1]

  211.         stmfd sp!, {r4-r10,lr} @ R14 is also called LR

  212.         adr r5, 5f

  213.         ands r4, r1, #3

  214.         ldr r12, [r5]

  215.         add r5, r5, r4, lsl #2

  216.         bic r1, r1, #3

  217.         ldrne pc, [r5]

  218. 1:

  219.         ldmia r1, {r4-r5, r10}

  220.         add r1, r1, r2

  221.         ADJ_ALIGN_DOUBLEWORD_D 1, r6, r7, r4, r5, r10

  222.         pld [r1]

  223.         RND_AVG32 r8, r9, r4, r5, r6, r7, r12

  224.         subs r3, r3, #1

  225.         stmia r0, {r8-r9}

  226.         add r0, r0, r2

  227.         bne 1b

  228.         ldmfd sp!, {r4-r10,pc}

  229.         .align 8

  230. 2:

  231.         ldmia r1, {r4-r5, r10}

  232.         add r1, r1, r2

  233.         ADJ_ALIGN_DOUBLEWORD_D 1, r6, r7, r4, r5, r10

  234.         ADJ_ALIGN_DOUBLEWORD_D 2, r8, r9, r4, r5, r10

  235.         pld [r1]

  236.         RND_AVG32 r4, r5, r6, r7, r8, r9, r12

  237.         subs r3, r3, #1

  238.         stmia r0, {r4-r5}

  239.         add r0, r0, r2

  240.         bne 2b

  241.         ldmfd sp!, {r4-r10,pc}

  242.         .align 8

  243. 3:

  244.         ldmia r1, {r4-r5, r10}

  245.         add r1, r1, r2

  246.         ADJ_ALIGN_DOUBLEWORD_D 2, r6, r7, r4, r5, r10

  247.         ADJ_ALIGN_DOUBLEWORD_D 3, r8, r9, r4, r5, r10

  248.         pld [r1]

  249.         RND_AVG32 r4, r5, r6, r7, r8, r9, r12

  250.         subs r3, r3, #1

  251.         stmia r0, {r4-r5}

  252.         add r0, r0, r2

  253.         bne 3b

  254.         ldmfd sp!, {r4-r10,pc}

  255.         .align 8

  256. 4:

  257.         ldmia r1, {r4-r5, r10}

  258.         add r1, r1, r2

  259.         ADJ_ALIGN_DOUBLEWORD_D 3, r6, r7, r4, r5, r10

  260.         pld [r1]

  261.         RND_AVG32 r8, r9, r6, r7, r5, r10, r12

  262.         subs r3, r3, #1

  263.         stmia r0, {r8-r9}

  264.         add r0, r0, r2

  265.         bne 4b

  266.         ldmfd sp!, {r4-r10,pc} @@ update PC with LR content.

  267.         .align 8

  268. 5:

  269.         .word 0xFEFEFEFE

  270.         .word 2b

  271.         .word 3b

  272.         .word 4b

  273. 

  274.         .align 8

  275.         .global put_no_rnd_pixels8_x2_arm

  276. put_no_rnd_pixels8_x2_arm:

  277.         @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h)

  278.         @ block = word aligned, pixles = unaligned

  279.         pld [r1]

  280.         stmfd sp!, {r4-r10,lr} @ R14 is also called LR

  281.         adr r5, 5f

  282.         ands r4, r1, #3

  283.         ldr r12, [r5]

  284.         add r5, r5, r4, lsl #2

  285.         bic r1, r1, #3

  286.         ldrne pc, [r5]

  287. 1:

  288.         ldmia r1, {r4-r5, r10}

  289.         add r1, r1, r2

  290.         ADJ_ALIGN_DOUBLEWORD_D 1, r6, r7, r4, r5, r10

  291.         pld [r1]

  292.         NO_RND_AVG32 r8, r9, r4, r5, r6, r7, r12

  293.         subs r3, r3, #1

  294.         stmia r0, {r8-r9}

  295.         add r0, r0, r2

  296.         bne 1b

  297.         ldmfd sp!, {r4-r10,pc}

  298.         .align 8

  299. 2:

  300.         ldmia r1, {r4-r5, r10}

  301.         add r1, r1, r2

  302.         ADJ_ALIGN_DOUBLEWORD_D 1, r6, r7, r4, r5, r10

  303.         ADJ_ALIGN_DOUBLEWORD_D 2, r8, r9, r4, r5, r10

  304.         pld [r1]

  305.         NO_RND_AVG32 r4, r5, r6, r7, r8, r9, r12

  306.         subs r3, r3, #1

  307.         stmia r0, {r4-r5}

  308.         add r0, r0, r2

  309.         bne 2b

  310.         ldmfd sp!, {r4-r10,pc}

  311.         .align 8

  312. 3:

  313.         ldmia r1, {r4-r5, r10}

  314.         add r1, r1, r2

  315.         ADJ_ALIGN_DOUBLEWORD_D 2, r6, r7, r4, r5, r10

  316.         ADJ_ALIGN_DOUBLEWORD_D 3, r8, r9, r4, r5, r10

  317.         pld [r1]

  318.         NO_RND_AVG32 r4, r5, r6, r7, r8, r9, r12

  319.         subs r3, r3, #1

  320.         stmia r0, {r4-r5}

  321.         add r0, r0, r2

  322.         bne 3b

  323.         ldmfd sp!, {r4-r10,pc}

  324.         .align 8

  325. 4:

  326.         ldmia r1, {r4-r5, r10}

  327.         add r1, r1, r2

  328.         ADJ_ALIGN_DOUBLEWORD_D 3, r6, r7, r4, r5, r10

  329.         pld [r1]

  330.         NO_RND_AVG32 r8, r9, r6, r7, r5, r10, r12

  331.         subs r3, r3, #1

  332.         stmia r0, {r8-r9}

  333.         add r0, r0, r2

  334.         bne 4b

  335.         ldmfd sp!, {r4-r10,pc} @@ update PC with LR content.

  336.         .align 8

  337. 5:

  338.         .word 0xFEFEFEFE

  339.         .word 2b

  340.         .word 3b

  341.         .word 4b

  342. 

  343. 

  344. @ ----------------------------------------------------------------

  345.         .align 8

  346.         .global put_pixels8_y2_arm

  347. put_pixels8_y2_arm:

  348.         @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h)

  349.         @ block = word aligned, pixles = unaligned

  350.         pld [r1]

  351.         stmfd sp!, {r4-r11,lr} @ R14 is also called LR

  352.         adr r5, 5f

  353.         ands r4, r1, #3

  354.         mov r3, r3, lsr #1

  355.         ldr r12, [r5]

  356.         add r5, r5, r4, lsl #2

  357.         bic r1, r1, #3

  358.         ldrne pc, [r5]

  359. 1:

  360.         ldmia r1, {r4-r5}

  361.         add r1, r1, r2

  362. 6:      ldmia r1, {r6-r7}

  363.         add r1, r1, r2

  364.         pld [r1]

  365.         RND_AVG32 r8, r9, r4, r5, r6, r7, r12

  366.         ldmia r1, {r4-r5}

  367.         add r1, r1, r2

  368.         stmia r0, {r8-r9}

  369.         add r0, r0, r2

  370.         pld [r1]

  371.         RND_AVG32 r8, r9, r6, r7, r4, r5, r12

  372.         subs r3, r3, #1

  373.         stmia r0, {r8-r9}

  374.         add r0, r0, r2

  375.         bne 6b

  376.         ldmfd sp!, {r4-r11,pc}

  377.         .align 8

  378. 2:

  379.         ldmia r1, {r4-r6}

  380.         add r1, r1, r2

  381.         pld [r1]

  382.         ADJ_ALIGN_DOUBLEWORD 1, r4, r5, r6

  383. 6:      ldmia r1, {r7-r9}

  384.         add r1, r1, r2

  385.         pld [r1]

  386.         ADJ_ALIGN_DOUBLEWORD 1, r7, r8, r9

  387.         RND_AVG32 r10, r11, r4, r5, r7, r8, r12

  388.         stmia r0, {r10-r11}

  389.         add r0, r0, r2

  390.         ldmia r1, {r4-r6}

  391.         add r1, r1, r2

  392.         pld [r1]

  393.         ADJ_ALIGN_DOUBLEWORD 1, r4, r5, r6

  394.         subs r3, r3, #1

  395.         RND_AVG32 r10, r11, r7, r8, r4, r5, r12

  396.         stmia r0, {r10-r11}

  397.         add r0, r0, r2

  398.         bne 6b

  399.         ldmfd sp!, {r4-r11,pc}

  400.         .align 8

  401. 3:

  402.         ldmia r1, {r4-r6}

  403.         add r1, r1, r2

  404.         pld [r1]

  405.         ADJ_ALIGN_DOUBLEWORD 2, r4, r5, r6

  406. 6:      ldmia r1, {r7-r9}

  407.         add r1, r1, r2

  408.         pld [r1]

  409.         ADJ_ALIGN_DOUBLEWORD 2, r7, r8, r9

  410.         RND_AVG32 r10, r11, r4, r5, r7, r8, r12

  411.         stmia r0, {r10-r11}

  412.         add r0, r0, r2

  413.         ldmia r1, {r4-r6}

  414.         add r1, r1, r2

  415.         pld [r1]

  416.         ADJ_ALIGN_DOUBLEWORD 2, r4, r5, r6

  417.         subs r3, r3, #1

  418.         RND_AVG32 r10, r11, r7, r8, r4, r5, r12

  419.         stmia r0, {r10-r11}

  420.         add r0, r0, r2

  421.         bne 6b

  422.         ldmfd sp!, {r4-r11,pc}

  423.         .align 8

  424. 4:

  425.         ldmia r1, {r4-r6}

  426.         add r1, r1, r2

  427.         pld [r1]

  428.         ADJ_ALIGN_DOUBLEWORD 3, r4, r5, r6

  429. 6:      ldmia r1, {r7-r9}

  430.         add r1, r1, r2

  431.         pld [r1]

  432.         ADJ_ALIGN_DOUBLEWORD 3, r7, r8, r9

  433.         RND_AVG32 r10, r11, r4, r5, r7, r8, r12

  434.         stmia r0, {r10-r11}

  435.         add r0, r0, r2

  436.         ldmia r1, {r4-r6}

  437.         add r1, r1, r2

  438.         pld [r1]

  439.         ADJ_ALIGN_DOUBLEWORD 3, r4, r5, r6

  440.         subs r3, r3, #1

  441.         RND_AVG32 r10, r11, r7, r8, r4, r5, r12

  442.         stmia r0, {r10-r11}

  443.         add r0, r0, r2

  444.         bne 6b

  445.         ldmfd sp!, {r4-r11,pc}

  446. 

  447.         .align 8

  448. 5:

  449.         .word 0xFEFEFEFE

  450.         .word 2b

  451.         .word 3b

  452.         .word 4b

  453. 

  454.         .align 8

  455.         .global put_no_rnd_pixels8_y2_arm

  456. put_no_rnd_pixels8_y2_arm:

  457.         @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h)

  458.         @ block = word aligned, pixles = unaligned

  459.         pld [r1]

  460.         stmfd sp!, {r4-r11,lr} @ R14 is also called LR

  461.         adr r5, 5f

  462.         ands r4, r1, #3

  463.         mov r3, r3, lsr #1

  464.         ldr r12, [r5]

  465.         add r5, r5, r4, lsl #2

  466.         bic r1, r1, #3

  467.         ldrne pc, [r5]

  468. 1:

  469.         ldmia r1, {r4-r5}

  470.         add r1, r1, r2

  471. 6:      ldmia r1, {r6-r7}

  472.         add r1, r1, r2

  473.         pld [r1]

  474.         NO_RND_AVG32 r8, r9, r4, r5, r6, r7, r12

  475.         ldmia r1, {r4-r5}

  476.         add r1, r1, r2

  477.         stmia r0, {r8-r9}

  478.         add r0, r0, r2

  479.         pld [r1]

  480.         NO_RND_AVG32 r8, r9, r6, r7, r4, r5, r12

  481.         subs r3, r3, #1

  482.         stmia r0, {r8-r9}

  483.         add r0, r0, r2

  484.         bne 6b

  485.         ldmfd sp!, {r4-r11,pc}

  486.         .align 8

  487. 2:

  488.         ldmia r1, {r4-r6}

  489.         add r1, r1, r2

  490.         pld [r1]

  491.         ADJ_ALIGN_DOUBLEWORD 1, r4, r5, r6

  492. 6:      ldmia r1, {r7-r9}

  493.         add r1, r1, r2

  494.         pld [r1]

  495.         ADJ_ALIGN_DOUBLEWORD 1, r7, r8, r9

  496.         NO_RND_AVG32 r10, r11, r4, r5, r7, r8, r12

  497.         stmia r0, {r10-r11}

  498.         add r0, r0, r2

  499.         ldmia r1, {r4-r6}

  500.         add r1, r1, r2

  501.         pld [r1]

  502.         ADJ_ALIGN_DOUBLEWORD 1, r4, r5, r6

  503.         subs r3, r3, #1

  504.         NO_RND_AVG32 r10, r11, r7, r8, r4, r5, r12

  505.         stmia r0, {r10-r11}

  506.         add r0, r0, r2

  507.         bne 6b

  508.         ldmfd sp!, {r4-r11,pc}

  509.         .align 8

  510. 3:

  511.         ldmia r1, {r4-r6}

  512.         add r1, r1, r2

  513.         pld [r1]

  514.         ADJ_ALIGN_DOUBLEWORD 2, r4, r5, r6

  515. 6:      ldmia r1, {r7-r9}

  516.         add r1, r1, r2

  517.         pld [r1]

  518.         ADJ_ALIGN_DOUBLEWORD 2, r7, r8, r9

  519.         NO_RND_AVG32 r10, r11, r4, r5, r7, r8, r12

  520.         stmia r0, {r10-r11}

  521.         add r0, r0, r2

  522.         ldmia r1, {r4-r6}

  523.         add r1, r1, r2

  524.         pld [r1]

  525.         ADJ_ALIGN_DOUBLEWORD 2, r4, r5, r6

  526.         subs r3, r3, #1

  527.         NO_RND_AVG32 r10, r11, r7, r8, r4, r5, r12

  528.         stmia r0, {r10-r11}

  529.         add r0, r0, r2

  530.         bne 6b

  531.         ldmfd sp!, {r4-r11,pc}

  532.         .align 8

  533. 4:

  534.         ldmia r1, {r4-r6}

  535.         add r1, r1, r2

  536.         pld [r1]

  537.         ADJ_ALIGN_DOUBLEWORD 3, r4, r5, r6

  538. 6:      ldmia r1, {r7-r9}

  539.         add r1, r1, r2

  540.         pld [r1]

  541.         ADJ_ALIGN_DOUBLEWORD 3, r7, r8, r9

  542.         NO_RND_AVG32 r10, r11, r4, r5, r7, r8, r12

  543.         stmia r0, {r10-r11}

  544.         add r0, r0, r2

  545.         ldmia r1, {r4-r6}

  546.         add r1, r1, r2

  547.         pld [r1]

  548.         ADJ_ALIGN_DOUBLEWORD 3, r4, r5, r6

  549.         subs r3, r3, #1

  550.         NO_RND_AVG32 r10, r11, r7, r8, r4, r5, r12

  551.         stmia r0, {r10-r11}

  552.         add r0, r0, r2

  553.         bne 6b

  554.         ldmfd sp!, {r4-r11,pc}

  555.         .align 8

  556. 5:

  557.         .word 0xFEFEFEFE

  558.         .word 2b

  559.         .word 3b

  560.         .word 4b

  561. 

  562. @ ----------------------------------------------------------------

  563. .macro  RND_XY2_IT align

  564.         @ l1=  (a & 0x03030303) + (b & 0x03030303) ?(+ 0x02020202)

  565.         @ h1= ((a & 0xFCFCFCFCUL) >> 2) + ((b & 0xFCFCFCFCUL) >> 2)

  566. .if \align == 0

  567.         ldmia r1, {r6-r8}

  568. .elseif \align == 3

  569.         ldmia r1, {r5-r7}

  570. .else

  571.         ldmia r1, {r8-r10}

  572. .endif

  573.         add r1, r1, r2

  574.         pld [r1]

  575. .if \align == 0

  576.         ADJ_ALIGN_DOUBLEWORD_D 1, r4, r5, r6, r7, r8

  577. .elseif \align == 1

  578.         ADJ_ALIGN_DOUBLEWORD_D 1, r4, r5, r8, r9, r10

  579.         ADJ_ALIGN_DOUBLEWORD_D 2, r6, r7, r8, r9, r10

  580. .elseif \align == 2

  581.         ADJ_ALIGN_DOUBLEWORD_D 2, r4, r5, r8, r9, r10

  582.         ADJ_ALIGN_DOUBLEWORD_D 3, r6, r7, r8, r9, r10

  583. .elseif \align == 3

  584.         ADJ_ALIGN_DOUBLEWORD_D 3, r4, r5, r5, r6, r7

  585. .endif

  586.         ldr r14, [r12, #0]      @ 0x03030303

  587.         tst r3, #1

  588.         and r8, r4, r14

  589.         and r9, r5, r14

  590.         and r10, r6, r14

  591.         and r11, r7, r14

  592.         ldreq r14, [r12, #16]   @ 0x02020202/0x01010101

  593.         add r8, r8, r10

  594.         add r9, r9, r11

  595.         addeq r8, r8, r14

  596.         addeq r9, r9, r14

  597.         ldr r14, [r12, #20]     @ 0xFCFCFCFC >> 2

  598.         and r4, r14, r4, lsr #2

  599.         and r5, r14, r5, lsr #2

  600.         and r6, r14, r6, lsr #2

  601.         and r7, r14, r7, lsr #2

  602.         add r10, r4, r6

  603.         add r11, r5, r7

  604.         subs r3, r3, #1

  605. .endm

  606. 

  607. .macro RND_XY2_EXPAND align

  608.         RND_XY2_IT \align

  609. 6:      stmfd sp!, {r8-r11}

  610.         RND_XY2_IT \align

  611.         ldmfd sp!, {r4-r7}

  612.         add r4, r4, r8

  613.         add r5, r5, r9

  614.         add r6, r6, r10

  615.         add r7, r7, r11

  616.         ldr r14, [r12, #24]     @ 0x0F0F0F0F

  617.         and r4, r14, r4, lsr #2

  618.         and r5, r14, r5, lsr #2

  619.         add r4, r4, r6

  620.         add r5, r5, r7

  621.         stmia r0, {r4-r5}

  622.         add r0, r0, r2

  623.         bge 6b

  624.         ldmfd sp!, {r4-r11,pc}

  625. .endm

  626. 

  627.         .align 8

  628.         .global put_pixels8_xy2_arm

  629. put_pixels8_xy2_arm:

  630.         @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h)

  631.         @ block = word aligned, pixles = unaligned

  632.         pld [r1]

  633.         stmfd sp!, {r4-r11,lr} @ R14 is also called LR

  634.         adrl r12, 5f

  635.         ands r4, r1, #3

  636.         add r5, r12, r4, lsl #2

  637.         bic r1, r1, #3

  638.         ldrne pc, [r5]

  639. 1:

  640.         RND_XY2_EXPAND 0

  641. 

  642.         .align 8

  643. 2:

  644.         RND_XY2_EXPAND 1

  645. 

  646.         .align 8

  647. 3:

  648.         RND_XY2_EXPAND 2

  649. 

  650.         .align 8

  651. 4:

  652.         RND_XY2_EXPAND 3

  653. 

  654. 5:

  655.         .word 0x03030303

  656.         .word 2b

  657.         .word 3b

  658.         .word 4b

  659.         .word 0x02020202

  660.         .word 0xFCFCFCFC >> 2

  661.         .word 0x0F0F0F0F

  662. 

  663.         .align 8

  664.         .global put_no_rnd_pixels8_xy2_arm

  665. put_no_rnd_pixels8_xy2_arm:

  666.         @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h)

  667.         @ block = word aligned, pixles = unaligned

  668.         pld [r1]

  669.         stmfd sp!, {r4-r11,lr} @ R14 is also called LR

  670.         adrl r12, 5f

  671.         ands r4, r1, #3

  672.         add r5, r12, r4, lsl #2

  673.         bic r1, r1, #3

  674.         ldrne pc, [r5]

  675. 1:

  676.         RND_XY2_EXPAND 0

  677. 

  678.         .align 8

  679. 2:

  680.         RND_XY2_EXPAND 1

  681. 

  682.         .align 8

  683. 3:

  684.         RND_XY2_EXPAND 2

  685. 

  686.         .align 8

  687. 4:

  688.         RND_XY2_EXPAND 3

  689. 

  690. 5:

  691.         .word 0x03030303

  692.         .word 2b

  693.         .word 3b

  694.         .word 4b

  695.         .word 0x01010101

  696.         .word 0xFCFCFCFC >> 2

  697.         .word 0x0F0F0F0F