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.
8579 Commits
32 Branches
1.1GB
Tree: 024c7378c3
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from '024c7378c3'
${ noResults }
FFmpeg/libavcodec/armv4l/dsputil_arm_s.S

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

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

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

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

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

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

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

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

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

  31. .endm

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

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

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

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

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

  37. .endm

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

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

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

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

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

  43. .endm

  44. 

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

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

  47.         @ Rmask = 0xFEFEFEFE

  48.         @ Rn = destroy

  49.         eor \Rd0, \Rn0, \Rm0

  50.         eor \Rd1, \Rn1, \Rm1

  51.         orr \Rn0, \Rn0, \Rm0

  52.         orr \Rn1, \Rn1, \Rm1

  53.         and \Rd0, \Rd0, \Rmask

  54.         and \Rd1, \Rd1, \Rmask

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

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

  57. .endm

  58. 

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

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

  61.         @ Rmask = 0xFEFEFEFE

  62.         @ Rn = destroy

  63.         eor \Rd0, \Rn0, \Rm0

  64.         eor \Rd1, \Rn1, \Rm1

  65.         and \Rn0, \Rn0, \Rm0

  66.         and \Rn1, \Rn1, \Rm1

  67.         and \Rd0, \Rd0, \Rmask

  68.         and \Rd1, \Rd1, \Rmask

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

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

  71. .endm

  72. 

  73. @ ----------------------------------------------------------------

  74.         .align 8

  75.         .global put_pixels16_arm

  76. put_pixels16_arm:

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

  78.         @ block = word aligned, pixles = unaligned

  79.         pld [r1]

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

  81.         adr r5, 5f

  82.         ands r4, r1, #3

  83.         bic r1, r1, #3

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

  85.         ldrne pc, [r5]

  86. 1:

  87.         ldmia r1, {r4-r7}

  88.         add r1, r1, r2

  89.         stmia r0, {r4-r7}

  90.         pld [r1]

  91.         subs r3, r3, #1

  92.         add r0, r0, r2

  93.         bne 1b

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

  95.         .align 8

  96. 2:

  97.         ldmia r1, {r4-r8}

  98.         add r1, r1, r2

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

  100.         pld [r1]

  101.         subs r3, r3, #1

  102.         stmia r0, {r9-r12}

  103.         add r0, r0, r2

  104.         bne 2b

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

  106.         .align 8

  107. 3:

  108.         ldmia r1, {r4-r8}

  109.         add r1, r1, r2

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

  111.         pld [r1]

  112.         subs r3, r3, #1

  113.         stmia r0, {r9-r12}

  114.         add r0, r0, r2

  115.         bne 3b

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

  117.         .align 8

  118. 4:

  119.         ldmia r1, {r4-r8}

  120.         add r1, r1, r2

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

  122.         pld [r1]

  123.         subs r3, r3, #1

  124.         stmia r0, {r9-r12}

  125.         add r0, r0, r2

  126.         bne 4b

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

  128.         .align 8

  129. 5:

  130.         .word 1b

  131.         .word 2b

  132.         .word 3b

  133.         .word 4b

  134. 

  135. @ ----------------------------------------------------------------

  136.         .align 8

  137.         .global put_pixels8_arm

  138. put_pixels8_arm:

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

  140.         @ block = word aligned, pixles = unaligned

  141.         pld [r1]

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

  143.         adr r5, 5f

  144.         ands r4, r1, #3

  145.         bic r1, r1, #3

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

  147.         ldrne pc, [r5]

  148. 1:

  149.         ldmia r1, {r4-r5}

  150.         add r1, r1, r2

  151.         subs r3, r3, #1

  152.         pld [r1]

  153.         stmia r0, {r4-r5}

  154.         add r0, r0, r2

  155.         bne 1b

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

  157.         .align 8

  158. 2:

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

  160.         add r1, r1, r2

  161.         ADJ_ALIGN_DOUBLEWORD 1, r4, r5, r12

  162.         pld [r1]

  163.         subs r3, r3, #1

  164.         stmia r0, {r4-r5}

  165.         add r0, r0, r2

  166.         bne 2b

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

  168.         .align 8

  169. 3:

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

  171.         add r1, r1, r2

  172.         ADJ_ALIGN_DOUBLEWORD 2, r4, r5, r12

  173.         pld [r1]

  174.         subs r3, r3, #1

  175.         stmia r0, {r4-r5}

  176.         add r0, r0, r2

  177.         bne 3b

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

  179.         .align 8

  180. 4:

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

  182.         add r1, r1, r2

  183.         ADJ_ALIGN_DOUBLEWORD 3, r4, r5, r12

  184.         pld [r1]

  185.         subs r3, r3, #1

  186.         stmia r0, {r4-r5}

  187.         add r0, r0, r2

  188.         bne 4b

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

  190.         .align 8

  191. 5:

  192.         .word 1b

  193.         .word 2b

  194.         .word 3b

  195.         .word 4b

  196. 

  197. @ ----------------------------------------------------------------

  198.         .align 8

  199.         .global put_pixels8_x2_arm

  200. put_pixels8_x2_arm:

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

  202.         @ block = word aligned, pixles = unaligned

  203.         pld [r1]

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

  205.         adr r5, 5f

  206.         ands r4, r1, #3

  207.         ldr r12, [r5]

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

  209.         bic r1, r1, #3

  210.         ldrne pc, [r5]

  211. 1:

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

  213.         add r1, r1, r2

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

  215.         pld [r1]

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

  217.         subs r3, r3, #1

  218.         stmia r0, {r8-r9}

  219.         add r0, r0, r2

  220.         bne 1b

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

  222.         .align 8

  223. 2:

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

  225.         add r1, r1, r2

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

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

  228.         pld [r1]

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

  230.         subs r3, r3, #1

  231.         stmia r0, {r4-r5}

  232.         add r0, r0, r2

  233.         bne 2b

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

  235.         .align 8

  236. 3:

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

  238.         add r1, r1, r2

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

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

  241.         pld [r1]

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

  243.         subs r3, r3, #1

  244.         stmia r0, {r4-r5}

  245.         add r0, r0, r2

  246.         bne 3b

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

  248.         .align 8

  249. 4:

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

  251.         add r1, r1, r2

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

  253.         pld [r1]

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

  255.         subs r3, r3, #1

  256.         stmia r0, {r8-r9}

  257.         add r0, r0, r2

  258.         bne 4b

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

  260.         .align 8

  261. 5:

  262.         .word 0xFEFEFEFE

  263.         .word 2b

  264.         .word 3b

  265.         .word 4b

  266. 

  267.         .align 8

  268.         .global put_no_rnd_pixels8_x2_arm

  269. put_no_rnd_pixels8_x2_arm:

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

  271.         @ block = word aligned, pixles = unaligned

  272.         pld [r1]

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

  274.         adr r5, 5f

  275.         ands r4, r1, #3

  276.         ldr r12, [r5]

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

  278.         bic r1, r1, #3

  279.         ldrne pc, [r5]

  280. 1:

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

  282.         add r1, r1, r2

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

  284.         pld [r1]

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

  286.         subs r3, r3, #1

  287.         stmia r0, {r8-r9}

  288.         add r0, r0, r2

  289.         bne 1b

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

  291.         .align 8

  292. 2:

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

  294.         add r1, r1, r2

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

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

  297.         pld [r1]

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

  299.         subs r3, r3, #1

  300.         stmia r0, {r4-r5}

  301.         add r0, r0, r2

  302.         bne 2b

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

  304.         .align 8

  305. 3:

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

  307.         add r1, r1, r2

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

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

  310.         pld [r1]

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

  312.         subs r3, r3, #1

  313.         stmia r0, {r4-r5}

  314.         add r0, r0, r2

  315.         bne 3b

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

  317.         .align 8

  318. 4:

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

  320.         add r1, r1, r2

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

  322.         pld [r1]

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

  324.         subs r3, r3, #1

  325.         stmia r0, {r8-r9}

  326.         add r0, r0, r2

  327.         bne 4b

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

  329.         .align 8

  330. 5:

  331.         .word 0xFEFEFEFE

  332.         .word 2b

  333.         .word 3b

  334.         .word 4b

  335. 

  336. 

  337. @ ----------------------------------------------------------------

  338.         .align 8

  339.         .global put_pixels8_y2_arm

  340. put_pixels8_y2_arm:

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

  342.         @ block = word aligned, pixles = unaligned

  343.         pld [r1]

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

  345.         adr r5, 5f

  346.         ands r4, r1, #3

  347.         mov r3, r3, lsr #1

  348.         ldr r12, [r5]

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

  350.         bic r1, r1, #3

  351.         ldrne pc, [r5]

  352. 1:

  353.         ldmia r1, {r4-r5}

  354.         add r1, r1, r2

  355. 6:      ldmia r1, {r6-r7}

  356.         add r1, r1, r2

  357.         pld [r1]

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

  359.         ldmia r1, {r4-r5}

  360.         add r1, r1, r2

  361.         stmia r0, {r8-r9}

  362.         add r0, r0, r2

  363.         pld [r1]

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

  365.         subs r3, r3, #1

  366.         stmia r0, {r8-r9}

  367.         add r0, r0, r2

  368.         bne 6b

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

  370.         .align 8

  371. 2:

  372.         ldmia r1, {r4-r6}

  373.         add r1, r1, r2

  374.         pld [r1]

  375.         ADJ_ALIGN_DOUBLEWORD 1, r4, r5, r6

  376. 6:      ldmia r1, {r7-r9}

  377.         add r1, r1, r2

  378.         pld [r1]

  379.         ADJ_ALIGN_DOUBLEWORD 1, r7, r8, r9

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

  381.         stmia r0, {r10-r11}

  382.         add r0, r0, r2

  383.         ldmia r1, {r4-r6}

  384.         add r1, r1, r2

  385.         pld [r1]

  386.         ADJ_ALIGN_DOUBLEWORD 1, r4, r5, r6

  387.         subs r3, r3, #1

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

  389.         stmia r0, {r10-r11}

  390.         add r0, r0, r2

  391.         bne 6b

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

  393.         .align 8

  394. 3:

  395.         ldmia r1, {r4-r6}

  396.         add r1, r1, r2

  397.         pld [r1]

  398.         ADJ_ALIGN_DOUBLEWORD 2, r4, r5, r6

  399. 6:      ldmia r1, {r7-r9}

  400.         add r1, r1, r2

  401.         pld [r1]

  402.         ADJ_ALIGN_DOUBLEWORD 2, r7, r8, r9

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

  404.         stmia r0, {r10-r11}

  405.         add r0, r0, r2

  406.         ldmia r1, {r4-r6}

  407.         add r1, r1, r2

  408.         pld [r1]

  409.         ADJ_ALIGN_DOUBLEWORD 2, r4, r5, r6

  410.         subs r3, r3, #1

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

  412.         stmia r0, {r10-r11}

  413.         add r0, r0, r2

  414.         bne 6b

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

  416.         .align 8

  417. 4:

  418.         ldmia r1, {r4-r6}

  419.         add r1, r1, r2

  420.         pld [r1]

  421.         ADJ_ALIGN_DOUBLEWORD 3, r4, r5, r6

  422. 6:      ldmia r1, {r7-r9}

  423.         add r1, r1, r2

  424.         pld [r1]

  425.         ADJ_ALIGN_DOUBLEWORD 3, r7, r8, r9

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

  427.         stmia r0, {r10-r11}

  428.         add r0, r0, r2

  429.         ldmia r1, {r4-r6}

  430.         add r1, r1, r2

  431.         pld [r1]

  432.         ADJ_ALIGN_DOUBLEWORD 3, r4, r5, r6

  433.         subs r3, r3, #1

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

  435.         stmia r0, {r10-r11}

  436.         add r0, r0, r2

  437.         bne 6b

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

  439. 

  440.         .align 8

  441. 5:

  442.         .word 0xFEFEFEFE

  443.         .word 2b

  444.         .word 3b

  445.         .word 4b

  446. 

  447.         .align 8

  448.         .global put_no_rnd_pixels8_y2_arm

  449. put_no_rnd_pixels8_y2_arm:

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

  451.         @ block = word aligned, pixles = unaligned

  452.         pld [r1]

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

  454.         adr r5, 5f

  455.         ands r4, r1, #3

  456.         mov r3, r3, lsr #1

  457.         ldr r12, [r5]

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

  459.         bic r1, r1, #3

  460.         ldrne pc, [r5]

  461. 1:

  462.         ldmia r1, {r4-r5}

  463.         add r1, r1, r2

  464. 6:      ldmia r1, {r6-r7}

  465.         add r1, r1, r2

  466.         pld [r1]

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

  468.         ldmia r1, {r4-r5}

  469.         add r1, r1, r2

  470.         stmia r0, {r8-r9}

  471.         add r0, r0, r2

  472.         pld [r1]

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

  474.         subs r3, r3, #1

  475.         stmia r0, {r8-r9}

  476.         add r0, r0, r2

  477.         bne 6b

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

  479.         .align 8

  480. 2:

  481.         ldmia r1, {r4-r6}

  482.         add r1, r1, r2

  483.         pld [r1]

  484.         ADJ_ALIGN_DOUBLEWORD 1, r4, r5, r6

  485. 6:      ldmia r1, {r7-r9}

  486.         add r1, r1, r2

  487.         pld [r1]

  488.         ADJ_ALIGN_DOUBLEWORD 1, r7, r8, r9

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

  490.         stmia r0, {r10-r11}

  491.         add r0, r0, r2

  492.         ldmia r1, {r4-r6}

  493.         add r1, r1, r2

  494.         pld [r1]

  495.         ADJ_ALIGN_DOUBLEWORD 1, r4, r5, r6

  496.         subs r3, r3, #1

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

  498.         stmia r0, {r10-r11}

  499.         add r0, r0, r2

  500.         bne 6b

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

  502.         .align 8

  503. 3:

  504.         ldmia r1, {r4-r6}

  505.         add r1, r1, r2

  506.         pld [r1]

  507.         ADJ_ALIGN_DOUBLEWORD 2, r4, r5, r6

  508. 6:      ldmia r1, {r7-r9}

  509.         add r1, r1, r2

  510.         pld [r1]

  511.         ADJ_ALIGN_DOUBLEWORD 2, r7, r8, r9

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

  513.         stmia r0, {r10-r11}

  514.         add r0, r0, r2

  515.         ldmia r1, {r4-r6}

  516.         add r1, r1, r2

  517.         pld [r1]

  518.         ADJ_ALIGN_DOUBLEWORD 2, r4, r5, r6

  519.         subs r3, r3, #1

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

  521.         stmia r0, {r10-r11}

  522.         add r0, r0, r2

  523.         bne 6b

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

  525.         .align 8

  526. 4:

  527.         ldmia r1, {r4-r6}

  528.         add r1, r1, r2

  529.         pld [r1]

  530.         ADJ_ALIGN_DOUBLEWORD 3, r4, r5, r6

  531. 6:      ldmia r1, {r7-r9}

  532.         add r1, r1, r2

  533.         pld [r1]

  534.         ADJ_ALIGN_DOUBLEWORD 3, r7, r8, r9

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

  536.         stmia r0, {r10-r11}

  537.         add r0, r0, r2

  538.         ldmia r1, {r4-r6}

  539.         add r1, r1, r2

  540.         pld [r1]

  541.         ADJ_ALIGN_DOUBLEWORD 3, r4, r5, r6

  542.         subs r3, r3, #1

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

  544.         stmia r0, {r10-r11}

  545.         add r0, r0, r2

  546.         bne 6b

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

  548.         .align 8

  549. 5:

  550.         .word 0xFEFEFEFE

  551.         .word 2b

  552.         .word 3b

  553.         .word 4b

  554. 

  555. @ ----------------------------------------------------------------

  556. .macro  RND_XY2_IT align, rnd

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

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

  559. .if \align == 0

  560.         ldmia r1, {r6-r8}

  561. .elseif \align == 3

  562.         ldmia r1, {r5-r7}

  563. .else

  564.         ldmia r1, {r8-r10}

  565. .endif

  566.         add r1, r1, r2

  567.         pld [r1]

  568. .if \align == 0

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

  570. .elseif \align == 1

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

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

  573. .elseif \align == 2

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

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

  576. .elseif \align == 3

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

  578. .endif

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

  580.         tst r3, #1

  581.         and r8, r4, r14

  582.         and r9, r5, r14

  583.         and r10, r6, r14

  584.         and r11, r7, r14

  585. .if \rnd == 1

  586.         ldreq r14, [r12, #16]   @ 0x02020202

  587. .else

  588.         ldreq r14, [r12, #28]   @ 0x01010101

  589. .endif

  590.         add r8, r8, r10

  591.         add r9, r9, r11

  592.         addeq r8, r8, r14

  593.         addeq r9, r9, r14

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

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

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

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

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

  599.         add r10, r4, r6

  600.         add r11, r5, r7

  601. .endm

  602. 

  603. .macro RND_XY2_EXPAND align, rnd

  604.         RND_XY2_IT \align, \rnd

  605. 6:      stmfd sp!, {r8-r11}

  606.         RND_XY2_IT \align, \rnd

  607.         ldmfd sp!, {r4-r7}

  608.         add r4, r4, r8

  609.         add r5, r5, r9

  610.         add r6, r6, r10

  611.         add r7, r7, r11

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

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

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

  615.         add r4, r4, r6

  616.         add r5, r5, r7

  617.         subs r3, r3, #1

  618.         stmia r0, {r4-r5}

  619.         add r0, r0, r2

  620.         bne 6b

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

  622. .endm

  623. 

  624.         .align 8

  625.         .global put_pixels8_xy2_arm

  626. put_pixels8_xy2_arm:

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

  628.         @ block = word aligned, pixles = unaligned

  629.         pld [r1]

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

  631.         adrl r12, 5f

  632.         ands r4, r1, #3

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

  634.         bic r1, r1, #3

  635.         ldrne pc, [r5]

  636. 1:

  637.         RND_XY2_EXPAND 0, 1

  638. 

  639.         .align 8

  640. 2:

  641.         RND_XY2_EXPAND 1, 1

  642. 

  643.         .align 8

  644. 3:

  645.         RND_XY2_EXPAND 2, 1

  646. 

  647.         .align 8

  648. 4:

  649.         RND_XY2_EXPAND 3, 1

  650. 

  651. 5:

  652.         .word 0x03030303

  653.         .word 2b

  654.         .word 3b

  655.         .word 4b

  656.         .word 0x02020202

  657.         .word 0xFCFCFCFC >> 2

  658.         .word 0x0F0F0F0F

  659.         .word 0x01010101

  660. 

  661.         .align 8

  662.         .global put_no_rnd_pixels8_xy2_arm

  663. put_no_rnd_pixels8_xy2_arm:

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

  665.         @ block = word aligned, pixles = unaligned

  666.         pld [r1]

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

  668.         adrl r12, 5f

  669.         ands r4, r1, #3

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

  671.         bic r1, r1, #3

  672.         ldrne pc, [r5]

  673. 1:

  674.         RND_XY2_EXPAND 0, 0

  675. 

  676.         .align 8

  677. 2:

  678.         RND_XY2_EXPAND 1, 0

  679. 

  680.         .align 8

  681. 3:

  682.         RND_XY2_EXPAND 2, 0

  683. 

  684.         .align 8

  685. 4:

  686.         RND_XY2_EXPAND 3, 0

  687. 

  688. 5:

  689.         .word 0x03030303

  690.         .word 2b

  691.         .word 3b

  692.         .word 4b

  693.         .word 0x02020202

  694.         .word 0xFCFCFCFC >> 2

  695.         .word 0x0F0F0F0F

  696.         .word 0x01010101