diff options
Diffstat (limited to 'common/arm')
30 files changed, 273 insertions, 895 deletions
diff --git a/common/arm/ih264_arm_memory_barrier.s b/common/arm/ih264_arm_memory_barrier.s index 523218f..3816409 100755..100644 --- a/common/arm/ih264_arm_memory_barrier.s +++ b/common/arm/ih264_arm_memory_barrier.s @@ -17,7 +17,7 @@ @ ***************************************************************************** @ * Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore @*/ -@/** +@** @******************************************************************************* @* @file @* ih264_arm_memory_barrier.s @@ -39,7 +39,6 @@ .text .p2align 2 - @***************************************************************************** @* @* Function Name : ih264_arm_dsb diff --git a/common/arm/ih264_deblk_chroma_a9.s b/common/arm/ih264_deblk_chroma_a9.s index 66102a7..8c9960a 100755..100644 --- a/common/arm/ih264_deblk_chroma_a9.s +++ b/common/arm/ih264_deblk_chroma_a9.s @@ -54,7 +54,7 @@ .text .p2align 2 -@/** +@** @******************************************************************************* @* @* @brief @@ -84,7 +84,7 @@ @* None @* @******************************************************************************* -@*/ +@* .global ih264_deblk_chroma_horz_bs4_bp_a9 @@ -130,7 +130,7 @@ ih264_deblk_chroma_horz_bs4_bp_a9: -@/** +@** @******************************************************************************* @* @* @brief @@ -160,7 +160,7 @@ ih264_deblk_chroma_horz_bs4_bp_a9: @* None @* @******************************************************************************* -@*/ +@* .global ih264_deblk_chroma_vert_bs4_bp_a9 @@ -224,7 +224,7 @@ ih264_deblk_chroma_vert_bs4_bp_a9: -@/** +@** @******************************************************************************* @* @* @brief @@ -260,7 +260,7 @@ ih264_deblk_chroma_vert_bs4_bp_a9: @* None @* @******************************************************************************* -@*/ +@* .global ih264_deblk_chroma_horz_bslt4_bp_a9 @@ -326,7 +326,7 @@ ih264_deblk_chroma_horz_bslt4_bp_a9: -@/** +@** @******************************************************************************* @* @* @brief @@ -362,7 +362,7 @@ ih264_deblk_chroma_horz_bslt4_bp_a9: @* None @* @******************************************************************************* -@*/ +@* .global ih264_deblk_chroma_vert_bslt4_bp_a9 @@ -465,7 +465,7 @@ ih264_deblk_chroma_vert_bslt4_bp_a9: -@/** +@** @******************************************************************************* @* @* @brief @@ -495,7 +495,7 @@ ih264_deblk_chroma_vert_bslt4_bp_a9: @* None @* @******************************************************************************* -@*/ +@* .global ih264_deblk_chroma_vert_bs4_mbaff_bp_a9 @@ -543,7 +543,7 @@ ih264_deblk_chroma_vert_bs4_mbaff_bp_a9: -@/** +@** @******************************************************************************* @* @* @brief @@ -579,7 +579,7 @@ ih264_deblk_chroma_vert_bs4_mbaff_bp_a9: @* None @* @******************************************************************************* -@*/ +@* .global ih264_deblk_chroma_vert_bslt4_mbaff_bp_a9 @@ -656,7 +656,7 @@ ih264_deblk_chroma_vert_bslt4_mbaff_bp_a9: -@/** +@** @******************************************************************************* @* @* @brief @@ -692,7 +692,7 @@ ih264_deblk_chroma_vert_bslt4_mbaff_bp_a9: @* None @* @******************************************************************************* -@*/ +@* .global ih264_deblk_chroma_horz_bs4_a9 @@ -743,7 +743,7 @@ ih264_deblk_chroma_horz_bs4_a9: -@/** +@** @******************************************************************************* @* @* @brief @@ -779,7 +779,7 @@ ih264_deblk_chroma_horz_bs4_a9: @* None @* @******************************************************************************* -@*/ +@* .global ih264_deblk_chroma_vert_bs4_a9 @@ -848,7 +848,7 @@ ih264_deblk_chroma_vert_bs4_a9: -@/** +@** @******************************************************************************* @* @* @brief @@ -893,7 +893,7 @@ ih264_deblk_chroma_vert_bs4_a9: @* None @* @******************************************************************************* -@*/ +@* .global ih264_deblk_chroma_horz_bslt4_a9 @@ -968,7 +968,7 @@ ih264_deblk_chroma_horz_bslt4_a9: -@/** +@** @******************************************************************************* @* @* @brief @@ -1013,7 +1013,7 @@ ih264_deblk_chroma_horz_bslt4_a9: @* None @* @******************************************************************************* -@*/ +@* .global ih264_deblk_chroma_vert_bslt4_a9 @@ -1119,7 +1119,7 @@ ih264_deblk_chroma_vert_bslt4_a9: -@/** +@** @******************************************************************************* @* @* @brief @@ -1155,7 +1155,7 @@ ih264_deblk_chroma_vert_bslt4_a9: @* None @* @******************************************************************************* -@*/ +@* .global ih264_deblk_chroma_vert_bs4_mbaff_a9 @@ -1206,7 +1206,7 @@ ih264_deblk_chroma_vert_bs4_mbaff_a9: -@/** +@** @******************************************************************************* @* @* @brief @@ -1251,7 +1251,7 @@ ih264_deblk_chroma_vert_bs4_mbaff_a9: @* None @* @******************************************************************************* -@*/ +@* .global ih264_deblk_chroma_vert_bslt4_mbaff_a9 diff --git a/common/arm/ih264_deblk_luma_a9.s b/common/arm/ih264_deblk_luma_a9.s index 3e6a4d9..9217ed2 100755..100644 --- a/common/arm/ih264_deblk_luma_a9.s +++ b/common/arm/ih264_deblk_luma_a9.s @@ -47,7 +47,7 @@ .text .p2align 2 -@/** +@** @******************************************************************************* @* @* @brief @@ -83,7 +83,7 @@ @* None @* @******************************************************************************* -@*/ +@* .global ih264_deblk_luma_horz_bslt4_a9 @@ -187,7 +187,7 @@ ih264_deblk_luma_horz_bslt4_a9: -@/** +@** @******************************************************************************* @* @* @brief @@ -217,7 +217,7 @@ ih264_deblk_luma_horz_bslt4_a9: @* None @* @******************************************************************************* -@*/ +@* .global ih264_deblk_luma_horz_bs4_a9 @@ -353,7 +353,7 @@ ih264_deblk_luma_horz_bs4_a9: -@/** +@** @******************************************************************************* @* @* @brief @@ -389,7 +389,7 @@ ih264_deblk_luma_horz_bs4_a9: @* None @* @******************************************************************************* -@*/ +@* .global ih264_deblk_luma_vert_bslt4_a9 @@ -574,7 +574,7 @@ ih264_deblk_luma_vert_bslt4_a9: -@/** +@** @******************************************************************************* @* @* @brief @@ -604,7 +604,7 @@ ih264_deblk_luma_vert_bslt4_a9: @* None @* @******************************************************************************* -@*/ +@* .global ih264_deblk_luma_vert_bs4_a9 @@ -800,7 +800,7 @@ ih264_deblk_luma_vert_bs4_a9: -@/** +@** @******************************************************************************* @* @* @brief @@ -830,7 +830,7 @@ ih264_deblk_luma_vert_bs4_a9: @* None @* @******************************************************************************* -@*/ +@* .global ih264_deblk_luma_vert_bs4_mbaff_a9 @@ -942,7 +942,7 @@ ih264_deblk_luma_vert_bs4_mbaff_a9: -@/** +@** @******************************************************************************* @* @* @brief @@ -978,7 +978,7 @@ ih264_deblk_luma_vert_bs4_mbaff_a9: @* None @* @******************************************************************************* -@*/ +@* .global ih264_deblk_luma_vert_bslt4_mbaff_a9 diff --git a/common/arm/ih264_default_weighted_pred_a9q.s b/common/arm/ih264_default_weighted_pred_a9q.s index 94cda46..a4688f2 100755..100644 --- a/common/arm/ih264_default_weighted_pred_a9q.s +++ b/common/arm/ih264_default_weighted_pred_a9q.s @@ -17,14 +17,13 @@ @ ***************************************************************************** @ * Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore @*/ -@/** +@** @****************************************************************************** @* @file @* ih264_default_weighted_pred_a9q.s @* @* @brief @* Contains function definitions for default weighted prediction. -@* Functions are coded using NEON intrinsics and can be compiled using ARM RVCT @* @* @author @* Kaushik Senthoor R @@ -38,7 +37,7 @@ @* None @* @******************************************************************************* -@*/ +@* @******************************************************************************* @* @function @* ih264_default_weighted_pred_luma_a9q() @@ -82,7 +81,7 @@ @* (ht,wd) can be (4,4), (4,8), (8,4), (8,8), (8,16), (16,8) or (16,16). @* @******************************************************************************* -@*/ +@* @void ih264_default_weighted_pred_luma_a9q(UWORD8 *pu1_src1, @ UWORD8 *pu1_src2, @ UWORD8 *pu1_dst, @@ -256,7 +255,7 @@ end_loops: @* (ht,wd) can be (2,2), (2,4), (4,2), (4,4), (4,8), (8,4) or (8,8). @* @******************************************************************************* -@*/ +@* @void ih264_default_weighted_pred_chroma_a9q(UWORD8 *pu1_src1, @ UWORD8 *pu1_src2, @ UWORD8 *pu1_dst, diff --git a/common/arm/ih264_ihadamard_scaling_a9.s b/common/arm/ih264_ihadamard_scaling_a9.s index 687099a..c7feddd 100755..100644 --- a/common/arm/ih264_ihadamard_scaling_a9.s +++ b/common/arm/ih264_ihadamard_scaling_a9.s @@ -17,7 +17,7 @@ @ ***************************************************************************** @ * Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore @*/ -@/** +@** @ ******************************************************************************* @ * @file @ * ih264_ihadamard_scaling_a9.s @@ -37,7 +37,7 @@ @ * None @ * @ ******************************************************************************* -@ */ +@ * @ * @brief This function performs a 4x4 inverse hadamard transform on the 4x4 DC coefficients @ * of a 16x16 intra prediction macroblock, and then performs scaling. @ * prediction buffer @@ -69,10 +69,10 @@ @ * @remarks none @ * @ ******************************************************************************* -@ */ +@ * @ * @ ******************************************************************************* -@ */ +@ * @ void ih264_ihadamard_scaling_4x4(WORD16* pi2_src, @ WORD16* pi2_out, @ const UWORD16 *pu2_iscal_mat, @@ -161,7 +161,7 @@ ih264_ihadamard_scaling_4x4_a9: @ ******************************************************************************* -@ */ +@ * @ * @brief This function performs a 2x2 inverse hadamard transform for chroma block @ * @ * @par Description: @@ -189,10 +189,10 @@ ih264_ihadamard_scaling_4x4_a9: @ * @remarks none @ * @ ******************************************************************************* -@ */ +@ * @ * @ ******************************************************************************* -@ */ +@ * @ void ih264_ihadamard_scaling_2x2_uv(WORD16* pi2_src, @ WORD16* pi2_out, @ const UWORD16 *pu2_iscal_mat, diff --git a/common/arm/ih264_inter_pred_chroma_a9q.s b/common/arm/ih264_inter_pred_chroma_a9q.s index afd2860..6681a7c 100755..100644 --- a/common/arm/ih264_inter_pred_chroma_a9q.s +++ b/common/arm/ih264_inter_pred_chroma_a9q.s @@ -17,7 +17,7 @@ @ ***************************************************************************** @ * Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore @*/ -@/** +@** @****************************************************************************** @* @file @* ih264_inter_pred_chroma_a9q.s @@ -36,16 +36,16 @@ @* None @* @******************************************************************************* -@*/ +@* -@/* All the functions here are replicated from ih264_inter_pred_filters.c +@* All the functions here are replicated from ih264_inter_pred_filters.c @ -@/** -@/** -@/** +@** +@** +@** @ -@/** +@** @******************************************************************************* @* @* @brief @@ -85,7 +85,7 @@ @* None @* @******************************************************************************* -@*/ +@* @void ih264_inter_pred_chroma(UWORD8 *pu1_src, @ UWORD8 *pu1_dst, @@ -112,8 +112,6 @@ ih264_inter_pred_chroma_a9q: - - stmfd sp!, {r4-r12, r14} @store register values to stack vstmdb sp!, {d8-d15} @push neon registers to stack ldr r4, [sp, #104] diff --git a/common/arm/ih264_inter_pred_filters_luma_horz_a9q.s b/common/arm/ih264_inter_pred_filters_luma_horz_a9q.s index ea6bba0..62b4b94 100755..100644 --- a/common/arm/ih264_inter_pred_filters_luma_horz_a9q.s +++ b/common/arm/ih264_inter_pred_filters_luma_horz_a9q.s @@ -17,7 +17,7 @@ @ ***************************************************************************** @ * Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore @*/ -@/** +@** @****************************************************************************** @* @file @* ih264_inter_pred_luma_horz_a9q.s @@ -36,13 +36,13 @@ @* None @* @******************************************************************************* -@*/ +@* -@/* All the functions here are replicated from ih264_inter_pred_filters.c +@* All the functions here are replicated from ih264_inter_pred_filters.c @ -@/** -@/** +@** +@** @******************************************************************************* @* @* @brief @@ -76,7 +76,7 @@ @* None @* @******************************************************************************* -@*/ +@* @void ih264_inter_pred_luma_horz ( @ UWORD8 *pu1_src, @@ -102,6 +102,9 @@ ih264_inter_pred_luma_horz_a9q: + + + stmfd sp!, {r4-r12, r14} @store register values to stack vstmdb sp!, {d8-d15} @push neon registers to stack ldr r5, [sp, #104] @Loads ht @@ -116,7 +119,7 @@ ih264_inter_pred_luma_horz_a9q: beq loop_4 loop_16: @when wd=16 - @// Processing row0 and row1 + @ Processing row0 and row1 vld1.8 {d2, d3, d4}, [r0], r2 @// Load row0 ;for checking loop vext.8 d31, d2, d3, #5 @//extract a[5] (column1,row0) vld1.8 {d5, d6, d7}, [r0], r2 @// Load row1 @@ -173,7 +176,7 @@ loop_16: @when wd=16 b loop_16 @ loop if height == 8 or 16 loop_8: -@// Processing row0 and row1 +@ Processing row0 and row1 vld1.8 {d5, d6}, [r0], r2 @// Load row1 vext.8 d28, d5, d6, #5 @//extract a[5] (column1,row1) vld1.8 {d2, d3}, [r0], r2 @// Load row0 @@ -204,7 +207,7 @@ loop_8: beq end_func @ Branch if height==4 - b loop_8 @looping if height =8 or 16 + b loop_8 @looping if height =8 or 16 loop_4: vld1.8 {d5, d6}, [r0], r2 @// Load row1 diff --git a/common/arm/ih264_inter_pred_filters_luma_vert_a9q.s b/common/arm/ih264_inter_pred_filters_luma_vert_a9q.s index 5b29e02..65c40a6 100755..100644 --- a/common/arm/ih264_inter_pred_filters_luma_vert_a9q.s +++ b/common/arm/ih264_inter_pred_filters_luma_vert_a9q.s @@ -17,7 +17,7 @@ @ ***************************************************************************** @ * Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore @*/ -@/** +@** @****************************************************************************** @* @file @* ih264_inter_pred_luma_vert_a9q.s @@ -36,14 +36,14 @@ @* None @* @******************************************************************************* -@*/ +@* -@/* All the functions here are replicated from ih264_inter_pred_filters.c +@* All the functions here are replicated from ih264_inter_pred_filters.c @ -@/** -@/** -@/** +@** +@** +@** @ ******************************************************************************* @ * @ * @brief @@ -195,10 +195,10 @@ loop_16: @when wd=16 subne r0, r0, r2 beq end_func @ Branch if height==4 - b loop_16 @ looping if height = 8 or 16 + b loop_16 @ looping if height = 8 or 16 loop_8: -@// Processing row0 and row1 +@ Processing row0 and row1 vld1.u32 d0, [r0], r2 @ Vector load from src[0_0] vld1.u32 d1, [r0], r2 @ Vector load from src[1_0] @@ -248,7 +248,7 @@ loop_8: loop_4: -@// Processing row0 and row1 +@ Processing row0 and row1 vld1.u32 d0[0], [r0], r2 @ Vector load from src[0_0] vld1.u32 d1[0], [r0], r2 @ Vector load from src[1_0] diff --git a/common/arm/ih264_inter_pred_luma_bilinear_a9q.s b/common/arm/ih264_inter_pred_luma_bilinear_a9q.s index 6a3c83d..8f049f8 100755..100644 --- a/common/arm/ih264_inter_pred_luma_bilinear_a9q.s +++ b/common/arm/ih264_inter_pred_luma_bilinear_a9q.s @@ -17,7 +17,7 @@ @ ***************************************************************************** @ * Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore @*/ -@/** +@** @****************************************************************************** @* @file @* ih264_inter_pred_luma_bilinear_a9q.s @@ -36,14 +36,14 @@ @* None @* @******************************************************************************* -@*/ +@* -@/* All the functions here are replicated from ih264_inter_pred_filters.c +@* All the functions here are replicated from ih264_inter_pred_filters.c @ -@/** -@/** -@/** +@** +@** +@** @ ******************************************************************************* @ * function:ih264_inter_pred_luma_bilinear @ * @@ -89,7 +89,7 @@ @* None @* @******************************************************************************* -@*/ +@* @void ih264_inter_pred_luma_bilinear(UWORD8 *pu1_src1, @ UWORD8 *pu1_src2, @@ -192,7 +192,7 @@ loop_16: @when wd=16 subs r12, r6, #8 vst1.8 {q15}, [r2], r5 @//Store dest row7 - beq end_func @ end function if ht=8 + beq end_func @ end function if ht=8 vld1.8 {q0}, [r0], r3 @// Load row8 ;src1 vaddl.u8 q10, d0, d4 @@ -275,7 +275,7 @@ loop_8: @wd=8; vqrshrun.s16 d31, q13, #1 subs r12, r6, #4 vst1.8 {d31}, [r2], r5 @//Store dest row3 - beq end_func @ end function if ht=4 + beq end_func @ end function if ht=4 vld1.8 {d12}, [r1], r4 @// Load row4 ;src2 vld1.8 {d8}, [r0], r3 @// Load row4 ;src1 @@ -298,7 +298,7 @@ loop_8: @wd=8; vqrshrun.s16 d31, q11, #1 subs r12, r6, #8 vst1.8 {d31}, [r2], r5 @//Store dest row7 - beq end_func @ end function if ht=8 + beq end_func @ end function if ht=8 vld1.8 {d0}, [r0], r3 @// Load row8 ;src1 vld1.8 {d4}, [r1], r4 @// Load row8 ;src2 @@ -367,7 +367,7 @@ loop_4: vqrshrun.s16 d31, q13, #1 subs r12, r6, #4 vst1.32 d31[0], [r2], r5 @//Store dest row3 - beq end_func @ end function if ht=4 + beq end_func @ end function if ht=4 vld1.32 d12[0], [r1], r4 @// Load row4 ;src2 vld1.32 d8[0], [r0], r3 @// Load row4 ;src1 diff --git a/common/arm/ih264_inter_pred_luma_copy_a9q.s b/common/arm/ih264_inter_pred_luma_copy_a9q.s index 8ba2fbf..c0b0568 100755..100644 --- a/common/arm/ih264_inter_pred_luma_copy_a9q.s +++ b/common/arm/ih264_inter_pred_luma_copy_a9q.s @@ -17,8 +17,8 @@ @ ***************************************************************************** @ * Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore @*/ -@/** -@/** +@** +@** @******************************************************************************* @* @* @brief @@ -53,7 +53,7 @@ @* None @* @******************************************************************************* -@*/ +@* @void ih264_inter_pred_luma_copy ( @ UWORD8 *pu1_src, @ UWORD8 *pu1_dst, @@ -182,7 +182,7 @@ end_inner_loop_wd_16: ldmfd sp!, {r4-r12, r15} @Reload the registers from SP -@ /* +@ * @ ******************************************************************************** @ * @ * @brief This function copies a 4x4 block to destination @@ -208,7 +208,7 @@ end_inner_loop_wd_16: @ * Currently wd and height is not used, ie a 4x4 block is always copied @ * @ ******************************************************************************* -@ */ +@ * @ void ih264_interleave_copy(WORD16 *pi2_src, @ UWORD8 *pu1_out, @ WORD32 pred_strd, diff --git a/common/arm/ih264_inter_pred_luma_horz_hpel_vert_hpel_a9q.s b/common/arm/ih264_inter_pred_luma_horz_hpel_vert_hpel_a9q.s index 43321a8..54183f0 100755..100644 --- a/common/arm/ih264_inter_pred_luma_horz_hpel_vert_hpel_a9q.s +++ b/common/arm/ih264_inter_pred_luma_horz_hpel_vert_hpel_a9q.s @@ -17,7 +17,7 @@ @ ***************************************************************************** @ * Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore @*/ -@/** +@** @****************************************************************************** @* @file @* ih264_inter_pred_luma_horz_hpel_vert_hpel_a9q.s @@ -36,14 +36,14 @@ @* None @* @******************************************************************************* -@*/ +@* -@/* All the functions here are replicated from ih264_inter_pred_filters.c +@* All the functions here are replicated from ih264_inter_pred_filters.c @ -@/** -@/** -@/** +@** +@** +@** @******************************************************************************* @* @* @brief @@ -88,7 +88,7 @@ @* None @* @******************************************************************************* -@*/; +@*; @void ih264_inter_pred_luma_horz_hpel_vert_hpel(UWORD8 *pu1_src, @ UWORD8 *pu1_dst, diff --git a/common/arm/ih264_inter_pred_luma_horz_hpel_vert_qpel_a9q.s b/common/arm/ih264_inter_pred_luma_horz_hpel_vert_qpel_a9q.s index 65a6de7..c8edf38 100755..100644 --- a/common/arm/ih264_inter_pred_luma_horz_hpel_vert_qpel_a9q.s +++ b/common/arm/ih264_inter_pred_luma_horz_hpel_vert_qpel_a9q.s @@ -17,7 +17,7 @@ @ ***************************************************************************** @ * Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore @*/ -@/** +@** @****************************************************************************** @* @file @* ih264_inter_pred_luma_horz_hpel_vert_qpel_a9q.s @@ -36,14 +36,14 @@ @* None @* @******************************************************************************* -@*/ +@* -@/* All the functions here are replicated from ih264_inter_pred_filters.c +@* All the functions here are replicated from ih264_inter_pred_filters.c @ -@/** -@/** -@/** +@** +@** +@** @******************************************************************************* @* @* @brief @@ -91,7 +91,7 @@ @* None @* @******************************************************************************* -@*/; +@*; @void ih264_inter_pred_luma_horz_hpel_vert_qpel(UWORD8 *pu1_src, @ UWORD8 *pu1_dst, @@ -835,7 +835,7 @@ loop_8: vmov q7, q14 vst1.32 d30, [r1], r3 @ store row 3 - bgt loop_8 @if height =8 or 16 loop + bgt loop_8 @if height =8 or 16 loop b end_func loop_4_start: diff --git a/common/arm/ih264_inter_pred_luma_horz_qpel_a9q.s b/common/arm/ih264_inter_pred_luma_horz_qpel_a9q.s index c39ae01..ab1d1d1 100755..100644 --- a/common/arm/ih264_inter_pred_luma_horz_qpel_a9q.s +++ b/common/arm/ih264_inter_pred_luma_horz_qpel_a9q.s @@ -17,7 +17,7 @@ @ ***************************************************************************** @ * Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore @*/ -@/** +@** @****************************************************************************** @* @file @* ih264_inter_pred_luma_horz_qpel_a9q.s @@ -30,19 +30,19 @@ @* @* @par List of Functions: @* -@* - ih264_inter_pred_luma_horz_qpe_a9ql() +@* - ih264_inter_pred_luma_horz_qpel_a9q() @* @* @remarks @* None @* @******************************************************************************* -@*/ +@* -@/* All the functions here are replicated from ih264_inter_pred_filters.c +@* All the functions here are replicated from ih264_inter_pred_filters.c @ -@/** -@/** +@** +@** @******************************************************************************* @* @* @brief @@ -79,7 +79,7 @@ @* None @* @******************************************************************************* -@*/ +@* @void ih264_inter_pred_luma_horz ( @ UWORD8 *pu1_src, @@ -126,7 +126,7 @@ ih264_inter_pred_luma_horz_qpel_a9q: beq loop_4 loop_16: @when wd=16 - @// Processing row0 and row1 + @ Processing row0 and row1 vld1.8 {d2, d3, d4}, [r0], r2 @// Load row0 vext.8 d31, d2, d3, #5 @//extract a[5] (column1,row0) vld1.8 {d5, d6, d7}, [r0], r2 @// Load row1 @@ -187,7 +187,7 @@ loop_16: @when wd=16 b loop_16 loop_8: -@// Processing row0 and row1 +@ Processing row0 and row1 vld1.8 {d5, d6}, [r0], r2 @// Load row1 vext.8 d28, d5, d6, #5 @//extract a[5] (column1,row1) @@ -221,7 +221,7 @@ loop_8: subs r5, r5, #2 @ 2 rows done, decrement by 2 beq end_func @ Branch if height==4 - b loop_8 @looping if height == 8 or 16 + b loop_8 @looping if height == 8 or 16 loop_4: vld1.8 {d5, d6}, [r0], r2 @// Load row1 diff --git a/common/arm/ih264_inter_pred_luma_horz_qpel_vert_hpel_a9q.s b/common/arm/ih264_inter_pred_luma_horz_qpel_vert_hpel_a9q.s index 565cc80..3c63ca3 100755..100644 --- a/common/arm/ih264_inter_pred_luma_horz_qpel_vert_hpel_a9q.s +++ b/common/arm/ih264_inter_pred_luma_horz_qpel_vert_hpel_a9q.s @@ -17,7 +17,7 @@ @ ***************************************************************************** @ * Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore @*/ -@/** +@** @****************************************************************************** @* @file @* ih264_inter_pred_luma_horz_qpel_vert_hpel_a9q.s @@ -36,14 +36,14 @@ @* None @* @******************************************************************************* -@*/ +@* -@/* All the functions here are replicated from ih264_inter_pred_filters.c +@* All the functions here are replicated from ih264_inter_pred_filters.c @ -@/** -@/** -@/** +@** +@** +@** @******************************************************************************* @* @* @brief @@ -91,7 +91,7 @@ @* None @* @******************************************************************************* -@*/; +@*; @void ih264_inter_pred_luma_horz_qpel_vert_hpel(UWORD8 *pu1_src, @ UWORD8 *pu1_dst, diff --git a/common/arm/ih264_inter_pred_luma_horz_qpel_vert_qpel_a9q.s b/common/arm/ih264_inter_pred_luma_horz_qpel_vert_qpel_a9q.s index 3c8b60a..cfe03a0 100755..100644 --- a/common/arm/ih264_inter_pred_luma_horz_qpel_vert_qpel_a9q.s +++ b/common/arm/ih264_inter_pred_luma_horz_qpel_vert_qpel_a9q.s @@ -17,7 +17,7 @@ @ ***************************************************************************** @ * Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore @*/ -@/** +@** @****************************************************************************** @* @file @* ih264_inter_pred_luma_horz_qpel_vert_qpel_a9q.s @@ -36,14 +36,11 @@ @* None @* @******************************************************************************* -@*/ +@* -@/* All the functions here are replicated from ih264_inter_pred_filters.c +@* All the functions here are replicated from ih264_inter_pred_filters.c @ -@/** -@/** -@/** @******************************************************************************* @* @* @brief @@ -90,7 +87,7 @@ @* None @* @******************************************************************************* -@*/; +@*; @void ih264_inter_pred_luma_horz_qpel_vert_qpel(UWORD8 *pu1_src, @ UWORD8 *pu1_dst, diff --git a/common/arm/ih264_inter_pred_luma_vert_qpel_a9q.s b/common/arm/ih264_inter_pred_luma_vert_qpel_a9q.s index d45055e..e2c68ef 100755..100644 --- a/common/arm/ih264_inter_pred_luma_vert_qpel_a9q.s +++ b/common/arm/ih264_inter_pred_luma_vert_qpel_a9q.s @@ -17,7 +17,7 @@ @ ***************************************************************************** @ * Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore @*/ -@/** +@** @****************************************************************************** @* @file @* ih264_inter_pred_luma_vert_qpel_a9q.s @@ -36,13 +36,11 @@ @* None @* @******************************************************************************* -@*/ +@* -@/* All the functions here are replicated from ih264_inter_pred_filters.c +@* All the functions here are replicated from ih264_inter_pred_filters.c @ -@/** -@/** @******************************************************************************* @* @* @brief @@ -79,7 +77,7 @@ @* None @* @******************************************************************************* -@*/ +@* @void ih264_inter_pred_luma_vert ( @ UWORD8 *pu1_src, @@ -211,12 +209,12 @@ loop_16: @when wd=16 subne r0, r0, r2 beq end_func @ Branch if height==4 - b loop_16 @ looping if height = 8 or 16 + b loop_16 @ looping if height = 8 or 16 loop_8: - @// Processing row0 and row1 + @ Processing row0 and row1 vld1.u32 d0, [r0], r2 @ Vector load from src[0_0] vld1.u32 d1, [r0], r2 @ Vector load from src[1_0] vld1.u32 d2, [r0], r2 @ Vector load from src[2_0] @@ -270,7 +268,7 @@ loop_8: b loop_8 @looping if height == 8 or 16 loop_4: -@// Processing row0 and row1 +@ Processing row0 and row1 vld1.u32 d0[0], [r0], r2 @ Vector load from src[0_0] vld1.u32 d1[0], [r0], r2 @ Vector load from src[1_0] diff --git a/common/arm/ih264_intra_pred_chroma_a9q.s b/common/arm/ih264_intra_pred_chroma_a9q.s index d03fc55..ccd5c0d 100755..100644 --- a/common/arm/ih264_intra_pred_chroma_a9q.s +++ b/common/arm/ih264_intra_pred_chroma_a9q.s @@ -17,7 +17,7 @@ @ ***************************************************************************** @ * Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore @*/ -@/** +@** @****************************************************************************** @* @file @* ih264_intra_pred_chroma_a9q.s @@ -39,15 +39,11 @@ @* None @* @******************************************************************************* -@*/ +@* -@/* All the functions here are replicated from ih264_chroma_intra_pred_filters.c +@* All the functions here are replicated from ih264_chroma_intra_pred_filters.c @ -@/** -@/** -@/** -@ .text .p2align 2 @@ -60,7 +56,7 @@ scratch_chroma_intrapred_addr1: scratch_intrapred_chroma_plane_addr1: .long ih264_gai1_intrapred_chroma_plane_coeffs2 - scrlblc2 - 8 -@/** +@** @******************************************************************************* @* @*ih264_intra_pred_chroma_8x8_mode_dc @@ -91,7 +87,7 @@ scratch_intrapred_chroma_plane_addr1: @* @remarks @* None @* -@*******************************************************************************/ +@******************************************************************************* @void ih264_intra_pred_chroma_8x8_mode_dc(UWORD8 *pu1_src, @ UWORD8 *pu1_dst, @ WORD32 src_strd, @@ -105,8 +101,6 @@ scratch_intrapred_chroma_plane_addr1: @ r3 => dst_strd @ r4 => ui_neighboravailability - - .global ih264_intra_pred_chroma_8x8_mode_dc_a9q ih264_intra_pred_chroma_8x8_mode_dc_a9q: @@ -191,10 +185,10 @@ str_pred: -@/****************************************************************************** +@****************************************************************************** -@/** +@** @******************************************************************************* @* @*ih264_intra_pred_chroma_8x8_mode_horz @@ -226,7 +220,7 @@ str_pred: @* None @* @******************************************************************************* -@*/ +@* @void ih264_intra_pred_chroma_8x8_mode_horz(UWORD8 *pu1_src, @ UWORD8 *pu1_dst, @ WORD32 src_strd, @@ -270,7 +264,7 @@ loop_8x8_horz: -@/** +@** @******************************************************************************* @* @*ih264_intra_pred_chroma_8x8_mode_vert @@ -339,10 +333,10 @@ ih264_intra_pred_chroma_8x8_mode_vert_a9q: -@/****************************************************************************** +@****************************************************************************** -@/** +@** @******************************************************************************* @* @*ih264_intra_pred_chroma_8x8_mode_plane @@ -373,7 +367,7 @@ ih264_intra_pred_chroma_8x8_mode_vert_a9q: @* @remarks @* None @* -@*******************************************************************************/ +@******************************************************************************* @void ih264_intra_pred_chroma_8x8_mode_plane(UWORD8 *pu1_src, @ UWORD8 *pu1_dst, @ WORD32 src_strd, @@ -393,7 +387,6 @@ ih264_intra_pred_chroma_8x8_mode_plane_a9q: stmfd sp!, {r4-r10, r12, lr} vpush {d8-d15} - vld1.32 d0, [r0] add r10, r0, #10 vld1.32 d1, [r10] @@ -542,7 +535,6 @@ scrlblc2: end_func_plane: - vpop {d8-d15} ldmfd sp!, {r4-r10, r12, pc} diff --git a/common/arm/ih264_intra_pred_luma_16x16_a9q.s b/common/arm/ih264_intra_pred_luma_16x16_a9q.s index e38e203..0dd82f3 100755..100644 --- a/common/arm/ih264_intra_pred_luma_16x16_a9q.s +++ b/common/arm/ih264_intra_pred_luma_16x16_a9q.s @@ -17,7 +17,7 @@ @ ***************************************************************************** @ * Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore @*/ -@/** +@** @****************************************************************************** @* @file @* ih264_intra_pred_luma_16x16_a9q.s @@ -39,14 +39,14 @@ @* None @* @******************************************************************************* -@*/ +@* -@/* All the functions here are replicated from ih264_intra_pred_filters.c +@* All the functions here are replicated from ih264_intra_pred_filters.c @ -@/** -@/** -@/** +@** +@** +@** @ .text @@ -57,10 +57,10 @@ .hidden ih264_gai1_intrapred_luma_plane_coeffs scratch_intrapred_addr1: .long ih264_gai1_intrapred_luma_plane_coeffs - scrlbl1 - 8 -@/** +@** @******************************************************************************* @* -@*ih264_intra_pred_luma_16x16_mode_vert_a9q +@*ih264_intra_pred_luma_16x16_mode_vert @* @* @brief @* Perform Intra prediction for luma_16x16 mode:vertical @@ -135,13 +135,13 @@ ih264_intra_pred_luma_16x16_mode_vert_a9q: -@/****************************************************************************** +@****************************************************************************** -@/** +@** @******************************************************************************* @* -@*ih264_intra_pred_luma_16x16_mode_horz_a9q +@*ih264_intra_pred_luma_16x16_mode_horz @* @* @brief @* Perform Intra prediction for luma_16x16 mode:horizontal @@ -170,7 +170,7 @@ ih264_intra_pred_luma_16x16_mode_vert_a9q: @* None @* @******************************************************************************* -@*/ +@* @void ih264_intra_pred_luma_16x16_mode_horz(UWORD8 *pu1_src, @ UWORD8 *pu1_dst, @ WORD32 src_strd, @@ -213,13 +213,13 @@ loop_16x16_horz: -@/****************************************************************************** +@****************************************************************************** -@/** +@** @******************************************************************************* @* -@*ih264_intra_pred_luma_16x16_mode_dc_a9q +@*ih264_intra_pred_luma_16x16_mode_dc @* @* @brief @* Perform Intra prediction for luma_16x16 mode:DC @@ -247,7 +247,7 @@ loop_16x16_horz: @* @remarks @* None @* -@*******************************************************************************/ +@******************************************************************************* @void ih264_intra_pred_luma_16x16_mode_dc(UWORD8 *pu1_src, @ UWORD8 *pu1_dst, @ WORD32 src_strd, @@ -300,7 +300,7 @@ top_available: @ONLY TOP AVAILABLE vdup.u8 q0, d0[0] b str_pred -left_available: @ONLY LEFT AVAILABLE +left_available: @ONLY LEFT AVAILABLE vld1.u8 {q0}, [r0] vpaddl.u8 q0, q0 vadd.u16 d0, d0, d1 @@ -337,13 +337,13 @@ str_pred: -@/****************************************************************************** +@****************************************************************************** -@/** +@** @******************************************************************************* @* -@*ih264_intra_pred_luma_16x16_mode_plane_a9q +@*ih264_intra_pred_luma_16x16_mode_plane @* @* @brief @* Perform Intra prediction for luma_16x16 mode:PLANE @@ -371,7 +371,7 @@ str_pred: @* @remarks @* None @* -@*******************************************************************************/ +@******************************************************************************* @void ih264_intra_pred_luma_16x16_mode_plane(UWORD8 *pu1_src, @ UWORD8 *pu1_dst, @ WORD32 src_strd, diff --git a/common/arm/ih264_intra_pred_luma_4x4_a9q.s b/common/arm/ih264_intra_pred_luma_4x4_a9q.s index cb386ea..5cc7e23 100755..100644 --- a/common/arm/ih264_intra_pred_luma_4x4_a9q.s +++ b/common/arm/ih264_intra_pred_luma_4x4_a9q.s @@ -17,7 +17,7 @@ @ ***************************************************************************** @ * Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore @*/ -@/** +@** @****************************************************************************** @* @file @* ih264_intra_pred_luma_4x4_a9q.s @@ -44,21 +44,16 @@ @* None @* @******************************************************************************* -@*/ - -@/* All the functions here are replicated from ih264_intra_pred_filters.c -@ +@* -@/** -@/** -@/** +@* All the functions here are replicated from ih264_intra_pred_filters.c @ .text .p2align 2 -@/** +@** @******************************************************************************* @* @*ih264_intra_pred_luma_4x4_mode_vert @@ -128,10 +123,10 @@ ih264_intra_pred_luma_4x4_mode_vert_a9q: -@/****************************************************************************** +@****************************************************************************** -@/** +@** @******************************************************************************* @* @*ih264_intra_pred_luma_4x4_mode_horz @@ -163,7 +158,7 @@ ih264_intra_pred_luma_4x4_mode_vert_a9q: @* None @* @******************************************************************************* -@*/ +@* @void ih264_intra_pred_luma_4x4_mode_horz(UWORD8 *pu1_src, @ UWORD8 *pu1_dst, @ WORD32 src_strd, @@ -210,10 +205,10 @@ ih264_intra_pred_luma_4x4_mode_horz_a9q: -@/****************************************************************************** +@****************************************************************************** -@/** +@** @******************************************************************************* @* @*ih264_intra_pred_luma_4x4_mode_dc @@ -244,7 +239,7 @@ ih264_intra_pred_luma_4x4_mode_horz_a9q: @* @remarks @* None @* -@*******************************************************************************/ +@******************************************************************************* @void ih264_intra_pred_luma_4x4_mode_dc(UWORD8 *pu1_src, @ UWORD8 *pu1_dst, @ WORD32 src_strd, @@ -352,7 +347,7 @@ end_func: -@/** +@** @******************************************************************************* @* @*ih264_intra_pred_luma_4x4_mode_diag_dl @@ -383,7 +378,7 @@ end_func: @* @remarks @* None @* -@*******************************************************************************/ +@******************************************************************************* @void ih264_intra_pred_luma_4x4_mode_diag_dl(UWORD8 *pu1_src, @ UWORD8 *pu1_dst, @ WORD32 src_strd, @@ -434,7 +429,7 @@ end_func_diag_dl: -@/** +@** @******************************************************************************* @* @*ih264_intra_pred_luma_4x4_mode_diag_dr @@ -465,7 +460,7 @@ end_func_diag_dl: @* @remarks @* None @* -@*******************************************************************************/ +@******************************************************************************* @void ih264_intra_pred_luma_4x4_mode_diag_dr(UWORD8 *pu1_src, @ UWORD8 *pu1_dst, @ WORD32 src_strd, @@ -514,7 +509,7 @@ end_func_diag_dr: -@/** +@** @******************************************************************************* @* @*ih264_intra_pred_luma_4x4_mode_vert_r @@ -545,7 +540,7 @@ end_func_diag_dr: @* @remarks @* None @* -@*******************************************************************************/ +@******************************************************************************* @void ih264_intra_pred_luma_4x4_mode_vert_r(UWORD8 *pu1_src, @ UWORD8 *pu1_dst, @ WORD32 src_strd, @@ -596,7 +591,7 @@ end_func_vert_r: -@/** +@** @******************************************************************************* @* @*ih264_intra_pred_luma_4x4_mode_horz_d @@ -627,7 +622,7 @@ end_func_vert_r: @* @remarks @* None @* -@*******************************************************************************/ +@******************************************************************************* @void ih264_intra_pred_luma_4x4_mode_horz_d(UWORD8 *pu1_src, @ UWORD8 *pu1_dst, @ WORD32 src_strd, @@ -659,7 +654,7 @@ ih264_intra_pred_luma_4x4_mode_horz_d_a9q: vqrshrun.s16 d5, q12, #2 sub r5, r3, #2 vmov.8 d6, d5 - vtrn.8 d4, d5 @ + vtrn.8 d4, d5 @ vst1.u16 {d5[1]}, [r1]! vst1.16 {d6[2]}, [r1], r5 vst1.u16 {d4[1]}, [r1]! @@ -678,7 +673,7 @@ end_func_horz_d: -@/** +@** @******************************************************************************* @* @*ih264_intra_pred_luma_4x4_mode_vert_l @@ -709,7 +704,7 @@ end_func_horz_d: @* @remarks @* None @* -@*******************************************************************************/ +@******************************************************************************* @void ih264_intra_pred_luma_4x4_mode_vert_l(UWORD8 *pu1_src, @ UWORD8 *pu1_dst, @ WORD32 src_strd, @@ -759,7 +754,7 @@ end_func_vert_l: -@/** +@** @******************************************************************************* @* @*ih264_intra_pred_luma_4x4_mode_horz_u @@ -790,7 +785,7 @@ end_func_vert_l: @* @remarks @* None @* -@*******************************************************************************/ +@******************************************************************************* @void ih264_intra_pred_luma_4x4_mode_horz_u(UWORD8 *pu1_src, @ UWORD8 *pu1_dst, @ WORD32 src_strd, @@ -825,9 +820,9 @@ ih264_intra_pred_luma_4x4_mode_horz_u_a9q: vext.8 d6, d5, d4, #1 vst1.8 {d4[2]}, [r1]! vst1.8 {d6[0]}, [r1]! - vtrn.8 d6, d5 @ + vtrn.8 d6, d5 @ sub r5, r3, #2 - vtrn.8 d4, d6 @ + vtrn.8 d4, d6 @ vdup.8 d7, r9 vst1.16 {d6[0]}, [r1], r5 vst1.16 {d6[0]}, [r1]! diff --git a/common/arm/ih264_intra_pred_luma_8x8_a9q.s b/common/arm/ih264_intra_pred_luma_8x8_a9q.s index 6da1c95..352d29d 100755..100644 --- a/common/arm/ih264_intra_pred_luma_8x8_a9q.s +++ b/common/arm/ih264_intra_pred_luma_8x8_a9q.s @@ -17,7 +17,7 @@ @ ***************************************************************************** @ * Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore @*/ -@/** +@** @****************************************************************************** @* @file @* ih264_intra_pred_luma_8x8_a9q.s @@ -45,17 +45,11 @@ @* None @* @******************************************************************************* -@*/ - -@/* All the functions here are replicated from ih264_intra_pred_filters.c -@ +@* -@/** -@/** -@/** +@* All the functions here are replicated from ih264_intra_pred_filters.c @ - .text .p2align 2 @@ -64,7 +58,7 @@ scratch_intrapred_addr_8x8: .long ih264_gai1_intrapred_luma_8x8_horz_u - scrlb8x8l2 - 8 -@/** +@** @******************************************************************************* @* @*ih264_intra_pred_luma_8x8_mode_ref_filtering @@ -95,7 +89,7 @@ scratch_intrapred_addr_8x8: @* @remarks @* None @* -@*******************************************************************************/ +@******************************************************************************* @void ih264_intra_pred_luma_8x8_mode_ref_filtering(UWORD8 *pu1_src, @ UWORD8 *pu1_dst) @@ -111,7 +105,6 @@ ih264_intra_pred_luma_8x8_mode_ref_filtering_a9q: stmfd sp!, {r4-r12, r14} @store register values to stack vpush {d8-d15} - vld1.u8 {q0}, [r0]! @ vld1.u8 {q1}, [r0] add r0, r0, #8 @ @@ -141,6 +134,7 @@ ih264_intra_pred_luma_8x8_mode_ref_filtering_a9q: end_func_ref_filt: + vpop {d8-d15} ldmfd sp!, {r4-r12, pc} @Restoring registers from stack @@ -149,7 +143,7 @@ end_func_ref_filt: -@/** +@** @******************************************************************************* @* @*ih264_intra_pred_luma_8x8_mode_vert @@ -219,10 +213,10 @@ ih264_intra_pred_luma_8x8_mode_vert_a9q: -@/****************************************************************************** +@****************************************************************************** -@/** +@** @******************************************************************************* @* @*ih264_intra_pred_luma_8x8_mode_horz @@ -254,7 +248,7 @@ ih264_intra_pred_luma_8x8_mode_vert_a9q: @* None @* @******************************************************************************* -@*/ +@* @void ih264_intra_pred_luma_8x8_mode_horz(UWORD8 *pu1_src, @ UWORD8 *pu1_dst, @ WORD32 src_strd, @@ -299,10 +293,10 @@ loop_8x8_horz: -@/****************************************************************************** +@****************************************************************************** -@/** +@** @******************************************************************************* @* @*ih264_intra_pred_luma_8x8_mode_dc @@ -333,7 +327,7 @@ loop_8x8_horz: @* @remarks @* None @* -@*******************************************************************************/ +@******************************************************************************* @void ih264_intra_pred_luma_8x8_mode_dc(UWORD8 *pu1_src, @ UWORD8 *pu1_dst, @ WORD32 src_strd, @@ -413,7 +407,7 @@ str_pred: -@/** +@** @******************************************************************************* @* @*ih264_intra_pred_luma_8x8_mode_diag_dl @@ -444,7 +438,7 @@ str_pred: @* @remarks @* None @* -@*******************************************************************************/ +@******************************************************************************* @void ih264_intra_pred_luma_8x8_mode_diag_dl(UWORD8 *pu1_src, @ UWORD8 *pu1_dst, @ WORD32 src_strd, @@ -506,7 +500,7 @@ end_func_diag_dl: -@/** +@** @******************************************************************************* @* @*ih264_intra_pred_luma_8x8_mode_diag_dr @@ -537,7 +531,7 @@ end_func_diag_dl: @* @remarks @* None @* -@*******************************************************************************/ +@******************************************************************************* @void ih264_intra_pred_luma_8x8_mode_diag_dr(UWORD8 *pu1_src, @ UWORD8 *pu1_dst, @ WORD32 src_strd, @@ -597,7 +591,7 @@ end_func_diag_dr: -@/** +@** @******************************************************************************* @* @*ih264_intra_pred_luma_8x8_mode_vert_r @@ -628,7 +622,7 @@ end_func_diag_dr: @* @remarks @* None @* -@*******************************************************************************/ +@******************************************************************************* @void ih264_intra_pred_luma_8x8_mode_vert_r(UWORD8 *pu1_src, @ UWORD8 *pu1_dst, @ WORD32 src_strd, @@ -717,7 +711,7 @@ end_func_vert_r: -@/** +@** @******************************************************************************* @* @*ih264_intra_pred_luma_8x8_mode_horz_d @@ -748,7 +742,7 @@ end_func_vert_r: @* @remarks @* None @* -@*******************************************************************************/ +@******************************************************************************* @void ih264_intra_pred_luma_8x8_mode_horz_d(UWORD8 *pu1_src, @ UWORD8 *pu1_dst, @ WORD32 src_strd, @@ -791,7 +785,7 @@ ih264_intra_pred_luma_8x8_mode_horz_d_a9q: vmov.8 q4, q2 vmov.8 q5, q3 sub r6, r3, #6 - vtrn.8 q4, q5 @ + vtrn.8 q4, q5 @ vmov.8 q6, q4 vmov.8 q7, q5 sub r5, r3, #4 @@ -835,7 +829,7 @@ end_func_horz_d: -@/** +@** @******************************************************************************* @* @*ih264_intra_pred_luma_8x8_mode_vert_l @@ -866,7 +860,7 @@ end_func_horz_d: @* @remarks @* None @* -@*******************************************************************************/ +@******************************************************************************* @void ih264_intra_pred_luma_8x8_mode_vert_l(UWORD8 *pu1_src, @ UWORD8 *pu1_dst, @ WORD32 src_strd, @@ -887,6 +881,7 @@ ih264_intra_pred_luma_8x8_mode_vert_l_a9q: stmfd sp!, {r4-r12, r14} @Restoring registers from stack vpush {d8-d15} + add r0, r0, #9 vld1.u8 {q0}, [r0] add r0, r0, #1 @@ -935,7 +930,7 @@ end_func_vert_l: -@/** +@** @******************************************************************************* @* @*ih264_intra_pred_luma_8x8_mode_horz_u @@ -966,7 +961,7 @@ end_func_vert_l: @* @remarks @* None @* -@*******************************************************************************/ +@******************************************************************************* @void ih264_intra_pred_luma_8x8_mode_horz_u(UWORD8 *pu1_src, @ UWORD8 *pu1_dst, @ WORD32 src_strd, diff --git a/common/arm/ih264_iquant_itrans_recon_a9.s b/common/arm/ih264_iquant_itrans_recon_a9.s index f71ca69..4e49f6a 100755..100644 --- a/common/arm/ih264_iquant_itrans_recon_a9.s +++ b/common/arm/ih264_iquant_itrans_recon_a9.s @@ -17,7 +17,7 @@ @ ***************************************************************************** @ * Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore @*/ -@/** +@** @ ******************************************************************************* @ * @file @ * ih264_iquant_itrans_recon_a9.s @@ -38,8 +38,8 @@ @ * None @ * @ ******************************************************************************* -@*/ -@/** +@* +@** @ ******************************************************************************* @ * @ * @brief @@ -82,7 +82,7 @@ @ * None @ * @ ******************************************************************************* -@ */ +@ * @void ih264_iquant_itrans_recon_4x4(WORD16 *pi2_src, @ UWORD8 *pu1_pred, @ UWORD8 *pu1_out, @@ -225,7 +225,7 @@ ih264_iquant_itrans_recon_4x4_a9: ldmfd sp!, {r4-r12, r15} @Reload the registers from SP - @/** +@** @ ******************************************************************************* @ * @ * @brief @@ -268,7 +268,7 @@ ih264_iquant_itrans_recon_4x4_a9: @ * None @ * @ ******************************************************************************* -@ */ +@ * @void ih264_iquant_itrans_recon_chroma_4x4(WORD16 *pi2_src, @ UWORD8 *pu1_pred, @ UWORD8 *pu1_out, @@ -416,7 +416,7 @@ ih264_iquant_itrans_recon_chroma_4x4_a9: ldmfd sp!, {r4-r12, r15} @Reload the registers from SP -@/* +@* @ ******************************************************************************* @ * @ * @brief @@ -459,7 +459,7 @@ ih264_iquant_itrans_recon_chroma_4x4_a9: @ * None @ * @ ******************************************************************************* -@ */ +@ * @void ih264_iquant_itrans_recon_8x8(WORD16 *pi2_src, @ UWORD8 *pu1_pred, @ UWORD8 *pu1_out, diff --git a/common/arm/ih264_iquant_itrans_recon_dc_a9.s b/common/arm/ih264_iquant_itrans_recon_dc_a9.s index 8d71bdb..97c4724 100755..100644 --- a/common/arm/ih264_iquant_itrans_recon_dc_a9.s +++ b/common/arm/ih264_iquant_itrans_recon_dc_a9.s @@ -17,7 +17,7 @@ @ ***************************************************************************** @ * Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore @*/ -@/** +@** @ ******************************************************************************* @ * @file @ * ih264_iquant_itrans_recon_dc_a9.s @@ -37,8 +37,8 @@ @ * None @ * @ ******************************************************************************* -@*/ -@/** +@* +@** @ ******************************************************************************* @ * @ * @brief @@ -83,7 +83,7 @@ @ * None @ * @ ******************************************************************************* -@ */ +@ * @void ih264_iquant_itrans_recon_4x4_dc(WORD16 *pi2_src, @ UWORD8 *pu1_pred, @ UWORD8 *pu1_out, @@ -167,7 +167,7 @@ ih264_iquant_itrans_recon_4x4_dc_a9: -@/* +@* @ ******************************************************************************* @ * @ * @brief @@ -212,7 +212,7 @@ ih264_iquant_itrans_recon_4x4_dc_a9: @ * None @ * @ ******************************************************************************* -@ */ +@ * @void ih264_iquant_itrans_recon_8x8_dc(WORD16 *pi2_src, @ UWORD8 *pu1_pred, @ UWORD8 *pu1_out, @@ -300,7 +300,7 @@ ih264_iquant_itrans_recon_8x8_dc_a9: ldmfd sp!, {r4-r8, r15} -@ /* +@ * @ ******************************************************************************** @ * @ * @brief This function reconstructs a 4x4 sub block from quantized resiude and @@ -328,7 +328,7 @@ ih264_iquant_itrans_recon_8x8_dc_a9: @ * @remarks none @ * @ ******************************************************************************* -@ */ +@ * @ void ih264_iquant_itrans_recon_chroma_4x4_dc(WORD16 *pi2_src, @ UWORD8 *pu1_pred, @ UWORD8 *pu1_out, @@ -368,6 +368,7 @@ ih264_iquant_itrans_recon_chroma_4x4_dc_a9: vmov.u16 q15, #0x00ff + vld1.u8 d18, [r2], r0 @load out [8 bit size) -8 coeffs vaddw.u8 q1, q0, d2 @Add pred vld1.u8 d19, [r2], r0 diff --git a/common/arm/ih264_itrans_recon_a9.s b/common/arm/ih264_itrans_recon_a9.s index 1d74da5..769d5d7 100755..100644 --- a/common/arm/ih264_itrans_recon_a9.s +++ b/common/arm/ih264_itrans_recon_a9.s @@ -17,7 +17,7 @@ @ ***************************************************************************** @ * Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore @*/ -@/** +@** @ ******************************************************************************* @ * @file @ * ih264_itrans_recon_neon_a9.s @@ -33,8 +33,8 @@ @ * None @ * @ ******************************************************************************* -@*/ -@/** +@* +@** @ ******************************************************************************* @ * @ * @brief @@ -72,7 +72,7 @@ @ * @ * @ ******************************************************************************* -@ */ +@ * @void ih264_itrans_recon_4x4( @ WORD16 *pi2_src, @ UWORD8 *pu1_pred, diff --git a/common/arm/ih264_mem_fns_neon.s b/common/arm/ih264_mem_fns_neon.s index 2808897..39ad9b3 100755..100644 --- a/common/arm/ih264_mem_fns_neon.s +++ b/common/arm/ih264_mem_fns_neon.s @@ -17,7 +17,7 @@ @ ***************************************************************************** @ * Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore @*/ -@/** +@** @ ******************************************************************************* @ * @file @ * ih264_mem_fns_neon.s @@ -40,9 +40,9 @@ @ * None @ * @ ******************************************************************************* -@*/ +@* -@/** +@** @******************************************************************************* @* @* @brief @@ -65,7 +65,7 @@ @* None @* @******************************************************************************* -@*/ +@* @void ih264_memcpy_mul_8(UWORD8 *pu1_dst, @ UWORD8 *pu1_src, @ UWORD8 num_bytes) @@ -94,7 +94,7 @@ loop_neon_memcpy_mul_8: @******************************************************************************* -@*/ +@* @void ih264_memcpy(UWORD8 *pu1_dst, @ UWORD8 *pu1_src, @ UWORD8 num_bytes) @@ -143,6 +143,8 @@ loop_memcpy: + + .global ih264_memset_mul_8_a9q ih264_memset_mul_8_a9q: @@ -208,6 +210,8 @@ loop_memset: + + .global ih264_memset_16bit_mul_8_a9q ih264_memset_16bit_mul_8_a9q: diff --git a/common/arm/ih264_padding_neon.s b/common/arm/ih264_padding_neon.s index 9bab268..e7a1f91 100755..100644 --- a/common/arm/ih264_padding_neon.s +++ b/common/arm/ih264_padding_neon.s @@ -17,7 +17,7 @@ @ ***************************************************************************** @ * Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore @*/ -@/** +@* @ ******************************************************************************* @ * @file @ * ih264_padding_neon.s @@ -39,10 +39,10 @@ @ * None @ * @ ******************************************************************************* -@*/ +@* -@/** +@** @******************************************************************************* @* @* @brief pad at the top of a 2d array @@ -67,7 +67,7 @@ @* @remarks none @* @******************************************************************************* -@*/ +@* @void ih264_pad_top(UWORD8 *pu1_src, @ WORD32 src_strd, @ WORD32 wd, @@ -110,7 +110,7 @@ loop_neon_pad_top: -@/** +@** @******************************************************************************* @* @* @brief @@ -147,7 +147,7 @@ loop_neon_pad_top: @* None @* @******************************************************************************* -@*/ +@* @#if PAD_LEFT_LUMA == C @void ih264_pad_left_luma(UWORD8 *pu1_src, @ WORD32 src_strd, @@ -160,6 +160,7 @@ loop_neon_pad_top: @ r3 => pad_size + .global ih264_pad_left_luma_a9q ih264_pad_left_luma_a9q: @@ -245,7 +246,7 @@ end_func: -@/** +@** @******************************************************************************* @* @* @brief @@ -282,7 +283,7 @@ end_func: @* None @* @******************************************************************************* -@*/ +@* @#if PAD_LEFT_CHROMA == C @void ih264_pad_left_chroma(UWORD8 *pu1_src, @ WORD32 src_strd, @@ -373,7 +374,7 @@ end_func_l_c: -@/** +@** @******************************************************************************* @* @* @brief @@ -410,7 +411,7 @@ end_func_l_c: @* None @* @******************************************************************************* -@*/ +@* @#if PAD_RIGHT_LUMA == C @void ih264_pad_right_luma(UWORD8 *pu1_src, @ WORD32 src_strd, @@ -519,7 +520,7 @@ end_func_r: -@/** +@** @******************************************************************************* @* @* @brief @@ -556,7 +557,7 @@ end_func_r: @* None @* @******************************************************************************* -@*/ +@* @#if PAD_RIGHT_CHROMA == C @void ih264_pad_right_chroma(UWORD8 *pu1_src, @ WORD32 src_strd, diff --git a/common/arm/ih264_platform_macros.h b/common/arm/ih264_platform_macros.h index 1f67403..1f67403 100755..100644 --- a/common/arm/ih264_platform_macros.h +++ b/common/arm/ih264_platform_macros.h diff --git a/common/arm/ih264_resi_trans_a9.s b/common/arm/ih264_resi_trans_a9.s deleted file mode 100755 index 08821f5..0000000 --- a/common/arm/ih264_resi_trans_a9.s +++ /dev/null @@ -1,604 +0,0 @@ -@/****************************************************************************** -@ * -@ * Copyright (C) 2015 The Android Open Source Project -@ * -@ * Licensed under the Apache License, Version 2.0 (the "License"); -@ * you may not use this file except in compliance with the License. -@ * You may obtain a copy of the License at: -@ * -@ * http://www.apache.org/licenses/LICENSE-2.0 -@ * -@ * Unless required by applicable law or agreed to in writing, software -@ * distributed under the License is distributed on an "AS IS" BASIS, -@ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -@ * See the License for the specific language governing permissions and -@ * limitations under the License. -@ * -@ ***************************************************************************** -@ * Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore -@*/ -@/** -@******************************************************************************* -@* @file -@* ih264_resi_trans_a9.s -@* -@* @brief -@* Contains function definitions for residual and forward trans -@* -@* @author -@* Ittiam -@* -@* @par List of Functions: -@* ih264_resi_trans_4x4_a9 -@* ih264_resi_trans_8x8_a9 -@* @remarks -@* None -@* -@******************************************************************************* - - -.text -.p2align 2 -@***************************************************************************** -@* -@* Function Name : ih264_resi_trans_4x4_a9 -@* Description : This function does cf4 of H264 followed by and approximate scaling -@* -@* Arguments : -@ R0 :pointer to src buffer -@ R1 :pointer to pred buffer -@ R2 :pointer to dst buffer -@ R3 :src_stride -@ STACk :pred_stride,dst_stride - -@* Values Returned : NONE -@* -@* Register Usage : -@* Stack Usage : -@* Cycles : Around -@* Interruptiaility : Interruptable -@* -@* Known Limitations -@* \Assumptions : -@* -@* Revision History : -@* DD MM YYYY Author(s) Changes -@* 30 12 2009 100633 First version -@* -@***************************************************************************** - - - .global ih264_resi_trans_4x4_a9 - .extern g_scal_coff_h264_4x4 -g_scal_coff_h264_4x4_addr: - .long g_scal_coff_h264_4x4 - 4x4lbl - 8 - -ih264_resi_trans_4x4_a9: - - @R0 :pointer to src buffer - @R1 :pointer to pred buffer - @R2 :pointer to dst buffer - @R3 :src_stride - @STACk :pred_stride,dst_stride - - push {r4-r12, lr} @push all the variables first - - mov r6, sp - add r6, r6, #40 @decrement stack pointer,to accomodate two variables - ldmfd r6, {r4-r5} @load the strides into registers - @R4 pred_stride - @R5 dst_stride - - - @we have to give the stride as post inrement in VLDR1 - @but since thr stride is from end of row 1 to start of row 2, - @we need to add the size of the curent row to strides ie we need to add 4 to it (4 bytes) - @ADD R3,#4 - @ADD R4,#4 - @ADD R5,#4 - @in case of dst the stride represnts 16 bit ie 2*8bits - @hence we need to add #4 to it and thenm multiply by 2 - @--------------------function loading done------------------------ - - @lets find residual - @data is like 1a -> d0[1:31] d0[32:64] - @ a b c d # # # # - vld1.u8 d30, [r0], r3 @load 4 pixels of row1 current buffer - vld1.u8 d31, [r1], r4 @load 4 pixels of row1 pred buffer - @ data is like 1a -> q4[1:63] q4[64:148] - @ d8[1:63] d9[1:63] - @ a b c d # # # # - - vld1.u8 d28, [r0], r3 @load row 2 of src to d28[0] - vld1.u8 d29, [r1], r4 @load row2 of pred to d29[0] - - vld1.u8 d26, [r0], r3 @load row 3 of src to d26[0] - vsubl.u8 q0, d30, d31 @curr - pred for row one - - vld1.u8 d27, [r1], r4 @load row 3of pred t0 d27[0] - vsubl.u8 q1, d28, d29 @find row 2 of src -pred to d0 - - vld1.u8 d24, [r0], r3 @load row 4 of src to d24[0] - - vld1.u8 d25, [r1], r4 @load row 4 of src tp d25[0] - vsubl.u8 q2, d26, d27 @load src-pred row 3 to d[2] - - lsl r5, r5, #2 @ multiply dst stride by since we are storing 32 bit values - ldr r6, g_scal_coff_h264_4x4_addr -4x4lbl: - add r6, r6, pc @ load the address of global array - - vsubl.u8 q3, d24, d25 @load row 4 of src - pred to q6 - - @after this - @D0 -> 1a - @D2 -> 2a - @D4 -> 3a - @D6 -> 4a - - @transpose the matrix so that we can do the horizontal transform first - @#1 #2 #3 #4 - @a b c d ---- D0 - @e f g h -----D2 - @i j k l -----D4 - @m n o p -----D6 - @transpose the inner 2x2 blocks - vtrn.16 d0, d2 - vld1.s16 {q10}, [r6]! @ load the scaling values 0-7; - vtrn.16 d4, d6 - @a e c g - @b f d h - @i m k o - @j n l p - vtrn.32 d0, d4 - vtrn.32 d2, d6 - @a e i m #1 -- D0 --- x4 - @b f j n #2 -- D2 --- x5 - @c g k o #3 -- D4 ----x6 - @d h l p #4 -- D6 ----x7 - - @we have loaded the residuals into the registers , now we need to add and subtract them - @let us do the horiz transform first - - vsub.s16 d5, d2, d4 @x2 = x5-x6 - vsub.s16 d7, d0, d6 @x3 = x4-x7; - - vadd.s16 d3, d2, d4 @x1 = x5+x6 - vadd.s16 d1, d0, d6 @x0 = x4+x7 - - - vshl.s16 d31, d7, #1 @ - vshl.s16 d30, d5, #1 @ - - vadd.s16 d0, d1, d3 @x0 + x1; - vsub.s16 d4, d1, d3 @x0 - x1; - - vadd.s16 d2, d31, d5 @U_SHIFT(x3,1,shft) + x2; - vsub.s16 d6, d7, d30 @x3 - U_SHIFT(x2,1,shft); - - @taking transform again so as to make do vert transform - vtrn.16 d0, d2 - vtrn.16 d4, d6 - - vtrn.32 d0, d4 - vtrn.32 d2, d6 - - @let us do vertical transform - @same code as horiz - - vadd.s16 d1, d0, d6 @x0 = x4+x7 - vadd.s16 d3, d2, d4 @x1 = x5+x6 - vsub.s16 d7, d0, d6 @x3 = x4-x7; - vsub.s16 d5, d2, d4 @x2 = x5-x6 - - -@Since we are going to do scal / quant or whatever, we are going to divide by -@a 32 bit number. So we have to expand the values - - @VADDL.S16 Q12,D1,D3;x0 + x1 - @VSUBL.S16 Q14,D1,D3;x0 - x1 - - @VSHL.S16 D8,D5,#1; - @VSHL.S16 D9,D7,#1; - - @VADDL.S16 Q13,D9,D5 ; + x2 - @VSUBL.S16 Q15,D7,D8 ;x3 - U_SHIFT(x2,1,shft) - -@scaling follows - -@now we need to do the scaling,so load the scaling matrix -@mutliplying by the scaling coeffient; store the results from q5-q8 ; - - vadd.s16 d24, d3, d1 @x4 = x0 + x1 - vsub.s16 d28, d1, d3 @x6 = x0 - x1 - - vshl.s16 d0, d7, #1 @ U_SHIFT(x3,1,shft) - vmull.s16 q4, d24, d20 @x4*s0 - - vshl.s16 d2, d5, #1 @ U_SHIFT(x2,1,shft) - - vadd.s16 d26, d0, d5 @x5 = U_SHIFT(x3,1,shft) + x2 - vmull.s16 q5, d26, d21 @x5*s1 - - vst1.s32 {q4}, [r2], r5 @save 4 pixels of row1 current buffer and increment pointer by stride - - vld1.s16 {q10}, [r6] @load 8-16 scaling coeffcients - - vsub.s16 d30, d7, d2 @x7 = x3 - U_SHIFT(x2,1,shft) - - vmull.s16 q6, d28, d20 @x6*s2 - vst1.s32 {q5}, [r2], r5 - - vmull.s16 q7, d30, d21 @x7*s3 - - - vst1.s32 {q6}, [r2], r5 - vst1.s32 {q7}, [r2] - - pop {r4-r12, pc} @pop back all variables - - - - -@***************************************************************************** -@* Function Name : ih264_resi_trans_8x8_a9 -@* Description : This function does cf8 followd by an approximate normalization of H264 -@* -@* Arguments : -@* R0 :pointer to src buffer -@ R1 :pointer to pred buffer -@ R2 :pointer to dst buffer -@ R3 :src_stride -@ STACk :pred_stride,dst_st -@* -@* -@* Values Returned : NONE -@* -@* Register Usage : -@* Stack Usage : -@* Cycles : Around -@* Interruptiaility : Interruptable -@* -@* Known Limitations -@* \Assumptions : -@* -@* Revision History : -@* DD MM YYYY Author(s) Changes -@* 30 12 2009 100633 First version -@* -@***************************************************************************** - - - .global ih264_resi_trans_8x8_a9 - .extern g_scal_coff_h264_8x8 -g_scal_coff_h264_8x8_addr: - .long g_scal_coff_h264_8x8 - 8x8lbl - 8 - - -ih264_resi_trans_8x8_a9: - - @R0 :pointer to src buffer - @R1 :pointer to pred buffer - @R2 :pointer to dst buffer - @R3 :src_stride - @STACk :pred_stride,dst_stride - - push {r4-r12, lr} @push all the variables first - - mov r6, sp - add r6, r6, #40 @decrement stack pointer,to accomodate two variables - ldmfd r6, {r4-r5} @load the strides into registers - @R4 pred_stride - @R5 dst_stride - - @we have to give the stride as post inrement in vst1 - @in case of dst the stride represnts 16 bit ie 2*8bits - @hence we need to add #4 to it and thenm multiply by 2 - @--------------------function loading done------------------------ - - @lets find residual - @data is like 1a -> d0[1:31] d0[32:64] - @ a b c d # # # # - vld1.u8 d30, [r0], r3 @load 4 pixels of row1 current buffer - vld1.u8 d31, [r1], r4 @load 4 pixels of row1 pred buffer - - vld1.u8 d28, [r0], r3 @src rw2 - vld1.u8 d29, [r1], r4 @pred rw2 - vsubl.u8 q0, d30, d31 @src-pred rw1 - - vld1.u8 d26, [r0], r3 - vld1.u8 d27, [r1], r4 - vsubl.u8 q1, d28, d29 - - vld1.u8 d24, [r0], r3 - vld1.u8 d25, [r1], r4 - vsubl.u8 q2, d26, d27 - - vld1.u8 d22, [r0], r3 - vld1.u8 d23, [r1], r4 - vsubl.u8 q3, d24, d25 - - vld1.u8 d20, [r0], r3 - vld1.u8 d21, [r1], r4 - vsubl.u8 q4, d22, d23 - - vld1.u8 d18, [r0], r3 - vld1.u8 d19, [r1], r4 - vsubl.u8 q5, d20, d21 - - vld1.u8 d16, [r0], r3 - vld1.u8 d17, [r1], r4 - vsubl.u8 q6, d18, d19 - - lsl r5, r5, #2 - - - vsubl.u8 q7, d16, d17 - - @after this - @Q0 -> 1a - @Q1 -> 2a - @Q2 -> 3a - @Q3 -> 4a - @Q4 -> 5a - @Q5 -> 6a - @Q6 -> 7a - @Q7 -> 8a - - @transpose the matrix so that we can do the horizontal transform first - - @transpose the inner 2x2 blocks - vtrn.16 q0, q1 - vtrn.16 q2, q3 - vtrn.16 q4, q5 - vtrn.16 q6, q7 - - @transpose the inner 4x4 blocks - vtrn.32 q0, q2 - vtrn.32 q1, q3 - - vtrn.32 q4, q6 - vtrn.32 q5, q7 - - @transpose the outer 8x8 blocks - vswp d1, d8 - vswp d7, d14 - vswp d3, d10 - vswp d5, d12 - @transpose done - -@@this point we will have data in Q0-Q7 -@Q7 will be populated within 2 clock cycle -@all others are availabe @ this clock cycle - - @we have loaded the residuals into the registers , now we need to add and subtract them - @let us do the horiz transform first - - vadd.s16 q8, q0, q7 @ a0 = r0 + r7; - vadd.s16 q9, q1, q6 @ a1 = r1 + r6; - vadd.s16 q10, q2, q5 @ a2 = r2 + r5; - vadd.s16 q11, q3, q4 @ a3 = r3 + r4; - - vsub.s16 q12, q0, q7 @ b0 = r0 - r7; - vsub.s16 q13, q1, q6 @ b1 = r1 - r6; - vsub.s16 q15, q3, q4 @ b3 = r3 - r4; - vsub.s16 q14, q2, q5 @ b2 = r2 - r5; - - vadd.s16 q1, q8, q11 @ a4 = a0 + a3; - vadd.s16 q3, q9, q10 @ a5 = a1 + a2; - vsub.s16 q7, q9, q10 @ a7 = a1 - a2; - vsub.s16 q5, q8, q11 @ a6 = a0 - a3; - - ldr r6, g_scal_coff_h264_8x8_addr -8x8lbl: - add r6, r6, pc @ load the address of global array - - vadd.s16 q0, q1, q3 @ pi2_res[0] = a4 + a5; - vshr.s16 q8, q7, #1 @ pi2_res[2] = a6 + D_SHIFT(a7,1,shft); - - vsub.s16 q4, q1, q3 @ pi2_res[4] = a4 - a5; - - vadd.s16 q2, q5, q8 @ - - - vshr.s16 q9, q5, #1 @ pi2_res[6] = D_SHIFT(a6,1,shft) - a7; - vsub.s16 q6, q9, q7 @ - -@do not change Q0,Q2.Q4,Q6 they contain results -@Q1,Q3,Q5,Q7 TO STORE RESULTS -@Q8 Q9 Q10 Q11 USE @WILL - - vshr.s16 q1, q12, #1 @ D_SHIFT(b0,1,shft) - vshr.s16 q3, q13, #1 @ D_SHIFT(b1,1,shft) - vshr.s16 q5, q14, #1 @ D_SHIFT(b2,1,shft) - vshr.s16 q7, q15, #1 @ D_SHIFT(b3,1,shft) - - vadd.s16 q8, q1, q12 @ (D_SHIFT(b0,1,shft) + b0); - vadd.s16 q9, q3, q13 @ (D_SHIFT(b1,1,shft) + b1); - vadd.s16 q10, q5, q14 @ (D_SHIFT(b2,1,shft) + b2); - vadd.s16 q11, q7, q15 @ (D_SHIFT(b3,1,shft) + b3); - - vadd.s16 q1, q14, q8 @ b2 + (D_SHIFT(b0,1,shft) + b0); - vsub.s16 q5, q15, q9 @ b3 - (D_SHIFT(b1,1,shft) + b1); - vadd.s16 q3, q15, q10 @ b3 + (D_SHIFT(b2,1,shft) + b2); - vsub.s16 q7, q11, q14 @ -b2 + (D_SHIFT(b3,1,shft) + b3); - - vadd.s16 q8, q13, q1 @ b4 = b1 + b2 + (D_SHIFT(b0,1,shft) + b0); - vsub.s16 q9, q12, q3 @ b5 = b0 - b3 - (D_SHIFT(b2,1,shft) + b2); - vadd.s16 q10, q12, q5 @ b6 = b0 + b3 - (D_SHIFT(b1,1,shft) + b1); - vadd.s16 q11, q13, q7 @ b7 = b1 - b2 + (D_SHIFT(b3,1,shft) + b3); - - vshr.s16 q15, q8, #2 @ D_SHIFT(b4,2,shft) - vshr.s16 q14, q9, #2 @ D_SHIFT(b5,2,shft); - vshr.s16 q13, q10, #2 @ D_SHIFT(b6,2,shft); - vshr.s16 q12, q11, #2 @ D_SHIFT(b7,2,shft); - - - vadd.s16 q3, q9, q13 @ pi2_res[3] = b5 + D_SHIFT(b6,2,shft); - vsub.s16 q5, q10, q14 @ pi2_res[5] = b6 - D_SHIFT(b5,2,shft); - vadd.s16 q1, q8, q12 @ pi2_res[1] = b4 + D_SHIFT(b7,2,shft); - vsub.s16 q7, q15, q11 @ pi2_res[7] = D_SHIFT(b4,2,shft) - b7; - - @------------horiz transform done------------------------- - @results are in Q0-Q7 - @all other neon registes can be used at will - -@doing vertical transform -@code exact copy of horiz transform above - - @transpose the inner 2x2 blocks - vtrn.16 q0, q1 - vtrn.16 q2, q3 - vtrn.16 q4, q5 - vtrn.16 q6, q7 - - @transpose the inner 4x4 blocks - vtrn.32 q0, q2 - vtrn.32 q1, q3 - - vtrn.32 q4, q6 - vtrn.32 q5, q7 - - @transpose the outer 8x8 blocks - vswp d1, d8 - vswp d3, d10 - vswp d5, d12 - vswp d7, d14 - - @transpose done - - vadd.s16 q8, q0, q7 @ a0 = r0 + r7; - vadd.s16 q9, q1, q6 @ a1 = r1 + r6; - vadd.s16 q10, q2, q5 @ a2 = r2 + r5; - vadd.s16 q11, q3, q4 @ a3 = r3 + r4; - - vsub.s16 q12, q0, q7 @ b0 = r0 - r7; - vsub.s16 q13, q1, q6 @ b1 = r1 - r6; - vsub.s16 q14, q2, q5 @ b2 = r2 - r5; - vsub.s16 q15, q3, q4 @ b3 = r3 - r4; - - vadd.s16 q1, q8, q11 @ a4 = a0 + a3; - vadd.s16 q3, q9, q10 @ a5 = a1 + a2; - vsub.s16 q5, q8, q11 @ a6 = a0 - a3; - vsub.s16 q7, q9, q10 @ a7 = a1 - a2; - - - vadd.s16 q0, q1, q3 @ pi2_res[0] = a4 + a5; - - vshr.s16 q8, q7, #1 @ pi2_res[2] = a6 + D_SHIFT(a7,1,shft); - @DSHIFT_TO_0 Q8,Q7,#1,#0 - vadd.s16 q2, q5, q8 @ - - vsub.s16 q4, q1, q3 @ pi2_res[4] = a4 - a5; - - vshr.s16 q9, q5, #1 @ pi2_res[6] = D_SHIFT(a6,1,shft) - a7; - vsub.s16 q6, q9, q7 @ - -@do not change Q0,Q2.Q4,Q6 they contain results -@Q1,Q3,Q5,Q7 TO STORE RESULTS -@Q8 Q9 Q10 Q11 USE @WILL - - vshr.s16 q1, q12, #1 @ D_SHIFT(b0,1,shft) - vshr.s16 q3, q13, #1 @ D_SHIFT(b1,1,shft) - vshr.s16 q5, q14, #1 @ D_SHIFT(b2,1,shft) - vshr.s16 q7, q15, #1 @ D_SHIFT(b3,1,shft) - - - vadd.s16 q8, q1, q12 @ (D_SHIFT(b0,1,shft) + b0); - vadd.s16 q9, q3, q13 @ (D_SHIFT(b1,1,shft) + b1); - vadd.s16 q10, q5, q14 @ (D_SHIFT(b2,1,shft) + b2); - vadd.s16 q11, q7, q15 @ (D_SHIFT(b3,1,shft) + b3); - - vadd.s16 q1, q14, q8 @ b2 + (D_SHIFT(b0,1,shft) + b0); - vadd.s16 q3, q15, q10 @ b3 + (D_SHIFT(b2,1,shft) + b2); - vsub.s16 q5, q15, q9 @ b3 - (D_SHIFT(b1,1,shft) + b1); - vsub.s16 q7, q11, q14 @ -b2 + (D_SHIFT(b3,1,shft) + b3); - - vadd.s16 q8, q13, q1 @ b4 = b1 + b2 + (D_SHIFT(b0,1,shft) + b0); - vsub.s16 q9, q12, q3 @ b5 = b0 - b3 - (D_SHIFT(b2,1,shft) + b2); - vadd.s16 q10, q12, q5 @ b6 = b0 + b3 - (D_SHIFT(b1,1,shft) + b1); - vadd.s16 q11, q13, q7 @ b7 = b1 - b2 + (D_SHIFT(b3,1,shft) + b3); - - vshr.s16 q15, q8, #2 @ D_SHIFT(b4,2,shft) - vshr.s16 q14, q9, #2 @ D_SHIFT(b5,2,shft); - vshr.s16 q13, q10, #2 @ D_SHIFT(b6,2,shft); - vshr.s16 q12, q11, #2 @ D_SHIFT(b7,2,shft); - - -@since we are going to scal by small values, we need not expand the guys to 32 bit bit values - vsub.s16 q5, q10, q14 @ pi2_res[5] = b6 - D_SHIFT(b5,2,shft); - vsub.s16 q7, q15, q11 @ pi2_res[7] = D_SHIFT(b4,2,shft) - b7; - vadd.s16 q3, q9, q13 @ pi2_res[3] = b5 + D_SHIFT(b6,2,shft); - vadd.s16 q1, q8, q12 @ pi2_res[1] = b4 + D_SHIFT(b7,2,shft); - - @------------vert transform done------------------------- - @results are in Q0-Q7 - @all other neon registes can be used at will - - @scaling - @since the 8x8 scaling matrix repeats in 1x4,1x4 block , - @we need only load 4 values for each row and in total 4 rows - vld1.s16 {q14-q15}, [r6] @ - - @since we need to get a 32 bit o/p for two 16 bit multiplications - @we need a VMULL instruction -@-----------------------------first and second row - - vmull.s16 q8, d0, d28 @scale the first row first 4 elem - vmull.s16 q9, d28, d1 @scale the second row last 4 elemts - - vmull.s16 q10, d2, d29 @ scale second row first 4 elem - vmull.s16 q11, d29, d3 @scale the second row last 4 elem - vmull.s16 q12, d4, d30 @scale third row first 4 elem - - vst1.s32 {q8, q9}, [r2], r5 @ write the first row complete - - vmull.s16 q13, d30, d5 @scale the third row last 4 elem - vmull.s16 q8, d6, d31 @scale the fourth row first 4 elem - - - vst1.s32 {q10, q11}, [r2], r5 @store the second row complete - -@------------------------------- 3rd and 4th row - - vmull.s16 q9, d31, d7 @scale the fourth row second column - - vst1.s32 {q12, q13}, [r2], r5 @store the third row complete - - vmull.s16 q10, d8, d28 @scale the 5th row fisrst 4 elms - vmull.s16 q11, d28, d9 @scale the 5th row second 4 elems - - vmull.s16 q12, d10, d29 @scale the 6th row first4 elements - - - vst1.s32 {q8, q9}, [r2], r5 @store fifth row - -@--------------------------------5th and 6th row - - vmull.s16 q13, d29, d11 @scale 6th row sendond 4 elems - - vmull.s16 q8, d12, d30 @scale 7th rw first 4 elms - - vst1.s32 {q10, q11}, [r2], r5 @store 6th row second 4 elements - - vmull.s16 q9, d30, d13 @scale 7th rw second 4 elms - vmull.s16 q10, d14, d31 @scale 8th rw forst 4 elms - - - vst1.s32 {q12, q13}, [r2], r5 @store 6th row - -@----------------------------------7th and 8th row - vmull.s16 q11, d31, d15 @scale 8th row second 4 elms - - vst1.s32 {q8, q9}, [r2], r5 @store 7th row - vst1.s32 {q10, q11}, [r2], r5 @store 8th row - -@----------------------------------done writing - - pop {r4-r12, pc} @pop back all variables - - - - - - diff --git a/common/arm/ih264_resi_trans_quant_a9.s b/common/arm/ih264_resi_trans_quant_a9.s index caf362e..bb836bd 100755..100644 --- a/common/arm/ih264_resi_trans_quant_a9.s +++ b/common/arm/ih264_resi_trans_quant_a9.s @@ -17,7 +17,7 @@ @ ***************************************************************************** @ * Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore @*/ -@/** +@** @******************************************************************************* @* @file @* ih264_resi_trans_quant_a9.s diff --git a/common/arm/ih264_weighted_bi_pred_a9q.s b/common/arm/ih264_weighted_bi_pred_a9q.s index ccae779..33859e6 100755..100644 --- a/common/arm/ih264_weighted_bi_pred_a9q.s +++ b/common/arm/ih264_weighted_bi_pred_a9q.s @@ -17,7 +17,7 @@ @ ***************************************************************************** @ * Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore @*/ -@/** +@** @****************************************************************************** @* @file @* ih264_weighted_bi_pred_a9q.s @@ -37,7 +37,7 @@ @* None @* @******************************************************************************* -@*/ +@* @******************************************************************************* @* @function @* ih264_weighted_bi_pred_luma_a9q() @@ -96,7 +96,7 @@ @* (ht,wd) can be (4,4), (4,8), (8,4), (8,8), (8,16), (16,8) or (16,16). @* @******************************************************************************* -@*/ +@* @void ih264_weighted_bi_pred_luma_a9q(UWORD8 *pu1_src1, @ UWORD8 *pu1_src2, @ UWORD8 *pu1_dst, @@ -411,7 +411,7 @@ end_loops: @* (ht,wd) can be (2,2), (2,4), (4,2), (4,4), (4,8), (8,4) or (8,8). @* @******************************************************************************* -@*/ +@* @void ih264_weighted_bi_pred_chroma_a9q(UWORD8 *pu1_src1, @ UWORD8 *pu1_src2, @ UWORD8 *pu1_dst, diff --git a/common/arm/ih264_weighted_pred_a9q.s b/common/arm/ih264_weighted_pred_a9q.s index 1ce94d0..81d26d4 100755..100644 --- a/common/arm/ih264_weighted_pred_a9q.s +++ b/common/arm/ih264_weighted_pred_a9q.s @@ -17,7 +17,7 @@ @ ***************************************************************************** @ * Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore @*/ -@/** +@** @****************************************************************************** @* @file @* ih264_weighted_pred_a9q.s @@ -37,7 +37,7 @@ @* None @* @******************************************************************************* -@*/ +@* @******************************************************************************* @* @function @* ih264_weighted_pred_luma_a9q() @@ -84,7 +84,7 @@ @* (ht,wd) can be (4,4), (4,8), (8,4), (8,8), (8,16), (16,8) or (16,16). @* @******************************************************************************* -@*/ +@* @void ih264_weighted_pred_luma_a9q(UWORD8 *pu1_src, @ UWORD8 *pu1_dst, @ WORD32 src_strd, @@ -314,7 +314,7 @@ end_loops: @* (ht,wd) can be (2,2), (2,4), (4,2), (4,4), (4,8), (8,4) or (8,8). @* @******************************************************************************* -@*/ +@* @void ih264_weighted_pred_chroma_a9q(UWORD8 *pu1_src, @ UWORD8 *pu1_dst, @ WORD32 src_strd, |