From: John Koleszar Date: Fri, 1 Mar 2013 00:25:38 +0000 (-0800) Subject: Merge master branch into experimental X-Git-Tag: v1.3.0~1151^2~92^2~2 X-Git-Url: http://review.tizen.org/git/?a=commitdiff_plain;h=69c67c9531b08c1b15985b351677a162739af43e;p=platform%2Fupstream%2Flibvpx.git Merge master branch into experimental Picks up some build system changes, compiler warning fixes, etc. Change-Id: I2712f99e653502818a101a72696ad54018152d4e --- 69c67c9531b08c1b15985b351677a162739af43e diff --cc build/make/configure.sh index e558ff2,1ac3035..1a78f27 --- a/build/make/configure.sh +++ b/build/make/configure.sh @@@ -1050,13 -1064,9 +1059,15 @@@ EO ;; esac + soft_enable runtime_cpu_detect + soft_enable mmx + soft_enable sse + soft_enable sse2 + soft_enable sse3 + soft_enable ssse3 + # We can't use 'check_cflags' until the compiler is configured and CC is + # populated. - if enabled gcc && ! disabled sse4_1 && ! check_cflags -msse4.1; then + if enabled gcc && ! disabled sse4_1 && ! check_cflags -msse4; then RTCD_OPTIONS="${RTCD_OPTIONS}--disable-sse4_1 " else soft_enable sse4_1 diff --cc vp8/decoder/onyxd_if.c index 24fc878,019c255..2db3096 --- a/vp8/decoder/onyxd_if.c +++ b/vp8/decoder/onyxd_if.c @@@ -367,6 -300,25 +300,25 @@@ int check_fragments_for_errors(VP8D_COM return 0; } + return 1; + } - ++ + int vp8dx_receive_compressed_data(VP8D_COMP *pbi, size_t size, + const uint8_t *source, + int64_t time_stamp) + { + #if HAVE_NEON + int64_t dx_store_reg[8]; + #endif + VP8_COMMON *cm = &pbi->common; + int retcode = -1; + + pbi->common.error.error_code = VPX_CODEC_OK; + + retcode = check_fragments_for_errors(pbi); + if(retcode <= 0) + return retcode; + #if HAVE_NEON #if CONFIG_RUNTIME_CPU_DETECT if (cm->cpu_caps & HAS_NEON) diff --cc vp9/common/x86/vp9_subpixel_8t_ssse3.asm index fa24f4c,b644da6..32f00e2 --- a/vp9/common/x86/vp9_subpixel_8t_ssse3.asm +++ b/vp9/common/x86/vp9_subpixel_8t_ssse3.asm @@@ -380,316 -424,21 +380,316 @@@ sym(vp9_filter_block1d8_v8_ssse3) %define k6k7 [rsp + 16*3] %define krd [rsp + 16*4] - mov rdx, arg(5) ;filter ptr - mov rsi, arg(0) ;src_ptr - mov rdi, arg(2) ;output_ptr - mov rcx, 0x0400040 - - movdqa xmm4, [rdx] ;load filters - movq xmm5, rcx - packsswb xmm4, xmm4 - pshuflw xmm0, xmm4, 0b ;k0_k1 - pshuflw xmm1, xmm4, 01010101b ;k2_k3 - pshuflw xmm2, xmm4, 10101010b ;k4_k5 - pshuflw xmm3, xmm4, 11111111b ;k6_k7 + VERTx8 0 - punpcklqdq xmm0, xmm0 - punpcklqdq xmm1, xmm1 + add rsp, 16*5 + pop rsp + pop rbx + ; begin epilog + pop rdi + pop rsi + RESTORE_XMM + UNSHADOW_ARGS + pop rbp + ret + +;void vp9_filter_block1d16_v8_ssse3 +;( +; unsigned char *src_ptr, +; unsigned int src_pitch, +; unsigned char *output_ptr, +; unsigned int out_pitch, +; unsigned int output_height, +; short *filter +;) +global sym(vp9_filter_block1d16_v8_ssse3) PRIVATE +sym(vp9_filter_block1d16_v8_ssse3): + push rbp + mov rbp, rsp + SHADOW_ARGS_TO_STACK 6 + SAVE_XMM 7 + push rsi + push rdi + push rbx + ; end prolog + + ALIGN_STACK 16, rax + sub rsp, 16*5 + %define k0k1 [rsp + 16*0] + %define k2k3 [rsp + 16*1] + %define k4k5 [rsp + 16*2] + %define k6k7 [rsp + 16*3] + %define krd [rsp + 16*4] + + VERTx16 0 + + add rsp, 16*5 + pop rsp + pop rbx + ; begin epilog + pop rdi + pop rsi + RESTORE_XMM + UNSHADOW_ARGS + pop rbp + ret + +;~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + + +global sym(vp9_filter_block1d4_v8_avg_ssse3) PRIVATE +sym(vp9_filter_block1d4_v8_avg_ssse3): + push rbp + mov rbp, rsp + SHADOW_ARGS_TO_STACK 6 + SAVE_XMM 7 + push rsi + push rdi + push rbx + ; end prolog + + ALIGN_STACK 16, rax + sub rsp, 16*5 + %define k0k1 [rsp + 16*0] + %define k2k3 [rsp + 16*1] + %define k4k5 [rsp + 16*2] + %define k6k7 [rsp + 16*3] + %define krd [rsp + 16*4] + + VERTx4 1 + + add rsp, 16*5 + pop rsp + pop rbx + ; begin epilog + pop rdi + pop rsi + RESTORE_XMM + UNSHADOW_ARGS + pop rbp + ret + +global sym(vp9_filter_block1d8_v8_avg_ssse3) PRIVATE +sym(vp9_filter_block1d8_v8_avg_ssse3): + push rbp + mov rbp, rsp + SHADOW_ARGS_TO_STACK 6 + SAVE_XMM 7 + push rsi + push rdi + push rbx + ; end prolog + + ALIGN_STACK 16, rax + sub rsp, 16*5 + %define k0k1 [rsp + 16*0] + %define k2k3 [rsp + 16*1] + %define k4k5 [rsp + 16*2] + %define k6k7 [rsp + 16*3] + %define krd [rsp + 16*4] + + VERTx8 1 + + add rsp, 16*5 + pop rsp + pop rbx + ; begin epilog + pop rdi + pop rsi + RESTORE_XMM + UNSHADOW_ARGS + pop rbp + ret + +global sym(vp9_filter_block1d16_v8_avg_ssse3) PRIVATE +sym(vp9_filter_block1d16_v8_avg_ssse3): + push rbp + mov rbp, rsp + SHADOW_ARGS_TO_STACK 6 + SAVE_XMM 7 + push rsi + push rdi + push rbx + ; end prolog + + ALIGN_STACK 16, rax + sub rsp, 16*5 + %define k0k1 [rsp + 16*0] + %define k2k3 [rsp + 16*1] + %define k4k5 [rsp + 16*2] + %define k6k7 [rsp + 16*3] + %define krd [rsp + 16*4] + + VERTx16 1 + + add rsp, 16*5 + pop rsp + pop rbx + ; begin epilog + pop rdi + pop rsi + RESTORE_XMM + UNSHADOW_ARGS + pop rbp + ret + +;~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +%macro HORIZx4 1 + mov rdx, arg(5) ;filter ptr + mov rsi, arg(0) ;src_ptr + mov rdi, arg(2) ;output_ptr + mov rcx, 0x0400040 + + movdqa xmm4, [rdx] ;load filters - movd xmm5, rcx ++ movq xmm5, rcx + packsswb xmm4, xmm4 + pshuflw xmm0, xmm4, 0b ;k0_k1 + pshuflw xmm1, xmm4, 01010101b ;k2_k3 + pshuflw xmm2, xmm4, 10101010b ;k4_k5 + pshuflw xmm3, xmm4, 11111111b ;k6_k7 + + punpcklqdq xmm0, xmm0 + punpcklqdq xmm1, xmm1 + punpcklqdq xmm2, xmm2 + punpcklqdq xmm3, xmm3 + + movdqa k0k1, xmm0 + movdqa k2k3, xmm1 + pshufd xmm5, xmm5, 0 + movdqa k4k5, xmm2 + movdqa k6k7, xmm3 + movdqa krd, xmm5 + + movsxd rax, dword ptr arg(1) ;src_pixels_per_line + movsxd rdx, dword ptr arg(3) ;output_pitch + movsxd rcx, dword ptr arg(4) ;output_height + +.loop: + movq xmm0, [rsi - 3] ; -3 -2 -1 0 1 2 3 4 + + movq xmm3, [rsi + 5] ; 5 6 7 8 9 10 11 12 + punpcklqdq xmm0, xmm3 + + movdqa xmm1, xmm0 + pshufb xmm0, [GLOBAL(shuf_t0t1)] + pmaddubsw xmm0, k0k1 + + movdqa xmm2, xmm1 + pshufb xmm1, [GLOBAL(shuf_t2t3)] + pmaddubsw xmm1, k2k3 + + movdqa xmm4, xmm2 + pshufb xmm2, [GLOBAL(shuf_t4t5)] + pmaddubsw xmm2, k4k5 + + pshufb xmm4, [GLOBAL(shuf_t6t7)] + pmaddubsw xmm4, k6k7 + + paddsw xmm0, xmm1 + paddsw xmm0, xmm4 + paddsw xmm0, xmm2 + paddsw xmm0, krd + psraw xmm0, 7 + packuswb xmm0, xmm0 +%if %1 + movd xmm1, [rdi] + pavgb xmm0, xmm1 +%endif + lea rsi, [rsi + rax] + movd [rdi], xmm0 + + lea rdi, [rdi + rdx] + dec rcx + jnz .loop +%endm + +%macro HORIZx8 1 + mov rdx, arg(5) ;filter ptr + mov rsi, arg(0) ;src_ptr + mov rdi, arg(2) ;output_ptr + mov rcx, 0x0400040 + + movdqa xmm4, [rdx] ;load filters + movd xmm5, rcx + packsswb xmm4, xmm4 + pshuflw xmm0, xmm4, 0b ;k0_k1 + pshuflw xmm1, xmm4, 01010101b ;k2_k3 + pshuflw xmm2, xmm4, 10101010b ;k4_k5 + pshuflw xmm3, xmm4, 11111111b ;k6_k7 + + punpcklqdq xmm0, xmm0 + punpcklqdq xmm1, xmm1 + punpcklqdq xmm2, xmm2 + punpcklqdq xmm3, xmm3 + + movdqa k0k1, xmm0 + movdqa k2k3, xmm1 + pshufd xmm5, xmm5, 0 + movdqa k4k5, xmm2 + movdqa k6k7, xmm3 + movdqa krd, xmm5 + + movsxd rax, dword ptr arg(1) ;src_pixels_per_line + movsxd rdx, dword ptr arg(3) ;output_pitch + movsxd rcx, dword ptr arg(4) ;output_height + +.loop: + movq xmm0, [rsi - 3] ; -3 -2 -1 0 1 2 3 4 + + movq xmm3, [rsi + 5] ; 5 6 7 8 9 10 11 12 + punpcklqdq xmm0, xmm3 + + movdqa xmm1, xmm0 + pshufb xmm0, [GLOBAL(shuf_t0t1)] + pmaddubsw xmm0, k0k1 + + movdqa xmm2, xmm1 + pshufb xmm1, [GLOBAL(shuf_t2t3)] + pmaddubsw xmm1, k2k3 + + movdqa xmm4, xmm2 + pshufb xmm2, [GLOBAL(shuf_t4t5)] + pmaddubsw xmm2, k4k5 + + pshufb xmm4, [GLOBAL(shuf_t6t7)] + pmaddubsw xmm4, k6k7 + + paddsw xmm0, xmm1 + paddsw xmm0, xmm4 + paddsw xmm0, xmm2 + paddsw xmm0, krd + psraw xmm0, 7 + packuswb xmm0, xmm0 +%if %1 + movq xmm1, [rdi] + pavgb xmm0, xmm1 +%endif + + lea rsi, [rsi + rax] + movq [rdi], xmm0 + + lea rdi, [rdi + rdx] + dec rcx + jnz .loop +%endm + +%macro HORIZx16 1 + mov rdx, arg(5) ;filter ptr + mov rsi, arg(0) ;src_ptr + mov rdi, arg(2) ;output_ptr + mov rcx, 0x0400040 + + movdqa xmm4, [rdx] ;load filters - movd xmm5, rcx ++ movq xmm5, rcx + packsswb xmm4, xmm4 + pshuflw xmm0, xmm4, 0b ;k0_k1 + pshuflw xmm1, xmm4, 01010101b ;k2_k3 + pshuflw xmm2, xmm4, 10101010b ;k4_k5 + pshuflw xmm3, xmm4, 11111111b ;k6_k7 + + punpcklqdq xmm0, xmm0 + punpcklqdq xmm1, xmm1 punpcklqdq xmm2, xmm2 punpcklqdq xmm3, xmm3 diff --cc vp9/decoder/vp9_decodframe.c index d4212a5,c3b9637..86806d2 --- a/vp9/decoder/vp9_decodframe.c +++ b/vp9/decoder/vp9_decodframe.c @@@ -565,9 -684,9 +565,9 @@@ static void decode_4x4_sb(VP9D_COMP *pb }; static void decode_superblock64(VP9D_COMP *pbi, MACROBLOCKD *xd, - int mb_row, unsigned int mb_col, + int mb_row, int mb_col, BOOL_DECODER* const bc) { - int i, n, eobtotal; + int n, eobtotal; TX_SIZE tx_size = xd->mode_info_context->mbmi.txfm_size; VP9_COMMON *const pc = &pbi->common; MODE_INFO *orig_mi = xd->mode_info_context; @@@ -688,9 -810,9 +688,9 @@@ } static void decode_superblock32(VP9D_COMP *pbi, MACROBLOCKD *xd, - int mb_row, unsigned int mb_col, + int mb_row, int mb_col, BOOL_DECODER* const bc) { - int i, n, eobtotal; + int n, eobtotal; TX_SIZE tx_size = xd->mode_info_context->mbmi.txfm_size; VP9_COMMON *const pc = &pbi->common; MODE_INFO *orig_mi = xd->mode_info_context;