Merge "Merge branch 'master' into experimental" into experimental

12 files changed, 17 insertions, 98 deletions

diff --git a/vp9/common/x86/vp9_idct_x86.c b/vp9/common/x86/vp9_idct_intrin_sse2.c
index 811ed9899..dd7e68aa3 100644
--- a/vp9/common/x86/vp9_idct_x86.c
+++ b/vp9/common/x86/vp9_idct_intrin_sse2.c
@@ -15,7 +15,6 @@
 #include "vp9/common/vp9_common.h"
 #include "vp9/common/vp9_idct.h"
 
-#if HAVE_SSE2
 // In order to improve performance, clip absolute diff values to [0, 255],
 // which allows to keep the additions/subtractions in 8 bits.
 void vp9_dc_only_idct_add_sse2(int input_dc, uint8_t *pred_ptr,
@@ -1972,4 +1971,3 @@ void vp9_short_idct32x32_sse2(int16_t *input, int16_t *output, int pitch) {
     }
   }
 }
-#endif
diff --git a/vp9/common/x86/vp9_idct_x86.h b/vp9/common/x86/vp9_idct_x86.h
deleted file mode 100644
index bd66d8c72..000000000
--- a/vp9/common/x86/vp9_idct_x86.h
+++ /dev/null
@@ -1,51 +0,0 @@
-/*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- *  Use of this source code is governed by a BSD-style license
- *  that can be found in the LICENSE file in the root of the source
- *  tree. An additional intellectual property rights grant can be found
- *  in the file PATENTS.  All contributing project authors may
- *  be found in the AUTHORS file in the root of the source tree.
- */
-
-
-#ifndef VP9_COMMON_X86_VP9_IDCT_X86_H_
-#define VP9_COMMON_X86_VP9_IDCT_X86_H_
-
-/* Note:
- *
- * This platform is commonly built for runtime CPU detection. If you modify
- * any of the function mappings present in this file, be sure to also update
- * them in the function pointer initialization code
- */
-
-#if HAVE_MMX
-extern prototype_second_order(vp9_short_inv_walsh4x4_mmx);
-extern prototype_second_order(vp9_short_inv_walsh4x4_1_mmx);
-
-#if !CONFIG_RUNTIME_CPU_DETECT
-#undef vp9_idct_iwalsh16
-#define vp9_idct_iwalsh16 vp9_short_inv_walsh4x4_mmx
-
-#undef vp9_idct_iwalsh1
-#define vp9_idct_iwalsh1 vp9_short_inv_walsh4x4_1_mmx
-
-#endif
-#endif
-
-#if HAVE_SSE2
-
-extern prototype_second_order(vp9_short_inv_walsh4x4_sse2);
-
-#if !CONFIG_RUNTIME_CPU_DETECT
-
-#undef vp9_idct_iwalsh16
-#define vp9_idct_iwalsh16 vp9_short_inv_walsh4x4_sse2
-
-#endif
-
-#endif
-
-
-
-#endif
diff --git a/vp9/decoder/vp9_decodframe.c b/vp9/decoder/vp9_decodframe.c
index ad3a48927..0b3d145fc 100644
--- a/vp9/decoder/vp9_decodframe.c
+++ b/vp9/decoder/vp9_decodframe.c
@@ -1331,7 +1331,7 @@ static void decode_tiles(VP9D_COMP *pbi,
   if (pbi->oxcf.inv_tile_order) {
     const int n_cols = pc->tile_columns;
     const uint8_t *data_ptr2[4][1 << 6];
-    vp9_reader UNINITIALIZED_IS_SAFE(bc_bak);
+    vp9_reader bc_bak = {0};
 
     // pre-initialize the offsets, we're going to read in inverse order
     data_ptr2[0][0] = data_ptr;
diff --git a/vp9/decoder/x86/vp9_dequantize_x86.c b/vp9/decoder/x86/vp9_dequantize_sse2.c
index cbe818143..1296b704a 100644
--- a/vp9/decoder/x86/vp9_dequantize_x86.c
+++ b/vp9/decoder/x86/vp9_dequantize_sse2.c
@@ -15,8 +15,6 @@
 #include "vp9/common/vp9_common.h"
 #include "vp9/common/vp9_idct.h"
 
-#if HAVE_SSE2
-
 void vp9_add_residual_4x4_sse2(const int16_t *diff, uint8_t *dest, int stride) {
   const int width = 4;
   const __m128i zero = _mm_setzero_si128();
@@ -445,4 +443,3 @@ void vp9_add_constant_residual_32x32_sse2(const int16_t diff, uint8_t *dest,
     dest += 4 * stride;
   } while (--i);
 }
-#endif
diff --git a/vp9/decoder/x86/vp9_idct_mmx.h b/vp9/decoder/x86/vp9_idct_mmx.h
deleted file mode 100644
index 7d9829175..000000000
--- a/vp9/decoder/x86/vp9_idct_mmx.h
+++ /dev/null
@@ -1,22 +0,0 @@
-/*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- *  Use of this source code is governed by a BSD-style license
- *  that can be found in the LICENSE file in the root of the source
- *  tree. An additional intellectual property rights grant can be found
- *  in the file PATENTS.  All contributing project authors may
- *  be found in the AUTHORS file in the root of the source tree.
- */
-
-#ifndef VP9_DECODER_X86_VP9_IDCT_MMX_H_
-#define VP9_DECODER_X86_VP9_IDCT_MMX_H_
-
-
-void vp9_dequant_dc_idct_add_mmx(short *input, const short *dq,
-                                 unsigned char *pred, unsigned char *dest,
-                                 int pitch, int stride, int Dc);
-
-void vp9_dequant_idct_add_mmx(short *input, const short *dq, unsigned char *pred,
-                              unsigned char *dest, int pitch, int stride);
-
-#endif /* VP9_DECODER_X86_VP9_IDCT_MMX_H_ */
diff --git a/vp9/encoder/vp9_dct.c b/vp9/encoder/vp9_dct.c
index ecd3e2dd3..ebf40e4e6 100644
--- a/vp9/encoder/vp9_dct.c
+++ b/vp9/encoder/vp9_dct.c
@@ -1286,4 +1286,3 @@ void vp9_short_fdct32x32_c(int16_t *input, int16_t *out, int pitch) {
       out[j + i * 32] = (temp_out[j] + 1 + (temp_out[j] < 0)) >> 2;
   }
 }
-
diff --git a/vp9/encoder/vp9_sad_c.c b/vp9/encoder/vp9_sad_c.c
index fa47f813c..b4cd19358 100644
--- a/vp9/encoder/vp9_sad_c.c
+++ b/vp9/encoder/vp9_sad_c.c
@@ -577,4 +577,3 @@ void vp9_sad4x4x4d_c(const uint8_t *src_ptr,
   sad_array[3] = vp9_sad4x4(src_ptr, src_stride,
                             ref_ptr[3], ref_stride, 0x7fffffff);
 }
-
diff --git a/vp9/encoder/x86/vp9_dct_sse2_intrinsics.c b/vp9/encoder/x86/vp9_dct_sse2.c
index 49cb837e0..49cb837e0 100644
--- a/vp9/encoder/x86/vp9_dct_sse2_intrinsics.c
+++ b/vp9/encoder/x86/vp9_dct_sse2.c
diff --git a/vp9/vp9_common.mk b/vp9/vp9_common.mk
index 8abd58399..aed6cc50b 100644
--- a/vp9/vp9_common.mk
+++ b/vp9/vp9_common.mk
@@ -83,7 +83,6 @@ VP9_COMMON_SRCS-$(CONFIG_POSTPROC_VISUALIZER) += common/vp9_textblit.c
 VP9_COMMON_SRCS-yes += common/vp9_treecoder.c
 VP9_COMMON_SRCS-$(CONFIG_IMPLICIT_SEGMENTATION) += common/vp9_implicit_segmentation.c
 
-VP9_COMMON_SRCS-$(ARCH_X86)$(ARCH_X86_64) += common/x86/vp9_idct_x86.h
 VP9_COMMON_SRCS-$(ARCH_X86)$(ARCH_X86_64) += common/x86/vp9_loopfilter_x86.h
 VP9_COMMON_SRCS-$(ARCH_X86)$(ARCH_X86_64) += common/x86/vp9_postproc_x86.h
 VP9_COMMON_SRCS-$(ARCH_X86)$(ARCH_X86_64) += common/x86/vp9_asm_stubs.c
@@ -113,13 +112,13 @@ VP9_COMMON_SRCS-yes += common/vp9_maskingmv.c
 VP9_COMMON_SRCS-$(HAVE_SSE3) += common/x86/vp9_mask_sse3.asm
 endif
 
-VP9_COMMON_SRCS-$(ARCH_X86)$(ARCH_X86_64) += common/x86/vp9_idct_x86.c
+VP9_COMMON_SRCS-$(HAVE_SSE2) += common/x86/vp9_idct_intrin_sse2.c
 VP9_COMMON_SRCS-$(HAVE_SSE2) += common/x86/vp9_sadmxn_sse2.c
 ifeq ($(HAVE_SSE2),yes)
-vp9/common/x86/vp9_idct_x86.c.o: CFLAGS += -msse2
+vp9/common/x86/vp9_idct_intrin_sse2.c.o: CFLAGS += -msse2
 vp9/common/x86/vp9_loopfilter_intrin_sse2.c.o: CFLAGS += -msse2
 vp9/common/x86/vp9_sadmxn_sse2.c.o: CFLAGS += -msse2
-vp9/common/x86/vp9_idct_x86.c.d: CFLAGS += -msse2
+vp9/common/x86/vp9_idct_intrin_sse2.c.d: CFLAGS += -msse2
 vp9/common/x86/vp9_loopfilter_intrin_sse2.c.d: CFLAGS += -msse2
 vp9/common/x86/vp9_sadmxn_sse2.c.d: CFLAGS += -msse2
 endif
diff --git a/vp9/vp9_dx_iface.c b/vp9/vp9_dx_iface.c
index 2426f2175..85022c951 100644
--- a/vp9/vp9_dx_iface.c
+++ b/vp9/vp9_dx_iface.c
@@ -408,9 +408,9 @@ static void parse_superframe_index(const uint8_t *data,
   *count = 0;
 
   if ((marker & 0xe0) == 0xc0) {
-    const int frames = (marker & 0x7) + 1;
-    const int mag = ((marker >> 3) & 3) + 1;
-    const int index_sz = 2 + mag  * frames;
+    const uint32_t frames = (marker & 0x7) + 1;
+    const uint32_t mag = ((marker >> 3) & 0x3) + 1;
+    const size_t index_sz = 2 + mag * frames;
 
     if (data_sz >= index_sz && data[data_sz - index_sz] == marker) {
       // found a valid superframe index
@@ -418,7 +418,7 @@ static void parse_superframe_index(const uint8_t *data,
       const uint8_t *x = data + data_sz - index_sz + 1;
 
       for (i = 0; i < frames; i++) {
-        int this_sz = 0;
+        uint32_t this_sz = 0;
 
         for (j = 0; j < mag; j++)
           this_sz |= (*x++) << (j * 8);
@@ -447,9 +447,9 @@ static vpx_codec_err_t vp9_decode(vpx_codec_alg_priv_t  *ctx,
     // Skip over the superframe index, if present
     if (data_sz && (*data_start & 0xe0) == 0xc0) {
       const uint8_t marker = *data_start;
-      const int frames = (marker & 0x7) + 1;
-      const int mag = ((marker >> 3) & 3) + 1;
-      const int index_sz = 2 + mag  * frames;
+      const uint32_t frames = (marker & 0x7) + 1;
+      const uint32_t mag = ((marker >> 3) & 0x3) + 1;
+      const uint32_t index_sz = 2 + mag * frames;
 
       if (data_sz >= index_sz && data_start[index_sz - 1] == marker) {
         data_start += index_sz;
diff --git a/vp9/vp9cx.mk b/vp9/vp9cx.mk
index 097211388..63c6ed8b1 100644
--- a/vp9/vp9cx.mk
+++ b/vp9/vp9cx.mk
@@ -108,10 +108,10 @@ VP9_CX_SRCS-$(ARCH_X86)$(ARCH_X86_64) += encoder/x86/vp9_quantize_mmx.asm
 VP9_CX_SRCS-$(ARCH_X86)$(ARCH_X86_64) += encoder/x86/vp9_encodeopt.asm
 VP9_CX_SRCS-$(ARCH_X86_64) += encoder/x86/vp9_ssim_opt.asm
 
-VP9_CX_SRCS-$(HAVE_SSE2) += encoder/x86/vp9_dct_sse2_intrinsics.c
+VP9_CX_SRCS-$(HAVE_SSE2) += encoder/x86/vp9_dct_sse2.c
 ifeq ($(HAVE_SSE2),yes)
-vp9/encoder/x86/vp9_dct_sse2_intrinsics.c.d: CFLAGS += -msse2
-vp9/encoder/x86/vp9_dct_sse2_intrinsics.c.o: CFLAGS += -msse2
+vp9/encoder/x86/vp9_dct_sse2.c.d: CFLAGS += -msse2
+vp9/encoder/x86/vp9_dct_sse2.c.o: CFLAGS += -msse2
 endif
 
 
diff --git a/vp9/vp9dx.mk b/vp9/vp9dx.mk
index 69b29976b..babdebb86 100644
--- a/vp9/vp9dx.mk
+++ b/vp9/vp9dx.mk
@@ -37,10 +37,10 @@ VP9_DX_SRCS-yes := $(filter-out $(VP9_DX_SRCS_REMOVE-yes),$(VP9_DX_SRCS-yes))
 
 VP9_DX_SRCS-$(HAVE_SSE2) += decoder/x86/vp9_idct_blk_sse2.c
 
-VP9_DX_SRCS-$(ARCH_X86)$(ARCH_X86_64) += decoder/x86/vp9_dequantize_x86.c
+VP9_DX_SRCS-$(HAVE_SSE2) += decoder/x86/vp9_dequantize_sse2.c
 ifeq ($(HAVE_SSE2),yes)
-vp9/decoder/x86/vp9_dequantize_x86.c.o: CFLAGS += -msse2
-vp9/decoder/x86/vp9_dequantize_x86.c.d: CFLAGS += -msse2
+vp9/decoder/x86/vp9_dequantize_sse2.c.o: CFLAGS += -msse2
+vp9/decoder/x86/vp9_dequantize_sse2.c.d: CFLAGS += -msse2
 endif
 
 $(eval $(call asm_offsets_template,\