Provide a Chrome-owned buffer to FFmpeg for video decoding, instead of

media/base/video_frame.cc

Index: media/base/video_frame.cc
===================================================================
--- media/base/video_frame.cc	(revision 140753)
+++ media/base/video_frame.cc	(working copy)
@@ -8,7 +8,12 @@
 #include "base/string_piece.h"
 #include "media/base/limits.h"
 #include "media/base/video_util.h"
+#if !defined(OS_ANDROID)
+#include "media/ffmpeg/ffmpeg_common.h"
+#endif
 
+#include <algorithm>
+
 namespace media {
 
 // static
@@ -16,18 +21,19 @@
     VideoFrame::Format format,
     size_t width,
     size_t height,
+    size_t alignment,
     base::TimeDelta timestamp,
     base::TimeDelta duration) {
-  DCHECK(IsValidConfig(format, width, height));
+  DCHECK(IsValidConfig(format, width, height, alignment));
   scoped_refptr<VideoFrame> frame(new VideoFrame(
       format, width, height, timestamp, duration));
   switch (format) {
     case VideoFrame::RGB32:
-      frame->AllocateRGB(4u);
+      frame->AllocateRGB(4u, alignment);
       break;
     case VideoFrame::YV12:
     case VideoFrame::YV16:
-      frame->AllocateYUV();
+      frame->AllocateYUV(alignment);
       break;
     default:
       LOG(FATAL) << "Unsupported frame format: " << format;
@@ -35,16 +41,23 @@
   return frame;
 }
 
+static inline bool IsPowerOfTwo(int alignment) {
+  // Check that |alignment| is a power of 2.
+  return (alignment + (alignment - 1)) == (alignment | (alignment - 1));
+}
+
 // static
 bool VideoFrame::IsValidConfig(
     VideoFrame::Format format,
     size_t width,
-    size_t height) {
+    size_t height,
+    size_t alignment) {
 
   return (format != VideoFrame::INVALID &&
           width > 0 && height > 0 &&
           width <= limits::kMaxDimension && height <= limits::kMaxDimension &&
-          width * height <= limits::kMaxCanvas);
+          width * height <= limits::kMaxCanvas &&
+          alignment > 0 && IsPowerOfTwo(alignment));
 }
 
 // static
@@ -88,42 +101,60 @@
 }
 
 static inline size_t RoundUp(size_t value, size_t alignment) {
-  // Check that |alignment| is a power of 2.
-  DCHECK((alignment + (alignment - 1)) == (alignment | (alignment - 1)));
+  DCHECK(IsPowerOfTwo(alignment));
   return ((value + (alignment - 1)) & ~(alignment-1));
 }
 
-void VideoFrame::AllocateRGB(size_t bytes_per_pixel) {
-  // Round up to align at a 64-bit (8 byte) boundary for each row.  This
-  // is sufficient for MMX reads (movq).
-  size_t bytes_per_row = RoundUp(width_ * bytes_per_pixel, 8);
+void VideoFrame::AllocateRGB(size_t bytes_per_pixel, size_t alignment) {
+  // Round up to align at least at a 64-bit (8 byte) boundary for each row.
+  // This is sufficient for MMX reads (movq).
+  size_t bytes_per_row = RoundUp(width_ * bytes_per_pixel,
+                                 std::max(alignment, static_cast<size_t>(8)));
+  size_t aligned_height = RoundUp(height_, alignment);
   strides_[VideoFrame::kRGBPlane] = bytes_per_row;
-  data_[VideoFrame::kRGBPlane] = new uint8[bytes_per_row * height_];
+#if !defined(OS_ANDROID)
+  // TODO(dalecurtis): use DataAligned or so, so this #ifdef hackery
+  // doesn't need to be repeated in every single user of aligned data.
+  data_[VideoFrame::kRGBPlane] = reinterpret_cast<uint8*>(
+    av_malloc(bytes_per_row * aligned_height));

[scherkus (not reviewing), 2012/06/14 19:21:20]

sorry -- "4 space indent" meant from the existing indentation for a total indent
of 6 spaces

[rbultje1, 2012/06/14 19:26:25]

Done.

+#else
+  data_[VideoFrame::kRGBPlane] = new uint8_t[bytes_per_row * aligned_height];
+#endif
   DCHECK(!(reinterpret_cast<intptr_t>(data_[VideoFrame::kRGBPlane]) & 7));
   COMPILE_ASSERT(0 == VideoFrame::kRGBPlane, RGB_data_must_be_index_0);
 }
 
 static const int kFramePadBytes = 15;  // Allows faster SIMD YUV convert.
 
-void VideoFrame::AllocateYUV() {
+void VideoFrame::AllocateYUV(size_t alignment) {
   DCHECK(format_ == VideoFrame::YV12 || format_ == VideoFrame::YV16);
-  // Align Y rows at 32-bit (4 byte) boundaries.  The stride for both YV12 and
-  // YV16 is 1/2 of the stride of Y.  For YV12, every row of bytes for U and V
-  // applies to two rows of Y (one byte of UV for 4 bytes of Y), so in the
-  // case of YV12 the strides are identical for the same width surface, but the
-  // number of bytes allocated for YV12 is 1/2 the amount for U & V as YV16.
-  // We also round the height of the surface allocated to be an even number
-  // to avoid any potential of faulting by code that attempts to access the Y
-  // values of the final row, but assumes that the last row of U & V applies to
-  // a full two rows of Y.
-  size_t y_height = rows(VideoFrame::kYPlane);
-  size_t y_stride = RoundUp(row_bytes(VideoFrame::kYPlane), 4);
-  size_t uv_stride = RoundUp(row_bytes(VideoFrame::kUPlane), 4);
-  size_t uv_height = rows(VideoFrame::kUPlane);
+  // Align Y rows at least at 32-bit (4 byte) boundaries.  The stride for both
+  // YV12 and YV16 is 1/2 of the stride of Y.  For YV12, every row of bytes for
+  // U and V applies to two rows of Y (one byte of UV for 4 bytes of Y), so in
+  // the case of YV12 the strides are identical for the same width surface, but
+  // the number of bytes allocated for YV12 is 1/2 the amount for U & V as
+  // YV16. We also round the height of the surface allocated to be an even
+  // number to avoid any potential of faulting by code that attempts to access
+  // the Y values of the final row, but assumes that the last row of U & V
+  // applies to a full two rows of Y.
+  size_t y_stride = RoundUp(row_bytes(VideoFrame::kYPlane),
+                            std::max(alignment, static_cast<size_t>(4)));
+  size_t uv_stride = RoundUp(row_bytes(VideoFrame::kUPlane),
+                             std::max(alignment, static_cast<size_t>(4)));
+  size_t y_height = RoundUp(height_, std::max(alignment,

[scherkus (not reviewing), 2012/06/14 19:21:20]

does the height need to be an even number or rounded up?

for example you pass in 16 here -- does that mean a height of 19 would be 32
instead of 20?

[rbultje1, 2012/06/14 19:26:25]

FFmpeg requires a multiple of 16, since each is decoded as a complete 16x16
macroblock.

+                                              static_cast<size_t>(2)));
+  size_t uv_height = format_ == VideoFrame::YV12 ? y_height / 2 : y_height;
   size_t y_bytes = y_height * y_stride;
   size_t uv_bytes = uv_height * uv_stride;
 
-  uint8* data = new uint8[y_bytes + (uv_bytes * 2) + kFramePadBytes];
+#if !defined(OS_ANDROID)
+  // TODO(dalecurtis): use DataAligned or so, so this #ifdef hackery
+  // doesn't need to be repeated in every single user of aligned data.
+  uint8* data = reinterpret_cast<uint8*>(
+    av_malloc(y_bytes + (uv_bytes * 2) + kFramePadBytes));

[scherkus (not reviewing), 2012/06/14 19:21:20]

indent here

[rbultje1, 2012/06/14 19:26:25]

Done.

+#else
+  uint8* data = new uint8_t[y_bytes + (uv_bytes * 2) + kFramePadBytes];
+#endif
   COMPILE_ASSERT(0 == VideoFrame::kYPlane, y_plane_data_must_be_index_0);
   data_[VideoFrame::kYPlane] = data;
   data_[VideoFrame::kUPlane] = data + y_bytes;
@@ -158,7 +189,13 @@
   // In multi-plane allocations, only a single block of memory is allocated
   // on the heap, and other |data| pointers point inside the same, single block
   // so just delete index 0.
-  delete[] data_[0];
+  if (data_[0]) {
+#if !defined(OS_ANDROID)
+    av_free(data_[0]);
+#else
+    delete[] data_[0];
+#endif
+  }
 }
 
 bool VideoFrame::IsValidPlane(size_t plane) const {