Changes the GLES2Implementation to use a RingBuffer...

Changes the GLES2Implementation to use a RingBuffer
to manage the transfer buffer.  This is significantly
faster than the FencedAllocator for our purposes.

TEST=some unit tests
BUG=none


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=45844

[apatrick, 2010-04-28 02:00:31]

lgtm

http://codereview.chromium.org/1796002/diff/1/7
File gpu/command_buffer/client/ring_buffer.h (right):

http://codereview.chromium.org/1796002/diff/1/7#newcode1
gpu/command_buffer/client/ring_buffer.h:1: // Copyright (c) 2009 The Chromium
Authors. All rights reserved.
2010

http://codereview.chromium.org/1796002/diff/1/7#newcode17
gpu/command_buffer/client/ring_buffer.h:17: // RingBuffer managed a piece of
memory as a ring buffer. Memory is allocated
A few typos in the comment

http://codereview.chromium.org/1796002/diff/1/7#newcode112
gpu/command_buffer/client/ring_buffer.h:112: //   base: The physical address the
corresponds to base_offset.
the -> that

http://codereview.chromium.org/1796002/diff/1/7#newcode118
gpu/command_buffer/client/ring_buffer.h:118: base_(static_cast<char*>(base) -
base_offset) {
char -> int8

http://codereview.chromium.org/1796002/diff/1/7#newcode164
gpu/command_buffer/client/ring_buffer.h:164: return static_cast<char *>(base_) +
offset;
char -> int8

http://codereview.chromium.org/1796002/diff/1/7#newcode169
gpu/command_buffer/client/ring_buffer.h:169: return static_cast<char *>(pointer)
- static_cast<char *>(base_);
char -> int8

http://codereview.chromium.org/1796002/diff/1/7#newcode183
gpu/command_buffer/client/ring_buffer.h:183: RingBuffer &allocator() { return
allocator_; }
Are we supposed to return a pointer in this case?

http://codereview.chromium.org/1796002/diff/1/4
File gpu/gpu.gyp (right):

http://codereview.chromium.org/1796002/diff/1/4#newcode233
gpu/gpu.gyp:233: 'command_buffer/client/fenced_allocator.cc',
why this change?

[greggman, 2010-04-28 18:39:35]

http://codereview.chromium.org/1796002/diff/1/7
File gpu/command_buffer/client/ring_buffer.h (right):

http://codereview.chromium.org/1796002/diff/1/7#newcode1
gpu/command_buffer/client/ring_buffer.h:1: // Copyright (c) 2009 The Chromium
Authors. All rights reserved.
On 2010/04/28 02:00:31, apatrick wrote:
> 2010

Done.

http://codereview.chromium.org/1796002/diff/1/7#newcode17
gpu/command_buffer/client/ring_buffer.h:17: // RingBuffer managed a piece of
memory as a ring buffer. Memory is allocated
On 2010/04/28 02:00:31, apatrick wrote:
> A few typos in the comment

Done.

http://codereview.chromium.org/1796002/diff/1/7#newcode112
gpu/command_buffer/client/ring_buffer.h:112: //   base: The physical address the
corresponds to base_offset.
On 2010/04/28 02:00:31, apatrick wrote:
> the -> that

Done.

http://codereview.chromium.org/1796002/diff/1/7#newcode118
gpu/command_buffer/client/ring_buffer.h:118: base_(static_cast<char*>(base) -
base_offset) {
On 2010/04/28 02:00:31, apatrick wrote:
> char -> int8

Done.

http://codereview.chromium.org/1796002/diff/1/7#newcode164
gpu/command_buffer/client/ring_buffer.h:164: return static_cast<char *>(base_) +
offset;
On 2010/04/28 02:00:31, apatrick wrote:
> char -> int8

Done.

http://codereview.chromium.org/1796002/diff/1/7#newcode169
gpu/command_buffer/client/ring_buffer.h:169: return static_cast<char *>(pointer)
- static_cast<char *>(base_);
On 2010/04/28 02:00:31, apatrick wrote:
> char -> int8

Done.

http://codereview.chromium.org/1796002/diff/1/4
File gpu/gpu.gyp (right):

http://codereview.chromium.org/1796002/diff/1/4#newcode233
gpu/gpu.gyp:233: 'command_buffer/client/fenced_allocator.cc',
On 2010/04/28 02:00:31, apatrick wrote:
> why this change?

Done.

[apatrick_chromium, 2010-04-28 19:35:44]

lgtm