Make transfer buffer reallocate at default size.

Make transfer buffer reallocate at default size.

TEST=unit tests
BUG=124109
R=vangelis@chromium.org


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=133781

[greggman, 2012-04-23 22:43:14]

Hey Vangelis, PTAL

[vangelis, 2012-04-23 23:46:35]

LGTM

A good bug to attach this to would be: 124109

http://codereview.chromium.org/10197002/diff/1/gpu/command_buffer/client/tran...
File gpu/command_buffer/client/transfer_buffer.cc (right):

http://codereview.chromium.org/10197002/diff/1/gpu/command_buffer/client/tran...
gpu/command_buffer/client/transfer_buffer.cc:48:
ReallocateRingBuffer(default_buffer_size_ - result_size);
Why do we subtract the result_size here?

http://codereview.chromium.org/10197002/diff/1/gpu/command_buffer/client/tran...
File gpu/command_buffer/client/transfer_buffer_unittest.cc (right):

http://codereview.chromium.org/10197002/diff/1/gpu/command_buffer/client/tran...
gpu/command_buffer/client/transfer_buffer_unittest.cc:442:
CreateTransferBuffer(kStartTransferBufferSize, _))
Should you pass a size < kStartTransferBufferSize to check for the regression
that this patch is fixing?

[greggman, 2012-04-24 20:04:46]

http://codereview.chromium.org/10197002/diff/1/gpu/command_buffer/client/tran...
File gpu/command_buffer/client/transfer_buffer.cc (right):

http://codereview.chromium.org/10197002/diff/1/gpu/command_buffer/client/tran...
gpu/command_buffer/client/transfer_buffer.cc:48:
ReallocateRingBuffer(default_buffer_size_ - result_size);
On 2012/04/23 23:46:35, vangelis wrote:
> Why do we subtract the result_size here?  

because the start of the buffer is reserved. I could change it to
default_buffer_size + result_size but it works this way now

http://codereview.chromium.org/10197002/diff/1/gpu/command_buffer/client/tran...
File gpu/command_buffer/client/transfer_buffer_unittest.cc (right):

http://codereview.chromium.org/10197002/diff/1/gpu/command_buffer/client/tran...
gpu/command_buffer/client/transfer_buffer_unittest.cc:442:
CreateTransferBuffer(kStartTransferBufferSize, _))
On 2012/04/23 23:46:35, vangelis wrote:
> Should you pass a size < kStartTransferBufferSize to check for the regression
> that this patch is fixing? 

No, the bug was that it was asking for a smaller transfer buffer. Before the
change this failed (tested), now that it allocates the starting size (default
size) this test passes.

[commit-bot: I haz the power, 2012-04-24 20:05:10]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/gman@chromium.org/10197002/1

[commit-bot: I haz the power, 2012-04-24 22:11:11]

Change committed as 133781