| bdd17cde8cf675f5d5703aad4dfb565932fb4c50 |
|
08-Feb-2012 |
Elliott Hughes <enh@google.com> |
Fix an ICS regression with hasArray on a JNI-allocated DirectByteBuffer.
Bug: http://code.google.com/p/android/issues/detail?id=24327
Change-Id: Ib31b60fe01e3d0b5a1ed7125ddf047e09fd64056
|
| f934c3d2c8dd9e6bc5299cef41adace2a671637d |
|
15-Mar-2011 |
Elliott Hughes <enh@google.com> |
Make OSMemory the semi-supported libcore.io.Memory.
Looks like we're not going to bother with a separate libcore.os package,
when libcore.io will do.
Change-Id: I2806c59349ed4b6410d768c4207c384ced973c54
|
| a1603838fe9e865575c87982e32c6343740e464c |
|
11-Dec-2010 |
Elliott Hughes <enh@google.com> |
Lots more bounds-checking/exception-throwing consistency.
Overflow-safe checks all round, plus better detail messages. This isn't
quite everything, but it's a large chunk of the work. Most notably, this
is all of io and nio.
There are numerous changes of exception priority here, and the harmony
tests noticed a subset of them in the nio code. I've modified our checked-out
copy of the tests to accept any of the throwable exceptions.
Change-Id: Id185f1228fb9a1d5fc9494e78375b5623fb0fe14
|
| 43a9f774d075e0e441d8b996e3f6c81ea483ec89 |
|
19-Oct-2010 |
Elliott Hughes <enh@google.com> |
BufferIterator support for byte[].
Bug: 3032515
Change-Id: I040a92c3ce241aa97719ff94e3cc31c62473bf46
|
| af301fb0e674665dae3e9cb94344558923ec8e0a |
|
19-Oct-2010 |
Elliott Hughes <enh@google.com> |
Factor out the code that peeks/pokes ints/longs/shorts from a byte[].
My intention here is that I'll want to reuse this code in my BufferIterator
for byte[]s (http://b/3032515).
This actually makes a noticeable nio benchmark improvement, presumably because
of the switch to static methods, so this helps http://b/2985452 too.
I plan on experimenting with native implementations for the new OSMemory
methods, but that can wait.
Change-Id: Ie5b34910b1e99930ba1f13c0e7002245b901039b
|
| 17b2f493468764dc1f237c02d94747d122dbd101 |
|
08-Oct-2010 |
Elliott Hughes <enh@google.com> |
Fix our byte order behavior for slice and wrap operations.
Our documentation was correct, but our implementation was not. Includes
extensive new tests.
Bug: 3066167
Change-Id: Ie7d6e6432f1d675f20af020cc662dab6eb181a50
|
| 8fbc397fc09158bee0bc0cb231c609c4c6e9fc15 |
|
18-Sep-2010 |
Elliott Hughes <enh@google.com> |
Optimize non-direct (heap) buffer bulk put.
I've also factored out the bounds checking, even though the JIT can't yet
inline it back. I'm unwilling to duplicate all that code again.
This patch also fixes bugs in the previous checkin relating to offsets
and positions. In real life (and, seemingly, in all our tests) these tend
to be 0, making it rare that anyone trips over these bugs. (The first nio
change I made was fixing just such a bug.) I've improved the test I added
last time so that it covers more of these cases.
Here are the "after" performance figures for heap buffer bulk copies:
CharBuffer_getCharArray 5797
CharBuffer_putCharArray 5818
DoubleBuffer_getDoubleArray 13053
DoubleBuffer_putDoubleArray 13026
FloatBuffer_getFloatArray 8216
FloatBuffer_putFloatArray 8283
IntBuffer_getIntArray 8253
IntBuffer_putIntArray 8371
LongBuffer_getLongArray 13053
LongBuffer_putLongArray 13020
ShortBuffer_getShortArray 5796
ShortBuffer_putShortArray 5864
This compares well with with direct buffer performance (just gets shown here):
CharBuffer_getCharArray 4412
DoubleBuffer_getDoubleArray 11594
FloatBuffer_getFloatArray 6787
IntBuffer_getIntArray 6791
LongBuffer_getLongArray 11672
ShortBuffer_getShortArray 4430
Bug: 2985452
Change-Id: I37307f9b304c87a87af5313790444dcdf6174ff1
|
| 0bc30ae8420b5b1abc1b2eefbdf8846309b5447d |
|
13-Sep-2010 |
Elliott Hughes <enh@google.com> |
More nio speedups.
Small improvements this time. Stop penalizing getChar/putChar with an extra
level of indirection through getShort/putShort; our VM can't win that
back yet, and the extra method call is a significant fraction of the
total time taken for these operations (1/3rd). This applies to direct, heap,
and mapped buffers. (One day we'll be able to come back and remove this.)
Also remove the float/double specializations from MemoryBlock, rewriting
the two callers of each to use int/long directly. Those callers are public
API, so this is as far as the removal of useless duplication can go here.
This applies to direct and mapped buffers.
Unroll loops (!) used to load/store short/int/long values from Java heap
byte[]s for performance. This applies to heap buffers.
Bug: 2935622
Change-Id: I093122af2b49d3ef864b87dce637c781727445dd
|
| 0e1b748ecabf720065a632f28330f5d4d037d5ae |
|
10-Sep-2010 |
Elliott Hughes <enh@google.com> |
Consistent naming for get/put/wrap arguments.
Bug: 2935622
Change-Id: Icb39392f9106f9fe5ba70ccdcb6b059dbf5617bc
|
| 0c53cf8b2c46deb41e91db50ddc17d598cc64a10 |
|
09-Sep-2010 |
Elliott Hughes <enh@google.com> |
Speed up MappedByteBuffer.
Make it possible to override limit(int) and position(int) internally, so I can
avoid updating MappedByteBufferAdapter's wrapped ByteBuffer's limit and
position on every get or put. I still need to do extra work with position if
you use the get/put overloads that don't take an index, but even that becomes
cheaper. The indexed overloads benefit the most, though. Before:
benchmark before after (both us)
ByteBuffer_getByte 830 757
ByteBuffer_getChar 940 875
ByteBuffer_getByte_indexed 800 509
ByteBuffer_getChar_indexed 936 643
Also factor out duplicated toString implementations.
Bug: 2935622
Change-Id: I4cfe206fde638bddb5a7b4a62223e2588f44b8ac
|
| 29dbfe19b113a13b712be2bc762ef1c81cd06c47 |
|
03-Sep-2010 |
Elliott Hughes <enh@google.com> |
Use finalizers to free direct buffers.
Also switch to a cheaper scheme for avoiding double-frees (which could only
happen if library code is broken; at the moment the only caller is the
iovec-style stuff). This halves the cost of direct buffer access, at the
unimportant cost of not throwing an exception if we try to access a buffer
we freed while we still had a reference to it.
Also use named constants for the sizes of primitive types.
Also kill BufferFactory.
Bug: 2935622
Change-Id: I6021ce7531f49d8cff7635d6d2cc54851bacaf69
|
| f33eae7e84eb6d3b0f4e86b59605bb3de73009f3 |
|
13-May-2010 |
Elliott Hughes <enh@google.com> |
Remove all trailing whitespace from the dalvik team-maintained parts of libcore.
Gentlemen, you may now set your editors to "strip trailing whitespace"...
Change-Id: I85b2f6c80e5fbef1af6cab11789790b078c11b1b
|
| 3ff29eb11a241d58c54ecc84230bb0672f4aa148 |
|
23-Apr-2010 |
Peter Hallam <peterhal@google.com> |
More merging of nio into luni
Change-Id: Ib4356b3cd57490df8ff246ea6a4bc1c269ba848b
|