Kostya Serebryany
|
4fd30769c1
|
[sanitizer] remove kBatchClassID that is not used any more; NFC
llvm-svn: 280185
|
2016-08-31 00:37:33 +00:00 |
|
Kostya Serebryany
|
8e7ea9ddb5
|
[sanitizer] re-apply r279572 and r279595 reverted in r279643: change the 64-bit allocator to use a single array for free-d chunks instead of a lock-free linked list of tranfer batches. This change simplifies the code, makes the allocator more 'hardened', and will allow simpler code to release RAM to OS. This may also slowdown malloc stress tests due to lock contension, but I did not observe noticeable slowdown on various real multi-threaded benchmarks.
llvm-svn: 279664
|
2016-08-24 21:20:10 +00:00 |
|
Vitaly Buka
|
769ec705bc
|
Revert r279572 "[sanitizer] change the 64-bit..." because of failures on ubsan
This reverts commit r279572 and r279595.
llvm-svn: 279643
|
2016-08-24 17:40:29 +00:00 |
|
Kostya Serebryany
|
f46d50e360
|
[sanitizer] change the 64-bit allocator to use a single array for free-d chunks instead of a lock-free linked list of tranfer batches. This change simplifies the code, makes the allocator more 'hardened', and will allow simpler code to release RAM to OS. This may also slowdown malloc stress tests due to lock contension, but I did not observe noticeable slowdown on various real multi-threaded benchmarks.
llvm-svn: 279572
|
2016-08-23 21:19:47 +00:00 |
|
Kostya Serebryany
|
ce2163459f
|
[sanitizer] allocator: split the local cache class into two, one for 32-bit allocator and one for 64-bit one. NFC. The two imlementations will diverge in the following changes.
llvm-svn: 279495
|
2016-08-23 00:30:43 +00:00 |
|
Kostya Serebryany
|
1ab649649d
|
[sanitizer] use 32-bit offset instead of 64-bit pointers in the 64-bit allocator's transfer batches. This saves 2x memory for the transfer batches (up to ~1.5% overall in some cases)
llvm-svn: 278179
|
2016-08-09 23:30:22 +00:00 |
|
Kostya Serebryany
|
d4a5749677
|
[sanitizer] minor refactoring in the allocator, NFC
llvm-svn: 278163
|
2016-08-09 20:54:50 +00:00 |
|
Kostya Serebryany
|
4cd2845e6a
|
[sanitizer] allocator: move TransferBatch into SizeClassAllocator64/SizeClassAllocator32 because we actually need different iplementations for the 64- and 32-bit case. NFC; the following patches will make the TransferBatch implementations differ
llvm-svn: 277899
|
2016-08-06 01:24:11 +00:00 |
|
Kostya Serebryany
|
dcaec77936
|
[sanitizer] remove one redundant loop from the allocator
llvm-svn: 277563
|
2016-08-03 03:42:55 +00:00 |
|
Kostya Serebryany
|
c49e296805
|
[sanitizer] refactor TransferBatch to hide the implementation. NFC expected. Second attempt after failed r276383 which was reverted.
llvm-svn: 277554
|
2016-08-03 00:14:10 +00:00 |
|
Kostya Serebryany
|
d251e94a01
|
[sanitizer] revert r276383 while investigating failures on bot
llvm-svn: 276456
|
2016-07-22 19:02:59 +00:00 |
|
Kostya Serebryany
|
5bc01c108d
|
[sanitizer] refactor TransferBatch to hide the implementation. NFC
llvm-svn: 276383
|
2016-07-22 02:21:12 +00:00 |
|
Kostya Serebryany
|
bfd69e0580
|
[sanitizer] allocator: introduce kUseSeparateSizeClassForBatch (false by default). When true, it will cause all TransferBatches to be allocated on a separate dedicated size class, which improves security and may potentially simplify memory reclamation. However in the current state this may cause up to 3% extra memory usage. Subsequent changes should bring this overhead down
llvm-svn: 276377
|
2016-07-22 01:13:13 +00:00 |
|
Kostya Serebryany
|
35eeea707e
|
[sanitizer] allocator: remove kPopulateSize and only use SizeClassMap::MaxCached; ensure that TransferBatch size is a power of two, refactor TransferBatch creation/destruction into separate functions.
llvm-svn: 276318
|
2016-07-21 18:47:53 +00:00 |
|
Kostya Serebryany
|
9835a81529
|
[sanitizers] split sanitizer_allocator.h into a number of smaller .h files; NFC
llvm-svn: 276195
|
2016-07-20 22:06:41 +00:00 |