mirrors/mimalloc
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
1,669 Commits 21 Branches 59 Tags 7.4 MiB
745a34f475
Commit Graph

489 Commits

Author SHA1 Message Date
Daan
ec2265486e bump version for further development 2022-02-14 16:47:57 -08:00
Daan
817569dfad bump to version x.0.5 2022-02-14 16:34:18 -08:00
Daan
c6f32c9533
Merge pull request #528 from michaeljclark/mimalloc-fixes 
mimalloc-types: amend comment adding medium to list of page kinds
 2022-02-02 19:57:50 -08:00
Daan
32da74e032
Merge pull request #539 from devnexen/obsd_asm_fix 
tls revives inline asm for openbsd
 2022-02-02 19:45:39 -08:00
Daan
f515b80093 update documentation on options 2022-02-02 19:19:03 -08:00
David Carlier
b1e1b0573c tls revives inline asm for openbsd 2022-01-27 07:01:32 +00:00
Michael Clark
9da8788dae
mimalloc-types: amend comment adding medium to list of page kinds 2022-01-19 10:14:08 +13:00
Daan
55b4d5e2b1
Merge pull request #517 from tiran/debug-uninit 
Allow overrides of MI_DEBUG memory constants
 2022-01-10 16:12:10 -08:00
Daan
4fc63358a6 update copyright year 2022-01-10 12:01:26 -08:00
Daan
4b63c76861 avoid conditional load on macos 2022-01-10 11:40:36 -08:00
Christian Heimes
0353fc38dd Allow overrides of MI_DEBUG memory constants 
CPython and Windows CRT debug builds use different values for uninit,
freed, and padding bytes. Make ``MI_DEBUG_*`` constants conditional to
allow embedders to override the constants.

Windows dbgheap:

```
_bNoMansLandFill = 0xFD
_bDeadLandFill   = 0xDD
_bCleanLandFill  = 0xCD
```

Python memory debug

```
PYMEM_CLEANBYTE      0xCD
PYMEM_DEADBYTE       0xDD
PYMEM_FORBIDDENBYTE  0xFD
```

Signed-off-by: Christian Heimes <christian@python.org>
 2022-01-05 13:22:47 +01:00
daan
43e5cd2671 revise assembly test for the _mi_threadid(); fix issue #495 with musl libc on arm; test specifically for bionic libc on Android 2022-01-01 15:54:06 -08:00
Daan
6ead2840ec
Merge pull request #504 from res2k/tweak-stats-doc 
Tweak documentation for mi_heap_area_t::used.
 2021-12-20 17:44:52 -08:00
Frank Richter
ff3bb5ce7b Tweak documentation for mi_heap_area_t::used. 
For microsoft/mimalloc#503.
 2021-12-19 14:51:15 +01:00
Daan Leijen
89090510bd update alignment tests 2021-12-18 11:11:44 -08:00
daan
30a99e2c51 default to portable C for the thread id on arm 32-bit, issue #495 2021-12-17 13:48:39 -08:00
daan
89f583a69b improve aligned allocation performance 2021-12-17 13:18:05 -08:00
daan
684c2c82a7 restrict max aligment boundary to prevent bug with segment determination (found by Matthew Parkinson). 2021-12-17 11:40:46 -08:00
daan
55ccc94707 wip: fix zero init for sbrk; add lock around sbrk 2021-12-15 19:28:34 -08:00
daan
82a0e56f35 fix warning on mi_reallocarr 2021-12-15 18:54:57 -08:00
Daan
4a586808f7 merge 2021-12-15 16:52:36 -08:00
Daan
317093d78b fix Android overriding (issue #381 and PR #487) 2021-12-15 16:48:57 -08:00
Daan
17456f18d1
Merge pull request #489 from devnexen/remove_dfbsd_warning 
removes dragonflybsd warning since the api and library overriding wor…
 2021-12-15 16:37:56 -08:00
Daan
1b09877bc4
Merge branch 'dev' into netbsd_reallocarr 2021-12-15 16:27:26 -08:00
Daan
523e6d5c9b fix thread slot on arm32 (issue #495) 2021-12-15 16:05:07 -08:00
Daan
69b6b24688 further improvements to installation directories 2021-12-14 18:29:14 -08:00
Daan
c858690dea Merge branch 'dev' of https://github.com/microsoft/mimalloc into dev 2021-12-12 10:35:19 -08:00
Daan
3a212d1895 fix assembly for mi_tls_slot_set on x32 and x64. Issue #488 2021-12-12 10:35:13 -08:00
David Carlier
d6a56dd99d introduces netbsd's reallocarr overriding. 
and couple of tests.
 2021-12-01 22:43:13 +00:00
DC
20e37bf7c1 removes dragonflybsd warning since the api and library overriding works, also 
for the os doc report purpose.
 2021-11-28 08:14:21 +00:00
daan
9183b1eec0 remove experiment with unsafe_free_with_threadid 2021-11-23 19:04:41 -08:00
daan
8b60a5ab70 add mi_unsafe_free_with_threadid and mi_get_current_threadid() 2021-11-23 17:59:27 -08:00
daan
85eccf324f Merge branch 'master' into dev 2021-11-14 16:44:54 -08:00
Daan
4a590b1447 bump version number to 1.7.3 2021-11-14 15:33:56 -08:00
Daan
d67ff1ca9f add include cstdef for std::size_t in C++ 2021-11-14 15:32:43 -08:00
Daan
9c9c3a0311
Merge pull request #485 from devnexen/builtin_overflow_sig_fix 
fix spurious build warning with overflow builtins
 2021-11-14 14:47:38 -08:00
daan
4b0cda2012 take overcommit into account for eager commit delay 2021-11-14 11:23:11 -08:00
daan
28896e5b19 prefix UNUSED,KiB,MiB,GiB; add mi_threadid_t type; add mi_ssize_t 2021-11-13 14:46:03 -08:00
daan
f9ac60a90f suppres eager commit delay for the first 4 threads 2021-11-12 19:03:20 -08:00
Daan
e853f530a0 add noexcept attributes to improve mi_free codegen 2021-11-04 18:54:57 -07:00
Daan
751a2249c8 add mi_decl_externc, nice layout 2021-11-02 21:53:20 -07:00
Daan
5360639748 add mi_malloc_good_size 2021-11-02 21:52:00 -07:00
dc
5b0e73281f fix spurious build warning with overflow builtins 2021-10-28 22:43:21 +01:00
Daan
f38956568c only set visibility attribute when compiling for a shared library 2021-10-27 17:01:12 -07:00
Christian Heimes
13de1920ae Rename _os_random_weak to _mi_os_random_weak 
The ``_os_random_weak`` function is the only non-static function
besides ``_ZSt15get_new_handlerv`` that is not prefixed with ``mi`` or
``_mi``.

The discrepancy was discovered by CPython's smelly script. The checker
looks for exported symbols that don't have well-defined prefixes.

Signed-off-by: Christian Heimes <christian@python.org>
 2021-10-21 21:40:45 +02:00
Daan
75987e4590
Merge pull request #410 from jserv/enforce-binary-prefix 
Distinguish SI and Binary Prefixes
 2021-10-19 12:28:33 -07:00
Daan
6068a8bb1b
Merge pull request #420 from hankluo6/typo 
Fix typo
 2021-10-19 12:17:44 -07:00
Daan
20a3eee6f0
Merge pull request #449 from bmalrat/FixTypo 
Fixed typo in header
 2021-10-19 10:24:54 -07:00
Daan
5b9409f4d6 add space after _Atomic to prevent errors on msvc without /TP (see PR #452) 2021-10-19 10:17:30 -07:00
Daan
1917fbd847
Merge pull request #467 from tiran/strict_prototypes 
Fix strict function prototype warnings
 2021-10-19 09:35:40 -07:00
Christian Heimes
7c73e3996d Fix strict function prototype warnings 
Fix warning ``warning: function declaration isn’t a prototype`` when
building mimalloc with ``-Wstrict-prototypes`` flag. In C argumentless
functions should be declared as ``func(void)``.

Reproducer:
```shell
$ cmake ../.. -DCMAKE_C_FLAGS="-Wstrict-prototypes"
$ make VERBOSE=1
```

Co-authored-by: Sam Gross <colesbury@gmail.com>
Co-authored-by: Neil Schemenauer <nas@arctrix.com>
Signed-off-by: Christian Heimes <christian@python.org>
 2021-10-19 10:48:26 +02:00
Daan Leijen
9c5928060e Merge branch 'dev' of https://github.com/microsoft/mimalloc into dev 2021-10-01 15:10:42 -07:00
Daan Leijen
679aad0659 update wasm support with emscripten compilation; now using sbrk instead of wasm_memory_grow 2021-10-01 15:05:01 -07:00
paulip1792
a2c3b0f8af add option to reserve huge os pages at a specific numa node. 2021-08-10 16:30:44 +08:00
bmalrat
30be78d97a Fixed typo in headers 2021-08-04 17:31:48 -04:00
Jim Huang
4369fe4323 Eliminate preprocessor warnings due to undefined "__GNUC__" with ClangCL 
When building some code against mimalloc with C inside Visual Studio
with ClangCL, the compiler complains about __GNUC__ being undefined.

Reported by Mojca Miklavec.

Close #422
 2021-06-24 17:29:06 +08:00
hank
1c1571742d fix typo 2021-06-21 22:36:47 +08:00
Daan Leijen
728be93977 fix for #414 making numa node count atomic 2021-06-17 19:38:51 -07:00
Daan Leijen
a83bca72b3 fixes for M1; disable interpose use zones; fix pedantic warnings 2021-06-17 19:15:09 -07:00
Daan Leijen
aeb62c2711 fix double quote includes 2021-06-07 16:50:31 -07:00
Daan Leijen
4ba32c3160 Revert "make all includes relative" 
This reverts commit 1feb6123d9.
 2021-06-07 16:47:57 -07:00
Daan Leijen
1feb6123d9 make all includes relative 2021-06-06 20:31:36 -07:00
Jim Huang
0f57425f80 Distinguish SI and Binary Prefixes 
SI prefixes [the decimal prefixes] refer strictly to powers of 10. They
should not be used to indicate powers of 2. e.g., one kilobit
represents 1000 bits instead of 1024 bits. IEC 60027‐2 symbols are
formed adding a "i" to the SI symbol (e.g. G + i = Gi).
 2021-05-30 20:13:28 +08:00
Jim Huang
5940d3bcce Bump copyright date 
Each source file has been changed according to relevant Git activities.
 2021-04-24 16:35:11 +00:00
Daan
766f1f9345
Merge pull request #388 from nico-abram/patch-2 
Fix typo in comment
 2021-04-22 10:34:13 -07:00
Daan
f941015928
Merge pull request #384 from kdrag0n/fix-android-thread-id 
Fix thread ID getter on Android ARM/AArch64
 2021-04-22 10:33:53 -07:00
Jim Huang
3402c6cc3f Revise the use of macOS predefined macro 
Quoted from "Porting UNIX/Linux Applications to OS X,"[1]
* macro __MACH__ is defined if Mach system calls are supported;
* macro __APPLE__ is defined in any Apple computer.

__MACH__ is not specific to macOS since GNU/Hurd runs on a Mach-based
microkernel (gnumach) [2]. __MACH__ is defined by the compiler,
leading to potential confusions. The solution is just changing the
checked identifier (i.e. __APPLE__), so it is really used only on
macOS.

[1] https://developer.apple.com/library/archive/documentation/Porting/Conceptual/PortingUnix/compiling/compiling.html
[2] https://www.gnu.org/software/hurd/microkernel/mach/gnumach.html
 2021-04-21 15:24:02 +08:00
unknown
8311cef0d1 Fix typo in comment 
it -> if in mimalloc-types.h
 2021-04-17 15:08:25 -03:00
Danny Lin
ad2fa2bf6f
Fix thread ID getter on Android ARM/AArch64 
Android's Bionic libc stores the thread ID in TLS slot 1 instead of 0
on 32-bit ARM and AArch64. Slot 0 contains a pointer to the ELF DTV
(Dynamic Thread Vector) instead, which is constant for each loaded DSO.

Because mimalloc uses the thread ID to determine whether operations are
thread-local or cross-thread (atomic), all threads having the same ID
causes internal data structures to get corrupted quickly when multiple
threads are using the allocator:

mimalloc: assertion failed: at "external/mimalloc/src/page.c":563, mi_page_extend_free
  assertion: "page->local_free == NULL"
mimalloc: assertion failed: at "external/mimalloc/src/page.c":74, mi_page_is_valid_init
  assertion: "page->used <= page->capacity"
mimalloc: assertion failed: at "external/mimalloc/src/page.c":100, mi_page_is_valid_init
  assertion: "page->used + free_count == page->capacity"
mimalloc: assertion failed: at "external/mimalloc/src/page.c":74, mi_page_is_valid_init
  assertion: "page->used <= page->capacity"

Add support for Android's alternate TLS layout to fix the crashes in
multi-threaded use cases.

Fixes #376.
 2021-04-07 01:59:47 -07:00
Daan Leijen
985f313b35 bump version to 1.7.1 2021-04-06 10:56:26 -07:00
Daan Leijen
71ac98ab08 rename <Windows.h> include to <windows.h> for mingw compatibility (see pr #367) 2021-02-22 13:04:11 -08:00
Daan Leijen
331491e1e8 build fix for Apple M1 (issue #354 and pr #356) 2021-02-02 10:46:30 -08:00
Daan Leijen
a7c33a3b0e fix getting the unique thread id on the Apple M1, see issue #354. 2021-02-01 15:47:22 -08:00
Daan Leijen
92ead2d880 bump version to 1.7.0 2021-01-31 13:51:19 -08:00
Daan Leijen
35c1fc2be9 limit memcpy as rep stosb to windows where the cpu supporst FSRM; add mi_memcpy_aligned for machine-word aligned copy. see issue #201 and pr #253 2021-01-30 14:33:46 -08:00
Daan Leijen
92ec493a5d possible fix for aligment warning (issue #341) 2021-01-29 16:21:50 -08:00
Daan Leijen
0a06884732 ensure memcpy with rep stosb is only used on windows 2021-01-29 16:09:09 -08:00
Daan
9b966c3492
Merge pull request #253 from haneefmubarak/memcpy-rep-movsb-windows-201 
resolve #201 with a platform-selective REP MOVSB implementation
 2021-01-29 16:00:00 -08:00
Daan
b759bcf5c7
Merge pull request #329 from asl/mi_stat_agg 
Unify statistics collection
 2021-01-29 12:52:29 -08:00
Daan Leijen
78ce716e2d add comment on use of tpidrro_el0 on macOS 2021-01-28 17:36:56 -08:00
Daan Leijen
8d4444ef00 remove spurious parenthesis (#350) 2021-01-28 17:36:35 -08:00
Daan
d9ae916a74
Merge pull request #350 from mr-c/patch-1 
add/improve atomic yields for SSE2, ARM*, PowerPC
 2021-01-28 17:29:54 -08:00
Uwe L. Korn
a753084f74 Use APPLE instead of MACH 2021-01-28 11:38:38 +01:00
Michael R. Crusoe
fb66ebea1d
add/improve atomic yields for SSE2, ARM*, PowerPC 2021-01-23 16:45:47 +01:00
Uwe L. Korn
88330cfc9f Use __APPLE__ instead of __MACH__ 2021-01-22 17:06:43 +01:00
Uwe L. Korn
ab3dac04c2 Use tpidrro_el0 for thread local storage in macOS-arm64 
Fixes #343
 2020-12-30 21:49:41 +01:00
Daan
03503ea4e5
Merge pull request #339 from devnexen/spin_impl_upd 
Restricts cpu yield instructions a little.
 2020-12-17 14:06:23 -08:00
Daan
33614cc054
Merge pull request #342 from wsmoses/fix2 
Fix strndup override
 2020-12-17 14:06:01 -08:00
Daan Leijen
bb386025b5 update override on macOS with interpose of malloc_default_zone (issues #313) 2020-12-15 16:03:54 -08:00
William S. Moses
30fc86cca9 Fix strndup override 2020-12-13 01:50:31 -05:00
David Carlier
33a10b4860 Restricts cpu yield instructions a little. 
adding clobber for ARM and preventing older 32 bits chips not supporting this instruction.
 2020-12-12 12:19:04 +00:00
Daan
9d0555c65b
Merge pull request #326 from asl/mi_stat_counter 
Honour MI_STAT in couple more places
 2020-12-10 11:53:30 -08:00
Daan
56a1c852ea
Merge pull request #325 from asl/max-error-counter 
Do not use the same counter for warnings and errors
 2020-12-10 10:23:08 -08:00
Daan
5bbe1c0216
Merge pull request #323 from devnexen/dfbsd_build_fix 
DragonFly support fix (for 5.8.x and forward).
 2020-12-10 10:19:05 -08:00
Daan
8b8011b4f0
Merge pull request #322 from Kokokokoka/x32_patch 
fix for x32 builds
 2020-12-10 10:14:04 -08:00
Anton Korobeynikov
765fc9c0ca Unify statistic collection: 
- For MI_STAT == 0 no allocation stats are collected
  - For MI_STAT == 1 only aggregated values (across normal, large and huge heaps) are collected
  - For MI_STAT == 1 separate per-bin collection for normal heap is done as well
 2020-11-11 11:41:39 +03:00
Anton Korobeynikov
00fb89f771 Rename the field 2020-11-11 11:41:33 +03:00
Anton Korobeynikov
39bcf8a6b0 Honour MI_STAT in couple more places. 2020-11-02 00:14:02 +03:00
Anton Korobeynikov
9c45221243 Do not use the same counter for warnings and errors. 
Warnings happen normally and could be safely ignored in the most cases,
however errors, if enabled, should not be ignored. Currently since warnings
and errors share the same counter we effectively stop showing errors after
16 warnings (which happen all the time).

Use different counters for errors and warnings.
 2020-11-01 23:57:42 +03:00
David Carlier
e6c2fd44fc DragonFly support fix (for 5.8.x and forward). 
The pthread slot approach is somewhat buggy (pretty visible
 with the stress unit test which segfault more or less randomly,
 but the stats never show up).
Using the default approach instead, the test passes eventough
 it s relatively slow (e.g 1.5 sec on FreeBSD vs 4.5 on DragonFly with same
 machine).
 2020-10-22 11:15:37 +01:00