1227a9e5a5
[atomic] Unify typedef
...
Removes volatile from fallback implementation. That was handwavy
anyway.
2018-08-08 22:45:49 -07:00
5cd1888c9d
[atomic] Change Mac atomic int type to int
...
If ever int is not 32bit, we'll fix this...
2018-08-08 22:43:53 -07:00
e1de86eac1
[atomic] Change Windows atomic int type to int
2018-08-08 22:42:33 -07:00
aee2d10b2b
[atomic] Change Solaris atomic int to signed
2018-08-08 22:40:10 -07:00
13f4c137c6
[atomic] Fix Solaris ones to add proper barriers
2018-08-01 14:15:00 -07:00
19dfaa3515
[atomic] Remove volatile from IBM impl signature
2018-08-01 14:15:00 -07:00
2093a3e0cb
[atomic] Oops
2018-08-01 14:15:00 -07:00
19b98348ff
[atomic] Use read-barrier for get()
2018-08-01 14:15:00 -07:00
006d4f031a
[atomic] Some more minor tweaks
2018-08-01 14:15:00 -07:00
dfc86e4b35
[atomic] Fix cast to fallback ptr_get()
2018-08-01 00:22:18 -07:00
ad27562742
[atomic] On IBM, use light-weight sync for everything
...
lwsync() is a full read/write-barrier. That's all we need, never
need sync(). I'm not sure why an isync() was used in fetch_and_add,
but since that's a read-modify-write, I just changed it to have
lwsync() on both sides.
2018-07-31 23:01:05 -07:00
fd638d215f
[atomic] Add XXX items around Solaris ops
...
Since add_int and cas are both read-modify-write, I wonder if we
also need a barrier after them.
2018-07-31 23:00:15 -07:00
896ff15ae6
[atomic] Fix get() impl
...
Originally, glib's atomic_get was implemented as "memory_barrier; load".
I copied this into cairo, fontconfig, and harfbuzz. However, that's
wrong. Correct way is "load; memory_barrier". The details are long
and hard to fully grasp. Best to read:
https://www.kernel.org/doc/Documentation/memory-barriers.txt
Also see my report against GNOME:
https://gitlab.gnome.org/GNOME/glib/issues/1449
Note that this is irrelevant if C++11-like atomic ops are available.
2018-07-31 22:51:38 -07:00
63c74e8d1d
[atomic] Fix fallback impl
2018-07-31 22:21:21 -07:00
4bc16aca47
[atomic] Add get_relaxed / set_relaxed
...
To help TSan and be more "correct".
2018-07-31 21:29:18 -07:00
3dd1b88765
[atomic] Use CONSUME, not ACQUIRE, memory-order for get()
...
Although, all implementations just elevate that to ACQUIRE.
But requirement for us is just CONSUME.
2018-07-31 20:28:36 -07:00
06b91d935d
Revert "[atomic] Make pointer get op relaxed instead of acquire"
...
This reverts commit b1e5650c67https://www.kernel.org/doc/Documentation/memory-barriers.txt
I've convinced myself that we need consume memory-ordering on get().
The location of memory-barrier in a load should be after, not before
the load. That needs fixing. I'll do that separately.
2018-07-31 20:28:36 -07:00
b1e5650c67
[atomic] Make pointer get op relaxed instead of acquire
...
We only use it before cmpexch, so relaxed is fine and faster for
common case.
2018-07-31 01:02:31 -07:00
04b7b81bcb
Reland "Implement C++11-style GCC builtin atomic ops"
...
Fixed the crasher in it.
2018-07-17 11:02:38 +02:00
3edef5a19b
Revert "Implement C++11-style GCC builtin atomic ops"
...
This reverts commit 21fa170f0b
2018-07-17 10:50:01 +02:00
21fa170f0b
Implement C++11-style GCC builtin atomic ops
...
Hopefully fixes bots.
2018-07-16 17:58:02 +02:00
804b59cf49
Relax C++11 atomic ops memory order to acquire-release
2018-07-16 15:45:22 +02:00
bda242409f
Implement C++11 <atomic> operations
...
Fixes https://github.com/harfbuzz/harfbuzz/issues/345
2018-07-16 15:45:22 +02:00
e36cd1dfd3
Remove trivial HB_ATOMIC_INT_INIT()
2018-05-31 19:31:39 -07:00
8a0952db7e
atomics: Favor compiler primitives over macOS APIs. ( #676 )
...
In macOS 10.12, the `OSMemoryBarrier` and related APIs were deprecated
in favor of using `std::atomic`. On the way to supporting `std::atomic`,
we can favor using the "Intel primitives" which are also available on
macOS.
2018-01-04 15:10:10 +03:30
42d518513c
Towards compiling with pre-C++11 compilers and nullptr fallback
...
https://github.com/behdad/harfbuzz/issues/585
2017-10-27 00:25:11 -06:00
1817221620
Minor
2017-05-16 14:26:28 -07:00
49e72634af
Limit use of AIX intrinsics to IBM's compiler
2015-12-10 17:44:19 +01:00
70b33edae7
Add atomic ops for AIX
...
Patch from Volker Simonis.
2015-12-10 15:54:42 +01:00
9c974360fe
Minor rename
2015-04-09 12:04:14 -07:00
2958f2c147
Fixup
2015-04-08 16:26:16 -07:00
3f174cd020
Minor refactoring to the atomics implementation
...
s/atomic_int/atomic_int_impl/ and s/atomic_ptr/atomic_ptr_impl/
to bring it in par with hb_mutex_impl_t, then re-introduce
hb_atomic_int_t as a wrapper around hb_atomic_int_impl_t.
In hb_reference_count_t, make it clear the non-atomic get and set
are intentional due to nature of the cases they are used in
(comparison to -1 and the debug output/tracing).
2015-04-08 13:04:22 -07:00
45a8b46f47
Allow implementing atomic and mutex ops in config
...
Motivated by
https://github.com/behdad/harfbuzz/pull/92
2015-04-08 12:49:38 -07:00
db30828048
[win] Consolidate windows.h include tips and tricks
2014-07-19 16:32:04 -04:00
f26d59d468
More fixing MemoryBarrier() on Mingw32
...
Set requested windows header to Vista. See discussion:
fbb2847f54 (commitcomment-7054700)
2014-07-19 16:10:21 -04:00
05870ed62e
Use __aarch64__ for 64-bit ARM detection, not __arm64__
...
Many GCC versions don't define __arm64__
2014-04-02 12:41:11 -07:00
fbb2847f54
Improve MemoryBarrier() implementation
...
See thread "[HarfBuzz] compilation error of 0.9.26 with MinGW"
started by Werner.
2014-02-05 08:22:06 -05:00
0bb31e4497
Bug 71845 - Use 64-bit cmpexch on ARM64 iOS
2013-11-20 14:21:07 -05:00
1f97060985
Put back MemoryBarrier fallback implementation on MINGW32
...
This almost reverts 2761e8a632
2013-04-04 15:02:03 -04:00
2761e8a632
[win32] Remove MemoryBarrier() fallback implementation
...
I added these because the older mingw32 toolchain didn't have
MemoryBarrier(). The newer mingw-w64 toolchain however has.
As reported by John Emmas this was causing build failure with
MSVC (on glib) because of inline issues. But that reminded me
that we may be taking this path even if the system implements
MemoryBarrier as a function, which is a waste. So, just remove
it.
2013-03-07 20:51:30 -05:00
a6c1e040e5
Improve check for Windows platforms
...
Instead of checking for compiler, check for platform.
2013-02-12 15:31:58 -05:00
b842780138
Minor
2013-02-11 17:02:17 -05:00
de649f07f1
Fix residuals from fontconfig changes
2013-01-14 00:26:43 -06:00
2dcb333f52
Add atomic ops for Solaris
...
Based on fontconfig patch from Raimund Steger.
2013-01-10 01:18:10 -06:00
11d2956553
Minor
2013-01-02 17:41:27 -06:00
0e9f0f3e5f
Fix atomic ops on iOS
...
Patch from John Ralls.
2012-12-10 15:25:21 -05:00
071d5b831e
Work around missing OSAtomicCompareAndSwapPtrBarrier() on OS X 10.4
...
Not sure how to handle iOS.
2012-12-10 00:57:00 -05:00
0e292eb2a2
Remove Glib thread-safety support
...
Now that we have pthread detection in configure, we don't need Glib
anymore. Glib will only be a Unicode data provider.
2012-10-02 15:09:38 -04:00
52ff2681d8
Use VisualStudio-style atomic intrinsics on mingw32
2012-08-28 18:03:35 -04:00
fa2bd9fb63
Further simplify atomic ops on Visual Studio
2012-07-14 12:15:54 -04:00
Behdad Esfahbod
Bruce Mitchener
Konstantin Ritt
Primiano Tucci