Added new atomic functions, renamed some old ones. Some #defines have
been added for backward compatibility.
Merged the pre- and post-ARMv6 implementations into a single file.
Renamed the semi-private __android_membar_full_smp to USE_SCREAMING_CAPS
since that's more appropriate for a macro.
Added lots of comments.
Note Mac OS X primitives have not been tested.
Change-Id: If827260750aeb61ad5c2b760e30658e29dbb26f2
Added atomic-inline.h. Added a platform-specific memory barrier call
there.
Added android_atomic_acquire_cmpxchg() and android_atomic_release_store().
Not tested on Mac OS X or SH.
Added memory barrier calls to linux-x86 atomics. Mac OS X has barrier
functions already. sh isn't really SMP-ready. linux-arm needs work
(to be done in a separate change).
Updated the makefile to make the SMP state visible to the code here.
Note that host binaries are NOT built with SMP enabled; while our hosts
are very likely SMP, it's not worth figuring out e.g. whether it's okay
to use the SSE2 mfence instruction or have to use something else. We
haven't had barriers enabled in host tools before, so there's probably
no need to stat now.
Removed quasiatomic 64-bit calls (now part of Dalvik).
Change-Id: I49e5e6c8abe70f304cdedb9d7b8e6e65f8925815
This routine allows creating a contiguous mspace from raw mapped memory.
In turn, this will enable preallocation of the 3 heap spaces, which will help
remembered sets and zygote/app checks given pointer values.
POSIX seems to have chosen open_memstream() over the BSD variant. We
want something for Dalvik that will work on both GNU/Linux and Android,
so this is open_memstream() implemented in terms of BSD funopen().
For Windows there's just a stub that calls abort().
I'm putting this in libcutils since it seems inappropriate for bionic
(which provides the BSD alternatives) but isn't Dalvik-specific.
Merge commit '38b2ddc4a7cf1c47397af118a6d466d45f59da04'
* commit '38b2ddc4a7cf1c47397af118a6d466d45f59da04':
cutils: Add support for reading a processes scheduler policy
This is needed by the MemoryFile changes in
https://android-git.corp.google.com/g/2714
where it is used to find out whether a file descriptor
refers to an ashmem region.
Merge commit 'ebefc48e61a7d5cf2a3228e6c8729feeeb42d1b4'
* commit 'ebefc48e61a7d5cf2a3228e6c8729feeeb42d1b4':
Expose the stable/frozen log-related definitions of <cutils/logd.h> into a new header: <android/log.h>
Merge commit 'fee77ec093f78c1bb0ce85aa16d7ee8e8fa06f8a'
* commit 'fee77ec093f78c1bb0ce85aa16d7ee8e8fa06f8a':
don't use cdefs.h as it breaks the windows build.
asocket_connect()
asocket_accept()
asocket_read()
asocket_write()
These calls are similar to the regular syscalls, but can be aborted with:
asocket_abort()
Calling close() on a regular POSIX socket does not abort blocked syscalls on
that socket in other threads.
After calling asocket_abort() the socket cannot be reused.
Call asocket_destory() *after* all threads have finished with the socket to
finish closing the socket and free the asocket structure.
The helper is implemented by setting the socket non-blocking to initiate
syscalls connect(), accept(), read(), write(), then using a blocking poll()
on both the primary socket and a local pipe. This makes the poll() abortable
by writing a byte to the local pipe in asocket_abort().
asocket_create() sets the fd to non-blocking mode. It must not be changed to
blocking mode.
Using asocket will triple the number of file descriptors required per
socket, due to the local pipe. It may be possible to use a global pipe per
process rather than per socket, but we have not been able to come up with a
race-free implementation yet.
All functions except asocket_init() and asocket_destroy() are thread safe.
Merge commit '414ff7d98ac8d7610a26206335954ad15f43f3ac'
* commit '414ff7d98ac8d7610a26206335954ad15f43f3ac':
Move fdevent from libcutils into adb directory. ADB is the only client of this API, and I intend to modify it extensively to clean its codebase soon.
Merge commit '722a5c0462f38827f4097065bfc3826b9e0e9fb4'
* commit '722a5c0462f38827f4097065bfc3826b9e0e9fb4':
Add support for "standalone months" to tztime's strftime().
The idea here is that some languages need a different form of the month
name in constructions like "January 2" than in "January 2009", since the
one in the "January 2" case really means "of January." So with this change,
a format string of "%-B" will use the standalone month, while "%B" will
continue to use the format month.
Andy McFadden