system/core/include is included in the global include path using
-isystem, which hides all warnings. cutils/trace.h has an ignored
return value warning, move the implementation to trace-dev.c so it
doesn't cause warnings in every module that includes it in preparation
for moving from -isystem to -I.
Test: m -j native
Bug: 31492149
Change-Id: If8b3fe13059c9e59c2d5208294d427d84fa6e588
system/core/include is included in the global include path using
-isystem, which hides all warnings. Fix warnings in cutils/atomic.h in
preparation for moving from -isystem to -I.
Test: m -j native
Bug: 31492149
Change-Id: I2a212f192309090a16d507e9ab6acf0fd7cac02d
I'll come back and remove klog_init when I've removed other calls to it.
Change-Id: Iad7fd26d853b4ddc54e9abd44516b6f138cbbfcb
Test: booted N9, looked at "adb shell dmesg" output.
clang doesn't do the frontend inlining/dead code elimination needed for
the fortified implementation to work. Therefore, turn it off.
Bug: http://b/28381737
Change-Id: Ie8dd970e3908b1daaa587ad2bd041e8f6e2089db
am: 6a29fe931d
* commit '6a29fe931d9fd3bf7f2aad3713dc70c080970763':
Add utility to prepare files in a similar way to directories
Change-Id: I288024d55e8cead1c902950938b03bfa8dcc3df3
Add comment that SharedBuffer is deprecated.
Both aref and SharedBuffer had memory ordering bugs. Aref has no
clients.
SharedBuffer had several bugs, which are fixed here:
mRefs was declared neither volatile, not atomic, allowing the
compiler to, for example, reuse a stale previously loaded value.
It used the default android_atomic release memory ordering, which
is insufficient for reference count decrements.
It used an ordinary memory read in onlyOwner() to check whether
an object is safe to deallocate, without any attempt to ensure
memory ordering.
Comments claimed that SharedBuffer was exactly 16 bytes, but
this was neither checked, nor correct on 64-bit platforms.
This turns mRef into a std::atomic and removes the android_atomic
dependency.
Bug: 28826227
Change-Id: I39fa0b4f70ac0471b14ad274806fc4e0c0802e78
(cherry picked from commit 3e4c076ef2)
Add comment that SharedBuffer is deprecated.
Both aref and SharedBuffer had memory ordering bugs. Aref has no
clients.
SharedBuffer had several bugs, which are fixed here:
mRefs was declared neither volatile, not atomic, allowing the
compiler to, for example, reuse a stale previously loaded value.
It used the default android_atomic release memory ordering, which
is insufficient for reference count decrements.
It used an ordinary memory read in onlyOwner() to check whether
an object is safe to deallocate, without any attempt to ensure
memory ordering.
Comments claimed that SharedBuffer was exactly 16 bytes, but
this was neither checked, nor correct on 64-bit platforms.
This turns mRef into a std::atomic and removes the android_atomic
dependency.
Bug: 28826227
Change-Id: I39fa0b4f70ac0471b14ad274806fc4e0c0802e78
Current profiles (the ones which have not been used for
compilation) are stored in /data/misc/profiles/cur/0/pkgname/.
Reference profiles (the merged of all user profiles, used for
compilation) are stored in /data/misc/profiles/ref/pkgname/.
Add a method to get the shared app gid from an uid or appid.
Bug: 26719109
Bug: 26563023
Change-Id: I89601d7dbeb3041df882c141a9127dac200a645e
Fixes libcutils multi-buffer write interface to be more friendly and
hooks into it from the fastboot Socket class.
Bug: http://b/26558551
Change-Id: Ibb3a8428fc379755602de52722c1260f9e345bc0
Unix and Windows both have functions to write multiple buffers to a
socket with a single call but they have very different signatures. This
CL creates some cross-platform functions to be able to perform these
operations in a uniform way, which will be required for upcoming
fastboot functionality.
This CL also fixes some inconsistent spacing in the touched files.
Bug: http://b/26558551
Change-Id: I8f14d52d3a1de1f3b464267666d6cd3b54263238
Tests that require a local server currently hardcode a test value,
which can run into conflicts depending on what's currently running on
the machine.
This CL adds socket_get_local_port(), which lets us pass 0 so the
system picks an open port and we can query which port it chose.
Bug: http://b/26236380
Change-Id: I01d1558884e7636081d9a357db6faa86929934f6
Allows ActivityManager to use the top-app cpuset to grant the currently
focused app exclusive access to a CPU core.
Change-Id: I45bca5170477e413dec6e5889338399d0859706c
Now that cutils has cross-platform socket functionality, we can
restructure fastboot to remove platform-dependent networking code.
This CL adds socket_set_receive_timeout() to libcutils and combines the
fastboot socket code into a single implementation. It also adds TCP
functionality to fastboot sockets, but nothing uses it yet except for
the unit tests. A future CL will add the TCP protocol which will use
this TCP socket implementation.
Bug: http://b/26558551
Change-Id: If613fb348f9332b31fa2c88d67fb1e839923768a
This CL moves Windows networking code from fastboot to libcutils so
that it can be shared with other host programs such as adb.
Not all libcutils networking functions have been implemented for
Windows, just those necessary for fastboot. In the next CL I will do
the same for adb, adding any additional required functions.
Unit tests have also been added to test the functions using a loopback
connection.
Bug: http://b/26236380.
Change-Id: Ibc51a67030fe69a04c23512eefa9d19b055c7c12
Certain apps decide that they want to chmod() their private data
directories to gain more security. We still want to carefully
enforce owner UID/GID, but relax the mode check for now.
Bug: 26549892
Change-Id: I362d530ba0b20fb23f427ac082ee003864adc57d
Some callers, such as installd, have stricter requirements around
directory preparation, where they want to assert ownership and mode
without quietly fixing the values.
Bug: 26466827
Change-Id: Id44db5f29a3326cfe178b443fb450ad2edeaefd8
Hash functions rely on overflow behavior, so whitelist them.
ATRACE_TAG_NOT_READY: use an unsigned constant when shifting bits.
Otherwise, the value overflows on shift. The users of this constant
assign it to a uint64_t variable.
Change-Id: I21c437ce2083525e906c3ead3259ec34a1ef4b66
and run fsck with -f on clean shutdown instead.
With -f, fsck.f2fs always performs a full scan of the /data
partition regardless of whether the partition is clean or not.
The full scan takes more than 2 seconds on volantis-userdebug
and delays the OS boot.
With -a, the command does almost nothing when the partition
is clean and finishes within 20-30ms on volantis-userdebug.
When the partition has an error or its check point has
CP_FSCK_FLAG (aka "need_fsck"), the command does exactly the
same full scan as -f to fix it.
Bug: 21853106
Change-Id: I126263caf34c0f5bb8f5e6794454d4e72526ce38
Colin Cross