* changes:
adb: add lock to remove_socket.
Revert "adb: fix deadlock between transport_unref and usb_close."
adb: allow selection of a specific transport.
adb: fix deadlock in kick_all_transports.
The comment that was previously here says that local_socket_list_lock
must be taken, but this function is exposed to external callers that
can't possibly take the lock.
Bug: http://b/65419665
Bug: 64709603 (presubmit balking at the line above)
Test: python test_device.py
Change-Id: I12d464933936b2a210a827ccf19ea201020d8d78
(cherry picked from commit 62c92f0c05)
This reverts commit 7e197ef833.
The mutex lock in transport_unref hides a race that seems otherwise
hard to fix. Specifically, there's no synchronization between acquiring
a transport and attaching it to an asocket*, leading to badness if the
transport is closed in between the two operations.
Fix the original problem the reverted patch addressed by manually
unlocking before calling unregister_usb_transport.
Bug: http://b/65419665
Bug: 64709603 (presubmit balking at the line above)
Test: python test_device.py
Change-Id: I0ed0044129b1671b2c5dd1b9fa2e70a9b4475dc5
(cherry picked from commit e48ecce671)
Extend device selection to allow selecting a specific transport via
monotonically increasing identifier (visible in devices -l).
This is useful when using multiple devices (like hikey960...) that
have identical bogus serial numbers like 0123456789ABCDEF.
Bug: http://b/37043226
Bug: 65419665
Test: adb -t {1, 2, 9999999} {get-serialno, shell, features}
Change-Id: I55e5dc5a406a4eeee0012e39b52e8cd232e608a6
(cherry picked from commit b122b17555)
Previously, kick_all_transports would deadlock if there were any
inaccessible transports, because the transport kick function would call
unregister_usb_transport, which attempts to take the already-held
transport lock.
Fix this by switching the transport lock over to a recursive mutex.
Bug: 65419665
Test: manual
Change-Id: If61296ff4745e1699f3e216811c1383582627604
(cherry picked from commit 1db71afe97)
Tombstones (especially ones with lots of VMAs) are regularly truncated.
We can at least show the number of VMAs, though, for anyone interested
in knowing whether they got close to the default 64Ki limit.
Bug: http://b/66911122
Bug: 64709603 (presubmit balking at the line above)
Test: ran crasher, examined tombstone
(cherry picked from commit 868d39a82f)
Change-Id: I286db66f28f132307d573dbe5164efc969dc6ddc
(cherry picked from commit c3ad75be84)
bootstat does not need root uid and root gid permissions to perform
its tasks. It appears that system uid and log gid are adequate and
appropriate.
Test: manual
Bug: 63736262
Bug: 65843095
Change-Id: I094c2cb054e441562fa8717a4d3dc0086fb70a7a
Pstore test cases should not belong to CTS.
The test cases should be a part of GTS if really required.
Test: manual
Bug: 66431569
Bug: 63969981
Bug: 63913130
Signed-off-by: Srinavasa Nagaraju <Srinavasa.Nagaraju@sony.com>
Signed-off-by: Yoshitaka Seto <yoshitaka.seto@sony.com>
Signed-off-by: Mark Salyzyn <salyzyn@google.com>
Tested-by: Srinavasa Nagaraju <Srinavasa.Nagaraju@sony.com>
Change-Id: Icd326403711ad268ce28a3863045fc9b2f211461
Get min adj score when needed, instead to propagate from caller funct.
Bug:65642829
Test: tested on gobo_512
Change-Id: I87f15c5c6206d471002828a24d0462b0abb0d196
Builtin commands may set the sys.powerctl property, which causes
reboot to be immediately processed. Unfortunately, part of the reboot
processing involves clearing the action queue, so when this scenario
happens, ActionManager::ExecuteOneCommand() can abort due to its state
being unexpectedly changed.
Longer term, the real fix here is to split init and property service.
In this case, the property sets will be sent to property service and
the reboot will only be processed once property service responds back
to init that the property has been set. Since that will not happen
within the action queue, there will be no risk of failure.
Short term, this change sets a flag in init to shutdown the device
before the next action is run, which defers the shutdown enough to fix
the crash, but continues to prevent any further commands from running.
Bug: 65374456
Test: force bullhead into the repro case and observe that it no longer
repros
Merged-In: I89c73dad8d7912a845d694b095cab061b8dcc05e
Change-Id: I89c73dad8d7912a845d694b095cab061b8dcc05e
(cherry picked from commit 3633a4014a)
If the system has enough memory the pressure events are ignored or
downgraded from critical to medium.
Bug: 65642829
Test: tested on gobo
Change-Id: I44e66d2e35508aceee5c1719313db217b80d582e
ExpandArgs() was factored out of Service::Start() to clean up init,
however this introduced a bug: the scope of expanded_args ends when
ExpandArgs() returns, yet pointers to the c strings contained within
those std::strings are returned from the function. These pointers are
invalid and have been seen to cause failures on real devices.
This change moves the execv() into ExpandArgs() and renames it
ExpandArgsAndExecv() to keep the clean separation of Service::Start()
but fix the variable scope issue.
Bug: 65303004
Test: boot fugu
Change-Id: I612128631f5b58d040bffcbc2220593ad16cd450
(cherry picked from commit 5e405cacb1)
Child processes inherit the signal handlers from their parent process.
In the case of init, fork()'ed processes, will attempt to reboot the
system if they receive a fatal signal). This is not the correct behavior;
these processes should terminate due to the provided signal like other
processes on the system.
This is particularly important as there are multiple LOG(FATAL) calls
in service.cpp for failures after fork() but before execv() when a
service is started.
Note, that pthread_atfork() is not a viable solution since clone() is
used in some cases instead of fork() and atfork handlers are not
called with clone().
Bug: 65637054
Test: LOG(FATAL) from a child process of init and see that it
terminates due to a signal correctly
Test: LOG(FATAL) from init proper and see that it reboots to the
bootloader
Change-Id: I875ebd7a5f6b3f5e3e2c028af3306917c4409db3
For Go devices fallback to memcg/memory pressure events for LMK.
Go devices will use memcg pressure events while inkernel module
is disabled.
Bug: 64852905
Test: tested on gobo
Change-Id: I267ab00be85e324331f6c91551ba013184de817e
Without an explicit check, the return value can wrap around and return
a value that is far too small to hold the data from the resulting
conversion.
No CTS test is provided because it would need to allocate at least
SSIZE_MAX / 2 bytes of UTF-16 data, which is unreasonable on 64-bit
devices.
Bug: 37723026
Test: run cts -p android.security
Change-Id: Ie2606b92b9eab1acfe8ce4663b43b81156a4cad7
Merged-In: I56ba5e31657633b7f33685dd8839d4b3b998e586
Video threads have soft realtime requirements that
must be met in order to maintain reliable frame delivery
even when the system is under high load. This CL
defines a new priority that can be used by video
threads, enabling them to be scheduled appropriately
relative to other system threads.
Change-Id: Idd57207e30309dfdff24389db0acf107532f9e5a
related-to-bug:63898080
Turns on the derive_gid feature for sdcardfs. This was moved
under a mount flag in the kernel.
Test: If the derive_gid flag is supported, the first mount
should succeed. If the flag is not, the second should
succeed.
Bug: 63245673
Change-Id: If1c1bce13d14120732e420252cb5605d33ce7c40
For devices where VNDK restrictions are all enforced, vendor apks are
recognized as unbundled; since system partition and vendor partition can
be updated independently from each other.
However, since vendor apks are still bundled in the vendor partition,
they are allowed to do more than ordinaly unbundled apks that are
downloaded and installed to the data partition.
1) /vendor/lib is allowed. So the path is added to the search_paths and
permitted paths of the classloader namespace.
2) LLNDK libs are allowed in addition to the NDK libs. So, LLNDK lib list
from llndk.libraries.txt is added to the list from public.libraries.txt.
3) VNDK-SP libs are allowed. To do so, the classloader namespace is
linked to the 'vndk' namespace where VNDK-SP libs are searched and
loaded from. The list of available VNDK-SP libs is read from
vndksp.libraries.txt file.
4) Name of the namespace is changed to 'vendor-classloader-namespace'
since the namespace is configured differently from the ordinary
'classloader-namespace'.
Bug: 63553457
Test: 2017 pixel devices build and boots to the UI
Test: a vendor apk (e.g. TimeService.apk) works. Turn the airplain mode on.
Set time. Reboot the device. The time is not reset.
Test: 1) set target as 2017 pixel
2) m -j CtsVendorJniTestCases
3) copy the built apk into /vendor/app/CtsVendorJniTestCases
4) reboot / factory reset
5) adb shell am instrument -w android.jni.vendor.cts
Change-Id: I447452eb025c0a0fd076b5c9ac081d453dc6074e
Without an explicit check, the return value can wrap around and return
a value that is far too small to hold the data from the resulting
conversion.
No CTS test is provided because it would need to allocate at least
SSIZE_MAX / 2 bytes of UTF-16 data, which is unreasonable on 64-bit
devices.
Bug: 37723026
Test: run cts -p android.security
Change-Id: I56ba5e31657633b7f33685dd8839d4b3b998e586
Ting-Yuan Huang