system-corennnn

Fork

Android 6.0.1 release 74
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1

iEYEABECAAYFAlgczu4ACgkQ6K0/gZqxDnitCgCePSzI1XB1U5uYTh/Zi64YsrlT
N+8An0LYcv78UQ/41XTrnjQDUrHXtObc
=d0tS
-----END PGP SIGNATURE-----

Merge tag 'android-6.0.1_r74' into marshmallow-x86

Android 6.0.1 release 74

ueventd: defer modules loading if explicitly listed

The patch extends the syntax of /etc/modules.blacklist.
The modules marked as deferred in this file will be loaded
after all other modules are loaded at coldboot stage.

liblog: add __android_log_close()

Bug: 30963384

(cherry picked from commit df7a4c6bae5f85532d79a93b7d9197a2aab17825)

Change-Id: Ide70df3c04e29301649a1ca234b1b0af687bcbfb
(cherry picked from commit b1b5d507cba4b37221016d308dbf9fdd39019108)

liblog: add __android_log_close()

Bug: 30963384

(cherry picked from commit df7a4c6bae5f85532d79a93b7d9197a2aab17825)

Change-Id: I2255486e84dd55af0f4e7fbbfb616c2deb1765d0
(cherry picked from commit 2d149691552676968b7de337f543463b920578b0)

init: set default value of ro.hardware to the target name

People usually forget to set the value of androidboot.hardware to
the target name in kernel cmdline which results in boot failure.

The change tries to reduce such an error in the future. The target
name is hardcoded to the image. This eliminates the necessity of
androidboot.hardware.

Android 6.0.1 release 66
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1

iEYEABECAAYFAlfO7mkACgkQ6K0/gZqxDniqrQCdFR+6OEohyLsDnMK3AuXJaQA+
rtQAniY0clzoDQSIpoOItDPjsmyGKvBN
=ocxE
-----END PGP SIGNATURE-----

Merge tag 'android-6.0.1_r66' into marshmallow-x86

Android 6.0.1 release 66

Revert "adb: sockets: fix race condition between input_thread and fdevent callback"

This reverts commit 3e872c4ce7d83fee2673694644559f6fa9bdcbe0.

Fix vold vulnerability in FrameworkListener

Modify FrameworkListener to ignore commands that exceed the maximum
buffer length and send an error message.

Bug: 29831647
Change-Id: I9e57d1648d55af2ca0191bb47868e375ecc26950
Signed-off-by: Connor O'Brien <connoro@google.com>
(cherry picked from commit baa126dc158a40bc83c17c6d428c760e5b93fb1a)
(cherry picked from commit 470484d2a25ad432190a01d1c763b4b36db33c7e)

ueventd: load all modules with the matched alias

Originally we only load the first module with the matched alias.
It causes some necessary modules are not loaded.

init: enlarge the time to wait coldboot done

We need more time to probe and load all modules.
If 10s is still not enough for you, let me know.

debuggerd: fix missed use of ptrace(PTRACE_ATTACH).

Bug: http://b/29555636
Change-Id: Ibd8a2e2b619b74aac667555b7085d6f28e367c07

Android 6.0.1 Release 61 (MOB30Z)
-----BEGIN PGP SIGNATURE-----

iEYEABECAAYFAlefxDQACgkQ6K0/gZqxDnip2ACfVQ2tY6lcGUlT0bwGUagtYIEy
ii4AniKmljbAlyO0M5eTydKej87c21Ea
=S5OJ
-----END PGP SIGNATURE-----

Merge tag 'android-6.0.1_r61' into marshmallow-x86

Android 6.0.1 Release 61 (MOB30Z)

ueventd: refactory module and firmware loading mechanism

Currently each module and firmware loading event is forked to
a child process to handle. It may cause some race conditions.

Now we fork ueventd into two processes. The parent process handles
normal uevents including modules loading, while the child process
only handles firmware loading events. All events are handled
sequentially.

init.rc: symlink /lib to /system/lib

Currently more linux drivers use request_firmware_nowait() to load
firmwares. It requires the firmwares exist in the standard path.

Create the symlink before ueventd started to ensure firmwares
could be loaded correctly.

init: Fix load_firmware error

In function 'load_firmware', parameter 'buf' passed to 'write'
is wrong. To fix it, use android::base::WriteFully to replace.

Change-Id: I13f79bdc9be9e5eb669f6bd975535b1dce965ef0
Signed-off-by: Biao Lu <biao.lu@intel.com>

DO NOT MERGE: debuggerd: verify that traced threads belong to the right process.

Fix two races in debuggerd's PTRACE_ATTACH logic:
1. The target thread in a crash dump request could exit between the
/proc/<pid>/task/<tid> check and the PTRACE_ATTACH.
2. Sibling threads could exit between listing /proc/<pid>/task and the
PTRACE_ATTACH.

Backport of NYC change I4dfe1ea30e2c211d2389321bd66e3684dd757591
Bug: http://b/29555636
Change-Id: I6c6efcf82a49bca140d761b2d1de04215ba4d252

adb: use asocket's close function when closing.

close_all_sockets was assuming that all registered local sockets used
local_socket_close as their close function. However, this is not true
for JDWP sockets.

Bug: http://b/28347842
Change-Id: I40a1174845cd33f15f30ce70828a7081cd5a087e
(cherry picked from commit 53eb31d87cb84a4212f4850bf745646e1fb12814)
(cherry picked from commit 014b01706cc64dc9c2ad94a96f62e07c058d0b5d)

adb: switch the socket list mutex to a recursive_mutex.

sockets.cpp was branching on whether a socket close function was
local_socket_close in order to avoid a potential deadlock if the socket
list lock was held while closing a peer socket.

Bug: http://b/28347842
Change-Id: I5e56f17fa54275284787f0f1dc150d1960256ab3
(functionally a cherrypick of 903b749f + 9b587dec, with windows disabled)

libutils/Unicode.cpp: Correct length computation and add checks for utf16->utf8

Inconsistent behaviour between utf16_to_utf8 and utf16_to_utf8_length
is causing a heap overflow.

Correcting the length computation and adding bound checks to the
conversion functions.

Test: ran libutils_tests
Bug: 29250543
Change-Id: I6115e3357141ed245c63c6eb25fc0fd0a9a7a2bb
(cherry picked from commit c4966a363e46d2e1074d1a365e232af0dcedd6a1)

Merge branch 'android-ia' into marshmallow-x86

Fix adb protocol issue when missed a packet

When there is an issue, adb re-initializes the connection and if it
misses the first 24 bytes "CNXN" packet, it will read the 8 bytes "host::"
packet instead followed by the first 17 bytes of the "CNXN packet" leading
to a new connection error; it reinitializes then the connection but
reads again the "host::" packet sent by the host etc and infinitely.

The device is then show offline or not shown at all on the host.

Change-Id: I4720c9494c7bede79becd5e916de50ace2115021
Tracked-On: https://jira01.devtools.intel.com/browse/OAM-19456
Signed-off-by: Jean-Christophe PINCE <jean-christophe.pince@intel.com>
Signed-off-by: Clement Calmels <clement.calmels@intel.com>
Reviewed-on: https://android.intel.com:443/468687

Create mixins configuration for app compatibility
enhancement feature

This feature is created for some cerner cases of
app compatibility issue. It's disabled by default.
To enable it, please add below lines to mixins.spec.
Currently, it aims to handle below issues:

1. All native libraries are put under assets directory.
Some apps put their libraries under the assets directory
in their APKs, instead of lib directory. It's not a
problem if the ABI of app libs is matched with platform
default ABI. But for apps with other ABIs, this feature
must be enabled to call native bridge for help.

Change-Id: I98ec4d47bee4596bfe0e9efd4cd0b6fe0c90fdf7
Tracked-On: https://jira01.devtools.intel.com/browse/OAM-24196
Signed-off-by: jgu21 <jinghui.gu@intel.com>
Reviewed-on: https://android.intel.com:443/481103

fs_mgr: fix encryptable=footer support

If the userdata partition has the encryptable=footer fstab option,
fs_mgr must leave room for the crypt footer.

Upstream: https://android-review.googlesource.com/206992
Change-Id: Id07818c5d93aafc27577f72fb0a780f26db51b16
Signed-off-by: Chuanxiao Dong <chuanxiao.dong@intel.com>
Tracked-On: https://jira01.devtools.intel.com/browse/OAM-23540
Reviewed-on: https://android.intel.com:443/479471

healthd: only kick animation once for charger unplug case

After animation done for charger unplug case, healthd may still
receive some uevent before the shutdown action. This change will
make sure battery animation will not be kicked for multiple times

Change-Id: I1e419b10b195d6919b5d5a83c87c334306896c2c
Signed-off-by: yangx.d.liu <yangx.d.liu@intel.com>
Signed-off-by: Wu, Hao <hao.wu@intel.com>
Tracked-On: https://jira01.devtools.intel.com/browse/OAM-21840
Reviewed-on: https://android.intel.com:443/477709

adb: fix klocwork isssue 'use of free memory possible'

Change-Id: I334076094c8bf9ab6c90bd170e141e190058ed72
Tracked-On: https://jira01.devtools.intel.com/browse/OAM-18056
Signed-off-by: Du, Changbin <changbin.du@intel.com>
Reviewed-on: https://android.intel.com:443/463913

adb: sockets: fix race condition between input_thread and fdevent callback

The transport's input_thread clean up all of its sockets by
close_all_sockets function before it exit. All sockets's callback is
invoked by local_socket_event_func by fdevent thread(the main thread).

But there is no synchronization between these two thread. Missing
synchronization can cause crash or unexpected behaviour. So far has
observed two crash issue:
1. after close_all_sockets return, still has sockets pending. Then when
the pending sockets try to access the removed transport cause segment
fault.
2. calling the socket's close callback(local_socket_close) but a socket
has been destroied. This will access a freed memory.

Change-Id: I9787b9ca7949946e977e102812518799051b9c67
Tracked-On: https://jira01.devtools.intel.com/browse/OAM-12574
Signed-off-by: Du, Changbin <changbin.du@intel.com>
Reviewed-on: https://android.intel.com:443/461670

adb: fix a spurious fatal condition

The fde may be removed just after rutern from select in fdevent_process
fucntion. This is not a fatal error but normal. So just igore the event
for this fde.

Change-Id: I08fca0c6e539755fbc15c82d46c6b3f3855a2f41
Tracked-On: https://jira01.devtools.intel.com/browse/OAM-12574
Signed-off-by: Du, Changbin <changbin.du@intel.com>
Reviewed-on: https://android.intel.com:443/460314

adb: fdevent: add synchronization logic to avoid crash

The fdevent part didn't implemente any synchronization protection. This
could introduce race condition issues. Espacially when adb is used for
Android stress test, this will be a problem that can cause server side
crash.

Change-Id: I93c311ed62770afc3061ccb8b60a0abd09f7453d
Tracked-On: https://jira01.devtools.intel.com/browse/OAM-12574
Signed-off-by: Du, Changbin <changbin.du@intel.com>
Reviewed-on: https://android.intel.com:443/460313

adb: fix 'adb wait-for-device' thread leak issue

After the "adb wait-for-device" command is killed, the thread created in
adb server doesn't exit and keep it there to monitor device. If keep
doing this(like a automation test tool), hundreds of threads will be
crated and make adb server crash with running out of resource.

Steps to reproduce the problem.
- run "adb -s xxxxx wait-for-device"
- use "pstree -p <pid-of-adb-server>" to check count of threads
- kill it by ctrl+C.
- use pstree check threads info.

Change-Id: I96599c7b43df1f5c92edd7065d562a1bb6bacb19
Tracked-On: https://jira01.devtools.intel.com/browse/OAM-12574
Signed-off-by: Du, Changbin <changbin.du@intel.com>
Reviewed-on: https://android.intel.com:443/460212

adb host: add device state in "adb wait-for-*"

The current implementation of the host commands "adb wait-for-*" allows
to specify only the transport layer (local, usb or any).
This patch allows the specification of the expected device state
(bootloader, recovery, device or sideload), this is usefull for
scripting purposes.

Use case:
$ adb reboot sideload-auto-reboot
$ adb wait-for-usb-sideload && adb sideload package.zip

This is a port of: https://android-review.googlesource.com/#/c/184290

Change-Id: I56a64b2d0f089cf1eb77424956296d660132b629
Signed-off-by: Leo Sartre <leox.sartre@intel.com>
Tracked-On: https://jira01.devtools.intel.com/browse/OAM-15622
Reviewed-on: https://android.intel.com:443/450503