This change allows applications to share RTLD_GLOBAL
native libraries between namespaces associated with
different classloaders.
The rule is - if a library is GLOBAL within namespace
associated with parent classloader, it is shared
with namespace associated with this classloader
Note that the sharing happens on create_namespace
event, which is tied to createClassloader in case of
application classloaders created by the framework, for
custom application classloaders it is tied to first
loadLibrary() event.
Bug: http://b/28560538
Bug: https://code.google.com/p/android/issues/detail?id=208458
Change-Id: I7ee701166f8ec5eff033b7acc0f80c7aa4ec5bda
Now that we have the fixed list of whitelisted libraries
there is no longer need to check if java_permitted_path is nullptr.
In fact checking it prevents custom classloaders created by apps
from loading libraries using absolute path.
Bug: http://b/28659864
Change-Id: I1bfc16a573b090f70f078ecccd0e4016939a5dc0
The dm-verity metadata contains the block device path that is given to
the dm-android-verity driver. If the device is using slot A/B, this
path is missing the slot suffix.
This patch makes fs_mgr replace all the block device path reference
with the one including the active slot suffix.
Change-Id: Ib624d0d18b12a8a287cac7f15605a2e0fb7c97c6
Signed-off-by: Jeremy Compostella <jeremy.compostella@intel.com>
This makes it possible to get a core dump from adbd when it decides to
exit.
Bug: http://b/28347842
Change-Id: I4cfe5f273f62b2c32e61232d3c39881ecdd6b582
adb implements its own file descriptor emulation layer on Windows,
which requires the use of adb_close instead of close throughout the
codebase. Add a template argument to unique_fd that allows for this.
Bug: http://b/28347842
Change-Id: I6397261f4973d49f2f8e04257bf67b348585bb63
Currently, ueventd.rc files only support /sys/devices/... paths
and don't support symlinked paths, specifically /sys/class/...
Supporting the class paths is necessary to have non-hardware
dependent paths. Some subsystems like IIO use /sys/bus/iio/, so
support that as well.
Change-Id: I29f3bf67b41664d1d75ac1820c46e13afe336d56
Signed-off-by: Rob Herring <robh@kernel.org>
When fs_mgr_do_format() is executed, fs_mgr creates a new
blank data file system. However, that filesystem is not
labeled and causes some unlabeled denials on early boot.
Example Denial:
avc: denied { search } for pid=2535 comm="logd" name="/" dev="mmcblk0p9" ino=2 scontext=u:r:logd:s0 tcontext=u:object_r:unlabeled:s0 tclass=dir permissive=1
To correct this, pass sehandle to the internal ext4 routine.
This way the ext4 filesystem will be labeled at creation and
the root inode will have a label.
Change-Id: Ieeecaa8bbc258e6d743d281dd956bdaca98b365f
Signed-off-by: William Roberts <william.c.roberts@intel.com>
It's no longer specific to make_ext4fs, will be used for mksquashfs as
well.
BUG: 27467028
Change-Id: I41b8c2b168ada45c8378dee522391edfb8f6b9a6
Signed-off-by: Mohamad Ayyash <mkayyash@google.com>
BUG: 28704419
Going too fast can result in the property not being set
and causing an extreme corner case failure being reported
and blamed on is_loggable functionality.
Bug: 25792367
Change-Id: Idda6f67d655d8a84d2809daf695394d5293e7431
fixup_sys_perms does the same matching as the function perm_path_matches
other than the leading '/sys'. Adjust the path and use the common
matching function. Replace the fixed length C strings with std::string.
Change-Id: Ic09d91aa4ac1f1096018617537d782b82ee9ef63
Signed-off-by: Rob Herring <robh@kernel.org>
Passing empty permitted path should result in no-op.
This addresses review comment on CL with commit
f334cbf0e1
Bug: http://b/28639227
Change-Id: I4a4540e522e90a145a374939921932c86f35e88d
(cherry picked from commit d0b1531929)
Apps on sdcard live under /mnt/expand and therefore
this directory should be under linker namespaces
permitted_path.
Bug: http://b/28639227
Change-Id: I462f9f23656c95d9c2a48bb3f513abcd9d08f340
(cherry picked from commit f334cbf0e1)
There is no point in reading/loading the list of public
libraries once public namespace is already initialized.
Change-Id: I5a53c46bb57b8de241ba521a6321c12e94a76476
They are intended for platform use only and we shouldn't have them
in the public header file. This CL moves them to libnativeloader
private header file.
Bug: http://b/28174921
Change-Id: I01cf446d2d8d7cf6f9e31130aa3d4e54fbbbb9fc
This CL adds support to forward or reverse TCP port 0 to allow the
system to automatically select an open port. The resolved port number
will be printed to stdout:
$ adb forward tcp:0 tcp:8000
12345
$ adb reverse tcp:0 tcp:9000
23456
This allows testing to be more robust by not hardcoding TCP ports which
may already be in use.
Forwarding port 0 is a host-only change and will work with any device,
but reversing port 0 requires the device to be updated with a new adbd
binary.
This CL also does a little bit of cleanup such as moving the alistener
class out of adb.h, and adds some error checking and additional tests.
Bug: 28051746
Test: python -m unittest discover
Test: adb_test
Test: `adb forward` and `adb reverse` with tcp:0
Change-Id: Icaa87346685b403ab5da7f0e6aa186aa091da572
When user creates custom classloader which does not extend
BaseDexClassLoader the librarySearchPath gets set to null
by java.lang.Runtime.doLoad().
This patch makes nativeloader correctly handle it.
Bug: http://b/28659864
Change-Id: I1b61c6bc952984d7c49775a9178fc3270948e62a
This is enabled only for builds with ro.debuggable=1
It is intended for use only in tests using dalvikvm
and needing access to platform libraries.
Bug: http://b/28449304
Change-Id: I402457d0da542996ccf265aeaa305f09881e4333
Elliott Hughes