There is a race in the very_long_name_35166374 test of
property_service. The test first sends a size value that is beyond
the limit that init will handle, then sends a dummy data value.
However, init closes the socket upon seeing the faulty size, and if
this happens before the test sends the dummy data, the test will crash
due to SIGPIPE.
Since there is no reason to send the dummy data at all, this change no
longer sends it to prevent the crash. It also now checks explicitly
that init returns an error through the socket.
Bug: 73619375
Test: the unit test in question
Change-Id: I2565a69fa54910cee0e15fc798445e18c91156ec
If create_service_thread fails, we'll leak `arg`.
This fixes a static analyzer complaint:
system/core/adb/services.cpp:298:13: warning: Potential leak of memory
pointed to by 'arg'
Bug: None
Test: Reran the static analyzer. No more complaints about this leak.
Change-Id: I5aec7fd78f2cc775b650501b02bdf0039d1647ca
Add a comment explaining why we define PROT_READ/PROT_WRITE, even
though a current libminijail supports both cosntants.
Bug: http://b/73273658
Test: treehugger
Change-Id: I51c1be1b1b569e94dbc9045a90bc28221b7dc9c7
When generating crash_dump.*.policy, replace PROT_READ and PROT_WRITE
to numeric constants to make the policy backward compatible with old
libminijail.so.
Bug: 73273658
Test: use the new policy in OMR1 devices
Change-Id: I936a733340ad4df8aef6562c03eb10c29ffdada2
I have changed the root DEX debug descriptor in ART to more
closely follow the JIT debug descriptor. Add the appropriate
offsets to correctly fetch the linked list head for DEX files.
Test: testrunner.py -t 137
Test: libunwindstack_test
Change-Id: I90402befc88fec42658f7330d51ee79756a7f872
The dynamic section contained an address, not an offset into the elf
file to indicate where the soname exists. Changed to use the strtab
entries in the section headers to map this address to the actual offset.
Refactor the soname test a bit to make it easier to verify the code.
Bug: 73499044
Test: Passes new unit tests.
Test: Ran unwind_info on the failing shared elf and verified the soinfo
Test: is correct.
Change-Id: I16ba148389bcb9aadd3566fb442dac27f89fe894
This process is now spawned from the main app_process zygote, rather
than being its own binary launched by init.
Bug: 63749735
Test: m
Change-Id: I87e00197e1ce64d77b4c306fcd702c6ad778b236
This change adds some additional flags to the /config mount. This is to
reduce the number of mounts with unnecessary privileges.
Bug: 73255020
Test: aosp_sailfish still boots
Test: CtsAppSecurityHostTestCases {ExternalStorageHostTest,StorageHostTest}
Merged-In: If3409d917cdf76a67ebfb7c4035a3ae8fee6189f
Change-Id: If3409d917cdf76a67ebfb7c4035a3ae8fee6189f
This change adds some additional flags to /mnt. This is to reduce
the number of mounts with unnecessary flags.
Bug: 73255020
Test: aosp_sailfish still boots
Test: CtsAppSecurityHostTestCases {StorageHostTest,ExternalStorageHostTest}
Test: CtsOsTestCases StorageManagerTest
Test: find /mnt | egrep -v '^/mnt/runtime/(default|read|write)/emulated' | \
xargs ls -lZd # Shows no character devices or executable files
Change-Id: I54739133119d9626ebeb2ef9a1c127f7a90fa098
Since we only want to change the ro flag on / (and leave all other mount
flags alone), this can also be achieved by passing MS_REMOUNT|MS_BIND,
even if the mount is not a bind-mount.
This aims to make running Android within a user namespace easier, since
remounts without the MS_BIND flag are forbidden.
Bug: 73255020
Test: aosp_sailfish still boots
Test: rootfs on / type rootfs (rw,seclabel)
/dev/root on / type ext4 (ro,seclabel,relatime,data=ordered)
Change-Id: I2f89a8badfc467db47304c9355648e8fd8ad1272
Since 'struct timespec' members (time_t and long) are both 32bit on
32bit systems, and std::chrono::{seconds,nanoseconds}::rep are both
>32bit, timespec members assigned in DurationToTimeSpec() can have a
negative value, especially when WaitForProperty() is called with the
default timeout value which is std::chrono::milliseconds::max().
Regarding functionality, passing a negative value to
__system_property_wait() is okay because WaitForProperty() still
waits for the property value (so unit tests are passing), but while
WaitForProperty() does that, the function, to be more exact,
SystemProperties::Wait() in bionic/, consumes ~100% of CPU time. This
happens because SystemProperties::Wait() which implements
__system_property_wait() has a tight while-loop with a __futex_wait()
call, and the futex call immediately returns EINVAL when the timespec
passed in has a negative value.
With this CL, WaitForProperty() will never pass a negative timespec
to __system_property_wait(), and therefore the __futex_wait() call
in bionic works as expected without consuming too much CPU time even
on 32bit systems.
Bug: None
Test: libbase_test32 still passes
Test: strace no longer shows repeated EINVALs from __futex_wait
Change-Id: Id1834fac8cd2876b02dbe4479bf3d3eda2fa7da1
Move the files out of runtime, leaving unfortunate dependences still
there on runtime/base (for the moment). Also separates the build logic
into two Android.bp files.
Bug: 22322814
Test: make -j 50 test-art-host
Change-Id: I54b06f035034e7564480cd5d4c33c4f62785682d
For non-VNDK-enforced devices, libz cannot be installed in vndk-sp
to avoid CTS test failure.
Move libz to llndk from vndk-sp for generating the namespace
configuration from ld.config.vndk_lite.txt.
Bug: 73296261
Test: run android.jni.cts test on marlin/sailfish
Change-Id: I1b8b43cd87dc8dfdd2667001ae1e50d9eaedf93d
This commit disables "bootstat" in PDK builds because "bootstat" depends
on "libstatslog" (from "frameworks/base") which is not included in PDK
builds as well.
Test: Build a target (described in http://b/72961456) with
`platform.zip` built from master FSK source tree.
Bug: 72961456
Change-Id: I06b1555694510e17ea82d5c6dfcdeaf99b905e4d
