This seccomp failure is in the fault handler:
05-25 12:03:25.042 10201 27425 27425 F DEBUG : backtrace:
05-25 12:03:25.042 10201 27425 27425 F DEBUG : #00 pc 00015380
/data/data/redacted/files/storage/lib/libcrashsdk.so
So whenever an app using this crash sdk crashes it looks like a seccomp
problem. Fixing this won't stop the apps crashing, but will make the
crash reports accurate and useful.
So yes, the bug below is already fixed, but this issue has come back 2
or 3 times with different apps (latest is b/62874867). This change
doesn't fix that crash either, but again it improves the reporting.
Bug: 62090571
Test: Device boots, app still fails but no longer with SECCOMP error
Change-Id: Ie0f8dc965001c8bc43f6a545b35bdcd38f006213
__libc_preinit sets up the stack protector global cookie value, and thus
cannot intialize a stack protector cookie for itself in the function
prologue. LTO compilation can inline functions requiring a stack
protector into __libc_preinit. This patch disables stack protection for
__libc_preinit and forces all potentially inlined functions into a
helper that can have a stack protector.
Test: run bionic-unit-tests
Change-Id: I45911611190f216c91eb6feff722967214c5f99f
No-one cares about seeing "async_safe_fatal" (which you have to admit is a
pretty confusing name for an app developer anyway).
On arm:
#00 pc 0001a43c /system/lib/libc.so (abort+63)
#01 pc 0001a627 /system/lib/libc.so (__assert+14)
And aarch64:
#00 pc 000000000001d75c /system/lib64/libc.so (abort+120)
#01 pc 000000000001dad0 /system/lib64/libc.so (__assert+44)
Bug: N/A
Test: ran `crasher assert` and `crasher64 assert`
Change-Id: I00be71c566c74cdb00f8e95d634777155bc3da03
With this, stack frame 0 is the abort, not tgkill.
arm:
#00 pc 0001a41c /system/lib/libc.so (abort+63)
arm64:
#00 pc 000000000001d75c /system/lib64/libc.so (abort+120)
Also "include what you use" for <sys/syscall.h>.
Bug: N/A
Test: ran `crasher abort` and `crasher64 abort`
Change-Id: I6517ac67b39b4133e890d52efc115071c812958b
linker_config#smoke and linker_config.asan_smoke are trying to find
paths under the /vendor directory. If there is no vendor partition,
the real path of them is started with /system/vendor.
This CL allows those paths in the tests by getting the resolved paths
for systems without a vendor partition.
Bug: http://b/62562515
Test: linker_config_test passes without a vendor partition.
Change-Id: Id6d16ef623efd81ab9083c3e819da2ad22a28bf8
This is no longer used in the platform, and shouldn't be used in NDK. Apps
should use the NDK's cpu-features module, which supports (a) more specific
queries and (b) all Android architectures, not just 32-bit ARM.
Bug: http://b/18556103
Test: builds
Change-Id: I544ef267a6d7d887223186180c77d9ad0321e605
Building ruby actually trips over both of these:
* if the RTLD_ constants aren't #defined, it uses its own incorrect values.
* if the REG_ constants aren't #defined, it confuses x86 with x86-64.
In all other places where we have enums in our headers, we already match
existing glibc practice.
Bug: http://b/62531921
Test: builds
Change-Id: I5b3aab25a1a24611bdc58f2eda4104a78e9f841c
Move all tests into stdlib_test.cpp since that's where the definition lives
in bionic.
Add a sweep test and a various size test.
Test: Run new unit tests on glibc and angler.
Change-Id: Ief1301f402bea82ce90240500dd6a01636dbdbae
No other C library expose these, and I couldn't find any callers.
Bug: http://b/62531921
Test: builds
Change-Id: I4a3505bc0897286a4036c48066b98d16665b573a
fstat64/fstatat64/_flush_cache were accidentally put in SYSCALLS.TXT in:
https://android-review.googlesource.com/#/c/390454/
This patch just moves them to SECCOMP_WHITELIST.TXT because we do not
want stubs accidenatally generated for the mentioned syscalls using
gensyscalls.py script.
This commit does not introduce any functional changes to mips64_policy.cpp.
Test: Run genseccomp.py -> File seccomp/mips64_policy.cpp not changed.
Test: Run gensyscalls.py -> INFO:root:no changes detected!
Change-Id: I3b527b3d9f18715c44a4e6ddc6db6e49f48f4890
Signed-off-by: Miodrag Dinic <miodrag.dinic@imgtec.com>
In a similar style to some of our other "not really, but enough" headers
like <sys/vt.h>.
Bug: N/A
Test: build GNU dd or BSD dd with a standalone toolchain
Change-Id: I8fbd1aac1d97e24b05e7aae8a55666300b5bf1ed
This, when used with BOARD_VNDK_VERSION flag, enables the
module to be built and installed twice, one for /system,
other for /vendor.
libasync_safe is used by libbacktrace which is in turn used by libs in
/vendor and thus marked as vendor_available. Therefore, libasync_safe
is also marked as vendor_available.
Details: https://android-review.googlesource.com/368372
Bug: 33241851
Test: BOARD_VNDK_VERSION=current m libasync_safe.vendor successful
Merged-In: I4b4de3b99a024d52612109cee3d66b4e5fc12dec
Change-Id: I4b4de3b99a024d52612109cee3d66b4e5fc12dec
(cherry picked from commit ec0d6b416e)
This breaks code that declares `errno` itself for whatever crazy reason:
b.c:22:12: error: conflicting types for '__errno'
extern int errno;
^
/usr/local/google/ndkports/toolchain/bin/../sysroot/usr/include/errno.h:47:20: note: expanded from macro 'errno'
#define errno (*__errno())
^
/usr/local/google/ndkports/toolchain/bin/../sysroot/usr/include/errno.h:44:15: note: previous declaration is here
volatile int* __errno(void) __attribute_const__;
^
Bug: N/A
Test: built various bits of GNU source
Change-Id: I27c03bf3bde419a001f98f1ea6c267c847f31271
This causes GNU make to fail to build:
main.o:main.c:(.text+0x18c): more undefined references to `bsd_signal' follow
Bug: N/A
Test: built GNU make with a hacked standalone toolchain
Change-Id: Icdfb4cf674df8abed44f6671454bed9f8d68e74b
r279663: enhance to handle 32-bit aligned data on 64-bit systems
r279666: qsort(3): small style(9) cleanups.
r288030: Use ANSI C prototypes. Eliminates -Wold-style-definition
warnings.
r318514: Use size_t.
r318515: Reduce the possible recursion level to log2(N) by always
recursing on the smaller partition instead.
Bug: 38450217
Test: builds, ./tests/run-on-host.sh 64
Change-Id: Ieb622e6f54f8f6d6f62e1bdb491d3b5c8e2f9965
The proper API for this isn't available until L, so expose this for
API levels earlier than that.
Test: make checkbuild
Bug: https://github.com/android-ndk/ndk/issues/420
Change-Id: I382b8f557be9530f3e13aaae353b4a6e7f9301ab
Treehugger Robot
Paul Lawrence
Josh Gao
Stephen Crane
Elliott Hughes
Jaesung Chung
Christopher Ferris
Nicolas Geoffray
Miodrag Dinic
Anders Lewis
Dan Albert
Neil Fuller
Jiyong Park
Xin Li