GWP-ASan's recoverable mode was landed upstream in
https://reviews.llvm.org/D140173.
This mode allows for a use-after-free or a buffer-overflow bug to be
detected by GWP-ASan, a crash report dumped, but then GWP-ASan (through
the preCrashReport() and postCrashReportRecoverableOnly() hooks) will
patch up the memory so that the process can continue, in spite of the
memory safety bug.
This is desirable, as it allows us to consider migrating non-system apps
from opt-in GWP-ASan to opt-out GWP-ASan. The major concern was "if we
make it opt-out, then bad apps will start crashing". If we don't crash,
problem solved :). Obviously, we'll need to do this with an amount of
process sampling to mitigate against the 70KiB memory overhead.
The biggest problem is that the debuggerd signal handler isn't the first
signal handler for apps, it's the sigchain handler inside of libart.
Clearly, the sigchain handler needs to ask us whether the crash is
GWP-ASan's fault, and if so, please patch up the allocator. Because of
linker namespace restrictions, libart can't directly ask the linker
(which is where debuggerd lies), so we provide a proxy function in libc.
Test: Build the platform, run sanitizer-status and various test apps
with recoverable gwp-asan. Assert that it doesn't crash, and we get a
debuggerd report.
Bug: 247012630
Change-Id: I86d5e27a9ca5531c8942e62647fd377c3cd36dfd
This is a no-op but will be used in upcoming scudo changes that allow to
change the buffer size at process startup time, and as such we will no
longer be able to call __scudo_get_ring_buffer_size in debuggerd.
Bug: 263287052
Change-Id: I18f166fc136ac8314d748eb80a806defcc25c9fd
I don't think we need or want to do this. Most other architectures
don't.
Test: bionic-unit-tests-static
Change-Id: I3ad31926909caf0a37e73ac6cbac1fecd02ea6de
Map all stacks (primary, thread, and sigaltstack) as PROT_MTE when the
binary requests it through the ELF note.
For the reference, the note is produced by the following toolchain changes:
https://reviews.llvm.org/D118948https://reviews.llvm.org/D119384https://reviews.llvm.org/D119381
Bug: b/174878242
Test: fvp_mini with ToT LLVM (more tests in a separate change)
Change-Id: I04a4e21c966e7309b47b1f549a2919958d93a872
Currently, the ELF note parsing in the loader is not permissive. This
patch relaxes the restrictions on the bits of the ELF note that could be
extended in the future.
This may allow more MTE options in apexes. If we add some extra metadata
bit (say, in bit 5) in the future, and then build MTE into every apex,
we don't want it to crash on Android 13 devices (we'd much rather it
just be a no-op).
Remove the (unused anywhere) NT_MEMTAG_DEFAULT and call it
NT_MEMTAG_NONE.
And finally, make the tests work on bionic-unit-tests-static. We
previously didn't deploy the test binaries, so add them as a data
dependency.
Bug: N/A
Test: atest bionic-unit-tests-static
Change-Id: I13530faad55c719c6eb848297f8ce378e18afbfc
Normally, platform-specific note types in the toolchain are prefixed
with the platform name. Because we're exposing the NT_TYPE_MEMTAG and
synthesizing the note in the toolchain in an upcoming patch
(https://reviews.llvm.org/D118948), it's been requested that we change
the name to include the platform prefix.
While NT_TYPE_IDENT and NT_TYPE_KUSER aren't known about or synthesized
by the toolchain, update those references as well for consistency.
Bug: N/A
Test: Build Android
Change-Id: I7742e4917ae275d59d7984991664ea48028053a1
The first time should_trace() returns true, bionic_trace_begin() calls
open() on trace_marker.
The problem is that open() can call bionic_trace_begin(). We've observed
this happening, for example when:
* fdtrack is enabled. dlopen("libfdtrack.so") can be used to enable
fdtrack on a process.
* ThreadA is busy unwinding inside fdtrack and is holding an fdtrack
internal mutex.
* ThreadB calls bionic_trace_begin() for the first time since the
property "debug.atrace.tags.enableflags" contains ATRACE_TAG_BIONIC.
* ThreadB calls open("/sys/kernel/tracing/trace_marker"). Since fdtrack
is enabled, ThreadB tries to do unwinding as well.
* ThreadB, inside fdtrack's unwinding tries to grab the same mutex that
ThreadA is holding.
* Mutex contention is reported using bionic_systrace, therefore
bionic_trace_begin() is called again on ThreadB.
* ThreadB tries to grab g_lock in bionin_systrace.cpp, but that's
already held by ThreadB itself, earlier on the stack. Therefore
ThreadB is stuck.
I managed to reproduce the above scenario by manually pausing ThreadA
inside unwinding with a debugger and letting ThreadB hitting
bionic_trace_begin() for the first time.
We could avoid using g_lock while calling open() (either by releasing
g_lock and reacquiring it later, or by using atomics), but
bionic_trace_begin() would try to call open() again. In my tests, open()
does not call bionic_trace_begin() a third time, because fdtrack has
reentrancy protection, but there might be another code path inside open
that calls bionic_trace_begin again (it could be racy or only happen in
certain configurations).
This commit fixes the problem by implementing reentrancy protection in
bionic_systrace.
Sample callstack from ThreadA deadlocked before the fix:
```
* frame #0: 0x0000007436db077c libc.so`syscall at syscall.S:41
frame #1: 0x0000007436db0ba0 libc.so`bionic_trace_begin(char const*) [inlined] __futex(ftx=0x000000743737a548, op=<unavailable>, value=2, timeout=0x0000000000000000, bitset=-1) at bionic_futex.h:45:16
frame #2: 0x0000007436db0b8c libc.so`bionic_trace_begin(char const*) [inlined] __futex_wait_ex(ftx=0x000000743737a548, value=2) at bionic_futex.h:66:10
frame #3: 0x0000007436db0b78 libc.so`bionic_trace_begin(char const*) [inlined] Lock::lock(this=0x000000743737a548) at bionic_lock.h:67:7
frame #4: 0x0000007436db0b74 libc.so`bionic_trace_begin(char const*) [inlined] should_trace() at bionic_systrace.cpp:38:10
frame #5: 0x0000007436db0b74 libc.so`bionic_trace_begin(message="Contending for pthread mutex") at bionic_systrace.cpp:59:8
frame #6: 0x0000007436e193e4 libc.so`NonPI::MutexLockWithTimeout(pthread_mutex_internal_t*, bool, timespec const*) [inlined] NonPI::NormalMutexLock(mutex=0x0000007296cae9f0, shared=0, use_realtime_clock=false, abs_timeout_or_null=0x0000000000000000) at pthread_mutex.cpp:592:17
frame #7: 0x0000007436e193c8 libc.so`NonPI::MutexLockWithTimeout(mutex=0x0000007296cae9f0, use_realtime_clock=false, abs_timeout_or_null=0x0000000000000000) at pthread_mutex.cpp:719:16
frame #8: 0x0000007436e1912c libc.so`::pthread_mutex_lock(mutex_interface=<unavailable>) at pthread_mutex.cpp:839:12 [artificial]
frame #9: 0x00000071a4e5b290 libfdtrack.so`std::__1::mutex::lock() [inlined] std::__1::__libcpp_mutex_lock(__m=<unavailable>) at __threading_support:256:10
frame #10: 0x00000071a4e5b28c libfdtrack.so`std::__1::mutex::lock(this=<unavailable>) at mutex.cpp:31:14
frame #11: 0x00000071a4e32634 libfdtrack.so`unwindstack::Elf::Step(unsigned long, unwindstack::Regs*, unwindstack::Memory*, bool*, bool*) [inlined] std::__1::lock_guard<std::__1::mutex>::lock_guard(__m=0x0000007296cae9f0) at __mutex_base:104:27
frame #12: 0x00000071a4e32618 libfdtrack.so`unwindstack::Elf::Step(this=0x0000007296cae9c0, rel_pc=66116, regs=0x0000007266ca0470, process_memory=0x0000007246caa130, finished=0x0000007ff910efb4, is_signal_frame=0x0000007ff910efb0) at Elf.cpp:206:31
frame #13: 0x00000071a4e2b3b0 libfdtrack.so`unwindstack::LocalUnwinder::Unwind(this=0x00000071a4ea1528, frame_info=<unavailable>, max_frames=34) at LocalUnwinder.cpp:102:22
frame #14: 0x00000071a4e2a3ec libfdtrack.so`fd_hook(event=<unavailable>) at fdtrack.cpp:119:18
frame #15: 0x0000007436dbf684 libc.so`::__open_2(pathname=<unavailable>, flags=<unavailable>) at open.cpp:72:10
frame #16: 0x0000007436db0a04 libc.so`bionic_trace_begin(char const*) [inlined] open(pathname=<unavailable>, flags=524289) at fcntl.h:63:12
frame #17: 0x0000007436db09f0 libc.so`bionic_trace_begin(char const*) [inlined] get_trace_marker_fd() at bionic_systrace.cpp:49:25
frame #18: 0x0000007436db09c0 libc.so`bionic_trace_begin(message="pthread_create") at bionic_systrace.cpp:63:25
```
Bug: 213642769
Change-Id: I10d331859045cb4a8609b007f5c6cf2577ff44df
musl libc doesn't provide fts, but elfutils and libabigail need it.
Export bionic's fts as a staic library that can be linked into elfutils
and libabigail when compiling against musl.
fts uses recallocarray, which musl doesn't provide, so also include
recallocarray.c in libfts.a.
Requires minor tweaks to fts.c and a wrapper around fts.h to make them
compatible with musl, primarily by providing local definitions of macros
provided in bionic's sys/cdefs.h.
Bug: 190084016
Test: m libfts
Change-Id: Ifac9a59e7504c0c1f5f8a3a5bd3c19a13980b83c
Revert submission 1403568-sysprop_trace
Reason for revert: makes property get/set non-reentrant
Reverted Changes:
I6f85f3f52:Add systrace tag for system property
Id2b93acb2:Adding system property tracing
Id78992d23:Add systrace tag for system property
I1ba9fc7bd:Add systrace tag for system property
Bug: 193050299
Test: build and boot a device
Change-Id: Ic7a83fb01a39113d408ed0c95d27f694d5a2649c
Merged-In: Ic7a83fb01a39113d408ed0c95d27f694d5a2649c
(cherry picked from commit 61a5a8380d)
Also delete some fdsan code that attempts to check for the post-fork
state, but never will, because we update the cached pid upon fork.
Bug: http://b/174542867
Test: /data/nativetest64/bionic-unit-tests/bionic-unit-tests
Test: treehugger
Change-Id: I9b748dac9de9b4c741897d93e64d31737e52bf8e
This works around buggy applications that read a few bytes past the
end of their allocation, which would otherwise cause a segfault with
the concurrent Scudo change that aligns large allocations to the right.
Because the implementation of
android_set_application_target_sdk_version() lives in the linker,
we need to introduce a hook so that libc is notified when the target
SDK version changes.
Bug: 181344545
Change-Id: Id4be6645b94fad3f64ae48afd16c0154f1de448f
This matches what we do for arm and arm64. 32-bit x86 is too big a mess
to warrant the effort still, but the more testing is done on cuttlefish,
the more value there is to making every stack frame count.
Before:
#00 pc 00000000000596d8 .../libc.so (syscall+24)
#01 pc 000000000005d072 .../libc.so (abort+194)
#02 pc 000000000005f1f0 .../libc.so (__fortify_fatal(char const*, ...)+160)
After:
#00 pc 000000000005d07d .../libc.so (abort+205)
#01 pc 000000000005f1e0 .../libc.so (__fortify_fatal(char const*, ...)+160)
Test: crasher64 fortify
Change-Id: Ib74cb8b36341093c268872e26020f35eb2d8ef66
Bug: http://b/157081822
If __libc_int0x80 is in a C/C++ file, Clang's coverage instrumentation
adds instructions to count the number of times it gets executed [1].
With coverage instrumentation, __libc_sysinfo, used on 32-bit x86, is
initialized to the wrong value, causing dl.preinit_system_calls to fail.
Moving the function to an assembly file leaves __libc_sysinfo properly
initialized.
[1] We could change clang so it doesn't instrument functions marked
__attribute__((naked)) as a followup.
Test: `m CLANG_COVERAGE=true NATIVE_COVERAGE_PATHS=bionic` and run
bionic-unit-tests
Change-Id: I73558253512392d345de8d5b66d38bb14b308fdf
With this change we can report memory errors involving secondary
allocations. Update the existing crasher tests to also test
UAF/overflow/underflow on allocations with sizes sufficient to trigger
the secondary allocator.
Bug: 135772972
Change-Id: Ic8925c1f18621a8f272e26d5630e5d11d6d34d38
"#" there is incorrect: macro parameter is expected following a hashmark.
Test: build libc with -mbranch-protection=standard
Change-Id: Ib8e7ddf260b4cdbd36246cc70f69970f33dee200
Technically, std::atomic does not have an operator(). Previously, this
code was relying on an implicit behavior of our C++ standard library in
order to compile. When compiling this code against a different C++
standard library, I encountered a compiler error on these lines.
This CL makes the std::atomic load() operation explicit, makes it
clearer what this code is actually doing and makes it conform better to
the C++ standard library spec rather than a particular implmentation.
Change-Id: I7f255dffc0a3d8e07c973c18e9ba4098c4b5843e
Use a note in executables to specify
(none|sync|async) heap tagging level. To be extended with (heap x stack x
globals) in the future. A missing note disables all tagging.
Bug: b/135772972
Test: bionic-unit-tests (in a future change)
Change-Id: Iab145a922c7abe24cdce17323f9e0c1063cc1321
In native build of libc it would be inlined and in native bridge mode
it's noinline, extern "C" and thus could be easily intercepted.
Test: m (without weak symbols in native bridge mode x86+arm build would be broken)
Change-Id: I67759858a5bc2174dce1db9732fdbd89ba7689cc
Now, when we can detect native bridge mode is sources we can do that.
Test: m (without weak symbols in native bridge mode x86+arm build would be broken)
Change-Id: I360e7df8211d03636bbe716dc14655ee8d765493
Let the assembler handle the low-level details of generating .eh_frame,
and use .cfi_escape to output the DWARF3 DW_CFA_def_cfa_expression and
DW_CFA_expression instructions.
Explicitly output a nop instruction between the FDE start
(.cfi_startproc) and the trampoline symbol.
x86_64: remove the rsp instruction, which is redundant with the CFA
itself.
Bug: http://b/169383888
Test: compare `objdump -Wf` output before and after
Test: bionic-unit-tests
Change-Id: Ie2a6d111cb473596be8c9f4fd64534e91d88f2a1
The instruction "bti c" is added through ENTRY*() macro,
using __bionic_asm_custom_entry(f).
The .note.gnu.property section is added with the new macro
NOTE_GNU_PROPERTY(). BTI and PAuth features are automatically
selected based on the presence of __ARM_FEATURE_* macros.
Furthermore, gensyscalls.py got updated to append the new
macro to the generated syscalls-arm64.S.
Test: Tested on FVP with BTI enabled.
Change-Id: I40ffe294b8426421125fffd0a9758567d919a09d
Introducing a new systrace tag, TRACE_TAG_SYSPROP, for use with
system property.
For property set, the tracing is added in __system_property_set() instead of
__system_property_update() / __system_property_add() so we can record
control properties like ctl.*, sys.powerctl.*, etc.., which won't be
updated via the latter two functions.
Bug: 147275573
Test: atest CtsBionicTestCases
Test: adb shell perfetto -o /data/misc/perfetto-traces/test_trace -t 10s bionic
Test: adb shell perfetto -o /data/misc/perfetto-traces/test_trace -t 10s sysprop
Test: adb shell /data/benchmarktest64/bionic-benchmarks/bionic-benchmarks \
--benchmark_filter=BM_property --bionic_cpu=4, then compares the results
of property benchmarks before and after the change, didn't see
significant difference.
Change-Id: Id2b93acb2ce02b308c0e4889f836159151af3b46
Merged-In: Id2b93acb2ce02b308c0e4889f836159151af3b46
(cherry picked from commit 26970c3493)
We can't have HWASAN enabled in these resolvers because they
may be called before HWASAN is initialized.
We were previously getting away with HWASAN being enabled on these
functions because ANDROID_EXPERIMENTAL_MTE was not defined, which
let the compiler simplify away any HWASAN instrumentation in the
resolvers. But with this build flag enabled (which is currently the
case in FVP, and eventually everywhere once the MTE kernel patches
land) the resolvers end up containing an HWASAN instrumented load of
the _hwcap2 field.
Bug: 135772972
Change-Id: Iac79d02ddc81630c955e107eb7f78389aeb33879
(Based on proposal at https://sourceware.org/glibc/wiki/ThreadPropertiesAPI)
This includes API to:
- locate static and dynamic TLS
- register thread-exit and dynamic TLS creation/destruction callbacks
Change-Id: Icd9d29a5b2f47495395645e19d3b2c96826f19c8
When glancing at libc.so, I noticed we had five copies of
`__fortify_fatal` sitting around. LLVM can't inline variadic functions;
allowing everything to reference the same one seems best to me. This
saves us a few hundred bytes. Whoopee.
Bug: None
Test: TreeHugger
Change-Id: I99188d49cd46d792e3167cb961f13b3c1b65bd61
Partly to buff our coverage numbers, but also for improved consistency
in error reporting.
Test: treehugger
Change-Id: Iffc32833a35f9e9535c1bc3e0f7cb3c4bbba5f7f
The SYNC tagging level enables stack trace collection for allocations and
deallocations, which allows allocation and deallocation stack traces to
appear in tombstones when encountering a tag check fault in synchronous tag
checking mode.
Bug: 135772972
Change-Id: Ibda9f51b29d2c8e2c993fc74425dea7bfa23ab1e
When a thread calls pthread_exit(3), ScopedSignalBlocker blocks all
user-visible signals, but leaves internal-use signals 33/34/36/36/37.
Signal 33 is used to unwind a thread for a backtrace, which can cause us
to access the stack after it's been unmapped. (Avoiding this was the
reason why we have the ScopedSignalBlocker in pthread_exit(3)!)
Fix this (and other potential issues) by changing ScopedSignalBlocker to
call __rt_sigprocmask(2) directly, so we don't mask out the internal-use
signals.
Bug: https://issuetracker.google.com/153624226
Test: not trivially reproducible
Change-Id: I9b125ed41ddee4c5d33b45920f1d142e52db47cb
Exports GWP-ASan allocator information callbacks to libdebuggerd so that
tombstoned can get information from the GWP-ASan allocator in the case
of a crash.
Bug: 135634846
Test: atest bionic-unit-tests
Change-Id: Ie16426af55602fb2a76c4e69217773354c365843
