The APIs that are tagged with # vndk are actually for LLNDK libraries.
Although LLNDK is part of VNDK, calling those APIs 'vndk' has given
users a wrong perception that the APIs don't need to be kept stable
because that's the norm for most of the VNDK libraries that are not
LLNDK.
In order to eliminate the misunderstanding, rename the tag to 'llndk' so
that people introducing new such API will realize what they are signing
themselves up for.
Bug: 143765505
Test: m
Merged-In: Iae2acdf1ff4097a64a5c6280797c66abb1d5a5e6
(cherry picked from commit 0e957b82c8)
Change-Id: Iae2acdf1ff4097a64a5c6280797c66abb1d5a5e6
This protocol documentation is spread out among various functions
where it is implemented. This change makes a single
README.protocol.md file with a high level overview, referencing the
struct names where useful.
Test: n/a
Change-Id: I83c9f484352b489b4a20cce241d92413f780f9ec
logger_entry and logger_entry_v2 were used for the kernel logger,
which we have long since deprecated. logger_entry_v3 is the same as
logger_entry_v4 without a uid field, so it is trivially removable,
especially since we're now always providing uids in log messages.
liblog and logd already get updated in sync with each other, so we
have no reason for backwards compatibility with their format.
Test: build, unit tests
Change-Id: I27c90609f28c8d826e5614fdb3fe59bde22b5042
I'm about to make a bunch of changes that are safe for backwards
compatibility, but would otherwise flag the checker, so it's time to
temporarily disable it.
Test: CLs pass despite making header changes
Change-Id: Ibc2d4ae51fb8e6125b9117ccd92bf821db945e67
This reverts commit 5f8162b086.
Reason for revert: Turns out they're being used.
Merged-In: Iad9010190c7a4140b69dc553df5debdd88dcf81a
Change-Id: Iad9010190c7a4140b69dc553df5debdd88dcf81a
These functions and headers were all mistakenly added to the vndk.
They should not be used by vendors.
Test: these symbols do not appear in vendor libraries
Merged-In: I03919b437c2d9f0e573b7a6b40249ed12fe874b9
Change-Id: I03919b437c2d9f0e573b7a6b40249ed12fe874b9
1) We don't need two copies of log_id_t
2) We don't need misleading sizeof_log_id_t or typeof_log_id_t macros
3) logd should use android_log_header_t explicitly for its recv buffer
size
4) Following on from b/129272512, we're settling that returning
LOG_ID_MAX is an acceptable return value from
android_name_to_log_id().
Bug: 129272512
Test: build, liblog, logcat unit tests
Change-Id: I67fb964a4a0ae9cb6e1514ca110e47e00dfcfa9a
There's no point in client side security checks in this library. If a
process has access to these files, then they'll be able to do any of
these operations themselves.
Test: liblog unit tests
Change-Id: I75d4e1509eb8ff0ac4579f820a8968f4f5ad4e06
Enable more of the disabled tests. These should not be flaky and have
value.
The dual_reader test is rewritten in the style of RunLogTests() as
well.
Test: 100 iterations of these tests on CF without failure
Change-Id: I15de9e21b066aa22635cc0bd71b51e2648198823
The log_time struct satisfies all of the requirements for an
implicitly created copy constructor to be present, so not defining one
here does not have any real effect.
We don't want to delete the copy constructor for the rationale given
either; modern C++ favors passing small types by value instead of by
reference as the compiler has more opportunity for optimization in
that case. That's especially true here, where the size of this struct
is the size of a pointer on 64 bit systems.
Test: the copy constructor exists for log_time
Change-Id: Id314ca7729f4b1ca02adb6c7f0ae759b22be2a5c
Continuing the speed up / clean up from the last change.
Enable a subtest that was previously disabled as well. It passes 100s
of cycles now on CF.
Test: liblog-unit-tests
Change-Id: Ifff6f400c3736a1a857a3fdaf22d7ef1794abf9b
A lot of liblog tests follow this pattern:
1) Write a log message
2) Sleep ~1 second
3) Use the non_blocking log reader to dump all log messages
4) Test those log messages
This causes running back to back tests to be very slow and still
allows for some amount of flakiness if the system is very loaded.
This change replaces that pattern with the following:
1) Write a log message
2) Set an alarm for 2 seconds as a test timeout
3) Read logs with the blocking reader until finding the expected log
messages
4) Test those log messages
5) Use the non_blocking reader to dump all log messages
6) Re-test those log messages, to ensure no duplicates, etc, which
isn't done in step 3).
Despite dumping the logs twice, the tests are orders of magnitude
faster in the good case, and should be less prone to flakes.
Test: liblog-unit-tests
Change-Id: Iedf473316576b8007746fe3560815bde1813787a
Per jmgao@, we still need to worry about unaligned integer accesses.
In any case, we already have the packed structs for all of the
unaligned data that we're reading, so this change favors using those
packed structs.
Bug: 142256213
Test: x86,arm32,arm64 liblog-unit-tests
Change-Id: I21fc629eac49895d03b5b31daa4cc494b0c4c230
There is an alarm() call that provides a 30 second timeout in case
logd is unavailable. The main reason for this is in the case that logd
is crashing, debuggerd is ptrace'ing logd, and tombstoned is
attempting to dump log messages. In this case, with no other checks
and without this alarm, tombstoned will deadlock when dumping log
messages.
However, tombstoned already has two mechanisms to prevent the above
situation from happening:
1) It checks that the thread name that is is dumping is either logd or
starts with "logd." and skips dumping logs in this case.
2) It does not dump logs if it is running in process for any process.
Calling alarm() or modifying signal handlers from general purpose
libraries is not recommended either, so without a strong reason to
keep this, this change removes it.
This also shortens the liblog.wrap_mode_blocks test time, since the 30
second issue that it ensures does not happen has been fundamentally
removed.
Test: `kill -8 `pidof logd`` succeeds without delay
Test: liblog, logd unit tests
Change-Id: Id8a40544645d220e49f7ba299201af80a0c44de9
Fix a typo where I check socket()'s return value for 0 instead of -1
as an error.
Remove code that checks if socket() returns 0 as the actual fd that it
returns. This should not happen in any well formed program.
Test: liblog unit tests
Change-Id: I1d878e85d9a39155d68c6c84e9cf9b0db8d1b3a4
We disabled then re-enabled a set of tests believing that they were
stable; they were not, so we disable them again while we investigate
their stability.
A majority of tests filter the logs from logd based on pid, so running
in isolation helps them not see unexpected information from other test
runs.
Bug: 138876729
Bug: 142041379
Test: run via gtest with gtest_repeat=10 without failure
Test: run via atest 5x times without failure
Test: observe that isolated applies to both gtest and atest
Change-Id: I757d52dd3233323be9519625868b2fd2aaa41aeb
These tests are flaky, but we're not seeing what the failure strings
are.
Bug: 142041379
Test: force a failure and see the right error message
Change-Id: Icd7777e5c309cac3b98ce65925980965a3cc3753
The old code did not free the transports, just unlink their nodes from
the linked list, so there was no race condition between close() and
other threads writing to the logs. By settings these to nullptr, I
introduced a race condition, but setting them to nullptr isn't
necessary, so let's simply not do that.
Test: liblog-unit-tests
Change-Id: Iec0b680addd13b5d30bd91ccef5bdeab6bf797b0
The "branchless" macro generates slightly worse code than std::min or
<sys/param.h>'s MIN.
Test: treehugger
Change-Id: I92c17325383bd6116c7b4736d42f01f7a9bb81a3
These checks are not needed; all callers are in this file and there's
only one possible situation: that we have valid pointers and sizes.
Test: liblog-unit-tests, log.tag.* properties still work
Change-Id: I8b385d5124c2d93cd99a352e30f745b075c8bd09
Now that we only have one transport in a given build configuration,
there is no need to have these lists.
Test: liblog-unit-tests
Change-Id: I3fdcdc33c1224a9522080ea06f36f14c83d946ac
There are no users; future versions will need to be more generic to
support the libbase style Logger/Aborter options.
Test: build, liblog-unit-tests
Change-Id: Ia4571cd39d3e499f18306920ecbf40ea8878fb81
There are no users of the liblog stderr logger, but 66 users of
android::base::StderrLogger. We'll prefer that format and work to
merge it into liblog.
Test: liblog-unit-tests
Change-Id: Ifad24b27ac0f51379b925f6644caa1f6f3d9ad81
Previously this test would rely on ro.debuggable being true, but there
are now some circumstances where ro.debuggable is true, yet `su` does
not exist. Therefore we only run this test when it is root.
Also simplify some of the test macros.
Bug: 140446213
Test: passes as root, skips as non-root
Change-Id: I0651963d5c7d4f5308989ab6b04aab1042094e5a
liblog has left over code from local_logger that allows for reading
from multiple sources and merging the contents. Since we've already
removed local_logger, this change removes the rest of this code.
Test: liblog-unit-tests
Change-Id: I5685ad6c1e7cbcaa0a660ed10f47714784a40791
Some tests were meant to be run with logd only and were ifdef'ed based
on USING_LOGGER_DEFAULT. Unfortunately a previous change removed the
appropriate #define, so these tests weren't being run at all. They
passed multiple times on CF, so they should still be stable.
Except for one test for __android_log_event_list. We removed that
struct, so this change removes the test as well.
Test: liblog-unit-tests on CF (5 runs)
Change-Id: Ifaf3dbfdc5c714aead3f26c333e9cf19a1ed86c6
Enable native bridge support for libbase, liblog,
libziparchive and libpropertyinfoparser.
This makes it possible to use them in binaries for translated
architectures.
Bug: http://b/77159578
Test: make
Change-Id: If67ce92288b17a052ea1e79a268e284f7d941439
This started as a change to use mbrtowc() instead of
utf8_character_length() as mbrtowc() does everything that
utf8_character_length() intends to do, but is a libc function. The
change was further intended to add unit tests to ensure that these
functions operate as intended.
It turns out that utf8_character_length() returned an error for the
utf8 characters that I tested, so this also has the side effect of
allowing valid utf8 characters to be printed in the 'printable' log
format, which was the original intention.
Also, print the binary data as hex instead of octal, since it is a
more suitable choice.
Test: new unit tests, existing unit tests, logcat -v printable
Change-Id: I4cc95aee81519411ef47892ca74eb31117c972d2
I'm not sure why I didn't remove this when I got rid of the other
functions marked as weak.
Test: build
Change-Id: I45e6bca7be0497e33be88d15afea8bb1d5165380
The system version of liblog has been versioned. Do the same for
LL-NDK verison as well.
Test: Build and examine the dynamic symbols.
Change-Id: Ie63cc845c891d9372fb1990e713b3fe4f0f2f723
There are a bunch of branches that check "id >= LOG_ID_MAX", but because
C++ hates you, this does a promotion to signed int despite the
fact that both sides of the comparison are the same enum with an
underlying type of unsigned int. (C++17 §7.6.3)
Return LOG_ID_MAX instead of a value that gets promoted to signed -1, to
avoid this.
Bug: http://b/129272512
Test: /data/nativetest64/logcat-unit-tests/logcat-unit-tests
Change-Id: I4b3ee662d76d5cc80d9a9625d17f7e5b5980de41
The code section is already marked as to be removed, simply suppress
the -Wstring-plus-int warning.
Test: m checkbuild
Bug: 128878287
Change-Id: I95a03aca90dbb5d27db49f5baf534cbaacf9b830
Remove a superfluous check in fs_mgr, since those return values are
already impossible for a non-debuggable build. Replace a one time
call to __android_log_is_debuggable(), since there's no reason to
cache the value or use this symbol.
Test: build
Change-Id: Icd4bef7b616c49d304303747388d7e3018c6fcfc
Ensure that only the symbols in liblog.map.txt can be used by the
platform.
Bug: 123349183
Test: build
Change-Id: I99ae5d0e8ba8f5061ec20701c941d861c9eb615d
socket_local_client() was copied from libcutils, but we only need a
small subset of its functionality. We instead create our own version
with just the needed parts.
Importantly, use CLOEXEC in this new function and other places where
it was missing previously.
Test: logging works, liblog-unit-tests
Change-Id: Ifb929227af67bafa13e391eab92358d9f6fe6450
liblog has a version script for the NDK and VNDK, however it doesn't
have symbols intended to be used by the platform. This change adds
those symbols to the version script and enables it, hiding the rest of
the symbols.
Test: build
Change-Id: I494d048d78bb47d763482eb56a6f79babfe163f4
- Zero initialize log_time instances by default
- Disable implicit conversions by making constructors explicit
- Explicitly initialize to EPOCH in most places
- Change sniffTime() to avoid in-place modification of a log_time
I stopped here, but we could consider following up with a more invasive
change to make log_time instances immutable and perhaps also remove the
default constructor to force explicit initialization to EPOCH.
Test: atest libbase_test netd_unit_test
Change-Id: I67e716ef74adaaf40ab2c6e2e0dddb8d309bc7ca
readv() isn't used by anyone, writev() has one easily replaced user.
uio.h can be left as a private header for windows compatibility with
struct iovec.
Test: build
Change-Id: I33d4c6bdee6fd818271f78ae06abdd2aa09430f2
After a few years of being available, there only ended up being one
user of this, so it is clear that logd's duplicate message mechanism
is the favored solution.
The one user of this rate limiting is questionable as is, since due to
the nature of storaged, the mainloop only runs once per minute by
default as is, so there is essentially nothing to be gained by rate
limiting any further.
Test: build
Change-Id: I0610d11efda1ce8b581b939bad11ff295cb2daa6
These headers are platforms headers, so by definition they'll never be
compiled for a platform less than the current one.
Test: build
Change-Id: Ic1c1535b935929d713806faf17fbf9cd82275329
Some liblog tests test symbols that aren't otherwise used on the
platform. Statically link the tests such that we can add version
scripts that expose only the symbols used by the platform, and still
allow the tests to test the rest.
Test: build + unit tests
Change-Id: I898280fa52d31a2c7ad4eafdaa85e8b5f2e87971
No users ever signed up to use this, so remove it to ease the
refactoring of liblog/libbase.
Bug: 119867234
Test: liblog unit tests
Change-Id: I37b99644112bae7b4a2e3f4d06749db08de4ea14
If realloc returns NULL, the memory passed to it will remain allocated.
That doesn't appear to be intended here.
Caught by our static analyzer.
Bug: None
Test: Ran the static analyzer.
Change-Id: I579bee6827146d404fc6f6b86074366aefbda63f
* Need a body and return value for this function
so the constexpr check can inline it.
* C mode can just use an extern function declaration, but not constexpr.
Bug: 116854606
Bug: 111614304
Test: build with WITH_TIDY=1 and enable static analyzer checks
Change-Id: Ie3f4efbcabed99416d196b6c361a772b8c6a4035
Add FALLTHROUGH_INTENDED for clang compiler.
Bug: 112564944
Test: build with global -Wimplicit-fallthrough.
Change-Id: I40f8bbf94e207c9dd90921e9b762ba51abab5777
With replaced AOSP system.img, adb can be root by other test.
The fail is caused by the reason that logdw is still active after
__android_log_close().
This test has a potential bug.
proc/net/unix may contain '\0', then "grep" command will treat this file
as a bianry file.
Using "grep -a" instead of "grep" could be a safer way to get the right
results from bianry file.
bug: 112250722
Test: Build CTS
Test: Pixel 2 XL PPP5.180610.010 passed the following tests
Test: run cts-on-gsi -m CtsLiblogTestCases --skip-preconditions
Change-Id: I9bfca1522e90ebb68e6cecbdb63e3b67248e47c4
Signed-off-by: Xinghua Yang <xinghua.yang@mediatek.com>
Add android_log_parser_reset and android_log_reset to reset reader/writer
log context. This allows to generate multiple log messages of the same
format without context reallocation.
Change-Id: Icb0f15855378aa4e66fc671ffefd15d9df6fee1e
* New clang compiler requires variadic function to have
at least one named parameter type.
* Use ##__VA_ARGS__ to work with empty __VA_ARGS__.
* Fix one ALOG_ASSERT parameter bug in lmkd/lmkd.c.
Bug: 111614304
Test: make with WITH_TIDY=1
Change-Id: I90f35aa88527a6897954f69a35b256a157a725c5
This avoids 2 static analysis warnings:
* assigned but unused variables in __VA_ARGS__
* unused expression result in __VA_ARGS__
Bug: 111614304
Test: make with WITH_TIDY=1
Change-Id: I4faae8787f6cc76de7ff4b6d08d25d0cb47324ea
Clang static analyzer can optimize out if (false) ...
and report unused variables in __VA_ARGS__.
Bug: 111614304
Test: make with WITH_TIDY=1
Change-Id: I214ced736230fda847031fd4eee23015fd988ffc
* changes:
Do not customize __format__ for Windows/MinGW to gnu_printf
Adapt to switch to Clang for Windows host builds
Update cflags for building Windows modules with Clang
Bug: http://b/69933090
Bug: http://b/69933068
MinGW uses gnu_printf to force C99 printf analysis with GCC, but clang
does not support gnu_printf. So just use the default that's used for
other platforms. This also mirrors upstream commit
015e637b4b/.
Test: m native-host-cross with Clang.
Change-Id: I3deb266d70e25296c4ae1d58637afbabc3d949c1
On svelte devices, CTS test liblog#android_logger_get_ fails on the
missing kernel buffer because it is not enabled to save space.
Test: gTest liblog_unit_tests
Bug: 109669791
Change-Id: Iaf46fe9713d3462a56885515a67db640544345ab
Global flag --pack-dyn-relocs=android is used with clang lld.
For b/24465209, we need to override that with pack_relocations:false.
Bug: 24465209
Bug: 80093890
Test: build with USE_CLANG_LLD=true and run dlext.compat_elf_hash_and_relocation_tables
Change-Id: I4edcdf49e184fa45ea2b6b6417654d81bcd09e1d
See longer explanation in b/80093890.
Clang lld does not generate expected DT_REL and DT_RELA tags
with --hash-style=both and --pack-dyn-relocs=android.
I am not sure about the extent of b/24465209, so
I would rather not to use lld for these .so files for now.
Bug: 80093890
Bug: 24465209
Test: build with USE_CLANG_LLD=true and run dlext.compat_elf_hash_and_relocation_tables
Change-Id: I645dbe25c5b9975605e3af5e717ed36c276f14d4
The static analyzer is concerned about the strcpys below this, since it
apparently doesn't try to model snprintf's potential behaviors. (In
particular, it was concerned that suffixLen might be >=
sizeof(suffixBuf)). While that's clearly suboptimal, this code can also
be simplified to make it more obvious what's happening and to appease
the analyzer.
No functionality change is intended.
Bug: None
Test: Ran the analyzer. It's no longer angry about strcpy overflows.
Change-Id: I4aa812144c90f6d3e833bbcb23c0694476a0e53e
Libraries that are direct or indirect dependencies of modules installed
to recovery partition (e.g. toybox) are marked as recovery_available:
true. This allows a recovery variant of the lib is created when it is
depended by other recovery or recovery_available modules.
Bug: 67916654
Bug: 64960723
Test: m -j
Change-Id: Ie59155c08890e96ce1893fa3687afcf763d7aea3
This is required in order to be able to log audit
events from wpa_supplicant. Only log writing functions
from log_event_list.h are made available.
Test: manually
Bug: 70886042
Change-Id: I097d4ad1de573662658678e6a9f1fc8f17820542
Revert "Remove obsolete workaround."
This reverts commit 1f3ac7583f.
Bug: 72143978
Bug: 24465209
Test: Tested failing case on sailfish, reverted back all CL's since
Test: 3471433 for b/24465209 and apps open
Add local BM_time_clock_gettime_*, BM_time_clock_getres_* and
BM_time_time benchmarks. Relates to the bionic benchmarks of
the same names, except adds CLOCK_MONOTONIC_RAW. Added here for
developer convenience whenever updates to the liblog or logd
code base need integration testing.
ToDo: add liblog gTests that analyse the benchmark data to confirm
that the specified integrated device has vdso access to all the
pertinent clock sources. Add liblog local benchmarks and tests to
measure the device clock drift of each possible liblog clock
source to help evaluate device configuration.
Test: liblog_benchmarks --benchmark_filter=BM_time*
Bug: 63737556
Bug: 69423514
Change-Id: Ibafe0880d976ef2b3885765f71e0ba6c99d56f2a
Remove our circa 2014 snapshot of the google benchmarking library, and
use the now very stable google-benchmark suite. Some porting effort,
and deal with some benchmarking saddle points that take too long to
sort out. Focus on minimal API changes, _odd_ new behaviors, and style.
Test: liblog_benchmarks, ensure results in about the same range
Bug: 69423514
Change-Id: I2add1df9cb664333bcf983b57121ae151b26935b
Currently used clang tidy does not filter the warnings caused
by macros, even if macros come from filtered headers.
This change allows projects using ALOGV to use
readability-implicit-bool-conversion without spreading //NOLINT everywhere.
Bug: 71533509
Change-Id: Id1b193d1e56d13f00171e04f600292877c5f6cb3
In the future, the sizes of tv_sec and tv_nsec (or even the size of
log_time struct itself) can change due to the 32-bit overflow expected
to happen in the year 2138. In order to hide such implementation details
to the clients of liblog, the two macros LOG_TIME_SEC and LOG_TIME_NSEC
are introduced.
Furthermore, vendors are provided with a simplified version of log_time.h
without C++ APIs. In doing so, log_time.h no longer includes time.h.
This breaks several modules that implicitly relied on the hidden
dependency, which should be fixed.
Bug: 37629934
Test: build with BOARD_VNDK_VERSION=current
Merged-In: If213fc291395554fd8de5f5d1fb005ceaaa5ca57
Change-Id: I01b36078c1d8f3f44824be20ae769ba1465b6feb
(cherry picked from commit 98c0d030c9)
libdl is part of system_shared_libs now. -ldl -lpthread -lm are now defaults
for host_ldlibs on Linux and Darwin. -lrt is a default for host_ldlibs on
Linux.
Test: m host
Change-Id: I0b3c147b00a8ab6ff289b85db55b88836c905f5c
Exempt-From-Owner-Approval: build system cleanup
