Additional benchmarks to check the effects and performance of the
frontend changes.
Test: manual liblog-benchmarks
Bug: 27405083
Change-Id: I2a9dba81b70e9d71cdb8de1b35e8dff099ad6038
android_set_log_frontend allows one to select LOGGER_DEFAULT,
LOGGER_LOGD or LOGGER_NULL, the latter allows a runtime mechanism to
turn logging on or off. LOGGER_DEFAULT is LOGGER_LOGD.
Test: gTest liblog-unit-tests notably liblog.android_set_log_frontend
and liblog-benchmarks before and after.
Bug: 27405083
Change-Id: I365dbee9f7d83112da0ef8d1cace748d247773c8
The behavior of passing NULL to c-string functions is undefined.
Test: passed clang-tidy checks
Bug: none
Change-Id: Ie7bcc43cd19bc4a1c314381af3929eae0a6154b9
If we don't output to stderr too, not only is it annoying to shell users (who
won't see anything), it prevents us from writing better gtests that actually
make assertions about the assert message.
Bug: http://b/23675822
Test: libutils tests still pass
Change-Id: I62b3144c385cba4dde485f0b0f9b42aeaef51e9a
android_lookupEventTagNum added. Adds support for creating a new
log tag at runtime, registered to the logd service.
Tested on Hikey, all services stopped, shell only access, CPUs not
locked (there is enough repeatability on this platform).
$ /data/nativetest64/liblog-benchmarks/liblog-benchmarks BM_lookupEventTagNum
iterations ns/op
Precharge: start
Precharge: stop 231
NB: only Tag matching, linear lookup (as reference, before unordered_map)
BM_lookupEventTagNum 1000000 1017
NB: unordered_map with full Tag & Format lookup, but with Tag hashing
BM_lookupEventTagNum 2000000 683
NB: with full Tag & Format hash and lookup for matching
BM_lookupEventTagNum 2000000 814
NB: only Tag matching (Hail Mary path)
BM_lookupEventTagNum 5000000 471
Because the database can now be dynamic, we added reader/writer locks
which adds a 65ns (uncontended) premium on lookups, and switch to
check for an allocation adds 25ns (either open code, or using
string_view, no difference) which means our overall speed takes 90%
as long as the requests did before we switched to unordered_map.
Faster than before where we originally utilized binary lookup on
static content, but not by much. Dynamic updates that are not cached
locally take the following times to acquire long path to logd to
generate.
BM_lookupEventTag 20000000 139
BM_lookupEventTag_NOT 20000000 87
BM_lookupEventFormat 20000000 139
BM_lookupEventTagNum_logd_new 5000 335936
BM_lookupEventTagNum_logd_existing 10000 249226
The long path pickups are mitigated by the built-in caching, and
the public mapping in /dev/event-log-tags.
SideEffects: Event tags and signal handlers do not mix
Test: liblog benchmarks
Bug: 31456426
Change-Id: I69e6489d899cf35cdccffcee0d8d7cad469ada0a
android_openEventTagMap(NULL) will open and mix content from
/system/etc/event-log-tags and /dev/event-log-tags. Only
reports identicals if global.
Test: gTest logd-unit-tests and liblog-unit-tests, liblog-benchmarks.
Bug: 31456426
Change-Id: Ic17d52a7829a4daaf013828d08fc1c09446ae8ef
Add a hidden -v nsec flag to logcat (actually logprint in liblog.so)
so that we can do more exacting tests of logcat.tail_time. Halve the
spam pressure of logcat.tail_* in inject() routine, and give us a few
more retries at the higher counts. Add instrumentation to
logcat.logrotate failures. Add inject for logcat.year test.
Test: gTest logcat-unit-tests
Bug: 34454772
Change-Id: If6f3bd21892c8a2b9ccee8c8bbf592a1ae0b2a57
Use refreshed android/log.h for NDK
This effectively a modern revert of the commit
b7c3996f7c
Test: build
Bug: 30465923
Bug: 34250038
Change-Id: I7d4a5134bb711051283c36a2d5bc803436dca12e
The visibility macros just expand to the defaults, so this isn't
needed anyway.
Test: mm
Bug: http://b/34740564
Change-Id: If0ba6b412554300823219405c1e801ceec747dd0
Move log reading definitions to their own home.
struct logger_entry
struct logger_entry_v2
struct logger_entry_v3
struct logger_entry_v4
LOGGER_ENTRY_MAX_PAYLOAD
LOGGER_ENTRY_MAX_LEN
struct log_msg
android_logger_get_id()
android_logger_clear()
android_logger_get_log_size()
android_logger_get_log_readable_size()
struct logger_list
android_logger_get_statistics()
android_logger_get_prune_list()
android_logger_set_prune_list()
android_logger_list_alloc()
android_logger_list_alloc_time()
android_logger_list_free()
android_logger_list_read()
android_logger_open()
android_logger_list_open()
Deal with a possible scenario where fcntl.h was not included at the
top of the source files, and log/log_read.h inclusion of fcntl.h
conflicts with a source file inclusion of sys/cdefs.h definition of
__unused macro.
Test: gTest liblog-unit-tests --gtest_filter=\
liblog.__android_log_write__android_logger_list_read:\
liblog.android_logger_get_
Bug: 34250038
Change-Id: Ib33544171563d6a351164754c254c4385686025b
Move common log_id_t and simple internal support function definitions
that use it from log/log.h to log/log_id.h.
log_id_t
__android_log_buf_write
__android_log_buf_print
android_name_to_log_id
android_log_id_to_name
Test: gTest liblog-unit-tests --gtest_filter=\
liblog.log_id:\
liblog.__android_log_buf_print:\
liblog.__android_log_buf_write:\
liblog.__android_log_buf_print__concurrent64
Bug: 34250038
Change-Id: Iad3704cc72943a3094e1193a6d032c7f29a6cd5c
move LOG macros to log/log_main.h
move include/android/log.h to liblog/include/android/log.h
Test: compile of all components and gTest liblog-unit-tests
Bug: 34250038
Bug: 30465923
Change-Id: If182dd9b83689e8b7bc1a44b2f5d913c7ee5eeee
ASan runtime library depends on liblog, so we have to disable ASan,
but nothing else.
Bug: 33091541
Test: SANITIZE_TARGET=address keeps working
Change-Id: I6e22925b7f1d5ec86fe1bd8c00dba4c3e86ddf3f
Moved headers from include/libcutils and include/liblog to
libcutils/include and liblog/include respectively, so they can be
exported via these libs. They needed to be moved since Soong does
not allow export from external folder.
Added symlink from old locations. They are needed since Soong
includes system/core/include by default. Once all modules are
cleaned up to explicitly add the required libs, the symlinks will be
removed.
Also added liblog_vndk_headers that exports a special log/log.h for
VNDK.
Moved headers of libcutils to libcutils_headers. They should be used
by modules for header-only inlines. Added libcutils_headers as
dependency of libcutils.
Added libcutils_vndk_headers that exports a special cutils/log.h
deprecating usage of the file. A later CL will deprecate the one in
libcutils_headers
Test: Add above libs to shared lib of local module
Change-Id: I6e1f9c5f23d8b6eae13dc3b7e5dfe7fae93b8510
This reduces BM_log_latency from ~300ns to ~100ns. Because, well, we
read the clock earlier and that is what BM_log_latency tries to
measure. The one major improvement in the switch from kernel logger
to user space logger was that we picked up the timestamp in the
context of the caller before doing anything else, but alas changes
over time neglected this fact and placed isloggable checking ahead
of this important fact.
Test: liblog_benchmarks, check results
Change-Id: I4bc9fc3cf8b1659e88417d967b1d0f3743f9e456
I'd like to mark setuid() WARN_UNUSED_RESULT to prevent security bugs.
This line would break because the file is compiled with -Werror.
Bug: 34251640
Test: Build, run unit test.
Change-Id: Ibb6fa5e5b899d47f77f2a32da762165de310101d
Deal with recovering after transitory failures surrounding logd
crash or recovery. Improve the chances that the logging functions
can work in a signal handler, not officially supported, but making
sure logging is not blamed for system lockups when misused.
Reorder gTests so that setuid(AID_SYSTEM) is performed after
liblog.enoent test, and that this occurs after other tests that
like to see buffers with content in them as we stop logd.
Test: gTest liblog-unit-tests --gtest_filter=liblog.enoent
Bug: 33755074
Change-Id: I66f88599534614b7b61da6b2ae5fe099ebaced3a
Some tests use hard-coded offsets to interpret the binary
events buffers. Switch to using the private event structures
to access the components of common event messages.
Test: gTest liblog-unit-tests
Bug: 33755074
Change-Id: I17447814583099d5ec417a54389e962158456005
toybox has a #define noreturn that trips over this.
Also move `format` out of the way, just in case.
Bug: https://github.com/android-ndk/ndk/issues/271
Test: builds
Change-Id: Ib8811136b4b422ff74625509539a5464a3c9af18
IF_ALOG_RATELIMIT()
ALOGE("Only one message globally using IF_ALOG_RATELIMIT()"
" in the process may appear ever ten seconds,"
" (%s)", strerror(errno));
static time_t last; /* initial state zero */
IF_ALOG_RATELIMIT_LOCAL(60, &last)
ALOGE("Only one message locally may appear every minute,"
" (%s)", strerror(errno));
These new calls are guaranteed not to affect the value of a
non-zero errno to simplify logging of errors. However, the
ALOGE calls in the above examples may update the errno value
upon their return.
Test: gTest liblog-unit-tests --gtest_filter=liblog.__android_log_ratelimit
Bug: 33535908
Change-Id: Id8cc192fc7d14504ffd418933cf88ae945c089f2
>3 times faster access for lookup functions. Code simplification and
C++ coding style conversion. android_lookupEventTagNum has been
expanded to utilize the Format match, and then drops down to just
Tag match.
Performance testing on Hikey, all services stopped, shell only access,
CPU not locked had good repeatability.
$ adb reboot ; adb wait-for-device ; adb shell su root stop
Before this adjustment using binary search:
$ /data/nativetest64/liblog-benchmarks/liblog-benchmarks BM_lookupEvent*
iterations ns/op
Precharge: start
Precharge: stop 231
BM_lookupEventTag 10000000 153
BM_lookupEventTag_NOT 20000000 139
BM_lookupEventFormat 10000000 153
After this adjustment strictly using std::unordered_map lookup:
$ /data/nativetest64/liblog-benchmarks/liblog-benchmarks BM_lookupEvent*
iterations ns/op
Precharge: start
Precharge: stop 231
BM_lookupEventTag 50000000 56
BM_lookupEventTag_NOT 100000000 21
BM_lookupEventFormat 50000000 56
Test: gTest logd-unit-tests and liblog-unit-tests, liblog-benchmarks.
Bug: 31456426
Change-Id: I1cfc3d641c3ea9ab08d455efe0c844d3a6381440
NB: recent activities on event_tag_map.c were to allow
this transition to be unremarkable.
Test: compile
Bug: 31456426
Change-Id: I13e1771a001874244711a5ac3023861c2a5f8c7d
Robustness fortification. Deal with gtest failure if event-log-tags
file is updated, adb sync for example, and thus will report
Shared_Dirty instead of Shared_Clear. Add check for Private_Dirty
and Private_Clean being clear at zero too as a bonus.
Test: gTest liblog-unit-tests --gtest_filter=liblog.event_log_tags
Bug: 31456426
Change-Id: Id912929fdd65499aa532a15653ef387a997a5054
Some of the __BIONIC__ checks were actually looking for an android
device. Host bionic has __BIONIC__, but not __ANDROID__.
Bug: 31559095
Test: Test linux_bionic compile
Change-Id: Ia0ca36fc9486fe8ea7e5d1b2e26ab0491903723c
Utilize liblog-unit-tests as a CTS test
Replacement for EventLogTest#testWriteEventWithOversizeValue
along with some other mandatory logging functionality.
Test: run cts tests
Bug: 26235244
Change-Id: I71f549bd0b93f6c41e1f91565b0fbddf042e4991
Mark Salyzyn