(cherry pick from commit ec4f5c776d)
Bionic getuid() and getpid() calls cache to reduce the need to
perform a syscall, and also reset their own cache after a vfork().
No more need for liblog to be performing this flawed cache operation.
Bug: 30085794
Change-Id: I70feed8bff0ddd919c2885a348ba67b14ddc0e0d
(cherry pick from commit efe8ecc1d9)
- periodic failures in apct, dropped second serial test
in refresh_cache, trusting check_cache or global.
- The retry loop to see if is_loggable recovers of 1000
was hiding subsequent tests, drop to 10 retries.
- On the whole, the average performance remains the same.
Bug: 25792367
Change-Id: I4110440ef46671d7a1c128689bde623808bed04f
(cherry pick from commit 7ef5249afa)
- add optimized & cached LIBLOG_HIDDEN __android_log_is_debuggable()
- check when writing, either LOG_ID_SECURITY, SafetyNet or
debuggable when pushing content to the pmsg buffer.
Bug: 27566046
Change-Id: I85f1b55ec329b38e00f4183836b6ed53046c323d
(cherry pick from commit c33103c440)
- Secure LOG_ID_KERNEL in writer
- Secure LOG_ID_SECURITY in reader and writer
- if writer transport says not available, do not write to that log id
Bug: 27566046
Bug: 27896341
Change-Id: If63a78a56fb94adfbf9979454c4cadb81af45c19
(cherry picked from commit 083c53462a)
Try to print as much content as possible should the application
logging only submit content as part of a tag with an empty message.
We search for the first non-printable ascii character in the tag, in
order to split it up for printing.
Applications (such as com.yahoo.mobile.client.android.weather) that
malform their log messages will no longer be punished by truncating
the content, but this should never be considered advocacy for their
bad behavior.
Bug: 27585978
Change-Id: Idb0680e8d6a6ad2bef5150661905acccb5b70afb
(cherry pick from commit 864e8e80e4)
- This is considered an Android Private function, not exported
for general use.
- goal is to retreive a file's content from a series of log
messages from pmsg, to be retrieved after a reboot for
transfer to a persistent location.
- files are presented in reverse sorted order, first based on
_any_ numerical content, then by alphanumeric order.
- Add a gTest for this function, relies on gTest for
liblog.__android_log_pmsg_file_write from prior to reboot.
Bug: 27176738
Change-Id: If37ef423009bd28b598b233af3bccef3429bdc22
(cherry pick from commit d4b061bde2)
- This is considered an Android Private function, not exported
for general use.
- goal is to record a file's content into a series of log
messages into pmsg, to be retrieved after a reboot for
transfer to a persistent location.
- filename reference is converted to a tag-unique
"<dirbase>:<filebase>".
- buffer and length representing the filename contents are
recorded, along with a sequence number placed into the nsec
time field to ANDROID_LOG_PMSG_FILE_MAX_SEQUENCE.
- Add a gTest for this function.
Bug: 27176738
Change-Id: If93df3ae8bfc1bb75516d4a1fd8dae0301af644b
(cherry pick from commit 018a96d03f)
Create config_logger, logger and logger_read to house the log
interfaces. Add fake_logger, logd_logger and pmsg_logger to
house the write and read transports. Allows for an easier and
direct path to add new transports to the library.
SideEffects: None, logger benchmark performance unaffected
Bug: 27176738
Bug: 27405083
Change-Id: I01b38637334a5242905c8c89f6ab0a92e2540008
(cherry pick from commit 50af7f8b1d)
- We actually are logging in a signal handler, the title is bluster
to remind developers. It is not a reliable path though as it can
lock up. Our goal is to minimize the chances of a lockup regardless
in the name of stability only. The test remains to catch regression
in the name of code quality and reliability. Expected to be
>99.999% reliable.
- Add a new _correct_ duplicate test that uses signal to release a
semaphore to a thread that performs the task. This path is expected
to be 100% reliable.
Bug: 27405083
Change-Id: Ibb7cf4b13e34ebfac2db2af8724b7db7a27f81a8
(cherry pick from commit be1d3c21b5)
- replace <sys/cdefs.h> with local "log_cdefs.h" which
fortifies and expands definitions, adding LIBLOG_ABI_PUBLIC,
LIBLOG_HIDDEN, LIBLOG_ABI_PRIVATE and LIBLOG_WEAK.
- clearly tag each interface as LIBLOG_ABI_PUBLIC, LIBLOG_HIDDEN,
LIBLOG_ABI_PRIVATE, LIBLOG_WEAK or static depending on scope
- Add -fvisibility=hidden to ensure nothing else leaks
- some code standard adjustments
Bug: 27566046
Change-Id: Ic14033c4e6d833d973beb035ddc1c6134fb35a3f
(cherry pick from commit 0085a135b9)
Use static space for long lived allocations as they
will appear to act like a memory leak. Resort to a
larger on-stack iovec to reduce the chances of an
allocation. Fix bug in writer where not enough size
was available for "security" buffer name. Minor
transitions to more consistent coding style.
Bug: 27107691
Change-Id: I68c918e7b916b1ae3b04829d48b3eddaa0a7e739
pid_t is 64-bit in 64-bit mingw, but the windows process/thread
functions return a DWORD(uint32_t). Instead of promoting to a pid_t and
fixing the format strings, just use a uint32_t to store the values.
android_thread_id also cannot be a 64-bit pointer, so for windows just
force it to be a uint32_t.
libutils/ProcessCallStack only works under Linux, since it makes heavy
use of /proc. Don't compile it under Windows or Darwin.
Bug: 26957718
(cherry picked from commit 86cf941c48)
Change-Id: I8d39d1951fea1b3011caf585c983e1da7959f7c0
(cherry pick from commit 81f407be36)
Switch to the event list library to compose the associated event.
SideEffects: Instead of composing event on a stack buffer of 512
bytes in size, a PAGE is allocated temporarily.
Bug: 27356456
Change-Id: Ic15a87f49385834c2287ed82c26439b2c5eb4f77
(cherry pick from commit 5cecedc6e8)
Add a test to confirm exact expected content using the testframe
setup for the events log handler. Remove dependency on 512 truncation
in liblog->
android_errorWriteWithInfoLog__android_logger_list_read__data_too_large
to something more liberal.
Bug: 27356456
Change-Id: I8a53ad3a16cf16b14856efe5b95417e857c7e09b
(cherry pick from commit 67d7eafd56)
android_log_write_string8_len(android_log_context ctx,
const char *value, size_t maxlen)
Caps the supplied string to a maxlen length. Alter API to handle
a NULL pointer for the value string for this and
android_log_write_string8() and instead of returning -EINVAL,
act like a null string "" was supplied to preserve the list
location. API is also changed to report the number of characters
actually placed into the android_log_context.
Bug: 27356456
Bug: 19235719
Change-Id: I6a03d405eac1d741555dd05555513ec691e7a46e
(cherry pick from commit bd1ad049b2)
Based off an initial request and effort by williamluh@google.com
- Added the following functions:
* Composing and Writing:
android_log_context create_android_logger(uint32_t tag)
int android_log_write_list_begin(android_log_context ctx)
int android_log_write_list_end(android_log_context ctx)
int android_log_write_int32(android_log_context ctx, int32_t value)
int android_log_write_int64(android_log_context ctx, int64_t value)
int android_log_write_string8(android_log_context ctx, const char *value)
int android_log_write_float32(android_log_context ctx, float value)
int android_log_write_list(android_log_context ctx, log_id_t id)
* Reading and Interpreting:
android_log_context create_android_log_parser(const char *msg, size_t len)
android_log_list_element android_log_read_next(android_log_context ctx)
android_log_list_element android_log_peek_next(android_log_context ctx)
* Destroy context used above:
int android_log_destroy(android_log_context *ctx);
- Added unit gTests
We moved implemented android_log_buffer_to_string() to the test since
it is an alternate for already existing logprint functionality.
Please move into liblog should it be of some common use, otherwise
as is it is a good means of stessing the reading and interpreting
handlers.
Signed-off-by: Mark Salyzyn <salyzyn@google.com>
Bug: 19235719
Change-Id: I4aa1927e8e6a75f0a129d15a27c891cf1ccd4f5c
(cherry pick from commit b525884edb)
- Fix bug in fake_log_devices when it can not allocate memory
failing to use the full on-stack buffer as backup.
- remove superfluous code.
Bug: 27265662
Change-Id: I97b6cca5f4ce8ecad9beb3a08353de596d6a4ad1
strdup to a static pointer is unnecessary, strlcpy into a static buffer
instead. Avoids allocations in the ALOG* path, allowing liblog to be
used by libmemleak.
Change-Id: Ie0986da27c1fc5eb8ce4ebb076b513be8e1ee676
- limit AID_SYSTEM uid or gid to read security buffer messages
- adjust liblog tests to reflect the reality of this adjustment
To fully test all security buffer paths and modes
$ su 0,0,0 /data/nativetest/liblog-unit-tests/liblog-unit-tests --gtest_filter=liblog.__security*
$ su 1000,1000,1000 /data/nativetest/liblog-unit-tests/liblog-unit-tests --gtest_filter=liblog.__security*
$ su 2000,2000,2000 /data/nativetest/liblog-unit-tests/liblog-unit-tests --gtest_filter=liblog.__security*
ToDo: Integrate the above individually into the gTest Q/A testing
Bug: 26029733
Change-Id: Idcf5492db78fa6934ef6fb43f3ef861052675651
Allow _one_ empty line to get through before evaluating
further. Add __android_log_bswrite testing for content
and contentless cases; checking for propagation and for
expected printing. As for printing which is fixed here,
security variants cover the same code paths as the events.
Bug: 26646213
Change-Id: I484718aa604e0a00afde4c34a00e87468ea93aa5
Callers will not guarantee that they can or will ratelimit, we need to
retain the ability to blacklist snet_event_log as a result.
This reverts commit 6aa21b225d.
Bug: 26178938
Change-Id: Ibf47d2e23a84c56f5f72d02312c698df7ff2b601
Dangerous bridge to cross to whitelist, who is special, who is not?
Rationalized as these events are used to catch exploits on platform.
As it stands no one should be allowed to block any messages in the
security context, not even for development purposes.
Bug: 26178938
Change-Id: Ibdc76bc0fe29ba05be168b623af1c9f41d7edbd2
Add __android_log_is_loggable() checking for all buffers except
LOG_ID_SECURITY. Return -EPERM if blocked. Since we are sniffing
the log tag, check validity and return -EINVAL.
NB: Try not to call __android_log_is_loggable() in native code within
a signal handler. Both here, and in the system properties, there
are locking paths that are not guaranteed to play well in that
environment. This has also been the case for the log writer path
even before this change. All attempts have been made to use trylock,
and to use a more expensive code path when contention occurs rather
than lead to deadlock.
Bug: 19544788
Bug: 26178938
Change-Id: I98738c662f6328189a6703251eb8721a05e956f9
Some compilers erroneously see uninitialized use, even despite all
accesses being behind !not_locked. Confirmed initialization does not
affect expected performance.
Bug: 26178938
Bug: 26029733
Bug: 17760225
Change-Id: Ib36ed8dd2c4b196ca84ef79a9531625dcee77e15
- Deal with __android_log_is_loggable inside signal handler by
treating a lock contention system call as more costly than reading
the associated property directly. Rather waiting around in a
contended stat to hit cache.
- Check both the individual known __system_property_serial() and
global __system_property_area_serial() to detect updates or
additions to the properties to respond in the most aggressive and
timely manner. __android_log_is_loggable() return at max CPU
clockrate on a N9 in 61ns.
- Craft a common do_cache2 inline that adds the above to the other
functions that utilize cache handling and preserves (within 3ns)
performance in android_log_clockid() and __android_log_security().
These functions return at max CPU clockrate on a N9 in 23ns.
Bug: 19544788
Bug: 25693940
Bug: 25792367
Bug: 26178938
Change-Id: I9cd94598f5c558e946b93977ad3714a4b03d0422
Primarily gives access to the Chattiest TIDs and TAGs
associated with a pid.
Has a secondary effect of allowing us to pull out the
command line, comm and in some cases the associated
PACKAGE for a specific pid while the logs are still
present even if the executable is gone.
Bug: 26029733
Bug: 21615139
Change-Id: I1ea63165a680a9318360579b70b1512078ed5682
This test tells us that scatter-gather (writev instead of write)
carries a small ~2% penalty. Unaligned buffer carry a similar
additional penalty. On N9 it takes 3us to 22us _just_ to write
the pmsg logs, depending on size. Some assumptions about the
socket read and write performance for the main logging can be made
from tracking these results and should improve design decisions.
Bug: 18771697
Bug: 23685592
Change-Id: Id4d64c449140e4f39078c62b0097e403df91fe0c
If we adjusted the maximum log payload, Resolve compile
issues resulting from the changes (gTest).
Bug: 25996918
Change-Id: I672b0f4d9d4a1394a5b2e27bf81a5e906bf92a10
