Michal Bloch [Tue, 9 Jun 2020 15:51:16 +0000 (17:51 +0200)]
Fix infinite sleeps in low-log environments
Change-Id: I3c3688d5bc05d648dfe2a38f349d2b403108a28a
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Michal Bloch [Tue, 9 Jun 2020 11:47:09 +0000 (11:47 +0000)]
Merge "Add current usage accessor to the daemon protocol" into tizen
Karol Lewandowski [Tue, 9 Jun 2020 11:46:04 +0000 (11:46 +0000)]
Merge changes Icf47adb1,Iec35237c into tizen
* changes:
Remove another lie
libdlogutil: allow sleeping forever
Agnieszka Baumann [Thu, 7 May 2020 09:16:52 +0000 (11:16 +0200)]
Add current usage accessor to the daemon protocol
Change-Id: Ia7446796068a8e06965fd6e8e44e416f1f066888
Agnieszka Baumann [Thu, 14 May 2020 08:40:41 +0000 (10:40 +0200)]
Clarify function names: getsize → get capacity
Change-Id: I28955dd2359498370756555cbd88d6ff5a1fb7e1
Mateusz Majewski [Tue, 9 Jun 2020 07:25:52 +0000 (09:25 +0200)]
Remove another lie
This time, I am the liar...
Change-Id: Icf47adb18ad1bd95ab4ecc16b51df200d829210a
Michal Bloch [Mon, 8 Jun 2020 18:43:32 +0000 (20:43 +0200)]
libdlogutil: allow sleeping forever
Change-Id: Iec35237c9a6ffbcf6b74f689675127617cbdddef
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Mateusz Majewski [Fri, 22 May 2020 12:49:47 +0000 (14:49 +0200)]
Improve server timestamp handling
Change-Id: I2242f8ded640817cc7e627743056a37d0811bd38
Mateusz Majewski [Fri, 5 Jun 2020 09:46:26 +0000 (11:46 +0200)]
Rename the test_libdlogutil.c file to ..._legacy.c
The only purpose of this is to make the next commit kinda readable.
Change-Id: I3e0e6472df46f6ce40e5a2e08dd7592e09e6a8fd
INSUN PYO [Fri, 5 Jun 2020 00:45:19 +0000 (09:45 +0900)]
Clean up: delete code that add and subtract zero
Change-Id: I3f6b20601f8d964717857aa3e128515dd043938b
Mateusz Majewski [Tue, 28 Apr 2020 10:45:08 +0000 (12:45 +0200)]
Make all the files in the dynamic directory respected
Change-Id: Id1f881bc8a22169ada9df672d291f080d78df3b0
Michal Bloch [Tue, 2 Jun 2020 15:56:23 +0000 (17:56 +0200)]
Fix garbage data in long-format strings
Change-Id: Ie271b8ca58f0cc018fb13b306f66f0fdf02af2e4
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Michal Bloch [Tue, 2 Jun 2020 11:09:28 +0000 (13:09 +0200)]
Initialize a value (fixes uninitialized access)
Change-Id: I1984ea4fe7c3ec1748d73ca546c8e550e31c9224
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Michal Bloch [Mon, 1 Jun 2020 10:17:53 +0000 (12:17 +0200)]
Release 6.0.12
This release allows the use of hybrid mode on pipe backend.
Change-Id: I74a4144996b1a34798c704f83f482227a91bfdec
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Mateusz Majewski [Thu, 28 May 2020 07:18:57 +0000 (09:18 +0200)]
Fix an ASAN complaint
Change-Id: I8a33a66a75162fb76d3994308545f2642ad921da
Mateusz Majewski [Fri, 29 May 2020 13:00:03 +0000 (15:00 +0200)]
Improve libdlogutil timeout handling
The current (lib-internal so far) API allows one to specify how much one
should wait for logs. Sometimes however, we should perform actions
despite this. In such cases, we should limit the timeout to a maximum
value. This commit implements this. Also, it allows us to remove the
unneccessary 100 ms fixed timeout.
Change-Id: Ia9ae0872098847e79c60607438bbc2a9438b40f8
Michal Bloch [Tue, 12 May 2020 12:28:47 +0000 (14:28 +0200)]
Allow hybrid mode for pipe backend
Partially reverts
44238daabe075f5b0c4852c61cfb9ee7f269bf97.
Change-Id: I49f75475f9285c24fde4f473b5e767c4953535be
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Mateusz Majewski [Mon, 18 May 2020 10:15:47 +0000 (12:15 +0200)]
logger: add socket connections to the socket epoll
Change-Id: I2ca5f500b02818454fa69c5ab4cfcfa4f8d3b3b9
Michal Bloch [Fri, 8 May 2020 19:19:52 +0000 (21:19 +0200)]
logger: add sockets to epoll
Change-Id: Ie6d4b87f62948973b2c95b9d735eeb265f9430dc
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Michal Bloch [Fri, 8 May 2020 16:10:05 +0000 (18:10 +0200)]
logger: fd entity manipulators work on epoll metadata
Change-Id: I26dab246c071bae6a6004ddc8516bf4e342f2ba2
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Michal Bloch [Thu, 7 May 2020 19:39:21 +0000 (21:39 +0200)]
logger: handle the (empty) socket epoll instead of sleeping
Change-Id: I666632213a1e99418a35ab5cde0d6ef1904e3650
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Michal Bloch [Thu, 7 May 2020 19:24:58 +0000 (21:24 +0200)]
logger: add a (so far unused) epoll for socket FDs
Change-Id: I87d93f99d4b76cf6759901822965d28964edcaa1
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Michal Bloch [Thu, 7 May 2020 19:16:47 +0000 (21:16 +0200)]
logger: bind epoll meta lifetime to logger
Change-Id: Ifddd234c30f215ac501faff550a24cea60838f07
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Michal Bloch [Thu, 7 May 2020 19:03:08 +0000 (21:03 +0200)]
logger: extract epoll handling function
Change-Id: I67e37cdceeab3a65643c6f8fc5872fef25fea35a
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Michal Bloch [Thu, 7 May 2020 17:26:21 +0000 (19:26 +0200)]
logger: extract epoll metadata to a struct
Change-Id: Idaedf8407fdb24260ba66f7bf39455c883a52e3c
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Michal Bloch [Thu, 7 May 2020 16:40:23 +0000 (18:40 +0200)]
logger: extract epoll dispatcher
Change-Id: I4202a56d7ee3b41ea9b2fe2c61d85112219ebff3
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Mateusz Majewski [Fri, 22 May 2020 10:51:32 +0000 (12:51 +0200)]
Fix a typo in Makefile
Change-Id: I8d382b0c620ffcb9a7e1373b550b1560c93f47bc
Karol Lewandowski [Thu, 21 May 2020 19:49:15 +0000 (19:49 +0000)]
Merge "logger: optimize reader time book-keeping" into tizen
Agnieszka Baumann [Thu, 16 Apr 2020 08:31:19 +0000 (10:31 +0200)]
sort_vector divided into positive and negative tests
Change-Id: Ibd6617f7d3e6d55deda7be859c33950d16734305
Agnieszka Baumann [Thu, 30 Apr 2020 16:32:32 +0000 (18:32 +0200)]
tests: logutil.c divided into positive and negative
Change-Id: I2f0c1078939005c8f95ec1cf36b2426d70347bba
Mateusz Majewski [Tue, 19 May 2020 10:31:35 +0000 (12:31 +0200)]
Add CPU usage measurement
Change-Id: I123a926433fd456e1860dd6cc9572a3ffe0e2d0c
Michal Bloch [Tue, 12 May 2020 12:51:09 +0000 (14:51 +0200)]
logger: optimize reader time book-keeping
Change-Id: I8449236f1c15261aa7550f175de1942b519ecbc1
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Hyotaek Shim [Wed, 13 May 2020 11:09:28 +0000 (20:09 +0900)]
Apply lazy_polling_sleep_ms=100
Change-Id: Ie2ce6ab2aae3b777306d19c9edac860a607e30ab
Signed-off-by: Hyotaek Shim <hyotaek.shim@samsung.com>
Michal Bloch [Mon, 4 May 2020 10:49:54 +0000 (12:49 +0200)]
Hybrid mode config: only use for Android Logger backend
Change-Id: Ifaac28dd807262cc27d0b46f9bffa56e734085f3
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Michal Bloch [Mon, 4 May 2020 10:30:39 +0000 (12:30 +0200)]
Revert "Revert "logger: introduce lazy/hybrid approach to early logging""
This reverts commit
064c096da525adedf0766d0b91de16b011841676.
Change-Id: I2cf0fecf8437766fe774a4f702622e3d8c0d0109
Mateusz Majewski [Mon, 4 May 2020 13:18:19 +0000 (15:18 +0200)]
Add secure logs test
Change-Id: I1157dae0c859b1b7b3e035e9c186efa553cc5aa8
Mateusz Majewski [Mon, 4 May 2020 12:36:46 +0000 (14:36 +0200)]
Add secure log flag to dlogsend
Change-Id: I1dec52e921afdbbf5a0ff402d55516117f614cfd
Michal Bloch [Tue, 5 May 2020 12:07:27 +0000 (14:07 +0200)]
Fix first secure log ignoring the config
Change-Id: Ifb575f0077fd35556e4263ef002535e274b95187
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Michal Bloch [Thu, 30 Apr 2020 16:23:24 +0000 (18:23 +0200)]
libdlog: add `enable_secure_logs` config entry
If it is set to 0, __dlog_sec_print will just
return 0. Defaults to 1.
Change-Id: Ie3a00c5aa25bd4035444ae27383d649acf0e1d23
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Hyotaek Shim [Mon, 4 May 2020 06:22:25 +0000 (06:22 +0000)]
Revert "Revert "Fix a bottleneck in log processing""
This reverts commit
161899c8835dbacedde7eafc36cbc94725dffddb.
Change-Id: I92027e73240b32569572ad244cee50d0c452f155
Hyotaek Shim [Mon, 4 May 2020 06:15:56 +0000 (06:15 +0000)]
Revert "Revert "Move some more getters around""
This reverts commit
146458f9b77c55345efbb3419ee82e423c66e993.
Change-Id: I20602b26aa154a277a6735209036edb658a42569
Hyotaek Shim [Mon, 4 May 2020 03:55:55 +0000 (03:55 +0000)]
Revert "Fix a bottleneck in log processing"
This reverts commit
e2144aefea4b4091a4f653c59f0fc4aeecf7cb42.t
Change-Id: I3333f245fb836316b582ac0c82ad987dfde48c84
Hyotaek Shim [Mon, 4 May 2020 03:04:37 +0000 (03:04 +0000)]
Revert "logger: introduce lazy/hybrid approach to early logging"
This reverts commit
4fe8f45ca0edf377bc5ef4fde82346e45c1a1145.
Change-Id: I47e25decfb352b6534cce03af3eb85414be43f6f
Hyotaek Shim [Mon, 4 May 2020 03:09:53 +0000 (03:09 +0000)]
Revert "Move some more getters around"
This reverts commit
f97985fd478b5e8229d58e84620349889ad9cd69.
Change-Id: I2313ea90bc0213aade119d2aff5aca0173c8ae72
Michal Bloch [Thu, 30 Apr 2020 17:39:29 +0000 (17:39 +0000)]
Merge "Fix a mistake" into tizen
Mateusz Majewski [Wed, 29 Apr 2020 15:09:56 +0000 (17:09 +0200)]
Fix a mistake
Oof!
Change-Id: Ia3a1f227645dce7b674792a6fee4298b50755703
Mateusz Majewski [Wed, 29 Apr 2020 14:30:01 +0000 (16:30 +0200)]
Make logctl lock the config file
Somebody pointed out that, if two logctls are launched simultanously,
it may result in a race condition and one of the changes dropped. Our
minds quickly flocked to a solution in which we use exclusive locks to
lock the file on write. This means we will always see the file in a
consistent state.
Change-Id: I5062bdcddc6c880bfd11ec66dad9cc4927e0536b
Mateusz Majewski [Wed, 29 Apr 2020 14:18:39 +0000 (16:18 +0200)]
Rewrite logctl to use open + fdopen instead of fopen
Sounds useless, but allows us to avoid a hack in which we open the file
in append mode despite only ever intenting to read from it. Also, this
gives us access to the fd, which will prove useful in the future.
Change-Id: I8f2e32decf363e7bcc4c726d0f3eeedbe6c5ce4d
Michal Bloch [Tue, 28 Apr 2020 09:46:47 +0000 (11:46 +0200)]
Release 6.0.11
This realease brings the lazy polling mode as a solution to CPU usage.
Change-Id: I776d9163505b985e738c537b5e1e62e6da86bd0a
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Michal Bloch [Wed, 22 Apr 2020 18:28:09 +0000 (20:28 +0200)]
logger: introduce lazy/hybrid approach to early logging
This mode is introduced in an attempt to limit CPU usage during boot time.
The logger will now wake up exactly after (instead of up to) the epoll delay
for some configurable amount of time (by default 10s). Ideally this would
just be waiting on BootingDone but the daemon doesn't handle DBus yet, which
would be a larger undertaking. After that delay, it switches back to usual
epoll behaviour (waking up on each log, limiting log loss).
Change-Id: I55555978902d18b2d922663a24fe1caeb6d330e8
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Michal Bloch [Tue, 28 Apr 2020 09:19:32 +0000 (11:19 +0200)]
Fix a bottleneck in log processing
Change-Id: I64cd95f2f6d6f1ba8b9967e4978f43e02bfdb311
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Mateusz Majewski [Tue, 14 Apr 2020 11:05:13 +0000 (13:05 +0200)]
Move some more getters around
Change-Id: I99b1cfc0964be3658e3059102dc00f11fe89f701
Mateusz Majewski [Wed, 8 Apr 2020 12:39:35 +0000 (14:39 +0200)]
Fix logger backend integration test
When we wrote the test, we assumed that the size of the buffer on the
logger backend is constant. This assumption turned out to be naïve and
completely wrong; the buffer on the emulator is much smaller.
Change-Id: Ied05dc9930688108c727969fb189e5cd3fb739bf
Michal Bloch [Wed, 22 Apr 2020 18:18:08 +0000 (20:18 +0200)]
logger: allow polling time control through config
"epoll_time_ms" entry.
Change-Id: I0c911978902d18b2d922663a24fe1caeb6d330e8
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Michal Bloch [Fri, 10 Apr 2020 14:16:35 +0000 (16:16 +0200)]
logger: allow disabling KMSG buffer through config
Done through the new "handle_kmsg" entry (defaults to 1).
Change-Id: I399605ae9a6513b233f0613f5330b4d4d1fa412d
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Mateusz Majewski [Thu, 9 Apr 2020 16:42:23 +0000 (18:42 +0200)]
Fix whitespace
Change-Id: Ibe66948c041cc1277e43d613bd6425503a6b5d04
Michal Bloch [Wed, 22 Apr 2020 10:53:03 +0000 (10:53 +0000)]
Merge changes I4ae42ed3,Ief2d8cc6 into tizen
* changes:
libdlogutil: make the epoll check stateful
libdlogutil: put epoll check in the state (nominally)
Mateusz Majewski [Mon, 20 Apr 2020 08:41:35 +0000 (10:41 +0200)]
Cover queued_entry_timestamp.c completely
Change-Id: I26471ddb94a04a716d08ec8fc71484c9c4d31982
Mateusz Majewski [Fri, 17 Apr 2020 12:59:18 +0000 (14:59 +0200)]
Cover logctl.c completely
Change-Id: I3a6a17976a463e2598f4d5f016e1c155481062f3
Mateusz Majewski [Fri, 17 Apr 2020 12:53:25 +0000 (14:53 +0200)]
Cover loglimiter.c completely
Change-Id: I847947c8f12b07f1013dc7cdca50a3b84c41f4c3
Michal Bloch [Thu, 16 Apr 2020 16:38:29 +0000 (18:38 +0200)]
libdlogutil: make the epoll check stateful
Change-Id: I4ae42ed3a94f644e7e63f55c79c17c62edd0f52b
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Michal Bloch [Thu, 16 Apr 2020 11:35:28 +0000 (13:35 +0200)]
libdlogutil: put epoll check in the state (nominally)
It's not really stateful yet, just the refactors needed to make it so.
Change-Id: Ief2d8cc6a9e89911eb775fea3e838e8fd8997990
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Agnieszka Baumann [Tue, 14 Apr 2020 15:25:45 +0000 (17:25 +0200)]
queued_entry divided into positive and negative tests
Change-Id: Ie41b5f119bca98f209a237e401e79dfaaf59d40a
Hyotaek Shim [Tue, 14 Apr 2020 12:26:38 +0000 (21:26 +0900)]
Apply Nice=1 to dlog_logger.service
Change-Id: I57961d4a336e2fc9736bc0ae5521b2f71090482c
Signed-off-by: Hyotaek Shim <hyotaek.shim@samsung.com>
Agnieszka Baumann [Tue, 7 Apr 2020 13:24:22 +0000 (15:24 +0200)]
fdi_pipe divided into positive and negative tests
Change-Id: I46bfa31f37391a478691428909b529e70cc89309
Agnieszka Baumann [Mon, 30 Mar 2020 11:22:53 +0000 (13:22 +0200)]
fdi_logger divided into positive and negative tests
Change-Id: I73c6729446d2025269c5517f4457fc68177fc4fa
Mateusz Majewski [Wed, 8 Apr 2020 12:11:50 +0000 (14:11 +0200)]
Remove another lie
Change-Id: I232d4b9883d66c496fd77a95430a3271f35ce1d5
Mateusz Majewski [Fri, 27 Mar 2020 14:08:35 +0000 (15:08 +0100)]
Move dlogutil_entry_s accessors
They are mostly unrelated to the rest of the library codewise. If we
move them somewhere else, rest of the code that might want to use them
won't have to pull in everything needed by the library (for example, the
capability library).
This is a temporary solution, in infinity we should have nice accessors
for dlogutil_entry_s all available in one spot and the whole DLog
accesing the struct only using those accessors. Right now, however, it's
nothing but a pipe dream since we have more important things to do than
pitful refactors.
Change-Id: Ifa05d056437e5d2143eb3a6fb25ad9247b08b1d6
Agnieszka Baumann [Thu, 2 Apr 2020 12:48:49 +0000 (14:48 +0200)]
fd_info divided into positive and negative tests
Change-Id: I36f179f6c0ac1f3f5fb2d9fb7c8bd07d198c43d7
Michal Bloch [Mon, 6 Apr 2020 07:41:42 +0000 (09:41 +0200)]
Release 6.0.10
This release fixes a memleak and brings
further refactors to libdlogutil internals.
Change-Id: I510d45297b1abe2dd97de4e345664c69ebb5a9df
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Michal Bloch [Fri, 3 Apr 2020 17:13:38 +0000 (19:13 +0200)]
libdlogutil: fix a memory leak
Change-Id: I8cc20443c9cd956e2f3fa00e0e2f4cc9a6c8682e
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Michal Bloch [Fri, 3 Apr 2020 17:12:57 +0000 (19:12 +0200)]
libdlogutil: leak safety
Change-Id: I9c18dfa4c0ba91b98e50ff5f43f111baa332680b
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Michal Bloch [Tue, 31 Mar 2020 17:38:09 +0000 (19:38 +0200)]
libdlogutil: split off flush target setting
Change-Id: Ic4cefeb97d836eaf41eb355c1907f963018e5bfd
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Michal Bloch [Tue, 31 Mar 2020 17:37:22 +0000 (19:37 +0200)]
libdlogutil: split off buffer replenishment
Change-Id: Ic4cefeb97d836eaf41eb355c1907f963018e5bfc
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Michal Bloch [Tue, 31 Mar 2020 17:35:34 +0000 (19:35 +0200)]
libdlogutil: split off buffer removal
Change-Id: Ic4cefeb97d836eaf41eb355c1907f963018e5bfb
Michal Bloch [Tue, 31 Mar 2020 17:34:26 +0000 (19:34 +0200)]
libdlogutil: split off continuous flushing
Change-Id: Ic4cefeb97d836eaf41eb355c1907f963018e5bfa
Michal Bloch [Tue, 31 Mar 2020 16:29:18 +0000 (18:29 +0200)]
libdlogutil: simplify -t mode sorting
Change-Id: I709b29b2b510d49410b790c648fa9eabe8802c18
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Michal Bloch [Thu, 2 Apr 2020 12:52:26 +0000 (14:52 +0200)]
libdlogutil: extract a common check to a function
Change-Id: I220e891e97a57bf13f775f1b4a928333b24710d2
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Michal Bloch [Fri, 27 Mar 2020 19:29:54 +0000 (20:29 +0100)]
libdlogutil: support timeout param (nominally)
Change-Id: I405093fb559b4f1935e404aa8ae3f597fcdf6cb0
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Michal Bloch [Tue, 31 Mar 2020 15:13:01 +0000 (15:13 +0000)]
Merge changes I138ad4f5,Ib6bc5b61 into tizen
* changes:
Const correctness
libdlogutil: remove callback/userdata from the state
Michal Bloch [Fri, 27 Mar 2020 19:52:10 +0000 (20:52 +0100)]
Const correctness
Change-Id: I138ad4f5b3f76006c88319a5c204d674de1d9cd9
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Agnieszka Baumann [Wed, 25 Mar 2020 09:21:52 +0000 (10:21 +0100)]
limiter divided into positive and negative tests
Change-Id: I25b01d7c093d80cf01bd6585a8ef526ceda915ef
Michal Bloch [Fri, 27 Mar 2020 14:39:21 +0000 (15:39 +0100)]
libdlogutil: remove callback/userdata from the state
Change-Id: Ib6bc5b618dd1fd95c6a77d6bf78bd5eda93ac91d
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Mateusz Majewski [Tue, 24 Mar 2020 15:00:13 +0000 (16:00 +0100)]
libdlogutil: refactor mainloop to work on single logs
The main worker function now calls the callback on a single log.
Backward compatibility is maintained through a wrapper.
Change-Id: Iaacc48977c360b12bba3b06ab0a51fe2ba18bf3f
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Michal Bloch [Wed, 25 Mar 2020 14:20:43 +0000 (15:20 +0100)]
libdlogutil: put logs into vector before polling
Eventually we want the API to return a single log. With poll at the
front, it would be done for every log returned, which is unnecessary.
Additionally, remove the "poll returned nothing" condition which
allowed all buffers to be drained at the same time. It shouldn't
be helping much in practice and can be simplified away.
The callback is temporarily not allowed to return 1. This will go
away in the final version of the API.
Change-Id: I4ccb0a613a902484303f651830a946ccc8718130
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Michal Bloch [Wed, 25 Mar 2020 13:38:34 +0000 (13:38 +0000)]
Merge "test_ptrs_list divided into positive and negative tests" into tizen
Agnieszka Baumann [Mon, 23 Mar 2020 11:57:54 +0000 (12:57 +0100)]
test_ptrs_list divided into positive and negative tests
Change-Id: I1e3391f6c1568a8a0adf4398401a7e875dfe133c
Agnieszka Baumann [Tue, 24 Mar 2020 14:47:00 +0000 (15:47 +0100)]
syslog_parser divided into positive and negative tests
Change-Id: I05366c4f2f884cde56e3b552fa01d94e84bd398d
Mateusz Majewski [Thu, 12 Mar 2020 09:23:34 +0000 (10:23 +0100)]
libdlogutil: flushing stateful, unlooped
This is done in order to move closer to the goal of making that
function reentrant. The block of code handling flushing still
forms a loop but now it is going to be trivial to change it to
return a single log by changing a 'continue' into a 'return'.
The block of code moved to the beginning so that the 'continue'
can emulate a loop as before (i.e. without epoll and sorting
happening in between) since the function is not reentrant yet.
Change-Id: I95248cfd3e8132dd14310219e93e3be887903605
Michal Bloch [Thu, 19 Mar 2020 13:36:32 +0000 (14:36 +0100)]
libdlogutil: move extra options to the state
Change-Id: I257424c7e1c4ec8667747a0e00e0c0594321a6db
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Michal Bloch [Mon, 23 Mar 2020 17:30:18 +0000 (18:30 +0100)]
Add license boilerplate
Change-Id: I5c107e25c8abba1fc8b1099f66a90aa09fcb0a0c
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Agnieszka Baumann [Wed, 4 Mar 2020 16:13:37 +0000 (17:13 +0100)]
test_common divided into positive and negative tests
Change-Id: I35fb99d45fdec21892019dc6268b7e1ed2a79c54
Michal Bloch [Mon, 23 Mar 2020 12:16:19 +0000 (13:16 +0100)]
Enforce correct Android Logger message size
Change-Id: Iabff9351c39f5a20c0ca6993730b23cb7ab301ca
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Yunmi Ha [Thu, 19 Mar 2020 06:18:42 +0000 (15:18 +0900)]
Measure code coverage only when gcov option is given
Change-Id: I9b4e2abfbacfc96e51bdf4d4cd48dd32c1c699bf
Signed-off-by: Yunmi Ha <yunmi.ha@samsung.com>
Michal Bloch [Wed, 18 Mar 2020 12:54:40 +0000 (13:54 +0100)]
Release 6.0.9
This release brings the fix of a SIGBUS crash and various refactoring preparations.
Change-Id: I4b3dcbdcf3ccefe2de7b4c875b8873324f269df7
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Mateusz Majewski [Wed, 18 Mar 2020 09:42:42 +0000 (10:42 +0100)]
Remove some dead code
Right now, literally only purpose of those two functions is confusing
people.
Change-Id: I1b4e9a814b4b6589abd5215ac79749a49ab219e9
Mateusz Majewski [Wed, 18 Mar 2020 14:09:50 +0000 (15:09 +0100)]
Fix a dlogutil regression
Change-Id: I0529ab65e04e2a19c7e326a9014cdd1bc4c22752
Michal Bloch [Wed, 18 Mar 2020 12:53:34 +0000 (12:53 +0000)]
Merge "Don't wait until lazy init to block SIGPIPE" into tizen
Michal Bloch [Wed, 18 Mar 2020 11:58:59 +0000 (12:58 +0100)]
Commentary for previous
Change-Id: I9b5f3f54f6089b4a2aa1713c698e922504371c28
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
Semun Lee [Wed, 18 Mar 2020 06:46:45 +0000 (15:46 +0900)]
Fix SIGBUS in pipe backend
make log entry to be aligned by 4
Change-Id: Ibe17b0f7e99dfcdf823b4dc8d328f4c65b4c436a
Signed-off-by: Semun Lee <semun.lee@samsung.com>