Mathieu Desnoyers [Fri, 21 Aug 2015 23:11:56 +0000 (16:11 -0700)]
Fix: add missing ust lock around objd_table_destroy()
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Mon, 17 Aug 2015 16:50:34 +0000 (09:50 -0700)]
Fix: application exit race with pthread cancel
Listener threads can be cancelled with ust lock held, which can hang the
following ust cleanup routine, because tracepoint probe unregister needs
to take the ust lock.
Fix this by disabling pthread cancellation for the entire duration of
the ust lock.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Jérémie Galarneau [Tue, 7 Jul 2015 18:35:54 +0000 (14:35 -0400)]
Fix: Cleanup local_apps sock_info in lttng_ust_cleanup
LTTng-UST will deadlock after a fork while waiting on the
"constructor_wait" semaphore if local apps, handled the session daemon
running under the current UID, are disabled or "not_allowed".
This deadlock can be triggered by setting an infinite registration
timeout, clearing the HOME environment variable and launching an app
which calls FORK(3). This will cause setup_local_apps() to fail to
determine the local_apps sock_path, thus leaving
local_apps.allowed == 0.
This, in turn, would cause lttng_ust_cleanup to skip the cleanup
of the local_apps sock_info after a fork,
leaving local_apps.constructor_sem_posted == 1. This would cause
handle_register_done() in the child to skip over the decrementation
of sem_count and post of the constructor_wait semaphore.
Signed-off-by: Jérémie Galarneau <jeremie.galarneau@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Mon, 6 Jul 2015 21:09:32 +0000 (17:09 -0400)]
Fix: handle sys_futex EINTR and EWOULDBLOCK
Handling sys_futex EINTR allows us to retry waiting on the futex
immediately if we are interrupted by a signal without having to try
connecting to the session daemon.
This should not impact correctness though, as the wait scheme would
simply try to connect, fail, and try waiting again.
Also handle errors on FUTEX_WAKE.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Mon, 22 Jun 2015 14:30:02 +0000 (10:30 -0400)]
Fix: perform volatile load of tracepoint state
Because tracepoint state is updated by lttng-ust threads, and read by
application threads, we need to perform a volatile load of that state.
Not having the volatile load can cause the compiler to optimize away the
reload of that state, keeping a local copy instead. For instance, a main
program consisting of a loop could keep the tracepoint state on its
stack or in registers without ever reloading it from memory.
Perform a volatile load (CMM_LOAD_SHARED()) of the tracepoint state.
Add the matching volatile store (CMM_STORE_SHARED()) in tracepoint.c for
the state update.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Thu, 18 Jun 2015 17:37:47 +0000 (13:37 -0400)]
Version 2.5.5
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Wed, 10 Jun 2015 11:27:18 +0000 (13:27 +0200)]
Fix: function instrumentation ip context
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Wed, 10 Jun 2015 11:20:31 +0000 (13:20 +0200)]
Fix: liblttng-ust-dl ip context
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Wed, 10 Jun 2015 11:15:28 +0000 (13:15 +0200)]
Fix: pthread wrapper ip context
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Wed, 10 Jun 2015 11:15:09 +0000 (13:15 +0200)]
Fix: lttng-ust-malloc ip context
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Wed, 10 Jun 2015 11:14:08 +0000 (13:14 +0200)]
Allow TP_IP_PARAM arg name to be configured
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Tue, 9 Jun 2015 09:51:06 +0000 (11:51 +0200)]
Fix: fetch caller address from tracef()
Don't fetch the address within the ust trampoline, but rather the caller
address.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Thu, 7 May 2015 17:25:04 +0000 (13:25 -0400)]
Fix: building probe providers with C++ compiler
Robert Daniels wrote:
> > I'm attempting to use lttng userspace tracing with a C++ application
> > on an ARM platform. I'm using GCC 4.8.4 on Linux 3.14 with the 2.6
> > release of lttng. I've compiled lttng-modules, lttng-ust, and
> > lttng-tools and have been able to get a simple test working with C
> > code. When I attempt to run the hello.cxx test on my target it will
> > segfault.
>
>
> I spent a little time digging into this issue and finally discovered the
> cause of my segfault with ARM C++ tracepoints.
>
> There is a struct called 'lttng_event' in ust-events.h which contains an
> empty union 'u'. This was the cause of my issue. Under C, this empty union
> compiles to a zero byte member while under C++ it compiles to a one byte
> member, and in my case was four-byte aligned which caused my C++ code to
> have the 'cds_list_head node' offset incorrectly by four bytes. This lead
> to an incorrect linked list structure which caused my issue.
>
> Since this union is empty, I simply removed it from the struct and everything
> worked correctly.
>
> I don't know the history or purpose behind this empty union so I'd like to
> know if this is a safe fix. If it is I can submit a patch with the union
> removed.
That's a very nice catch!
We do not support building tracepoint probe provider with
g++ yet, as stated in lttng-ust(3):
"- Note for C++ support: although an application instrumented with
tracepoints can be compiled with g++, tracepoint probes should be
compiled with gcc (only tested with gcc so far)."
However, if it works fine with this fix, then I'm tempted to take it,
especially because removing the empty union does not appear to affect
the layout of struct lttng_event as seen from liblttng-ust, which must
be compiled with a C compiler, and from probe providers compiled with
a C compiler. So all we are changing is the layout of a probe provider
compiled with a C++ compiler, which is anyway buggy at the moment,
because it is not compatible with the layout expected by liblttng-ust
compiled with a C compiler.
Reported-by: Robert Daniels <robert.daniels@vantagecontrols.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Wed, 29 Apr 2015 15:20:23 +0000 (11:20 -0400)]
Version 2.5.4
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Thu, 23 Apr 2015 22:45:05 +0000 (18:45 -0400)]
Fix: use lttng_secure_getenv to handle env. vars. involving paths
This is a security fix for applications linked against liblttng-ust
which are exposed as setuid binaries.
A malicious user which can run those applications could target those
environment variable paths to locations that would allow it to create
files in various areas of the filesystem.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Thu, 23 Apr 2015 22:48:43 +0000 (18:48 -0400)]
Fix: add missing getenv wrapper
Required for old glibc.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Jérémie Galarneau [Thu, 23 Apr 2015 17:07:38 +0000 (13:07 -0400)]
Fix: Don't wait during registration if clock_gettime() fails
get_constructor_timeout() currently returns -1 which, according to
the lttng-ust(3) man page and lttng_ust_init() implementation,
"waits forever".
This changes the behavior to match what is expressed in the comments.
Comments in get_constructor_timeout() and get_timeout() are also
modified to match the following convention:
-1: wait forever
0: don't wait
1: wait for "constructor_delay_ms"
Signed-off-by: Jérémie Galarneau <jeremie.galarneau@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Mon, 26 Jan 2015 20:07:27 +0000 (15:07 -0500)]
Version 2.5.3
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Fri, 16 Jan 2015 16:33:00 +0000 (11:33 -0500)]
Fix: missing parenthesis in offset_align_floor
Is currently unused. Triggers a compile error when used.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Mon, 12 Jan 2015 18:27:53 +0000 (13:27 -0500)]
Fix: add urcu-bp TLS fixup
When performing a library base address statedump, the UST threads can
grab the RCU read-side lock, which touch a TLS. We should "pre-fault"
this TLS in the constructor to eliminate a deadlock involving libc
dynamic loader lock.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Mon, 12 Jan 2015 16:09:18 +0000 (11:09 -0500)]
Fix: add missing poll.h include
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Mon, 5 Jan 2015 02:06:19 +0000 (21:06 -0500)]
Fix: get_subbuf: bound number of consumerd retry
When an application traced with per-UID buffers vanishes between its
reserve and commit ring buffer operations (e.g. segfault), it may cause
the consumer daemon to enter an endless loop, awaiting for the
application to complete its commit.
Fix this situation by bounding the number of retry using an adaptative
approach: try actively for a few times, and then try a few more times
with a 10ms delay between attempts.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Fri, 12 Dec 2014 17:14:40 +0000 (12:14 -0500)]
Version 2.5.2
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Philippe Proulx [Wed, 26 Nov 2014 21:50:56 +0000 (16:50 -0500)]
Fix: man: you can't link a lib with a static lib
Fixes: #825
Signed-off-by: Philippe Proulx <eeppeliteloop@gmail.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Mon, 1 Dec 2014 23:18:13 +0000 (18:18 -0500)]
Fix: context alignment not properly handled
This issue affects only architectures without efficient unaligned
accesses, only when a context field with larger alignment follows a
context field with smaller alignment. It generates unreadable traces
when such context fields are enabled in this configuration.
Fixes #867
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Christian Babeux [Fri, 28 Nov 2014 21:39:50 +0000 (16:39 -0500)]
Fix: Shorthand option -cp not supported in all version of Java
Signed-off-by: Christian Babeux <christian.babeux@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Mon, 17 Nov 2014 16:36:10 +0000 (17:36 +0100)]
Fix: allow tracepoints to build with --std=c99
std=c99 without #define __USE_BSD does not work for tracepoint header,
due to missing endianness macros. Force their definition in the
ust-endian.h wrapper if they are missing.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Wed, 12 Nov 2014 23:08:04 +0000 (18:08 -0500)]
Fix: filter attach vs event enable race
In order to correctly handle the use-case where events are enabled
_after_ trace is started, and _after_ applications are already being
traced, the event should be created in a "disabled" state, so that it
does not trace events until its filter is attached.
This fix needs to be done both in lttng-tools and lttng-ust. In order to
keep ABI compatibility between tools and ust within a stable release
cycle, we introduce a new "disabled" within struct lttng_ust_event
padding (previously zeroed). Newer LTTng-UST checks this flag, and
fallback on the old racy behavior (enabling the event on creation) if it
is unset.
Therefore, old session daemon works with newer lttng-ust of the same
stable release, and vice-versa. However, building lttng-tools requires
an upgraded lttng-ust, which contains the communication protocol with
the new "disabled" field.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Christian Babeux [Wed, 12 Nov 2014 19:35:21 +0000 (14:35 -0500)]
Fix: add a configure check for weak symbols support
Fixes #236
Signed-off-by: Christian Babeux <christian.babeux@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Wed, 12 Nov 2014 21:05:36 +0000 (16:05 -0500)]
Fix coverity warning about sizeof(void **) != sizeof(void *)
The size of the copied data is actually void *, not void **.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Sat, 8 Nov 2014 17:16:41 +0000 (12:16 -0500)]
Fix: off-by-one in sequence filter comparator
If we pass the following sequence as event input:
- field name "seqfield2"
- seq. len.: 4
- sequence array: "testA" (5 characters),
The following filters do not have the intended effect:
* 'seqfield2=="testA"'
- expected: no match.
- actual behavior: match.
* 'seqfield2=="test"'
- expected: match.
- actual behavior: no match.
This is caused by an off-by-one in the comparison with sequence length
in the filter bytecode interpreter.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Mon, 20 Oct 2014 20:59:51 +0000 (16:59 -0400)]
Version 2.5.1
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Jérémie Galarneau [Thu, 16 Oct 2014 13:45:19 +0000 (09:45 -0400)]
Fix: preserve example files' timestamps when copying
This fixes an issue where examples were always being rebuilt
when performing an out of tree build since the examples were
being copied to the build directory with a timestamp more
recent than the already-built example objects.
Signed-off-by: Jérémie Galarneau <jeremie.galarneau@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Philippe Proulx [Tue, 2 Sep 2014 20:12:55 +0000 (16:12 -0400)]
configure.ac: fix "missing jni.h" error message
DIR in --with-java-jdk=DIR must point to the JDK base
directory, not the actual JDK include directory.
Signed-off-by: Philippe Proulx <eeppeliteloop@gmail.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Wed, 16 Jul 2014 22:18:11 +0000 (18:18 -0400)]
Version 2.5.0
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Wed, 16 Jul 2014 21:40:53 +0000 (17:40 -0400)]
Mutex while updating statedump_pending is not needed
Take care of coverity warning ** CID
1226392: Data race condition (MISSING_LOCK)
/liblttng-ust/lttng-ust-comm.c: 1704 in lttng_ust_sockinfo_session_enabled()
It's not that the lock is missing, it's that we don't need to hold the
lock when doing the other modification.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Paul Woegerer [Mon, 14 Jul 2014 08:38:25 +0000 (10:38 +0200)]
Revert "Turn base address dump into experimental feature"
This reverts commit
b11abb674e50c67e3410ab2bd5d0a263e88b73ba.
Signed-off-by: Paul Woegerer <paul_woegerer@mentor.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Wed, 16 Jul 2014 14:35:43 +0000 (10:35 -0400)]
Fix: document ust_fork_mutex nesting
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Paul Woegerer [Mon, 14 Jul 2014 08:38:24 +0000 (10:38 +0200)]
Bugfix for #745 deadlock with baddr statedump+fork
Fixes #745
Signed-off-by: Paul Woegerer <paul_woegerer@mentor.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Francis Deslauriers [Thu, 10 Jul 2014 22:22:14 +0000 (18:22 -0400)]
Move enablers syncing after the channel registration
Fixes possible race condition where an event registration is received
before its channel registration by the consumer.
Signed-off-by: Francis Deslauriers <francis.deslauriers@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Paul Woegerer [Mon, 7 Jul 2014 14:04:21 +0000 (16:04 +0200)]
Add lttng_ust_notrace to static inline functions
The static inline functions
__tracepoint_provider_mismatch_<TRACEPOINT_PROVIDER> and
__tracepoint_provider_check_<TRACEPOINT_PROVIDER> are missing the
lttng_ust_notrace attribute to prevent them from getting instrumented
with -finstrument-functions (if e.g. defined in CFLAGS). This patch is
adding the attribute to those functions.
Signed-off-by: Paul Woegerer <paul_woegerer@mentor.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Michel Normand [Mon, 30 Jun 2014 16:12:19 +0000 (12:12 -0400)]
lttng ust: support ppc64le within configure
Takes care of https://bugzilla.redhat.com/show_bug.cgi?id=
1114691
Signed-off-by: Michel Normand <normand@linux.vnet.ibm.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Fri, 27 Jun 2014 14:07:32 +0000 (10:07 -0400)]
Version 2.5.0-rc2
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
David Goulet [Thu, 26 Jun 2014 19:07:54 +0000 (15:07 -0400)]
Fix: lttng-ust-jul: set thread in daemon mode
This way, when the JVM stops the main thread, the agent's thread will be
automatically shutdown thus making the .dispose() call optional.
Signed-off-by: David Goulet <dgoulet@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
David Goulet [Wed, 18 Jun 2014 18:08:24 +0000 (14:08 -0400)]
JUL: use root logger to capture events
The JUL agent now uses the root logger ("") to capture all events. This
allows us to remove the Timer thread and cleanup a huge portion of the
code base. It simplifies a great deal the internal structure of the
agent since we don't have to monitor the Logger object anymore.
Since tracepoint filtering is done in UST, we just the LTTng handler to
the root logger and send everything to UST.
Signed-off-by: David Goulet <dgoulet@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Christian Babeux [Thu, 26 Jun 2014 15:42:28 +0000 (11:42 -0400)]
Fix: Wrong configure check for UST perf event counters context support
When building on non-x86 platforms, the build errors out with the following:
CC lttng-context-perf-counters.lo
lttng-context-perf-counters.c:95:2: error: #error "Perf event counters are only supported on x86 so far."
lttng-context-perf-counters.c: In function 'read_perf_counter':
lttng-context-perf-counters.c:114:4: warning: implicit declaration of function 'rdpmc' [-Wimplicit-function-declaration]
make[2]: *** [lttng-context-perf-counters.lo] Error 1
The configure script checks for the presence of the "perf_event.h" header to enable support for
perf events counters. However, the current implementation is only available on x86,
hence the build fails on platform where the perf header is available.
Fix this issue by detecting the architecture we are currently building and wether
to enable or not the support for perf event counters context.
Signed-off-by: Christian Babeux <christian.babeux@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Tue, 10 Jun 2014 16:17:08 +0000 (12:17 -0400)]
Fix: multiple providers in same C file failure
If we have a C file in an application that is meant to contain both the
instrumentation and the provider, build fails if we have more than one
provider.
Fixed by putting back definitions to the state they were when defined by
tracepoint.h after a probe provider is created. This allows the next
tracepoint provider to define its tracepoints.
Fixes #805
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Gerlando Falauto [Tue, 3 Jun 2014 08:01:58 +0000 (10:01 +0200)]
Fix: header inclusion guard typo in lttng-ust-tracef-provider.h
Fix typo potentially allowing for multiple inclusions.
Signed-off-by: Gerlando Falauto <gerlando.falauto@keymile.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Wed, 28 May 2014 14:24:08 +0000 (10:24 -0400)]
Fix: 2.5.0-rc1 changelog has off-by-one date
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Wed, 28 May 2014 14:18:27 +0000 (10:18 -0400)]
Version 2.5.0-rc1
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
David Goulet [Tue, 27 May 2014 15:03:42 +0000 (11:03 -0400)]
Fix: out of tree build for lib Java and JUL
Reported-by: Daniel Olivier <daniel.olivier@ericsson.com>
Signed-off-by: David Goulet <dgoulet@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Fri, 23 May 2014 14:28:41 +0000 (10:28 -0400)]
Fix: default loglevel is DEBUG
Include "tracef" events when no explicit loglevel is specified. (for
usability)
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Fri, 23 May 2014 13:27:01 +0000 (09:27 -0400)]
Fix: refuse overwrite mode buffers with less than two subbuffers
Fixes #793
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Tue, 13 May 2014 13:34:03 +0000 (09:34 -0400)]
Fix: tracef.h: add missing c++ guard
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
David Goulet [Wed, 7 May 2014 15:33:41 +0000 (11:33 -0400)]
Fix: remove JUL loglevel filtering from the agent
The loglevel filtering is now done on the UST side setup by
liblttng-ctl. To achieve that, the agent has a new structure to support
two semantic concepts being a LTTng event and a JUL logger that have a
relation of N:1. This addition does not impact the current interfaces of
the agent.
Signed-off-by: David Goulet <dgoulet@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
David Goulet [Tue, 6 May 2014 15:56:25 +0000 (11:56 -0400)]
Fix: cleanup JUL agent state on sessiond disconnect
Signed-off-by: David Goulet <dgoulet@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Jérémie Galarneau [Wed, 7 May 2014 15:41:46 +0000 (11:41 -0400)]
Fix: Move generated headers to the BUILT_SOURCES target
This fixes the out of tree build when enabling Java support. The
problem in using an LTTngUst.c target is that automake will then
assume the file is generated and thus be found in the builddir.
Using the BUILT_SOURCES target will ensure the
org_lttng_ust_LTTngUst.h and org_lttng_ust_jul_LTTngUst.h headers
are generated before LTTngUst.c is compiled.
Tested-by: David Goulet <dgoulet@efficios.com>
Signed-off-by: Jérémie Galarneau <jeremie.galarneau@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Wed, 30 Apr 2014 16:09:59 +0000 (18:09 +0200)]
Fix: perf ust: check close return value
Fix coverity defect:
** CID
1208905: Unchecked return value (CHECKED_RETURN)
/liblttng-ust/lttng-context-perf-counters.c: 147 in setup_perf()
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Wed, 30 Apr 2014 07:08:27 +0000 (09:08 +0200)]
Cleanup: work-around missing inet.h struct mmsghdr fwd declaration
Removes build warning.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Fri, 7 Feb 2014 16:07:44 +0000 (11:07 -0500)]
Implement LTTng-UST perf counters support on x86
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Adrien Vergé [Fri, 25 Apr 2014 16:12:24 +0000 (12:12 -0400)]
Fix linking failure when cross-compiling
Examples binaries in doc/examples require libs in liblttng-ust/.libs/,
and those libs require other libs in the same directory. For instance
liblttng-ust.so uses liblttng-ust-tracepoint.so symbols, so the latter
is needed to properly link example programs. Currently, ld is given
the lib dir only via -rpath (used for linking executables dependencies)
but not via -rpath-link (used for linking libs dependencies).
When compiling regularly, this does not seem to be a problem (maybe ld
is smart enough). However when cross-compiling with the
arm-linux-gnueabihf toolchain, -rpath-link is needed otherwise linking
fails with:
arm-linux-gnueabihf/bin/ld: warning: liblttng-ust-tracepoint.so.0, needed by ../../../liblttng-ust/.libs//liblttng-ust.so, not found (try using -rpath or -rpath-link)
This patch adds -rpath-link to doc/examples/Makefile.am.
Signed-off-by: Adrien Vergé <adrienverge@gmail.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Romain Naour [Wed, 9 Apr 2014 21:55:07 +0000 (23:55 +0200)]
Fix: disable liblttng-ust-dl if dlinfo is not available in C library
According to uClibc commit [1], dlinfo is not available in recent uClibc
(>0.9.33), but available in older version of uClibc (<=0.9.33) whith a
different prototype than glibc.
dlinfo used by liblttng-ust-dl require RTLD_DI_LINKMAP that is not
available in uClibc.
To be able to use LTTng UST with uClibc, we need to disable the Dynamic
Linker Tracing functionality [2] if RTLD_DI_LINKMAP is not detected.
[1] http://git.uclibc.org/uClibc/commit/?id=
f3c9dc499c5c787ddd8c4320f2d44d2ae6e40c22
[2] http://lists.lttng.org/pipermail/lttng-dev/2014-February/022423.html
Signed-off-by: Romain Naour <romain.naour@openwide.fr>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Tue, 15 Apr 2014 17:05:34 +0000 (13:05 -0400)]
Fix: pass proper args when writing commit counter
lib_ring_buffer_write_commit_counter()'s 'buf_offset' argument should
contain offset of beginning of area used by the record being comitted.
However, lib_ring_buffer_commit() passes ctx->buf_offset, that gets
advanced by lib_ring_buffer_write() and thus points to just-after-
end-of-record at lib_ring_buffer_commit() time. This causes
lib_ring_buffer_write_commit_counter() to return without changing
commit_hot[idx].seq, due to
if (unlikely(subbuf_offset(offset - commit_count, chan)))
return;
Since after-crash data extraction tool checks 'seq' field to find out
how much data is in buffer, this results into inavailability of
data from partially-filled subbuffer for after-crash analysis.
This patch modifies lib_ring_buffer_write_commit_counter() and all its
callers to pass and expect the end of the area. So code works as it
should, and complete information becomes visible in crash dump.
[ Fix ported from lttng-modules. Changelog inspired from Nikita
Yushchenko's original patch. ]
Fixes #785
Reported-by: Nikita Yushchenko <nyoushchenko@mvista.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Mon, 14 Apr 2014 14:58:52 +0000 (10:58 -0400)]
Fix: pad strings that are modified concurrently with tracing
Fixes #734
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Jérémie Galarneau [Tue, 8 Apr 2014 18:43:32 +0000 (14:43 -0400)]
Fix: Use $AM_CC when it is defined instead of gcc in demo-tracef
Signed-off-by: Jérémie Galarneau <jeremie.galarneau@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Tue, 8 Apr 2014 18:29:29 +0000 (14:29 -0400)]
Revert "Fix: disable liblttng-ust-dl if dlinfo is not available in C library"
This reverts commit
88a15b8f075177056017dbcc2e0c6d624805a34e.
It has been reported that this commit breaks the standard case.
Fixes: #778
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Jérémie Galarneau [Tue, 8 Apr 2014 15:28:23 +0000 (11:28 -0400)]
Fix: .split() the CC environment variable in lttng-gen-tp
The CC variable might contain command line arguments which lttng-gen-tp
currently does not handle. This fix ensures that the arguments are
properly passed on to the compiler binary as if a normal Makefile
was used.
Signed-off-by: Jérémie Galarneau <jeremie.galarneau@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Romain Naour [Sun, 6 Apr 2014 21:02:41 +0000 (23:02 +0200)]
Fix: disable liblttng-ust-dl if dlinfo is not available in C library
According to uClibc commit [1], dlinfo is not available.
To be able to use LTTng UST with uClibc, we need to disable
the Dynamic Linker Tracing functionality [2].
[1] http://git.uclibc.org/uClibc/commit/?id=
f3c9dc499c5c787ddd8c4320f2d44d2ae6e40c22
[2] http://lists.lttng.org/pipermail/lttng-dev/2014-February/022423.html
Signed-off-by: Romain Naour <romain.naour@openwide.fr>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Jesper Derehag [Thu, 27 Mar 2014 07:57:06 +0000 (08:57 +0100)]
Fix: python invocation through env
This modification calls the interpreter through env instead of
directly, which should be more portable for those distros which do not install
python into /usr/bin/.
Signed-off-by: Jesper Derehag <jderehag@hotmail.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Wed, 2 Apr 2014 20:02:37 +0000 (16:02 -0400)]
Change default loglevel to TRACE_DEBUG_LINE
We do not want default loglevel to include tracef() debug output.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Jérémie Galarneau [Tue, 1 Apr 2014 15:49:58 +0000 (11:49 -0400)]
Fix: Override AM_PATH_PYTHON's default action-if-not-found
AM_PATH_PYTHON's default action aborts configure. This fix ensures that
configure follows through with a noop.
Signed-off-by: Jérémie Galarneau <jeremie.galarneau@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Tue, 25 Mar 2014 14:00:28 +0000 (10:00 -0400)]
Fix: don't accept configure --disable-shared
Building LTTng-UST with --disable-shared compiles fine, but the
resulting installed static libraries silently fails to trace, because
tracepoint fails to dlopen liblttng-ust-tracepoint.so. So don't allow
this unsupported configure option.
Also document this requirement in the README.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Tue, 25 Mar 2014 13:46:40 +0000 (09:46 -0400)]
Fix: configure.ac: add missing result to alignment req. check
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Tue, 18 Mar 2014 14:27:03 +0000 (10:27 -0400)]
Fix: malloc wrapper: infinite recursion with compat TLS
Fixes #765
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Sat, 15 Mar 2014 13:43:26 +0000 (09:43 -0400)]
Fix: lttng-ust needs -ldl for tracef()
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Sat, 15 Mar 2014 13:36:01 +0000 (09:36 -0400)]
Fix: typo in lttng-ust.3
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Sat, 15 Mar 2014 00:31:13 +0000 (20:31 -0400)]
Fix: liblttng-ust-libc-wrapper recursive use of calloc
When built with URCU with --disable-compiler-tls, URCU tls-compat.h uses
calloc in its TLS compat layer. This is used by RCU read-side, which is
used by tracepoint. So using a tracepoint in calloc triggers an infinite
recursion, since its first use of the TLS (with TLS compat) calls
calloc.
Fix this issue by keeping a per-thread nesting counter, and as soon as
we are allocating memory from a context nested within the memory
allocator, don't trace.
Don't trace from the static allocator: keep it as simple as possible,
since this is our fallback.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Fri, 14 Mar 2014 22:24:16 +0000 (18:24 -0400)]
Fix: add demo-tracef to dist tarball
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Fri, 14 Mar 2014 21:34:41 +0000 (17:34 -0400)]
Fix: add missing header files for tracef
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Thu, 12 Dec 2013 12:29:26 +0000 (07:29 -0500)]
Implement tracef() instrumentation API
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Zifei Tong [Fri, 14 Mar 2014 13:23:21 +0000 (21:23 +0800)]
Add 'unused' attribute to tracepoint callback
Compiling tracepoint provider with clang generates 'Wunused-function'
warnings. Add 'unused' attribute to silence these warnings.
Fixes #760
Signed-off-by: Zifei Tong <zifeitong@gmail.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Lars Persson [Wed, 12 Mar 2014 09:28:49 +0000 (10:28 +0100)]
Use autoconf AM_MAINTAINER_MODE
Give distribution maintainers the option to skip rebuilding
autoconf and automake generated files. The default behaviour
is still to have the rebuild rules enabled.
Signed-off-by: Lars Persson <larper@axis.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Zifei Tong [Tue, 4 Mar 2014 12:50:14 +0000 (20:50 +0800)]
Fix: mismatch between code and comments
Signed-off-by: Zifei Tong <zifeitong@gmail.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Mon, 10 Mar 2014 15:16:35 +0000 (11:16 -0400)]
Fix: incorrect urcu git URL in README
Reported-by: Ruslan Uvashev <ruslan.uvashev@gmail.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Fri, 28 Feb 2014 18:33:51 +0000 (13:33 -0500)]
Version 2.4.0
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Fri, 28 Feb 2014 18:26:28 +0000 (13:26 -0500)]
Fix: JUL test: update to API change
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Fri, 28 Feb 2014 15:40:33 +0000 (10:40 -0500)]
Fix: Cast from pointer to different integer size warning
Occurs on a 32-bit build. Caused by casting pointer arguments to 64-bit
integers (in the filter stack preparation code).
Fixes #691
Reviewed-by: Jérémie Galarneau <jeremie.galarneau@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Paul Woegerer [Fri, 28 Feb 2014 09:02:00 +0000 (10:02 +0100)]
Turn base address dump into experimental feature
Instead of having base address state dump enabled by default make it
necessary to explicitly enable it with environment variable
LTTNG_UST_WITH_EXPERIMENTAL_BADDR_STATEDUMP.
Documentation is updated to reflect this change and explain about the
experimental nature of this feature.
Signed-off-by: Paul Woegerer <paul_woegerer@mentor.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
David Goulet [Thu, 27 Feb 2014 19:51:55 +0000 (14:51 -0500)]
Fix: JUL support multiple loglevels
Fixes #744
Signed-off-by: David Goulet <dgoulet@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Jérémie Galarneau [Thu, 27 Feb 2014 17:30:41 +0000 (12:30 -0500)]
Fix: Skip lttng-gen-tp example build if Python is missing
Fixes #665
Signed-off-by: Jérémie Galarneau <jeremie.galarneau@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
David Goulet [Thu, 27 Feb 2014 16:53:22 +0000 (11:53 -0500)]
Fix: dynamically fetch the session daemon port
When connecting to the session daemon, fetch the port at each try. This
moves the read port file functions to the TCP session daemon client
where the connect happens.
If the port file is not found, simply retry 3 seconds later. Also, this
adds a timeout to the acquire semaphore of 3 seconds.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: David Goulet <dgoulet@efficios.com>
David Goulet [Thu, 27 Feb 2014 15:06:52 +0000 (10:06 -0500)]
Fix: JUL agent connect to user and root sessiond
In order to achieve such a thing, two tracepoints are used where one is
for the regular user and the other one for root. This is to avoid
duplicating the payload in the user and root traces.
Furthermore, if a root port is found, a new thread is started to handle
the second session daemon.
This is lockstep commit with lttng-tools:
f43f95a9a82e01eed34593260d510bd32e2083ec
Fixes #732
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: David Goulet <dgoulet@efficios.com>
David Goulet [Wed, 26 Feb 2014 21:06:31 +0000 (16:06 -0500)]
Fix: handle registration done command
This makes the getLTTngAgent() call wait until we receive that message
so that in the meantime the session daemon can enable events.
Fixes #692
Signed-off-by: David Goulet <dgoulet@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Jérémie Galarneau [Tue, 25 Feb 2014 21:32:05 +0000 (16:32 -0500)]
Fix: Unsynchronized access in LTTngTCPSessiondClient
enabledEventList is shared between the LTTngThread and eventTimer
threads but is not synchronized.
This patch changes enabledEventList's type from an ArrayList to a
synchronized HashSet.
Signed-off-by: Jérémie Galarneau <jeremie.galarneau@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Tue, 25 Feb 2014 11:43:46 +0000 (06:43 -0500)]
Cleanup: fix comment
Fixes #652
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Paul Woegerer [Tue, 25 Feb 2014 09:47:00 +0000 (10:47 +0100)]
Fix: Dump executable base-address with readlink
The previous approach only worked if the traced executable was invoked
via its fully qualified path. Using readlink to determine the full
qualified path of the traced executable works reliably even when the
traced executable is started via PATH lookup.
Signed-off-by: Paul Woegerer <paul_woegerer@mentor.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Fri, 21 Feb 2014 18:16:20 +0000 (13:16 -0500)]
Fix: add assert for NULL dereference
CID
1021248 (#1 of 1): Dereference null return value (NULL_RETURNS)3.
dereference: Dereferencing a null pointer "obj".
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Fri, 21 Feb 2014 18:13:14 +0000 (13:13 -0500)]
Fix: tracepoint out of memory handling
CID
1021247 (#1 of 1): Dereference null return value (NULL_RETURNS)3.
dereference: Dereferencing a null pointer "pl".
Handle more gracefully add_callsite memory allocation failure too.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Fri, 21 Feb 2014 18:00:59 +0000 (13:00 -0500)]
Fix: dereference before null check
Found by Coverity:
CID
1090602 (#1 of 1): Dereference before null check (REVERSE_INULL)
check_after_deref: Null-checking "stream" suggests that it may be null,
but it has already been dereferenced on all paths leading to the check
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Mathieu Desnoyers [Fri, 21 Feb 2014 17:54:07 +0000 (12:54 -0500)]
Cleanup: remove logically dead code
CID
1021235 (#1 of 1): Logically dead code (DEADCODE)
dead_error_line:
Execution cannot reach this statement "case LTTNG_UST_LOGLEVEL_ALL:".
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
This page took 0.043709 seconds and 4 git commands to generate.