Fix: channels can be _enabled_ after tracing is started, but not created

Fixes #619

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix filter parser segmentation fault with bison 3.0

Replace deprecated YYLEX_PARAM with %lex-param

Tested with bison 2.4, 2.5, 2.7 and 3.0.

Signed-off-by: Zifei Tong <soariez@gmail.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: typo in error msg

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: imprecise error message about root sessiond/tracing group

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: LTTNG_ERR_NEED_ROOT_SESSIOND error message

A root session daemon _and_ "tracing" group user membership are required
to trace the kernel.

Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Jérémie Galarneau <jeremie.galarneau@efficios.com>
Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: don't skip chmod if tracing group is not found

Fixes #620

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Tracepoint probes don't need extern C

tracepoint.h has the proper extern C guards, so every liblttng-ust API
called from the probe has the proper linkage.

The rest of the symbols are only meaningful within the probe, so we
don't care if they are mangled or not.

This fixes all UST applications in tests/

Fixes #597

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: reset out_fd_offset when we rotate the trace file

This value is only used for the sync_file_range, but it has to be set
to 0 when we start to write in a new trace file, otherwise the values
passed to this call are bogus.

Applied to 2.3, but could probably be backported to 2.2.

Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Julien Desfossez <jdesfossez@efficios.com>
Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: kernel ctl error codes are based on errno

ioctl returns -1, and error codes are based on -errno.

Signed-off-by: Julien Desfossez <jdesfossez@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Version 2.2.3

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

test: test_periodical_metadata_flush shrink timer period, kill app

Decrease the switch timer period to 100ms (instead of 1s), since we're
only getting 2s worth of app run. If the system is busy, nothing really
guarantee that the timer will indeed fire during this time-frame (so
strictly speaking, this test could still fail).

Also, kill the application before trying to read the trace: this should
ensure that the trace is not appended to concurrently with validation.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: format string type mismatch

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: tests: provide channel name when a non-default channel exists: per-pid

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: tests: provide channel name when a non-default channel exists

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: data pending race

There is a data pending race involving late population of the streams in
the stream hash table, and applying flush on streams that are not yet
globally visible.

This is caused by the fact that streams are added to the hash table only
when received by the data-handling consumer thread.

This results in data_pending() incorrectly returning that there is no
data pending in some cases.

This has been discovered by adding 1s delay in read subbuffer function
for testing.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Cleanup: ust-consumer: wrong indentation

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: print dots while waiting for data availability

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: push metadata on stop for per-UID buffers

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Version 2.2.2

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: set tracefile size test with PID buffers

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: Babelstats fail to parse output with no process name or pid

The process name and pid are optionnal components in the babeltrace
text output.

Signed-off-by: Christian Babeux <christian.babeux@efficios.com>

Fix: kernel data unit test

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: don't send error to sessiond on orderly shutdown

The send always fail if the sessiond has closed the socket so no point
of trying to send an error.

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: consumer: incorrect size zmalloc

Also, check return value.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: RCU read side lock unbalanced

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix deadlock: don't take channel lock in timer

Reviewed-by: Julien Desfossez <julien.desfossez@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Introduce channel timer lock

Currently a shadow of the channel lock. Will eventually be used to
protect channel timer handler from concurrent channel updates without
being held when the timer is stopped (future commit).

Reviewed-by: Julien Desfossez <julien.desfossez@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Conflicts:
src/common/consumer-stream.c

document lttng_ustconsumer_request_metadata locking constraints

Reviewed-by: Julien Desfossez <julien.desfossez@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

document metadata_switch_timer() deadlock

Reviewed-by: Julien Desfossez <julien.desfossez@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

document metadata_switch_timer() locking constraints

Reviewed-by: Julien Desfossez <julien.desfossez@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

consumer: replace DBG2() instances by DBG()

The consumer daemon does not currently support DBG2() printout.

Reviewed-by: Julien Desfossez <julien.desfossez@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

ust consumer: data_pending check is endpoint active

Reviewed-by: Julien Desfossez <julien.desfossez@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: kernel consumer: data_pending check if endpoint active

Reviewed-by: Julien Desfossez <julien.desfossez@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

consumer: explicitly set endpoint status to active at init

Ensures we explicitly set active at init for channels that don't
interact with a relayd.

Reviewed-by: Julien Desfossez <julien.desfossez@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

document consumer_metadata_cache_flushed use of consumer_data.lock

Reviewed-by: Julien Desfossez <julien.desfossez@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

consumer: introduce channel lock

Reviewed-by: Julien Desfossez <julien.desfossez@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Missing NULL pointer init in tap.c

Could lead to free() of unitialized pointer.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: Unchecked asprintf/vasprintf return values

strp may be modified even in case of failure according to the
ASPRINTF(3) man page.

Signed-off-by: Jérémie Galarneau <jeremie.galarneau@efficios.com>
Signed-off-by: David Goulet <dgoulet@efficios.com>

Update version to v2.2.1

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: consumer err_sock cloexec

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: possible consumer sockets double close on cleanup

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: add gpl and lgpl files to tarball

Fixes #581

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: don't install libtap system wide

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: close consumer sockets in sessiond cleanup

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: print errno message on connect() error

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: possible double-close on stream out_fd

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: session ID signess to uin64_t in sessiond

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: consumer_add_relayd_socket() report errors to sessiond

Wait until the side-effect to relayd is actually performed before
replying that "all is fine" to sessiond.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: UST per-UID channels persist across application teardown

Fixes #582

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: add missing enum lttcomm_return_code entries

Those should not appear as unknown.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: detect the correct version of LTTng-UST

Fixes #580

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: sessiond: use uint64_t for all session ids

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: consumer: use uint64_t for all sessiond_id

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: periodical flush check trace before stop

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: consumer: 64-bit index for relayd rather than 32-bit (v2)

Relayd "unique" ids wrap every 32-bit, and in some cases, negative
values are considered as error.

Change this to make the error value specifically -1ULL, use a direct
comparison (since we use an unsigned 64-bit integer, comparison with 0
becomes incorrect).

Since we now use a 64-bit ID, it is assumed to _never_ wrap-around
(remember, value -1ULL is an _error_). Therefore,
consumer_add_relayd_socket() can become much more strict than it was:
instead of accepting re-use of net_seq_idx, we can now assert that upon
LTTNG_STREAM_CONTROL socket, we have indeed allocated a relayd object,
and upon LTTNG_STREAM_DATA, we have found a relayd object.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: send per-pid session id in channel creation

The registry indexing for per-pid sessions is done with a per-pid
session id. So for per-pid buffers, we need to send the per-pid session
id as well as the global session id to the consumer in order to give it
enough information if it needs to request metadata later.

This patch adds the session_id_per_pid to the channel creation message
and to the consumer. When the sessiond receives a metadata_request,
depending on the buffer type (per-pid or per-uid), it selects the right
id to do the registry lookup.

Signed-off-by: Julien Desfossez <jdesfossez@efficios.com>
Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: consumer double-close on error

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Update URCU detection to correctly check for a 0.7 version

Signed-off-by: Yannick Brosseau <yannick.brosseau@gmail.com>

Fix: set hidden attribute to utils_* calls

Signed-off-by: David Goulet <dgoulet@efficios.com>

Update version to v2.2.0

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: build path for unit test to build out of tree

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: if relayd is unreachable, disable consumer for the session

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: possible infinite loop in disable ust event

Fixes #575

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: don't enable a channel if a session was already started

This removes also the check done to start the session if it was started
before.

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: bad type for the relayd id

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: add debug statement in kernel flush metadata

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: destroy default created channel if add-context fails

Fixes #554

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: clarify tracefile size/count in lttng.1 man

Fixes #572

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: don't stop a session that was not previously started

Fixes #567

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: update lttng.1 man page

Fixes #571

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: bad protocol flow between sessiond and consumerd

Also, the error handling was wrong for some error path.

Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: kernel memory leak in error path

Signed-off-by: David Goulet <dgoulet@efficios.com>

Update version to v2.2.0-rc3

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: --disable-static option

Fixes #562

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: add missing file for make dist

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: channel and stream leak in consumerd

Multiple error paths were possibly leaking channel and stream structure
in the lttng-consumerd.

This has been detected by the stress test killing the relayd randomly
during a long period of time.

Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: consumer add channel return value was positive on error

Reported-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: wrong check after malloc in buffer registry

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: change an ERR message to a DBG

Signed-off-by: David Goulet <dgoulet@efficios.com>

Tests: Fix wrong number of tests in thread stall as root test

Signed-off-by: Christian Babeux <christian.babeux@efficios.com>
Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: set +x to test helper script

Signed-off-by: David Goulet <dgoulet@efficios.com>

Test: add new stress test that kills the relayd randomly

Signed-off-by: David Goulet <dgoulet@efficios.com>

Using LTTNG_HOME environment variable if exists, with fallback to HOME

Fix: remove use of stat()

1019896 Time of check time of use
In utils_mkdir_recursive: A check occurs on a file's attributes before
the file is used in a privileged operation, but things may have changed
(CWE-367)

mkdir() is already doing the EEXIST check.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: out of bound strcpy

1019904 Copy into fixed size buffer
In add_context: A source buffer of statically unknown size is copied
into a fixed-size destination buffer (CWE-120)

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Remove logically dead code

1019963 Logically dead code
In list_sessions: Code can never be reached because of a logical
contradiction (CWE-561)

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: handle close return value

1019968 Unchecked return value
In lttng_health_check: Value returned from a function is not checked for
errors before being used (CWE-252)

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Remove useless if() check

1022350 Dereference before null check
In ust_registry_session_destroy: All paths that lead to this null
pointer comparison already dereference the pointer earlier (CWE-476)

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: goto error_fatal on lttng_kconsumer_recv_cmd errors

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: goto error_fatal on lttng_ustconsumer_recv_cmd errors

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: unnecessarily large memory allocation

1019911 Wrong sizeof argument
In consumer_thread_data_poll: The sizeof operator is invoked on the
wrong argument (CWE-569)

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: comparing array against NULL

1019939 Array compared against 0
In process_client_msg: Array compared against NULL pointer (CWE-398)

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: Remove reference to lttng-probe-lttng kernel module

The lttng-probe-lttng module was removed as part of lttng-modules
commit 662ed1665428b38ee00d4be5751581108bf35390

Fixes #553

Signed-off-by: Jérémie Galarneau <jeremie.galarneau@efficios.com>
Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: memleak in liblttng-ctl visit_node_load()

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: memleak in the UST registry

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: add missing free causing a memleak

Also remove a unused variable in the ust registry event structure.

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: memleaks in channel/event notify

Signed-off-by: David Goulet <dgoulet@efficios.com>

Fix: sanitize wait queue in the dispatch thread

This is to avoid memory leaks when the notify socket is never received
thus cleaning the wait node for command socket that are invalid.

Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: David Goulet <dgoulet@efficios.com>

Implement health check for app registration dispatch

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: David Goulet <dgoulet@efficios.com>

Implement health check for app notification thread

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: David Goulet <dgoulet@efficios.com>