Fix: ring buffer: get_subbuf() checks should be performed on "consumed" parameter

This triggers lots of false-positive -EAGAIN errors in flight recorder
snapshots.

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

Fix: SWITCH_FLUSH new sub-buffer checks

The SWITCH_FLUSH, when performed on a completely empty sub-buffer, was
missing some checks (imported from space reservation).

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

Fix: ring buffer: handle concurrent update in nested buffer wrap around check

With stress-test loads that trigger sub-buffer switch very frequently
(small 4kB sub-buffers, frequent flush) in lttng-modules, we currently
observe this kind of warnings once every few minutes:

[65335.896208] ring buffer relay-overwrite-mmap, cpu 5: records were lost. Caused by:
[65335.896208]   [ 0 buffer full, 1 nest buffer wrap-around, 0 event too big ]

It appears that the check for nested buffer wrap-around does not take
into account that a concurrent execution contexts (either nested for
per-cpu buffers, or from another CPU or nested for global buffers) can
update the commit_count value concurrently.

What we really want to do with this check is to ensure that if we enter
a sub-buffer that had an unbalanced reserve/commit count, assuming there
is no hope that this gets rebalanced promptly, we detect this and drop
the current event. However, in the case where the commit counter has
been concurrently updated by another reserve or a switch, we want to
retry the entire reserve operation.

One way to detect this is to sample the reserve offset twice, around the
commit counter read, along with the appropriate memory barriers.
Therefore, we can detect if the mismatch between reserve and commit
counter is actually caused by a concurrent update, which necessarily has
updated the reserve counter.

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

Fix: handle writes of length 0

lib_ring_buffer_write() could be passed a length of 0. This typically
has no side-effect as far as writing into the buffers is concerned,
except for one detail: in overwrite mode, there is a check to make sure
the sub-buffer can be written into. This check is performed even if
length is 0. In the case where this would fall exactly at the end of a
sub-buffer, the check would fail, because the offset would fall exactly
at the beginning of the next sub-buffer.

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

Version 2.1.3

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

Tracepoint.h: replace assertion by fprintf and abort()

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

Fix: tracepoint.h incorrect assumption about constructor order

Incorrect assumption about constructor execution order can trigger a
segfault when trying to execute tracepoint_register_lib.

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

Fix: tracepoint instrumentation constructor order issue

If the linker decides to run a constructor from a tracepoint probe
before the constructor from the application, a recent modification
(commit 558b9d86247004f8e9bbaf8c982f3b2b182093d1) allowed that the wrong
constructor execution order could prohibit the program's tracepoints
from being registered.

Fix this by making two separate constructors.

Fixes #487

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

Performance: add unlikely to tracepoint dynamic linking test

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

Fix: ensure all probe providers have their symbols

Original commit log from stable-2.2:
Following the removal of metadata generation from UST, the probes cannot
find the tracepoint_dlopen() symbol. Ensure those probes can be built by
declaring this symbol within each probe.

stable-2.1 backport:
Similar issue was experienced with the "demo" test program when built
outside of the UST tree.

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

Clarify probe registration documentation/errors

Fixes #513

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

Add mising include in ust.h

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

lttng-gen-tp: Fix include guard name with file using non valid characters

Now, the include guard will be generated with all the non
alpha-numeric characters replace with an '_'

Fixes #511

Signed-off-by: Yannick Brosseau <yannick.brosseau@gmail.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: missing dependency for liblttng-ust-tracepoint.so

Fixes #538

Reported-by: Alexandros Frantzis <alexandros.frantzis@canonical.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

manpage: Document probe provider compatibility

Fixes #502

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

Typo fix in README

Fixes #505

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

Safety-check: only call register_lib if found in liblttng-ust-tracepoint.so.0

If a liblttng-ust-tracepoint.so.0 would happen not to contain the
expected symbols, don't trace (rather than segfault).

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

Fix: filter string comparison should check for literal

The prior fix "Fix: filter string wildcard comparison" missed check for
string literal.

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

Fix: filter string wildcard comparison

wildcards * should match 0 or more characters (not 1 or more).

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

Fix: _LGPL_SOURCE rcu dereference fix

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

Version 2.1.2

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

Fix uninitialized has_loglevel variable

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

tracepoint: Don't add NULL probes

Reported-by: Sahara <keun-o.park@windriver.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Remove mention of locking issues associated with dlopen usage

This notice has been removed from the README as part of revision
b834dead.

Add a warning relating to dlclose() usage detailed in issue 447 in both
the man page and the README file.

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

Fix: refcount issue in lttng-ust-abi.c

The following scenario could lead to a segmentation fault in
applications when the sessiond disappears or when the application try to
exit. This is caused by incorrect handling of reference counts.

This can happen in very particular race scenario, described as follows:

1) The sessiond asks for "release" of one or more objects (e.g. a
   session object), but _without_ asking for release of _all_ objects
   referencing the session.
2) The application exits, thus calling objd_table_destroy(). It walks on
   all objects, decrementing their reference count, freeing memory when
   their reference count reaches "1".

However, here is the issue: since "release" has already been performed
by sessiond on the session object, this extra reference count unref
performed by objd_table_destroy() can make the session object disappear
while it is still needed by either the channel object or the enabler
object. Therefore, we can experience a segmentation fault when we try to
unref and free the channel or enabler objects within
objd_table_destroy().

Fix this issue by adding the concept of an "owner reference". Only
objd_table_destroy(), lttng_ust_objd_table_owner_cleanup(), "release"
commands, and failure paths of object creation are allowed to decrement
the owner reference. We restrict objd_table_destroy() and
lttng_ust_objd_table_owner_cleanup() to _only_ decrement refcount of
objects _if_ their owner reference is still held.

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

Fix: don't flush-final for offset 0 if reader is on sub-buffer

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

Use tp rcu link test in provider

Ensure we never trigger a null pointer exception.

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

Remove direct dependency of probes on urcu-bp

The filter feature added a direct dependency of probes providers against
urcu-bp .so, while linking them against liblttng-ust should suffice. Fix
this by using the tracepoint.h wrapper for rcu_dereference().

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

Fix package: don't distribute generated headers

Fixes #421

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

Fix: don't cancel already exited threads

Ensure we don't issue pthread_cancel() on already exited threads.

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

Scalability fix: tracepoint.c hash table size increase

Speed up process startup for applications with large number of
tracepoints (e.g. 16k in J9 vm) by increasing the tracepoint.c hash
table size from 64 to 4096.

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

Scalability fix for many events: event hash table size

Increase number of buckets for event hash table (per session) from 64 to
4096. This improves process startup time for J9 vm (with 16k
tracepoints) very significantly.

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

Speed up probe registration for large amount of events

LTTng-UST probe registration is O(n^2). I actually left a comment that
describes this, also implying that we can improve this if it becomes an
issue.

I've had a report from Yang Wang, who works on instrumenting the J9 VM,
that O(n^2) does not agree well with 16000 probes and tracepoints.

It appears that lttng_probe_register() is being too paranoid for its own
good. There are now many things that are guaranteed by the way probe
providers are being built.

We actually need to keep a check that no provider with the same name has
been registered (O(n) on the number of registered providers, could be
improved with a hash table if it ever becomes necessary).

Use an assert to check that each event name starts with its own provider
name (it would be an internal error within the provider if it's not the
case). (O(n) on the number of events within a provider)

The rest is just useless, so remove this O(n^2) check.

While we are there, remove the now unused
lttng_event_get()/lttng_event_put() functions.

Reported-by: Yang Wang <yangw.wang5@unb.ca>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Add missing demo-trace shell script to dist tarball

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

Fix: missing test for lttng_ust_comm_should_quit in lttng-ust-comm.c

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

Documentation: clarify debian package name for uuid in README

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

Document dependency on libuuid

Fixes #418

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

Fix: comment in ust-ctl.h

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

Fix I/O-related error values in ustctl

These internal ustctl API members document >=0 as OK values. Make sure
reply recv errors are returned as errors.

Clarify ustctl_create_stream() error values.

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

Version 2.1.1

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

Introduce hash table for lttng_create_event_if_missing()

lttng_create_event_if_missing() takes a lot of CPU time with stress-test
applications containing 1000 different TRACEPOINT_EVENT() and 1000
individual tracepoint() call-site.

With tracing disabled:

time ./AppWith1000_lines_TP 0

real    0m2.487s
user    0m2.424s
sys     0m0.000s

Introducing this hash table cuts the overhead very significantly when
tracing is enabled:

Before patch:

 72.16%  AppWith1000_lin  liblttng-ust.so.0.0.0             [.] lttng_create_event_if_missing
 25.64%  AppWith1000_lin  AppWith1000_lines_TP              [.] a(int)

time ./AppWith1000_lines_TP 0

real    0m7.946s
user    0m7.864s
sys     0m0.000s

After patch:

 89.13%  AppWith1000_lin  AppWith1000_lines_TP              [.] a(int)
  3.30%  AppWith1000_lin  liblttng-ust.so.0.0.0             [.] lttng_create_event_if_missing

time ./AppWith1000_lines_TP 0

real    0m2.762s
user    0m2.692s
sys     0m0.004s

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

Fix: don't build C++ example if a C++ compiler isn't available

By default lttng-ust builds a hello.cxx C++ example that demonstrates
the usage of the userspace tracing library in a C++ program.
Unfortunately, when no C++ support is available, the build of lttng-ust
fails just because of this example code. So we make the compilation of
this code conditional on whether a working C++ compiler was found.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Version 2.1.0

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

Bump liblttng-ust-ctl lib version major number

This internal library between ust and tools has had its ABI changed
between 2.0 and 2.1.

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

Bump UST proto version minor number

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

Helper to debug: add object name

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

Tests: Fix rundir not created in ust-basic-tracing

Signed-off-by: Christian Babeux <christian.babeux@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

ust test: add missing return -1

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

Test: update ust tracing unit tests to 2.1 internal ABI

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

Tests: Fix rundir not created in ust-multi-test

Signed-off-by: Christian Babeux <christian.babeux@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Adapt internal files and examples to TRACEPOINT_INCLUDE

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

Adapt tests to TRACEPOINT_INCLUDE

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

Cleanup: tests remove hello.cxx/ust_tests_demo.h

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

Fix: don't do macro expansion in tracepoint file name

Change the API in a backward compatible way: previously, probe headers
would do e.g.

but e.g. "linux" here would be expanded to "1", which is bad.

Fix this by allowing those headers to do:

instead. We still support the old TRACEPOINT_INCLUDE_FILE for backward
compatibility.

Reported-by: Sébastien Barthélémy <barthelemy@crans.org>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix lttng-gen-tp: Template file must end in .tp

Fixes #297

Signed-off-by: Christian Babeux <christian.babeux@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Pack structures in comm protocol between UST and sessiond

Ensure robustness with respect to 32-bit vs 64-bit apps vs sessiond.

Since we are updating the ABI, change the order of overwrite field in
channel and channel attributes, to remove some unneeded padding.

This breaks compatibility between sessiond 2.1 and ust 2.0 (and
vice-versa), but sessiond refuses applications with version number that
does not match.

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

Cleanup lttng-gen-tp: Help not showing when using -h,--help

Fixes #299

Signed-off-by: Christian Babeux <christian.babeux@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Cleanup lttng-gen-tp: remove leading underscore before include guard

Fixes #298

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

fix memleak: ustctl free shadow chan on ustctl_unmap_channel()

Also on error path within map channel.

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

Update license text

- Add LICENSE file
- all public headers are now under MIT license,
- add missing text from MIT license into each file using this license.

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

Fix: check if event enabled for bytecode-less events

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

Fix: handle enablers without bytecode

to reproduce issue with the upstream "hello" program:

lttng create
lttng enable-event -u -a --filter 'intfield==42'
lttng enable-event -u -a
lttng start

./hello
lttng stop
lttng view

only shows event with 42 intfield. We expect all events.

Fixes #406

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

Print probe provider mismatch error even without -Wsystem-headers

Suggested-by: David OShea <David.OShea@quantum.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

filter: Add missing padding

This will make future extensions to filter much easier.

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

Version 2.1.0-rc2

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

filter interpreter cleanup: use uint64_t for retval

Reported-by: Christian Babeux <christian.babeux@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: filter linking can dereference NULL pointer on alloc failure

Reported-by: Christian Babeux <christian.babeux@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Cleanup: remove whitespaces and EOL in tests

Reviewed-by: David Goulet <dgoulet@efficios.com>
Reviewed-by: Christian Babeux <christian.babeux@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Filter: use only single lower bit of filter return value

Implementation change: use a uint64_t as filter return value, and test
with a mask for this bit. This reserves other bits for future use.

Reviewed-by: David Goulet <dgoulet@efficios.com>
Reviewed-by: Christian Babeux <christian.babeux@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: filter: var len array at end of external structure

Reviewed-by: David Goulet <dgoulet@efficios.com>
Reviewed-by: Christian Babeux <christian.babeux@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: filter link fail handling

Reviewed-by: David Goulet <dgoulet@efficios.com>
Reviewed-by: Christian Babeux <christian.babeux@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: add missing seqnum field to filter

Reviewed-by: David Goulet <dgoulet@efficios.com>
Reviewed-by: Christian Babeux <christian.babeux@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

filters: perform union rather than intersection

Reviewed-by: David Goulet <dgoulet@efficios.com>
Reviewed-by: Christian Babeux <christian.babeux@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Implement support for overlapping wildcard/events

Reviewed-by: David Goulet <dgoulet@efficios.com>
Reviewed-by: Christian Babeux <christian.babeux@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Cleanup: add lttng_/lttng-/LTTNG_ prefixes

ltt_event -> lttng_event mass rename
Rename ltt_chan -> lttng_chan
Rename ltt_session -> lttng_session
Rename enum abstract_types to enum lttng_abstract_types
Rename ltt_transport to lttng_transport
Rename rest of ltt_ prefixes to lttng_
Complete file and symbol renames from LTT/ltt to LTTNG/lttng
Finish ltt->lttng symbol conversion

Reviewed-by: David Goulet <dgoulet@efficios.com>
Reviewed-by: Christian Babeux <christian.babeux@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

filter: add seqnum field to filter command

Reviewed-by: David Goulet <dgoulet@efficios.com>
Reviewed-by: Christian Babeux <christian.babeux@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Filter iteration: iterate on list of filters

Reviewed-by: David Goulet <dgoulet@efficios.com>
Reviewed-by: Christian Babeux <christian.babeux@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

document that tracepoint names should ideally not be re-used

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

Remove LIBFORMAT config declaration, unused

Remove leftover from UST 0.x. Not needed anymore.

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

Add libc errno translation layer to UST error code

Also add four new possible code being EEXIST, EINVAL, ENOSYS and EPERM.

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

Revert "Fix: configure.ac should use "--target", not "--host""

This reverts commit 98c8a1d90d1052980365353eef201124863128b0.

No. --target is for building cross-compilers. --host was  appropriate.

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

Fix: configure.ac should use "--target", not "--host"

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

Fix: add const qualifier for filter local void *

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

Fix: re-allow non-lvalue string, sequence, array parameters

Issue introduced by upstream commit:

  commit 27f4b6094f399f2fe231b58801dce98cbd21baa9
  Author: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
  Date: Tue Sep 4 12:17:07 2012 -0400

  Fix filter: pointer to string, not string, should be on stack

  Fixes #329

Reported-by: David Bryant <david.bryant@quantum.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Tear down handles associated with a closed sessiond socket

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

Distinguish UST return codes from transport return codes

One API change visible from lttng-tools: -ENOENT return values for
tracepoint list and tracepoint field list are now -LTTNG_UST_ERR_NOENT.

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

Fix: Conditionally disable tests requiring shared libs support

When building lttng-ust with shared library support explicitly
disabled (e.g.: ./configure --disable-shared), libtool fail with
a fatal error:

  CC       tp.lo
  CC       tp2.lo
  CCLD     liblttng-ust-provider-ust-tests-demo.la
libtool: link: can not build a shared library
libtool: link: See the libtool documentation for more information.
libtool: link: Fatal configuration error.

The build should not fail because some tests require explicit shared
library support, instead they should be skipped.

This patch detect that the --disable-shared flag was passed to the
configure script and toggle the "NO_SHARED" Automake variable.
Thus, the tests that require explicit shared library support can
be skipped when the NO_SHARED variable is true.

[ Edit by Mathieu Desnoyers: add "" in configure.ac to follow the local
  coding style. ]

Signed-off-by: Christian Babeux <christian.babeux@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Cleanup: don't spawn per-user thread if HOME is not set

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

Manpage: document supported UST contexts

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

Fix: procname context semantic

Cache the procname per-thread rather than per-process to take into
account that prctl() can be used to set thread names.

prctl() should be issued before tracing each thread's first event if we
care about the thread name.

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

Fix: Fix self-assign warning on struct ustfork_clone_info init

Use the proper field designator syntax (C99) to initialize the
ustfork_clone_info struct.

Signed-off-by: Christian Babeux <christian.babeux@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: memcpy of string is larger than source

Hollis Blanchard <hollis_blanchard@mentor.com> wrote:
> I seem to have hit a little problem with a "hello world" test app and
> lttng-ust 2.0.3. lttng-ust.git seems to be affected as well. Basically,
> I created a single UST tracepoint, but as soon as I run "lttng
> enable-event -u -a", my app segfaults. The problem seems to be that when
> creating the event to pass to ltt_event_create(), we try to memcpy the
> full 256 bytes of name. However, the name might be shorter, and if we
> get unlucky it falls within 256 bytes of the segment boundary...

Fixing the 3 sites where this issue arise. Manually inspecting all
memcpy in the UST code returned by grep did the job.

Reported-by: Hollis Blanchard <hollis_blanchard@mentor.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Implement liblttng-ust-fork daemon() override test

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

liblttng-ust-fork: override daemon() call

Fixes #252

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

ustfork: set errno to ENOSYS if symbol lookup fails

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

Fix: be quiet on filter linker error

Could be reproduced by starting the application before enabling the
event with filter.

Fixes #362

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

Build out of src tree

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

Fix: filter bytecode specializer stack leak

Fixes #356

Signed-off-by: Christian Babeux <christian.babeux@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: reloc offset validation error out on filters with no reloc table

The reloc table is currently appended at the end of the bytecode data.
With this scheme, the reloc table offset will be equal to the length
of the bytecode data.

<- length ->
+----------+-------------+
| BYTECODE | RELOC TABLE |
+----------+-------------+
           |
           +--> Reloc table offset

A special case arise with filters with no reloc table.

Example:

Filter: "myString" == "yourString"
./filter-grammar-test -p -B -i -b < bogus
<root>
        <op type="==">
                <expression>
                        <string value="myString"/>
                </expression>
                <expression>
                        <string value="yourString"/>
                </expression>
        </op>
</root>
Generating IR... done
Validating IR... done
Generating bytecode... done
Size of bytecode generated: 24 bytes.
Bytecode:

Val. Operator
---- --------
0x40 (FILTER_OP_LOAD_STRING)
0x6D m
0x79 y
0x53 S
0x74 t
0x72 r
0x69 i
0x6E n
0x67 g
0x00 \0
0x40 (FILTER_OP_LOAD_STRING)
0x79 y
0x6F o
0x75 u
0x72 r
0x53 S
0x74 t
0x72 r
0x69 i
0x6E n
0x67 g
0x00 \0
0x0C (FILTER_OP_EQ)
0x01 (FILTER_OP_RETURN)

Reloc table (offset: 24):
Empty

<-   24   ->
+----------+
| BYTECODE | <- No reloc table
+----------+
           |
           +--> Reloc table offset

In this case, we see that the reloc table offset (24) is indeed equal to
the length of the bytecode (24), but the reloc table is _empty_. Thus,
the reloc_offset received in handle_message() will be equal to the
data_size and will be wrongly flagged as not within the data even thought
the filter is entirely valid.

The fix is to simply allow a reloc_offset to be equal to the data_size.

Fixes #342

Signed-off-by: Christian Babeux <christian.babeux@efficios.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Perform calculation on bit size in 64-bit

allow 32-bit architectures to have sub-buffers larger than 256MB.

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