Introduce __dynamic_array_enc_ext_2() and tp_memcpy_dyn_2()

Allow copying into a single variable-length buffer from two sources.
Needed to fix printk support for Linux kernel < 3.5.

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

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), 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>

Cleanup: lib_ring_buffer_switch_new_end() only calls subbuffer_set_data_size()

lib_ring_buffer_switch_new_end() is always called when an event exactly
fills a sub-buffer, which makes padding_size always 0. However, there is
one side-effect that lib_ring_buffer_switch_new_end() needs to have: it
calls subbuffer_set_data_size() to update the size of the data to be
read from the sub-buffer.

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

Revert "Cleanup: ring buffer: remove lib_ring_buffer_switch_new_end()"

This reverts commit 5fb66f07aae4884426f1706d0281bd242a38c2a7.

The equivalent UST commit of this change broke regression tests of
lttng-tools.

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

Fix: handle writes of length 0

lib_ring_buffer_write(), lib_ring_buffer_memset() and
lib_ring_buffer_copy_from_user_inatomic() 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.

It triggers this warning:

[65356.890016] ------------[ cut here ]------------
[65356.890016] WARNING: at /home/compudj/git/lttng-modules/wrapper/ringbuffer/../../lib/ringbuffer/../../wrapper/ringbuffer/../../lib/ringbuffer/backend.h:110 lttng_event_write+0x118/0x140 [lttng_ring_buffer_client_mmap_overwrite]()
[65356.890016] Hardware name: X7DAL
[65356.890016] Modules linked in: lttng_probe_writeback(O) lttng_probe_workqueue(O) lttng_probe_vmscan(O) lttng_probe_udp(O) lttng_probe_timer(O) lttng_probe_sunrpc(O) lttng_probe_statedump(O) lttng_probe_sock(O) lttng_probe_skb(O) lttng_probe_signal(O) lttng_probe_scsi(O) lttng_probe_sched(O) lttng_probe_rcu(O) lttng_probe_random(O) lttng_probe_printk(O) lttng_probe_power(O) lttng_probe_net(O) lttng_probe_napi(O) lttng_probe_module(O) lttng_probe_kvm(O) lttng_probe_kmem(O) lttng_probe_jbd2(O) lttng_probe_jbd(O) lttng_probe_irq(O) lttng_probe_ext4(O) lttng_probe_ext3(O) lttng_probe_compaction(O) lttng_probe_btrfs(O) lttng_probe_block(O) lttng_types(O) lttng_ring_buffer_metadata_mmap_client(O) lttng_ring_buffer_client_mmap_overwrite(O) lttng_ring_buffer_client_mmap_discard(O) lttng_ring_buffer_metadata_client(O) lttng_ring_buffer_client_overwrite(O) lttng_ring_buffer_client_discard(O) lttng_tracer(O) lttng_kretprobes(O) lttng_ftrace(O) lttng_kprobes(O) lttng_statedump(O) lttng_lib_ring_buffer(O) cpufreq_ondemand loop e1000e kvm_intel kvm ptp pps_core [last unloaded: lttng_lib_ring_buffer]
[65357.287529] Pid: 0, comm: swapper/7 Tainted: G           O 3.9.4-trace-test #143
[65357.309694] Call Trace:
[65357.317022]  <IRQ>  [<ffffffff8103a3ef>] warn_slowpath_common+0x7f/0xc0
[65357.336893]  [<ffffffff8103a44a>] warn_slowpath_null+0x1a/0x20
[65357.354368]  [<ffffffffa0ff17b8>] lttng_event_write+0x118/0x140 [lttng_ring_buffer_client_mmap_overwrite]
[65357.383025]  [<ffffffffa100134f>] __event_probe__block_rq_with_error+0x1bf/0x220 [lttng_probe_block]
[65357.410376]  [<ffffffff812ea134>] blk_update_request+0x324/0x720
[65357.428364]  [<ffffffff812ea561>] blk_update_bidi_request+0x31/0x90
[65357.447136]  [<ffffffff812eb68c>] blk_end_bidi_request+0x2c/0x80
[65357.465127]  [<ffffffff812eb6f0>] blk_end_request+0x10/0x20
[65357.481822]  [<ffffffff81406b7c>] scsi_io_completion+0x9c/0x670
[65357.499555]  [<ffffffff813fe320>] scsi_finish_command+0xb0/0xe0
[65357.517283]  [<ffffffff81406965>] scsi_softirq_done+0xa5/0x140
[65357.534758]  [<ffffffff812f1d30>] blk_done_softirq+0x80/0xa0
[65357.551710]  [<ffffffff81043b00>] __do_softirq+0xe0/0x440
[65357.567881]  [<ffffffff81043ffe>] irq_exit+0x9e/0xb0
[65357.582754]  [<ffffffff81026465>] smp_call_function_single_interrupt+0x35/0x40
[65357.604388]  [<ffffffff8167be2f>] call_function_single_interrupt+0x6f/0x80
[65357.624976]  <EOI>  [<ffffffff8100ac06>] ? default_idle+0x46/0x300
[65357.643541]  [<ffffffff8100ac04>] ? default_idle+0x44/0x300
[65357.660235]  [<ffffffff8100b899>] cpu_idle+0x89/0xe0
[65357.675109]  [<ffffffff81664911>] start_secondary+0x220/0x227

Always from an event that can write a 0-length field as last field of
its payload, and it always happen directly on a sub-buffer boundary.

While we are there, check for length 0 in lib_ring_buffer_read_cstr()
too.

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

Fix: ring buffer: RING_BUFFER_FLUSH ioctl buffer corruption

lib_ring_buffer_switch_slow() clearly states:

 * Note, however, that as a v_cmpxchg is used for some atomic
 * operations, this function must be called from the CPU which owns the
 * buffer for a ACTIVE flush.

But unfortunately, the RING_BUFFER_FLUSH ioctl does not follow these
important directives. Therefore, whenever the consumer daemon or session
daemon explicitly triggers a "flush" on a buffer, it can race with data
being written to the buffer, leading to corruption of the reserve/commit
counters, and therefore corruption of data in the buffer. It triggers
these warnings for overwrite mode buffers:

[65356.890016] WARNING: at
/home/compudj/git/lttng-modules/wrapper/ringbuffer/../../lib/ringbuffer/../../wrapper/ringbuffer/../../lib/ringbuffer/backend.h:110 lttng_event_write+0x118/0x140 [lttng_ring_buffer_client_mmap_overwrite]()

Which indicates that we are trying to write into a sub-buffer for which
we don't have exclusive access. It also causes the following warnings to
show up:

[65335.896208] ring buffer relay-overwrite-mmap, cpu 5: records were lost. Caused by:
[65335.896208]   [ 0 buffer full, 80910 nest buffer wrap-around, 0 event too big ]

Which is caused by corrupted commit counter.

Fix this by sending an IPI to the CPU owning the flushed buffer for
per-cpu synchronization. For global synchronization, no IPI is needed,
since we allow writes from remote CPUs.

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

Cleanup: ring buffer: remove lib_ring_buffer_switch_new_end()

lib_ring_buffer_switch_new_end() is a leftover from the days where an
event that would exactly fill the current sub-buffer would automatically
trigger a sub-buffer switch into the next sub-buffer.

Even before the ring buffer code has been moved into lttng-modules, this
behavior had been changed: an event that exactly fills a sub-buffer only
fills this current sub-buffer, and does not need to switch into the
next one to populate the sub-buffer header. This change had been done so
periodical timer switch, which shares the same semantic as an event
exactly filling a sub-buffer, would not create tons of empty
sub-buffers.

However, when doing this change, lib_ring_buffer_switch_new_end() has
not been removed, but clearly should have been. Its job is now performed
by the event "commit".

lib_ring_buffer_switch_new_end() has no effect, since padding_size is
always 0.

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

print warning and return -EBUSY on double get_subbuf()

Each get_subbuf() should be paired with a put_subbuf().

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

Add header to fix compilation of syscalls-extractor on 3.9 kernel

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

Fix: don't do 0 byte event reservation

Caused sequence of:

lttng create
lttng enable-event -k sched_switch
lttng start
lttng stop
lttng start
lttng stop

to hang on waiting for data......

Issue appeared in

commit d83004aa718065e5221d8ef4b99e4302b70b3df9
Author: Julien Desfossez <jdesfossez@efficios.com>
Date:   Thu Jun 6 16:38:48 2013 -0400

    Metadata cache and multiple metadata channels

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

Document that payload size need to be greater than 0

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

Fix build and load against linux-2.6.33.x

* lttng-event.h declared but did not implement
  lttng_add_perf_counter_to_ctx on kernel >=2.6.33, the implementation
  was in lttng-context-perf-counters.c, which was only included for
  kernel >=2.6.34.  This prevented the module from being loaded.

* on kernel 2.6.33.x, lttng-context-perf-counters.c complains about
  implicit declaration for {get,put}_online_cpus and
  {,un}register_cpu_notifier; so fix header inclusion.

Signed-off-by: Samuel Martin <smartin@aldebaran-robotics.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: add missing CONFIG_COMPAT around lib_ring_buffer_compat_ioctl

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

Metadata cache and multiple metadata channels

Implement a cache for the kernel metadata and a list of metadata
channels.

When new metadata is appended, all metadata channels are awakened so
they can return from poll and get the newly added metadata
This allows to request the metadata multiple times by creating multiple
metadata channels (useful for snapshots).

With this new feature, the poll and get_subbuf ring buffer operations
are now overridden by lttng-abi for the metadata channels, to check the
cache before doing these operations.

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

fix block instrumentation probe signature mismatch for 3.9.x kernels

Was causing OOPS when tracing the block layer.

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

fix: block instrumentation: NULL pointer dereference

This backtrace appears with kernel 3.9.4:

[19311.313455] BUG: unable to handle kernel NULL pointer dereference at           (null)
[19311.316589] IP: [<ffffffffa076392b>] __event_probe__block_bio+0x8b/0x1a0 [lttng_probe_block]
[19311.316589] PGD 41808a067 PUD 4190a3067 PMD 0
[19311.316589] Oops: 0000 [#1] PREEMPT SMP
[19311.316589] Modules linked in: lttng_probe_writeback(O) lttng_probe_workqueue(O) lttng_probe_vmscan(O) lttng_probe_udp(O) lm
[19311.316589] CPU 4
[19311.316589] Pid: 1423, comm: kjournald Tainted: G           O 3.9.4-trace-test #143 Supermicro X7DAL/X7DAL
[19311.316589] RIP: 0010:[<ffffffffa076392b>]  [<ffffffffa076392b>] __event_probe__block_bio+0x8b/0x1a0 [lttng_probe_block]
[19311.316589] RSP: 0018:ffff88041c0d5aa8  EFLAGS: 00010246
[19311.316589] RAX: 0000000000000000 RBX: ffff88041a994c00 RCX: ffff88041b610800
[19311.316589] RDX: 0000000000000004 RSI: ffff88041c0d5b00 RDI: ffff88041c0d5ab0
[19311.316589] RBP: ffff88041c0d5b18 R08: fffffffffffc0000 R09: 0000000000000000
[19311.316589] R10: ffff88041a9947e8 R11: 0000000000000001 R12: ffff880419252c18
[19311.316589] R13: ffff88041c9e0a48 R14: ffff88041749bd10 R15: ffff88041749bd20
[19311.316589] FS:  0000000000000000(0000) GS:ffff88042fd00000(0000) knlGS:0000000000000000
[19311.316589] CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[19311.316589] CR2: 0000000000000000 CR3: 0000000417a0a000 CR4: 00000000000007e0
[19311.316589] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[19311.316589] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
[19311.316589] Process kjournald (pid: 1423, threadinfo ffff88041c0d4000, task ffff88041daa0000)
[19311.316589] Stack:
[19311.316589]  0000000000000000 ffff88041c3e4000 ffff880408987518 0000000000000024
[19311.316589]  0000000400000001 ffff88042fd19e30 000000000000002a 00000000000126c1
[19311.316589]  00000000000126bb 0000119043fa4efc ffff880400000000 0000000000000246
[19311.316589] Call Trace:
[19311.316589]  [<ffffffff812e79fb>] bio_attempt_back_merge+0xfb/0x1e0
[19311.316589]  [<ffffffff812ec96f>] blk_queue_bio+0x43f/0x570
[19311.316589]  [<ffffffff812ea68a>] generic_make_request+0xca/0x100
[19311.316589]  [<ffffffff81495bfe>] raid1_unplug+0x12e/0x170
[19311.316589]  [<ffffffff812ec36c>] blk_flush_plug_list+0x9c/0x210
[19311.316589]  [<ffffffff812ec4f8>] blk_finish_plug+0x18/0x50
[19311.316589]  [<ffffffff81243a48>] journal_commit_transaction+0x728/0x1f30
[19311.316589]  [<ffffffff8104e463>] ? try_to_del_timer_sync+0x53/0x70
[19311.316589]  [<ffffffff8104e485>] ? del_timer_sync+0x5/0xd0
[19311.316589]  [<ffffffff8124967d>] kjournald+0xed/0x290
[19311.316589]  [<ffffffff8106a270>] ? __init_waitqueue_head+0x60/0x60
[19311.316589]  [<ffffffff81249590>] ? commit_timeout+0x10/0x10
[19311.316589]  [<ffffffff810695fb>] kthread+0xdb/0xe0
[19311.316589]  [<ffffffff81672c60>] ? _raw_spin_unlock_irq+0x30/0x60
[19311.316589]  [<ffffffff81069520>] ? __init_kthread_worker+0x70/0x70
[19311.316589]  [<ffffffff8167af1c>] ret_from_fork+0x7c/0xb0
[19311.316589]  [<ffffffff81069520>] ? __init_kthread_worker+0x70/0x70

Fixes #559

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

Update rcu instrumentation to 3.9.[0-4]

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

Fix CPU hotplug section mismatches

Get rid of the following section mismatches:

WARNING: /home/jang/temp/lttng-modules-2.2.0-r0/git/lttng-tracer.o(.text+0x19dc0): Section mismatch in reference from the function lttng_add_perf_counter_to_ctx() to the function .cpuinit.text:lttng_perf_counter_cpu_hp_callback()
The function lttng_add_perf_counter_to_ctx() references
the function __cpuinit lttng_perf_counter_cpu_hp_callback().
This is often because lttng_add_perf_counter_to_ctx lacks a __cpuinit
annotation or the annotation of lttng_perf_counter_cpu_hp_callback is wrong.

WARNING: /home/jang/temp/lttng-modules-2.2.0-r0/git/lib/lttng-lib-ring-buffer.o(.text+0x1204): Section mismatch in reference from the function channel_backend_init() to the function .cpuinit.text:lib_ring_buffer_cpu_hp_callback()
The function channel_backend_init() references
the function __cpuinit lib_ring_buffer_cpu_hp_callback().
This is often because channel_backend_init lacks a __cpuinit
annotation or the annotation of lib_ring_buffer_cpu_hp_callback is wrong.

WARNING: /home/jang/temp/lttng-modules-2.2.0-r0/git/lib/lttng-lib-ring-buffer.o(.text+0x269c): Section mismatch in reference from the function channel_create() to the function .cpuinit.text:lib_ring_buffer_cpu_hp_callback()
The function channel_create() references
the function __cpuinit lib_ring_buffer_cpu_hp_callback().
This is often because channel_create lacks a __cpuinit
annotation or the annotation of lib_ring_buffer_cpu_hp_callback is wrong.

WARNING: /home/jang/temp/lttng-modules-2.2.0-r0/git/lib/lttng-lib-ring-buffer.o(.text+0x4a1c): Section mismatch in reference from the function channel_iterator_init() to the function .cpuinit.text:channel_iterator_cpu_hotplug()
The function channel_iterator_init() references
the function __cpuinit channel_iterator_cpu_hotplug().
This is often because channel_iterator_init lacks a __cpuinit
annotation or the annotation of channel_iterator_cpu_hotplug is wrong.

Signed-off-by: Jan Glauber <jan.glauber@gmail.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Add system calls definitions for powerpc-32

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

Remove bashism in lttng-syscalls-generate-headers.sh

Options to echo are not portable. In particular, the 'echo -e' option is
implemented by some shells, including bash, to expand escape sequences.
However, dash is one of the other family of shells that instead expands
escape sequences by default.

The printf command is portable and much more reliable.

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

Update ARM (32 bit) syscall tracepoints to 3.4

Update from 2.6.38 to 3.4. Added syscalls:
sys_acct
sys_swapon
sys_swapoff
sys_quotactl
sys_perf_event_open
sys_fanotify_init
sys_name_to_handle_at
sys_open_by_handle_at
sys_clock_adjtime
sys_syncfs
sys_sendmmsg
sys_setns
sys_process_vm_readv
sys_process_vm_writev

Signed-off-by: Jan Glauber <jan.glauber@gmail.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Version 2.2.0-rc2

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

Remove useless lttng_metadata probe

We don't use this event anymore since we write the metadata directly
into the ring buffer, no need for an external event. This probe was
the only one in the lttng-probe-lttng module, so we can get rid of
this module as well.

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

Fix: warning kvm_trace_exit_reason redefined for 3.9 kernel

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

Fix: 3.9.0 block layer instrumentation build

Linux upstream commit:

commit 0a82a8d132b26d438eb90b3ab35a7016e7227a1d
Author: Linus Torvalds <torvalds@linux-foundation.org>
Date:   Thu Apr 18 09:00:26 2013 -0700

    Revert "block: add missing block_bio_complete() tracepoint"

Requires us to revert in lttng-modules too.

Fixes #512

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

Revert "Stop build if CONFIG_TRACEPOINTS is not set"

This reverts commit 699f9673466e1a6537ad23cc697b79660639f260.

It breaks make clean.

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

Stop build if CONFIG_TRACEPOINTS is not set

Previously it just tries to compile with zero modules, which is
confusing if you thought you had configured everything correctly. Now it
throws an error which tells which went wrong.

Fixes #506

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

Add comments to endifs.

This should make it easier to read construction of the makefile.

Refs #506

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

Fix: remove ARM set_tls system call override

We'll need to find a better way to instrument ARM-specific system calls
located at a far offset from the standard systems calls. A 16MB
lttng-modules kernel module is really not acceptable.

Removing this instrumentation for now. sys_set_tls will appear as
sys_unknown.

Fixes #472

CC: Ryan Kyser <Ryan.Kyser@jci.com>
Ref: http://lists.lttng.org/pipermail/lttng-dev/2013-April/019990.html
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Cleanup: arm sys_mmap whitespace

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

Clean up using global_dirty_limit wrapper for writeback probe

Move the wrapper around reading of global_dirty_limit to /wrapper/
directory.  Introduce a new kallsyms_lookup_dataptr function for
obtaining the address unchanged and use it in global_dirty_limit
wrapper.  Since the data address is available only if
CONFIG_KALLSYMS_ALL is set, omit the whole probe from building if this
config is missing.

[ Edit by Mathieu Desnoyers: small coding style fixes ]

Signed-off-by: Andrew Gabbasov <andrew_gabbasov@mentor.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Version 2.2.0-rc1

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

Update README

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

Fix: vmscan instrumentation build

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

writeback instrumentation: remove duplicated structures

Don't replicate internal structures from the kernel: this is asking for
serious trouble, and could lead to breakage if building on newer kernels
that have modified structures.

The proper approach, if we really need to extract this information,
would be to add APIs to the Linux kernel workqueue that allow getting
this information.

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

block: update instrumentation to 3.9 rc kernel

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

rcu: update to 3.9 rc kernel instrumentation

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

workqueue instrumentation: remove duplicated structures

Don't replicate internal structures from the kernel: this is asking for
serious trouble, and could lead to breakage if building on newer kernels
that have modified structures.

The proper approach, if we really need to extract this information,
would be to add APIs to the Linux kernel workqueue that allow getting
this information.

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

Rename odd event names

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

Introduce API to remap event names exposed by LTTng

Some kernel event providers don't follow naming hierarchy. Rename them
internally within LTTng.

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

lttng-module: sched.h: Fix compilation on 3.9 kernel

With commit 8bd75c77b7c6a3954140dd2e20346aef3efe4a35
included in 3.9-rc1 kernel, rt specific bits in "linux/sched.h"
were moved into new header file "linux/sched/rt.h".

Fixes this compilation error:
 CC [M]  /home/majo/lttng/lttng-modules/probes/lttng-probe-sched.o
...
/home/majo/lttng/lttng-modules/probes/../instrumentation/events/lttng-module
/../../../probes/../instrumentation/events/lttng-module/sched.h:
In function '__event_probe__sched_switch':
/home/majo/lttng/lttng-modules/probes/../instrumentation/events/lttng-module
/../../../probes/../instrumentation/events/lttng-module/sched.h:164:1:
error: 'MAX_RT_PRIO' undeclared (first use in this function)
...

Signed-off-by: Maxin B. John <maxin.john@enea.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix Linux 3.8 btrfs probe warning

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

Fix Linux 3.8 ext4 probe support

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

Fix 3.8 kernel support: namespace lttng_is_signed_type()

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

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

Triggers rare sub-buffer corruption when a buffer is completely full and
a final flush is then performed.

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

instrumentation sched: add missing fork pid info

Fixes #434

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

Fix check in lttng_strlen_user_inatomic

__copy_from_user_inatomic returns the number of bytes that could not be
copied, not an error code. This fixes the test accordingly.

[ Edit by Mathieu Desnoyers: change "ret" type to unsigned long too. ]

Signed-off-by: Simon Marchi <simon.marchi@polymtl.ca>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

instrumentation: sync with Linux commit a1ce392

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

Fix: statedump: missing locking, use lttng_iterate_fd

- Take a spinlock around use of dentry->d_name.name.
- Use lttng_iterate_fd wrapper (and thus iterate_fd() from Linux kernels
  from v3.7 and onward).

Suggested-by: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Implement lttng_iterate_fd wrapper

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

Instrumentation cleanup: kmem events should print hex pointers

Fixes #422

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

Fix compilation of ext4 and btrfs tracepoints on 2.6.40 kernels

See f95480cf62e9acb42b74fce162544694b76c9704.

The new btrfs and ext4 tracepoints used 3.0.0 in the compilation logic.

Signed-off-by: Simon Marchi <simon.marchi@polymtl.ca>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: statedump hang/too early completion due to logic error

The previous "Fix: statedump hang due to incorrect wait/wakeup use" was
not actually fixing the real problem.

The issue is that we should pass the expected condition to wait_event()
rather than its contrary.

This bug has been sitting there for a while. I suspect that a recent
change in the Linux scheduler behavior for newly spawned worker threads
might have contributed to trigger the hang more reliably.

The effects of this bugs are:
- possible hang of the lttng-sessiond (within the kernel) at tracing
  start,
- the statedump end event is traced before all worker threads have
  actually completed, which can confuse LTTng viewer state systems.

Reported-by: Phil Wilshire <sysdcs@gmail.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: statedump hang due to incorrect wait/wakeup use

Reported-by: Phil Wilshire <sysdcs@gmail.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

put_ulong calls in lib_ring_buffer_compat_ioctl() should be compat

I was noticing the that some of the put_ulong calls in
lib_ring_buffer_compat_ioctl() were not compat_put_ulong.  I can't
imagine this is correct.

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

wrapper/perf.h: Fix kernel version condition

The pid_t parameter of function perf_event_create_kernel_counter
was changed to task_struct pointer starting from 2.6.37.x.

Signed-off-by: Andrew Gabbasov <andrew_gabbasov@mentor.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Add new kernel probes instrumentation

Add kernel probes for btrfs, compaction, ext4, printk, random, rcu,
regmap, rpm, sunrpc, workqueue, writeback.

Signed-off-by: Andrew Gabbasov <andrew_gabbasov@mentor.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Update kernel probes to more detailed match to kernel versions

Some ifdef's are added to kernel probes instrumentation to make them
more close to original tracepoints in different kernel versions.
Supported kernel version are from 2.6.32 to 3.7.

Signed-off-by: Andrew Gabbasov <andrew_gabbasov@mentor.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix possible kernel build errors with linux-patches

Kernel sources of version 2.6.32 - 2.6.34 with applied patches
from linux-patches may fail to compile if tracepoint samples
are configured to build. Some part of backported commits
are added to the kernel patches to avoid those errors.

Signed-off-by: Andrew Gabbasov <andrew_gabbasov@mentor.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Add missing MIT license text to 3 files under this license

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

Update sites using kernel version checking macro to new range

Signed-off-by: Andrew Gabbasov <andrew_gabbasov@mentor.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Make upper bound of kernel version checking macro exclusive

It's more usable to have the upper limit exclusive. It helps to avoid
hardcoding of stable branch highest version number, i.e. having a range
from 3.1.0 up to 3.2.0 (exclusively) gives us all 3.1.x versions.

Signed-off-by: Andrew Gabbasov <andrew_gabbasov@mentor.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

sock instrumentation: fix fields to get referenced values

Due to specific of passing values in lttng-modules, if it is supposed
to display the values, passing a pointer will not be enough,
we need to store the actual values.

Signed-off-by: Andrew Gabbasov <andrew_gabbasov@mentor.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

ext3 instrumentation: fix of assignment code conversion

Due to specifics of handling assignment code in lttng-modules,
plain code in TP_fast_assign (outside tp_* macros) will not be reached.
Everything should be enclosed into tp_* fragments.

Signed-off-by: Andrew Gabbasov <andrew_gabbasov@mentor.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

sched instrumentation: rename "pid" fields in sched_process_exec

Rename "pid" to "tid" fields in new code, similarly to what was
done earlier for all sched tracepoints.

Signed-off-by: Andrew Gabbasov <andrew_gabbasov@mentor.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Remove remaining semicolons in TP_fast_assign blocks

Signed-off-by: Andrew Gabbasov <andrew_gabbasov@mentor.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix compilation for 3.0 kernels that are named 2.6.40

Since some distro released the 3.0 kernel as 2.6.40, it might be useful
to adjust some checks to treat 2.6.40 kernels as 3.0.

Signed-off-by: Simon Marchi <simon.marchi@polymtl.ca>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix compilation for 3.0 branch (>= 3.0.39)

The isolate_mode_t type that appeared in 3.2 was backported to 3.0.39 so
the version check must be fixed. It was not backported to the 3.1 branch
though, so it must be excluded.

Signed-off-by: Simon Marchi <simon.marchi@polymtl.ca>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Add kernel version checking macro

It contains a macro to help checking for kernel version ranges.

Signed-off-by: Simon Marchi <simon.marchi@polymtl.ca>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Remove semicolons in TP_fast_assign blocks

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

Add ifdefs to net probe to support Linux 2.6.39

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

Add kernel probes for supplementary subsystems

Add probes for: asoc, ext3, gpio, jbd, jbd2, kmem, lock, module, napi,
net, power, regulator, scsi, skb, sock, udp, vmscan subsystems.

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

Check if interrupt action name is not null in statedump

Protection against poorly written drivers who give a null name to their
irq handlers.

Signed-off-by: Simon Marchi <simon.marchi@polymtl.ca>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix exec instrumentation: only for kernels 3.4.0 and higher

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

Add TRACE_EVENT(sched_process_exec) to sched.h

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

lib/ringbuffer/frontend_api.h: Include linux/prefetch.h.

prefetch() is used in this file so make sure we include the appropriate
header before using it.

Signed-off-by: Will Newton <will.newton@imgtec.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix warning with 3.6.0 kernel

No actual runtime impact, but neater anyway.

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

ABI with support for compat 32/64 bits

The current ABI does not work for compat 32/64 bits.
This patch moves the current ABI as old-abi and provides a new ABI in
which all the structures exchanged between user and kernel-space are
packed. Also this new ABI moves the "int overwrite" member of the
struct lttng_kernel_channel to remove the alignment added by the
compiler.

A patch for lttng-tools has been developed in parallel to this one to
support the new ABI. These 2 patches have been tested in all
possible configurations (applied or not) on 64-bit and 32-bit kernels
(with CONFIG_COMPAT) and a user-space in 32 and 64-bit.

Here are the results of the tests :
k 64 compat     | u 32 compat     | OK
k 64 compat     | u 64 compat     | OK
k 64 compat     | u 32 non-compat | KO
k 64 compat     | u 64 non-compat | OK

k 64 non-compat | u 64 compat     | OK
k 64 non-compat | u 32 compat     | KO
k 64 non-compat | u 64 non-compat | OK
k 64 non-compat | u 32 non-compat | KO

k 32 compat     | u compat        | OK
k 32 compat     | u non-compat    | OK

k 32 non-compat | u compat        | OK
k 32 non-compat | u non-compat    | OK

The results are as expected :
- on 32-bit user-space and kernel, every configuration works.
- on 64-bit user-space and kernel, every configuration works.
- with 32-bit user-space on a 64-bit kernel the only configuration
  where it works is when the compat patch is applied everywhere.

Signed-off-by: Julien Desfossez <jdesfossez@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>

Use uint64_t for packet header content size and packet size

Allow packets larger than 256MB (4GB / 8 bits per byte).

Fixes #354

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

Fix: compat layer only on supported architectures

The compat_ulong_t type is only defined on some architectures, so we
need to protect the compat defines with the ifdef CONFIG_COMPAT.

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

Fix ring buffer 32/64 bits compat

The type of fields exchanged between kernel and userspace must be
compat_ulong_t instead of unsigned long in case of compat where
userspace is 32 bits and kernel is 64 bits.

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

Fix cleanup: move lttng-tracer-core.h include to lib ringbuffer config.h

This is where it is used. Add a #ifdef check for the LTTNG_TRACER_CORE_H
define to ensure that the header it depends on is always included.

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

Fix ring_buffer_frontend.c: missing include lttng-tracer-core.h

In lib/ringbuffer/ring_buffer_frontend.c, RING_BUFFER_ALIGN is undefined,
leading to no alignment offset being recorded after the call to
config->cb.record_header_size() in lib_ring_buffer_try_reserve_slow().

However, lttng-ring-buffer-client.h does define RING_BUFFER_ALIGN, so
the alignment offset will be produced when the packet header is written
in lttng_write_event_header().

This discrepancy may be observed on architectures that don't set
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS, such as ARM, with a babeltrace
error such as:

babeltrace: ../../include/babeltrace/ctf/types.h:206: ctf_pos_get_event:
Assertion `pos->offset <= pos->content_size' failed.
Aborted

indicating that the actual content size differs from the calculated one
due to the difference in alignment.  Including the appropriate header
file in ring_buffer_frontend.c solves the problem.

Signed-off-by: Wade Farnsworth <wade_farnsworth@mentor.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Fix: statedump missing check for NULL namespace

Cause kernel OOPS. statedump PID namespace support is newly introduced
in lttng-modules 2.1. Happens when iterating on a zombie process.

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

Update version to 2.1.0-rc1

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

fix timestamps on architectures without CONFIG_KTIME_SCALAR

trace_clock_monotonic_wrapper() should return a u64 representing the
number of nanoseconds since system startup.
ktime_get() provides that value directly within its .tv64 field only
on those architectures defining CONFIG_KTIME_SCALAR, whereas in all
other cases (e.g. PowerPC) a ktime_to_ns() conversion (which
translates back to .tv64 when CONFIG_KTIME_SCALAR is defined)
becomes necessary.

Signed-off-by: Gerlando Falauto <gerlando.falauto@keymile.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

Support for linux kernels 2.6.32 through 2.6.37

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

Document limitation of vppid and ppid context wrt eventual RCU instrumentation

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

Cleanup: no need to hold RCU read-side lock when reading current nsproxy

As documented in include/linux/nsproxy.h.

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

Add env hostname information

Approximation in case of hostname change or in case of
contextualisation, but it does the job for 95% of use-cases.

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

Fix: lttng_statedump_process_state for each PID NS has infinite loop

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

lttng_statedump_process_state for each PID NS

When a process is in a namespace, its pid, tid and ppid are relative to
the namespace. Since namespaces can be nested, we need to know the
representation of each process in each namespace.

This patch changes the lttng_enumerate_task_fd to iterate over each
PID namespace of a process if any, that way we generate, in the
statedump, an entry for each process in each namespace it belongs.
To know the nesting level, the field "level" is added to the
lttng_statedump_process_state event, 0 being the top-level.
For processes running on the top-level namespace, the statedump
behaviour is unchanged (except the added "level" field).

For example (no nesting, just one level of namespace) :
lttng_statedump_process_state: {
        tid = 32185, vtid = 1, pid = 32185,
        vpid = 1, ppid = 32173, vppid = 0,
        level = 1, name = "init" }
lttng_statedump_process_state: {
        tid = 32185, vtid = 32185, pid = 32185,
        vpid = 32185, ppid = 32173, vppid = 32173,
        level = 0, name = "init" }

Confirmed that the process 32173 in the top-level namespace is indeed
the lxc-start command that created the container and its namespace.

[ Edit by Mathieu Desnoyers: Minor cleanups ]

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

Support the hostname context

This is particularly useful when tracing a machine with containers
(lxc), that way we can easily distinguish events generated inside
a container.

[ Edit by Mathieu Desnoyers: various fix. ]

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

Fix: statedump namespaced pid, tid and ppid

Use the target process namespace to lookup the value (pid, tid or ppid),
not the current task.

[ Edit by Mathieu Desnoyers: removed extra p->nsproxy check. Already
  performed within rcu read-side critical section. ]

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

Fix: statedump: disable vm maps enumeration

We need the tasklist lock to correctly dump these, which is unavailable
to modules. Disable this feature for now.

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

Fix: ensure userspace accesses are done with _inatomic

Otherwise, triggers scheduling while atomic (might_sleep()) warnings,
since we call those from a tracepoint probe (with preemption disabled).

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

Fix: vppid context should test for current nsproxy

Triggers a NULL pointer exception with tracing scheduling events with
vppid context activated.

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