diff options
author | Steven Rostedt <srostedt@redhat.com> | 2012-05-14 17:02:33 -0400 |
---|---|---|
committer | Steven Rostedt <rostedt@goodmis.org> | 2012-05-16 19:50:23 -0400 |
commit | 659f451ff21315ebfeeb46b9adccee8ce1b52c25 (patch) | |
tree | da818329ad9c47e3a5c5414e37e836cb8bda5bec | |
parent | 5040b4b7bcc26a311c799d46f67174bcb20d05dd (diff) | |
download | op-kernel-dev-659f451ff21315ebfeeb46b9adccee8ce1b52c25.zip op-kernel-dev-659f451ff21315ebfeeb46b9adccee8ce1b52c25.tar.gz |
ring-buffer: Add integrity check at end of iter read
There use to be ring buffer integrity checks after updating the
size of the ring buffer. But now that the ring buffer can modify
the size while the system is running, the integrity checks were
removed, as they require the ring buffer to be disabed to perform
the check.
Move the integrity check to the reading of the ring buffer via the
iterator reads (the "trace" file). As reading via an iterator requires
disabling the ring buffer, it is a perfect place to have it.
If the ring buffer happens to be disabled when updating the size,
we still perform the integrity check.
Cc: Vaibhav Nagarnaik <vnagarnaik@google.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
-rw-r--r-- | kernel/trace/ring_buffer.c | 29 |
1 files changed, 29 insertions, 0 deletions
diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index d673ef0..e0573c5 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -1599,6 +1599,29 @@ int ring_buffer_resize(struct ring_buffer *buffer, unsigned long size, } out: + /* + * The ring buffer resize can happen with the ring buffer + * enabled, so that the update disturbs the tracing as little + * as possible. But if the buffer is disabled, we do not need + * to worry about that, and we can take the time to verify + * that the buffer is not corrupt. + */ + if (atomic_read(&buffer->record_disabled)) { + atomic_inc(&buffer->record_disabled); + /* + * Even though the buffer was disabled, we must make sure + * that it is truly disabled before calling rb_check_pages. + * There could have been a race between checking + * record_disable and incrementing it. + */ + synchronize_sched(); + for_each_buffer_cpu(buffer, cpu) { + cpu_buffer = buffer->buffers[cpu]; + rb_check_pages(cpu_buffer); + } + atomic_dec(&buffer->record_disabled); + } + mutex_unlock(&buffer->mutex); return size; @@ -3750,6 +3773,12 @@ ring_buffer_read_finish(struct ring_buffer_iter *iter) { struct ring_buffer_per_cpu *cpu_buffer = iter->cpu_buffer; + /* + * Ring buffer is disabled from recording, here's a good place + * to check the integrity of the ring buffer. + */ + rb_check_pages(cpu_buffer); + atomic_dec(&cpu_buffer->record_disabled); atomic_dec(&cpu_buffer->buffer->resize_disabled); kfree(iter); |