diff options
| author | Steven Rostedt (Red Hat) <rostedt@goodmis.org> | 2014-02-11 13:38:54 -0500 |
|---|---|---|
| committer | Steven Rostedt <rostedt@goodmis.org> | 2014-02-11 13:38:54 -0500 |
| commit | d651aa1d68a2f0a7ee65697b04c6a92f8c0a12f2 (patch) | |
| tree | 0261f1a636d376a4e61f0e09b0f5e69aea3fc3b0 /kernel/trace | |
| parent | b28a960c42fcd9cfc987441fa6d1c1a471f0f9ed (diff) | |
ring-buffer: Fix first commit on sub-buffer having non-zero delta
Each sub-buffer (buffer page) has a full 64 bit timestamp. The events on
that page use a 27 bit delta against that timestamp in order to save on
bits written to the ring buffer. If the time between events is larger than
what the 27 bits can hold, a "time extend" event is added to hold the
entire 64 bit timestamp again and the events after that hold a delta from
that timestamp.
As a "time extend" is always paired with an event, it is logical to just
allocate the event with the time extend, to make things a bit more efficient.
Unfortunately, when the pairing code was written, it removed the "delta = 0"
from the first commit on a page, causing the events on the page to be
slightly skewed.
Fixes: 69d1b839f7ee "ring-buffer: Bind time extend and data events together"
Cc: stable@vger.kernel.org # 2.6.37+
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Diffstat (limited to 'kernel/trace')
| -rw-r--r-- | kernel/trace/ring_buffer.c | 7 |
1 files changed, 7 insertions, 0 deletions
diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 294b8a271a0..fc4da2d97f9 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -2397,6 +2397,13 @@ __rb_reserve_next(struct ring_buffer_per_cpu *cpu_buffer, write &= RB_WRITE_MASK; tail = write - length; + /* + * If this is the first commit on the page, then it has the same + * timestamp as the page itself. + */ + if (!tail) + delta = 0; + /* See if we shot pass the end of this buffer page */ if (unlikely(write > BUF_PAGE_SIZE)) return rb_move_tail(cpu_buffer, length, tail, |