[lkml]   [2010]   [Dec]   [13]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
    SubjectRe: [PATCH 1/2] perf tools: Add reference timestamp to perf header
    On Mon, 2010-12-13 at 10:49 -0700, David S. Ahern wrote:
    > On 12/13/10 10:47, Peter Zijlstra wrote:
    > > On Mon, 2010-12-13 at 18:35 +0100, Peter Zijlstra wrote:
    > >> * this CLOCK_MONOTONIC offset (for what little good that does, since
    > >> our clock isn't strictly sync'ed to CLOCK_MONOTONIC so we can incur
    > >> arbitrary drift).
    > >
    > > In fact, the only sane way to do that is by creating a software counter
    > > that represents CLOCK_MONOTONIC and sample that say once a minute (or
    > > more often if you want smaller drift).
    > >
    > >
    > What about creating a PERF_RECORD_TIME and generate an event when the
    > counter is opened? It contains a PERF_SAMPLE_TIME and say
    > PERF_SAMPLE_TOD (time-of-day)? We're not sending rockets to saturn; we
    > just need the timestamps to match other log files.

    That's similar to the first thing I proposed. The problem is with long
    record sessions your drift can become quite significant, then when you
    merge sort your other log events stuff can get out of order. Which can
    lead to some serious head-scratching..

    Another problem with this approach is things like flight-recorder mode
    where you constantly over-write your old data, you'd have to build some
    trigger to always output a new record before you over-write the old one,
    so there's always one consistent record around. Drift is an even more
    serious problem here since flight-record more could be running for days
    before (if ever) you dump it.

     \ /
      Last update: 2010-12-13 18:59    [W:0.027 / U:46.936 seconds]
    ©2003-2017 Jasper Spaans. hosted at Digital OceanAdvertise on this site