A discussion follow-up with Linus made me change the implementation to put the fadvise after the synchronous page writeback, to ensure better streaming behavior.
Mathieu * Mathieu Desnoyers ([email protected]) wrote: > Hi, > > I just released ltt-control 0.85 which waits for previous subbuffers to be > written to disk (using sync_file_range()) and uses fadvise to tell the kernel > that pages won't be reaccessed. This lessens the tracer impact on the page > cache. > > Thanks to Linus for hints on how to use fadvise and sync_file_range to do > this, c.f.: > > http://marc.info/?l=linux-kernel&m=127430576309253&w=2 > > Mathieu > > -- > Mathieu Desnoyers > Operating System Efficiency R&D Consultant > EfficiOS Inc. > http://www.efficios.com -- Mathieu Desnoyers Operating System Efficiency R&D Consultant EfficiOS Inc. http://www.efficios.com _______________________________________________ ltt-dev mailing list [email protected] http://lists.casi.polymtl.ca/cgi-bin/mailman/listinfo/ltt-dev
