Re: [PATCH 5/5] proc: export more page flags in /proc/kpageflags

[Wu Fengguang <fengguang.wu@intel.com>]

On Tue, Apr 28, 2009 at 05:25:06PM +0800, Pekka Enberg wrote:
> On Tue, 2009-04-28 at 09:40 +0200, Andi Kleen wrote:
> >>> > > I think i have to NAK this kind of ad-hoc instrumentation of kernel
> >>> > > internals and statistics until we clear up why such instrumentation
> 
> * Pekka Enberg <penberg@cs.helsinki.fi> wrote:
> >>> > I think because it has zero fast path overhead and can be used
> >>> > any time without enabling anything special.
> >
> > On Tue, Apr 28, 2009 at 12:15 PM, Ingo Molnar <mingo@elte.hu> wrote:
> >> ( That's a dubious claim in any case - tracepoints are very cheap.
> >>  And they could be made even cheaper and such efforts would benefit
> >>  all the tracepoint users so it's a prime focus of interest.
> >>  Andi is a SystemTap proponent, right? I saw him oppose pretty much
> >>  everything built-in kernel tracing related. I consider that a
> >>  pretty extreme position. )
> 
> On Tue, Apr 28, 2009 at 12:19 PM, Pekka Enberg <penberg@cs.helsinki.fi> wrote:
> > I have no idea how expensive tracepoints are but I suspect they don't
> > make too much sense for this particular scenario. After all, kmemtrace
> > is mainly interested in _allocation patterns_ whereas this patch seems
> > to be more interested in "memory layout" type of things.
> 
> That said, I do foresee a need to be able to turn on more detailed
> tracing after you've identified problematic areas from kpageflags type
> of overview report. And for that, you almost certainly want
> kmemtrace/tracepoints style solution with pid/function/whatever regexp
> matching ftrace already provides.

Exactly - kmemtrace is the tool I looked for when hunting down the
page flags of the leaked ring buffer memory :-)

Thanks,
Fengguang