Messages in this thread | | | Date | Wed, 2 Mar 2011 19:49:53 +0100 | From | Ingo Molnar <> | Subject | Re: [RFC PATCH 4/5] mm: Add hit/miss accounting for Page Cache |
| |
* Dave Hansen <dave@linux.vnet.ibm.com> wrote:
> On Wed, 2011-03-02 at 09:45 +0100, Ingo Molnar wrote: > > But, instead of trying to improve those aspects of our existing instrumentation > > frameworks, mm/* is gradually growing its own special instrumentation hacks, missing > > the big picture and fragmenting the instrumentation space some more. > > > > That trend is somewhat sad. > > Go any handy examples of how you'd like to see these done?
There's a very, very old branch in tip:tracing/mm (by Steve) that shows off some of the concepts that could be introduced, to 'dump' current MM state via an extension to the tracepoints APIs:
3383e37ea796: tracing, page-allocator: Add a postprocessing script for page-allocator-related ftrace events c33b3596bc38: tracing, page-allocator: Add trace event for page traffic related to the buddy lists 0d524fb734bc: tracing, mm: Add trace events for anti-fragmentation falling back to other migratetypes b9a28177eedf: tracing, page-allocator: Add trace events for page allocation and page freeing 807243eb20b2: Merge branch 'perfcounters/urgent' into tracing/mm 08b6cb88eeb5: perf_counter tools: Provide default bfd_demangle() function in case it's not around eb4671011887: tracing/mm: rename 'trigger' file to 'dump_range' 1487a7a1ff99: tracing/mm: fix mapcount trace record field dcac8cdac1d4: tracing/mm: add page frame snapshot trace
That's just a demo in essence - showing what things could be done in this area.
You can pick those commits up via running:
http://people.redhat.com/mingo/tip.git/README
Thanks,
Ingo
| |