Re: [PATCH v2 4/5] mm/compaction: more trace to understand when/why compaction start/finish

[Vlastimil Babka]

On 01/13/2015 08:16 AM, Joonsoo Kim wrote:
> On Mon, Jan 12, 2015 at 04:53:53PM +0100, Vlastimil Babka wrote:
>> On 01/12/2015 09:21 AM, Joonsoo Kim wrote:
>> > It is not well analyzed that when/why compaction start/finish or not. With
>> > these new tracepoints, we can know much more about start/finish reason of
>> > compaction. I can find following bug with these tracepoint.
>> > 
>> > http://www.spinics.net/lists/linux-mm/msg81582.html
>> > 
>> > Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@xxxxxxx>
>> > ---
>> >  include/linux/compaction.h        |    3 ++
>> >  include/trace/events/compaction.h |   94 +++++++++++++++++++++++++++++++++++++
>> >  mm/compaction.c                   |   41 ++++++++++++++--
>> >  3 files changed, 134 insertions(+), 4 deletions(-)
>> > 
>> > diff --git a/include/linux/compaction.h b/include/linux/compaction.h
>> > index a9547b6..d82181a 100644
>> > --- a/include/linux/compaction.h
>> > +++ b/include/linux/compaction.h
>> > @@ -12,6 +12,9 @@
>> >  #define COMPACT_PARTIAL		3
>> >  /* The full zone was compacted */
>> >  #define COMPACT_COMPLETE	4
>> > +/* For more detailed tracepoint output */
>> > +#define COMPACT_NO_SUITABLE_PAGE	5
>> > +#define COMPACT_NOT_SUITABLE_ZONE	6
>> >  /* When adding new state, please change compaction_status_string, too */
>> >  
>> >  /* Used to signal whether compaction detected need_sched() or lock contention */
>> > diff --git a/include/trace/events/compaction.h b/include/trace/events/compaction.h
>> > index 139020b..839dd4f 100644
>> > --- a/include/trace/events/compaction.h
>> > +++ b/include/trace/events/compaction.h
>> > @@ -164,6 +164,100 @@ TRACE_EVENT(mm_compaction_end,
>> >  		compaction_status_string[__entry->status])
>> >  );
>> >  
>> > +TRACE_EVENT(mm_compaction_try_to_compact_pages,
>> > +
>> > +	TP_PROTO(
>> > +		int order,
>> > +		gfp_t gfp_mask,
>> > +		enum migrate_mode mode,
>> > +		int alloc_flags,
>> > +		int classzone_idx),
>> 
>> I wonder if alloc_flags and classzone_idx is particularly useful. It affects the
>> watermark checks, but those are a bit of blackbox anyway.
> 
> Yes, I think so. How about printing gfp_flag rather than these? It would
> tell us migratetype and other information so would be useful.

Yeah gfp_mask should be enough.

> 
> Thanks.
> 

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/