Re: [PATCH v1 3/9] perf util: Reconstruct rblist for supporting per-thread shadow stats

[Jin, Yao]

On 11/21/2017 11:17 PM, Jiri Olsa wrote:
> On Mon, Nov 20, 2017 at 10:43:38PM +0800, Jin Yao wrote:
> > In current stat-shadow.c, the rblist deleting is ignored.
> >
> > The patch reconstructs the codes of rblist init/free, and adds
> > the implementation to node_delete method of rblist.
> >
> > This patch also does:
> >
> > 1. Add the ctx/type/stat into rbtree keys because we will
> > use this rbtree to maintain shadow metrics to replace original
> > a couple of static arrays for supporting per-thread shadow stats.
> >
> > 2. Create a static runtime_stat variable 'rt_stat' which
> > will log the shadow metrics by default.
> >
> > Signed-off-by: Jin Yao <yao.jin@xxxxxxxxxxxxxxx>
> > ---
> >   tools/perf/util/stat-shadow.c | 62 ++++++++++++++++++++++++++++++++++++++++---
> >   tools/perf/util/stat.h        |  2 ++
> >   2 files changed, 60 insertions(+), 4 deletions(-)
> >
> > diff --git a/tools/perf/util/stat-shadow.c b/tools/perf/util/stat-shadow.c
> > index 5853901..045e129 100644
> > --- a/tools/perf/util/stat-shadow.c
> > +++ b/tools/perf/util/stat-shadow.c
> > @@ -40,12 +40,16 @@ static struct stats runtime_aperf_stats[NUM_CTX][MAX_NR_CPUS];
> >   static struct rblist runtime_saved_values;
> >   static bool have_frontend_stalled;
> >   
> > +static struct runtime_stat rt_stat;
> >   struct stats walltime_nsecs_stats;
> >   
> >   struct saved_value {
> >   	struct rb_node rb_node;
> >   	struct perf_evsel *evsel;
> > +	enum stat_type type;
> > +	int ctx;
> >   	int cpu;
> > +	struct runtime_stat *stat;
> >   	struct stats stats;
> >   };
> >   
> > @@ -58,6 +62,23 @@ static int saved_value_cmp(struct rb_node *rb_node, const void *entry)
> >   
> >   	if (a->cpu != b->cpu)
> >   		return a->cpu - b->cpu;
> > +
> > +	if (a->type != b->type)
> > +		return a->type - b->type;
> > +
> > +	if (a->ctx != b->ctx)
> > +		return a->ctx - b->ctx;
> > +
>
> could you please comment in here on the cases where
> evsel is defined and when not
>
> jirka

Thanks for reminding.

I will add comments to describe the cases where and when the evsel is 
set or not.

Thanks
Jin Yao

> > +	if (a->evsel == NULL && b->evsel == NULL) {
> > +		if (a->stat == b->stat)
> > +			return 0;
> > +
> > +		if ((char *)a->stat < (char *)b->stat)
> > +			return -1;
> > +
> > +		return 1;
> > +	}
> > +
>
> SNIP