Re: [PATCH v5 2/2] perf tool: improves DSO long names lookup speed with rbtree

[Waiman Long <waiman.long@hp.com>]

On 09/30/2014 11:21 AM, Arnaldo Carvalho de Melo wrote:
> Em Mon, Sep 29, 2014 at 04:07:29PM -0400, Waiman Long escreveu:
>> With workload that spawns and destroys many threads and processes,
>> it was found that perf-mem could took a long time to post-process
>> the perf data after the target workload had completed its operation.
>> The performance bottleneck was found to be the lookup and insertion
>> of the new DSO structures (thousands of them in this case).
>>
>> In a dual-socket Ivy-Bridge E7-4890 v2 machine (30-core, 60-thread),
>> the perf profile below shows what perf was doing after the profiled
>> AIM7 shared workload completed:
>>
>> -     83.94%  perf  libc-2.11.3.so     [.] __strcmp_sse42
>>     - __strcmp_sse42
>>        - 99.82% map__new
>>             machine__process_mmap_event
>>             perf_session_deliver_event
>>             perf_session__process_event
>>             __perf_session__process_events
>>             cmd_record
>>             cmd_mem
>>             run_builtin
>>             main
>>             __libc_start_main
>> -     13.17%  perf  perf               [.] __dsos__findnew
>>       __dsos__findnew
>>       map__new
>>       machine__process_mmap_event
>>       perf_session_deliver_event
>>       perf_session__process_event
>>       __perf_session__process_events
>>       cmd_record
>>       cmd_mem
>>       run_builtin
>>       main
>>       __libc_start_main
>>
>> So about 97% of CPU times were spent in the map__new() function
>> trying to insert new DSO entry into the DSO linked list. The whole
>> post-processing step took about 9 minutes.
>>
>> The DSO structures are currently searched linearly. So the total
>> processing time will be proportional to n^2.
>>
>> To overcome this performance problem, the DSO code is modified to
>> also put the DSO structures in a RB tree sorted by its long name
>> in additional to being in a simple linked list. With this change,
>> the processing time will become proportional to n*log(n) which will
>> be much quicker for large n. However, the short name will still be
>> searched using the old linear searching method.  With that patch
>> in place, the same perf-mem post-processing step took less than 30
>> seconds to complete.
>>
>> Signed-off-by: Waiman Long<Waiman.Long@hp.com>
>> ---
>>   tools/perf/util/dso.c     |   72 ++++++++++++++++++++++++++++++++++++++++++--
>>   tools/perf/util/dso.h     |    1 +
>>   tools/perf/util/machine.c |    1 +
>>   tools/perf/util/machine.h |    4 ++-
>>   4 files changed, 73 insertions(+), 5 deletions(-)
>>
>> diff --git a/tools/perf/util/dso.c b/tools/perf/util/dso.c
>> index 901a58f..9a81c03 100644
>> --- a/tools/perf/util/dso.c
>> +++ b/tools/perf/util/dso.c
>> @@ -653,6 +653,67 @@ struct dso *dso__kernel_findnew(struct machine *machine, const char *name,
>>   	return dso;
>>   }
>>
>> +/*
>> + * Find a matching entry and/or link current entry to RB tree.
>> + * Either one of the dso or name parameter must be non-NULL or the
>> + * function will not work.
>> + */
>> +static struct dso *dso__findlink_by_longname(struct rb_root *root,
>> +					     struct dso *dso, const char *name)
>> +{
>> +	struct rb_node **p =&root->rb_node;
>> +	struct rb_node  *parent = NULL;
>> +	int warned = false;
>> +
>> +	if (!name)
>> +		name = dso->long_name;
>> +	/*
>> +	 * Find node with the matching name
>> +	 */
>> +	while (*p) {
>> +		struct dso *this = rb_entry(*p, struct dso, rb_node);
>> +		int rc = strcmp(name, this->long_name);
>> +
>> +		parent = *p;
>> +		if (rc == 0) {
>> +			/*
>> +			 * In case the new DSO is a duplicate of an existing
>> +			 * one, print an one-time warning&  put the new entry
>> +			 * at the end of the list of duplicates.
>> +			 */
>> +			if (!dso || (dso == this))
>> +				return this;	/* Find matching dso */
>> +			/*
>> +			 * The core kernel DSOs may have duplicated long name.
>> +			 * (See dso__load_sym()). Don't print warning for them.
>> +			 */
>> +			if (!warned&&  !strstr(name, "kernel.kallsyms")
>> +				&&  !strstr(name, "/vmlinux")) {
>> +				pr_warning("Duplicated dso long name: %s\n",
>> +					   name);
>> +				warned = true;
> I still wonder if in this case we should just return, i.e. why would we
> want to have multiple entries with the same name here? Anyway, I guess
> it doesn't hurt, right?
>
> Something to be further investigated to find a better solution, but I
> guess that the patch as-is now should provide that speedup without
> introducing any new oddities. Will apply.

If I don't add the kernel name check, I will get a warning every time I 
run mem recording with the workloads that I am using. So it is happening 
in the current code. I think the short name may be different. I will do 
more test to find out. If that is the case, an alternative is to do a 
short name comparison if the long name match.

>> +			}
>> +			rc = 1;
>> +		}
>> +		if (rc<  0)
>> +			p =&parent->rb_left;
>> +		else
>> +			p =&parent->rb_right;
>> +	}
>> +	if (dso) {
>> +		/* Add new node and rebalance tree */
>> +		rb_link_node(&dso->rb_node, parent, p);
>> +		rb_insert_color(&dso->rb_node, root);
>> +	}
>> +	return NULL;
>> +}
>> +
>> +static inline struct dso *
>> +dso__find_by_longname(struct rb_root *root, const char *name)
>> +{
>> +	return dso__findlink_by_longname(root, NULL, name);
>> +}
>> +
>>   void dso__set_long_name(struct dso *dso, const char *name, bool name_allocated)
>>   {
>>   	if (name == NULL)
>> @@ -755,6 +816,7 @@ struct dso *dso__new(const char *name)
>>   		dso->a2l_fails = 1;
>>   		dso->kernel = DSO_TYPE_USER;
>>   		dso->needs_swap = DSO_SWAP__UNSET;
>> +		RB_CLEAR_NODE(&dso->rb_node);
>>   		INIT_LIST_HEAD(&dso->node);
>>   		INIT_LIST_HEAD(&dso->data.open_entry);
>>   	}
>> @@ -765,6 +827,10 @@ struct dso *dso__new(const char *name)
>>   void dso__delete(struct dso *dso)
>>   {
>>   	int i;
>> +
>> +	if (!RB_EMPTY_NODE(&dso->rb_node))
>> +		pr_err("DSO %s is still in rbtree when being deleted!\n",
>> +		       dso->long_name);
>>   	for (i = 0; i<  MAP__NR_TYPES; ++i)
>>   		symbols__delete(&dso->symbols[i]);
>>
>> @@ -854,6 +920,7 @@ bool __dsos__read_build_ids(struct list_head *head, bool with_hits)
>>   void dsos__add(struct dsos *dsos, struct dso *dso)
>>   {
>>   	list_add_tail(&dso->node,&dsos->head);
>> +	dso__findlink_by_longname(&dsos->root, dso, NULL);
>>   }
>>
>>   struct dso *dsos__find(const struct dsos *dsos, const char *name,
>> @@ -867,10 +934,7 @@ struct dso *dsos__find(const struct dsos *dsos, const char *name,
>>   				return pos;
>>   		return NULL;
>>   	}
>> -	list_for_each_entry(pos,&dsos->head, node)
>> -		if (strcmp(pos->long_name, name) == 0)
>> -			return pos;
>> -	return NULL;
>> +	return dso__find_by_longname((struct rb_root *)&dsos->root, name);
> Why do you need this cast? Humm, because in the end it will get to a
> function that either does insertion or does a simple search. Ok, I think
> that dso__find_by_longname is the closest to that thing where the cast
> should be applied, after making dso__find_by_longname receive a const
> rb_root pointer.
>
> I.e. the dso__find_by_longname name implies it will not change any of
> its parameters, its supposed to be a simple search. I will do this
> change while applying it.
>
> - Arnaldo

Yes, you are right. I should do the casting in dso__find_by_longname(). 
Please make the adjustment.

Thanks,
Longman