linux-kernel - Re: [PATCH 5/6 v4] cfq-iosched: CFQ group hierarchical scheduling and use

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <4D55EEBB.8060304@cn.fujitsu.com>
Date:	Sat, 12 Feb 2011 10:21:47 +0800
From:	Gui Jianfeng <guijianfeng@...fujitsu.com>
To:	Vivek Goyal <vgoyal@...hat.com>
CC:	Jens Axboe <axboe@...nel.dk>, Shaohua Li <shaohua.li@...el.com>,
	lkml <linux-kernel@...r.kernel.org>,
	Chad Talbott <ctalbott@...gle.com>,
	Divyesh Shah <dpshah@...gle.com>
Subject: Re: [PATCH 5/6 v4] cfq-iosched: CFQ group hierarchical scheduling
 and use_hierarchy interface

Vivek Goyal wrote:
> On Thu, Feb 10, 2011 at 03:47:45PM +0800, Gui Jianfeng wrote:
>> CFQ group hierarchical scheduling and use_hierarchy interface.
>>
> 
> Hi Gui,
> 
> I have done a quick high level review. Some minor comments inline.
> 
> [..]
>>  struct cfq_data {
>>  	struct request_queue *queue;
>> -	/* Root service tree for cfq_groups */
>> -	struct cfq_rb_root grp_service_tree;
>>  	struct cfq_group root_group;
>>  
>> +	/* cfq group schedule in flat or hierarchy manner. */
>> +	bool use_hierarchy;
>> +
> 
> This seems to be redundant now? Nobody is using it?

Ahh, I think so.

> 
>>  	/*
>>  	 * The priority currently being served
>>  	 */
>> @@ -246,6 +251,9 @@ struct cfq_data {
>>  	unsigned long workload_expires;
>>  	struct cfq_group *serving_group;
>>  
>> +	/* Service tree for cfq group flat scheduling mode. */
>> +	struct cfq_rb_root grp_service_tree;
> 
> Above comment is misleading. This service tree is now used both for
> flat as well as hierarhical mode.

Will modify.

> 
> [..]
>>  static void
>>  cfq_group_service_tree_add(struct cfq_data *cfqd, struct cfq_group *cfqg)
>>  {
>> -	struct cfq_rb_root *st = &cfqd->grp_service_tree;
>>  	struct cfq_entity *cfqe = &cfqg->cfqe;
>> -	struct cfq_entity *__cfqe;
>>  	struct rb_node *n;
>> +	struct cfq_entity *entity;
>> +	struct cfq_rb_root *st;
>> +	struct cfq_group *__cfqg;
>>  
>>  	cfqg->nr_cfqq++;
>> +
>>  	if (!RB_EMPTY_NODE(&cfqe->rb_node))
>>  		return;
>>  
>>  	/*
>> -	 * Currently put the group at the end. Later implement something
>> -	 * so that groups get lesser vtime based on their weights, so that
>> -	 * if group does not loose all if it was not continously backlogged.
>> +	 * Enqueue this group and its ancestors onto their service tree.
>>  	 */
>> -	n = rb_last(&st->rb);
>> -	if (n) {
>> -		__cfqe = rb_entry_entity(n);
>> -		cfqe->vdisktime = __cfqe->vdisktime + CFQ_IDLE_DELAY;
>> -	} else
>> -		cfqe->vdisktime = st->min_vdisktime;
>> +	while (cfqe) {
>> +		if (!RB_EMPTY_NODE(&cfqe->rb_node))
>> +			return;
>>  
>> -	cfq_entity_service_tree_add(st, cfqe);
>> +		/*
>> +		 * Currently put the group at the end. Later implement
>> +		 * something so that groups get lesser vtime based on
>> +		 * their weights, so that if group does not loose all
>> +		 * if it was not continously backlogged.
>> +		 */
> 
> Can we use vdisktime boost logic for groups also? I think it can be a separate
> patch in the series (the last one). Keeping it as a separate patch will
> also help you to coordinate with chad's patch.

Make a separete patch make more sense, will do it as soon as this series gets merged.

> 
>> +		st = cfqe->service_tree;
> 
> Group entity set their service tree when they get allocated and retain
> this pointer even when they get deleted from serivce tree. Queue entities
> seem to have it NULL when they get deleted from service tree and it
> gets set again when queue is getting inserted. It would be nice if we
> can fix this discrepancy and keep it consistent. I think clearing up
> cfqe->service_tree is a better idea and then calculate it again for
> group also.

Ok, will consider to change.

> 
> [..]
>>  
>> -static struct cfq_group *
>> -cfq_find_alloc_cfqg(struct cfq_data *cfqd, struct cgroup *cgroup, int create)
>> +static void init_cfqe(struct blkio_cgroup *blkcg,
>> +				    struct cfq_group *cfqg)
> 
> As you are using this function for initializing group entity, possibly
> rename it to init_group_entity() or init_group_cfqe() etc.

Sure.

> 
> [..]
>> +static struct cfq_group *
>> +cfq_find_alloc_cfqg(struct cfq_data *cfqd, struct cgroup *cgroup, int create)
>> +{
>> +	struct blkio_cgroup *blkcg = cgroup_to_blkio_cgroup(cgroup);
>> +	struct cfq_group *cfqg = NULL;
>> +	void *key = cfqd;
>> +	struct backing_dev_info *bdi = &cfqd->queue->backing_dev_info;
>> +	unsigned int major, minor;
>> +
>> +	cfqg = cfqg_of_blkg(blkiocg_lookup_group(blkcg, key));
>> +	if (cfqg && !cfqg->blkg.dev && bdi->dev && dev_name(bdi->dev)) {
>> +		sscanf(dev_name(bdi->dev), "%u:%u", &major, &minor);
>> +		cfqg->blkg.dev = MKDEV(major, minor);
>> +		goto done;
>> +	}
> 
> Should we make this updation of this info hierarhical?

IMHO, it's fine to defer the updation when we really get the cfqg.

Will post an updated version.

Thanks,
Gui

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