linux-kernel - Re: [PATCH 5/7] blk-mq-sched: add framework for MQ capable IO schedulers

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]

Date:   Tue, 13 Dec 2016 15:29:18 +0100
From:   Bart Van Assche <bart.vanassche@...disk.com>
To:     Jens Axboe <axboe@...com>, <axboe@...nel.dk>,
        <linux-block@...r.kernel.org>, <linux-kernel@...r.kernel.org>
CC:     <paolo.valente@...aro.org>, <osandov@...com>
Subject: Re: [PATCH 5/7] blk-mq-sched: add framework for MQ capable IO
 schedulers

On 12/08/2016 09:13 PM, Jens Axboe wrote:
> +static inline void blk_mq_sched_put_request(struct request *rq)
> +{
> +	struct request_queue *q = rq->q;
> +	struct elevator_queue *e = q->elevator;
> +
> +	if (e && e->type->mq_ops.put_request)
> +		e->type->mq_ops.put_request(rq);
> +	else
> +		blk_mq_free_request(rq);
> +}

blk_mq_free_request() always triggers a call of blk_queue_exit(). 
dd_put_request() only triggers a call of blk_queue_exit() if it is not a 
shadow request. Is that on purpose?

> +static inline struct request *
> +blk_mq_sched_get_request(struct request_queue *q, unsigned int op,
> +			 struct blk_mq_alloc_data *data)
> +{
> +	struct elevator_queue *e = q->elevator;
> +	struct blk_mq_hw_ctx *hctx;
> +	struct blk_mq_ctx *ctx;
> +	struct request *rq;
> +
> +	blk_queue_enter_live(q);
> +	ctx = blk_mq_get_ctx(q);
> +	hctx = blk_mq_map_queue(q, ctx->cpu);
> +
> +	blk_mq_set_alloc_data(data, q, 0, ctx, hctx);
> +
> +	if (e && e->type->mq_ops.get_request)
> +		rq = e->type->mq_ops.get_request(q, op, data);
> +	else
> +		rq = __blk_mq_alloc_request(data, op);
> +
> +	if (rq)
> +		data->hctx->queued++;
> +
> +	return rq;
> +
> +}

Some but not all callers of blk_mq_sched_get_request() call 
blk_queue_exit() if this function returns NULL. Please consider to move 
the blk_queue_exit() call from the blk_mq_alloc_request() error path 
into this function. I think that will make it a lot easier to verify 
whether or not the blk_queue_enter() / blk_queue_exit() calls are 
balanced properly.

Additionally, since blk_queue_enter() / blk_queue_exit() calls by 
blk_mq_sched_get_request() and blk_mq_sched_put_request() must be 
balanced and since the latter function only calls blk_queue_exit() for 
non-shadow requests, shouldn't blk_mq_sched_get_request() call 
blk_queue_enter_live() only if __blk_mq_alloc_request() is called?

Thanks,

Bart.