linux-kernel - Re: [PATCH] virtio_blk: set the default scheduler to none

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: <20231207013617-mutt-send-email-mst@kernel.org>
Date:   Thu, 7 Dec 2023 01:38:37 -0500
From:   "Michael S. Tsirkin" <mst@...hat.com>
To:     Jason Wang <jasowang@...hat.com>
Cc:     Li Feng <fengli@...rtx.com>, Jens Axboe <axboe@...nel.dk>,
        Paolo Bonzini <pbonzini@...hat.com>,
        Stefan Hajnoczi <stefanha@...hat.com>,
        Xuan Zhuo <xuanzhuo@...ux.alibaba.com>,
        "open list:BLOCK LAYER" <linux-block@...r.kernel.org>,
        open list <linux-kernel@...r.kernel.org>,
        "open list:VIRTIO BLOCK AND SCSI DRIVERS" 
        <virtualization@...ts.linux.dev>
Subject: Re: [PATCH] virtio_blk: set the default scheduler to none

On Thu, Dec 07, 2023 at 02:02:36PM +0800, Jason Wang wrote:
> On Thu, Dec 7, 2023 at 12:33 PM Li Feng <fengli@...rtx.com> wrote:
> >
> > virtio-blk is generally used in cloud computing scenarios, where the
> > performance of virtual disks is very important. The mq-deadline scheduler
> > has a big performance drop compared to none with single queue.
> 
> At least you can choose the scheduler based on if mq is supported or not?
> 
> Thanks

This is already the case:

static struct elevator_type *elevator_get_default(struct request_queue *q)
{
        if (q->tag_set && q->tag_set->flags & BLK_MQ_F_NO_SCHED_BY_DEFAULT)
                return NULL;

        if (q->nr_hw_queues != 1 &&
            !blk_mq_is_shared_tags(q->tag_set->flags))
                return NULL;

        return elevator_find_get(q, "mq-deadline");
}

I guess I agree blk is typically kind of similar to loopback
so none by default makes sense here same as for loopback.

Stefan care to comment?

-- 
MST