qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Fam Zheng <famz@redhat.com>
To: Stefan Hajnoczi <stefanha@redhat.com>
Cc: qemu-devel@nongnu.org, Paolo Bonzini <pbonzini@redhat.com>,
	Ming Lei <ming.lei@canonical.com>, Kevin Wolf <kwolf@redhat.com>
Subject: Re: [Qemu-devel] [PATCH 5/9] virtio-blk: multiqueue batch notify
Date: Mon, 23 May 2016 10:43:23 +0800	[thread overview]
Message-ID: <20160523024323.GA9980@ad.usersys.redhat.com> (raw)
In-Reply-To: <1463787632-7241-6-git-send-email-stefanha@redhat.com>

On Fri, 05/20 16:40, Stefan Hajnoczi wrote:
> The batch notification BH needs to know which virtqueues to notify when
> multiqueue is enabled.  Use a bitmap to track the virtqueues that with
> pending notifications.

This approach works great as long as VQs are in the same iothread. An
alternative way would be using separate BH per VQ, which will naturely work with
multi queue block layer in the future.  Should we just go for that in the first
place?  Seems less code churn, and no imaginable disadvantage compared to this
patch.

Fam

> 
> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
> ---
>  hw/block/virtio-blk.c          | 20 ++++++++++++++++----
>  include/hw/virtio/virtio-blk.h |  1 +
>  2 files changed, 17 insertions(+), 4 deletions(-)
> 
> diff --git a/hw/block/virtio-blk.c b/hw/block/virtio-blk.c
> index d16a1b7..ab0f589 100644
> --- a/hw/block/virtio-blk.c
> +++ b/hw/block/virtio-blk.c
> @@ -54,11 +54,19 @@ static void virtio_blk_batch_notify_bh(void *opaque)
>  {
>      VirtIOBlock *s = opaque;
>      VirtIODevice *vdev = VIRTIO_DEVICE(s);
> +    unsigned nvqs = s->conf.num_queues;
> +    unsigned i = 0;
>  
> -    if (s->dataplane_started && !s->dataplane_disabled) {
> -        virtio_blk_data_plane_notify(s->dataplane, s->vq);
> -    } else {
> -        virtio_notify(vdev, s->vq);
> +    while ((i = find_next_bit(s->batch_notify_vqs, nvqs, i)) < nvqs) {
> +        VirtQueue *vq = virtio_get_queue(vdev, i);
> +
> +        bitmap_clear(s->batch_notify_vqs, i, 1);
> +
> +        if (s->dataplane_started && !s->dataplane_disabled) {
> +            virtio_blk_data_plane_notify(s->dataplane, vq);
> +        } else {
> +            virtio_notify(vdev, vq);
> +        }
>      }
>  }
>  
> @@ -77,6 +85,8 @@ static void virtio_blk_req_complete(VirtIOBlockReq *req, unsigned char status)
>  
>      stb_p(&req->in->status, status);
>      virtqueue_push(req->vq, &req->elem, req->in_len);
> +
> +    bitmap_set(s->batch_notify_vqs, virtio_queue_get_id(req->vq), 1);
>      qemu_bh_schedule(s->batch_notify_bh);
>  }
>  
> @@ -922,6 +932,7 @@ static void virtio_blk_device_realize(DeviceState *dev, Error **errp)
>  
>      s->batch_notify_bh = aio_bh_new(blk_get_aio_context(s->blk),
>                                      virtio_blk_batch_notify_bh, s);
> +    s->batch_notify_vqs = bitmap_new(conf->num_queues);
>  
>      s->change = qemu_add_vm_change_state_handler(virtio_blk_dma_restart_cb, s);
>      register_savevm(dev, "virtio-blk", virtio_blk_id++, 2,
> @@ -938,6 +949,7 @@ static void virtio_blk_device_unrealize(DeviceState *dev, Error **errp)
>      VirtIOBlock *s = VIRTIO_BLK(dev);
>  
>      qemu_bh_delete(s->batch_notify_bh);
> +    g_free(s->batch_notify_vqs);
>      virtio_blk_data_plane_destroy(s->dataplane);
>      s->dataplane = NULL;
>      qemu_del_vm_change_state_handler(s->change);
> diff --git a/include/hw/virtio/virtio-blk.h b/include/hw/virtio/virtio-blk.h
> index 487b28d..b6e7860 100644
> --- a/include/hw/virtio/virtio-blk.h
> +++ b/include/hw/virtio/virtio-blk.h
> @@ -51,6 +51,7 @@ typedef struct VirtIOBlock {
>      void *rq;
>      QEMUBH *bh;
>      QEMUBH *batch_notify_bh;
> +    unsigned long *batch_notify_vqs;
>      VirtIOBlkConf conf;
>      unsigned short sector_mask;
>      bool original_wce;
> -- 
> 2.5.5
> 

  parent reply	other threads:[~2016-05-23  2:43 UTC|newest]

Thread overview: 25+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-05-20 23:40 [Qemu-devel] [PATCH 0/9] virtio-blk: multiqueue support Stefan Hajnoczi
2016-05-20 23:40 ` [Qemu-devel] [PATCH 1/9] virtio-blk: use batch notify in non-dataplane case Stefan Hajnoczi
2016-05-20 23:40 ` [Qemu-devel] [PATCH 2/9] virtio-blk: tell dataplane which vq to notify Stefan Hajnoczi
2016-05-20 23:40 ` [Qemu-devel] [PATCH 3/9] virtio-blk: associate request with a virtqueue Stefan Hajnoczi
2016-05-20 23:40 ` [Qemu-devel] [PATCH 4/9] virtio-blk: add VirtIOBlockConf->num_queues Stefan Hajnoczi
2016-05-20 23:40 ` [Qemu-devel] [PATCH 5/9] virtio-blk: multiqueue batch notify Stefan Hajnoczi
2016-05-21 16:02   ` Paolo Bonzini
2016-05-27 21:38     ` Stefan Hajnoczi
2016-05-23  2:43   ` Fam Zheng [this message]
2016-05-23  8:17     ` Paolo Bonzini
2016-05-23  8:56       ` Fam Zheng
2016-05-20 23:40 ` [Qemu-devel] [PATCH 6/9] vmstate: add VMSTATE_VARRAY_UINT32_ALLOC Stefan Hajnoczi
2016-05-20 23:40 ` [Qemu-devel] [PATCH 7/9] virtio-blk: live migrate s->rq with multiqueue Stefan Hajnoczi
2016-05-21 15:37   ` Paolo Bonzini
2016-05-27 21:42     ` Stefan Hajnoczi
2016-05-20 23:40 ` [Qemu-devel] [PATCH 8/9] virtio-blk: dataplane multiqueue support Stefan Hajnoczi
2016-05-20 23:40 ` [Qemu-devel] [PATCH 9/9] virtio-blk: add num-queues device property Stefan Hajnoczi
2016-05-24 12:51 ` [Qemu-devel] [PATCH 0/9] virtio-blk: multiqueue support Christian Borntraeger
2016-05-27 21:44   ` Stefan Hajnoczi
2016-05-31  0:44   ` Stefan Hajnoczi
2016-05-27 11:55 ` [Qemu-devel] [RFC] virtio-blk: simple multithreaded MQ implementation for bdrv_raw Roman Pen
2016-05-27 22:27   ` Stefan Hajnoczi
2016-05-30  6:40     ` Alexandre DERUMIER
2016-05-30 12:14       ` Roman Penyaev
2016-05-30 11:59     ` Roman Penyaev

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20160523024323.GA9980@ad.usersys.redhat.com \
    --to=famz@redhat.com \
    --cc=kwolf@redhat.com \
    --cc=ming.lei@canonical.com \
    --cc=pbonzini@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=stefanha@redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).