From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:45905) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1XCUSP-0007YM-Rp for qemu-devel@nongnu.org; Wed, 30 Jul 2014 10:02:23 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1XCUSK-0005m2-3t for qemu-devel@nongnu.org; Wed, 30 Jul 2014 10:02:17 -0400 Received: from mx1.redhat.com ([209.132.183.28]:25645) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1XCUSJ-0005kS-Qj for qemu-devel@nongnu.org; Wed, 30 Jul 2014 10:02:12 -0400 Message-ID: <53D8FAD7.1070509@redhat.com> Date: Wed, 30 Jul 2014 16:01:59 +0200 From: Paolo Bonzini MIME-Version: 1.0 References: <1406720388-18671-1-git-send-email-ming.lei@canonical.com> <1406720388-18671-15-git-send-email-ming.lei@canonical.com> In-Reply-To: <1406720388-18671-15-git-send-email-ming.lei@canonical.com> Content-Type: text/plain; charset=ISO-8859-15 Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [PATCH 14/15] hw/block/virtio-blk: create num_queues vqs if dataplane is enabled List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Ming Lei , qemu-devel@nongnu.org, Peter Maydell , Stefan Hajnoczi Cc: Kevin Wolf , Fam Zheng , "Michael S. Tsirkin" Il 30/07/2014 13:39, Ming Lei ha scritto: > Now we only support multi vqs for dataplane case. > > Signed-off-by: Ming Lei > --- > hw/block/virtio-blk.c | 18 +++++++++++++++++- > include/hw/virtio/virtio-blk.h | 1 + > 2 files changed, 18 insertions(+), 1 deletion(-) > > diff --git a/hw/block/virtio-blk.c b/hw/block/virtio-blk.c > index ab99156..44e9956 100644 > --- a/hw/block/virtio-blk.c > +++ b/hw/block/virtio-blk.c > @@ -556,6 +556,7 @@ static void virtio_blk_update_config(VirtIODevice *vdev, uint8_t *config) > blkcfg.physical_block_exp = get_physical_block_exp(s->conf); > blkcfg.alignment_offset = 0; > blkcfg.wce = bdrv_enable_write_cache(s->bs); > + stw_p(&blkcfg.num_queues, s->blk.num_queues); > memcpy(config, &blkcfg, sizeof(struct virtio_blk_config)); > } > > @@ -590,6 +591,12 @@ static uint32_t virtio_blk_get_features(VirtIODevice *vdev, uint32_t features) > if (bdrv_is_read_only(s->bs)) > features |= 1 << VIRTIO_BLK_F_RO; > > +#ifdef CONFIG_VIRTIO_BLK_DATA_PLANE > + if (s->blk.num_queues > 1) { > + features |= 1 << VIRTIO_BLK_F_MQ; > + } > +#endif > + > return features; > } > > @@ -739,8 +746,13 @@ static void virtio_blk_device_realize(DeviceState *dev, Error **errp) > #ifdef CONFIG_VIRTIO_BLK_DATA_PLANE > Error *err = NULL; > #endif > + int i; > static int virtio_blk_id; > > +#ifndef CONFIG_VIRTIO_BLK_DATA_PLANE > + blk->num_queues = 1; > +#endif > + > if (!blk->conf.bs) { > error_setg(errp, "drive property not set"); > return; > @@ -765,7 +777,10 @@ static void virtio_blk_device_realize(DeviceState *dev, Error **errp) > s->rq = NULL; > s->sector_mask = (s->conf->logical_block_size / BDRV_SECTOR_SIZE) - 1; > > - s->vq = virtio_add_queue(vdev, 128, virtio_blk_handle_output); > + s->vqs = g_malloc0(sizeof(VirtQueue *) * blk->num_queues); > + for (i = 0; i < blk->num_queues; i++) > + s->vqs[i] = virtio_add_queue(vdev, 128, virtio_blk_handle_output); > + s->vq = s->vqs[0]; > s->complete_request = virtio_blk_complete_request; > #ifdef CONFIG_VIRTIO_BLK_DATA_PLANE > virtio_blk_data_plane_create(vdev, blk, &s->dataplane, &err); > @@ -802,6 +817,7 @@ static void virtio_blk_device_unrealize(DeviceState *dev, Error **errp) > qemu_del_vm_change_state_handler(s->change); > unregister_savevm(dev, "virtio-blk", s); > blockdev_mark_auto_del(s->bs); > + g_free(s->vqs); > virtio_cleanup(vdev); > } > > diff --git a/include/hw/virtio/virtio-blk.h b/include/hw/virtio/virtio-blk.h > index ad70c9a..b4609dd 100644 > --- a/include/hw/virtio/virtio-blk.h > +++ b/include/hw/virtio/virtio-blk.h > @@ -132,6 +132,7 @@ typedef struct VirtIOBlock { > VirtIODevice parent_obj; > BlockDriverState *bs; > VirtQueue *vq; > + VirtQueue **vqs; > void *rq; > QEMUBH *bh; > BlockConf *conf; > Dataplane must not be a change to the guest ABI. If you implement this feature you have to implement it for both dataplane and non-dataplne. Paolo