From: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
To: Jason Wang <jasowang@redhat.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>,
virtualization@lists.linux-foundation.org,
linux-kernel@vger.kernel.org
Subject: Re: [PATCH v4 14/14] virtio_pci: queue_reset: support VIRTIO_F_RING_RESET
Date: Fri, 11 Feb 2022 15:21:13 +0800 [thread overview]
Message-ID: <1644564073.3777697-1-xuanzhuo@linux.alibaba.com> (raw)
In-Reply-To: <8b0b5f06-74eb-d6cb-07e2-38249e4cda92@redhat.com>
On Fri, 11 Feb 2022 15:05:40 +0800, Jason Wang <jasowang@redhat.com> wrote:
>
> 在 2022/2/9 下午8:29, Xuan Zhuo 写道:
> > This patch implements virtio pci support for QUEUE RESET.
> >
> > Performing reset on a queue is divided into these steps:
> >
> > 1. reset_vq: reset one vq
> > 2. recycle the buffer from vq by virtqueue_detach_unused_buf()
> > 3. release the ring of the vq by vring_release_virtqueue()
> > 4. enable_reset_vq: re-enable the reset queue
> >
> > This patch implements reset_vq, enable_reset_vq in the pci scenario
> >
> > Signed-off-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
> > ---
> > drivers/virtio/virtio_pci_common.c | 8 +--
> > drivers/virtio/virtio_pci_modern.c | 80 ++++++++++++++++++++++++++++--
> > drivers/virtio/virtio_ring.c | 2 +
> > include/linux/virtio.h | 1 +
> > 4 files changed, 85 insertions(+), 6 deletions(-)
> >
> > diff --git a/drivers/virtio/virtio_pci_common.c b/drivers/virtio/virtio_pci_common.c
> > index cb01eb0cb2e4..303637ac4914 100644
> > --- a/drivers/virtio/virtio_pci_common.c
> > +++ b/drivers/virtio/virtio_pci_common.c
> > @@ -255,9 +255,11 @@ static void vp_del_vq(struct virtqueue *vq)
> > struct virtio_pci_vq_info *info = vp_dev->vqs[vq->index];
> > unsigned long flags;
> >
> > - spin_lock_irqsave(&vp_dev->lock, flags);
> > - list_del(&info->node);
> > - spin_unlock_irqrestore(&vp_dev->lock, flags);
> > + if (!vq->reset) {
> > + spin_lock_irqsave(&vp_dev->lock, flags);
> > + list_del(&info->node);
> > + spin_unlock_irqrestore(&vp_dev->lock, flags);
> > + }
> >
> > vp_dev->del_vq(info);
> > kfree(info);
> > diff --git a/drivers/virtio/virtio_pci_modern.c b/drivers/virtio/virtio_pci_modern.c
> > index d29d40bf0b45..cc45515eda50 100644
> > --- a/drivers/virtio/virtio_pci_modern.c
> > +++ b/drivers/virtio/virtio_pci_modern.c
> > @@ -34,6 +34,9 @@ static void vp_transport_features(struct virtio_device *vdev, u64 features)
> > if ((features & BIT_ULL(VIRTIO_F_SR_IOV)) &&
> > pci_find_ext_capability(pci_dev, PCI_EXT_CAP_ID_SRIOV))
> > __virtio_set_bit(vdev, VIRTIO_F_SR_IOV);
> > +
> > + if (features & BIT_ULL(VIRTIO_F_RING_RESET))
> > + __virtio_set_bit(vdev, VIRTIO_F_RING_RESET);
> > }
> >
> > /* virtio config->finalize_features() implementation */
> > @@ -176,6 +179,70 @@ static void vp_reset(struct virtio_device *vdev)
> > vp_disable_cbs(vdev);
> > }
> >
> > +static int vp_modern_reset_vq(struct virtqueue *vq)
> > +{
> > + struct virtio_pci_device *vp_dev = to_vp_device(vq->vdev);
> > + struct virtio_pci_modern_device *mdev = &vp_dev->mdev;
> > + struct virtio_pci_vq_info *info;
> > + unsigned long flags;
> > + u16 msix_vec;
> > +
> > + if (!virtio_has_feature(vq->vdev, VIRTIO_F_RING_RESET))
> > + return -ENOENT;
> > +
> > + vp_modern_set_queue_reset(mdev, vq->index);
> > +
> > + info = vp_dev->vqs[vq->index];
> > + msix_vec = info->msix_vector;
> > +
> > + /* Disable VQ callback. */
> > + if (vp_dev->per_vq_vectors && msix_vec != VIRTIO_MSI_NO_VECTOR)
> > + disable_irq(pci_irq_vector(vp_dev->pci_dev, msix_vec));
>
>
> I think we need a comment to explain why per_vq_mode needs to be dealt
> with differently.
OK.
>
>
> > +
> > + /* delete vq */
> > + spin_lock_irqsave(&vp_dev->lock, flags);
> > + list_del(&info->node);
> > + spin_unlock_irqrestore(&vp_dev->lock, flags);
>
>
> So I don't see where vring is freed and vp_setup_vq() may try to
> allocate new memory, won't it be a memory leak in this case?
1. reset_vq: reset one vq
2. recycle the buffer from vq by virtqueue_detach_unused_buf()
3. release the ring of the vq by vring_release_virtqueue()
4. enable_reset_vq: re-enable the reset queue
vring_release_virtqueue() (#8 patch) will release the vring.
That is called by the driver.
I think I should add a check to vp_modern_enable_reset_vq() that
vring_release_virtqueue() has already been called.
Thanks
>
> Thanks
>
>
> > +
> > + vq->reset = true;
> > +
> > + INIT_LIST_HEAD(&info->node);
> > +
> > + return 0;
> > +}
> > +
> > +static int vp_modern_enable_reset_vq(struct virtqueue *vq, u16 ring_num)
> > +{
> > + struct virtio_pci_device *vp_dev = to_vp_device(vq->vdev);
> > + struct virtio_pci_modern_device *mdev = &vp_dev->mdev;
> > + struct virtio_pci_vq_info *info;
> > + struct virtqueue *_vq;
> > + u16 msix_vec;
> > +
> > + if (!vq->reset)
> > + return -EPERM;
> > +
> > + /* check queue reset status */
> > + if (vp_modern_get_queue_reset(mdev, vq->index) != 1)
> > + return -EBUSY;
> > +
> > + info = vp_dev->vqs[vq->index];
> > + _vq = vp_setup_vq(vq->vdev, vq->index, NULL, NULL, NULL,
> > + info->msix_vector, ring_num);
> > + if (IS_ERR(_vq)) {
> > + vq->reset = true;
> > + return PTR_ERR(_vq);
> > + }
> > +
> > + vp_modern_set_queue_enable(&vp_dev->mdev, vq->index, true);
> > +
> > + msix_vec = vp_dev->vqs[vq->index]->msix_vector;
> > + if (vp_dev->per_vq_vectors && msix_vec != VIRTIO_MSI_NO_VECTOR)
> > + enable_irq(pci_irq_vector(vp_dev->pci_dev, msix_vec));
> > +
> > + return 0;
> > +}
> > +
> > static u16 vp_config_vector(struct virtio_pci_device *vp_dev, u16 vector)
> > {
> > return vp_modern_config_vector(&vp_dev->mdev, vector);
> > @@ -231,10 +298,13 @@ static struct virtqueue *setup_vq(struct virtio_pci_device *vp_dev,
> > virtqueue_get_avail_addr(vq),
> > virtqueue_get_used_addr(vq));
> >
> > - vq->priv = (void __force *)vp_modern_map_vq_notify(mdev, index, NULL);
> > if (!vq->priv) {
> > - err = -ENOMEM;
> > - goto err_map_notify;
> > + vq->priv = (void __force *)vp_modern_map_vq_notify(mdev, index,
> > + NULL);
> > + if (!vq->priv) {
> > + err = -ENOMEM;
> > + goto err_map_notify;
> > + }
>
>
> This seems unrelated or an artifact of previous patches?
>
> Thanks
>
>
> > }
> >
> > if (msix_vec != VIRTIO_MSI_NO_VECTOR) {
> > @@ -402,6 +472,8 @@ static const struct virtio_config_ops virtio_pci_config_nodev_ops = {
> > .set_vq_affinity = vp_set_vq_affinity,
> > .get_vq_affinity = vp_get_vq_affinity,
> > .get_shm_region = vp_get_shm_region,
> > + .reset_vq = vp_modern_reset_vq,
> > + .enable_reset_vq = vp_modern_enable_reset_vq,
> > };
> >
> > static const struct virtio_config_ops virtio_pci_config_ops = {
> > @@ -420,6 +492,8 @@ static const struct virtio_config_ops virtio_pci_config_ops = {
> > .set_vq_affinity = vp_set_vq_affinity,
> > .get_vq_affinity = vp_get_vq_affinity,
> > .get_shm_region = vp_get_shm_region,
> > + .reset_vq = vp_modern_reset_vq,
> > + .enable_reset_vq = vp_modern_enable_reset_vq,
> > };
> >
> > /* the PCI probing function */
> > diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c
> > index b8747df8dc1f..4f6028e1e2d9 100644
> > --- a/drivers/virtio/virtio_ring.c
> > +++ b/drivers/virtio/virtio_ring.c
> > @@ -1731,6 +1731,7 @@ static struct virtqueue *vring_create_virtqueue_packed(
> > vq->vq.vdev = vdev;
> > vq->vq.num_free = num;
> > vq->vq.index = index;
> > + vq->vq.reset = false;
> > vq->we_own_ring = true;
> > vq->notify = notify;
> > vq->weak_barriers = weak_barriers;
> > @@ -2220,6 +2221,7 @@ static int __vring_init_virtqueue(struct virtqueue *_vq,
> > vq->vq.vdev = vdev;
> > vq->vq.num_free = vring.num;
> > vq->vq.index = index;
> > + vq->vq.reset = false;
> > vq->we_own_ring = false;
> > vq->notify = notify;
> > vq->weak_barriers = weak_barriers;
> > diff --git a/include/linux/virtio.h b/include/linux/virtio.h
> > index dd1657c3a488..5d4817d79f3f 100644
> > --- a/include/linux/virtio.h
> > +++ b/include/linux/virtio.h
> > @@ -32,6 +32,7 @@ struct virtqueue {
> > unsigned int index;
> > unsigned int num_free;
> > void *priv;
> > + bool reset;
> > };
> >
> > int virtqueue_add_outbuf(struct virtqueue *vq,
>
next prev parent reply other threads:[~2022-02-11 7:24 UTC|newest]
Thread overview: 26+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-02-09 12:28 [PATCH v4 00/14] virtio pci support VIRTIO_F_RING_RESET Xuan Zhuo
2022-02-09 12:28 ` [PATCH v4 01/14] virtio_pci: struct virtio_pci_common_cfg add queue_notify_data Xuan Zhuo
2022-02-09 12:28 ` [PATCH v4 02/14] virtio: queue_reset: add VIRTIO_F_RING_RESET Xuan Zhuo
2022-02-09 12:28 ` [PATCH v4 03/14] virtio_ring: queue_reset: add function vring_setup_virtqueue() Xuan Zhuo
2022-02-09 12:28 ` [PATCH v4 04/14] virtio_ring: queue_reset: split: add __vring_init_virtqueue() Xuan Zhuo
2022-02-09 12:28 ` [PATCH v4 05/14] virtio_ring: queue_reset: split: support enable reset queue Xuan Zhuo
2022-02-09 12:28 ` [PATCH v4 06/14] virtio_ring: queue_reset: packed: " Xuan Zhuo
2022-02-09 12:28 ` [PATCH v4 07/14] virtio_ring: queue_reset: extract the release function of the vq ring Xuan Zhuo
2022-02-09 12:28 ` [PATCH v4 08/14] virtio_ring: queue_reset: add vring_release_virtqueue() Xuan Zhuo
2022-02-09 12:28 ` [PATCH v4 09/14] virtio: queue_reset: struct virtio_config_ops add callbacks for queue_reset Xuan Zhuo
2022-02-11 6:49 ` Jason Wang
2022-02-11 7:07 ` Xuan Zhuo
2022-02-11 7:36 ` Jason Wang
2022-02-09 12:28 ` [PATCH v4 10/14] virtio_pci: queue_reset: update struct virtio_pci_common_cfg and option functions Xuan Zhuo
2022-02-09 12:28 ` [PATCH v4 11/14] virtio_pci: queue_reset: release vq by vp_dev->vqs Xuan Zhuo
2022-02-09 12:28 ` [PATCH v4 12/14] virtio_pci: queue_reset: setup_vq() support vring_setup_virtqueue() Xuan Zhuo
2022-02-09 12:29 ` [PATCH v4 13/14] virtio_pci: queue_reset: vp_setup_vq() support ring_num Xuan Zhuo
2022-02-09 12:29 ` [PATCH v4 14/14] virtio_pci: queue_reset: support VIRTIO_F_RING_RESET Xuan Zhuo
2022-02-11 7:05 ` Jason Wang
2022-02-11 7:21 ` Xuan Zhuo [this message]
2022-02-11 7:45 ` Jason Wang
2022-02-11 8:24 ` Xuan Zhuo
2022-02-14 2:50 ` Xuan Zhuo
2022-02-11 5:40 ` [PATCH v4 00/14] virtio pci " Jason Wang
2022-02-11 6:27 ` Xuan Zhuo
2022-02-11 7:09 ` Jason Wang
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=1644564073.3777697-1-xuanzhuo@linux.alibaba.com \
--to=xuanzhuo@linux.alibaba.com \
--cc=jasowang@redhat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mst@redhat.com \
--cc=virtualization@lists.linux-foundation.org \
/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).