From: Eli Cohen <elic@nvidia.com>
To: Jason Wang <jasowang@redhat.com>, "mst@redhat.com" <mst@redhat.com>
Cc: "virtualization@lists.linux-foundation.org"
<virtualization@lists.linux-foundation.org>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Subject: RE: [PATCH 2/2] vdpa: mlx5: synchronize driver status with CVQ
Date: Mon, 21 Mar 2022 06:56:24 +0000 [thread overview]
Message-ID: <DM8PR12MB5400B6489C7FEF8B950AF1C2AB169@DM8PR12MB5400.namprd12.prod.outlook.com> (raw)
In-Reply-To: <20220321060429.10457-2-jasowang@redhat.com>
> From: Jason Wang <jasowang@redhat.com>
> Sent: Monday, March 21, 2022 8:04 AM
> To: mst@redhat.com; jasowang@redhat.com
> Cc: Eli Cohen <elic@nvidia.com>; virtualization@lists.linux-foundation.org; linux-kernel@vger.kernel.org
> Subject: [PATCH 2/2] vdpa: mlx5: synchronize driver status with CVQ
>
> Currently, CVQ doesn't have any synchronization with the driver
> status. Then CVQ emulation code run in the middle of:
>
> 1) device reset
> 2) device status changed
> 3) map updating
>
> The will lead several unexpected issue like trying to execute CVQ
> command after the driver has been teared down.
>
> Fixing this by using reslock to synchronize CVQ emulation code with
> the driver status changing:
>
> - protect the whole device reset, status changing and map updating
> with reslock
> - protect the CVQ handler with the reslock and check
> VIRTIO_CONFIG_S_DRIVER_OK in the CVQ handler
>
> This will guarantee that:
>
> 1) CVQ handler won't work if VIRTIO_CONFIG_S_DRIVER_OK is not set
> 2) CVQ handler will see a consistent state of the driver instead of
> the partial one when it is running in the middle of the
> teardown_driver() or setup_driver().
>
> Cc: 5262912ef3cfc ("vdpa/mlx5: Add support for control VQ and MAC setting")
> Signed-off-by: Jason Wang <jasowang@redhat.com>
> ---
Reviewed-by: Eli Cohen <elic@nvidia.com>
> drivers/vdpa/mlx5/net/mlx5_vnet.c | 42 +++++++++++++++++++++++--------
> 1 file changed, 31 insertions(+), 11 deletions(-)
>
> diff --git a/drivers/vdpa/mlx5/net/mlx5_vnet.c b/drivers/vdpa/mlx5/net/mlx5_vnet.c
> index d5a6fb3f9c41..524240f55c1c 100644
> --- a/drivers/vdpa/mlx5/net/mlx5_vnet.c
> +++ b/drivers/vdpa/mlx5/net/mlx5_vnet.c
> @@ -1618,11 +1618,17 @@ static void mlx5_cvq_kick_handler(struct work_struct *work)
> mvdev = wqent->mvdev;
> ndev = to_mlx5_vdpa_ndev(mvdev);
> cvq = &mvdev->cvq;
> +
> + mutex_lock(&ndev->reslock);
> +
> + if (!(mvdev->status & VIRTIO_CONFIG_S_DRIVER_OK))
> + goto done;
> +
> if (!(ndev->mvdev.actual_features & BIT_ULL(VIRTIO_NET_F_CTRL_VQ)))
> - return;
> + goto done;
>
> if (!cvq->ready)
> - return;
> + goto done;
>
> while (true) {
> err = vringh_getdesc_iotlb(&cvq->vring, &cvq->riov, &cvq->wiov, &cvq->head,
> @@ -1663,6 +1669,9 @@ static void mlx5_cvq_kick_handler(struct work_struct *work)
> break;
> }
> }
> +
> +done:
> + mutex_unlock(&ndev->reslock);
> }
>
> static void mlx5_vdpa_kick_vq(struct vdpa_device *vdev, u16 idx)
> @@ -2125,6 +2134,8 @@ static int mlx5_vdpa_change_map(struct mlx5_vdpa_dev *mvdev, struct vhost_iotlb
> struct mlx5_vdpa_net *ndev = to_mlx5_vdpa_ndev(mvdev);
> int err;
>
> + mutex_lock(&ndev->reslock);
> +
> suspend_vqs(ndev);
> err = save_channels_info(ndev);
> if (err)
> @@ -2137,18 +2148,20 @@ static int mlx5_vdpa_change_map(struct mlx5_vdpa_dev *mvdev, struct vhost_iotlb
> goto err_mr;
>
> if (!(mvdev->status & VIRTIO_CONFIG_S_DRIVER_OK))
> - return 0;
> + goto err_mr;
>
> restore_channels_info(ndev);
> err = setup_driver(mvdev);
> if (err)
> goto err_setup;
>
> + mutex_unlock(&ndev->reslock);
> return 0;
>
> err_setup:
> mlx5_vdpa_destroy_mr(mvdev);
> err_mr:
> + mutex_unlock(&ndev->reslock);
> return err;
> }
>
> @@ -2157,7 +2170,8 @@ static int setup_driver(struct mlx5_vdpa_dev *mvdev)
> struct mlx5_vdpa_net *ndev = to_mlx5_vdpa_ndev(mvdev);
> int err;
>
> - mutex_lock(&ndev->reslock);
> + WARN_ON(!mutex_is_locked(&ndev->reslock));
> +
> if (ndev->setup) {
> mlx5_vdpa_warn(mvdev, "setup driver called for already setup driver\n");
> err = 0;
> @@ -2187,7 +2201,6 @@ static int setup_driver(struct mlx5_vdpa_dev *mvdev)
> goto err_fwd;
> }
> ndev->setup = true;
> - mutex_unlock(&ndev->reslock);
>
> return 0;
>
> @@ -2198,23 +2211,22 @@ static int setup_driver(struct mlx5_vdpa_dev *mvdev)
> err_rqt:
> teardown_virtqueues(ndev);
> out:
> - mutex_unlock(&ndev->reslock);
> return err;
> }
>
> static void teardown_driver(struct mlx5_vdpa_net *ndev)
> {
> - mutex_lock(&ndev->reslock);
> +
> + WARN_ON(!mutex_is_locked(&ndev->reslock));
> +
> if (!ndev->setup)
> - goto out;
> + return;
>
> remove_fwd_to_tir(ndev);
> destroy_tir(ndev);
> destroy_rqt(ndev);
> teardown_virtqueues(ndev);
> ndev->setup = false;
> -out:
> - mutex_unlock(&ndev->reslock);
> }
>
> static void clear_vqs_ready(struct mlx5_vdpa_net *ndev)
> @@ -2235,6 +2247,8 @@ static void mlx5_vdpa_set_status(struct vdpa_device *vdev, u8 status)
>
> print_status(mvdev, status, true);
>
> + mutex_lock(&ndev->reslock);
> +
> if ((status ^ ndev->mvdev.status) & VIRTIO_CONFIG_S_DRIVER_OK) {
> if (status & VIRTIO_CONFIG_S_DRIVER_OK) {
> err = setup_driver(mvdev);
> @@ -2244,16 +2258,19 @@ static void mlx5_vdpa_set_status(struct vdpa_device *vdev, u8 status)
> }
> } else {
> mlx5_vdpa_warn(mvdev, "did not expect DRIVER_OK to be cleared\n");
> - return;
> + goto err_clear;
> }
> }
>
> ndev->mvdev.status = status;
> + mutex_unlock(&ndev->reslock);
> return;
>
> err_setup:
> mlx5_vdpa_destroy_mr(&ndev->mvdev);
> ndev->mvdev.status |= VIRTIO_CONFIG_S_FAILED;
> +err_clear:
> + mutex_unlock(&ndev->reslock);
> }
>
> static int mlx5_vdpa_reset(struct vdpa_device *vdev)
> @@ -2263,6 +2280,8 @@ static int mlx5_vdpa_reset(struct vdpa_device *vdev)
>
> print_status(mvdev, 0, true);
> mlx5_vdpa_info(mvdev, "performing device reset\n");
> +
> + mutex_lock(&ndev->reslock);
> teardown_driver(ndev);
> clear_vqs_ready(ndev);
> mlx5_vdpa_destroy_mr(&ndev->mvdev);
> @@ -2275,6 +2294,7 @@ static int mlx5_vdpa_reset(struct vdpa_device *vdev)
> if (mlx5_vdpa_create_mr(mvdev, NULL))
> mlx5_vdpa_warn(mvdev, "create MR failed\n");
> }
> + mutex_unlock(&ndev->reslock);
>
> return 0;
> }
> --
> 2.18.1
next prev parent reply other threads:[~2022-03-21 6:56 UTC|newest]
Thread overview: 18+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-03-21 6:04 [PATCH 1/2] vdpa: mlx5: prevent cvq work from hogging CPU Jason Wang
2022-03-21 6:04 ` [PATCH 2/2] vdpa: mlx5: synchronize driver status with CVQ Jason Wang
2022-03-21 6:56 ` Eli Cohen [this message]
2022-03-21 7:23 ` Michael S. Tsirkin
2022-03-21 7:36 ` Jason Wang
2022-03-21 6:31 ` [PATCH 1/2] vdpa: mlx5: prevent cvq work from hogging CPU Eli Cohen
2022-03-21 7:20 ` Michael S. Tsirkin
2022-03-21 7:35 ` Jason Wang
[not found] ` <20220321085317.3148-1-hdanton@sina.com>
2022-03-21 8:59 ` Jason Wang
2022-03-21 9:00 ` Jason Wang
[not found] ` <20220321123420.3207-1-hdanton@sina.com>
2022-03-22 1:59 ` Jason Wang
[not found] ` <20220324005345.3623-1-hdanton@sina.com>
2022-03-24 2:34 ` Jason Wang
[not found] ` <20220324060419.3682-1-hdanton@sina.com>
2022-03-24 6:17 ` Michael S. Tsirkin
2022-03-24 8:20 ` Jason Wang
[not found] ` <20220324120217.3746-1-hdanton@sina.com>
2022-03-24 12:24 ` Eli Cohen
2022-03-25 3:22 ` Jason Wang
2022-03-25 6:45 ` Michael S. Tsirkin
2022-03-25 7:53 ` 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=DM8PR12MB5400B6489C7FEF8B950AF1C2AB169@DM8PR12MB5400.namprd12.prod.outlook.com \
--to=elic@nvidia.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).