DPDK patches and discussions
 help / color / mirror / Atom feed
From: Xueming Li <xuemingl@nvidia.com>
To: <dev@dpdk.org>
Cc: <xuemingl@nvidia.com>, Matan Azrad <matan@nvidia.com>,
	"Viacheslav Ovsiienko" <viacheslavo@nvidia.com>
Subject: [PATCH 6/7] vdpa/mlx5: support device cleanup callback
Date: Thu, 24 Feb 2022 21:28:19 +0800	[thread overview]
Message-ID: <20220224132820.1939650-7-xuemingl@nvidia.com> (raw)
In-Reply-To: <20220224132820.1939650-1-xuemingl@nvidia.com>

This patch supports device cleanup callback API which called when device
disconected with VM. Cached resources like VM MR and VQ memory are
released.

Signed-off-by: Xueming Li <xuemingl@nvidia.com>
---
 drivers/vdpa/mlx5/mlx5_vdpa.c | 23 +++++++++++++++++++++++
 drivers/vdpa/mlx5/mlx5_vdpa.h |  1 +
 2 files changed, 24 insertions(+)

diff --git a/drivers/vdpa/mlx5/mlx5_vdpa.c b/drivers/vdpa/mlx5/mlx5_vdpa.c
index dbaa590d5d1..c83b1141482 100644
--- a/drivers/vdpa/mlx5/mlx5_vdpa.c
+++ b/drivers/vdpa/mlx5/mlx5_vdpa.c
@@ -270,6 +270,8 @@ mlx5_vdpa_dev_close(int vid)
 	if (priv->lm_mr.addr)
 		mlx5_os_wrapped_mkey_destroy(&priv->lm_mr);
 	priv->state = MLX5_VDPA_STATE_PROBED;
+	if (!priv->connected)
+		mlx5_vdpa_dev_cache_clean(priv);
 	priv->vid = 0;
 	/* The mutex may stay locked after event thread cancel - initiate it. */
 	pthread_mutex_init(&priv->vq_config_lock, NULL);
@@ -294,6 +296,7 @@ mlx5_vdpa_dev_config(int vid)
 		return -1;
 	}
 	priv->vid = vid;
+	priv->connected = true;
 	if (mlx5_vdpa_mtu_set(priv))
 		DRV_LOG(WARNING, "MTU cannot be set on device %s.",
 				vdev->device->name);
@@ -431,12 +434,32 @@ mlx5_vdpa_reset_stats(struct rte_vdpa_device *vdev, int qid)
 	return mlx5_vdpa_virtq_stats_reset(priv, qid);
 }
 
+static int
+mlx5_vdpa_dev_clean(int vid)
+{
+	struct rte_vdpa_device *vdev = rte_vhost_get_vdpa_device(vid);
+	struct mlx5_vdpa_priv *priv;
+
+	if (vdev == NULL)
+		return -1;
+	priv = mlx5_vdpa_find_priv_resource_by_vdev(vdev);
+	if (priv == NULL) {
+		DRV_LOG(ERR, "Invalid vDPA device: %s.", vdev->device->name);
+		return -1;
+	}
+	if (priv->state == MLX5_VDPA_STATE_PROBED)
+		mlx5_vdpa_dev_cache_clean(priv);
+	priv->connected = false;
+	return 0;
+}
+
 static struct rte_vdpa_dev_ops mlx5_vdpa_ops = {
 	.get_queue_num = mlx5_vdpa_get_queue_num,
 	.get_features = mlx5_vdpa_get_vdpa_features,
 	.get_protocol_features = mlx5_vdpa_get_protocol_features,
 	.dev_conf = mlx5_vdpa_dev_config,
 	.dev_close = mlx5_vdpa_dev_close,
+	.dev_cleanup = mlx5_vdpa_dev_clean,
 	.set_vring_state = mlx5_vdpa_set_vring_state,
 	.set_features = mlx5_vdpa_features_set,
 	.migration_done = NULL,
diff --git a/drivers/vdpa/mlx5/mlx5_vdpa.h b/drivers/vdpa/mlx5/mlx5_vdpa.h
index 540bf87a352..24bafe85b44 100644
--- a/drivers/vdpa/mlx5/mlx5_vdpa.h
+++ b/drivers/vdpa/mlx5/mlx5_vdpa.h
@@ -121,6 +121,7 @@ enum mlx5_dev_state {
 
 struct mlx5_vdpa_priv {
 	TAILQ_ENTRY(mlx5_vdpa_priv) next;
+	bool connected;
 	enum mlx5_dev_state state;
 	pthread_mutex_t vq_config_lock;
 	uint64_t no_traffic_counter;
-- 
2.35.1


  parent reply	other threads:[~2022-02-24 13:29 UTC|newest]

Thread overview: 43+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-02-24 13:28 [PATCH 0/7] vdpa/mlx5: improve device shutdown time Xueming Li
2022-02-24 13:28 ` [PATCH 1/7] vdpa/mlx5: fix interrupt trash that leads to segment fault Xueming Li
2022-02-24 13:28 ` [PATCH 2/7] vdpa/mlx5: fix dead loop when process interrupted Xueming Li
2022-02-24 13:28 ` [PATCH 3/7] vdpa/mlx5: no kick handling during shutdown Xueming Li
2022-02-24 13:28 ` [PATCH 4/7] vdpa/mlx5: reuse resources in reconfiguration Xueming Li
2022-02-24 13:28 ` [PATCH 5/7] vdpa/mlx5: cache and reuse hardware resources Xueming Li
2022-02-24 13:28 ` Xueming Li [this message]
2022-02-24 13:28 ` [PATCH 7/7] vdpa/mlx5: make statistics counter persistent Xueming Li
2022-02-24 14:38 ` [PATCH v1 0/7] vdpa/mlx5: improve device shutdown time Xueming Li
2022-02-24 14:38   ` [PATCH v1 1/7] vdpa/mlx5: fix interrupt trash that leads to segment fault Xueming Li
2022-02-24 14:38   ` [PATCH v1 2/7] vdpa/mlx5: fix dead loop when process interrupted Xueming Li
2022-02-24 14:38   ` [PATCH v1 3/7] vdpa/mlx5: no kick handling during shutdown Xueming Li
2022-02-24 14:38   ` [PATCH v1 4/7] vdpa/mlx5: reuse resources in reconfiguration Xueming Li
2022-02-24 14:38   ` [PATCH v1 5/7] vdpa/mlx5: cache and reuse hardware resources Xueming Li
2022-02-24 14:38   ` [PATCH v1 6/7] vdpa/mlx5: support device cleanup callback Xueming Li
2022-02-24 14:38   ` [PATCH v1 7/7] vdpa/mlx5: make statistics counter persistent Xueming Li
2022-02-24 15:50 ` [PATCH v2 0/7] vdpa/mlx5: improve device shutdown time Xueming Li
2022-02-24 15:50   ` [PATCH v2 1/7] vdpa/mlx5: fix interrupt trash that leads to segment fault Xueming Li
2022-04-20 10:39     ` Maxime Coquelin
2022-02-24 15:50   ` [PATCH v2 2/7] vdpa/mlx5: fix dead loop when process interrupted Xueming Li
2022-04-20 10:33     ` Maxime Coquelin
2022-02-24 15:50   ` [PATCH v2 3/7] vdpa/mlx5: no kick handling during shutdown Xueming Li
2022-04-20 12:37     ` Maxime Coquelin
2022-04-20 13:23       ` Xueming(Steven) Li
2022-02-24 15:50   ` [PATCH v2 4/7] vdpa/mlx5: reuse resources in reconfiguration Xueming Li
2022-04-20 14:49     ` Maxime Coquelin
2022-02-24 15:50   ` [PATCH v2 5/7] vdpa/mlx5: cache and reuse hardware resources Xueming Li
2022-04-20 15:03     ` Maxime Coquelin
2022-04-25 13:28       ` Xueming(Steven) Li
2022-05-05 20:01         ` Maxime Coquelin
2022-02-24 15:51   ` [PATCH v2 6/7] vdpa/mlx5: support device cleanup callback Xueming Li
2022-04-21  8:19     ` Maxime Coquelin
2022-02-24 15:51   ` [PATCH v2 7/7] vdpa/mlx5: make statistics counter persistent Xueming Li
2022-04-21  8:22     ` Maxime Coquelin
2022-05-08 14:25 ` [PATCH v3 0/7] vdpa/mlx5: improve device shutdown time Xueming Li
2022-05-08 14:25   ` [PATCH v3 1/7] vdpa/mlx5: fix interrupt trash that leads to segment fault Xueming Li
2022-05-08 14:25   ` [PATCH v3 2/7] vdpa/mlx5: fix dead loop when process interrupted Xueming Li
2022-05-08 14:25   ` [PATCH v3 3/7] vdpa/mlx5: no kick handling during shutdown Xueming Li
2022-05-08 14:25   ` [PATCH v3 4/7] vdpa/mlx5: reuse resources in reconfiguration Xueming Li
2022-05-08 14:25   ` [PATCH v3 5/7] vdpa/mlx5: cache and reuse hardware resources Xueming Li
2022-05-08 14:25   ` [PATCH v3 6/7] vdpa/mlx5: support device cleanup callback Xueming Li
2022-05-08 14:25   ` [PATCH v3 7/7] vdpa/mlx5: make statistics counter persistent Xueming Li
2022-05-09 19:38   ` [PATCH v3 0/7] vdpa/mlx5: improve device shutdown time Maxime Coquelin

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=20220224132820.1939650-7-xuemingl@nvidia.com \
    --to=xuemingl@nvidia.com \
    --cc=dev@dpdk.org \
    --cc=matan@nvidia.com \
    --cc=viacheslavo@nvidia.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).