patches for DPDK stable branches
 help / color / mirror / Atom feed
From: longli@linuxonhyperv.com
To: dev@dpdk.org, Stephen Hemminger <sthemmin@microsoft.com>
Cc: Long Li <longli@microsoft.com>, stable@dpdk.org
Subject: [PATCH] net/netvsc: fix issues with hot adding multiple VF PCI devices
Date: Thu, 24 Mar 2022 10:46:17 -0700
Message-ID: <1648143977-18034-1-git-send-email-longli@linuxonhyperv.com> (raw)

From: Long Li <longli@microsoft.com>

This patch fixes two issues with hot removing/adding a VF PCI device:
1. The original device argument is lost when it's hot added
2. If there are multiple VFs hot adding at the same time, some of the VFs may
not get added successfully because only one single VF status is stored in the
netvsc.

Fix these by storing the original device arguments and maintain a list of hot
add contexts to deal with multiple VF devices.

Fixes: a2a23a794b ("net/netvsc: support VF device hot add/remove")
Cc: stable@dpdk.org
Signed-off-by: Long Li <longli@microsoft.com>
---
 drivers/net/netvsc/hn_ethdev.c | 78 +++++++++++++++++++++++++++-------
 drivers/net/netvsc/hn_var.h    | 12 +++++-
 drivers/net/netvsc/hn_vf.c     |  4 ++
 3 files changed, 77 insertions(+), 17 deletions(-)

diff --git a/drivers/net/netvsc/hn_ethdev.c b/drivers/net/netvsc/hn_ethdev.c
index 8a950403ac..0a357d3645 100644
--- a/drivers/net/netvsc/hn_ethdev.c
+++ b/drivers/net/netvsc/hn_ethdev.c
@@ -554,9 +554,10 @@ static int hn_subchan_configure(struct hn_data *hv,
 static void netvsc_hotplug_retry(void *args)
 {
 	int ret;
-	struct hn_data *hv = args;
+	struct hv_hotadd_context *hot_ctx = args;
+	struct hn_data *hv = hot_ctx->hv;
 	struct rte_eth_dev *dev = &rte_eth_devices[hv->port_id];
-	struct rte_devargs *d = &hv->devargs;
+	struct rte_devargs *d = &hot_ctx->da;
 	char buf[256];
 
 	DIR *di;
@@ -566,10 +567,13 @@ static void netvsc_hotplug_retry(void *args)
 	int s;
 
 	PMD_DRV_LOG(DEBUG, "%s: retry count %d",
-		    __func__, hv->eal_hot_plug_retry);
+		    __func__, hot_ctx->eal_hot_plug_retry);
 
-	if (hv->eal_hot_plug_retry++ > NETVSC_MAX_HOTADD_RETRY)
-		return;
+	if (hot_ctx->eal_hot_plug_retry++ > NETVSC_MAX_HOTADD_RETRY) {
+		PMD_DRV_LOG(NOTICE, "Failed to parse PCI device retry=%d",
+			    hot_ctx->eal_hot_plug_retry);
+		goto free_hotadd_ctx;
+	}
 
 	snprintf(buf, sizeof(buf), "/sys/bus/pci/devices/%s/net", d->name);
 	di = opendir(buf);
@@ -602,7 +606,7 @@ static void netvsc_hotplug_retry(void *args)
 		}
 		if (req.ifr_hwaddr.sa_family != ARPHRD_ETHER) {
 			closedir(di);
-			return;
+			goto free_hotadd_ctx;
 		}
 		memcpy(eth_addr.addr_bytes, req.ifr_hwaddr.sa_data,
 		       RTE_DIM(eth_addr.addr_bytes));
@@ -611,8 +615,13 @@ static void netvsc_hotplug_retry(void *args)
 			PMD_DRV_LOG(NOTICE,
 				    "Found matching MAC address, adding device %s network name %s",
 				    d->name, dir->d_name);
+
+			/* If this device has been hot removed from this
+			 * parent device, restore its args.
+			 */
 			ret = rte_eal_hotplug_add(d->bus->name, d->name,
-						  d->args);
+						  hv->vf_devargs ?
+						  hv->vf_devargs : "");
 			if (ret) {
 				PMD_DRV_LOG(ERR,
 					    "Failed to add PCI device %s",
@@ -624,12 +633,20 @@ static void netvsc_hotplug_retry(void *args)
 		 * the device, or its MAC address did not match.
 		 */
 		closedir(di);
-		return;
+		goto free_hotadd_ctx;
 	}
 	closedir(di);
 retry:
 	/* The device is still being initialized, retry after 1 second */
-	rte_eal_alarm_set(1000000, netvsc_hotplug_retry, hv);
+	rte_eal_alarm_set(1000000, netvsc_hotplug_retry, hot_ctx);
+	return;
+
+free_hotadd_ctx:
+	rte_spinlock_lock(&hv->hotadd_lock);
+	LIST_REMOVE(hot_ctx, list);
+	rte_spinlock_unlock(&hv->hotadd_lock);
+
+	rte_free(hot_ctx);
 }
 
 static void
@@ -637,7 +654,8 @@ netvsc_hotadd_callback(const char *device_name, enum rte_dev_event_type type,
 		       void *arg)
 {
 	struct hn_data *hv = arg;
-	struct rte_devargs *d = &hv->devargs;
+	struct hv_hotadd_context *hot_ctx;
+	struct rte_devargs *d;
 	int ret;
 
 	PMD_DRV_LOG(INFO, "Device notification type=%d device_name=%s",
@@ -649,26 +667,42 @@ netvsc_hotadd_callback(const char *device_name, enum rte_dev_event_type type,
 		if (hv->vf_ctx.vf_state > vf_removed)
 			break;
 
+		hot_ctx = rte_zmalloc("NETVSC-HOTADD", sizeof(*hot_ctx),
+				      rte_mem_page_size());
+
+		if (!hot_ctx) {
+			PMD_DRV_LOG(ERR, "Failed to allocate hotadd context");
+			return;
+		}
+
+		hot_ctx->hv = hv;
+		d = &hot_ctx->da;
+
 		ret = rte_devargs_parse(d, device_name);
 		if (ret) {
 			PMD_DRV_LOG(ERR,
 				    "devargs parsing failed ret=%d", ret);
-			return;
+			goto free_ctx;
 		}
 
 		if (!strcmp(d->bus->name, "pci")) {
 			/* Start the process of figuring out if this
 			 * PCI device is a VF device
 			 */
-			hv->eal_hot_plug_retry = 0;
-			rte_eal_alarm_set(1000000, netvsc_hotplug_retry, hv);
+			rte_spinlock_lock(&hv->hotadd_lock);
+			LIST_INSERT_HEAD(&hv->hotadd_list, hot_ctx, list);
+			rte_spinlock_unlock(&hv->hotadd_lock);
+			rte_eal_alarm_set(1000000, netvsc_hotplug_retry, hot_ctx);
+			return;
 		}
 
 		/* We will switch to VF on RDNIS configure message
 		 * sent from VSP
 		 */
-
+free_ctx:
+		rte_free(hot_ctx);
 		break;
+
 	default:
 		break;
 	}
@@ -1003,12 +1037,20 @@ hn_dev_close(struct rte_eth_dev *dev)
 {
 	int ret;
 	struct hn_data *hv = dev->data->dev_private;
+	struct hv_hotadd_context *hot_ctx;
 
 	PMD_INIT_FUNC_TRACE();
 	if (rte_eal_process_type() != RTE_PROC_PRIMARY)
 		return 0;
 
-	rte_eal_alarm_cancel(netvsc_hotplug_retry, &hv->devargs);
+	rte_spinlock_lock(&hv->hotadd_lock);
+	while (!LIST_EMPTY(&hv->hotadd_list)) {
+		hot_ctx = LIST_FIRST(&hv->hotadd_list);
+		rte_eal_alarm_cancel(netvsc_hotplug_retry, hot_ctx);
+		LIST_REMOVE(hot_ctx, list);
+		rte_free(hot_ctx);
+	}
+	rte_spinlock_unlock(&hv->hotadd_lock);
 
 	ret = hn_vf_close(dev);
 	hn_dev_free_queues(dev);
@@ -1097,6 +1139,9 @@ eth_hn_dev_init(struct rte_eth_dev *eth_dev)
 
 	PMD_INIT_FUNC_TRACE();
 
+	rte_spinlock_init(&hv->hotadd_lock);
+	LIST_INIT(&hv->hotadd_list);
+
 	vmbus = container_of(device, struct rte_vmbus_device, device);
 	eth_dev->dev_ops = &hn_eth_dev_ops;
 	eth_dev->rx_queue_count = hn_dev_rx_queue_count;
@@ -1221,6 +1266,9 @@ eth_hn_dev_uninit(struct rte_eth_dev *eth_dev)
 	ret_stop = hn_dev_stop(eth_dev);
 	hn_dev_close(eth_dev);
 
+	free(hv->vf_devargs);
+	hv->vf_devargs = NULL;
+
 	hn_detach(hv);
 	hn_chim_uninit(eth_dev);
 	rte_vmbus_chan_close(hv->primary->chan);
diff --git a/drivers/net/netvsc/hn_var.h b/drivers/net/netvsc/hn_var.h
index fbb3995507..416c042a27 100644
--- a/drivers/net/netvsc/hn_var.h
+++ b/drivers/net/netvsc/hn_var.h
@@ -126,6 +126,13 @@ struct hn_vf_ctx {
 	enum vf_device_state	vf_state;
 };
 
+struct hv_hotadd_context {
+	LIST_ENTRY(hv_hotadd_context) list;
+	struct hn_data *hv;
+	struct rte_devargs da;
+	int eal_hot_plug_retry;
+};
+
 struct hn_data {
 	struct rte_vmbus_device *vmbus;
 	struct hn_rx_queue *primary;
@@ -175,8 +182,9 @@ struct hn_data {
 
 	struct vmbus_channel *channels[HN_MAX_CHANNELS];
 
-	struct rte_devargs devargs;
-	int		eal_hot_plug_retry;
+	rte_spinlock_t	hotadd_lock;
+	LIST_HEAD(hotadd_list, hv_hotadd_context) hotadd_list;
+	char		*vf_devargs;
 };
 
 static inline struct vmbus_channel *
diff --git a/drivers/net/netvsc/hn_vf.c b/drivers/net/netvsc/hn_vf.c
index fead8eba5d..72bd899683 100644
--- a/drivers/net/netvsc/hn_vf.c
+++ b/drivers/net/netvsc/hn_vf.c
@@ -129,6 +129,10 @@ static void hn_remove_delayed(void *args)
 		PMD_DRV_LOG(ERR, "rte_eth_dev_stop failed port_id=%u ret=%d",
 			    port_id, ret);
 
+	/* Record the device parameters for possible hotplug events */
+	if (dev->devargs && dev->devargs->args)
+		hv->vf_devargs = strdup(dev->devargs->args);
+
 	ret = rte_eth_dev_close(port_id);
 	if (ret)
 		PMD_DRV_LOG(ERR, "rte_eth_dev_close failed port_id=%u ret=%d",
-- 
2.32.0


             reply	other threads:[~2022-03-24 17:46 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-03-24 17:46 longli [this message]
2022-05-03 10:01 ` Ferruh Yigit

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=1648143977-18034-1-git-send-email-longli@linuxonhyperv.com \
    --to=longli@linuxonhyperv.com \
    --cc=dev@dpdk.org \
    --cc=longli@microsoft.com \
    --cc=stable@dpdk.org \
    --cc=sthemmin@microsoft.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

patches for DPDK stable branches

This inbox may be cloned and mirrored by anyone:

	git clone --mirror http://inbox.dpdk.org/stable/0 stable/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 stable stable/ http://inbox.dpdk.org/stable \
		stable@dpdk.org
	public-inbox-index stable

Example config snippet for mirrors.
Newsgroup available over NNTP:
	nntp://inbox.dpdk.org/inbox.dpdk.stable


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git