net/netvsc: add support for mtu_set

Message ID MN0PR21MB3120FDA48E7C122E74FF58F0B5CAA@MN0PR21MB3120.namprd21.prod.outlook.com (mailing list archive)
State Superseded, archived
Delegated to: Ferruh Yigit
Headers
Series net/netvsc: add support for mtu_set |

Checks

Context Check Description
ci/checkpatch warning coding style issues
ci/loongarch-compilation success Compilation OK
ci/loongarch-unit-testing success Unit Testing PASS
ci/github-robot: build success github build: passed
ci/iol-mellanox-Performance success Performance Testing PASS
ci/iol-broadcom-Performance success Performance Testing PASS
ci/iol-intel-Performance success Performance Testing PASS
ci/iol-broadcom-Functional success Functional Testing PASS
ci/iol-compile-amd64-testing success Testing PASS
ci/iol-intel-Functional success Functional Testing PASS
ci/iol-sample-apps-testing warning Testing issues
ci/iol-unit-amd64-testing success Testing PASS
ci/iol-unit-arm64-testing success Testing PASS
ci/iol-compile-arm64-testing success Testing PASS

Commit Message

Sam Andrew Oct. 5, 2023, 11:17 p.m. UTC
  Add support for changing the netvsc MTU. The MTU can only be set at nvs
initialization, therefore to change the MTU the underlying vmbus
channel(s) are torn down and the vmbus device unmapped and remapped. The
existing rx and tx queue(s) are reconnected to the new vmbus channel(s).

Signed-off-by: Sam Andrew <samandrew@microsoft.com>
---
 drivers/bus/vmbus/vmbus_common.c |   4 +-
 drivers/net/netvsc/hn_ethdev.c   | 185 +++++++++++++++++++++++++------
 drivers/net/netvsc/hn_rndis.c    |   7 ++
 drivers/net/netvsc/hn_rndis.h    |   1 +
 drivers/net/netvsc/hn_var.h      |   3 +-
 drivers/net/netvsc/hn_vf.c       |  23 +++-
 6 files changed, 188 insertions(+), 35 deletions(-)
  

Comments

Stephen Hemminger Oct. 9, 2023, 3:17 p.m. UTC | #1
On Thu, 5 Oct 2023 23:17:28 +0000
Sam Andrew <samandrew@microsoft.com> wrote:

> +
> +static int
> +hn_dev_mtu_set(struct rte_eth_dev *dev, uint16_t mtu)
> +{
> +	struct hn_data *hv = dev->data->dev_private;
> +	unsigned int orig_mtu = dev->data->mtu;
> +	uint32_t rndis_mtu;
> +	int ret = 0;
> +	int i;
> +
> +	if (dev->data->dev_started) {
> +		PMD_DRV_LOG(ERR, "Device must be stopped before changing MTU");
> +		return -EIO;
> +	}
> +

It looks like this proposed patch does not have the accelerated networking
virtual device case.  The driver needs to first update the MTU of the underlying
VF device (and handle errors), then reinit the vmbus device.
  
Sam Andrew Oct. 9, 2023, 4:41 p.m. UTC | #2
Hi Stephen,

From: Stephen Hemminger <stephen@networkplumber.org> 
Sent: Monday, October 9, 2023 8:18 AM

> On Thu, 5 Oct 2023 23:17:28 +0000
> Sam Andrew <samandrew@microsoft.com> wrote:

> > +
> > +static int
> > +hn_dev_mtu_set(struct rte_eth_dev *dev, uint16_t mtu) {
> > +	struct hn_data *hv = dev->data->dev_private;
> > +	unsigned int orig_mtu = dev->data->mtu;
> > +	uint32_t rndis_mtu;
> > +	int ret = 0;
> > +	int i;
> > +
> > +	if (dev->data->dev_started) {
> > +		PMD_DRV_LOG(ERR, "Device must be stopped before changing MTU");
> > +		return -EIO;
> > +	}
> > +

> It looks like this proposed patch does not have the accelerated networking
> virtual device case.  The driver needs to first update the MTU of the underlying
> VF device (and handle errors), then reinit the vmbus device.

The following line should handle the AN virtual device case:

+	/* Change MTU of underlying VF dev first, if it exists */
+	ret = hn_vf_mtu_set(dev, mtu);
+	if (ret)
+		return ret;

This is the new method added in hn_vf.c:

--- a/drivers/net/netvsc/hn_vf.c
+++ b/drivers/net/netvsc/hn_vf.c
@@ -778,3 +784,18 @@ int hn_vf_reta_hash_update(struct rte_eth_dev *dev,
 
 	return ret;
 }
+
+int hn_vf_mtu_set(struct rte_eth_dev *dev, uint16_t mtu) {
+	struct hn_data *hv = dev->data->dev_private;
+	struct rte_eth_dev *vf_dev;
+	int ret = 0;
+
+	rte_rwlock_read_lock(&hv->vf_lock);
+	vf_dev = hn_get_vf_dev(hv);
+	if (hv->vf_ctx.vf_vsc_switched && vf_dev)
+		ret = vf_dev->dev_ops->mtu_set(vf_dev, mtu);
+	rte_rwlock_read_unlock(&hv->vf_lock);
+
+	return ret;
+}
  
Stephen Hemminger Oct. 10, 2023, 3:55 p.m. UTC | #3
On Mon, 9 Oct 2023 16:41:38 +0000
Sam Andrew <samandrew@microsoft.com> wrote:

> > It looks like this proposed patch does not have the accelerated networking
> > virtual device case.  The driver needs to first update the MTU of the underlying
> > VF device (and handle errors), then reinit the vmbus device.  
> 
> The following line should handle the AN virtual device case:
> 
> +	/* Change MTU of underlying VF dev first, if it exists */
> +	ret = hn_vf_mtu_set(dev, mtu);
> +	if (ret)
> +		return ret;
> 
> This is the new method added in hn_vf.c:

Yes, that should cover it.
  

Patch

diff --git a/drivers/bus/vmbus/vmbus_common.c b/drivers/bus/vmbus/vmbus_common.c
index 95f3ad78bc..fd205e8ac3 100644
--- a/drivers/bus/vmbus/vmbus_common.c
+++ b/drivers/bus/vmbus/vmbus_common.c
@@ -39,7 +39,9 @@  vmbus_map_resource(void *requested_addr, int fd, off_t offset, size_t size,
 			  "mmap(%d, %p, %zu, %ld) failed: %s",
 			  fd, requested_addr, size, (long)offset,
 			  strerror(errno));
-	}
+	} else
+		VMBUS_LOG(DEBUG, "  VMBUS memory mapped at %p",
+			  mapaddr);
 	return mapaddr;
 }
 
diff --git a/drivers/net/netvsc/hn_ethdev.c b/drivers/net/netvsc/hn_ethdev.c
index d0bbc0a4c0..92e4c72902 100644
--- a/drivers/net/netvsc/hn_ethdev.c
+++ b/drivers/net/netvsc/hn_ethdev.c
@@ -1059,37 +1059,6 @@  hn_dev_close(struct rte_eth_dev *dev)
 	return ret;
 }
 
-static const struct eth_dev_ops hn_eth_dev_ops = {
-	.dev_configure		= hn_dev_configure,
-	.dev_start		= hn_dev_start,
-	.dev_stop		= hn_dev_stop,
-	.dev_close		= hn_dev_close,
-	.dev_infos_get		= hn_dev_info_get,
-	.txq_info_get		= hn_dev_tx_queue_info,
-	.rxq_info_get		= hn_dev_rx_queue_info,
-	.dev_supported_ptypes_get = hn_vf_supported_ptypes,
-	.promiscuous_enable     = hn_dev_promiscuous_enable,
-	.promiscuous_disable    = hn_dev_promiscuous_disable,
-	.allmulticast_enable    = hn_dev_allmulticast_enable,
-	.allmulticast_disable   = hn_dev_allmulticast_disable,
-	.set_mc_addr_list	= hn_dev_mc_addr_list,
-	.reta_update		= hn_rss_reta_update,
-	.reta_query             = hn_rss_reta_query,
-	.rss_hash_update	= hn_rss_hash_update,
-	.rss_hash_conf_get      = hn_rss_hash_conf_get,
-	.tx_queue_setup		= hn_dev_tx_queue_setup,
-	.tx_queue_release	= hn_dev_tx_queue_release,
-	.tx_done_cleanup        = hn_dev_tx_done_cleanup,
-	.rx_queue_setup		= hn_dev_rx_queue_setup,
-	.rx_queue_release	= hn_dev_rx_queue_release,
-	.link_update		= hn_dev_link_update,
-	.stats_get		= hn_dev_stats_get,
-	.stats_reset            = hn_dev_stats_reset,
-	.xstats_get		= hn_dev_xstats_get,
-	.xstats_get_names	= hn_dev_xstats_get_names,
-	.xstats_reset		= hn_dev_xstats_reset,
-};
-
 /*
  * Setup connection between PMD and kernel.
  */
@@ -1129,12 +1098,158 @@  hn_detach(struct hn_data *hv)
 	hn_rndis_detach(hv);
 }
 
+/*
+ * Connects EXISTING rx/tx queues to NEW vmbus channel(s), and
+ * re-initializes NDIS and RNDIS, including re-sending initial
+ * NDIS/RNDIS configuration. To be used after the underlying vmbus
+ * has been un- and re-mapped, e.g. as must happen when the device
+ * MTU is changed.
+ */
+static int
+hn_reinit(struct rte_eth_dev *dev, uint16_t mtu)
+{
+	struct hn_data *hv = dev->data->dev_private;
+	struct hn_rx_queue **rxqs = (struct hn_rx_queue **)dev->data->rx_queues;
+	struct hn_tx_queue **txqs = (struct hn_tx_queue **)dev->data->tx_queues;
+	int i, ret = 0;
+
+	/* Point primary queues at new primary channel */
+	rxqs[0]->chan = hv->channels[0];
+	txqs[0]->chan = hv->channels[0];
+
+	ret = hn_attach(hv, mtu);
+	if (ret)
+		return ret;
+
+	/* Create vmbus subchannels, additional RNDIS configuration */
+	ret = hn_dev_configure(dev);
+	if (ret)
+		return ret;
+
+	/* Point any additional queues at new subchannels */
+	for (i = 1; i < dev->data->nb_rx_queues; i++)
+		rxqs[i]->chan = hv->channels[i];
+	for (i = 1; i < dev->data->nb_tx_queues; i++)
+		txqs[i]->chan = hv->channels[i];
+
+	return ret;
+}
+
+static int
+hn_dev_mtu_set(struct rte_eth_dev *dev, uint16_t mtu)
+{
+	struct hn_data *hv = dev->data->dev_private;
+	unsigned int orig_mtu = dev->data->mtu;
+	uint32_t rndis_mtu;
+	int ret = 0;
+	int i;
+
+	if (dev->data->dev_started) {
+		PMD_DRV_LOG(ERR, "Device must be stopped before changing MTU");
+		return -EIO;
+	}
+
+	/* Change MTU of underlying VF dev first, if it exists */
+	ret = hn_vf_mtu_set(dev, mtu);
+	if (ret)
+		return ret;
+
+	/* Release channel resources */
+	hn_detach(hv);
+
+	/* Close any secondary vmbus channels */
+	for (i = 1; i < hv->num_queues; i++)
+		rte_vmbus_chan_close(hv->channels[i]);
+
+	/* Close primary vmbus channel */
+	rte_free(hv->channels[0]);
+
+	/* Unmap and re-map vmbus device */
+	rte_vmbus_unmap_device(hv->vmbus);
+	ret = rte_vmbus_map_device(hv->vmbus);
+	if (ret) {
+		/* This is a catastrophic error - the device is unusable */
+		PMD_DRV_LOG(ERR, "Could not re-map vmbus device!");
+		return ret;
+	}
+
+	/* Update pointers to re-mapped UIO resources */
+	hv->rxbuf_res = hv->vmbus->resource[HV_RECV_BUF_MAP];
+	hv->chim_res  = hv->vmbus->resource[HV_SEND_BUF_MAP];
+
+	/* Re-open the primary vmbus channel */
+	ret = rte_vmbus_chan_open(hv->vmbus, &hv->channels[0]);
+	if (ret) {
+		/* This is a catastrophic error - the device is unusable */
+		PMD_DRV_LOG(ERR, "Could not re-open vmbus channel!");
+		return ret;
+	}
+
+	rte_vmbus_set_latency(hv->vmbus, hv->channels[0], hv->latency);
+
+	ret = hn_reinit(dev, mtu);
+	if (!ret)
+		goto out;
+
+	/* In case of error, attempt to restore original MTU */
+	ret = hn_reinit(dev, orig_mtu);
+	if (ret)
+		PMD_DRV_LOG(ERR, "Restoring original MTU failed for netvsc");
+
+	ret = hn_vf_mtu_set(dev, orig_mtu);
+	if (ret)
+		PMD_DRV_LOG(ERR, "Restoring original MTU failed for VF");
+
+out:
+	if (hn_rndis_get_mtu(hv, &rndis_mtu))
+		PMD_DRV_LOG(ERR, "Could not get MTU via RNDIS");
+	else {
+		dev->data->mtu = (uint16_t)rndis_mtu;
+		PMD_DRV_LOG(DEBUG, "RNDIS MTU is %u", dev->data->mtu);
+	}
+
+	return ret;
+}
+
+static const struct eth_dev_ops hn_eth_dev_ops = {
+	.dev_configure		= hn_dev_configure,
+	.dev_start		= hn_dev_start,
+	.dev_stop		= hn_dev_stop,
+	.dev_close		= hn_dev_close,
+	.dev_infos_get		= hn_dev_info_get,
+	.txq_info_get		= hn_dev_tx_queue_info,
+	.rxq_info_get		= hn_dev_rx_queue_info,
+	.dev_supported_ptypes_get = hn_vf_supported_ptypes,
+	.promiscuous_enable     = hn_dev_promiscuous_enable,
+	.promiscuous_disable    = hn_dev_promiscuous_disable,
+	.allmulticast_enable    = hn_dev_allmulticast_enable,
+	.allmulticast_disable   = hn_dev_allmulticast_disable,
+	.set_mc_addr_list	= hn_dev_mc_addr_list,
+	.mtu_set                = hn_dev_mtu_set,
+	.reta_update		= hn_rss_reta_update,
+	.reta_query             = hn_rss_reta_query,
+	.rss_hash_update	= hn_rss_hash_update,
+	.rss_hash_conf_get      = hn_rss_hash_conf_get,
+	.tx_queue_setup		= hn_dev_tx_queue_setup,
+	.tx_queue_release	= hn_dev_tx_queue_release,
+	.tx_done_cleanup        = hn_dev_tx_done_cleanup,
+	.rx_queue_setup		= hn_dev_rx_queue_setup,
+	.rx_queue_release	= hn_dev_rx_queue_release,
+	.link_update		= hn_dev_link_update,
+	.stats_get		= hn_dev_stats_get,
+	.stats_reset            = hn_dev_stats_reset,
+	.xstats_get		= hn_dev_xstats_get,
+	.xstats_get_names	= hn_dev_xstats_get_names,
+	.xstats_reset		= hn_dev_xstats_reset,
+};
+
 static int
 eth_hn_dev_init(struct rte_eth_dev *eth_dev)
 {
 	struct hn_data *hv = eth_dev->data->dev_private;
 	struct rte_device *device = eth_dev->device;
 	struct rte_vmbus_device *vmbus;
+	uint32_t mtu;
 	unsigned int rxr_cnt;
 	int err, max_chan;
 
@@ -1218,6 +1333,12 @@  eth_hn_dev_init(struct rte_eth_dev *eth_dev)
 	if (err)
 		goto failed;
 
+	err = hn_rndis_get_mtu(hv, &mtu);
+	if (err)
+		goto failed;
+	eth_dev->data->mtu = (uint16_t)mtu;
+	PMD_INIT_LOG(DEBUG, "RNDIS MTU is %u", eth_dev->data->mtu);
+
 	err = hn_rndis_get_eaddr(hv, eth_dev->data->mac_addrs->addr_bytes);
 	if (err)
 		goto failed;
@@ -1272,7 +1393,7 @@  eth_hn_dev_uninit(struct rte_eth_dev *eth_dev)
 
 	hn_detach(hv);
 	hn_chim_uninit(eth_dev);
-	rte_vmbus_chan_close(hv->primary->chan);
+	rte_vmbus_chan_close(hv->channels[0]);
 	rte_free(hv->primary);
 	ret = rte_eth_dev_owner_delete(hv->owner.id);
 	if (ret != 0)
diff --git a/drivers/net/netvsc/hn_rndis.c b/drivers/net/netvsc/hn_rndis.c
index 29c6009b2c..778f86f2bf 100644
--- a/drivers/net/netvsc/hn_rndis.c
+++ b/drivers/net/netvsc/hn_rndis.c
@@ -1111,6 +1111,13 @@  hn_rndis_get_eaddr(struct hn_data *hv, uint8_t *eaddr)
 	return 0;
 }
 
+int
+hn_rndis_get_mtu(struct hn_data *hv, uint32_t *mtu)
+{
+	return hn_rndis_query(hv, OID_GEN_MAXIMUM_FRAME_SIZE, NULL, 0,
+			       mtu, sizeof(uint32_t));
+}
+
 int
 hn_rndis_get_linkstatus(struct hn_data *hv)
 {
diff --git a/drivers/net/netvsc/hn_rndis.h b/drivers/net/netvsc/hn_rndis.h
index 9a8251fc2f..7f40f6221d 100644
--- a/drivers/net/netvsc/hn_rndis.h
+++ b/drivers/net/netvsc/hn_rndis.h
@@ -10,6 +10,7 @@  void	hn_rndis_link_status(struct rte_eth_dev *dev, const void *msg);
 int	hn_rndis_attach(struct hn_data *hv);
 void	hn_rndis_detach(struct hn_data *hv);
 int	hn_rndis_get_eaddr(struct hn_data *hv, uint8_t *eaddr);
+int	hn_rndis_get_mtu(struct hn_data *hv, uint32_t *mtu);
 int	hn_rndis_get_linkstatus(struct hn_data *hv);
 int	hn_rndis_get_linkspeed(struct hn_data *hv);
 int	hn_rndis_set_rxfilter(struct hn_data *hv, uint32_t filter);
diff --git a/drivers/net/netvsc/hn_var.h b/drivers/net/netvsc/hn_var.h
index e1f8e69a28..e37946804d 100644
--- a/drivers/net/netvsc/hn_var.h
+++ b/drivers/net/netvsc/hn_var.h
@@ -13,7 +13,7 @@ 
  * Tunable ethdev params
  */
 #define HN_MIN_RX_BUF_SIZE	1024
-#define HN_MAX_XFER_LEN		2048
+#define HN_MAX_XFER_LEN		RTE_ETHER_MAX_JUMBO_FRAME_LEN
 #define	HN_MAX_MAC_ADDRS	1
 #define HN_MAX_CHANNELS		64
 
@@ -287,6 +287,7 @@  int	hn_vf_rss_hash_update(struct rte_eth_dev *dev,
 int	hn_vf_reta_hash_update(struct rte_eth_dev *dev,
 			       struct rte_eth_rss_reta_entry64 *reta_conf,
 			       uint16_t reta_size);
+int hn_vf_mtu_set(struct rte_eth_dev *dev, uint16_t mtu);
 int	hn_eth_rmv_event_callback(uint16_t port_id,
 				  enum rte_eth_event_type event __rte_unused,
 				  void *cb_arg, void *out __rte_unused);
diff --git a/drivers/net/netvsc/hn_vf.c b/drivers/net/netvsc/hn_vf.c
index 782395d805..90cb6f6923 100644
--- a/drivers/net/netvsc/hn_vf.c
+++ b/drivers/net/netvsc/hn_vf.c
@@ -239,7 +239,7 @@  int hn_vf_add(struct rte_eth_dev *dev, struct hn_data *hv)
 
 	port = hv->vf_ctx.vf_port;
 
-	/* If the primary device has started, this is a VF host add.
+	/* If the primary device has started, this is a VF hot add.
 	 * Configure and start VF device.
 	 */
 	if (dev->data->dev_started) {
@@ -264,6 +264,12 @@  int hn_vf_add(struct rte_eth_dev *dev, struct hn_data *hv)
 			goto exit;
 		}
 
+		ret = hn_vf_mtu_set(dev, dev->data->mtu);
+		if (ret) {
+			PMD_DRV_LOG(ERR, "Failed to set VF MTU");
+			goto exit;
+		}
+
 		PMD_DRV_LOG(NOTICE, "Starting VF port %d", port);
 		ret = rte_eth_dev_start(port);
 		if (ret) {
@@ -778,3 +784,18 @@  int hn_vf_reta_hash_update(struct rte_eth_dev *dev,
 
 	return ret;
 }
+
+int hn_vf_mtu_set(struct rte_eth_dev *dev, uint16_t mtu)
+{
+	struct hn_data *hv = dev->data->dev_private;
+	struct rte_eth_dev *vf_dev;
+	int ret = 0;
+
+	rte_rwlock_read_lock(&hv->vf_lock);
+	vf_dev = hn_get_vf_dev(hv);
+	if (hv->vf_ctx.vf_vsc_switched && vf_dev)
+		ret = vf_dev->dev_ops->mtu_set(vf_dev, mtu);
+	rte_rwlock_read_unlock(&hv->vf_lock);
+
+	return ret;
+}