[patch net-next v3 0/3] export device physical port id to userspace

[patch net-next v3 0/3] export device physical port id to userspace

[Jiri Pirko]

This patchset is based on patch by Narendra_K@Dell.com
Once device which can change phys port id during its lifetime adopts this,
NETDEV_CHANGEPHYSPORTID event will be added and driver will call
call_netdevice_notifiers(NETDEV_NETDEV_CHANGEPHYSPORTID, dev) to propagate
the change to userspace.

v1->v2: as suggested by Ben, handle -EOPNOTSUPP in rtnl code (wrapped up ndo call)
v2->v3: adjusted patch 1 commit message

Jiri Pirko (3):
  net: add ndo to get id of physical port of the device
  rtnl: export physical port id via RT netlink
  net: export physical port id via sysfs

 include/linux/netdevice.h    | 20 ++++++++++++++++++++
 include/uapi/linux/if_link.h |  1 +
 net/core/dev.c               | 18 ++++++++++++++++++
 net/core/net-sysfs.c         | 42 ++++++++++++++++++++++++++++++++++++++++++
 net/core/rtnetlink.c         | 25 ++++++++++++++++++++++++-
 5 files changed, 105 insertions(+), 1 deletion(-)

-- 
1.8.1.4

[patch net-next v3 1/3] net: add ndo to get id of physical port of the device

[Jiri Pirko]

This patch adds a ndo for getting physical port of the device. Driver
which is aware of being virtual function of some physical port should
implement this ndo. This is applicable not only for IOV, but for other
solutions (NPAR, multichannel) as well. Basically if there is possible
to have multiple netdevs on the single hw port.

Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Acked-by: Ben Hutchings <bhutchings@solarflare.com>
---
 include/linux/netdevice.h | 20 ++++++++++++++++++++
 net/core/dev.c            | 18 ++++++++++++++++++
 2 files changed, 38 insertions(+)

diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index 0741a1e..726dec2 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h
@@ -728,6 +728,16 @@ struct netdev_fcoe_hbainfo {
 };
 #endif
 
+#define MAX_PHYS_PORT_ID_LEN 32
+
+/* This structure holds a unique identifier to identify the
+ * physical port used by a netdevice.
+ */
+struct netdev_phys_port_id {
+	unsigned char id[MAX_PHYS_PORT_ID_LEN];
+	unsigned char id_len;
+};
+
 /*
  * This structure defines the management hooks for network devices.
  * The following hooks can be defined; unless noted otherwise, they are
@@ -932,6 +942,12 @@ struct netdev_fcoe_hbainfo {
  *	that determine carrier state from physical hardware properties (eg
  *	network cables) or protocol-dependent mechanisms (eg
  *	USB_CDC_NOTIFY_NETWORK_CONNECTION) should NOT implement this function.
+ *
+ * int (*ndo_get_phys_port_id)(struct net_device *dev,
+ *			       struct netdev_phys_port_id *ppid);
+ *	Called to get ID of physical port of this device. If driver does
+ *	not implement this, it is assumed that the hw is not able to have
+ *	multiple net devices on single physical port.
  */
 struct net_device_ops {
 	int			(*ndo_init)(struct net_device *dev);
@@ -1060,6 +1076,8 @@ struct net_device_ops {
 						      struct nlmsghdr *nlh);
 	int			(*ndo_change_carrier)(struct net_device *dev,
 						      bool new_carrier);
+	int			(*ndo_get_phys_port_id)(struct net_device *dev,
+							struct netdev_phys_port_id *ppid);
 };
 
 /*
@@ -2317,6 +2335,8 @@ extern int		dev_set_mac_address(struct net_device *,
 					    struct sockaddr *);
 extern int		dev_change_carrier(struct net_device *,
 					   bool new_carrier);
+extern int		dev_get_phys_port_id(struct net_device *dev,
+					     struct netdev_phys_port_id *ppid);
 extern int		dev_hard_start_xmit(struct sk_buff *skb,
 					    struct net_device *dev,
 					    struct netdev_queue *txq);
diff --git a/net/core/dev.c b/net/core/dev.c
index 26755dd..90172eb 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -4989,6 +4989,24 @@ int dev_change_carrier(struct net_device *dev, bool new_carrier)
 EXPORT_SYMBOL(dev_change_carrier);
 
 /**
+ *	dev_get_phys_port_id - Get device physical port ID
+ *	@dev: device
+ *	@ppid: port ID
+ *
+ *	Get device physical port ID
+ */
+int dev_get_phys_port_id(struct net_device *dev,
+			 struct netdev_phys_port_id *ppid)
+{
+	const struct net_device_ops *ops = dev->netdev_ops;
+
+	if (!ops->ndo_get_phys_port_id)
+		return -EOPNOTSUPP;
+	return ops->ndo_get_phys_port_id(dev, ppid);
+}
+EXPORT_SYMBOL(dev_get_phys_port_id);
+
+/**
  *	dev_new_index	-	allocate an ifindex
  *	@net: the applicable net namespace
  *
-- 
1.8.1.4

Re: [patch net-next v3 1/3] net: add ndo to get id of physical port of the device

[Narendra_K]

On Mon, Jul 22, 2013 at 06:35:15PM +0530, Jiri Pirko wrote:
> 
> This patch adds a ndo for getting physical port of the device. Driver
> which is aware of being virtual function of some physical port should
> implement this ndo. This is applicable not only for IOV, but for other
> solutions (NPAR, multichannel) as well. Basically if there is possible
> to have multiple netdevs on the single hw port.
> 
> Signed-off-by: Jiri Pirko <jiri@resnulli.us>
> Acked-by: Ben Hutchings <bhutchings@solarflare.com>
> ---

Signed-off-by: Narendra K <narendra_k@dell.com>

-- 
With regards,
Narendra K
Linux Engineering
Dell Inc.

Re: [patch net-next v3 1/3] net: add ndo to get id of physical port of the device

[Or Gerlitz]

On Mon, Jul 22, 2013 at 4:05 PM, Jiri Pirko <jiri@resnulli.us> wrote:
> This patch adds a ndo for getting physical port of the device. Driver
> which is aware of being virtual function of some physical port should
> implement this ndo. This is applicable not only for IOV, but for other
> solutions (NPAR, multichannel) as well. Basically if there is possible
> to have multiple netdevs on the single hw port.

BTW - one thing which can be helpful for the discussion, is to have
this new ndo actually implemented for a certain driver. Also typically
we prefer to avoid adding new features at the networking core layer
without any concrete user. Making some driver to implement a new
feature is always a good way to see it flies well.

Or.

>
> Signed-off-by: Jiri Pirko <jiri@resnulli.us>
> Acked-by: Ben Hutchings <bhutchings@solarflare.com>

Re: [patch net-next v3 1/3] net: add ndo to get id of physical port of the device

[Jiri Pirko]

Mon, Jul 22, 2013 at 05:59:37PM CEST, or.gerlitz@gmail.com wrote:
>On Mon, Jul 22, 2013 at 4:05 PM, Jiri Pirko <jiri@resnulli.us> wrote:
>> This patch adds a ndo for getting physical port of the device. Driver
>> which is aware of being virtual function of some physical port should
>> implement this ndo. This is applicable not only for IOV, but for other
>> solutions (NPAR, multichannel) as well. Basically if there is possible
>> to have multiple netdevs on the single hw port.
>
>BTW - one thing which can be helpful for the discussion, is to have
>this new ndo actually implemented for a certain driver. Also typically
>we prefer to avoid adding new features at the networking core layer
>without any concrete user. Making some driver to implement a new
>feature is always a good way to see it flies well.

I have almost done implementation for igb/igbvf (consulting with Intel
guys) and I'm working on bnx2x NPAR as well (consulting with Broadcom). I will
post these patches soon.

After that, I plan to work on more drivers.

>
>Or.
>
>>
>> Signed-off-by: Jiri Pirko <jiri@resnulli.us>
>> Acked-by: Ben Hutchings <bhutchings@solarflare.com>

Re: [patch net-next v3 1/3] net: add ndo to get id of physical port of the device

[Or Gerlitz]

On Mon, Jul 22, 2013 at 9:01 PM, Jiri Pirko <jiri@resnulli.us> wrote:

> I have almost done implementation for igb/igbvf (consulting with Intel
> guys) and I'm working on bnx2x NPAR as well (consulting with Broadcom). I
> will post these patches soon.

good, I would prefer to have at least one driver patch as part of the
initial series, as typically done when adding e.g new ndo entry, makes
sense?

Or.

[patch net-next v3 2/3] rtnl: export physical port id via RT netlink

[Jiri Pirko]

Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Acked-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: Narendra K <narendra_k@dell.com>
---
 include/uapi/linux/if_link.h |  1 +
 net/core/rtnetlink.c         | 25 ++++++++++++++++++++++++-
 2 files changed, 25 insertions(+), 1 deletion(-)

diff --git a/include/uapi/linux/if_link.h b/include/uapi/linux/if_link.h
index 03f6170..04c0e7a 100644
--- a/include/uapi/linux/if_link.h
+++ b/include/uapi/linux/if_link.h
@@ -143,6 +143,7 @@ enum {
 	IFLA_NUM_TX_QUEUES,
 	IFLA_NUM_RX_QUEUES,
 	IFLA_CARRIER,
+	IFLA_PHYS_PORT_ID,
 	__IFLA_MAX
 };
 
diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c
index 3de7408..0b2972c 100644
--- a/net/core/rtnetlink.c
+++ b/net/core/rtnetlink.c
@@ -767,7 +767,8 @@ static noinline size_t if_nlmsg_size(const struct net_device *dev,
 	       + rtnl_vfinfo_size(dev, ext_filter_mask) /* IFLA_VFINFO_LIST */
 	       + rtnl_port_size(dev) /* IFLA_VF_PORTS + IFLA_PORT_SELF */
 	       + rtnl_link_get_size(dev) /* IFLA_LINKINFO */
-	       + rtnl_link_get_af_size(dev); /* IFLA_AF_SPEC */
+	       + rtnl_link_get_af_size(dev) /* IFLA_AF_SPEC */
+	       + nla_total_size(MAX_PHYS_PORT_ID_LEN); /* IFLA_PHYS_PORT_ID */
 }
 
 static int rtnl_vf_ports_fill(struct sk_buff *skb, struct net_device *dev)
@@ -846,6 +847,24 @@ static int rtnl_port_fill(struct sk_buff *skb, struct net_device *dev)
 	return 0;
 }
 
+static int rtnl_phys_port_id_fill(struct sk_buff *skb, struct net_device *dev)
+{
+	int err;
+	struct netdev_phys_port_id ppid;
+
+	err = dev_get_phys_port_id(dev, &ppid);
+	if (err) {
+		if (err == -EOPNOTSUPP)
+			return 0;
+		return err;
+	}
+
+	if (nla_put(skb, IFLA_PHYS_PORT_ID, ppid.id_len, ppid.id))
+		return -EMSGSIZE;
+
+	return 0;
+}
+
 static int rtnl_fill_ifinfo(struct sk_buff *skb, struct net_device *dev,
 			    int type, u32 pid, u32 seq, u32 change,
 			    unsigned int flags, u32 ext_filter_mask)
@@ -913,6 +932,9 @@ static int rtnl_fill_ifinfo(struct sk_buff *skb, struct net_device *dev,
 			goto nla_put_failure;
 	}
 
+	if (rtnl_phys_port_id_fill(skb, dev))
+		goto nla_put_failure;
+
 	attr = nla_reserve(skb, IFLA_STATS,
 			sizeof(struct rtnl_link_stats));
 	if (attr == NULL)
@@ -1113,6 +1135,7 @@ const struct nla_policy ifla_policy[IFLA_MAX+1] = {
 	[IFLA_PROMISCUITY]	= { .type = NLA_U32 },
 	[IFLA_NUM_TX_QUEUES]	= { .type = NLA_U32 },
 	[IFLA_NUM_RX_QUEUES]	= { .type = NLA_U32 },
+	[IFLA_PHYS_PORT_ID]	= { .type = NLA_BINARY, .len = MAX_PHYS_PORT_ID_LEN },
 };
 EXPORT_SYMBOL(ifla_policy);
 
-- 
1.8.1.4

[patch net-next v3 3/3] net: export physical port id via sysfs

[Jiri Pirko]

Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Acked-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: Narendra K <narendra_k@dell.com>
---
 net/core/net-sysfs.c | 42 ++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 42 insertions(+)

diff --git a/net/core/net-sysfs.c b/net/core/net-sysfs.c
index 981fed3..56b4635 100644
--- a/net/core/net-sysfs.c
+++ b/net/core/net-sysfs.c
@@ -334,6 +334,47 @@ static ssize_t store_group(struct device *dev, struct device_attribute *attr,
 	return netdev_store(dev, attr, buf, len, change_group);
 }
 
+static size_t _format_port_id(char *buf, int buflen,
+			      const unsigned char *id, int len)
+{
+	int i;
+	char *cp = buf;
+
+	for (i = 0; i < len; i++)
+		cp += scnprintf(cp, buflen - (cp - buf), "%02x", id[i]);
+	return cp - buf;
+}
+
+ssize_t sysfs_format_port_id(char *buf, const unsigned char *id, int len)
+{
+	size_t l;
+
+	l = _format_port_id(buf, PAGE_SIZE, id, len);
+	l += scnprintf(buf + l, PAGE_SIZE - l, "\n");
+	return (ssize_t)l;
+}
+
+static ssize_t show_phys_port_id(struct device *dev,
+				 struct device_attribute *attr, char *buf)
+{
+	struct net_device *netdev = to_net_dev(dev);
+	ssize_t ret = -EINVAL;
+
+	if (!rtnl_trylock())
+		return restart_syscall();
+
+	if (dev_isalive(netdev)) {
+		struct netdev_phys_port_id ppid;
+
+		ret = dev_get_phys_port_id(netdev, &ppid);
+		if (!ret)
+			ret = sysfs_format_port_id(buf, ppid.id, ppid.id_len);
+	}
+	rtnl_unlock();
+
+	return ret;
+}
+
 static struct device_attribute net_class_attributes[] = {
 	__ATTR(addr_assign_type, S_IRUGO, show_addr_assign_type, NULL),
 	__ATTR(addr_len, S_IRUGO, show_addr_len, NULL),
@@ -355,6 +396,7 @@ static struct device_attribute net_class_attributes[] = {
 	__ATTR(tx_queue_len, S_IRUGO | S_IWUSR, show_tx_queue_len,
 	       store_tx_queue_len),
 	__ATTR(netdev_group, S_IRUGO | S_IWUSR, show_group, store_group),
+	__ATTR(phys_port_id, S_IRUGO, show_phys_port_id, NULL),
 	{}
 };
 
-- 
1.8.1.4

Re: [patch net-next v3 3/3] net: export physical port id via sysfs

[David Miller]

From: Jiri Pirko <jiri@resnulli.us>
Date: Mon, 22 Jul 2013 15:05:17 +0200

> +static size_t _format_port_id(char *buf, int buflen,
> +			      const unsigned char *id, int len)
> +{
> +	int i;
> +	char *cp = buf;
> +
> +	for (i = 0; i < len; i++)
> +		cp += scnprintf(cp, buflen - (cp - buf), "%02x", id[i]);
> +	return cp - buf;
> +}

This is "%phN", please use the printf strings we have support for
in the kernel instead of reimplementing them :-)

Re: [patch net-next v3 0/3] export device physical port id to userspace

[David Miller]

From: Jiri Pirko <jiri@resnulli.us>
Date: Mon, 22 Jul 2013 15:05:14 +0200

> This patchset is based on patch by Narendra_K@Dell.com
> Once device which can change phys port id during its lifetime adopts this,
> NETDEV_CHANGEPHYSPORTID event will be added and driver will call
> call_netdevice_notifiers(NETDEV_NETDEV_CHANGEPHYSPORTID, dev) to propagate
> the change to userspace.
> 
> v1->v2: as suggested by Ben, handle -EOPNOTSUPP in rtnl code (wrapped up ndo call)
> v2->v3: adjusted patch 1 commit message

As mentioned we'd like to have at least one driver implementing the new
interface as part of the patch set.

Once you have that sorted I think this is mostly ready.

Re: [patch net-next v3 0/3] export device physical port id to userspace

[Jiri Pirko]

Thu, Jul 25, 2013 at 01:00:37AM CEST, davem@davemloft.net wrote:
>From: Jiri Pirko <jiri@resnulli.us>
>Date: Mon, 22 Jul 2013 15:05:14 +0200
>
>> This patchset is based on patch by Narendra_K@Dell.com
>> Once device which can change phys port id during its lifetime adopts this,
>> NETDEV_CHANGEPHYSPORTID event will be added and driver will call
>> call_netdevice_notifiers(NETDEV_NETDEV_CHANGEPHYSPORTID, dev) to propagate
>> the change to userspace.
>> 
>> v1->v2: as suggested by Ben, handle -EOPNOTSUPP in rtnl code (wrapped up ndo call)
>> v2->v3: adjusted patch 1 commit message
>
>As mentioned we'd like to have at least one driver implementing the new
>interface as part of the patch set.
>
>Once you have that sorted I think this is mostly ready.


Allright. Will repost with most probably igb/igbvf support implemented
as well. Thanks!