Hi,
ok, at least some progress has happened:
-Replaced device-specific oper_state method with NETIF_F_STACKED flag to
select between IF_OPER_DOWN or IF_OPER_LOWERLAYERDOWN
-sysfs support to read operstate
-completed netlink support (Jamal, Thomas, can you verify the code?)
-added netif_oper_up() query function
-treat IF_OPER_UNKNOWN equivalent to IF_OPER_UP in some cases to have compat
to devices that do not set carrier state
-adopted vlan drivers
-verified operation with starfire, vlan, loopback. IF_OPER_UNKNOWN won't
propagate to upper layers when stacking devices.
Todo:
-complete sysfs
-test netlink userspace interaction
-docs
Stefan
diff -X dontdiff -ur linux-2.6.14/include/linux/if.h linux-2.6.14-rfc2863/include/linux/if.h
--- linux-2.6.14/include/linux/if.h 2005-11-02 11:07:32.000000000 +0100
+++ linux-2.6.14-rfc2863/include/linux/if.h 2005-11-18 20:14:25.000000000 +0100
@@ -33,7 +33,7 @@
#define IFF_LOOPBACK 0x8 /* is a loopback net */
#define IFF_POINTOPOINT 0x10 /* interface is has p-p link */
#define IFF_NOTRAILERS 0x20 /* avoid use of trailers */
-#define IFF_RUNNING 0x40 /* interface running and carrier ok */
+#define IFF_RUNNING 0x40 /* interface RFC2863 OPER_UP */
#define IFF_NOARP 0x80 /* no ARP protocol */
#define IFF_PROMISC 0x100 /* receive all packets */
#define IFF_ALLMULTI 0x200 /* receive all multicast packets*/
@@ -43,12 +43,16 @@
#define IFF_MULTICAST 0x1000 /* Supports multicast */
-#define IFF_VOLATILE (IFF_LOOPBACK|IFF_POINTOPOINT|IFF_BROADCAST|IFF_MASTER|IFF_SLAVE|IFF_RUNNING)
-
#define IFF_PORTSEL 0x2000 /* can set media type */
#define IFF_AUTOMEDIA 0x4000 /* auto media select active */
#define IFF_DYNAMIC 0x8000 /* dialup device with changing addresses*/
+#define IFF_CARRIER 0x10000 /* driver signals carrier */
+#define IFF_DORMANT 0x20000 /* driver signals dormant */
+
+#define IFF_VOLATILE (IFF_LOOPBACK|IFF_POINTOPOINT|IFF_BROADCAST|\
+ IFF_MASTER|IFF_SLAVE|IFF_RUNNING|IFF_CARRIER|IFF_DORMANT)
+
/* Private (from user) interface flags (netdevice->priv_flags). */
#define IFF_802_1Q_VLAN 0x1 /* 802.1Q VLAN device. */
#define IFF_EBRIDGE 0x2 /* Ethernet bridging device. */
@@ -80,6 +84,22 @@
#define IF_PROTO_FR_ETH_PVC 0x200B
#define IF_PROTO_RAW 0x200C /* RAW Socket */
+/* RFC 2863 operational status */
+enum {
+ IF_OPER_UNKNOWN,
+ IF_OPER_NOTPRESENT,
+ IF_OPER_DOWN,
+ IF_OPER_LOWERLAYERDOWN,
+ IF_OPER_TESTING,
+ IF_OPER_DORMANT,
+ IF_OPER_UP,
+};
+
+/* link modes */
+enum {
+ IF_LINK_MODE_DEFAULT,
+ IF_LINK_MODE_DORMANT, /* limit upward transition to dormant */
+};
/*
* Device mapping structure. I'd just gone off and designed a
diff -X dontdiff -ur linux-2.6.14/include/linux/if_vlan.h linux-2.6.14-rfc2863/include/linux/if_vlan.h
--- linux-2.6.14/include/linux/if_vlan.h 2005-11-02 11:08:10.000000000 +0100
+++ linux-2.6.14-rfc2863/include/linux/if_vlan.h 2005-11-28 13:50:07.000000000 +0100
@@ -64,6 +64,8 @@
/* found in socket.c */
extern void vlan_ioctl_set(int (*hook)(void __user *));
+extern void vlan_transfer_operstate(const struct net_device *dev, struct net_device *vlandev);
+
#define VLAN_NAME "vlan"
/* if this changes, algorithm will have to be reworked because this
diff -X dontdiff -ur linux-2.6.14/include/linux/netdevice.h linux-2.6.14-rfc2863/include/linux/netdevice.h
--- linux-2.6.14/include/linux/netdevice.h 2005-11-02 11:08:10.000000000 +0100
+++ linux-2.6.14-rfc2863/include/linux/netdevice.h 2005-11-28 13:33:35.000000000 +0100
@@ -230,7 +230,8 @@
__LINK_STATE_SCHED,
__LINK_STATE_NOCARRIER,
__LINK_STATE_RX_SCHED,
- __LINK_STATE_LINKWATCH_PENDING
+ __LINK_STATE_LINKWATCH_PENDING,
+ __LINK_STATE_DORMANT,
};
@@ -308,6 +309,7 @@
#define NETIF_F_VLAN_CHALLENGED 1024 /* Device cannot handle VLAN packets */
#define NETIF_F_TSO 2048 /* Can offload TCP/IP segmentation */
#define NETIF_F_LLTX 4096 /* LockLess TX */
+#define NETIF_F_STACKED 8192 /* Interface is stacked */
struct net_device *next_sched;
@@ -334,11 +336,14 @@
*/
- unsigned short flags; /* interface flags (a la BSD) */
+ unsigned int flags; /* interface flags (a la BSD) */
unsigned short gflags;
unsigned short priv_flags; /* Like 'flags' but invisible to userspace. */
unsigned short padded; /* How much padding added by alloc_netdev() */
+ unsigned char operstate; /* RFC2863 operstate */
+ unsigned char link_mode; /* mapping policy to operstate */
+
unsigned mtu; /* interface MTU value */
unsigned short type; /* interface hardware type */
unsigned short hard_header_len; /* hardware hdr length */
@@ -712,6 +717,10 @@
/* Carrier loss detection, dial on demand. The functions netif_carrier_on
* and _off may be called from IRQ context, but it is caller
* who is responsible for serialization of these calls.
+ *
+ * The name carrier is inappropriate, these functions should really be
+ * called netif_lowerlayer_*() because they represent the state of any
+ * kind of lower layer not just hardware media.
*/
extern void linkwatch_fire_event(struct net_device *dev);
@@ -727,6 +736,29 @@
extern void netif_carrier_off(struct net_device *dev);
+static inline void netif_dormant_on(struct net_device *dev)
+{
+ if (!test_and_set_bit(__LINK_STATE_DORMANT, &dev->state))
+ linkwatch_fire_event(dev);
+}
+
+static inline void netif_dormant_off(struct net_device *dev)
+{
+ if (test_and_clear_bit(__LINK_STATE_DORMANT, &dev->state))
+ linkwatch_fire_event(dev);
+}
+
+static inline int netif_dormant(const struct net_device *dev)
+{
+ return test_bit(__LINK_STATE_DORMANT, &dev->state);
+}
+
+
+static inline int netif_oper_up(const struct net_device *dev) {
+ return (dev->operstate == IF_OPER_UP ||
+ dev->operstate == IF_OPER_UNKNOWN /* backward compat */);
+}
+
/* Hot-plugging. */
static inline int netif_device_present(struct net_device *dev)
{
diff -X dontdiff -ur linux-2.6.14/include/linux/rtnetlink.h linux-2.6.14-rfc2863/include/linux/rtnetlink.h
--- linux-2.6.14/include/linux/rtnetlink.h 2005-11-02 11:08:11.000000000 +0100
+++ linux-2.6.14-rfc2863/include/linux/rtnetlink.h 2005-11-18 20:14:05.000000000 +0100
@@ -733,6 +733,8 @@
#define IFLA_MAP IFLA_MAP
IFLA_WEIGHT,
#define IFLA_WEIGHT IFLA_WEIGHT
+ IFLA_OPERSTATE,
+ IFLA_LINKMODE,
__IFLA_MAX
};
diff -X dontdiff -ur linux-2.6.14/net/8021q/vlan.c linux-2.6.14-rfc2863/net/8021q/vlan.c
--- linux-2.6.14/net/8021q/vlan.c 2005-11-02 11:07:35.000000000 +0100
+++ linux-2.6.14-rfc2863/net/8021q/vlan.c 2005-11-28 13:53:27.000000000 +0100
@@ -68,7 +68,7 @@
/* Bits of netdev state that are propagated from real device to virtual */
#define VLAN_LINK_STATE_MASK \
- ((1<<__LINK_STATE_PRESENT)|(1<<__LINK_STATE_NOCARRIER))
+ ((1<<__LINK_STATE_PRESENT)|(1<<__LINK_STATE_NOCARRIER)|(1<<__LINK_STATE_DORMANT))
/* End of global variables definitions. */
@@ -325,6 +325,8 @@
*/
new_dev->get_stats = vlan_dev_get_stats;
+ new_dev->features |= NETIF_F_STACKED;
+
/* Make this thing known as a VLAN device */
new_dev->priv_flags |= IFF_802_1Q_VLAN;
@@ -343,6 +345,26 @@
new_dev->do_ioctl = vlan_dev_ioctl;
}
+void vlan_transfer_operstate(const struct net_device *dev, struct net_device *vlandev)
+{
+ if (netif_carrier_ok(dev)) {
+ if (!netif_carrier_ok(vlandev))
+ netif_carrier_on(vlandev);
+ } else {
+ if (netif_carrier_ok(vlandev))
+ netif_carrier_off(vlandev);
+ }
+
+ /* Have to respect userspace enforced dormant state
+ * of real device, also must allow supplicant running
+ * on VLAN device
+ */
+ if (dev->operstate == IF_OPER_DORMANT)
+ netif_dormant_on(vlandev);
+ else
+ netif_dormant_off(vlandev);
+}
+
/* Attach a VLAN device to a mac address (ie Ethernet Card).
* Returns the device that was created, or NULL if there was
* an error of some kind.
@@ -578,13 +600,7 @@
if (!vlandev)
continue;
- if (netif_carrier_ok(dev)) {
- if (!netif_carrier_ok(vlandev))
- netif_carrier_on(vlandev);
- } else {
- if (netif_carrier_ok(vlandev))
- netif_carrier_off(vlandev);
- }
+ vlan_transfer_operstate(dev, vlandev);
if ((vlandev->state & VLAN_LINK_STATE_MASK) != flgs) {
vlandev->state = (vlandev->state &~ VLAN_LINK_STATE_MASK)
diff -X dontdiff -ur linux-2.6.14/net/8021q/vlan_dev.c linux-2.6.14-rfc2863/net/8021q/vlan_dev.c
--- linux-2.6.14/net/8021q/vlan_dev.c 2005-11-02 11:08:11.000000000 +0100
+++ linux-2.6.14-rfc2863/net/8021q/vlan_dev.c 2005-11-28 13:56:33.000000000 +0100
@@ -786,6 +786,8 @@
{
if (!(VLAN_DEV_INFO(dev)->real_dev->flags & IFF_UP))
return -ENETDOWN;
+ vlan_transfer_operstate(VLAN_DEV_INFO(dev)->real_dev, dev);
+ linkwatch_fire_event(dev); /* force event to setup operstate */
return 0;
}
diff -X dontdiff -ur linux-2.6.14/net/core/dev.c linux-2.6.14-rfc2863/net/core/dev.c
--- linux-2.6.14/net/core/dev.c 2005-11-06 17:35:22.000000000 +0100
+++ linux-2.6.14-rfc2863/net/core/dev.c 2005-11-28 13:33:58.000000000 +0100
@@ -2141,12 +2141,20 @@
flags = (dev->flags & ~(IFF_PROMISC |
IFF_ALLMULTI |
- IFF_RUNNING)) |
+ IFF_RUNNING |
+ IFF_CARRIER |
+ IFF_DORMANT)) |
(dev->gflags & (IFF_PROMISC |
IFF_ALLMULTI));
- if (netif_running(dev) && netif_carrier_ok(dev))
- flags |= IFF_RUNNING;
+ if (netif_running(dev)) {
+ if (netif_oper_up(dev))
+ flags |= IFF_RUNNING;
+ if (netif_carrier_ok(dev))
+ flags |= IFF_CARRIER;
+ if (netif_dormant(dev))
+ flags |= IFF_DORMANT;
+ }
return flags;
}
diff -X dontdiff -ur linux-2.6.14/net/core/link_watch.c linux-2.6.14-rfc2863/net/core/link_watch.c
--- linux-2.6.14/net/core/link_watch.c 2005-06-17 21:48:29.000000000 +0200
+++ linux-2.6.14-rfc2863/net/core/link_watch.c 2005-11-27 15:09:05.000000000 +0100
@@ -49,6 +49,34 @@
/* Avoid kmalloc() for most systems */
static struct lw_event singleevent;
+static inline unsigned char default_operstate(const struct net_device *dev) {
+ if (!netif_carrier_ok(dev))
+ return dev->features&NETIF_F_STACKED?IF_OPER_LOWERLAYERDOWN:IF_OPER_DOWN;
+ if (netif_dormant(dev)) return IF_OPER_DORMANT;
+ return IF_OPER_UP;
+}
+
+
+static void rfc2863_policy(struct net_device *dev) {
+ unsigned char operstate = default_operstate(dev);
+
+ if (operstate == dev->operstate) return;
+
+ switch(dev->link_mode) {
+ case IF_LINK_MODE_DORMANT:
+ if (operstate == IF_OPER_UP) operstate = IF_OPER_DORMANT;
+ break;
+ case IF_LINK_MODE_DEFAULT:
+ default:
+ break;
+ }
+
+ write_lock_bh(&dev_base_lock);
+ dev->operstate = operstate;
+ write_unlock_bh(&dev_base_lock);
+}
+
+
/* Must be called with the rtnl semaphore held */
void linkwatch_run_queue(void)
{
@@ -81,6 +109,7 @@
} else
dev_deactivate(dev);
+ rfc2863_policy(dev);
netdev_state_change(dev);
}
diff -X dontdiff -ur linux-2.6.14/net/core/net-sysfs.c linux-2.6.14-rfc2863/net/core/net-sysfs.c
--- linux-2.6.14/net/core/net-sysfs.c 2005-06-17 21:48:29.000000000 +0200
+++ linux-2.6.14-rfc2863/net/core/net-sysfs.c 2005-11-28 12:57:45.000000000 +0100
@@ -136,9 +136,34 @@
return -EINVAL;
}
+static const char *operstates[] = {
+ "unknown",
+ NULL, /* notpresent, currently unused */
+ "down",
+ "lowerlayerdown",
+ NULL, /* testing, currently unused */
+ "dormant",
+ "up"
+};
+
+static ssize_t show_operstate(struct class_device *dev, char *buf)
+{
+ const struct net_device *netdev = to_net_dev(dev);
+ unsigned char operstate;
+
+ read_lock(&dev_base_lock);
+ operstate = netdev->operstate;
+ if (!netif_running(netdev)) operstate = IF_OPER_DOWN;
+ read_unlock(&dev_base_lock);
+
+ if (operstate >= sizeof(operstates)) return -EINVAL; /* should not happen */
+ return sprintf(buf, "%s\n", operstates[operstate]);
+}
+
static CLASS_DEVICE_ATTR(address, S_IRUGO, show_address, NULL);
static CLASS_DEVICE_ATTR(broadcast, S_IRUGO, show_broadcast, NULL);
static CLASS_DEVICE_ATTR(carrier, S_IRUGO, show_carrier, NULL);
+static CLASS_DEVICE_ATTR(operstate, S_IRUGO, show_operstate, NULL);
/* read-write attributes */
NETDEVICE_SHOW(mtu, fmt_dec);
@@ -215,6 +240,7 @@
&class_device_attr_address,
&class_device_attr_broadcast,
&class_device_attr_carrier,
+ &class_device_attr_operstate,
NULL
};
diff -X dontdiff -ur linux-2.6.14/net/core/rtnetlink.c linux-2.6.14-rfc2863/net/core/rtnetlink.c
--- linux-2.6.14/net/core/rtnetlink.c 2005-11-02 11:08:12.000000000 +0100
+++ linux-2.6.14-rfc2863/net/core/rtnetlink.c 2005-11-28 13:33:31.000000000 +0100
@@ -178,6 +178,31 @@
}
+static void set_operstate(struct net_device *dev, unsigned char transition) {
+ unsigned char operstate = dev->operstate;
+ ASSERT_RTNL();
+
+ switch(transition) {
+ case IF_OPER_UP:
+ if (operstate == IF_OPER_DORMANT ||
+ operstate == IF_OPER_UNKNOWN)
+ operstate = IF_OPER_UP;
+ break;
+ case IF_OPER_DORMANT:
+ if (operstate == IF_OPER_UP ||
+ operstate == IF_OPER_UNKNOWN)
+ operstate = IF_OPER_DORMANT;
+ break;
+ }
+
+ if (dev->operstate != operstate) {
+ write_lock_bh(&dev_base_lock);
+ dev->operstate = operstate;
+ write_unlock_bh(&dev_base_lock);
+ netdev_state_change(dev);
+ }
+}
+
static int rtnetlink_fill_ifinfo(struct sk_buff *skb, struct net_device *dev,
int type, u32 pid, u32 seq, u32 change,
unsigned int flags)
@@ -208,6 +233,13 @@
}
if (1) {
+ u8 operstate = dev->operstate;
+ u8 link_mode = dev->link_mode;
+ RTA_PUT(skb, IFLA_OPERSTATE, sizeof(operstate), &operstate);
+ RTA_PUT(skb, IFLA_LINKMODE, sizeof(link_mode), &link_mode);
+ }
+
+ if (1) {
struct rtnl_link_ifmap map = {
.mem_start = dev->mem_start,
.mem_end = dev->mem_end,
@@ -398,6 +430,22 @@
dev->weight = *((u32 *) RTA_DATA(ida[IFLA_WEIGHT - 1]));
}
+ if (ida[IFLA_OPERSTATE - 1]) {
+ if (ida[IFLA_OPERSTATE - 1]->rta_len != RTA_LENGTH(sizeof(u8)))
+ goto out;
+
+ set_operstate(dev, *((u8 *) RTA_DATA(ida[IFLA_OPERSTATE - 1])));
+ }
+
+ if (ida[IFLA_LINKMODE - 1]) {
+ if (ida[IFLA_LINKMODE - 1]->rta_len != RTA_LENGTH(sizeof(u8)))
+ goto out;
+
+ write_lock_bh(&dev_base_lock);
+ dev->link_mode = *((u8 *) RTA_DATA(ida[IFLA_LINKMODE - 1]));
+ write_unlock_bh(&dev_base_lock);
+ }
+
if (ifm->ifi_index >= 0 && ida[IFLA_IFNAME - 1]) {
char ifname[IFNAMSIZ];
