[PATCHSET] IFF_RUNNING related fixes

[PATCHSET] IFF_RUNNING related fixes

[Thomas Graf]

Fixes various drivers still relying on IFF_RUNNING. I don't have
any of these cards so all patches completely untested.

Unresolved broken users of IFF_RUNNING are:

  drivers/wan/syncppp.c
     The stats update based on IFF_RUNNING has been broken
     for a long time. I'm not sure whether it should be
     changed to check netif_running or carrier_ok so I
     left it as-is for now.

  drivers/s390/net/*
     Various usages I'm not sure about the intention of the
     original code. Authors contacted but no reply so far.

Please do a

	bk pull bk://kernel.bkbits.net/tgraf/netdev-2.6-iff_running

This will update the following files:

 drivers/net/au1000_eth.c       |   10 ++--------
 drivers/net/bmac.c             |    2 --
 drivers/net/sk98lin/skge.c     |    8 +++-----
 drivers/net/tlan.c             |    4 ++--
 drivers/net/tokenring/ibmtr.c  |   11 +++--------
 drivers/net/wan/lmc/lmc_main.c |    8 ++------
 include/linux/if.h             |    2 +-
 7 files changed, 13 insertions(+), 32 deletions(-)

through these ChangeSets:

<tgraf@suug.ch> (05/03/28 1.2039)
   [NET]: Change comment to describe the real meaning of IFF_RUNNING
   
   Signed-off-by: Thomas Graf <tgraf@suug.ch>

<tgraf@suug.ch> (05/03/28 1.2038)
   [sk98lin]: Use netif_carrier_* instead of IFF_RUNNING and remove obsolete usage of IFF_RUNNING
   
   Signed-off-by: Thomas Graf <tgraf@suug.ch>

<tgraf@suug.ch> (05/03/28 1.2037)
   [ibmtr]: Use netif_carrier_* instead of IFF_RUNNING and remove obsolete usage of IFF_RUNNING
   
   Signed-off-by: Thomas Graf <tgraf@suug.ch>

<tgraf@suug.ch> (05/03/28 1.2036)
   [lmc]: Use netif_carrier_* instead of IFF_RUNNING
   
   Signed-off-by: Thomas Graf <tgraf@suug.ch>

<tgraf@suug.ch> (05/03/28 1.2035)
   [tlan]: Use netif_carrier_* intead of IFF_RUNNING
   
   Signed-off-by: Thomas Graf <tgraf@suug.ch>

<tgraf@suug.ch> (05/03/28 1.2034)
   [bmac]: Remove obsolete usage of IFF_RUNING
   
   Signed-off-by: Thomas Graf <tgraf@suug.ch>

<tgraf@suug.ch> (05/03/28 1.2033)
   [au1000]: Use netif_carrier_* instead of IFF_RUNNING
   
   Signed-off-by: Thomas Graf <tgraf@suug.ch>

[PATCH 1/7] [au1000]: Use netif_carrier_* instead of IFF_RUNNING

[Thomas Graf]

# This is a BitKeeper generated diff -Nru style patch.
#
# ChangeSet
#   2005/03/28 00:06:22+02:00 tgraf@suug.ch 
#   [au1000]: Use netif_carrier_* instead of IFF_RUNNING
#   
#   Signed-off-by: Thomas Graf <tgraf@suug.ch>
# 
# drivers/net/au1000_eth.c
#   2005/03/28 00:06:10+02:00 tgraf@suug.ch +2 -8
#   [au1000]: Use netif_carrier_* instead of IFF_RUNNING
# 
diff -Nru a/drivers/net/au1000_eth.c b/drivers/net/au1000_eth.c
--- a/drivers/net/au1000_eth.c	2005-03-28 01:08:16 +02:00
+++ b/drivers/net/au1000_eth.c	2005-03-28 01:08:16 +02:00
@@ -1681,10 +1681,6 @@
 		control |= MAC_FULL_DUPLEX;
 	}
 
-	/* fix for startup without cable */
-	if (!link) 
-		dev->flags &= ~IFF_RUNNING;
-
 	aup->mac->control = control;
 	aup->mac->vlan1_tag = 0x8100; /* activate vlan support */
 	au_sync();
@@ -1709,16 +1705,14 @@
 	if_port = dev->if_port;
 	if (aup->phy_ops->phy_status(dev, aup->phy_addr, &link, &speed) == 0) {
 		if (link) {
-			if (!(dev->flags & IFF_RUNNING)) {
+			if (!netif_carrier_ok) {
 				netif_carrier_on(dev);
-				dev->flags |= IFF_RUNNING;
 				printk(KERN_INFO "%s: link up\n", dev->name);
 			}
 		}
 		else {
-			if (dev->flags & IFF_RUNNING) {
+			if (netif_carrier_ok) {
 				netif_carrier_off(dev);
-				dev->flags &= ~IFF_RUNNING;
 				dev->if_port = 0;
 				printk(KERN_INFO "%s: link down\n", dev->name);
 			}

Re: [PATCH 1/7] [au1000]: Use netif_carrier_* instead of IFF_RUNNING

[Thomas Graf]

Correct patch checked-in this time.

# This is a BitKeeper generated diff -Nru style patch.
#
# ChangeSet
#   2005/03/28 02:43:27+02:00 tgraf@suug.ch 
#   [au1000]: Use netif_carrier_* instead of IFF_RUNNING
#   
#   Signed-off-by: Thomas Graf <tgraf@suug.ch>
# 
# drivers/net/au1000_eth.c
#   2005/03/28 02:43:16+02:00 tgraf@suug.ch +2 -8
#   [au1000]: Use netif_carrier_* instead of IFF_RUNNING
# 
diff -Nru a/drivers/net/au1000_eth.c b/drivers/net/au1000_eth.c
--- a/drivers/net/au1000_eth.c	2005-03-28 03:15:43 +02:00
+++ b/drivers/net/au1000_eth.c	2005-03-28 03:15:43 +02:00
@@ -1681,10 +1681,6 @@
 		control |= MAC_FULL_DUPLEX;
 	}
 
-	/* fix for startup without cable */
-	if (!link) 
-		dev->flags &= ~IFF_RUNNING;
-
 	aup->mac->control = control;
 	aup->mac->vlan1_tag = 0x8100; /* activate vlan support */
 	au_sync();
@@ -1709,16 +1705,14 @@
 	if_port = dev->if_port;
 	if (aup->phy_ops->phy_status(dev, aup->phy_addr, &link, &speed) == 0) {
 		if (link) {
-			if (!(dev->flags & IFF_RUNNING)) {
+			if (!netif_carrier_ok(dev)) {
 				netif_carrier_on(dev);
-				dev->flags |= IFF_RUNNING;
 				printk(KERN_INFO "%s: link up\n", dev->name);
 			}
 		}
 		else {
-			if (dev->flags & IFF_RUNNING) {
+			if (netif_carrier_ok(dev)) {
 				netif_carrier_off(dev);
-				dev->flags &= ~IFF_RUNNING;
 				dev->if_port = 0;
 				printk(KERN_INFO "%s: link down\n", dev->name);
 			}

Re: [PATCH 1/7] [au1000]: Use netif_carrier_* instead of IFF_RUNNING

[Jeff Garzik]

Applied patches 1-7 with two notes:

1) Include signed-off-by line with ALL patches in the future.

2) I'll let these patches hang out in netdev-2.6 (and thus -mm) for a 
little while.  If you can solicit testers, that would be most helpful.

	Jeff

[PATCH 2/7] [bmac]: Remove obsolete usage of IFF_RUNING

[Thomas Graf]

# This is a BitKeeper generated diff -Nru style patch.
#
# ChangeSet
#   2005/03/28 00:10:11+02:00 tgraf@suug.ch 
#   [bmac]: Remove obsolete usage of IFF_RUNING
#   
#   Signed-off-by: Thomas Graf <tgraf@suug.ch>
# 
# drivers/net/bmac.c
#   2005/03/28 00:09:27+02:00 tgraf@suug.ch +0 -2
#   [bmac]: Remove obsolete usage of IFF_RUNING
# 
diff -Nru a/drivers/net/bmac.c b/drivers/net/bmac.c
--- a/drivers/net/bmac.c	2005-03-28 01:08:09 +02:00
+++ b/drivers/net/bmac.c	2005-03-28 01:08:09 +02:00
@@ -1412,7 +1412,6 @@
 	bp->opened = 1;
 	bmac_reset_and_enable(dev);
 	enable_irq(dev->irq);
-	dev->flags |= IFF_RUNNING;
 	return 0;
 }
 
@@ -1425,7 +1424,6 @@
 	int i;
 
 	bp->sleeping = 1;
-	dev->flags &= ~(IFF_UP | IFF_RUNNING);
 
 	/* disable rx and tx */
 	config = bmread(dev, RXCFG);

[PATCH 3/7] [tlan]: Use netif_carrier_* intead of IFF_RUNNING

[Thomas Graf]

# This is a BitKeeper generated diff -Nru style patch.
#
# ChangeSet
#   2005/03/28 00:13:23+02:00 tgraf@suug.ch 
#   [tlan]: Use netif_carrier_* intead of IFF_RUNNING
#   
#   Signed-off-by: Thomas Graf <tgraf@suug.ch>
# 
# drivers/net/tlan.c
#   2005/03/28 00:13:12+02:00 tgraf@suug.ch +2 -2
#   [tlan]: Use netif_carrier_* intead of IFF_RUNNING
# 
diff -Nru a/drivers/net/tlan.c b/drivers/net/tlan.c
--- a/drivers/net/tlan.c	2005-03-28 01:08:02 +02:00
+++ b/drivers/net/tlan.c	2005-03-28 01:08:02 +02:00
@@ -2807,7 +2807,7 @@
  	       if (priv->link) {
 		      priv->link = 0;
 	              printk(KERN_DEBUG "TLAN: %s has lost link\n", dev->name);
-	              dev->flags &= ~IFF_RUNNING;
+		      netif_carrier_off(dev);
 		      TLan_SetTimer( dev, (2*HZ), TLAN_TIMER_LINK_BEAT );
 		      return;
 		}
@@ -2817,7 +2817,7 @@
         if ((phy_status & MII_GS_LINK) && !priv->link) {
  		priv->link = 1;
         	printk(KERN_DEBUG "TLAN: %s has reestablished link\n", dev->name);
-        	dev->flags |= IFF_RUNNING;
+		netif_carrier_on(dev);
         }
 
 	/* Setup a new monitor */

[PATCH 4/7] [lmc]: Use netif_carrier_* instead of IFF_RUNNING

[Thomas Graf]

# This is a BitKeeper generated diff -Nru style patch.
#
# ChangeSet
#   2005/03/28 00:15:07+02:00 tgraf@suug.ch 
#   [lmc]: Use netif_carrier_* instead of IFF_RUNNING
#   
#   Signed-off-by: Thomas Graf <tgraf@suug.ch>
# 
# drivers/net/wan/lmc/lmc_main.c
#   2005/03/28 00:14:55+02:00 tgraf@suug.ch +2 -6
#   [lmc]: Use netif_carrier_* instead of IFF_RUNNING
# 
diff -Nru a/drivers/net/wan/lmc/lmc_main.c b/drivers/net/wan/lmc/lmc_main.c
--- a/drivers/net/wan/lmc/lmc_main.c	2005-03-28 01:07:55 +02:00
+++ b/drivers/net/wan/lmc/lmc_main.c	2005-03-28 01:07:55 +02:00
@@ -723,7 +723,7 @@
         /* lmc_reset (sc); Why reset??? The link can go down ok */
 
         /* Inform the world that link has been lost */
-        dev->flags &= ~IFF_RUNNING;
+	netif_carrier_off(dev);
     }
 
     /*
@@ -736,7 +736,7 @@
          /* lmc_reset (sc); Again why reset??? */
 
          /* Inform the world that link protocol is back up. */
-         dev->flags |= IFF_RUNNING;
+	 netif_carrier_on(dev);
 
          /* Now we have to tell the syncppp that we had an outage
           * and that it should deal.  Calling sppp_reopen here
@@ -1168,8 +1168,6 @@
     sc->lmc_media->set_link_status (sc, 1);
     sc->lmc_media->set_status (sc, NULL);
 
-    //dev->flags |= IFF_RUNNING;
-    
     netif_wake_queue(dev);
 
     sc->lmc_txfull = 0;
@@ -1232,8 +1230,6 @@
     csr6 &= ~LMC_DEC_ST;		/* Turn off the Transmission bit */
     csr6 &= ~LMC_DEC_SR;		/* Turn off the Receive bit */
     LMC_CSR_WRITE (sc, csr_command, csr6);
-
-    dev->flags &= ~IFF_RUNNING;
 
     sc->stats.rx_missed_errors +=
         LMC_CSR_READ (sc, csr_missed_frames) & 0xffff;

[PATCH 5/7] [ibmtr]: Use netif_carrier_* instead of IFF_RUNNING and remove obsolete usage of IFF_RUNNING

[Thomas Graf]

# This is a BitKeeper generated diff -Nru style patch.
#
# ChangeSet
#   2005/03/28 00:16:54+02:00 tgraf@suug.ch 
#   [ibmtr]: Use netif_carrier_* instead of IFF_RUNNING and remove obsolete usage of IFF_RUNNING
#   
#   Signed-off-by: Thomas Graf <tgraf@suug.ch>
# 
# drivers/net/tokenring/ibmtr.c
#   2005/03/28 00:16:44+02:00 tgraf@suug.ch +3 -8
#   [ibmtr]: Use netif_carrier_* instead of IFF_RUNNING and remove obsolete usage of IFF_RUNNING
# 
diff -Nru a/drivers/net/tokenring/ibmtr.c b/drivers/net/tokenring/ibmtr.c
--- a/drivers/net/tokenring/ibmtr.c	2005-03-28 01:07:47 +02:00
+++ b/drivers/net/tokenring/ibmtr.c	2005-03-28 01:07:47 +02:00
@@ -888,11 +888,6 @@
 	ti->sap_status   = CLOSED; /* CLOSED or OPEN      */
 	ti->open_failure =     NO; /* NO     or YES       */
 	ti->open_mode    = MANUAL; /* MANUAL or AUTOMATIC */
-	/* 12/2000 not typical Linux, but we can use RUNNING to let us know when
-	the network has crapped out or cables are disconnected. Useful because
-	the IFF_UP flag stays up the whole time, until ifconfig tr0 down.
-	*/
-	dev->flags &= ~IFF_RUNNING;
 
 	ti->sram_phys &= ~1; /* to reverse what we do in tok_close */
 	/* init the spinlock */
@@ -1242,7 +1237,7 @@
 		ti->open_status = CLOSED;
 		ti->sap_status  = CLOSED;
 		ti->open_mode   = AUTOMATIC;
-		dev->flags &= ~IFF_RUNNING;
+		netif_carrier_off(dev);
 		netif_stop_queue(dev);
 		ti->open_action = RESTART;
 		outb(0, dev->base_addr + ADAPTRESET);
@@ -1323,7 +1318,7 @@
 				break;
 			}
 			netif_wake_queue(dev);
-			dev->flags |= IFF_RUNNING;/*BMS 12/2000*/
+			netif_carrier_on(dev);
 			break;
 		case DIR_INTERRUPT:
 		case DIR_MOD_OPEN_PARAMS:
@@ -1427,7 +1422,7 @@
 								ring_status);
 			if(ring_status& (REMOVE_RECV|AUTO_REMOVAL|LOBE_FAULT)){
 				netif_stop_queue(dev);
-				dev->flags &= ~IFF_RUNNING;/*not typical Linux*/
+				netif_carrier_off(dev);
 				DPRINTK("Remove received, or Auto-removal error"
 					", or Lobe fault\n");
 				DPRINTK("We'll try to reopen the closed adapter"

[PATCH 6/7] [sk98lin]: Use netif_carrier_* instead of IFF_RUNNING and remove obsolete usage of IFF_RUNNING

[Thomas Graf]

# This is a BitKeeper generated diff -Nru style patch.
#
# ChangeSet
#   2005/03/28 00:18:49+02:00 tgraf@suug.ch 
#   [sk98lin]: Use netif_carrier_* instead of IFF_RUNNING and remove obsolete usage of IFF_RUNNING
#   
#   Signed-off-by: Thomas Graf <tgraf@suug.ch>
# 
# drivers/net/sk98lin/skge.c
#   2005/03/28 00:18:37+02:00 tgraf@suug.ch +3 -5
#   [sk98lin]: Use netif_carrier_* instead of IFF_RUNNING and remove obsolete usage of IFF_RUNNING
# 
diff -Nru a/drivers/net/sk98lin/skge.c b/drivers/net/sk98lin/skge.c
--- a/drivers/net/sk98lin/skge.c	2005-03-28 01:07:40 +02:00
+++ b/drivers/net/sk98lin/skge.c	2005-03-28 01:07:40 +02:00
@@ -4212,7 +4212,7 @@
 			Flags);
 
 		SkGeStopPort(pAC, IoC, FromPort, SK_STOP_ALL, SK_HARD_RST);
-		pAC->dev[Param.Para32[0]]->flags &= ~IFF_RUNNING;
+		netif_carrier_off(pAC->dev[Param.Para32[0]]);
 		spin_unlock_irqrestore(
 			&pAC->TxPort[FromPort][TX_PRIO_LOW].TxDesRingLock,
 			Flags);
@@ -4355,7 +4355,7 @@
 		}
 
 		/* Inform the world that link protocol is up. */
-		pAC->dev[Param.Para32[0]]->flags |= IFF_RUNNING;
+		netif_carrier_on(pAC->dev[Param.Para32[0]]);
 
 		break;
 	case SK_DRV_NET_DOWN:	 /* SK_U32 Reason */
@@ -4368,7 +4368,7 @@
 		} else {
 			DoPrintInterfaceChange = SK_TRUE;
 		}
-		pAC->dev[Param.Para32[1]]->flags &= ~IFF_RUNNING;
+		netif_carrier_off(pAC->dev[Param.Para32[1]]);
 		break;
 	case SK_DRV_SWITCH_HARD: /* SK_U32 FromPortIdx SK_U32 ToPortIdx */
 		SK_DBG_MSG(NULL, SK_DBGMOD_DRV, SK_DBGCAT_DRV_EVENT,
@@ -4961,7 +4961,6 @@
 #ifdef CONFIG_NET_POLL_CONTROLLER
 	dev->poll_controller =	&SkGePollController;
 #endif
-	dev->flags &= 		~IFF_RUNNING;
 	SET_NETDEV_DEV(dev, &pdev->dev);
 	SET_ETHTOOL_OPS(dev, &SkGeEthtoolOps);
 
@@ -5035,7 +5034,6 @@
 		dev->set_mac_address    = &SkGeSetMacAddr;
 		dev->do_ioctl           = &SkGeIoctl;
 		dev->change_mtu         = &SkGeChangeMtu;
-		dev->flags             &= ~IFF_RUNNING;
 		SET_NETDEV_DEV(dev, &pdev->dev);
 		SET_ETHTOOL_OPS(dev, &SkGeEthtoolOps);
 

[PATCH 7/7] [NET]: Change comment to describe the real meaning of IFF_RUNNING

[Thomas Graf]

# This is a BitKeeper generated diff -Nru style patch.
#
# ChangeSet
#   2005/03/28 00:46:34+02:00 tgraf@suug.ch 
#   [NET]: Change comment to describe the real meaning of IFF_RUNNING
#   
#   Signed-off-by: Thomas Graf <tgraf@suug.ch>
# 
# include/linux/if.h
#   2005/03/28 00:46:23+02:00 tgraf@suug.ch +1 -1
#   [NET]: Change comment to describe the real meaning of IFF_RUNNING
# 
diff -Nru a/include/linux/if.h b/include/linux/if.h
--- a/include/linux/if.h	2005-03-28 01:07:33 +02:00
+++ b/include/linux/if.h	2005-03-28 01:07:33 +02:00
@@ -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		/* resources allocated		*/
+#define	IFF_RUNNING	0x40		/* interface running and carrier ok */
 #define	IFF_NOARP	0x80		/* no ARP protocol		*/
 #define	IFF_PROMISC	0x100		/* receive all packets		*/
 #define	IFF_ALLMULTI	0x200		/* receive all multicast packets*/

Re: [PATCHSET] IFF_RUNNING related fixes

[Thomas Graf]

* Thomas Graf <20050328000610.GN3086@postel.suug.ch> 2005-03-28 02:06
> Fixes various drivers still relying on IFF_RUNNING. I don't have
> any of these cards so all patches completely untested.

Oops, one of the changesets has a wrong patch checked in which
will not compile. Let me fix this up first.