projects / pve-kernel-3.10.0.git / blame
| Commit | Line | Data |
|---|---|---|
| 8aa49427 DM |
1 | From d2615bf450694c1302d86b9cc8a8958edfe4c3a4 Mon Sep 17 00:00:00 2001 |
| 2 | From: Vlad Yasevich <vyasevic@redhat.com> | |
| 3 | Date: Wed, 20 Nov 2013 01:47:15 +0000 | |
| 4 | Subject: net: core: Always propagate flag changes to interfaces | |
| 5 | ||
| 6 | The following commit: | |
| 7 | b6c40d68ff6498b7f63ddf97cf0aa818d748dee7 | |
| 8 | net: only invoke dev->change_rx_flags when device is UP | |
| 9 | ||
| 10 | tried to fix a problem with VLAN devices and promiscuouse flag setting. | |
| 11 | The issue was that VLAN device was setting a flag on an interface that | |
| 12 | was down, thus resulting in bad promiscuity count. | |
| 13 | This commit blocked flag propagation to any device that is currently | |
| 14 | down. | |
| 15 | ||
| 16 | A later commit: | |
| 17 | deede2fabe24e00bd7e246eb81cd5767dc6fcfc7 | |
| 18 | vlan: Don't propagate flag changes on down interfaces | |
| 19 | ||
| 20 | fixed VLAN code to only propagate flags when the VLAN interface is up, | |
| 21 | thus fixing the same issue as above, only localized to VLAN. | |
| 22 | ||
| 23 | The problem we have now is that if we have create a complex stack | |
| 24 | involving multiple software devices like bridges, bonds, and vlans, | |
| 25 | then it is possible that the flags would not propagate properly to | |
| 26 | the physical devices. A simple examle of the scenario is the | |
| 27 | following: | |
| 28 | ||
| 29 | eth0----> bond0 ----> bridge0 ---> vlan50 | |
| 30 | ||
| 31 | If bond0 or eth0 happen to be down at the time bond0 is added to | |
| 32 | the bridge, then eth0 will never have promisc mode set which is | |
| 33 | currently required for operation as part of the bridge. As a | |
| 34 | result, packets with vlan50 will be dropped by the interface. | |
| 35 | ||
| 36 | The only 2 devices that implement the special flag handling are | |
| 37 | VLAN and DSA and they both have required code to prevent incorrect | |
| 38 | flag propagation. As a result we can remove the generic solution | |
| 39 | introduced in b6c40d68ff6498b7f63ddf97cf0aa818d748dee7 and leave | |
| 40 | it to the individual devices to decide whether they will block | |
| 41 | flag propagation or not. | |
| 42 | ||
| 43 | Reported-by: Stefan Priebe <s.priebe@profihost.ag> | |
| 44 | Suggested-by: Veaceslav Falico <vfalico@redhat.com> | |
| 45 | Signed-off-by: Vlad Yasevich <vyasevic@redhat.com> | |
| 46 | Signed-off-by: David S. Miller <davem@davemloft.net> | |
| 47 | --- | |
| 48 | (limited to 'net/core/dev.c') | |
| 49 | ||
| 50 | diff --git a/net/core/dev.c b/net/core/dev.c | |
| 51 | index 7e00a73..ba3b7ea 100644 | |
| 52 | --- a/net/core/dev.c | |
| 53 | +++ b/net/core/dev.c | |
| 54 | @@ -4996,7 +4996,7 @@ static void dev_change_rx_flags(struct net_device *dev, int flags) | |
| 55 | { | |
| 56 | const struct net_device_ops *ops = dev->netdev_ops; | |
| 57 | ||
| 58 | - if ((dev->flags & IFF_UP) && ops->ndo_change_rx_flags) | |
| 59 | + if (ops->ndo_change_rx_flags) | |
| 60 | ops->ndo_change_rx_flags(dev, flags); | |
| 61 | } | |
| 62 | ||
| 63 | -- | |
| 64 | cgit v0.9.2 |