IPv6: Fix incorrect disable_ipv6 behavior
Subsystems (4)
PATCH 1/1 IPv6: Fix incorrect disable_ipv6 behavior
It doesn't look like the disable_ipv6 devconf setting works as described in ip-sysctl.txt. I think it was originally designed to only be set by addrconf_dad_failure(), but since it's mode 0644 it can be controlled through sysctl -w.
Here's an example:
# sysctl -w net.ipv6.conf.eth0.disable_ipv6=1 net.ipv6.conf.eth0.disable_ipv6 = 1 # sysctl -w net.ipv6.conf.eth0.accept_dad=2 net.ipv6.conf.eth0.accept_dad = 2
# ifdown eth0 # ifup eth0
Mar 5 11:23:45 titan kernel: eth0: duplicate address detected!
# ip -6 a show dev eth0
5: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qlen 1000
inet6 fe80::21f:29ff:fe5a:ef04/64 scope link tentative
valid_lft forever preferred_lft foreverThe address didn't actually fail DAD, but since disable_ipv6 was set, addrconf_dad_timer() called addrconf_dad_failure() before checking if it succeeded.
The following patch fixes the behavior. The other option is to make this value read-only, but disable_ipv6 can be useful.
---
Fix the behavior of allowing both sysctl and addrconf_dad_failure() to set the disable_ipv6 parameter without any bad side-effects. If DAD fails and accept_dad > 1, we will still set disable_ipv6=1, but then instead of allowing an RA to add an address then immediately fail DAD, we simply don't allow the address to be added in the first place. This also lets the user set this flag and disable all IPv6 addresses on the interface, or on the entire system.
--- Documentation/networking/ip-sysctl.txt | 4 +++- net/ipv6/addrconf.c | 21 ++++++++++++++------- 2 files changed, 17 insertions(+), 8 deletions(-)
Documentation/networking/ip-sysctl.txt +3 -1
@@ -1043,7 +1043,9 @@ max_addresses - INTEGER Default: 16 disable_ipv6 - BOOLEAN - Disable IPv6 operation. + Disable IPv6 operation. If accept_dad is set to 2, this value + will be dynamically set to TRUE if DAD fails for the link-local + address. Default: FALSE (enable IPv6 operation) accept_dad - INTEGER
net/ipv6/addrconf.c +14 -7
@@ -590,6 +590,7 @@ ipv6_add_addr(struct inet6_dev *idev, const struct in6_addr *addr, int pfxlen, {struct inet6_ifaddr *ifa =NULL ;struct rt6_info *rt; +struct net *net =dev_net (idev->dev);int hash;int err =0 ;int addr_type =ipv6_addr_type (addr);
@@ -606,6 +607,11 @@ ipv6_add_addr(struct inet6_dev *idev, const struct in6_addr *addr, int pfxlen, goto out2; } +if (idev->cnf.disable_ipv6 || net->ipv6.devconf_all->disable_ipv6) { + err = -EACCES ; + goto out2; + } +write_lock (&addrconf_hash_lock);/* Ignore adding duplicate addresses on an interface */
@@ -1433,6 +1439,11 @@ static void addrconf_dad_stop(struct inet6_ifaddr *ifp)void addrconf_dad_failure (struct inet6_ifaddr *ifp) {struct inet6_dev *idev = ifp->idev; + +if (net_ratelimit ()) +printk (KERN_INFO "%s: IPv6 duplicate address detected!\n" , + ifp->idev->dev->name); +if (idev->cnf.accept_dad >1 && !idev->cnf.disable_ipv6) {struct in6_addr addr;
@@ -1443,11 +1454,12 @@ void addrconf_dad_failure(struct inet6_ifaddr *ifp) ipv6_addr_equal(&ifp->addr, &addr)) {/* DAD failed for link-local based on MAC address */ idev->cnf.disable_ipv6 =1 ; + +printk (KERN_INFO "%s: IPv6 being disabled!\n" , + ifp->idev->dev->name); } } -if (net_ratelimit ()) -printk (KERN_INFO "%s: duplicate address detected!\n" , ifp->idev->dev->name);addrconf_dad_stop (ifp); }
@@ -2823,11 +2835,6 @@ static void addrconf_dad_timer(unsigned long data)read_unlock_bh (&idev->lock); goto out; } -if (idev->cnf.accept_dad >1 && idev->cnf.disable_ipv6) { -read_unlock_bh (&idev->lock); -addrconf_dad_failure (ifp); -return ; - }spin_lock_bh (&ifp->lock);if (ifp->probes ==0 ) {/*
Signature · 1 line
1.5.4.3
Quoted 8 lines · Brian Haley
Fix the behavior of allowing both sysctl and
addrconf_dad_failure()to set thedisable_ipv6parameter without any bad side-effects. If DAD fails andaccept_dad> 1, we will still set disable_ipv6=1, but then instead of allowing an RA to add an address then immediately fail DAD, we simply don't allow the address to be added in the first place. This also lets the user set this flag and disable all IPv6 addresses on the interface, or on the entire system.
So I haven't seen an Ack/Nack for this patch, and the patchwork queue has shrunk to just a handful in the past few hours :)
-Brian
David Miller
maintainerFrom: Brian Haley <[email protected]>
Date: Fri, 13 Mar 2009 20:52:40 -0400
Quoted 10 lines
Quoted 8 lines · Brian Haley
Fix the behavior of allowing both sysctl and
addrconf_dad_failure()to set thedisable_ipv6parameter without any bad side-effects. If DAD fails andaccept_dad> 1, we will still set disable_ipv6=1, but then instead of allowing an RA to add an address then immediately fail DAD, we simply don't allow the address to be added in the first place. This also lets the user set this flag and disable all IPv6 addresses on the interface, or on the entire system.
Quoted 2 lines
So I haven't seen an Ack/Nack for this patch, and the patchwork queue has shrunk to just a handful in the past few hours :)
It's near the top of my queue, I'm just debating what tree to stick it into.
Probably I'll put it into net-next-2.6
From: Brian Haley <[email protected]>
Date: Thu, 5 Mar 2009 17:28:40 -0500
Quoted 10 lines
Fix the behavior of allowing both sysctl and
addrconf_dad_failure()to set thedisable_ipv6parameter without any bad side-effects. If DAD fails andaccept_dad> 1, we will still set disable_ipv6=1, but then instead of allowing an RA to add an address then immediately fail DAD, we simply don't allow the address to be added in the first place. This also lets the user set this flag and disable all IPv6 addresses on the interface, or on the entire system.Signed-off-by Brian Haley <[email protected]>
Applied to net-next-2.6, thanks!