Skip to contentSkip to conversation

ipv6: avoid divide by zero in rt6_multipath_rebalance

PATCH Applied Cen Zhang (Microsoft) 3 participants 4 messages Aug 17 – Aug 24

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 ipv6: avoid divide by zero in rt6_multipath_rebalance

rt6_multipath_rebalance() calculates the total eligible nexthop weight in one pass and programs upper bounds in a second pass. Since RTM_NEWROUTE is RTNL-free, a concurrent ignore_routes_with_linkdown update can make the first pass return zero while the second sees an eligible nexthop, causing rt6_upper_bound_set() to divide by zero.

UBSAN: division-overflow in net/ipv6/route.c:4845:17
Oops: divide error: 0000 [#1] SMP KASAN NOPTI
  rt6_upper_bound_set() net/ipv6/route.c:4845
  rt6_multipath_rebalance()
  fib6_add_rt2node()
  ip6_route_multipath_add()
  inet6_rtm_newroute()

Skip upper-bound calculation when the first pass reports a zero total. This respects the lock-free performance considerations here and solves insecure scenarios.

Fixes bd11ff421d36 ("ipv6: Get rid of RTNL for SIOCDELRT and RTM_DELROUTE.")
Reported-by [email protected]
---
 net/ipv6/route.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
net/ipv6/route.c +1 -1
index 5968ce5ad150..68924cff9205 100644
@@ -4840,7 +4840,7 @@ static void rt6_upper_bound_set(struct fib6_info *rt, int *weight, int total)
 {
 	int upper_bound = -1;
 
-	if (!rt6_is_dead(rt)) {
+	if (total && !rt6_is_dead(rt)) {
 		*weight += rt->fib6_nh->fib_nh_weight;
 		upper_bound = DIV_ROUND_CLOSEST_ULL((u64) (*weight) << 31,
 						    total) - 1;
Signature · 1 line
2.52.0
maintainer Reviewed-by on v1 1/1
#1 lore raw
Quoted 19 lines · Cen Zhang (Microsoft)

rt6_multipath_rebalance() calculates the total eligible nexthop weight in one pass and programs upper bounds in a second pass. Since RTM_NEWROUTE is RTNL-free, a concurrent ignore_routes_with_linkdown update can make the first pass return zero while the second sees an eligible nexthop, causing rt6_upper_bound_set() to divide by zero.

UBSAN: division-overflow in net/ipv6/route.c:4845:17
Oops: divide error: 0000 [#1] SMP KASAN NOPTI
  rt6_upper_bound_set() net/ipv6/route.c:4845
  rt6_multipath_rebalance()
  fib6_add_rt2node()
  ip6_route_multipath_add()
  inet6_rtm_newroute()

Skip upper-bound calculation when the first pass reports a zero total. This respects the lock-free performance considerations here and solves insecure scenarios.

Quoted 5 lines
Fixes bd11ff421d36 ("ipv6: Get rid of RTNL for SIOCDELRT and RTM_DELROUTE.")
Reported-by [email protected]

FYI:

"Patch authors are expected to proactively look into the AI-generated reviews and handle such feedback as any other kind of review: either debate it or address it. In both cases a reply on the mailing list is expected."

https://docs.kernel.org/next/process/maintainer-netdev.html#review-timelines

I went over the feedback from Sashiko [1]. Most of it is correct, but irrelevant.

Changing the value of "ignore_routes_with_linkdown" on the fly never worked correctly and it is not interesting: It is expected that the user configures it during initialization, not after configuring routes.

When this sysctl is changed, the kernel does not iterate over all the nexthop groups and rebalances them. For example, if we have two nexthops in a group, one is down and now we set "ignore_routes_with_linkdown" to 1, the kernel can still pick the down nexthop until some event causes the group to be rebalanced.

Also note that this only affects legacy nexthop groups. Nexthop objects do not make use of "ignore_routes_with_linkdown".

[1] https://netdev-ai.bots.linux.dev/sashiko/#/patchset/20260817013237.2797-1-blbllhy%40gmail.com

#3 lore raw
In reply to Cen Zhang (Microsoft)
This patch was applied to netdev/net.git (main) by Jakub Kicinski <[email protected]>:

Hello:

This patch was applied to netdev/net.git (main)
by Jakub Kicinski <[email protected]>:

Quoted 8 lines · On Sun, 16 Aug 2026 21:32:37 -0400 you wrote:

rt6_multipath_rebalance() calculates the total eligible nexthop weight in one pass and programs upper bounds in a second pass. Since RTM_NEWROUTE is RTNL-free, a concurrent ignore_routes_with_linkdown update can make the first pass return zero while the second sees an eligible nexthop, causing rt6_upper_bound_set() to divide by zero.

[...]

Here is the summary with links:
- [net] ipv6: avoid divide by zero in rt6_multipath_rebalance

    https://git.kernel.org/netdev/net/c/d2c26c2911dd

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html