Skip to contentSkip to conversation

net/ipv6: do not copy dst flags on rt init

PATCH In review Peter Oskolkov 3 participants 3 messages Sep 17, 2018 – Sep 18, 2018

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

PATCH v2 1/1 net/ipv6: do not copy dst flags on rt init

DST_NOCOUNT in dst_entry::flags tracks whether the entry counts toward route cache size (net->ipv6.sysctl.ip6_rt_max_size).

If the flag is NOT set, dst_ops::pcpuc_entries counter is incremented in dist_init() and decremented in dst_destroy().

This flag is tied to allocation/deallocation of dst_entry and should not be copied from another dst/route. Otherwise it can happen that dst_ops::pcpuc_entries counter grows until no new routes can be allocated because the counter reached ip6_rt_max_size due to DST_NOCOUNT not set and thus no counter decrements on gc-ed routes.

Fixes 3b6761d18bc1 ("net/ipv6: Move dst flags to booleans in fib entries")
Cc David Ahern <[email protected]>
---
 net/ipv6/route.c | 2 --
 1 file changed, 2 deletions(-)
net/ipv6/route.c +0 -2
index 3eed045c65a5..480a79f47c52 100644
@@ -946,8 +946,6 @@ static void ip6_rt_init_dst_reject(struct rt6_info *rt, struct fib6_info *ort)
 
 static void ip6_rt_init_dst(struct rt6_info *rt, struct fib6_info *ort)
 {
-	rt->dst.flags |= fib6_info_dst_flags(ort);
-
 	if (ort->fib6_flags & RTF_REJECT) {
 		ip6_rt_init_dst_reject(rt, ort);
 		return;
Signature · 1 line
2.19.0.397.gdd90340f6a-goog
Reviewed-by on v2 1/1
#1 lore raw
Quoted 26 lines · Peter Oskolkov

DST_NOCOUNT in dst_entry::flags tracks whether the entry counts toward route cache size (net->ipv6.sysctl.ip6_rt_max_size).

If the flag is NOT set, dst_ops::pcpuc_entries counter is incremented in dist_init() and decremented in dst_destroy().

This flag is tied to allocation/deallocation of dst_entry and should not be copied from another dst/route. Otherwise it can happen that dst_ops::pcpuc_entries counter grows until no new routes can be allocated because the counter reached ip6_rt_max_size due to DST_NOCOUNT not set and thus no counter decrements on gc-ed routes.

Fixes 3b6761d18bc1 ("net/ipv6: Move dst flags to booleans in fib entries")
Cc David Ahern <[email protected]>
---
 net/ipv6/route.c | 2 --
 1 file changed, 2 deletions(-)
net/ipv6/route.c +0 -0
index 3eed045c65a5..480a79f47c52 100644

net/ipv6/route.c static void ip6_rt_init_dst_reject(struct rt6_info *rt, struct fib6_info *ort) · line 951

Quoted 8 lines
net/ipv6/route.c +0 -2
@@ -946,8 +946,6 @@ static void ip6_rt_init_dst_reject(struct rt6_info *rt, struct fib6_info *ort)
 
 static void ip6_rt_init_dst(struct rt6_info *rt, struct fib6_info *ort)
 {
-	rt->dst.flags |= fib6_info_dst_flags(ort);
-
 	if (ort->fib6_flags & RTF_REJECT) {
 		ip6_rt_init_dst_reject(rt, ort);
 		return;

Thanks for the patch.

maintainer
#2 lore raw
In reply to Peter Oskolkov

From: Peter Oskolkov <[email protected]>
Date: Mon, 17 Sep 2018 10:20:53 -0700

Quoted 12 lines

DST_NOCOUNT in dst_entry::flags tracks whether the entry counts toward route cache size (net->ipv6.sysctl.ip6_rt_max_size).

If the flag is NOT set, dst_ops::pcpuc_entries counter is incremented in dist_init() and decremented in dst_destroy().

This flag is tied to allocation/deallocation of dst_entry and should not be copied from another dst/route. Otherwise it can happen that dst_ops::pcpuc_entries counter grows until no new routes can be allocated because the counter reached ip6_rt_max_size due to DST_NOCOUNT not set and thus no counter decrements on gc-ed routes.

Quoted 4 lines
Fixes 3b6761d18bc1 ("net/ipv6: Move dst flags to booleans in fib entries")
Cc David Ahern <[email protected]>

Applied and queued up for -stable, thank you.