[PATCH] ipv6: prevent a possible race condition with lifetimes

valid_lft, prefered_lft and tstamp are always accessed under the lock
"lock" in other places. Reading these without taking the lock may result
in inconsistencies regarding the calculation of the valid and preferred
variables since decisions are taken on these fields for those variables.

Signed-off-by: Niels Dossche <redacted>
---
 net/ipv6/addrconf.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c
index 3f23da8c0b10..6c8ab3e6e6fe 100644
--- a/net/ipv6/addrconf.c
+++ b/net/ipv6/addrconf.c

@@ -4998,6 +4998,7 @@ static int inet6_fill_ifaddr(struct sk_buff *skb, struct inet6_ifaddr *ifa,
            nla_put_s32(skb, IFA_TARGET_NETNSID, args->netnsid))
                goto error;
 
+       spin_lock_bh(&ifa->lock);
        if (!((ifa->flags&IFA_F_PERMANENT) &&
              (ifa->prefered_lft == INFINITY_LIFE_TIME))) {
                preferred = ifa->prefered_lft;

@@ -5019,6 +5020,7 @@ static int inet6_fill_ifaddr(struct sk_buff *skb, struct inet6_ifaddr *ifa,
                preferred = INFINITY_LIFE_TIME;
                valid = INFINITY_LIFE_TIME;
        }
+       spin_unlock_bh(&ifa->lock);
 
        if (!ipv6_addr_any(&ifa->peer_addr)) {
                if (nla_put_in6_addr(skb, IFA_LOCAL, &ifa->addr) < 0 ||

-- 
2.35.1

On 2/20/22 10:54 AM, Niels Dossche wrote:

valid_lft, prefered_lft and tstamp are always accessed under the lock
"lock" in other places. Reading these without taking the lock may result
in inconsistencies regarding the calculation of the valid and preferred
variables since decisions are taken on these fields for those variables.

Signed-off-by: Niels Dossche <redacted>
---
 net/ipv6/addrconf.c | 2 ++
 1 file changed, 2 insertions(+)

Reviewed-by: David Ahern <dsahern@kernel.org>

On Sun, 20 Feb 2022 18:54:40 +0100 Niels Dossche wrote:

valid_lft, prefered_lft and tstamp are always accessed under the lock
"lock" in other places. Reading these without taking the lock may result
in inconsistencies regarding the calculation of the valid and preferred
variables since decisions are taken on these fields for those variables.

Signed-off-by: Niels Dossche <redacted>

Looks like your email client has replaced tabs with spaces, 
so the patch won't apply. Could you try resending with git send-email?
Please add Dave's review tag in the next version, and the subject
tag should be [PATCH net v2]. Thanks!

It appears that the mail server of the university is changing tabs to spaces.
I will resend it from my personal gmail address, since that does not seem to
change tabs to spaces. Sorry for the inconvenience. Thanks!

On 23/02/2022 01:43, Jakub Kicinski wrote:

On Sun, 20 Feb 2022 18:54:40 +0100 Niels Dossche wrote:

quoted

valid_lft, prefered_lft and tstamp are always accessed under the lock
"lock" in other places. Reading these without taking the lock may result
in inconsistencies regarding the calculation of the valid and preferred
variables since decisions are taken on these fields for those variables.

Signed-off-by: Niels Dossche <redacted>

Looks like your email client has replaced tabs with spaces, 
so the patch won't apply. Could you try resending with git send-email?
Please add Dave's review tag in the next version, and the subject
tag should be [PATCH net v2]. Thanks!