On Fri, Jul 10, 2026 at 8:53 PM Allison Henderson [off-list ref] wrote:

On Fri, 2026-07-10 at 15:30 -0700, Xiang Mei wrote:

quoted

rds_tcp_laddr_check() looks up a scoped IPv6 interface with
dev_get_by_index_rcu(), drops the RCU read-side lock, and only then
passes the bare struct net_device * into ipv6_chk_addr().

dev_get_by_index_rcu() only keeps the device alive within the same RCU
read-side section. After rcu_read_unlock(), a concurrent RTM_DELLINK can
free the net_device; ipv6_chk_addr() then dereferences the stale pointer
in __ipv6_chk_addr_and_flags() (e.g. l3mdev_master_dev_rcu(dev)), reading
freed memory.

Keep the RCU read-side lock held across the ipv6_chk_addr() call instead
of dropping it right after the lookup, so the device cannot be freed
while it is in use.

  BUG: KASAN: slab-use-after-free in __ipv6_chk_addr_and_flags (... net/ipv6/addrconf.c:1998)
  Read of size 8 at addr ffff8880106ec000 by task exploit/153
  Call Trace:
   ...
   kasan_report (mm/kasan/report.c:595)
   __ipv6_chk_addr_and_flags (... net/ipv6/addrconf.c:1998)
   ipv6_chk_addr (net/ipv6/addrconf.c:2031 net/ipv6/addrconf.c:1972)
   rds_tcp_laddr_check (net/rds/tcp.c:370)
   rds_bind (net/rds/bind.c:248)
   __sys_bind (net/socket.c:1920)
   __x64_sys_bind (net/socket.c:1956)
   do_syscall_64 (arch/x86/entry/syscall_64.c:63)
   entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:121)

Thanks Xiang!  Just more thing to make sure your patches gets through. Add a change log here like this:

Changes since v1:
  Use rcu_read_locks instead of dev_hold/put
  Rebased on [PATCH net v2] rds: Fix inet6_addr_lst NULL dereference when IPv6 is disabled

Changes since v2:
  Add change log


Otherwise, it might get bounced if they happen to try your patch first.
With that fixed:
Reviewed-by: Allison Henderson <achender@kernel.org>

Thanks!
Allison

Thanks for the tips; I learned it!

v3 is ready:
https://lore.kernel.org/netdev/20260714212239.1511269-1-xmei5@asu.edu/T/#u (local)

Xiang

quoted

Thanks Xiang!  Just more thing to make sure your patches gets through. Add a change log here like this:

Changes since v1:
  Use rcu_read_locks instead of dev_hold/put
  Rebased on [PATCH net v2] rds: Fix inet6_addr_lst NULL dereference when IPv6 is disabled

Changes since v2:
  Add change log

Reviewed-by: Allison Henderson <achender@kernel.org>
Fixes: eee2fa6ab322 ("rds: Changing IP address internal representation to struct in6_addr")
Reported-by: Weiming Shi <redacted>
Assisted-by: Claude:claude-opus-4-8
Signed-off-by: Xiang Mei <redacted>
---
 net/rds/tcp.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/net/rds/tcp.c b/net/rds/tcp.c
index 955d92277d5a..30cfb0087f2c 100644
--- a/net/rds/tcp.c
+++ b/net/rds/tcp.c

@@ -355,23 +355,25 @@ int rds_tcp_laddr_check(struct net *net, const struct in6_addr *addr,
      /* If the scope_id is specified, check only those addresses
       * hosted on the specified interface.
       */
+     rcu_read_lock();
      if (scope_id != 0) {
-             rcu_read_lock();
              dev = dev_get_by_index_rcu(net, scope_id);
              /* scope_id is not valid... */
              if (!dev) {
                      rcu_read_unlock();
                      return -EADDRNOTAVAIL;
              }
-             rcu_read_unlock();
      }
 #if IS_ENABLED(CONFIG_IPV6)
      if (ipv6_mod_enabled()) {
              ret = ipv6_chk_addr(net, addr, dev, 0);
-             if (ret)
+             if (ret) {
+                     rcu_read_unlock();
                      return 0;
+             }
      }
 #endif
+     rcu_read_unlock();
      return -EADDRNOTAVAIL;
 }