From a687d1d501f6c35aa5bebd9a0aae9545d46e4f65 Mon Sep 17 00:00:00 2001 From: Eric Dumazet <edumazet@google.com> Date: Tue, 8 Oct 2024 14:31:10 +0000 Subject: [PATCH] net: do not delay dst_entries_add() in dst_release() commit ac888d58869bb99753e7652be19a151df9ecb35d upstream. dst_entries_add() uses per-cpu data that might be freed at netns dismantle from ip6_route_net_exit() calling dst_entries_destroy() Before ip6_route_net_exit() can be called, we release all the dsts associated with this netns, via calls to dst_release(), which waits an rcu grace period before calling dst_destroy() dst_entries_add() use in dst_destroy() is racy, because dst_entries_destroy() could have been called already. Decrementing the number of dsts must happen sooner. Notes: 1) in CONFIG_XFRM case, dst_destroy() can call dst_release_immediate(child), this might also cause UAF if the child does not have DST_NOCOUNT set. IPSEC maintainers might take a look and see how to address this. 2) There is also discussion about removing this count of dst, which might happen in future kernels. Fixes: f88649721268 ("ipv4: fix dst race in sk_dst_get()") Closes: https://lore.kernel.org/lkml/CANn89iLCCGsP7SFn9HKpvnKu96Td4KD08xf7aGtiYgZnkjaL=w@mail.gmail.com/T/ Reported-by: Naresh Kamboju <naresh.kamboju@linaro.org> Tested-by: Linux Kernel Functional Testing <lkft@linaro.org> Tested-by: Naresh Kamboju <naresh.kamboju@linaro.org> Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Xin Long <lucien.xin@gmail.com> Cc: Steffen Klassert <steffen.klassert@secunet.com> Reviewed-by: Xin Long <lucien.xin@gmail.com> Link: https://patch.msgid.link/20241008143110.1064899-1-edumazet@google.com Signed-off-by: Paolo Abeni <pabeni@redhat.com> [ resolved conflict due to bc9d3a9f2afc ("net: dst: Switch to rcuref_t reference counting") is not in the tree ] Signed-off-by: Abdelkareem Abdelsaamad <kareemem@amazon.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> --- net/core/dst.c | 17 ++++++++++++----- 1 file changed, 12 insertions(+), 5 deletions(-) diff --git a/net/core/dst.c b/net/core/dst.c index fb3bcba87..fe7158b8a 100755 --- a/net/core/dst.c +++ b/net/core/dst.c @@ -113,9 +113,6 @@ struct dst_entry *dst_destroy(struct dst_entry * dst) child = xdst->child; } #endif - if (!(dst->flags & DST_NOCOUNT)) - dst_entries_add(dst->ops, -1); - if (dst->ops->destroy) dst->ops->destroy(dst); if (dst->dev) @@ -166,6 +163,12 @@ void dst_dev_put(struct dst_entry *dst) } EXPORT_SYMBOL(dst_dev_put); +static void dst_count_dec(struct dst_entry *dst) +{ + if (!(dst->flags & DST_NOCOUNT)) + dst_entries_add(dst->ops, -1); +} + void dst_release(struct dst_entry *dst) { if (dst) { @@ -175,8 +178,10 @@ void dst_release(struct dst_entry *dst) if (WARN_ONCE(newrefcnt < 0, "dst_release underflow")) net_warn_ratelimited("%s: dst:%p refcnt:%d\n", __func__, dst, newrefcnt); - if (!newrefcnt) + if (!newrefcnt){ + dst_count_dec(dst); call_rcu(&dst->rcu_head, dst_destroy_rcu); + } } } EXPORT_SYMBOL(dst_release); @@ -190,8 +195,10 @@ void dst_release_immediate(struct dst_entry *dst) if (WARN_ONCE(newrefcnt < 0, "dst_release_immediate underflow")) net_warn_ratelimited("%s: dst:%p refcnt:%d\n", __func__, dst, newrefcnt); - if (!newrefcnt) + if (!newrefcnt){ + dst_count_dec(dst); dst_destroy(dst); + } } } EXPORT_SYMBOL(dst_release_immediate);