Re: [PATCH net-next v3] soreuseport: publish num_socks with acquire/release

[Eric Dumazet]

On Wed, Sep 2, 2026 at 9:18 AM Jinjie Ruan <ruanjinjie@xxxxxxxxxx> wrote:
>
> Replace the smp_wmb()/smp_rmb() barrier pair with
> smp_store_release()/smp_load_acquire() on reuse->num_socks.
> Writers publish socks[] updates via release before incrementing
> or decrementing the count; readers acquire the count before
> accessing socks[], ensuring they observe a consistent view.
>
> The detach path gains proper ordering between the socks[] write and
> the decrement, which was previously unordered.
>
> No functional change intended.
>
> Cc: Eric Dumazet <edumazet@xxxxxxxxxx>
> Cc: Kuniyuki Iwashima <kuniyu@xxxxxxxxxx>
> Cc: Paolo Abeni <pabeni@xxxxxxxxxx>
> Cc: Willem de Bruijn <willemb@xxxxxxxxxx>
> Cc: "David S. Miller" <davem@xxxxxxxxxxxxx>
> Cc: Jakub Kicinski <kuba@xxxxxxxxxx>
> Cc: Simon Horman <horms@xxxxxxxxxx>
> Assisted-by: DeepSeek:DeepSeek-V3
> Signed-off-by: Jinjie Ruan <ruanjinjie@xxxxxxxxxx>
> ---
> v3:
> - Split out from following patch set as Kuniyuki suggested.
> Link: https://lore.kernel.org/all/20260901024234.135119-1-ruanjinjie@xxxxxxxxxx/
> ---
>  net/core/sock_reuseport.c | 20 ++++++++------------
>  1 file changed, 8 insertions(+), 12 deletions(-)
>
> diff --git a/net/core/sock_reuseport.c b/net/core/sock_reuseport.c
> index 29948cb44b7d..6d3c511d1def 100644
> --- a/net/core/sock_reuseport.c
> +++ b/net/core/sock_reuseport.c
> @@ -125,9 +125,8 @@ static void __reuseport_add_sock(struct sock *sk,
>                                  struct sock_reuseport *reuse)
>  {
>         reuse->socks[reuse->num_socks] = sk;
> -       /* paired with smp_rmb() in reuseport_(select|migrate)_sock() */
> -       smp_wmb();
> -       reuse->num_socks++;
> +       /* paired with smp_load_acquire() in reuseport_(select|migrate)_sock() */
> +       smp_store_release(&reuse->num_socks, reuse->num_socks + 1);
>         reuseport_get_incoming_cpu(sk, reuse);
>  }
>
> @@ -140,7 +139,8 @@ static bool __reuseport_detach_sock(struct sock *sk,
>                 return false;
>
>         reuse->socks[i] = reuse->socks[reuse->num_socks - 1];

OK, but I think we should fix these problematic accesses to reuse->socks[X]
to avoid load/store tearing before your patch?

This might avoid some KCSAN / AI-review reports.

Please squash the following parts (shown before your changes) into
your v4 patch:

I think that this could target the net tree, with
Fixes: ef456144da8e ("soreuseport: define reuseport groups")

Please wait ~24 hours before sending a V4.

Thanks.

diff --git a/net/core/sock_reuseport.c b/net/core/sock_reuseport.c
index 29948cb44b7d1a8864ff680aaeb2b4582ede5ff4..8380936fae736337b902f70f0f5ebc1155a3bf37
100644
--- a/net/core/sock_reuseport.c
+++ b/net/core/sock_reuseport.c
@@ -139,7 +139,7 @@ static bool __reuseport_detach_sock(struct sock *sk,
        if (i == -1)
                return false;

-       reuse->socks[i] = reuse->socks[reuse->num_socks - 1];
+       WRITE_ONCE(reuse->socks[i], reuse->socks[reuse->num_socks - 1]);
        reuse->num_socks--;
        reuseport_put_incoming_cpu(sk, reuse);

@@ -521,7 +521,7 @@ static struct sock *run_bpf_filter(struct
sock_reuseport *reuse, u16 socks,
        if (index >= socks)
                return NULL;

-       return reuse->socks[index];
+       return READ_ONCE(reuse->socks[index]);
 }

 static struct sock *reuseport_select_sock_by_hash(struct sock_reuseport *reuse,
@@ -532,7 +532,7 @@ static struct sock
*reuseport_select_sock_by_hash(struct sock_reuseport *reuse,

        i = j = reciprocal_scale(hash, num_socks);
        do {
-               struct sock *sk = reuse->socks[i];
+               struct sock *sk = READ_ONCE(reuse->socks[i]);

                if (sk->sk_state != TCP_ESTABLISHED) {
                        /* Paired with WRITE_ONCE() in
__reuseport_(get|put)_incoming_cpu(). */