Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Mon, 11 Jan 2021 18:29:44 +0000
To:     "David S. Miller" <davem@davemloft.net>,
        Jakub Kicinski <kuba@kernel.org>
From:   Alexander Lobakin <alobakin@pm.me>
Cc:     Eric Dumazet <edumazet@google.com>,
        Edward Cree <ecree@solarflare.com>,
        Jonathan Lemon <jonathan.lemon@gmail.com>,
        Willem de Bruijn <willemb@google.com>,
        Miaohe Lin <linmiaohe@huawei.com>,
        Alexander Lobakin <alobakin@pm.me>,
        Steffen Klassert <steffen.klassert@secunet.com>,
        Guillaume Nault <gnault@redhat.com>,
        Yadu Kishore <kyk.segfault@gmail.com>,
        Al Viro <viro@zeniv.linux.org.uk>, netdev@vger.kernel.org,
        linux-kernel@vger.kernel.org
Reply-To: Alexander Lobakin <alobakin@pm.me>
Subject: [PATCH net-next 5/5] skbuff: refill skb_cache early from deferred-to-consume entries
Message-ID: <20210111182801.12609-5-alobakin@pm.me>
In-Reply-To: <20210111182801.12609-1-alobakin@pm.me>
References: <20210111182655.12159-1-alobakin@pm.me> <20210111182801.12609-1-alobakin@pm.me>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

Instead of unconditional queueing of ready-to-consume skbuff_heads
to flush_skb_cache, feed skb_cache with them instead if it's not
full already.
This greatly reduces the frequency of kmem_cache_alloc_bulk() calls.

Signed-off-by: Alexander Lobakin <alobakin@pm.me>
---
 net/core/skbuff.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index 57a7307689f3..ba0d5611635e 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -904,6 +904,11 @@ static inline void _kfree_skb_defer(struct sk_buff *sk=
b)
 =09/* drop skb->head and call any destructors for packet */
 =09skb_release_all(skb);
=20
+=09if (nc->skb_count < NAPI_SKB_CACHE_SIZE) {
+=09=09nc->skb_cache[nc->skb_count++] =3D skb;
+=09=09return;
+=09}
+
 =09/* record skb to CPU local list */
 =09nc->flush_skb_cache[nc->flush_skb_count++] =3D skb;
=20
--=20
2.30.0