Received-SPF: pass (google.com: domain of linux-kernel+bounces-211965-linux.lists.archive=gmail.com@vger.kernel.org designates 139.178.88.99 as permitted sender) client-ip=139.178.88.99;
From: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
To: linux-kernel@vger.kernel.org,
	netdev@vger.kernel.org
Cc: "David S. Miller" <davem@davemloft.net>,
	Daniel Bristot de Oliveira <bristot@kernel.org>,
	Boqun Feng <boqun.feng@gmail.com>,
	Daniel Borkmann <daniel@iogearbox.net>,
	Eric Dumazet <edumazet@google.com>,
	Frederic Weisbecker <frederic@kernel.org>,
	Ingo Molnar <mingo@redhat.com>,
	Jakub Kicinski <kuba@kernel.org>,
	Paolo Abeni <pabeni@redhat.com>,
	Peter Zijlstra <peterz@infradead.org>,
	Thomas Gleixner <tglx@linutronix.de>,
	Waiman Long <longman@redhat.com>,
	Will Deacon <will@kernel.org>,
	Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Subject: [PATCH v6 net-next 04/15] net: Use nested-BH locking for napi_alloc_cache.
Date: Wed, 12 Jun 2024 18:44:30 +0200
Message-ID: <20240612170303.3896084-5-bigeasy@linutronix.de>
In-Reply-To: <20240612170303.3896084-1-bigeasy@linutronix.de>
References: <20240612170303.3896084-1-bigeasy@linutronix.de>
Precedence: bulk
MIME-Version: 1.0
Content-Transfer-Encoding: quoted-printable

napi_alloc_cache is a per-CPU variable and relies on disabled BH for its
locking. Without per-CPU locking in local_bh_disable() on PREEMPT_RT
this data structure requires explicit locking.

Add a local_lock_t to the data structure and use local_lock_nested_bh()
for locking. This change adds only lockdep coverage and does not alter
the functional behaviour for !PREEMPT_RT.

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
---
 net/core/skbuff.c | 29 ++++++++++++++++++++++++-----
 1 file changed, 24 insertions(+), 5 deletions(-)

diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index 656b298255c5f..0c13cfa20658f 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -277,6 +277,7 @@ static void *page_frag_alloc_1k(struct page_frag_1k *nc=
, gfp_t gfp_mask)
 #endif
=20
 struct napi_alloc_cache {
+	local_lock_t bh_lock;
 	struct page_frag_cache page;
 	struct page_frag_1k page_small;
 	unsigned int skb_count;
@@ -284,7 +285,9 @@ struct napi_alloc_cache {
 };
=20
 static DEFINE_PER_CPU(struct page_frag_cache, netdev_alloc_cache);
-static DEFINE_PER_CPU(struct napi_alloc_cache, napi_alloc_cache);
+static DEFINE_PER_CPU(struct napi_alloc_cache, napi_alloc_cache) =3D {
+	.bh_lock =3D INIT_LOCAL_LOCK(bh_lock),
+};
=20
 /* Double check that napi_get_frags() allocates skbs with
  * skb->head being backed by slab, not a page fragment.
@@ -306,11 +309,16 @@ void napi_get_frags_check(struct napi_struct *napi)
 void *__napi_alloc_frag_align(unsigned int fragsz, unsigned int align_mask)
 {
 	struct napi_alloc_cache *nc =3D this_cpu_ptr(&napi_alloc_cache);
+	void *data;
=20
 	fragsz =3D SKB_DATA_ALIGN(fragsz);
=20
-	return __page_frag_alloc_align(&nc->page, fragsz, GFP_ATOMIC,
+	local_lock_nested_bh(&napi_alloc_cache.bh_lock);
+	data =3D __page_frag_alloc_align(&nc->page, fragsz, GFP_ATOMIC,
 				       align_mask);
+	local_unlock_nested_bh(&napi_alloc_cache.bh_lock);
+	return data;
+
 }
 EXPORT_SYMBOL(__napi_alloc_frag_align);
=20
@@ -338,16 +346,20 @@ static struct sk_buff *napi_skb_cache_get(void)
 	struct napi_alloc_cache *nc =3D this_cpu_ptr(&napi_alloc_cache);
 	struct sk_buff *skb;
=20
+	local_lock_nested_bh(&napi_alloc_cache.bh_lock);
 	if (unlikely(!nc->skb_count)) {
 		nc->skb_count =3D kmem_cache_alloc_bulk(net_hotdata.skbuff_cache,
 						      GFP_ATOMIC,
 						      NAPI_SKB_CACHE_BULK,
 						      nc->skb_cache);
-		if (unlikely(!nc->skb_count))
+		if (unlikely(!nc->skb_count)) {
+			local_unlock_nested_bh(&napi_alloc_cache.bh_lock);
 			return NULL;
+		}
 	}
=20
 	skb =3D nc->skb_cache[--nc->skb_count];
+	local_unlock_nested_bh(&napi_alloc_cache.bh_lock);
 	kasan_mempool_unpoison_object(skb, kmem_cache_size(net_hotdata.skbuff_cac=
he));
=20
 	return skb;
@@ -740,9 +752,13 @@ struct sk_buff *__netdev_alloc_skb(struct net_device *=
dev, unsigned int len,
 		pfmemalloc =3D nc->pfmemalloc;
 	} else {
 		local_bh_disable();
+		local_lock_nested_bh(&napi_alloc_cache.bh_lock);
+
 		nc =3D this_cpu_ptr(&napi_alloc_cache.page);
 		data =3D page_frag_alloc(nc, len, gfp_mask);
 		pfmemalloc =3D nc->pfmemalloc;
+
+		local_unlock_nested_bh(&napi_alloc_cache.bh_lock);
 		local_bh_enable();
 	}
=20
@@ -806,11 +822,11 @@ struct sk_buff *napi_alloc_skb(struct napi_struct *na=
pi, unsigned int len)
 		goto skb_success;
 	}
=20
-	nc =3D this_cpu_ptr(&napi_alloc_cache);
-
 	if (sk_memalloc_socks())
 		gfp_mask |=3D __GFP_MEMALLOC;
=20
+	local_lock_nested_bh(&napi_alloc_cache.bh_lock);
+	nc =3D this_cpu_ptr(&napi_alloc_cache);
 	if (NAPI_HAS_SMALL_PAGE_FRAG && len <=3D SKB_WITH_OVERHEAD(1024)) {
 		/* we are artificially inflating the allocation size, but
 		 * that is not as bad as it may look like, as:
@@ -832,6 +848,7 @@ struct sk_buff *napi_alloc_skb(struct napi_struct *napi=
, unsigned int len)
 		data =3D page_frag_alloc(&nc->page, len, gfp_mask);
 		pfmemalloc =3D nc->page.pfmemalloc;
 	}
+	local_unlock_nested_bh(&napi_alloc_cache.bh_lock);
=20
 	if (unlikely(!data))
 		return NULL;
@@ -1429,6 +1446,7 @@ static void napi_skb_cache_put(struct sk_buff *skb)
 	if (!kasan_mempool_poison_object(skb))
 		return;
=20
+	local_lock_nested_bh(&napi_alloc_cache.bh_lock);
 	nc->skb_cache[nc->skb_count++] =3D skb;
=20
 	if (unlikely(nc->skb_count =3D=3D NAPI_SKB_CACHE_SIZE)) {
@@ -1440,6 +1458,7 @@ static void napi_skb_cache_put(struct sk_buff *skb)
 				     nc->skb_cache + NAPI_SKB_CACHE_HALF);
 		nc->skb_count =3D NAPI_SKB_CACHE_HALF;
 	}
+	local_unlock_nested_bh(&napi_alloc_cache.bh_lock);
 }
=20
 void __napi_kfree_skb(struct sk_buff *skb, enum skb_drop_reason reason)
--=20
2.45.1