netdev - [v3 net-next 09/10] skbuff: reuse NAPI skb cache on allocation path (__alloc

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Tue, 09 Feb 2021 20:49:16 +0000
From:   Alexander Lobakin <alobakin@...me>
To:     "David S. Miller" <davem@...emloft.net>,
        Jakub Kicinski <kuba@...nel.org>
Cc:     Jonathan Lemon <jonathan.lemon@...il.com>,
        Eric Dumazet <edumazet@...gle.com>,
        Dmitry Vyukov <dvyukov@...gle.com>,
        Willem de Bruijn <willemb@...gle.com>,
        Alexander Lobakin <alobakin@...me>,
        Randy Dunlap <rdunlap@...radead.org>,
        Kevin Hao <haokexin@...il.com>,
        Pablo Neira Ayuso <pablo@...filter.org>,
        Jakub Sitnicki <jakub@...udflare.com>,
        Marco Elver <elver@...gle.com>,
        Dexuan Cui <decui@...rosoft.com>,
        Paolo Abeni <pabeni@...hat.com>,
        Alexei Starovoitov <ast@...nel.org>,
        Daniel Borkmann <daniel@...earbox.net>,
        Andrii Nakryiko <andriin@...com>,
        Taehee Yoo <ap420073@...il.com>,
        Cong Wang <xiyou.wangcong@...il.com>,
        Björn Töpel <bjorn@...nel.org>,
        Miaohe Lin <linmiaohe@...wei.com>,
        Guillaume Nault <gnault@...hat.com>,
        Yonghong Song <yhs@...com>, zhudi <zhudi21@...wei.com>,
        Michal Kubecek <mkubecek@...e.cz>,
        Marcelo Ricardo Leitner <marcelo.leitner@...il.com>,
        Dmitry Safonov <0x7f454c46@...il.com>,
        Yang Yingliang <yangyingliang@...wei.com>,
        linux-kernel@...r.kernel.org, netdev@...r.kernel.org
Subject: [v3 net-next 09/10] skbuff: reuse NAPI skb cache on allocation path (__alloc_skb())

Try to use the same technique for obtaining skbuff_head from NAPI
cache in {,__}alloc_skb(). Two points here:
 - __alloc_skb() can be used for allocating clones or allocating skbs
   for distant nodes. Try to grab head from the cache only for
   non-clones and for local nodes;
 - can be called from any context, so napi_safe == false.

Signed-off-by: Alexander Lobakin <alobakin@...me>
---
 net/core/skbuff.c | 12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)

diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index 8747566a8136..8850086f8605 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -354,15 +354,19 @@ struct sk_buff *__alloc_skb(unsigned int size, gfp_t gfp_mask,
 	struct sk_buff *skb;
 	u8 *data;
 	bool pfmemalloc;
+	bool clone;
 
-	cache = (flags & SKB_ALLOC_FCLONE)
-		? skbuff_fclone_cache : skbuff_head_cache;
+	clone = !!(flags & SKB_ALLOC_FCLONE);
+	cache = clone ? skbuff_fclone_cache : skbuff_head_cache;
 
 	if (sk_memalloc_socks() && (flags & SKB_ALLOC_RX))
 		gfp_mask |= __GFP_MEMALLOC;
 
 	/* Get the HEAD */
-	skb = kmem_cache_alloc_node(cache, gfp_mask & ~__GFP_DMA, node);
+	if (clone || unlikely(node != NUMA_NO_NODE && node != numa_mem_id()))
+		skb = kmem_cache_alloc_node(cache, gfp_mask & ~GFP_DMA, node);
+	else
+		skb = napi_skb_cache_get(false);
 	if (unlikely(!skb))
 		return NULL;
 	prefetchw(skb);
@@ -393,7 +397,7 @@ struct sk_buff *__alloc_skb(unsigned int size, gfp_t gfp_mask,
 	__build_skb_around(skb, data, 0);
 	skb->pfmemalloc = pfmemalloc;
 
-	if (flags & SKB_ALLOC_FCLONE) {
+	if (clone) {
 		struct sk_buff_fclones *fclones;
 
 		fclones = container_of(skb, struct sk_buff_fclones, skb1);
-- 
2.30.0