From patchwork Mon Jan 11 18:29:22 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Alexander Lobakin X-Patchwork-Id: 12011383 X-Patchwork-Delegate: kuba@kernel.org Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-12.9 required=3.0 tests=BAYES_00,DKIM_SIGNED, DKIM_VALID,DKIM_VALID_AU,HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER, INCLUDES_PATCH,MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS, UNWANTED_LANGUAGE_BODY,URIBL_BLOCKED autolearn=unavailable autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 32A12C4332B for ; Mon, 11 Jan 2021 18:30:35 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id 04B102222D for ; Mon, 11 Jan 2021 18:30:34 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S2390548AbhAKSaM (ORCPT ); Mon, 11 Jan 2021 13:30:12 -0500 Received: from mail-40133.protonmail.ch ([185.70.40.133]:23265 "EHLO mail-40133.protonmail.ch" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S2390530AbhAKSaM (ORCPT ); Mon, 11 Jan 2021 13:30:12 -0500 Date: Mon, 11 Jan 2021 18:29:22 +0000 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=pm.me; s=protonmail; t=1610389769; bh=zIOKkbaTUZEnQc1ND01YiT6/uvlP61oysJn+nnvfaLw=; h=Date:To:From:Cc:Reply-To:Subject:In-Reply-To:References:From; b=XHMFEB4eGdt6GGKVfUxQkAFPCfAP/Ync4+3YHTB5iNOmp52136/LyuTJL4aHcLZEf +L8zKUPLopwF0Xx4VP5OyW9X5M557N7D/QhqzGdNOFCksnHd7Mf8K8lc0Sv6OK3t3u Dgy/mTOsXLiQrtKh82kPoudz4tKIVtfVNDy0714alvvKSTXEh4qCSu+im5OJagS+ZR 6GH724zqXQS2J09KHS3dPdv7/Z+Y8dFHSn2Tr860zUbn1NGqb3Frw+fnvXc6+v23Bg Zc3QKJMToB4fu5h8hKvFhQ3CQ5TSoBo5qq4asMcwyXCrZ/E8Zl3CjBh3Cxt4E4ylK+ Lx43tCLUNh3PA== To: "David S. Miller" , Jakub Kicinski From: Alexander Lobakin Cc: Eric Dumazet , Edward Cree , Jonathan Lemon , Willem de Bruijn , Miaohe Lin , Alexander Lobakin , Steffen Klassert , Guillaume Nault , Yadu Kishore , Al Viro , netdev@vger.kernel.org, linux-kernel@vger.kernel.org Reply-To: Alexander Lobakin Subject: [PATCH net-next 4/5] skbuff: allocate skbuff_heads by bulks instead of one by one Message-ID: <20210111182801.12609-4-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 Precedence: bulk List-ID: X-Mailing-List: netdev@vger.kernel.org X-Patchwork-Delegate: kuba@kernel.org Use the same napi_alloc_cache struct and the same approach as used for bulk-freeing skbuff_heads to allocate them for new skbs. The new skb_cache will store up to NAPI_SKB_CACHE_SIZE (currently 64, which equals to NAPI_POLL_WEIGHT to be capable to serve one polling cycle) and will be refilled by bulks in case of full depletion or after completing network softirqs. Signed-off-by: Alexander Lobakin --- net/core/skbuff.c | 20 +++++++++++++++++++- 1 file changed, 19 insertions(+), 1 deletion(-) diff --git a/net/core/skbuff.c b/net/core/skbuff.c index 0e8c597ff6ce..57a7307689f3 100644 --- a/net/core/skbuff.c +++ b/net/core/skbuff.c @@ -367,6 +367,8 @@ EXPORT_SYMBOL(build_skb_around); struct napi_alloc_cache { struct page_frag_cache page; + u32 skb_count; + void *skb_cache[NAPI_SKB_CACHE_SIZE]; u32 flush_skb_count; void *flush_skb_cache[NAPI_SKB_CACHE_SIZE]; }; @@ -490,7 +492,15 @@ static struct sk_buff *__napi_decache_skb(struct napi_alloc_cache *nc) if (nc->flush_skb_count) return nc->flush_skb_cache[--nc->flush_skb_count]; - return kmem_cache_alloc(skbuff_head_cache, GFP_ATOMIC); + if (unlikely(!nc->skb_count)) + nc->skb_count = kmem_cache_alloc_bulk(skbuff_head_cache, + GFP_ATOMIC, + NAPI_SKB_CACHE_SIZE, + nc->skb_cache); + if (unlikely(!nc->skb_count)) + return NULL; + + return nc->skb_cache[--nc->skb_count]; } /** @@ -870,6 +880,7 @@ void __consume_stateless_skb(struct sk_buff *skb) void __kfree_skb_flush(void) { struct napi_alloc_cache *nc = this_cpu_ptr(&napi_alloc_cache); + u32 num; /* flush flush_skb_cache if containing objects */ if (nc->flush_skb_count) { @@ -877,6 +888,13 @@ void __kfree_skb_flush(void) nc->flush_skb_cache); nc->flush_skb_count = 0; } + + num = NAPI_SKB_CACHE_SIZE - nc->skb_count; + if (num) + nc->skb_count += kmem_cache_alloc_bulk(skbuff_head_cache, + GFP_ATOMIC, num, + nc->skb_cache + + nc->skb_count); } static inline void _kfree_skb_defer(struct sk_buff *skb)