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;
Subject: Re: [Linuxarm] Re: [PATCH RFC 0/7] add socket to netdev page frag
 recycling support
To:     Eric Dumazet <edumazet@google.com>,
        Yunsheng Lin <linyunsheng@huawei.com>
Cc:     David Miller <davem@davemloft.net>,
        Jakub Kicinski <kuba@kernel.org>,
        Alexander Duyck <alexander.duyck@gmail.com>,
        Russell King <linux@armlinux.org.uk>,
        Marcin Wojtas <mw@semihalf.com>, linuxarm@openeuler.org,
        Yisen Zhuang <yisen.zhuang@huawei.com>,
        Salil Mehta <salil.mehta@huawei.com>,
        Thomas Petazzoni <thomas.petazzoni@bootlin.com>,
        Jesper Dangaard Brouer <hawk@kernel.org>,
        Ilias Apalodimas <ilias.apalodimas@linaro.org>,
        Alexei Starovoitov <ast@kernel.org>,
        Daniel Borkmann <daniel@iogearbox.net>,
        John Fastabend <john.fastabend@gmail.com>,
        Andrew Morton <akpm@linux-foundation.org>,
        Peter Zijlstra <peterz@infradead.org>,
        Will Deacon <will@kernel.org>,
        Matthew Wilcox <willy@infradead.org>,
        Vlastimil Babka <vbabka@suse.cz>,
        Fenghua Yu <fenghua.yu@intel.com>,
        Roman Gushchin <guro@fb.com>, Peter Xu <peterx@redhat.com>,
        "Tang, Feng" <feng.tang@intel.com>, Jason Gunthorpe <jgg@ziepe.ca>,
        mcroce@microsoft.com, Hugh Dickins <hughd@google.com>,
        Jonathan Lemon <jonathan.lemon@gmail.com>,
        Alexander Lobakin <alobakin@pm.me>,
        Willem de Bruijn <willemb@google.com>,
        wenxu <wenxu@ucloud.cn>, Cong Wang <cong.wang@bytedance.com>,
        Kevin Hao <haokexin@gmail.com>,
        Aleksandr Nogikh <nogikh@google.com>,
        Marco Elver <elver@google.com>, Yonghong Song <yhs@fb.com>,
        kpsingh@kernel.org, Andrii Nakryiko <andrii@kernel.org>,
        Martin KaFai Lau <kafai@fb.com>,
        Song Liu <songliubraving@fb.com>,
        netdev <netdev@vger.kernel.org>,
        LKML <linux-kernel@vger.kernel.org>, bpf <bpf@vger.kernel.org>,
        chenhao288@hisilicon.com,
        Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>,
        David Ahern <dsahern@kernel.org>, memxor@gmail.com,
        linux@rempel-privat.de, Antoine Tenart <atenart@kernel.org>,
        Wei Wang <weiwan@google.com>, Taehee Yoo <ap420073@gmail.com>,
        Arnd Bergmann <arnd@arndb.de>,
        Mat Martineau <mathew.j.martineau@linux.intel.com>,
        aahringo@redhat.com, ceggers@arri.de, yangbo.lu@nxp.com,
        Florian Westphal <fw@strlen.de>, xiangxia.m.yue@gmail.com,
        linmiaohe <linmiaohe@huawei.com>, Christoph Hellwig <hch@lst.de>
References: <1629257542-36145-1-git-send-email-linyunsheng@huawei.com>
 <CANn89iJDf9uzSdqLEBeTeGB1uAxvmruKfK5HbeZWp+Cdc+qggQ@mail.gmail.com>
 <2cf4b672-d7dc-db3d-ce90-15b4e91c4005@huawei.com>
 <4b2ad6d4-8e3f-fea9-766e-2e7330750f84@huawei.com>
 <CANn89iK0nMG3qq226aL-urrtPF5jBN6UQCV=ckTmAFqWgy5kiA@mail.gmail.com>
From:   David Ahern <dsahern@gmail.com>
Message-ID: <5fdc5223-7d67-fed7-f691-185dcb2e3d80@gmail.com>
Date:   Wed, 25 Aug 2021 09:29:06 -0700
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:78.0)
 Gecko/20100101 Thunderbird/78.13.0
MIME-Version: 1.0
In-Reply-To: <CANn89iK0nMG3qq226aL-urrtPF5jBN6UQCV=ckTmAFqWgy5kiA@mail.gmail.com>
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 8/23/21 8:04 AM, Eric Dumazet wrote:
>>
>>
>> It seems PAGE_ALLOC_COSTLY_ORDER is mostly related to pcp page, OOM, memory
>> compact and memory isolation, as the test system has a lot of memory installed
>> (about 500G, only 3-4G is used), so I used the below patch to test the max
>> possible performance improvement when making TCP frags twice bigger, and
>> the performance improvement went from about 30Gbit to 32Gbit for one thread
>> iperf tcp flow in IOMMU strict mode,
> 
> This is encouraging, and means we can do much better.
> 
> Even with SKB_FRAG_PAGE_ORDER  set to 4, typical skbs will need 3 mappings
> 
> 1) One for the headers (in skb->head)
> 2) Two page frags, because one TSO packet payload is not a nice power-of-two.

interesting observation. I have noticed 17 with the ZC API. That might
explain the less than expected performance bump with iommu strict mode.

> 
> The first issue can be addressed using a piece of coherent memory (128
> or 256 bytes per entry in TX ring).
> Copying the headers can avoid one IOMMU mapping, and improve IOTLB
> hits, because all
> slots of the TX ring buffer will use one single IOTLB slot.
> 
> The second issue can be solved by tweaking a bit
> skb_page_frag_refill() to accept an additional parameter
> so that the whole skb payload fits in a single order-4 page.
> 
>