Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8;
 format=flowed
Content-Transfer-Encoding: 8bit
Date:   Sun, 17 Nov 2019 12:52:24 +0300
From:   Alexander Lobakin <alobakin@dlink.ru>
To:     David Miller <davem@davemloft.net>
Cc:     ecree@solarflare.com, jiri@mellanox.com, edumazet@google.com,
        idosch@mellanox.com, pabeni@redhat.com, petrm@mellanox.com,
        sd@queasysnail.net, f.fainelli@gmail.com,
        jaswinder.singh@linaro.org, manishc@marvell.com,
        GR-Linux-NIC-Dev@marvell.com, johannes.berg@intel.com,
        emmanuel.grumbach@intel.com, luciano.coelho@intel.com,
        linuxwifi@intel.com, kvalo@codeaurora.org, netdev@vger.kernel.org,
        linux-wireless@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v2 net-next] net: core: allow fast GRO for skbs with
 Ethernet header in head
In-Reply-To: <20191116.130101.268806870571558138.davem@davemloft.net>
References: <20191115091135.13487-1-alobakin@dlink.ru>
 <20191116.130101.268806870571558138.davem@davemloft.net>
User-Agent: Roundcube Webmail/1.4.0
Message-ID: <5622418d39ce3ebc3d526d3e16c8546b@dlink.ru>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

David Miller wrote 17.11.2019 00:01:

> From: Alexander Lobakin <alobakin@dlink.ru>
> Date: Fri, 15 Nov 2019 12:11:35 +0300
> 
>> Commit 78d3fd0b7de8 ("gro: Only use skb_gro_header for completely
>> non-linear packets") back in May'09 (v2.6.31-rc1) has changed the
>> original condition '!skb_headlen(skb)' to
>> 'skb->mac_header == skb->tail' in gro_reset_offset() saying: "Since
>> the drivers that need this optimisation all provide completely
>> non-linear packets" (note that this condition has become the current
>> 'skb_mac_header(skb) == skb_tail_pointer(skb)' later with commmit
>> ced14f6804a9 ("net: Correct comparisons and calculations using
>> skb->tail and skb-transport_header") without any functional changes).
>> 
>> For now, we have the following rough statistics for v5.4-rc7:
>> 1) napi_gro_frags: 14
>> 2) napi_gro_receive with skb->head containing (most of) payload: 83
>> 3) napi_gro_receive with skb->head containing all the headers: 20
>> 4) napi_gro_receive with skb->head containing only Ethernet header: 2
>> 
>> With the current condition, fast GRO with the usage of
>> NAPI_GRO_CB(skb)->frag0 is available only in the [1] case.
>> Packets pushed by [2] and [3] go through the 'slow' path, but
>> it's not a problem for them as they already contain all the needed
>> headers in skb->head, so pskb_may_pull() only moves skb->data.
>> 
>> The layout of skbs in the fourth [4] case at the moment of
>> dev_gro_receive() is identical to skbs that have come through [1],
>> as napi_frags_skb() pulls Ethernet header to skb->head. The only
>> difference is that the mentioned condition is always false for them,
>> because skb_put() and friends irreversibly alter the tail pointer.
>> They also go through the 'slow' path, but now every single
>> pskb_may_pull() in every single .gro_receive() will call the *really*
>> slow __pskb_pull_tail() to pull headers to head. This significantly
>> decreases the overall performance for no visible reasons.
>  ...
>> Signed-off-by: Alexander Lobakin <alobakin@dlink.ru>
> 
> Applied to net-next, thanks.

Thank you!

Regards,
ᚷ ᛖ ᚢ ᚦ ᚠ ᚱ