netdev - Re: [PATCH v2 net-next v2 5/5] net: add heuristic for enabling TCP fraglist GRO

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] [day] [month] [year] [list]

Message-ID: <554bd11b-376e-4d68-ad79-01e5202deaf8@nbd.name>
Date: Thu, 25 Apr 2024 17:29:04 +0200
From: Felix Fietkau <nbd@....name>
To: Eric Dumazet <edumazet@...gle.com>
Cc: netdev@...r.kernel.org, "David S. Miller" <davem@...emloft.net>,
 David Ahern <dsahern@...nel.org>, Jakub Kicinski <kuba@...nel.org>,
 Paolo Abeni <pabeni@...hat.com>, willemdebruijn.kernel@...il.com,
 linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 net-next v2 5/5] net: add heuristic for enabling TCP
 fraglist GRO

On 25.04.24 17:26, Eric Dumazet wrote:
> On Thu, Apr 25, 2024 at 5:04 PM Felix Fietkau <nbd@....name> wrote:
>>
>> When forwarding TCP after GRO, software segmentation is very expensive,
>> especially when the checksum needs to be recalculated.
>> One case where that's currently unavoidable is when routing packets over
>> PPPoE. Performance improves significantly when using fraglist GRO
>> implemented in the same way as for UDP.
>>
>> When NETIF_F_GRO_FRAGLIST is enabled, perform a lookup for an established
>> socket in the same netns as the receiving device. While this may not
>> cover all relevant use cases in multi-netns configurations, it should be
>> good enough for most configurations that need this.
>>
>> Here's a measurement of running 2 TCP streams through a MediaTek MT7622
>> device (2-core Cortex-A53), which runs NAT with flow offload enabled from
>> one ethernet port to PPPoE on another ethernet port + cake qdisc set to
>> 1Gbps.
>>
>> rx-gro-list off: 630 Mbit/s, CPU 35% idle
>> rx-gro-list on:  770 Mbit/s, CPU 40% idle
>>
>> Signe-off-by: Felix Fietkau <nbd@....name>
>> ---
>>  net/ipv4/tcp_offload.c   | 48 +++++++++++++++++++++++++++++++++++++
>>  net/ipv6/tcpv6_offload.c | 51 ++++++++++++++++++++++++++++++++++++++++
>>  2 files changed, 99 insertions(+)
>>
>> diff --git a/net/ipv4/tcp_offload.c b/net/ipv4/tcp_offload.c
>> index 520fd425ab19..3bb96a110402 100644
>> --- a/net/ipv4/tcp_offload.c
>> +++ b/net/ipv4/tcp_offload.c
>> @@ -405,6 +405,52 @@ void tcp_gro_complete(struct sk_buff *skb)
>>  }
>>  EXPORT_SYMBOL(tcp_gro_complete);
>>
>> +static void tcp4_check_fraglist_gro(struct list_head *head, struct sk_buff *skb)
>> +{
>> +       const struct iphdr *iph = skb_gro_network_header(skb);
> 
> I do not think loading iph before all skb_gro_header() and
> skb_gro_header_slow() calls is wise.
> 
> pskb_may_pull() can re-allocate skb->head

Will fix.

>> +       struct net *net = dev_net(skb->dev);
>> +       unsigned int off, hlen, thlen;
>> +       struct sk_buff *p;
>> +       struct tcphdr *th;
>> +       struct sock *sk;
>> +       int iif, sdif;
>> +
>> +       if (!(skb->dev->features & NETIF_F_GRO_FRAGLIST))
>> +               return;
>> +
>> +       off = skb_gro_offset(skb);
>> +       hlen = off + sizeof(*th);
>> +       th = skb_gro_header(skb, hlen, off);
>> +       if (unlikely(!th))
>> +               return;
>> +
>> +       thlen = th->doff * 4;
>> +       if (thlen < sizeof(*th))
>> +               return;
>> +
>> +       hlen = off + thlen;
>> +       if (!skb_gro_may_pull(skb, hlen)) {
>> +               th = skb_gro_header_slow(skb, hlen, off);
>> +               if (unlikely(!th))
>> +                       return;
>> +       }
>> +
>> +       p = tcp_gro_lookup(head, th);
>> +       if (p) {
>> +               NAPI_GRO_CB(skb)->is_flist = NAPI_GRO_CB(p)->is_flist;
>> +               return;
>> +       }
>> +
>> +       inet_get_iif_sdif(skb, &iif, &sdif);
>> +       sk = __inet_lookup_established(net, net->ipv4.tcp_death_row.hashinfo,
>> +                                      iph->saddr, th->source,
>> +                                      iph->daddr, ntohs(th->dest),
>> +                                      iif, sdif);
>> +       NAPI_GRO_CB(skb)->is_flist = !sk;
>> +       if (sk)
>> +               sock_put(sk);
>> +}
>> +
>>  INDIRECT_CALLABLE_SCOPE
>>  struct sk_buff *tcp4_gro_receive(struct list_head *head, struct sk_buff *skb)
>>  {
>> @@ -416,6 +462,8 @@ struct sk_buff *tcp4_gro_receive(struct list_head *head, struct sk_buff *skb)
>>                 return NULL;
>>         }
>>
> 
> I would probably pull the whole TCP header here, before calling
> tcp4_check_fraglist_gro(head, skb)
> and no longer do this twice from tcp4_check_fraglist_gro() and tcp_gro_receive()
> 
> Perhaps define a new inline helper, that will be called from
> tcp4_gro_receive() and tcp6_gro_receive(),
> and not anymore from  tcp_gro_receive()
> 
> static inline struct tcphdr *tcp_gro_pull_header(...)
> {
>       ....
>         off = skb_gro_offset(skb);
>         hlen = off + sizeof(*th);
>         th = skb_gro_header(skb, hlen, off);
>         if (unlikely(!th))
>                 return NULL;
> 
>         thlen = th->doff * 4;
>         if (thlen < sizeof(*th))
>                 return NULL;
> 
>         hlen = off + thlen;
>         if (!skb_gro_may_pull(skb, hlen))
>                 th = skb_gro_header_slow(skb, hlen, off);
> 
>        return th;
> }

Makes sense

Thanks,

- Felix