netdev - Re: [PATCH net-next 0/2] udp_tunnel: GRO optimizations

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]

Message-ID: <0dc78208-ad16-4d83-a315-ae2c8bd3bbb8@redhat.com>
Date: Thu, 6 Mar 2025 21:59:47 +0100
From: Paolo Abeni <pabeni@...hat.com>
To: Jakub Kicinski <kuba@...nel.org>
Cc: netdev@...r.kernel.org, Willem de Bruijn
 <willemdebruijn.kernel@...il.com>, "David S. Miller" <davem@...emloft.net>,
 Eric Dumazet <edumazet@...gle.com>, Simon Horman <horms@...nel.org>,
 David Ahern <dsahern@...nel.org>
Subject: Re: [PATCH net-next 0/2] udp_tunnel: GRO optimizations

On 3/6/25 7:50 PM, Jakub Kicinski wrote:
> On Thu,  6 Mar 2025 16:56:51 +0100 Paolo Abeni wrote:
>> The UDP tunnel GRO stage is source of measurable overhead for workload
>> based on UDP-encapsulated traffic: each incoming packets requires a full
>> UDP socket lookup and an indirect call.
>>
>> In the most common setups a single UDP tunnel device is used. In such
>> case we can optimize both the lookup and the indirect call.
>>
>> Patch 1 tracks per netns the active UDP tunnels and replaces the socket
>> lookup with a single destination port comparison when possible.
>>
>> Patch 2 tracks the different types of UDP tunnels and replaces the
>> indirect call with a static one when there is a single UDP tunnel type
>> active.
>>
>> I measure ~5% performance improvement in TCP over UDP tunnel stream
>> tests on top of this series.
> 
> Breaks the build with NET_UDP_TUNNEL=n (in contest) :(
> 
> net/ipv4/udp_offload.c: In function ‘udp_tunnel_gro_rcv’:
> net/ipv4/udp_offload.c:172:16: error: returning ‘struct sk_buff *’ from a function with incompatible return type ‘struct skbuff *’ [-Werror=incompatible-pointer-types]
>   172 |         return call_gro_receive_sk(udp_sk(sk)->gro_receive, sk, head, skb);
>       |                ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> net/ipv4/udp_offload.c: In function ‘udp_gro_receive’:
> net/ipv4/udp_offload.c:786:12: error: assignment to ‘struct sk_buff *’ from incompatible pointer type ‘struct skbuff *’ [-Werror=incompatible-pointer-types]
>   786 |         pp = udp_tunnel_gro_rcv(sk, head, skb);
>       |            ^
> In file included from ./include/linux/seqlock.h:19,
>                  from ./include/linux/dcache.h:11,
>                  from ./include/linux/fs.h:8,
>                  from ./include/linux/highmem.h:5,
>                  from ./include/linux/bvec.h:10,
>                  from ./include/linux/skbuff.h:17,
>                  from net/ipv4/udp_offload.c:9:
> net/ipv4/udp_offload.c: In function ‘udpv4_offload_init’:
> net/ipv4/udp_offload.c:936:21: error: ‘udp_tunnel_gro_type_lock’ undeclared (first use in this function); did you mean ‘udp_tunnel_gro_rcv’?
>   936 |         mutex_init(&udp_tunnel_gro_type_lock);
>       |                     ^~~~~~~~~~~~~~~~~~~~~~~~
> ./include/linux/mutex.h:64:23: note: in definition of macro ‘mutex_init’
>    64 |         __mutex_init((mutex), #mutex, &__key);                          \
>       |                       ^~~~~
> net/ipv4/udp_offload.c:936:21: note: each undeclared identifier is reported only once for each function it appears in
>   936 |         mutex_init(&udp_tunnel_gro_type_lock);
>       |                     ^~~~~~~~~~~~~~~~~~~~~~~~
> ./include/linux/mutex.h:64:23: note: in definition of macro ‘mutex_init’
>    64 |         __mutex_init((mutex), #mutex, &__key);                          \
>       |                       ^~~~~
> cc1: all warnings being treated as errors

whoops.. thanks will fix in v2!

/P