Received-SPF: pass (google.com: domain of linux-kernel+bounces-157076-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:40f1:3f00::1 as permitted sender) client-ip=2604:1380:40f1:3f00::1;
Date: Wed, 24 Apr 2024 10:30:23 -0400
From: Willem de Bruijn <willemdebruijn.kernel@gmail.com>
To: Felix Fietkau <nbd@nbd.name>, 
 Willem de Bruijn <willemdebruijn.kernel@gmail.com>, 
 Paolo Abeni <pabeni@redhat.com>, 
 Eric Dumazet <edumazet@google.com>
Cc: netdev@vger.kernel.org, 
 "David S. Miller" <davem@davemloft.net>, 
 Jakub Kicinski <kuba@kernel.org>, 
 David Ahern <dsahern@kernel.org>, 
 linux-kernel@vger.kernel.org
Message-ID: <6629177f6b4c7_1a76072949b@willemb.c.googlers.com.notmuch>
In-Reply-To: <eb0e1eb6-306c-4cf7-922d-203db09d8d29@nbd.name>
References: <20240423094117.93206-1-nbd@nbd.name>
 <CANn89i+6xRe4V6aDmD-9EM0uD7A87f6rzg3S7Xq6-NaB_Mb4nw@mail.gmail.com>
 <63abfa26-d990-46c3-8982-3eaf7b8f8ee5@nbd.name>
 <CANn89iJZvoKVB+AK1_44gki2pHyigyMLXFkyevSQpH3iDbnCvw@mail.gmail.com>
 <7476374f-cf0c-45d0-8100-1b2cd2f290d5@nbd.name>
 <CANn89iLddm704LHPDnnoF2RbCfvrivAz0e6HTeiBARmvzoUBjA@mail.gmail.com>
 <ebe85dca-e0e9-4c55-a15d-20d340f66848@nbd.name>
 <97f10c8b5b615eac8f65d67ef10928d97b6b760d.camel@redhat.com>
 <328ab7b3-4890-4e0d-8b9a-fed7700f1a6a@nbd.name>
 <66285f637bdec_1924cc2941b@willemb.c.googlers.com.notmuch>
 <eb0e1eb6-306c-4cf7-922d-203db09d8d29@nbd.name>
Subject: Re: [RFC] net: add TCP fraglist GRO support
Precedence: bulk
Mime-Version: 1.0
Content-Type: text/plain;
 charset=utf-8
Content-Transfer-Encoding: quoted-printable

Felix Fietkau wrote:
> On 24.04.24 03:24, Willem de Bruijn wrote:
> > Felix Fietkau wrote:
> >> On 23.04.24 16:34, Paolo Abeni wrote:
> >> > On Tue, 2024-04-23 at 14:23 +0200, Felix Fietkau wrote:
> >> >> On 23.04.24 14:11, Eric Dumazet wrote:
> >> >> > On Tue, Apr 23, 2024 at 1:55=E2=80=AFPM Felix Fietkau <nbd@nbd.=
name> wrote:
> >> >> > > =

> >> >> > > In the world of consumer-grade WiFi devices, there are a lot =
of chipsets
> >> >> > > with limited or nonexistent SG support, and very limited chec=
ksum
> >> >> > > offload capabilities on Ethernet. The WiFi side of these devi=
ces is
> >> >> > > often even worse. I think fraglist GRO is a decent fallback f=
or the
> >> >> > > inevitable corner cases.
> >> >> > =

> >> >> > What about netfilter and NAT ? Are they okay with NETIF_F_FRAGL=
IST_GRO already ?
> >> >> > =

> >> >> > Many of these devices are probably using NAT.
> >> >> =

> >> >> In my tests, nftables NAT works just fine, both with and without =

> >> >> flowtable offloading. I didn't see anything in netfilter that wou=
ld have =

> >> >> a problem with this.
> >> > =

> >> > I see you handle explicitly NAT changes in __tcpv4_gso_segment_csu=
m(),
> >> > like the current UDP code.
> >> > =

> >> > The TCP header has many other fields that could be updated affecti=
ng
> >> > the TCP csum.
> >> > Handling every possible mutation looks cumbersome and will likely
> >> > reduce the performance benefits.
> >> > =

> >> > What is your plan WRT other TCP header fields update?
> >> =

> >> I think that should be easy enough to handle. My patch already only =

> >> combines packets where tcp_flag_word(th) is identical. So when =

> >> segmenting, I could handle all flags changes with a single =

> >> inet_proto_csum_replace4 call.
> >> =

> >> > Strictly WRT the patch, I guess it deserves to be split in series,=

> >> > moving UDP helpers in common code and possibly factoring out more
> >> > helpers with separate patches.
> >> Will do.
> > =

> > A significant chunk of the complexity is in the
> > tcp[46]_check_fraglist_gro sk match. Is this heuristic worth the
> > complexity?
> > =

> > It seems that the platforms that will enable NETIF_F_FRAGLIST will
> > be mainly forwarding planes.
> =

> There are people using their devices as file servers and routers at the=
 =

> same time. The heuristic helps for those cases.

Ok.

> > If keeping, this refinement can probably a separate follow-on patch i=
n
> > the series too:
> > =

> > - refactor existing udp code
> > - add segmentation support to handle such packets on tx
> > - add coalescing support that starts building such packets on rx
> > - refine coalescing choice
> I don't really understand what you're suggesting. With my patch, the GR=
O =

> code handles coalescing of packets. Segmentation on output is also =

> supported. The next version of my patch will fix the cases that were to=
o =

> similar to the UDP code, so I guess refactoring to share code doesn't =

> really make sense there.
> Am I missing something?

I mean if breaking up into a series. First have the refactoring patch
which should be easy to review to be a noop. Then add the segmentation
code, which needs to exist before packets may arrive that depend on
it. Then add the code that produces such packets. To break up into
manageable chunks.