Re: [PATCH net-next 3/5] net: mvpp2: cls: Use RSS contexts to handle RSS tables

From: Matteo Croce <mcroce@redhat.com>
To: Maxime Chevallier <maxime.chevallier@bootlin.com>
Cc: "David S . Miller" <davem@davemloft.net>,
	netdev <netdev@vger.kernel.org>,
	LKML <linux-kernel@vger.kernel.org>,
	Antoine Tenart <antoine.tenart@bootlin.com>,
	Thomas Petazzoni <thomas.petazzoni@bootlin.com>,
	gregory.clement@bootlin.com, miquel.raynal@bootlin.com,
	Nadav Haklai <nadavh@marvell.com>,
	Stefan Chulski <stefanc@marvell.com>,
	Marcin Wojtas <mw@semihalf.com>,
	Russell King <linux@armlinux.org.uk>,
	Linux ARM <linux-arm-kernel@lists.infradead.org>
Subject: Re: [PATCH net-next 3/5] net: mvpp2: cls: Use RSS contexts to handle RSS tables
Date: Tue, 14 Apr 2020 01:43:02 +0200	[thread overview]
Message-ID: <CAGnkfhzsx_uEPkZQC-_-_NamTigD8J0WgcDioqMLSHVFa3V6GQ@mail.gmail.com> (raw)
In-Reply-To: <20190524100554.8606-4-maxime.chevallier@bootlin.com>

On Tue, Apr 14, 2020 at 1:21 AM Maxime Chevallier
<maxime.chevallier@bootlin.com> wrote:
>
> The PPv2 controller has 8 RSS tables that are shared across all ports on
> a given PPv2 instance. The previous implementation allocated one table
> per port, leaving others unused.
>
> By using RSS contexts, we can make use of multiple RSS tables per
> port, one being the default table (always id 0), the other ones being
> used as destinations for flow steering, in the same way as rx rings.
>
> This commit introduces RSS contexts management in the PPv2 driver. We
> always reserve one table per port, allocated when the port is probed.
>
> The global table list is stored in the struct mvpp2, as it's a global
> resource. Each port then maintains a list of indices in that global
> table, that way each port can have it's own numbering scheme starting
> from 0.
>
> One limitation that seems unavoidable is that the hashing parameters are
> shared across all RSS contexts for a given port. Hashing parameters for
> ctx 0 will be applied to all contexts.
>
> Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>

Hi all,

I noticed that enabling rxhash blocks the RX on my Macchiatobin. It
works fine with the 10G ports (the RX rate goes 4x up) but it
completely kills the gigabit interface.

# 10G port
root@macchiatobin:~# iperf3 -c 192.168.0.2
Connecting to host 192.168.0.2, port 5201
[  5] local 192.168.0.1 port 42394 connected to 192.168.0.2 port 5201
[ ID] Interval           Transfer     Bitrate         Retr  Cwnd
[  5]   0.00-1.00   sec   941 MBytes  7.89 Gbits/sec  4030    250 KBytes
[  5]   1.00-2.00   sec   933 MBytes  7.82 Gbits/sec  4393    240 KBytes
root@macchiatobin:~# ethtool -K eth0 rxhash on
root@macchiatobin:~# iperf3 -c 192.168.0.2
Connecting to host 192.168.0.2, port 5201
[  5] local 192.168.0.1 port 42398 connected to 192.168.0.2 port 5201
[ ID] Interval           Transfer     Bitrate         Retr  Cwnd
[  5]   0.00-1.00   sec   860 MBytes  7.21 Gbits/sec  428    410 KBytes
[  5]   1.00-2.00   sec   859 MBytes  7.20 Gbits/sec  185    563 KBytes

# gigabit port
root@macchiatobin:~# iperf3 -c turbo
Connecting to host turbo, port 5201
[  5] local 192.168.85.42 port 45144 connected to 192.168.85.6 port 5201
[ ID] Interval           Transfer     Bitrate         Retr  Cwnd
[  5]   0.00-1.00   sec   113 MBytes   948 Mbits/sec    0    407 KBytes
[  5]   1.00-2.00   sec   112 MBytes   942 Mbits/sec    0    428 KBytes
root@macchiatobin:~# ethtool -K eth2 rxhash on
root@macchiatobin:~# iperf3 -c turbo
iperf3: error - unable to connect to server: Resource temporarily unavailable

I've bisected and it seems that this commit causes the issue. I tried
to revert it on nex-next as a second test, but the code has changed a
lot much since, generating too much conflicts.
Can you have a look into this?

Thanks,
-- 
Matteo Croce
per aspera ad upstream