[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAKErNvrfirDasvmDDbGDu=301tOE43RaTW9jVjBL8=pngPH6YQ@mail.gmail.com>
Date: Thu, 20 Mar 2025 09:43:39 +0200
From: Maxim Mikityanskiy <maxtram95@...il.com>
To: Tariq Toukan <ttoukan.linux@...il.com>
Cc: Saeed Mahameed <saeedm@...dia.com>, Leon Romanovsky <leon@...nel.org>,
Andrew Lunn <andrew+netdev@...n.ch>, "David S. Miller" <davem@...emloft.net>,
Eric Dumazet <edumazet@...gle.com>, Jakub Kicinski <kuba@...nel.org>, Paolo Abeni <pabeni@...hat.com>,
netdev@...r.kernel.org, Maxim Mikityanskiy <maxim@...valent.com>,
Tariq Toukan <tariqt@...dia.com>
Subject: Re: [PATCH net] net/mlx5e: Fix ethtool -N flow-type ip4 to RSS context
On Thu, 20 Mar 2025 at 09:32, Tariq Toukan <ttoukan.linux@...il.com> wrote:
>
>
>
> On 19/03/2025 14:45, Maxim Mikityanskiy wrote:
> > There commands can be used to add an RSS context and steer some traffic
> > into it:
> >
> > # ethtool -X eth0 context new
> > New RSS context is 1
> > # ethtool -N eth0 flow-type ip4 dst-ip 1.1.1.1 context 1
> > Added rule with ID 1023
> >
> > However, the second command fails with EINVAL on mlx5e:
> >
> > # ethtool -N eth0 flow-type ip4 dst-ip 1.1.1.1 context 1
> > rmgr: Cannot insert RX class rule: Invalid argument
> > Cannot insert classification rule
> >
> > It happens when flow_get_tirn calls flow_type_to_traffic_type with
> > flow_type = IP_USER_FLOW or IPV6_USER_FLOW. That function only handles
> > IPV4_FLOW and IPV6_FLOW cases, but unlike all other cases which are
> > common for hash and spec, IPv4 and IPv6 defines different contants for
> > hash and for spec:
> >
> > #define TCP_V4_FLOW 0x01 /* hash or spec (tcp_ip4_spec) */
> > #define UDP_V4_FLOW 0x02 /* hash or spec (udp_ip4_spec) */
> > ...
> > #define IPV4_USER_FLOW 0x0d /* spec only (usr_ip4_spec) */
> > #define IP_USER_FLOW IPV4_USER_FLOW
> > #define IPV6_USER_FLOW 0x0e /* spec only (usr_ip6_spec; nfc only) */
> > #define IPV4_FLOW 0x10 /* hash only */
> > #define IPV6_FLOW 0x11 /* hash only */
> >
> > Extend the switch in flow_type_to_traffic_type to support both, which
> > fixes the failing ethtool -N command with flow-type ip4 or ip6.
> >
>
> Hi Maxim,
> Thanks for your patch!
>
> > Fixes: 248d3b4c9a39 ("net/mlx5e: Support flow classification into RSS contexts")
>
> Seems that the issue originates in commit 756c41603a18 ("net/mlx5e:
> ethtool, Support user configuration for RX hash fields"),
Not really; commit 756c41603a18 configures the hash (not flow
direction), and IPV4_FLOW/IPV6_FLOW are already correct constants for
IP-based hashes. Moreover, we don't support them anyway, see
mlx5e_set_rss_hash_opt:
/* RSS does not support anything other than hashing to queues
* on src IP, dest IP, TCP/UDP src port and TCP/UDP dest
* port.
*/
if (flow_type != TCP_V4_FLOW &&
flow_type != TCP_V6_FLOW &&
flow_type != UDP_V4_FLOW &&
flow_type != UDP_V6_FLOW)
return -EOPNOTSUPP;
> when directly
> classifying into an RQ, before the multi RSS context support.
Direct classification into an RQ actually works before my fix, because
it goes to another branch in flow_get_tirn, that doesn't call
flow_type_to_traffic_type. It's only steering to an RSS context that
was broken for flow-type ip4/ip6.
> > Signed-off-by: Maxim Mikityanskiy <maxim@...valent.com>
> > ---
> > drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c | 2 ++
> > 1 file changed, 2 insertions(+)
> >
> > diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c
> > index 773624bb2c5d..d68230a7b9f4 100644
> > --- a/drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c
> > +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c
> > @@ -884,8 +884,10 @@ static int flow_type_to_traffic_type(u32 flow_type)
> > case ESP_V6_FLOW:
> > return MLX5_TT_IPV6_IPSEC_ESP;
> > case IPV4_FLOW:
> > + case IP_USER_FLOW:
> > return MLX5_TT_IPV4;
> > case IPV6_FLOW:
> > + case IPV6_USER_FLOW:
> > return MLX5_TT_IPV6;
> > default:
> > return -EINVAL;
>
Powered by blists - more mailing lists