[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <dd3d10e5b14a1ccceed52e19e83179993efc1f8e.camel@mellanox.com>
Date: Thu, 13 Dec 2018 17:08:59 +0000
From: Saeed Mahameed <saeedm@...lanox.com>
To: "nikola.ciprich@...uxbox.cz" <nikola.ciprich@...uxbox.cz>
CC: "davem@...emloft.net" <davem@...emloft.net>,
"netdev@...r.kernel.org" <netdev@...r.kernel.org>,
"saeedm@....mellanox.co.il" <saeedm@....mellanox.co.il>,
"xiyou.wangcong@...il.com" <xiyou.wangcong@...il.com>,
"edumazet@...gle.com" <edumazet@...gle.com>
Subject: Re: [Patch net v2] mlx5: fixup checksum for short ethernet frame
padding
On Thu, 2018-12-13 at 09:40 +0100, Nikola Ciprich wrote:
> Hi Saeed,
>
> > for now i feel that you don't want csum complete in your servers
> > and we
> > already have the tool for that, just do:
> >
> > ethtool --set-priv-flags <ifname> rx_no_csum_complete on
> > ethtool --show-priv-flags <ifname>
> > Private flags for p5p1:
> > rx_cqe_moder : on
> > tx_cqe_moder : off
> > rx_cqe_compress : off
> > rx_striding_rq : off
> > rx_no_csum_complete: on
> >
> > this will disable csum complete and avoid the csum error for padded
> > short packets.
>
> we're experiencing the same issue with 4.19 and I'd like to hotfix
> it now before the issue is fixed in the kernel.. however I don't see
> the flags in ethtool (even with latest 4.19 version..)
Hi Nikola
4.19 is missing this priv flag, it was only submitted to 4.20
commit b856df28f9230a47669efbdd57896084caadb2b3
Author: Or Gerlitz <ogerlitz@...lanox.com>
Date: Sun Jul 1 08:58:38 2018 +0000
net/mlx5e: Allow reporting of checksum unnecessary
> I'm using following cards:
>
> d8:00.0 Ethernet controller: Mellanox Technologies MT27710 Family
> [ConnectX-4 Lx]
>
> Am i doing something wrong, or is this different issue?
>
Most likely the same issue, we are finalizing the patch initially
proposed by Cong, you can find it here, I plan to submit it next week,
after all the regression tests.
https://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux.git/commit/?h=topic/csum-fix
It would be nice if you verify this fixes your issue.
Thanks,
Saeed.
> thanks a lot in advance
>
> BR
>
> nik
>
>
>
>
>
>
>
> >
> > Thanks,
> > Saeed.
> >
Powered by blists - more mailing lists