[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CAM_iQpVFxe5RUjmRha6k-_6UiOA1mKhZG5-jnZ-fuRU9N3r3dg@mail.gmail.com>
Date: Tue, 4 Dec 2018 12:44:03 -0800
From: Cong Wang <xiyou.wangcong@...il.com>
To: Linux Kernel Network Developers <netdev@...r.kernel.org>
Cc: Saeed Mahameed <saeedm@...lanox.com>,
Eric Dumazet <edumazet@...gle.com>,
Tariq Toukan <tariqt@...lanox.com>
Subject: Re: [Patch net v3] mlx5: force CHECKSUM_NONE for short ethernet frames
On Mon, Dec 3, 2018 at 10:14 PM Cong Wang <xiyou.wangcong@...il.com> wrote:
>
> When an ethernet frame is padded to meet the minimum ethernet frame
> size, the padding octets are not covered by the hardware checksum.
> Fortunately the padding octets are ususally zero's, which don't affect
> checksum. However, we have a switch which pads non-zero octets, this
> causes kernel hardware checksum fault repeatedly.
>
> Prior to commit 88078d98d1bb ("net: pskb_trim_rcsum() and CHECKSUM_COMPLETE are friends"),
> skb checksum was forced to be CHECKSUM_NONE when padding is detected.
> After it, we need to keep skb->csum updated, like what we do for RXFCS.
> However, fixing up CHECKSUM_COMPLETE requires to verify and parse IP
> headers, it is not worthy the effort as the packets are so small that
> CHECKSUM_COMPLETE can't save anything.
>
> I tested this patch with RXFCS on and off, it works fine without any
> warning in both cases.
>
> Fixes: 88078d98d1bb ("net: pskb_trim_rcsum() and CHECKSUM_COMPLETE are friends"),
> Cc: Saeed Mahameed <saeedm@...lanox.com>
> Cc: Eric Dumazet <edumazet@...gle.com>
> Cc: Tariq Toukan <tariqt@...lanox.com>
> Signed-off-by: Cong Wang <xiyou.wangcong@...il.com>
Nacked-by: Cong Wang <xiyou.wangcong@...il.com>
This patch has no value for upstream. Let's discard it.
Thanks!
Powered by blists - more mailing lists