[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <09f73d4d-efa3-479d-96b5-fd51d8687a21@lunn.ch>
Date: Tue, 27 May 2025 21:08:21 +0200
From: Andrew Lunn <andrew@...n.ch>
To: Ricard Bejarano <ricard@...arano.io>
Cc: Mika Westerberg <mika.westerberg@...ux.intel.com>,
netdev@...r.kernel.org, michael.jamet@...el.com,
YehezkelShB@...il.com, andrew+netdev@...n.ch, davem@...emloft.net,
edumazet@...gle.com, kuba@...nel.org, pabeni@...hat.com
Subject: Re: Poor thunderbolt-net interface performance when bridged
> Then set up iperf3 and tcpdump, but kernel panics:
>
> May 27 18:30:32 blue kernel: skbuff: skb_over_panic: text:ffffffffc0c1b9e7 len:1545195755 put:1545195755 head:ffff9c9dcc652000 data:ffff9c9dcc65200c tail:0x5c19d0f7 end:0x1ec0 dev:<NULL>
O.K. It could be the description does not actually contain any data,
just the flag indicating bad things happened.
So, another idea, just to see if skb with frags are an
issue. Somewhere need the beginning of tbnet_start_xmit() add:
if (skb_is_nonlinear(skb))
skb = skb_linearize(skb);
That should convert an skb with fragments to a skb without
fragments. It will be bad for performance, so if it does work it is
not a fix, but it will confirm we are in the right area.
Andrew
Powered by blists - more mailing lists