[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <d0902829-c588-4fba-93c0-9c0dfcc221f6@intel.com>
Date: Tue, 22 Apr 2025 15:20:20 -0700
From: Jacob Keller <jacob.e.keller@...el.com>
To: Vishal Badole <Vishal.Badole@....com>, <Shyam-sundar.S-k@....com>,
<andrew+netdev@...n.ch>, <davem@...emloft.net>, <edumazet@...gle.com>,
<kuba@...nel.org>, <pabeni@...hat.com>, <Thomas.Lendacky@....com>,
<netdev@...r.kernel.org>, <linux-kernel@...r.kernel.org>
CC: <stable@...r.kernel.org>
Subject: Re: [PATCH net V2] amd-xgbe: Fix to ensure dependent features are
toggled with RX checksum offload
On 4/21/2025 7:04 AM, Vishal Badole wrote:
> According to the XGMAC specification, enabling features such as Layer 3
> and Layer 4 Packet Filtering, Split Header, Receive Side Scaling (RSS),
> and Virtualized Network support automatically selects the IPC Full
> Checksum Offload Engine on the receive side.
>
> When RX checksum offload is disabled, these dependent features must also
> be disabled to prevent abnormal behavior caused by mismatched feature
> dependencies.
>
> Ensure that toggling RX checksum offload (disabling or enabling) properly
> disables or enables all dependent features, maintaining consistent and
> expected behavior in the network device.
>
My understanding based on previous changes I've made to Intel drivers,
the netdev community opinion here is that the driver shouldn't
automatically change user configuration like this. Instead, it should
reject requests to disable a feature if that isn't possible due to the
other requirements.
In this case, that means checking and rejecting disable of Rx checksum
offload whenever the features which depend on it are enabled, and reject
requests to enable the features when Rx checksum is disabled.
Powered by blists - more mailing lists