[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <f8c9cc3f-4cba-488b-9c93-c31b404f4d63@lunn.ch>
Date: Tue, 11 Feb 2025 15:16:46 +0100
From: Andrew Lunn <andrew@...n.ch>
To: "Jagielski, Jedrzej" <jedrzej.jagielski@...el.com>
Cc: Paul Menzel <pmenzel@...gen.mpg.de>,
"intel-wired-lan@...ts.osuosl.org" <intel-wired-lan@...ts.osuosl.org>,
"Nguyen, Anthony L" <anthony.l.nguyen@...el.com>,
"netdev@...r.kernel.org" <netdev@...r.kernel.org>,
Simon Horman <horms@...nel.org>,
"Kitszel, Przemyslaw" <przemyslaw.kitszel@...el.com>,
"Polchlopek, Mateusz" <mateusz.polchlopek@...el.com>
Subject: Re: [Intel-wired-lan] [PATCH iwl-next v3] ixgbe: add support for
thermal sensor event reception
> Actually there is only one adapter across all portfolio of ixgbe adapters
> which supports this feature. That is 82599, none other supports it.
> Even next generations (x540, x550) didn't provide support for reading thermal
> data sensor.
> As E610 is some type of extending x550 it also follows this path at this point.
It is something you should consider. The machine disappears off the
net for no obvious reason, and needs a cold boot to get it back? vs
HWMON entries you can monitor, a warning when the critical value is
reached, which probably reaches the network console and so gets logged
somewhere, and then the emergency value which shuts down the NIC
without any notification getting out of the box.
Also, if there is temperature information, Linux can take an active
part in managing it. If the critical value is reached, it could
downshift to a lower link mode. Better to have a slower link than no
link and a cold boot.
Andrew
Powered by blists - more mailing lists