[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20191105174836.4df162dd@cakuba.netronome.com>
Date: Tue, 5 Nov 2019 17:48:36 -0800
From: Jakub Kicinski <jakub.kicinski@...ronome.com>
To: Jeff Kirsher <jeffrey.t.kirsher@...el.com>
Cc: davem@...emloft.net,
Anirudh Venkataramanan <anirudh.venkataramanan@...el.com>,
netdev@...r.kernel.org, nhorman@...hat.com, sassmann@...hat.com,
Tony Nguyen <anthony.l.nguyen@...el.com>,
Andrew Bowers <andrewx.bowers@...el.com>
Subject: Re: [net-next 03/15] ice: Add support for FW recovery mode
detection
On Tue, 5 Nov 2019 16:46:08 -0800, Jeff Kirsher wrote:
> From: Anirudh Venkataramanan <anirudh.venkataramanan@...el.com>
>
> This patch adds support for firmware recovery mode detection.
>
> The idea behind FW recovery mode is to recover from a bad FW state,
> due to corruption or misconfiguration. The FW triggers recovery mode
> by setting the right bits in the GL_MNG_FWSM register and issuing
> an EMP reset.
>
> The driver may or may not be loaded when recovery mode is triggered. So
> on module load, the driver first checks if the FW is already in recovery
> mode. If so, it drops into recovery mode as well, where it creates and
> registers a single netdev that only allows a very small set of repair/
> diagnostic operations (like updating the FW, checking version, etc.)
> through ethtool.
>
> If recovery mode is triggered when the driver is loaded/operational,
> the first indication of this in the driver is via the EMP reset event.
> As part of processing the reset event, the driver checks the GL_MNG_FWSM
> register to determine if recovery mode was triggered. If so, traffic is
> stopped, peers are closed and the ethtool ops are updated to allow only
> repair/diagnostic operations.
>
> Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@...el.com>
> Signed-off-by: Tony Nguyen <anthony.l.nguyen@...el.com>
> Tested-by: Andrew Bowers <andrewx.bowers@...el.com>
> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@...el.com>
You shouldn't need to spawn a fake netdev just to recover the device.
Implement devlink, you can have a devlink instance with full debug
info and allow users to update FW via the flash op, even if driver is
unable to bring up any port.
Powered by blists - more mailing lists