netdev - Re: [PATCH v3 net-next 13/19] ionic: Add initial ethtool support

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <2fc052de-2d08-5517-3ee2-094364c4284f@pensando.io>
Date:   Thu, 18 Jul 2019 10:14:59 -0700
From:   Shannon Nelson <snelson@...sando.io>
To:     Andrew Lunn <andrew@...n.ch>
Cc:     netdev@...r.kernel.org
Subject: Re: [PATCH v3 net-next 13/19] ionic: Add initial ethtool support

On 7/17/19 8:31 PM, Andrew Lunn wrote:
> On Fri, Jul 12, 2019 at 10:32:38PM -0700, Shannon Nelson wrote:
>> On 7/8/19 7:30 PM, Andrew Lunn wrote:
>>>> +static int ionic_nway_reset(struct net_device *netdev)
>>>> +{
>>>> +	struct lif *lif = netdev_priv(netdev);
>>>> +	int err = 0;
>>>> +
>>>> +	if (netif_running(netdev))
>>>> +		err = ionic_reset_queues(lif);
>>> What does ionic_reset_queues() do? It sounds nothing like restarting
>>> auto negotiation?
>>>
>>>       Andrew
>> Basically, it's a rip-it-all-down-and-start-over way of restarting the
>> connection, and is also useful for fixing queues that are misbehaving.  It's
>> a little old-fashioned, taken from the ixgbe example, but is effective when
>> there isn't an actual "restart auto-negotiation" command in the firmware.
> O.K. More comments please.

After a little more discussion with the nic connection folks, they've 
said to just flap the link to force a new auto-negotiation, so that's 
what I'm adding - set state to off, then back on.  And yes, another 
comment here.

>
> Did you consider throwing the firmware away and just letting Linux
> control the hardware? It would make this all much more transparent and
> debuggable.
>

Yes, the tension between direct driver bit twiddling and smartnic 
offloading continues.

sln