[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <1491515218.4166.124.camel@kernel.crashing.org>
Date: Fri, 07 Apr 2017 07:46:58 +1000
From: Benjamin Herrenschmidt <benh@...nel.crashing.org>
To: Florian Fainelli <f.fainelli@...il.com>, netdev@...r.kernel.org
Subject: Re: [PATCH v2 00/13] ftgmac100: Rework batch 1 - Link & Interrupts
On Thu, 2017-04-06 at 12:46 -0700, Florian Fainelli wrote:
> > I thought a while ago we could add some dev flag to prevent the link
> > watch from doing that, but never got to look into it myself and
> > apparently neither did Gavin.
>
> It sounds like a similar situation to e.g: 802.1x, you want to let some
> frames make it through and you need a working link for that, but you
> can't quite flag the device as UP entirely. Maybe you can find a way to
> re-use IFF_LOWER_UP to that purpose and only set that flag based on the
> NC-SI frames indicating link state?
Possibly yes, I wasn't familiar with the difference betweenÂ
"carrier" and IFF_LOWER_UP so far. I'll have a look or have Gavin do
so.
It would be nice to reflect that properly so things like fallover or
bonding do the right thing. On the other hand I yet have to see a
system where the BMC has more than one NC-SI link to a NIC (what they
usually have is that NC-SI is connected to multiple NICs which is a
different matter and should be handled within the NC-SI stack).
Anyway, I'll look into it further down the track along with other
improvements I'd like to do to the NC-SI stack such as properly setting
up multicast and vlan filters in the peer NIC and various cleanups.
Cheers,
Ben.
Powered by blists - more mailing lists