Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Thu, 9 May 2019 10:44:03 -0700
From:   Jakub Kicinski <jakub.kicinski@netronome.com>
To:     Maxime Chevallier <maxime.chevallier@bootlin.com>
Cc:     davem@davemloft.net, netdev@vger.kernel.org,
        linux-kernel@vger.kernel.org,
        Antoine Tenart <antoine.tenart@bootlin.com>,
        thomas.petazzoni@bootlin.com, gregory.clement@bootlin.com,
        miquel.raynal@bootlin.com, nadavh@marvell.com, stefanc@marvell.com,
        mw@semihalf.com, Russell King <linux@armlinux.org.uk>,
        linux-arm-kernel@lists.infradead.org
Subject: Re: [PATCH net] net: mvpp2: cls: Add missing NETIF_F_NTUPLE flag
Message-ID: <20190509104403.64c9c45b@cakuba.hsd1.ca.comcast.net>
In-Reply-To: <20190509071408.23eae42a@bootlin.com>
References: <20190507123635.17782-1-maxime.chevallier@bootlin.com>
        <20190507102803.09fcb56c@cakuba.hsd1.ca.comcast.net>
        <20190509071408.23eae42a@bootlin.com>
Organization: Netronome Systems, Ltd.
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Thu, 9 May 2019 07:14:08 +0200, Maxime Chevallier wrote:
> Hello Jakub, David,
> 
> On Tue, 7 May 2019 10:28:03 -0700
> Jakub Kicinski <jakub.kicinski@netronome.com> wrote:
> 
> >> -	if (mvpp22_rss_is_supported())
> >> +	if (mvpp22_rss_is_supported()) {
> >>  		dev->hw_features |= NETIF_F_RXHASH;
> >> +		dev->features |= NETIF_F_NTUPLE;    
> >
> >Hm, why not in hw_features?  
> 
> Because as of today, there's nothing implemented to disable
> classification offload in the driver, so the feature can't be toggled.
> 
> Is this an issue ? Sorry if I'm doing this wrong, but I didn't see any
> indication that this feature has to be host-writeable.

No I don't think it's an issue, I was expecting you'd flush all the
filters when feature is disabled (remove them entirely), I didn't
expect that to be too hard.

> I can make so that it's toggle-able, but it's not as straightforward as
> we would think, since the classifier is also used for RSS (so, we can't
> just disable the classifier as a whole, we would have to invalidate
> each registered flow).

Ack, I don't think disabling the hardware components is required.
Just remove the existing filters, and don't allow new ones.  
But no strong feelings here, feel free to repost with:

Acked-by: Jakub Kicinski <jakub.kicinski@netronome.com>

if flushing the filters is too much hassle.