linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Cristian Ciocaltea <cristian.ciocaltea@gmail.com>
To: Andrew Lunn <andrew@lunn.ch>
Cc: "David S. Miller" <davem@davemloft.net>,
	"Jakub Kicinski" <kuba@kernel.org>,
	"Rob Herring" <robh+dt@kernel.org>,
	"Andreas Färber" <afaerber@suse.de>,
	"Manivannan Sadhasivam" <mani@kernel.org>,
	"Philipp Zabel" <p.zabel@pengutronix.de>,
	netdev@vger.kernel.org, devicetree@vger.kernel.org,
	linux-arm-kernel@lists.infradead.org,
	linux-actions@lists.infradead.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v3 2/3] net: ethernet: actions: Add Actions Semi Owl Ethernet MAC driver
Date: Tue, 23 Mar 2021 10:34:09 +0200	[thread overview]
Message-ID: <20210323083409.GA1559969@BV030612LT> (raw)
In-Reply-To: <YFkqYqgwDhV/bBlc@lunn.ch>

On Tue, Mar 23, 2021 at 12:38:10AM +0100, Andrew Lunn wrote:
> > +static void owl_emac_set_multicast(struct net_device *netdev, int count)
> > +{
> > +	struct owl_emac_priv *priv = netdev_priv(netdev);
> > +	struct netdev_hw_addr *ha;
> > +	int index = 0;
> > +
> > +	if (count <= 0) {
> > +		priv->mcaddr_list.count = 0;
> > +		return;
> > +	}
> > +
> > +	netdev_for_each_mc_addr(ha, netdev) {
> > +		if (!is_multicast_ether_addr(ha->addr))
> > +			continue;
> 
> Is this possible?

I remember I've seen this in one of the drivers I have studied, but
I'm not really sure it is actually necessary. I added it to be on the
safe side..

> > +
> > +		WARN_ON(index >= OWL_EMAC_MAX_MULTICAST_ADDRS);
> > +		ether_addr_copy(priv->mcaddr_list.addrs[index++], ha->addr);
> > +	}
> > +
> > +	priv->mcaddr_list.count = index;
> > +
> > +	owl_emac_setup_frame_xmit(priv);
> > +}
> > +
> > +static void owl_emac_ndo_set_rx_mode(struct net_device *netdev)
> > +{
> > +	struct owl_emac_priv *priv = netdev_priv(netdev);
> > +	u32 status, val = 0;
> > +	int mcast_count = 0;
> > +
> > +	if (netdev->flags & IFF_PROMISC) {
> > +		val = OWL_EMAC_BIT_MAC_CSR6_PR;
> > +	} else if (netdev->flags & IFF_ALLMULTI) {
> > +		val = OWL_EMAC_BIT_MAC_CSR6_PM;
> > +	} else if (netdev->flags & IFF_MULTICAST) {
> > +		mcast_count = netdev_mc_count(netdev);
> > +
> > +		if (mcast_count > OWL_EMAC_MAX_MULTICAST_ADDRS) {
> > +			val = OWL_EMAC_BIT_MAC_CSR6_PM;
> > +			mcast_count = 0;
> > +		}
> > +	}
> > +
> > +	spin_lock_bh(&priv->lock);
> > +
> > +	/* Temporarily stop DMA TX & RX. */
> > +	status = owl_emac_dma_cmd_stop(priv);
> > +
> > +	/* Update operation modes. */
> > +	owl_emac_reg_update(priv, OWL_EMAC_REG_MAC_CSR6,
> > +			    OWL_EMAC_BIT_MAC_CSR6_PR | OWL_EMAC_BIT_MAC_CSR6_PM,
> > +			    val);
> > +
> > +	/* Restore DMA TX & RX status. */
> > +	owl_emac_dma_cmd_set(priv, status);
> > +
> > +	spin_unlock_bh(&priv->lock);
> > +
> > +	/* Set/reset multicast addr list. */
> > +	owl_emac_set_multicast(netdev, mcast_count);
> > +}
> 
> I think this can be simplified. At least, you had me going around in
> circles a while trying to see if WARN_ON() could be triggered from
> user space.
> 
> If you have more than OWL_EMAC_MAX_MULTICAST_ADDRS MC addresses, you
> go into promisc mode. Can you then skip calling
> owl_emac_set_multicast(), which appears not to do too much useful when
> passed 0?

The main purpose of always calling owl_emac_set_multicast() is to ensure
the size of the mcaddr_list is correctly updated (either set or reset).
This prevents owl_emac_setup_frame_xmit() using obsolete data, when
invoked from different contexts (i.e. MAC address changed).

A conditional call involves splitting the mcaddr_list management logic
(i.e. moving the 'reset' operation from the callee to the caller), which
IMO would make the usage of a separate function less justified.

Thanks,
Cristi

> 
>        Andrew

  reply	other threads:[~2021-03-23  8:35 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-03-21 23:29 [PATCH v3 0/3] Add support for Actions Semi Owl Ethernet MAC Cristian Ciocaltea
2021-03-21 23:29 ` [PATCH v3 1/3] dt-bindings: net: Add Actions Semi Owl Ethernet MAC binding Cristian Ciocaltea
2021-03-21 23:29 ` [PATCH v3 2/3] net: ethernet: actions: Add Actions Semi Owl Ethernet MAC driver Cristian Ciocaltea
2021-03-22  3:30   ` Florian Fainelli
2021-03-22  8:44     ` Cristian Ciocaltea
2021-06-28  8:25       ` Amit Tomer
2021-06-28  9:57         ` Cristian Ciocaltea
2021-03-22 23:38   ` Andrew Lunn
2021-03-23  8:34     ` Cristian Ciocaltea [this message]
2021-03-21 23:29 ` [PATCH v3 3/3] MAINTAINERS: Add entries for Actions Semi Owl Ethernet MAC Cristian Ciocaltea
2021-03-22 20:00 ` [PATCH v3 0/3] Add support " patchwork-bot+netdevbpf

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20210323083409.GA1559969@BV030612LT \
    --to=cristian.ciocaltea@gmail.com \
    --cc=afaerber@suse.de \
    --cc=andrew@lunn.ch \
    --cc=davem@davemloft.net \
    --cc=devicetree@vger.kernel.org \
    --cc=kuba@kernel.org \
    --cc=linux-actions@lists.infradead.org \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mani@kernel.org \
    --cc=netdev@vger.kernel.org \
    --cc=p.zabel@pengutronix.de \
    --cc=robh+dt@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).