netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Florian Fainelli <f.fainelli@gmail.com>
To: netdev@vger.kernel.org
Cc: Florian Fainelli <f.fainelli@gmail.com>,
	Doug Berger <opendmb@gmail.com>,
	"David S. Miller" <davem@davemloft.net>,
	Jakub Kicinski <kuba@kernel.org>,
	bcm-kernel-feedback-list@broadcom.com (open list:BROADCOM GENET
	ETHERNET DRIVER), linux-kernel@vger.kernel.org (open list)
Subject: [PATCH net-next] net: bcmgenet: Patch PHY interface for dedicated PHY driver
Date: Fri, 17 Sep 2021 14:55:38 -0700	[thread overview]
Message-ID: <20210917215539.3020216-1-f.fainelli@gmail.com> (raw)

When we are using a dedicated PHY driver (not the Generic PHY driver)
chances are that it is going to configure RGMII delays and do that in a
way that is incompatible with our incorrect interpretation of the
phy_interface value.

Add a quirk in order to reverse the PHY_INTERFACE_MODE_RGMII to the
value of PHY_INTERFACE_MODE_RGMII_ID such that the MAC continues to be
configured the way it used to be, but the PHY driver can account for
adding delays. Conversely when PHY_INTERFACE_MODE_RGMII_TXID is
specified, return PHY_INTERFACE_MODE_RGMII_RXID to the PHY since we will
have enabled a TXC MAC delay (id_mode_dis=0, meaning there is a delay
inserted).

This is not considered a bug fix at this point since it only affects
Broadcom STB platforms shipping with a Device Tree blob that is not
updatable in the field (quite a few devices out there) and which was
generated using the scripted Device Tree environment shipped with those
platforms' SDK.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
---
 drivers/net/ethernet/broadcom/genet/bcmmii.c | 38 ++++++++++++++++++--
 1 file changed, 36 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/broadcom/genet/bcmmii.c b/drivers/net/ethernet/broadcom/genet/bcmmii.c
index 89d16c587bb7..2d29de9a33e3 100644
--- a/drivers/net/ethernet/broadcom/genet/bcmmii.c
+++ b/drivers/net/ethernet/broadcom/genet/bcmmii.c
@@ -286,6 +286,7 @@ int bcmgenet_mii_probe(struct net_device *dev)
 	struct bcmgenet_priv *priv = netdev_priv(dev);
 	struct device *kdev = &priv->pdev->dev;
 	struct device_node *dn = kdev->of_node;
+	phy_interface_t phy_iface = priv->phy_interface;
 	struct phy_device *phydev;
 	u32 phy_flags = 0;
 	int ret;
@@ -300,9 +301,42 @@ int bcmgenet_mii_probe(struct net_device *dev)
 	priv->old_duplex = -1;
 	priv->old_pause = -1;
 
+	/* This is an ugly quirk but we have not been correctly interpreting
+	 * the phy_interface values and we have done that across different
+	 * drivers, so at least we are consistent in our mistakes.
+	 *
+	 * When the Generic PHY driver is in use either the PHY has been
+	 * strapped or programmed correctly by the boot loader so we should
+	 * stick to our incorrect interpretation since we have validated it.
+	 *
+	 * Now when a dedicated PHY driver is in use, we need to reverse the
+	 * meaning of the phy_interface_mode values to something that the PHY
+	 * driver will interpret and act on such that we have two mistakes
+	 * canceling themselves so to speak. We only do this for the two
+	 * modes that GENET driver officially supports on Broadcom STB chips:
+	 * PHY_INTERFACE_MODE_RGMII and PHY_INTERFACE_MODE_RGMII_TXID. Other
+	 * modes are not *officially* supported with the boot loader and the
+	 * scripted environment generating Device Tree blobs for those
+	 * platforms.
+	 *
+	 * Note that internal PHY, MoCA and fixed-link configurations are not
+	 * affected because they use different phy_interface_t values or the
+	 * Generic PHY driver.
+	 */
+	switch (priv->phy_interface) {
+	case PHY_INTERFACE_MODE_RGMII:
+		phy_iface = PHY_INTERFACE_MODE_RGMII_ID;
+		break;
+	case PHY_INTERFACE_MODE_RGMII_TXID:
+		phy_iface = PHY_INTERFACE_MODE_RGMII_RXID;
+		break;
+	default:
+		break;
+	}
+
 	if (dn) {
 		phydev = of_phy_connect(dev, priv->phy_dn, bcmgenet_mii_setup,
-					phy_flags, priv->phy_interface);
+					phy_flags, phy_iface);
 		if (!phydev) {
 			pr_err("could not attach to PHY\n");
 			return -ENODEV;
@@ -332,7 +366,7 @@ int bcmgenet_mii_probe(struct net_device *dev)
 		phydev->dev_flags = phy_flags;
 
 		ret = phy_connect_direct(dev, phydev, bcmgenet_mii_setup,
-					 priv->phy_interface);
+					 phy_iface);
 		if (ret) {
 			pr_err("could not attach to PHY\n");
 			return -ENODEV;
-- 
2.25.1


             reply	other threads:[~2021-09-17 21:55 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-09-17 21:55 Florian Fainelli [this message]
2021-09-18 11:00 ` [PATCH net-next] net: bcmgenet: Patch PHY interface for dedicated PHY driver 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=20210917215539.3020216-1-f.fainelli@gmail.com \
    --to=f.fainelli@gmail.com \
    --cc=bcm-kernel-feedback-list@broadcom.com \
    --cc=davem@davemloft.net \
    --cc=kuba@kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=netdev@vger.kernel.org \
    --cc=opendmb@gmail.com \
    /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).