netdev - Re: [RFC net-next 2/5] net: phylink: add separate pcs operations structure

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: <20200324194642.GY25745@shell.armlinux.org.uk>
Date:   Tue, 24 Mar 2020 19:46:42 +0000
From:   Russell King - ARM Linux admin <linux@...linux.org.uk>
To:     Andrew Lunn <andrew@...n.ch>
Cc:     Florian Fainelli <f.fainelli@...il.com>,
        Heiner Kallweit <hkallweit1@...il.com>,
        "David S. Miller" <davem@...emloft.net>, netdev@...r.kernel.org
Subject: Re: [RFC net-next 2/5] net: phylink: add separate pcs operations
 structure

On Thu, Mar 19, 2020 at 04:06:52PM +0100, Andrew Lunn wrote:
> > Oh, I forgot to mention on the library point - that's what has already
> > been created in:
> > 
> > "net: phylink: pcs: add 802.3 clause 45 helpers"
> > "net: phylink: pcs: add 802.3 clause 22 helpers"
> > 
> > which add library implementations for the pcs_get_state(), pcs_config()
> > and pcs_an_restart() methods.
> > 
> > What remains is vendor specific - for pcs_link_up(), there is no
> > standard, since it requires fiddling with vendor specific registers to
> > program, e.g. the speed in SGMII mode when in-band is not being used.
> > The selection between different PCS is also vendor specific.
> > 
> > It would have been nice to use these helpers for Marvell DSA switches
> > too, but the complexities of DSA taking a multi-layered approach rather
> > than a library approach, plus the use of paging makes it very
> > difficult.
> > 
> > So, basically on the library point, "already considered and
> > implemented".
> 
> Hi Russell
> 
> The 6390X family of switches has two PCSs, one for 1000BaseX/SGMII and
> a second one for 10GBaseR. So at some point there is going to be a
> mux, but maybe it will be internal to mv88e6xxx and not shareable. Or
> internal to DSA, and shareable between DSA drivers. We will see.

This is what I came up with, but I'm not really able to test it with
my ZII platforms. See the attached patch and the patch below. I
haven't polished them yet - been a little otherwise occupied as I'm
sure you can imagine given the situation.

diff --git a/drivers/net/dsa/mv88e6xxx/serdes.c b/drivers/net/dsa/mv88e6xxx/serdes.c
index 8cef46ee1d2f..59bd3f447386 100644
--- a/drivers/net/dsa/mv88e6xxx/serdes.c
+++ b/drivers/net/dsa/mv88e6xxx/serdes.c
@@ -631,8 +631,8 @@ int mv88e6390_serdes_pcs_config(struct mv88e6xxx_chip *chip, int port,
 				      MV88E6390_SGMII_BMCR, bmcr);
 }
 
-int mv88e6390_serdes_pcs_get_state(struct mv88e6xxx_chip *chip, int port,
-				   u8 lane, struct phylink_link_state *state)
+static int mv88e6390_serdes_pcs_get_state_sgmii(struct mv88e6xxx_chip *chip,
+	int port, u8 lane, struct phylink_link_state *state)
 {
 	u16 lpa, status;
 	int err;
@@ -654,6 +654,45 @@ int mv88e6390_serdes_pcs_get_state(struct mv88e6xxx_chip *chip, int port,
 	return mv88e6xxx_serdes_pcs_get_state(chip, status, lpa, state);
 }
 
+static int mv88e6390_serdes_pcs_get_state_10g(struct mv88e6xxx_chip *chip,
+	int port, u8 lane, struct phylink_link_state *state)
+{
+	u16 status;
+	int err;
+
+	err = mv88e6390_serdes_read(chip, lane, MDIO_MMD_PHYXS,
+				    MV88E6390_10G_STAT1, &status);
+	if (err)
+		return err;
+
+	state->link = !!(status & MDIO_STAT1_LSTATUS);
+	if (state->link) {
+		state->speed = SPEED_10000;
+		state->duplex = DUPLEX_FULL;
+	}
+
+	return 0;
+}
+
+int mv88e6390_serdes_pcs_get_state(struct mv88e6xxx_chip *chip, int port,
+				   u8 lane, struct phylink_link_state *state)
+{
+	switch (state->interface) {
+	case PHY_INTERFACE_MODE_SGMII:
+	case PHY_INTERFACE_MODE_1000BASEX:
+	case PHY_INTERFACE_MODE_2500BASEX:
+		return mv88e6390_serdes_pcs_get_state_sgmii(chip, port, lane,
+							    state);
+	case PHY_INTERFACE_MODE_XAUI:
+	case PHY_INTERFACE_MODE_RXAUI:
+		return mv88e6390_serdes_pcs_get_state_10g(chip, port, lane,
+							  state);
+
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
 int mv88e6390_serdes_pcs_an_restart(struct mv88e6xxx_chip *chip, int port,
 				    u8 lane)
 {
diff --git a/drivers/net/dsa/mv88e6xxx/serdes.h b/drivers/net/dsa/mv88e6xxx/serdes.h
index 12013f22df10..3af236f4b3c8 100644
--- a/drivers/net/dsa/mv88e6xxx/serdes.h
+++ b/drivers/net/dsa/mv88e6xxx/serdes.h
@@ -41,6 +41,7 @@
 
 /* 10GBASE-R and 10GBASE-X4/X2 */
 #define MV88E6390_10G_CTRL1		(0x1000 + MDIO_CTRL1)
+#define MV88E6390_10G_STAT1		(0x1000 + MDIO_STAT1)
 
 /* 1000BASE-X and SGMII */
 #define MV88E6390_SGMII_BMCR		(0x2000 + MII_BMCR)

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTC broadband for 0.8mile line in suburbia: sync at 10.2Mbps down 587kbps up

View attachment "0001-net-dsa-mv88e6xxx-use-generic-clause-45-definitions.patch" of type "text/x-diff" (2032 bytes)