* [PATCH net] net: ibm: emac: fix fixed-link speed and duplex
@ 2026-11-11 0:30 Rosen Penev
0 siblings, 0 replies; only message in thread
From: Rosen Penev @ 2026-11-11 0:30 UTC (permalink / raw)
To: netdev
Cc: Andrew Lunn, David S. Miller, Eric Dumazet, Jakub Kicinski,
Paolo Abeni, Heiner Kallweit, open list
Since commit d99b408ed8e5 ("net: phy: fixed_phy: remove not needed
initialization of phy_device members"), fixed_phy_register() no longer
copies the fixed-link state into the new phy_device, phylib fills it in
once the PHY state machine runs.
EMAC never attaches or starts the fixed PHY. emac_init_phy() copies
phydev->speed and phydev->duplex right after registering it, so it now
gets SPEED_UNKNOWN and DUPLEX_UNKNOWN and emac_configure() programs the
MAC and RGMII bridge for 10 Mbps half duplex.
On the Meraki MX60, whose EMAC has a 1000 Mbps RGMII fixed-link to a
QCA8327 switch, every transmitted frame reports carrier loss, the
switch CPU port counts FCS errors and nothing is received.
Take the fixed-link parameters from the device tree, in both the
"fixed-link" subnode and the legacy property form.
Fixes: d99b408ed8e5 ("net: phy: fixed_phy: remove not needed initialization of phy_device members")
Signed-off-by: Rosen Penev <rosenp@gmail.com>
---
drivers/net/ethernet/ibm/emac/core.c | 36 +++++++++++++++++++++++++++-
1 file changed, 35 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/ibm/emac/core.c b/drivers/net/ethernet/ibm/emac/core.c
index 9fa18ad99f4d..f097b6e49eeb 100644
--- a/drivers/net/ethernet/ibm/emac/core.c
+++ b/drivers/net/ethernet/ibm/emac/core.c
@@ -2696,6 +2696,40 @@ static int emac_dt_phy_probe(struct emac_instance *dev)
return res;
}
+/* The fixed PHY is never started, so phylib never fills in its link state.
+ * Take the link parameters straight from the device tree.
+ */
+static void emac_fixed_link_state(struct emac_instance *dev,
+ struct device_node *np)
+{
+ struct device_node *fixed_link;
+ u32 prop[5];
+
+ dev->phy.autoneg = AUTONEG_DISABLE;
+ dev->phy.advertising = 0;
+
+ fixed_link = of_get_child_by_name(np, "fixed-link");
+ if (fixed_link) {
+ u32 speed = SPEED_UNKNOWN;
+
+ of_property_read_u32(fixed_link, "speed", &speed);
+ dev->phy.speed = speed;
+ dev->phy.duplex = of_property_read_bool(fixed_link, "full-duplex") ?
+ DUPLEX_FULL : DUPLEX_HALF;
+ dev->phy.pause = of_property_read_bool(fixed_link, "pause");
+ dev->phy.asym_pause = of_property_read_bool(fixed_link,
+ "asym-pause");
+ of_node_put(fixed_link);
+ } else if (!of_property_read_u32_array(np, "fixed-link", prop,
+ ARRAY_SIZE(prop))) {
+ /* legacy <phy_id duplex speed pause asym_pause> */
+ dev->phy.duplex = prop[1] ? DUPLEX_FULL : DUPLEX_HALF;
+ dev->phy.speed = prop[2];
+ dev->phy.pause = prop[3];
+ dev->phy.asym_pause = prop[4];
+ }
+}
+
static int emac_init_phy(struct emac_instance *dev)
{
struct device_node *np = dev->ofdev->dev.of_node;
@@ -2730,7 +2764,7 @@ static int emac_init_phy(struct emac_instance *dev)
ndev->phydev = of_phy_find_device(np);
if (res || !ndev->phydev)
return res ? res : -EINVAL;
- emac_adjust_link(dev->ndev);
+ emac_fixed_link_state(dev, np);
put_device(&ndev->phydev->mdio.dev);
}
return 0;
--
2.56.0
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2026-11-11 0:30 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-11-11 0:30 [PATCH net] net: ibm: emac: fix fixed-link speed and duplex Rosen Penev
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®