net: spacemit: Check for netif_carrier_ok() in emac_stats_update()

Some PHYs stop the refclk for power saving, usually while link down.
This causes reading stats to time out.

Therefore, in emac_stats_update(), also don't update and reschedule if
!netif_carrier_ok(). But that means we could be missing later updates if
the link comes back up, so also reschedule when link up is detected in
emac_adjust_link().

While we're at it, improve the comments and error message prints around
this to reflect the better understanding of how this could happen.
Hopefully if this happens again on new hardware, these comments will
direct towards a solution.

Closes: https://lore.kernel.org/r/[email protected]/
Fixes: bfec6d7 ("net: spacemit: Add K1 Ethernet MAC")
Co-developed-by: Chukun Pan <[email protected]>
Signed-off-by: Chukun Pan <[email protected]>
Signed-off-by: Vivian Wang <[email protected]>
Link: https://patch.msgid.link/20260123-k1-ethernet-clarify-stat-timeout-v3-1-93b9df627e87@iscas.ac.cn
Signed-off-by: Jakub Kicinski <[email protected]>

Author: dramforever

drivers/net/ethernet/spacemit/k1_emac.c

@@ -1099,7 +1099,13 @@ static int emac_read_stat_cnt(struct emac_priv *priv, u8 cnt, u32 *res,
 					100, 10000);
 
 	if (ret) {
-		netdev_err(priv->ndev, "Read stat timeout\n");
+		/*
+		 * This could be caused by the PHY stopping its refclk even when
+		 * the link is up, for power saving. See also comments in
+		 * emac_stats_update().
+		 */
+		dev_err_ratelimited(&priv->ndev->dev,
+				    "Read stat timeout. PHY clock stopped?\n");
 		return ret;
 	}
 
@@ -1147,17 +1153,25 @@ static void emac_stats_update(struct emac_priv *priv)
 
 	assert_spin_locked(&priv->stats_lock);
 
-	if (!netif_running(priv->ndev) || !netif_device_present(priv->ndev)) {
-		/* Not up, don't try to update */
+	/*
+	 * We can't read statistics if the interface is not up. Also, some PHYs
+	 * stop their reference clocks for link down power saving, which also
+	 * causes reading statistics to time out. Don't update and don't
+	 * reschedule in these cases.
+	 */
+	if (!netif_running(priv->ndev) ||
+	    !netif_carrier_ok(priv->ndev) ||
+	    !netif_device_present(priv->ndev)) {
 		return;
 	}
 
 	for (i = 0; i < sizeof(priv->tx_stats) / sizeof(*tx_stats); i++) {
 		/*
-		 * If reading stats times out, everything is broken and there's
-		 * nothing we can do. Reading statistics also can't return an
-		 * error, so just return without updating and without
-		 * rescheduling.
+		 * If reading stats times out anyway, the stat registers will be
+		 * stuck, and we can't really recover from that.
+		 *
+		 * Reading statistics also can't return an error, so just return
+		 * without updating and without rescheduling.
 		 */
 		if (emac_tx_read_stat_cnt(priv, i, &res))
 			return;
@@ -1636,6 +1650,12 @@ static void emac_adjust_link(struct net_device *dev)
 		emac_wr(priv, MAC_GLOBAL_CONTROL, ctrl);
 
 		emac_set_fc_autoneg(priv);
+
+		/*
+		 * Reschedule stats updates now that link is up. See comments in
+		 * emac_stats_update().
+		 */
+		mod_timer(&priv->stats_timer, jiffies);
 	}
 
 	phy_print_status(phydev);