[PATCH v2 2/2] e1000e: fix link fluctuations problem

From: Ricky Wu
Date: Fri May 03 2024 - 06:19:03 EST


As described in https://bugzilla.kernel.org/show_bug.cgi?id=218642,
Intel I219-LM reports link up -> link down -> link up after hot-plugging
the Ethernet cable.

The problem is because the unstable behavior of Link Status bit in
PHY Status Register of some e1000e NIC. When we re-plug the cable,
the e1000e_phy_has_link_generic() (called after the Link-Status-Changed
interrupt) has read this bit with 1->0->1 (1=link up, 0=link down)
and e1000e reports it to net device layer respectively.

This patch solves the problem by passing polling delays on
e1000e_phy_has_link_generic() so that it will not get the unstable
states of Link Status bit.

Link: https://bugzilla.kernel.org/show_bug.cgi?id=218642
Fixes: 7d3cabbcc86 ("e1000e: disable K1 at 1000Mbps for 82577/82578")
Signed-off-by: Ricky Wu <en-wei.wu@xxxxxxxxxxxxx>
---

In v2:
* Split the sleep codes part into PATCHSET [1/2]
---
drivers/net/ethernet/intel/e1000e/ich8lan.c | 12 +++++++++++-
1 file changed, 11 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/intel/e1000e/ich8lan.c b/drivers/net/ethernet/intel/e1000e/ich8lan.c
index f9e94be36e97..68f5698a22b0 100644
--- a/drivers/net/ethernet/intel/e1000e/ich8lan.c
+++ b/drivers/net/ethernet/intel/e1000e/ich8lan.c
@@ -1428,7 +1428,17 @@ static s32 e1000_check_for_copper_link_ich8lan(struct e1000_hw *hw)
* link. If so, then we want to get the current speed/duplex
* of the PHY.
*/
- ret_val = e1000e_phy_has_link_generic(hw, 1, 0, &link);
+ /* We've seen that I219-LM sometimes has link fluctuations
+ * (link up -> link down -> link up) after hot-plugging the cable.
+ * The problem is caused by the instability of the Link Status bit
+ * (BMSR_LSTATUS) in MII Status Register. The average time between
+ * the first link up and link down is between 3~4 ms.
+ * Increasing the iteration times and setting up the delay to
+ * 100ms (which is safe) solves the problem.
+ * This behavior hasn't been seen on other NICs and also not being
+ * documented in datasheet/errata.
+ */
+ ret_val = e1000e_phy_has_link_generic(hw, COPPER_LINK_UP_LIMIT, 100000, &link);
if (ret_val)
goto out;

--
2.40.1