Re: [Intel-wired-lan] [PATCH iwl-net v1 1/1] e1000e: move force SMBUS from enable ulp function to avoid PHY loss issue

[Paul Menzel]

Dear Vitaly,

Thank you for your reply.


Am 29.02.24 um 11:34 schrieb Lifshits, Vitaly:

On 2/12/2024 11:00 AM, Paul Menzel wrote:

Am 11.02.24 um 16:12 schrieb Vitaly Lifshits:
Forcing SMBUS inside the ULP enabling flow leads to sporadic PHY loss on
some systems.

Please list the systems you know of.
On some Meteor-lake systems, we also suspect that on some legacy 
platforms this issue happened with some low probability.

It’d be great if you could be more specific about the Meteor Lake 
systems. What model and what firmware versions?

It is suspected to be caused by initiating PHY transactions before the
the interface settles.

No new paragraph is needed here. Please do not break the line, just 
because a sentence ends. (If you use paragraphs, please separate them 
by a blank line.
Will be fixed in a v2.

Thank you.

Separating this configuration from the ULP enabling flow avoids this
issue.

Fixes: 6607c99e7034 ("e1000e: i219 - fix to enable both ULP and EEE in Sx 
state")
Signed-off-by: Vitaly Lifshits <vitaly.lifsh...@intel.com>
Co-developed-by: Dima Ruinskiy <dima.ruins...@intel.com>
Signed-off-by: Dima Ruinskiy <dima.ruins...@intel.com>
---
  drivers/net/ethernet/intel/e1000e/ich8lan.c | 19 -------------------
  drivers/net/ethernet/intel/e1000e/netdev.c  | 18 ++++++++++++++++++
  2 files changed, 18 insertions(+), 19 deletions(-)

diff --git a/drivers/net/ethernet/intel/e1000e/ich8lan.c 
b/drivers/net/ethernet/intel/e1000e/ich8lan.c
index 717c52913e84..4d83c9a0c023 100644
--- a/drivers/net/ethernet/intel/e1000e/ich8lan.c
+++ b/drivers/net/ethernet/intel/e1000e/ich8lan.c
@@ -1165,25 +1165,6 @@ s32 e1000_enable_ulp_lpt_lp(struct e1000_hw *hw, bool 
to_sx)
      if (ret_val)
          goto out;
-    /* Switching PHY interface always returns MDI error
-     * so disable retry mechanism to avoid wasting time
-     */
-    e1000e_disable_phy_retry(hw);
-
-    /* Force SMBus mode in PHY */
-    ret_val = e1000_read_phy_reg_hv_locked(hw, CV_SMB_CTRL, &phy_reg);
-    if (ret_val)
-        goto release;
-    phy_reg |= CV_SMB_CTRL_FORCE_SMBUS;
-    e1000_write_phy_reg_hv_locked(hw, CV_SMB_CTRL, phy_reg);
-
-    e1000e_enable_phy_retry(hw);
-
-    /* Force SMBus mode in MAC */
-    mac_reg = er32(CTRL_EXT);
-    mac_reg |= E1000_CTRL_EXT_FORCE_SMBUS;
-    ew32(CTRL_EXT, mac_reg);
-
      /* Si workaround for ULP entry flow on i127/rev6 h/w.  Enable
       * LPLU and disable Gig speed when entering ULP
       */
diff --git a/drivers/net/ethernet/intel/e1000e/netdev.c 
b/drivers/net/ethernet/intel/e1000e/netdev.c
index af5d9d97a0d6..8fcf8f11f5a4 100644
--- a/drivers/net/ethernet/intel/e1000e/netdev.c
+++ b/drivers/net/ethernet/intel/e1000e/netdev.c
@@ -6622,6 +6622,7 @@ static int __e1000_shutdown(struct pci_dev 
*pdev, bool runtime)
      struct e1000_adapter *adapter = netdev_priv(netdev);
      struct e1000_hw *hw = &adapter->hw;
      u32 ctrl, ctrl_ext, rctl, status, wufc;
+    u16 smb_ctrl;
      int retval = 0;
      /* Runtime suspend should only enable wakeup for link changes */
@@ -6696,6 +6697,23 @@ static int __e1000_shutdown(struct pci_dev *pdev, bool 
runtime)
          if (retval)
              return retval;
+
+        /* Force SMBUS to allow WOL */
+        /* Switching PHY interface always returns MDI error
+         * so disable retry mechanism to avoid wasting time
+         */
+        e1000e_disable_phy_retry(hw);
+
+        e1e_rphy(hw, CV_SMB_CTRL, &smb_ctrl);
+        smb_ctrl |= CV_SMB_CTRL_FORCE_SMBUS;
+        e1e_wphy(hw, CV_SMB_CTRL, smb_ctrl);
+
+        e1000e_enable_phy_retry(hw);
+
+        /* Force SMBus mode in MAC */
+        ctrl_ext = er32(CTRL_EXT);
+        ctrl_ext |= E1000_CTRL_EXT_FORCE_SMBUS;
+        ew32(CTRL_EXT, ctrl_ext);
      }
      /* Ensure that the appropriate bits are set in LPI_CTRL

Why is `__e1000_shutdown()` the correct place, where it was in 
`e1000_enable_ulp_lpt_lp()` before?
First of all because forcing the SMBUS is not related to ULP enabling 
flow, but rather a configuration that is related to shutdown. Secondly, 
moving the function to shutdown function allows enough time for the 
interface to settle and avoids adding a delay.

Please add that clarification to the commit message.


Kind regards,

Paul