[Intel-wired-lan] [PATCH v4] e1000e: Increase polling timeout on MDIC ready bit

[Neftin, Sasha]

Hello Kai-Heng,
On 9/29/2020 16:31, Kai-Heng Feng wrote:
> Hi Sasha,
> 
>> On Sep 29, 2020, at 21:08, Neftin, Sasha <sasha.neftin at intel.com> wrote:
>>
>> On 9/28/2020 11:36, Kai-Heng Feng wrote:
>>> We are seeing the following error after S3 resume:
>>> [  704.746874] e1000e 0000:00:1f.6 eno1: Setting page 0x6020
>>> [  704.844232] e1000e 0000:00:1f.6 eno1: MDI Write did not complete
>>> [  704.902817] e1000e 0000:00:1f.6 eno1: Setting page 0x6020
>>> [  704.903075] e1000e 0000:00:1f.6 eno1: reading PHY page 769 (or 0x6020 shifted) reg 0x17
>>> [  704.903281] e1000e 0000:00:1f.6 eno1: Setting page 0x6020
>>> [  704.903486] e1000e 0000:00:1f.6 eno1: writing PHY page 769 (or 0x6020 shifted) reg 0x17
>>> [  704.943155] e1000e 0000:00:1f.6 eno1: MDI Error
>>> ...
>>> [  705.108161] e1000e 0000:00:1f.6 eno1: Hardware Error
>>> As Andrew Lunn pointed out, MDIO has nothing to do with phy, and indeed
>>> increase polling iteration can resolve the issue.
>>> This patch only papers over the symptom, as we don't really know the
>>> root cause of the issue. The most possible culprit is Intel ME, which
>>> may do its own things that conflict with software.
>>> Signed-off-by: Kai-Heng Feng <kai.heng.feng at canonical.com>
>>> ---
>>> v4:
>>>   - States that this patch just papers over the symptom.
>>> v3:
>>>   - Moving delay to end of loop doesn't save anytime, move it back.
>>>   - Point out this is quitely likely caused by Intel ME.
>>> v2:
>>>   - Increase polling iteration instead of powering down the phy.
>>>   drivers/net/ethernet/intel/e1000e/phy.c | 2 +-
>>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>> diff --git a/drivers/net/ethernet/intel/e1000e/phy.c b/drivers/net/ethernet/intel/e1000e/phy.c
>>> index e11c877595fb..e6d4acd90937 100644
>>> --- a/drivers/net/ethernet/intel/e1000e/phy.c
>>> +++ b/drivers/net/ethernet/intel/e1000e/phy.c
>>> @@ -203,7 +203,7 @@ s32 e1000e_write_phy_reg_mdic(struct e1000_hw *hw, u32 offset, u16 data)
>>>   	 * Increasing the time out as testing showed failures with
>>>   	 * the lower time out
>>>   	 */
>>> -	for (i = 0; i < (E1000_GEN_POLL_TIMEOUT * 3); i++) {
>>> +	for (i = 0; i < (E1000_GEN_POLL_TIMEOUT * 10); i++) {
>> As we discussed (many threads) - AMT/ME systems not supported on Linux as properly. I do not think increasing polling iteration will solve the problem. Rather mask it.
> 
> I am aware of the status quo of no proper support on Intel ME.
> 
>> I prefer you check option to disable ME vi BIOS on your system.
> 
> We can't ask user to change the BIOS to accommodate Linux. So before a proper solution comes out, masking the problem is good enough for me.
> Until then, I'll carry it as a downstream distro patch.
What will you do with system that even after increasing polling time 
will run into HW error?
> 
> Kai-Heng
> 
>>>   		udelay(50);
>>>   		mdic = er32(MDIC);
>>>   		if (mdic & E1000_MDIC_READY)
>> Thanks,
>> Sasha
> 
Thanks,
Sasha