On 6/19/24 20:17, Lifshits, Vitaly wrote:
Hi Hui,
On 6/14/2024 4:53 AM, Hui Wang wrote:
Hi Vitaly
I tested the patch on a laptop with the ethernet card [8086:550A],
the system suspend and resume worked well with the cable plugged or
unplugged.
But I still think that reverting 2 regression commits is a better
solution.
Reverting the patches will bring us back to the starting point where
we had a PHY access loss issue during suspend flows on MTL systems. I
don't think that it is what we want to accomplish.
OK, got it.
Thanks.
...
+ if (hw->mac.type == e1000_pch_mtp) {
Maybe we should also check ret_val like below:
No, because we want to unconditionally force the MAC and the PHY to
transition to SMBUS mode. This is how it used to work previously
before moving this part of the code to the end of the function.
OK, got it.
Thanks.
if (!ret_val && hw->mac.type == e1000_pch_mtp) {