On 6/10/24 14:36, Paul Menzel wrote:
Dear Hui,
Thank you for your patch.
Am 10.06.24 um 03:32 schrieb Hui Wang:
This reverts commit bfd546a552e140b0a4c8a21527c39d6d21addb28
Commit bfd546a552e1 ("e1000e: move force SMBUS near the end of
enable_ulp function") introduces system suspend failure on some
ethernet cards, at the moment, the pciid of the affected ethernet
cards include [8086:15b8] and [8086:15bc].
About the regression the commit bfd546a552e1 ("e1000e: move force
… regression introduced by commit …
Got it.
SMBUS near the end of enable_ulp function") tried to fix, looks like
it is not trivial to fix, we need to find a better way to resolve it.
Please send a revert for commit 861e8086029e (e1000e: move force SMBUS
from enable ulp function to avoid PHY loss issue), present since Linux
v6.9-rc3 and not containing enough information in the commit messsage,
so we have a proper baseline. (That’s also why I originally suggested
to split it into two commits (revert + your change).)
In regards to reverting the commit 861e8086029e (e1000e: move force
SMBUS from enable ulp function to avoid PHY loss issue), the author is
Vitaly, let him evaluate how to act.
Thanks.