[SRU] [Unstable/OEM-B] [PATCH 0/2] r8169 doesn't get woken up by ethernet cable plugging, no PME generated

[Seth Forshee]

On Tue, Feb 26, 2019 at 04:56:55PM +0800, Kai-Heng Feng wrote:
> BugLink: https://bugs.launchpad.net/bugs/1817676
> 
> [Impact]
> Once r8169 (Realtek ethernet) and its root port enter to D3, plugging
> network cable doesn't wake r8169 up.
> 
> [Fix]
> Revert "PCI/PME: Implement runtime PM callbacks".
> Quote the original commit message:
> "
> The commit tried to prevent root port waking up from D3cold immediately but
> looks like disabing root port PME interrupt is not the right way to fix
> that issue so revert it now.  The patch following proposes an alternative
> solution to that issue.
> "
> 
> [Test]
> After applying the fix, plugging the ethernet cable can generate PME IRQ
> correctly, hence waking up r8169.
> 
> [Regression Potential]
> Medium. The commits are relative new, rely on some PCI detail that I am
> not familiar with (PCI spec isn't publicly available).
> Fortunately we only need this fix for Unstable/OEM-B, so it won't hit
> most users if any regression happens.

I think it would be safer to apply only the revert if the only problem
we're trying to fix is the wakeup on cable plug, since that commit was
new in 4.20 anyhow. If you think we also need the second commit, please
explain why and let me know what kind of testing you've done with it.

Thanks,
Seth