ACK/cmt / APPLIED[D]: [PATCH 0/1][SRU][B][C][D][OEM-B]PC SN720 NVMe WDC 256GB consumes more power in S2Idle than during long idle

[AceLan Kao]

This patch is not going upstream.

Maintainer doesn't like the quirks, doesn't agree this should be fixed
in drivers,
and think it should be a firmware issue that doesn't behave well during D3.

We've setup many meetings with nvme vendors, and one vendor says from the spec.
there are no words that describes they should enter low power mode when they are
put into suspended state, and their firmware design doesn't allow to
add code to enter
low power mode.

We have more meetings with nvme vendors to figure out a better solution,
but currently adding quirk and preserve them in ubuntu kernel is the
only feasible way
we can do.
Seth Forshee <seth.forshee at canonical.com> 於 2018年12月5日 週三 下午9:57寫道：
>
> On Thu, Nov 29, 2018 at 03:44:23PM +0800, AceLan Kao wrote:
> > To avoid conflicts, this patch assumes below commits have been applied
> > on top of master-next branch on Bionic, Cosmic, and Disco kernel,
> > since both patches modify the same files on the same position for the
> > same issue.
> > https://lists.ubuntu.com/archives/kernel-team/2018-November/096832.html
> >
> > BugLink: https://bugs.launchpad.net/bugs/1805775
> >
> > [Impact]
> > On new systems that facilitate s2idle, we observed the power consumption
> > raises higher than long idle does during s2idle with Western Digital PC
> > SN720 NVMe SSD SDAPNTW-256G.
> >
> > Short idle: 5.3
> > Long idle: 3.0
> > S2I: 5.07
> >
> > [Fix]
> > Windows doesn't put nvme to D3 in modern standby, and uses its own APST
> > feature to do the power management. To leverage its APST feature during
> > s2idle, we can't disable nvme device while suspending, too.
> > So, here is what we did on the driver, 1. prevent nvme from entering D3,
> > 2. prevent nvme from being disabled when suspending.
> >
> > [Test]
> > Verified on the WD NVMe, it fixes the power consumption issue with no
> > regression. And the power consumption decreases to 1.66W during s2idle.
> >
> > [Regression Potential]
> > Low, the patches only applied to specific nvme module, and from our test,
> > the system is still stable.
>
> Scope limited to a single device, positive testing. Note though that the
> patches should be prefixed with "UBUNTU: SAUCE:" and not just "SAUCE:".
>
> Acked-by: Seth Forshee <seth.forshee at canonical.com>
>
> Is this patch going upstream?