[Intel-gfx] [PATCH] Revert "nvme-pci: use host managed power state for suspend"

[Chris Wilson]

Quoting Keith Busch (2019-08-16 13:46:41)
> On Fri, Aug 16, 2019 at 6:38 AM Christoph Hellwig <hch at lst.de> wrote:
> > On Fri, Aug 16, 2019 at 01:30:29PM +0100, Chris Wilson wrote:
> > > Quoting Christoph Hellwig (2019-08-16 13:26:44)
> > > > Please, report the actual problem.  Blindly reverting a patch without
> > > > even an explanation of your regressions is not the way to do it.
> > >
> > > As stated, the system doesn't suspend.
> > >
> > > If you would like to wait, you will get test results from our CI
> > > giving the current failed state and the outcome of the patch.
> >
> > Platform type, SSD vendor and type, firmware version?

Which platform were you interested in, and is that information not
present in the debug log? The issue is observed across multiple different
SSD and vendors.
 
> Also not a fan of knee-jerk reverts. Even if it may turn out to be
> necessary, let's at least start with a bug report for an opportunity
> to fix first!

I just did report that we successfully bisected the earlier bug we
reported, and was testing a revert for our CI.

> Could you please try Rafael's solution? These two commits here:
> 
> https://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm.git/commit/?h=xps13-9380-20190812&id=accd2dd72c8f087441d725dd916688171519e4e6
> https://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm.git/commit/?h=xps13-9380-20190812&id=4eaefe8c621c6195c91044396ed8060c179f7aae

Which indeed work, thank you.
-Chris