Re: [PATCH][v4.2.y-ckt]ACPI / LPSS: allow to use specific PM domain during ->probe()

[Greg KH <gregkh@linuxfoundation.org>]

On Mon, May 09, 2016 at 02:21:15PM +0300, Andy Shevchenko wrote:
> On Sun, 2016-05-08 at 20:04 -0400, Joseph Salisbury wrote:
> > On 05/08/2016 01:11 PM, Greg KH wrote:
> > > 
> > > On Fri, May 06, 2016 at 10:05:52AM -0400, Joseph Salisbury wrote:
> > > > 
> > > > Hello,
> > > > ����
> > > > Please consider including mainline commit
> > > > de16d55206c5ce82ad44590b01718fbe2323853c in the next v4.2.y-ckt
> > > > release.��It was
> > > > included in the mainline tree as of v4.5-rc1.��It has been tested
> > > > and
> > > > confirmed
> > > > to resolve http://bugs.launchpad.net/bugs/1549354 .
> > > > ����
> > > > commit de16d55206c5ce82ad44590b01718fbe2323853c
> > > > Author: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
> > > > Date:���Fri Dec 4 23:49:19 2015 +0200
> > > > 
> > > > ����ACPI / LPSS: allow to use specific PM domain during ->probe()
> > > > 
> > > It doesn't apply to 4.4-stable, why would it apply to 4.2?
> > > 
> > > confused,
> > > 
> > > greg k-h
> > Hi Greg,
> > 
> > Commit de16d55 is an amendment to commit 01ac170, per the commit
> > message. Commit 01ac170 landed in mainline as of 3.19-rc1 and commit
> > de16d55 landed in mainline as of v4.5-rc1. However, neither commits
> > were
> > cc'd to stable.
> > 
> > This request for 4.2.y was to prevent a revert of a revert due to a
> > bug[0].��In 4.2.y-ckt(And in mainline), commit 01ac170 was reverted by
> > commit 05c7343(Commit b5f88dd in mainline).��This introduced bug
> > 1549354[0] in Ubuntu.��This bug reports that the Elantech touchpad
> > stopped working.��
> > 
> > Commit 01ac170 was not reverted in v4.4.y, so it shouldn't have
> > introduced the same bug.��The commit message for b5f88dd says: "Fixes:
> > 3df2da968744",��but it was not cc'd to stable, so I'm not sure if you
> > plan on including it in 4.4.y.��If you do, commit 01ac170 will be
> > needed to prevent the bug.
> > 
> > I also should have mentioned commit de16d55 will cause a build failure
> > because it requires commit 14b6257 as a prerequisite.��So if you
> > include commit b5f88dd1d
> > in 4.4.y, you will need commits de16d55 and 14b6257a5 to prevent the
> > indicated bug.
> 
> It might be not the best idea to cherry-pick only those two since
> culprit commit in your case is a part of a bigger fix, i.e.�
> commit eebb3e8d8aaf ("ACPI / LPSS: override power state for LPSS DMA
> device"). So, it would be better to have that fix to land in the kernel
> if you care.

This is crazy.  I think I've taken that last patch, so I'm going to
leave 4.4-stable alone for now unless someone tells me I've messed
something up...

thanks,

greg k-h