Re: [PATCH v2] platform/x86: intel-hid: Add support for Device Specific Methods

[Darren Hart <dvhart@infradead.org>]

On Mon, Jul 09, 2018 at 04:38:16AM +0000, Mario.Limonciello@dell.com wrote:
> > -----Original Message-----
> > From: Srinivas Pandruvada [mailto:srinivas.pandruvada@linux.intel.com]
> > Sent: Saturday, July 7, 2018 8:44 AM
> > To: Darren Hart; Andy Shevchenko
> > Cc: Limonciello, Mario; Alex Hung; Andy Shevchenko; Platform Driver; Linux Kernel
> > Mailing List; Rafael J. Wysocki
> > Subject: Re: [PATCH v2] platform/x86: intel-hid: Add support for Device Specific
> > Methods
> > 
> > On Fri, 2018-07-06 at 16:59 -0700, Darren Hart wrote:
> > > On Mon, Jul 02, 2018 at 05:06:14PM +0300, Andy Shevchenko wrote:
> > > > On Mon, Jul 2, 2018 at 4:51 PM,  <Mario.Limonciello@dell.com>
> > > > wrote:
...
> > Mario can add more.
> > But I think Dell has released a BIOS fix, so that power button can
> > still work using non _DSM way. So I think we can wait for normal
> > release cycle.
> 
> Yeah, I added some comments on a separate reply already indicating this.
> 
> The affected system we know about has reverted the new interface.
> 
> This all stems from an expectation that the _DSM has been there for "many"
> releases on the Windows side.  Long enough that even all the corporate downgrade
> scenarios to older Windows versions and the drivers with them all worked properly.
> 
> If this doesn't end up being a candidate for backporting to stable we'll probably
> end up asking our various OS partners to backport it as a SAUCE type patch in distro
> kernels to eventually be able to turn this back on.

Thanks Mario. We'll skip the RC. We can discuss -stable as a separate
issue.  This patch fixes a demonstrable bug, the biggest obstacle here
is the size. At 240 lines with context, it more than doubles the maximum
patch size for stable.

My recommendation is that we treat the stable backport of this as a
special case, e.g. Andy and I don't tag it, but you or Srinivas propose
it and specifically make the case for why an exception should be made to
Greg.

How self contained and isolated this change is will be an important part
of the argument. On the one hand, it's all one file, on the other hand,
this is a fairly generic driver in use in more and more platforms, with
the potential for widespread impact if this introduces a new bug.

-- 
Darren Hart
VMware Open Source Technology Center