[Intel-gfx] [PATCH v2] drm/i915: Invoke BXT _DSM to enable MUX on HP Workstation laptops

[Ville Syrjälä]

On Mon, Apr 26, 2021 at 07:10:06PM +0800, Kai-Heng Feng wrote:
> On Fri, Apr 23, 2021 at 8:41 PM Ville Syrjälä
> <ville.syrjala at linux.intel.com> wrote:
> >
> > On Fri, Apr 23, 2021 at 12:46:54PM +0800, Kai-Heng Feng wrote:
> > > On HP Fury G7 Workstations, graphics output is re-routed from Intel GFX
> > > to discrete GFX after S3. This is not desirable, because userspace will
> > > treat connected display as a new one, losing display settings.
> > >
> > > The expected behavior is to let discrete GFX drives all external
> > > displays.
> > >
> > > The platform in question uses ACPI method \_SB.PCI0.HGME to enable MUX.
> > > The method is inside the BXT _DSM, so add the _DSM and call it
> > > accordingly.
> > >
> > > I also tested some MUX-less and iGPU only laptops with the BXT _DSM, no
> > > regression was found.
> > >
> > > v2:
> > >  - Forward declare struct pci_dev.
> > >
> > > Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/3113
> > > References: https://lore.kernel.org/intel-gfx/1460040732-31417-4-git-send-email-animesh.manna@intel.com/
> > > Signed-off-by: Kai-Heng Feng <kai.heng.feng at canonical.com>
> > > ---
> > >  drivers/gpu/drm/i915/display/intel_acpi.c | 17 +++++++++++++++++
> > >  drivers/gpu/drm/i915/display/intel_acpi.h |  3 +++
> > >  drivers/gpu/drm/i915/i915_drv.c           |  5 +++++
> > >  3 files changed, 25 insertions(+)
> > >
> > > diff --git a/drivers/gpu/drm/i915/display/intel_acpi.c b/drivers/gpu/drm/i915/display/intel_acpi.c
> > > index 833d0c1be4f1..c7b57c22dce3 100644
> > > --- a/drivers/gpu/drm/i915/display/intel_acpi.c
> > > +++ b/drivers/gpu/drm/i915/display/intel_acpi.c
> > > @@ -14,11 +14,16 @@
> > >
> > >  #define INTEL_DSM_REVISION_ID 1 /* For Calpella anyway... */
> > >  #define INTEL_DSM_FN_PLATFORM_MUX_INFO 1 /* No args */
> > > +#define INTEL_DSM_FN_PLATFORM_BXT_MUX_INFO 0 /* No args */
> > >
> > >  static const guid_t intel_dsm_guid =
> > >       GUID_INIT(0x7ed873d3, 0xc2d0, 0x4e4f,
> > >                 0xa8, 0x54, 0x0f, 0x13, 0x17, 0xb0, 0x1c, 0x2c);
> > >
> > > +static const guid_t intel_bxt_dsm_guid =
> > > +     GUID_INIT(0x3e5b41c6, 0xeb1d, 0x4260,
> > > +               0x9d, 0x15, 0xc7, 0x1f, 0xba, 0xda, 0xe4, 0x14);
> > > +
> >
> > I think this dsm is just supposed to be more or less an
> > alternative to the opregion SCI stuff. Why there are two
> > ways to do the same things I have no idea. The opregion
> > spec does not tell us such mundane details.
> 
> Right now I think it's HP specific and from what I can see it doesn't
> touch opregion.

It's part of the opregion spec.

> 
> >
> > It's also not documented to do anything except list the
> > supported functions:
> > "Get BIOS Data Functions Supported “Function #0"
> >  This function can be called to discover which “_DSM” Functions are
> >  supported. It may only return success if the return value accurately
> >  lists supported Functions."
> >
> > But what you're apparently saying is that calling this changes
> > the behaviour of the system somehow? That is troubling.
> 
> It flips a bit in BIOS-reserved Intel GPIO, and EC/hardware will
> change the MUX based on the GPIO bit.
> 
> We can add a DMI check to match "HP" to minimize the potential
> regression factor.

I'm rather thinking that calling it always may be the right thing to do,
assuming Windows does it as well. Maybe more vendors use it to backdoor in
random junk like this :(

-- 
Ville Syrjälä
Intel