[Intel-gfx] [PATCH] drm/i915/display: split out intel_vga_client.[ch]

[Ville Syrjälä]

On Tue, Oct 01, 2019 at 05:28:51PM +0300, Jani Nikula wrote:
> On Tue, 01 Oct 2019, Ville Syrjälä <ville.syrjala at linux.intel.com> wrote:
> > On Tue, Oct 01, 2019 at 03:12:49PM +0100, Chris Wilson wrote:
> >> Quoting Jani Nikula (2019-10-01 14:43:53)
> >> > Split out code related to vga client and vga switcheroo
> >> > register/unregister and state handling from i915_drv.c and
> >> > intel_display.c.
> >> > 
> >> > It's a bit difficult to draw the line how much to move to the new file
> >> > from i915_drv.c, but it seemed to me keeping i915_suspend_switcheroo()
> >> > and i915_resume_switcheroo() in place was cleanest.
> >> > 
> >> > No functional changes.
> >> > 
> >> > Cc: Ville Syrjälä <ville.syrjala at linux.intel.com>
> >> > Cc: Chris Wilson <chris at chris-wilson.co.uk>
> >> > Signed-off-by: Jani Nikula <jani.nikula at intel.com>
> >> > 
> >> > ---
> >> > 
> >> > It's also a bit fuzzy if this is a sensible split anyway. Could also
> >> > name it intel_vga and move these from intel_display.c there?
> >> 
> >> My initial thought that the switcheroo interface would remain in core,
> >
> > Yeah the switcheroo stuff should perhaps stays with the rest of the pm hooks.
> 
> Okay, so keep all of switcheroo in i915_drv.c, and move all the vgaarb
> stuff (incl the ones I mentioned from intel_display.c) to
> intel_vga.[ch]?

I can get behind that plan.

> 
> >
> >> that it is more of a global power state that we currently just use for
> >> the legacy vga switching.
> >> 
> >> The patch looks fine, on a pure mechanical pov,
> >> Reviewed-by: Chris Wilson <chris at chris-wilson.co.uk>
> >> 
> >> For the sake of argument, could you float the split in the other
> >> direction?
> 
> Please elaborate. Move switcheroo higher in the call chain?
> 
> BR,
> Jani.
> 
> 
> >> 
> >> And maybe Ville has a good opinion on how it is meant to work :)
> >> -Chris
> 
> -- 
> Jani Nikula, Intel Open Source Graphics Center

-- 
Ville Syrjälä
Intel