[Intel-gfx] [PATCH 3/4] drm/i915: Flatten DRIVER_MODESET checks in i915_irq.c

[Imre Deak]

Hi Dave,

On to, 2015-02-19 at 15:42 +0200, Imre Deak wrote:
> On to, 2015-02-19 at 15:39 +0200, Imre Deak wrote:
> > On to, 2015-02-19 at 12:25 +0000, Dave Gordon wrote:
> > > On 12/02/15 22:38, Imre Deak wrote:
> > > > On Tue, 2015-02-03 at 11:30 +0100, Daniel Vetter wrote:
> > > >> UMS is no more!
> > > >>
> > > >> Signed-off-by: Daniel Vetter <daniel.vetter at intel.com>
> > > 
> > > Some machines now won't boot in "recovery mode", which specifies
> > > "nomodeset" and therefore results in various important bits of code not
> > > being executed. Will we eventually ignore "modeset" completely, or just
> > > refuse to load at all if "nomodeset" is explicitly specified?
> > 
> > The driver will already refuse to load with nomodeset for GEN6+ for
> > quite some time now. On old platforms UMS would still work before this
> > patch, but afaik there was a decision to stop supporting UMS. Note that
> > this doesn't mean "recovery mode" or equivalently nomodeset will break
> > booting, it just means user space will fall back to vesa/vga or text
> > mode.
> 
> Ah, or did you mean after this patch we should refuse loading the driver
> in case of nomodeset even for old platforms? That would make sense
> indeed.

I was wrong here, I was thinking only about the GEN6 MODESET check in
i915_driver_load. As Daniel pointed out on IRC in addition to that we
also silently fail to load driver in i915_init for !MODESET always,
regardless of the platform, so the check in i915_driver_load is
redundant. Based on this it's safe to remove the !MODESET parts.

--Imre