[Intel-gfx] [PATCH 3/6] drm/i915: Add IS_POULSBO and IS_MRST helper macros

[Richard Purdie]

On Thu, 2009-05-14 at 13:37 -0700, Jesse Barnes wrote:
> On Wed, 13 May 2009 15:02:59 +0100
> Richard Purdie <rpurdie at linux.intel.com> wrote:
> 
> > Add IS_POULSBO and IS_MRST helper macros to i915_drv.h, then use
> > these to enable the correct code paths for these pieces of hardware
> > in the common VDC code.
> > 
> > Signed-off-by: Richard Purdie <rpurdie at linux.intel.com>
> > 
> > Index: git/drivers/gpu/drm/i915/i915_drv.h
> > ===================================================================
> > --- git.orig/drivers/gpu/drm/i915/i915_drv.h	2009-05-13
> > 13:13:00.000000000 +0100 +++
> > git/drivers/gpu/drm/i915/i915_drv.h	2009-05-13
> > 13:22:13.000000000 +0100 @@ -691,17 +691,23 @@ #define IS_IGDGM(dev)
> > ((dev)->pci_device == 0xa011) #define IS_IGD(dev) (IS_IGDG(dev) ||
> > IS_IGDGM(dev)) 
> > +#define IS_POULSBO(dev) (((dev)->pci_device == 0x8108) || \
> > +			 ((dev)->pci_device == 0x8109))
> > +
> > +#define IS_MRST(dev) (((dev)->pci_device & 0xfffc) == 0x4100)
> > +
> >  #define IS_G33(dev)    ((dev)->pci_device == 0x29C2 ||	\
> >  			(dev)->pci_device == 0x29B2 ||	\
> >  			(dev)->pci_device == 0x29D2 ||  \
> >  			(IS_IGD(dev)))
> >  
> >  #define IS_I9XX(dev) (IS_I915G(dev) || IS_I915GM(dev) ||
> > IS_I945G(dev) || \
> > -		      IS_I945GM(dev) || IS_I965G(dev) || IS_G33(dev))
> > +		      IS_I945GM(dev) || IS_I965G(dev) || IS_G33(dev)
> > || \
> > +		      IS_POULSBO(dev))
> >  
> >  #define IS_MOBILE(dev) (IS_I830(dev) || IS_I85X(dev) ||
> > IS_I915GM(dev) || \ IS_I945GM(dev) || IS_I965GM(dev) || IS_GM45(dev)
> > || \
> > -			IS_IGD(dev))
> > +			IS_IGD(dev) || IS_POULSBO(dev))
> >  
> >  #define I915_NEED_GFX_HWS(dev) (IS_G33(dev) || IS_GM45(dev) ||
> 
> I guess you've tested these so we know they work.  I just wonder if it
> would be better to have IS_9XX_VDC and IS_965_VDC macros instead (not
> that I can think of any cases where that would be necessary vs. what's
> here).

Yes, this is known to work. Its really just a question of whether we
need two sets of similar macros. I can't see a case for it at present
and I didn't want to add complexity we don't need...

-- 
Richard Purdie
Intel Open Source Technology Centre