svn commit: r274040 - in head/sys: conf dev/agp

[Konstantin Belousov]

On Mon, Nov 03, 2014 at 07:50:15PM +0100, Tijl Coosemans wrote:
> On Mon, 3 Nov 2014 11:49:36 +0000 (UTC) Konstantin Belousov <kib at FreeBSD.org> wrote:
> > Author: kib
> > Date: Mon Nov  3 11:49:35 2014
> > New Revision: 274040
> > URL: https://svnweb.freebsd.org/changeset/base/274040
> > 
> > Log:
> >   agp_intel.c provides support for AGP on Intel chipsets from 440/BX to
> >   875.  This intersects with the agp_i810.c, which supports all Intels
> >   from i810 to Core i5/7.  Both agp_intel.c and agp_i810.c are compiled
> >   into kernel when device agp is specified in config, and agp_i810
> >   attach seems to be selected by chance due to linking order.
> 
> I think agp_intel are the plain AGP chipsets while agp_i810 are the
> integrated graphics.  None of the device IDs you deleted from agp_intel
> appear in agp_i810.

Hm, I am not aware of any i810 variants without IGP.  The commit above
sit in my local drm branch for long time, and was based on actual report
by user who did experienced problems due to the agp_intel taking over
aperture, instead of agp_i810.

I see your point, and for now I reverted the pciid part of the change,
274084.

I still think that removing agp_intel.c from amd64 static agp compilation
is right thing to do.  AFAIR EMT64-capable Pentium 4 (Prescotts ? I do not
remember anymore) were compatible with 865/875 series, but it hardly
matters.