Re: removing the agp(4) driver

From: Cy Schubert <Cy.Schubert_at_cschubert.com>
Date: Sun, 06 Apr 2025 04:13:02 UTC
In message <20250317070241.36571d8c632f2caa6a733dc1@dec.sakura.ne.jp>, 
Tomoaki
AOKI writes:
> On Sun, 16 Mar 2025 21:20:19 +0300
> Artem Bunichev <tembun@bk.ru> wrote:
>
> > > Just a FYI:
> > > It could be because even MINIMAL kernel configuration has
> > > 
> > >  device		agp		# support several AGP chipsets
> > > 
> > > line.
> > 
> > Yes, I noticed that both GENERIC and MINIMAL kernels do have the agp(4) dri
> ver
> > included.  But as I can see, it doesn't make sense to compile this driver i
> f
> > your machine doesn't need it (I mean, no actual hardware for the driver;
> > `dmesg |grep agp' outputs nothing on my ThinkPad X220).  As I understand fr
> om
> > Emmanuel Vadot's message, the problem can be solved with conditional
> > compilation, but I just want to sort of bump it, because I found that it
> > hasn't been solved yet.  I think that it's better to do this before the dri
> ver
> > will actually be removed.
> > 
> > Artem.
>
> 100% agree with you.
> But as graphics/drm-*-kmod are ported from Linux, I suspect if Linux
> still support agp even on amd64 (would be nonsense, though) and ask agp
> driver whether there's any GPU or not, agp driver would be needed just
> to reply "there's none!".
>
> Not read the codes, so I would be wrong. But just a possible reason to
> keep it.

Coming late to the party, my base kernel config contains,

fgit@slippy$ ag -s agp
amd64/KOMQUATS
59:nodevice  agp         # support several AGP chipsets
fgit@slippy$ 

None of my machines use it and kldstat shows it not loaded.


-- 
Cheers,
Cy Schubert <Cy.Schubert@cschubert.com>
FreeBSD UNIX:  <cy@FreeBSD.org>   Web:  https://FreeBSD.org
NTP:           <cy@nwtime.org>    Web:  https://nwtime.org

			e^(i*pi)+1=0