Re: X problem on PMAGB-B

David Brownlee <[email protected]> Mon, 17 Nov 2025 19:11:33 +0000
Newsgroups gmane.os.netbsd.ports.mips.devel
Message-ID <CAGN_6pauj8Cef_ruqrADHZW8U1R1yp+RkXv=xgvT+N1QtZ4niw@mail.gmail.com>
On Mon, 17 Nov 2025 at 10:50, Michael <[email protected]> wrote:
>
> Hello,
>
> On Sun, 16 Nov 2025 23:39:18 -0500
> Dave McGuire <[email protected]> wrote:
>
> >    But there is a strange offset to the video.  I've attached two pics
> > which show what I mean.  See where my pointer has grabbed the window
> > title bar in the second picture; I've actually grabbed the window and am
> > dragging it, so that's where the X server thinks the window is.  But the
> > actual displayed objects are shifted to the left.
>
> Hmm, sfb_common_init() adds another 4096 to the VRAM address, let's do
> the same:
>
> Index: sfb.c
> ===================================================================
> RCS file: /cvsroot/src/sys/dev/tc/sfb.c,v
> retrieving revision 1.89
> diff -u -w -r1.89 sfb.c
> --- sfb.c       6 Dec 2021 16:00:07 -0000       1.89
> +++ sfb.c       17 Nov 2025 10:45:52 -0000
> @@ -485,9 +485,9 @@
>  {
>         struct sfb_softc *sc = v;
>
> -       if (offset >= SFB_SIZE || offset < 0)
> +       if (offset >= SFB_FB_SIZE || offset < 0)
>                 return (-1);
> -       return machine_btop(sc->sc_vaddr + offset);
> +       return machine_btop(sc->sc_vaddr + SFB_FB_OFFSET + 4096 + offset);
>  }

Updated builds with the + 4096 at https://sync.absd.org/pmax/  -
should only need to update the kernel :)

David