Re: Latest evbmips/LOONGSON (Lemote Yeeloong) observations

[Michael <macallan%netbsd.org@localhost>]

Hello,

On Thu, 23 Jul 2015 11:30:59 +1000
matthew green <mrg%eterna.com.au@localhost> wrote:

> > > (==) Using default built-in configuration (12 lines)
> > > (EE) Failed to load /usr/X11R7/lib/modules/drivers/siliconmotion_drv.so: /usr/X11R7/lib/modules/drivers/siliconmotion_drv.so: Undefined symbol "vgaHWddc1Se
> > tSpeed" (symnum = 33)
> > 
> > This one is trivial to fix, we don't need or want the vgahw module on MIPS.
> > Heck, the SM50x isn't even VGA compatible.
> > 
> > > Is this yet another example of the "siliconmotion" driver being built
> > > for PC hardware when the target hardware is anything but?
> > 
> > Yes and no. Some newer SiliconMotion chips do have VGA compatibility
> > goo, and if we were using VGA text mode on them then the vgahw module
> > would actually be useful. We don't though.
> > 
> > I'm going to fix that in the next few days.
> 
> FWIW, the siliconmotion driver itself seems to actually depend on the
> vgaHW module on purpose.  there are many refs in smi_driver.c,
> smilynx_crtc.c and smilynx_hw.c (and others.)

Of course it's on purpose. Doesn't do anything useful though.

> it might just be easier to make it provided for linkage, even if
> those functions won't actually be called.

On r128 I just #ifdefed them away.

have fun
Michael