Subject: Re: toe in the water
To: None <>
From: der Mouse <mouse@Rodents.Montreal.QC.CA>
List: port-next68k
Date: 08/14/1999 21:56:14
>> - What is the state of SCSI support?  [...]
> [...probe, try to do I/O, dropped bytes...]  [W]e still have no
> documentation on the dma chip or how it interacts with the ncr53c90a
> scsi device.

Sigh.  One would think nobody cared about the proprietary aspects of
the black hardware any more...I guess no such luck. :(

> If people are interested, perhaps I should begin a discussion here
> that includes my understanding of the dma hardware.  Someone else
> co-hacking on the dma driver and scsi support will hopefully keep the
> effort from stalling on my slow progress.

Like most of us (I suspect), my amount of free time is normally zero to
negative, so I'm not likely to make spectacular progress either.  But
every now and then I have a day or two with nothing better to do, and
sometimes manage to get bursts of stuff done.  But I have a vested
interest in making SCSI work, even if comparatively slowly, because it
will allow me to finally turf NeXT's OS on my Cube. :)  I've even
considered running it diskless, because its main raison d'etre is
running a mud, and swapping was such a performance hit that I loaded
the machine with enough RAM that it doesn't swap - in which case
running diskless wouldn't be a big deal.  The main reason I don't do
this is that database checkpoints are already slow enough; doing them
over nfs would be execrable.  (Secondary reason is that I don't want to
have to try to bring up NFS and tighten it down enough for me to be
content that it's secure.  The NeXT I'm currently running diskless is
being served off a NetBSD/sun3 machine via a crossed 10baseT cable and
a microtransceiver; neither one is accessible to the net at large.)

>> - At boot time the framebuffer appears as
>> 	nextdisplay0 at mainbus0: 1152 x 832, 2bpp
>>     ...but I know that hardware, and it's really 1120x832.
> #define	VIDEO_W		1120		/* Visible pixels per scanline */
> #define	VIDEO_MW	1152		/* Actual pixels per scanline */

Hm, that's true; wscons will need to know the stride size.  I'm tempted
to offer a patch to print the actual size, though, rather than the
stride size.  Once I have such a patch, of course. :-)

					der Mouse

		     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B