[ale] Re: DiamondStealth and frame buffer...

[Robert L. Harris]

Harald Koenig wrote:

> On Aug 18, Robert L. Harris wrote:
>
> >   I have a AHA3940UW scsi controller in my machine now.  I also have
> a
> > Diamond Stealth VRam (3000 series) card.  Since I put the 3940 in I
> can
> > not run X.  THIS IS NOT AN X PROBLEM  NO FLAMES PLEASE.
>
> it might be a X problem (well, there is a hardware bug in the S3 968
> chip
> which can show such symptoms for some configurations with SCSI
> controllers).
>
> make sure you're using XFree86 3.3 or newer (3.3.1 is recent)
> which works around this bug!  if you're already using this version,
> please send the output of "cat /proc/pci" after reboot plus
> "X -probeonly" which might not crash the system...

I've been running MetroX for quite some time.  I couldn't find any
documentation on the S3 chip.  I dug it out of DaimondMM today.
I've notified Adaptec of this cutie so when people go digging for info
in the future, they'll know about it.  I talked to a guy on the
linux-kernel
list who explained who/why it's an X problem and not kernel based
instead
of just sending me a messages saying it's an X problem.

> > Adaptec says it's a problem with the video card.  Diamond admits
> it's a
> > problem with the video card.  They say there is a "bug" in the
> > framebuffering and if it can be "readjusted" it would solve the
> problem.
> > I don't know how many other people have had this problem and spent
> $$$
> > we don't have to buy new cards thinking the card is toast but I
> can't
> > afford to.  Does anyone know who/how/where I should look about the
> > framebuffering and how trivial it is?  Could this be a reccomended
>
> well, your signature gives the pointer:  man XF86Config ;-)
>
I had "rtfm" and didn't see anything related to this.  That was part
ofthe problem.  It would have been nice if this bug had been publicized
on
the web page or such...  Would have saved alot of time/energy/email...

>
>
> > perl -e 'print
> > $i=pack(c5,(41*2),sqrt(7056),(unpack(c,H)-2),oct(115),10);'
>
> true, but most people usually don't :-(
>
> another usual omission in problem/bug reports (like your's;) is not to
>
> mention any version number of the software and/or drivers being used
> or including debug output.   having that information from your Xserver
>
> I could be more detailed (and could save time typing common blurb)...
>

I had specified XFree86 (aparently I didn't have THE LATEST but what
I was told was the latest by someone usually  reliable) Latest MetroX
and
Accell-X.  There was no output.  Once the server started, it hit that
bug and
the machine and scsi controller locked tight, and nothing was written
after that.

Robert

--
----------------------------------------------------------------------------

Robert L. Harris             |   If NT is the answer,
System Engineer For Hire.    \_    You don't understand the question.

Email:
  Robert at ast.lmco.com
http://www.orci.com/~nomad

DISCLAIMER:
  These are MY OPINIONS ALONE.  I speak for no-one else.

perl -e 'print
$i=pack(c5,(41*2),sqrt(7056),(unpack(c,H)-2),oct(115),10);'