*BSD News Article 8313


Return to BSD News archive

Xref: sserve comp.unix.sysv386:25913 comp.unix.bsd:8366 comp.windows.x:47921
Newsgroups: comp.unix.sysv386,comp.unix.bsd,comp.windows.x
Path: sserve!manuel.anu.edu.au!munnari.oz.au!network.ucsd.edu!usc!zaphod.mps.ohio-state.edu!pacific.mps.ohio-state.edu!linac!att!cbnewse!cbnewsd!att-out!cbnewsj!dwex
From: dwex@cbnewsj.cb.att.com (david.e.wexelblat)
Subject: Re: I need 8514/ATI Graphics Ultra support for ISC UNIX SVR3.2 & X11R4
Organization: AT&T
Date: Tue, 24 Nov 1992 17:29:15 GMT
Message-ID: <1992Nov24.172915.25813@cbnewsj.cb.att.com>
References: <1992Nov16.213215.4701@netcom.com> <1992Nov23.073800.6592@Informatik.TU-Muenchen.DE> <1eshe9INNeum@agate.berkeley.edu>
Lines: 52

In article <1eshe9INNeum@agate.berkeley.edu> curtis@cs.berkeley.edu (Curtis Yarvin) writes:
> In article <1992Nov23.073800.6592@Informatik.TU-Muenchen.DE> roell@informatik.tu-muenchen.de (Thomas Roell) writes:
> >>   >We are using the Metrolink X11R4 software with Motif 1.1.3 on ISC 3.0.
> >>   >They support the ATI ultra very well. For information call
> >>   >
> >>   >(305) 970-7353
> >>   >
> >>   >They also have X11R5 with Motif 1.1.4 available.
> >>   >
> >>   It is trivial to port XBSD8514 to ISC.
> >>   XBSD8514 is known to work with the ATI Ultra and has been clocked
> >>   at 60k xstones.
> >
> >Well, the issue here is that MetroLink (and SGCS) have access to the
> >full documents about the Mach8/Mach32 chipsets.
> 
> Mach8 documentation is now publicly available.
> 

This is interesting.  How does one get it?

> Roell has, in the past, claimed that SysV use of X8514 requires kernel
> mods; this is also false.  SVR3 has a call to enable user-level port
> access, and I believe SVR4 has it too.
> 
> c

No.  What Thomas has claimed is that SVR4 needs kernel mods to work correctly
with devices that use I/O portts above 0x3FF, and this is in fact the case.
Both SGCS and MetroLink provide patches to make it work.  ISC does not need
the patch.  Basically, SVR4 has the same mechanism as SVR3 to get the 
I/O port access.  But if your process receives a signal, it loses the
permission.  The patch makes the kernel leave the permissions alone.  We
are beta-testing our own version of this patch for XFree86, as we have
a couple of people looking at high-end monochrome boards, like the Sigma
Designs boards, and these use extended ports.

This bug is reportedly fixed in SVR4.2.

Another little caveat - the back door that allows the I/O access is
an undocumented part of the VPIX support code.  Since this is a conditional
compilation area, a kernel that is built without VPIX support will NOT
be able to access extended I/O ports without driver support.  Of course,
I can't imagine that the kernel would be built without VPIX support, but
it is possible.

--
David Wexelblat <dwex@mtgzfs3.att.com>  (908) 957-5871
AT&T Bell Laboratories, 200 Laurel Ave - 4B-421, Middletown, NJ  07748

"The meaning of life?  That's simple.  Try to be happy, try not to hurt
 other people, and hope to fall in love."  -- Mallory Keaton