Re: New XFree86-3DLabs-3.3.6 RPM available!

On Mon, Feb 24, 2003 at 09:04:05AM -0600, Tom 'spot' Callaway wrote:
[...]
> So, the Raptor is a TI produced card? This is very much a sledgehammer
> fix (although, its the right fix).

I never pretended that it was an *elegant* fix... ;-)


> Does the Raptor have a PCI_CHIP entry that differs from the TI_PERMEDIA,
> or is it using the TI_PERMEDIA? If it (the PCI_CHIP entry) is not
> unique, then the attached patch is a more appropriate fix (and yes, I'm
> being semantic on a dead branch of XFree86 that we're dragging along
> instead of doing driver ports ;)

See, that's where you have more experience than I do... <g>
With regard to PCI_CHIP: That's how I figured out my "fix" in the
first place - I added statements to the server to show me what is
actually returned for pcrp->_vendor and pcrp->_device. The vendor code
was the same as PCI_VENDOR_TI (0x104C), while the device code was the same
as PCI_CHIP_TI_PERMEDIA2 (0x3d07). Hence the idea of trying to treat
it the same as the 3DLabs version. Question is (I honestly don't
know): Isn't there an (almost) identlical ix86 based PCI card with the
same chipset? In that case it would make sense that the same codes get
returned. I also was wondering whether the difference wrt
PCI_CMD_IO_ENABLE might be something related to the Sparc architecture
(its PCI behaving differently or suchlike), but I'm lacking the
background to be able to say that. I *did* find it interesting that
the PCI_CMD_IO_ENABLE test was commented out already for
PCI_VENDOR_3DLABS and I did wonder why the two chipsets wouldn't
behave the same (after all, both are Permedia 2), despite being from
different manufacturers.


> The fix that you've proposed effectively forks XFree86 3.3.6, which is
> something I'm hesitant to do. If XFree86 3.3.6 was current, we'd be
> dragging our own branch forward, since they'd never take this patch (it
> undoubtedly breaks the Permedia 2). The patch I've attached would be
> likely be accepted, since it would only affect Linux/SPARC builds of
> XFree86 3.3.6.

Ah! Ok, admittedly, I didn't spend thought on this, but I can see the
logic behind your statement. I did change the spec file so that the
patch only gets applied for Sparc, but I didn't think it through...
Makes sense to have the code behave the same way.
New RPMs are already available at:

http://www.ribbrock.org/Aurora/1.0/XFree86-3.3.6/


> Please, don't misconstrue this response as harsh.
[...]

Heck no, why would I? If I was able to be at least of some help,
that's plenty for me - I'm just happy that I can contribute at least a
little bit (now that I finally got that U5/400 swapped for a working
specimen... ;-) ). The upstream issue is something that hasn't crossed
my mind, but I'm always open to new insights, so you feedback is
appreciated. :-)


> The compat-modules dep was something I assumed was erroneous anyways on
> sparc (since it exploded in about 200 places trying to build), your spec
> fix looks fine.

The only thing I'm wondering about is: The 3DLabs server isn't the
only one that has that requirement. The others probably should be
changed as well. Did you ever try to find out which frame buffers are
actually working properly with XFree86 4.x on Sparc? I'm asking 'cause
I'm curious on what base you made the selection of 3.3.6 servers.
Here's what I can understand so far:

3DLabs:
  - Raptor GFX8P/PGX32

Mach64:
  - PGX8/PGX24, I presume - I have no experience with those (my U5 has a
    PGX24 onboard, but the PGX32/Raptor is better, so I don't bother using
    it).

Sun24:
  - e.g. S24 and CG14 (aka SS20 VSIMM), right? My SS20 with a 4MB VSIMM
    works perfectly and I had a SS5/170 with a S24 running X with no
    problems under Aurora 0.32 a while ago (still have that box if testing
    is needed).
    
Sun:
  - Hm - CG6 and CG3? CG6 most certainly works.

SunMono:
  - BG1 (?) - no idea

FBDev:
  - ???

There are quite a few missing (Creator, Elite, ...) - are all of them
working with XFree86 4.x? (just curious)

Cheerio,

Thomas
-- 
-----------------------------------------------------------------------------
                Thomas Ribbrock    http://www.ribbrock.org 
  "You have to live on the edge of reality - to make your dreams come true!"