Bug 133236

Summary: (EE) No devices detected.
Product: [openSUSE] SUSE Linux 10.1 Reporter: Andreas Schwab <schwab>
Component: SaX2Assignee: Marcus Schaefer <ms>
Status: RESOLVED FIXED QA Contact: Stefan Dirsch <sndirsch>
Severity: Normal    
Priority: P5 - None CC: ms, rw, sndirsch, susedev
Version: Alpha 2plus   
Target Milestone: ---   
Hardware: IA64   
OS: Other   
Whiteboard: ia64-graphics
Found By: Other Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---
Attachments: xorg.conf from sax2 -m 0=radeon -l
Xorg.0.log
Working X config

Description Andreas Schwab 2005-11-10 10:45:02 UTC
This is xorg from STABLE with the patch from bug 132308.
Comment 1 Andreas Schwab 2005-11-10 10:46:59 UTC
Created attachment 56957 [details]
xorg.conf from sax2 -m 0=radeon -l
Comment 2 Andreas Schwab 2005-11-10 10:47:41 UTC
Created attachment 56958 [details]
Xorg.0.log
Comment 3 Stefan Dirsch 2005-11-10 10:57:51 UTC
Thanks for testing.

*** This bug has been marked as a duplicate of 132308 ***
Comment 4 Andreas Schwab 2005-11-10 12:10:01 UTC
This is a SaX bug, X -configure produces a working configuration.
Comment 5 Andreas Schwab 2005-11-10 12:11:08 UTC
Created attachment 56966 [details]
Working X config
Comment 6 Stefan Dirsch 2005-11-10 14:33:28 UTC
I would call it an up-to-now unknown new feature, which is currently missing.

BusID       "PCI:1@5:0:0"

seems to be the trick.
Comment 7 Raymund Will 2005-11-10 15:37:04 UTC
BTW, 'lspci' shows

  0005:01:00.0 VGA compatible controller: 
     ATI Technologies Inc R420 JM [FireGL X3] (rev 80)

for this device...
Comment 8 Marcus Schaefer 2005-11-14 14:06:04 UTC
sorry something like

  BusID       "PCI:1@5:0:0"

cannot be handled by sax without an understanding what this crappy
format is all about ??
Comment 9 Raymund Will 2005-11-14 14:39:52 UTC
See comment #7: PCI domains appear to be the problem -- or better,
the lack of support for them in X/sax ...
(I'd guess it means bus#1 of domain#5 and suppose, that "this crappy
format" comes from some PCI-specific code of X.org.)
Comment 10 Marcus Schaefer 2005-11-14 14:56:36 UTC
Ok I changed the parser rule for the bus id setup but I still don't
understand the semantic of the setup ? on an ix86 architecture I have
something like:

   01:00.0 Class 0300: 10de:0309 (rev a2)
 

 BusID 1:0:0

how to express that in the new syntax:
 
 BusID 1@1:0:0  ?

an explanation or specification would be usefull
Comment 11 Raymund Will 2005-11-14 15:08:51 UTC
For me it looks like
  BusID 1:0.0
           ^ watch out!
is the usual 'Bus' 'Device' 'Subdevice' thing -- which is extended by
a leading "PCI:" to that new format
  BusID "PCI:1@0:0:0"
                  ^ strange!
IMHO it should have been kept consistent with 'lspci' (see man-page)
and look more like
  BusID 0:1:0.0
(with or without leading zeros and consistently decimal *or* hex, but
not mixed, as it appears to be in Xorg.*.log!)

BTW, I'd expect to see those PCI domains on newer/bigger ix86 (maybe AMD64)
machines as well, so I guess it isn't really IA64-only after all.
Comment 12 Egbert Eich 2005-11-14 15:48:43 UTC
(In reply to comment #8)
> sorry something like
> 
>   BusID       "PCI:1@5:0:0"
> 
> cannot be handled by sax without an understanding what this crappy
> format is all about ??
> 

Please calm down. 
PCI:<bus>@<domain>:<device>:<function>.
<domain> is the PCI domain.
Comment 13 Marcus Schaefer 2005-11-14 17:40:25 UTC
has been fixed now. I added the domain concept to sax2 as well
Comment 14 Gerald Pfeifer 2005-11-14 18:51:47 UTC
Cool, thanks a lot for the ultra-quick response, Marcus!