Bug 133503

Summary:	No room for fake Host-to-PCI bridge
Product:	[openSUSE] SUSE Linux 10.1	Reporter:	Andreas Schwab <schwab>
Component:	X.Org	Assignee:	Egbert Eich <eich>
Status:	RESOLVED INVALID	QA Contact:	Stefan Dirsch <sndirsch>
Severity:	Normal
Priority:	P2 - High	CC:	eich, rw, sndirsch
Version:	unspecified
Target Milestone:	---
Hardware:	IA64
OS:	Other
Whiteboard:
Found By:	Other	Services Priority:
Business Priority:		Blocker:	---
Marketing QA Status:	---	IT Deployment:	---
Attachments:	Xorg.0.log Logfile of a good server (10.0)

Description Andreas Schwab 2005-11-11 17:20:59 UTC

X doesn't start.

Comment 1 Andreas Schwab 2005-11-11 17:21:37 UTC

Created attachment 57137 [details]
Xorg.0.log

Comment 3 Andreas Schwab 2005-11-12 17:17:03 UTC

Egbert should be able to test it.

Comment 4 Stefan Dirsch 2005-11-14 21:52:47 UTC

xc/programs/Xserver/hw/xfree86/os-support/bus/zx1PCI.c: xf86PostScanZX1()
[...]
    /* Add the fake bus' host bridge */
    if (++idx >= MAX_PCI_DEVICES)
        FatalError("HP ZX1:  No room for fake Host-to-PCI bridge\n");

xc/programs/Xserver/hw/xfree86/os-support/bus/Pci.h: 
[...]
#define MAX_PCI_DEVICES 128       /* Max number of devices accomodated */

I remember this value having increased from 64 to 128 in the past.

Comment 5 Stefan Dirsch 2005-11-16 08:11:48 UTC

Could you figure out, what's the value of idx in your case?

Comment 6 Andreas Schwab 2005-11-21 12:28:43 UTC

See the log file.

$ grep -c 'PCI.*chip' Xorg.0.log
128

Comment 7 Andreas Schwab 2005-11-21 12:31:28 UTC

Created attachment 57866 [details]
Logfile of a good server (10.0)

Comment 8 Stefan Dirsch 2005-11-21 13:43:52 UTC

Thanks. So the PCI devices with BusID > c0 are bogus (duplicates). I thought you had some new crazy hardware with more than 128 devices. I ddn't look closer at the Vendor/Device IDs.

Comment 9 Stefan Dirsch 2005-11-21 15:39:05 UTC

Egbert will take care of this.

Comment 10 Egbert Eich 2005-11-24 17:34:49 UTC

Andreas, before I drag this bug bulky thing out from underneath the table and trun it on killing my ears - can't you give me a box where I can log in and test this?
Preferrably with a built tree somewhere?

Comment 11 Andreas Schwab 2005-11-24 17:56:58 UTC

You should have one yourself.  It happens on the HP zx2000.

Comment 12 Egbert Eich 2005-11-25 07:51:45 UTC

Andreas, please reread #10 for reason for asking. If you confirm that you have *no* zx2000 available I will try mine. 
Furthermore: currently I have SLES9 installed. If I need 10.1 for testing I may not be able to test this before January.

Comment 13 Andreas Schwab 2005-11-25 15:19:46 UTC

I'm sorry for wasting your time, but this is actually the effect of the patch from bug 132308.  I'll move this issue to the other bug.