Bug 1962 - ATI Mach8/32/64 probe in ATI driver makes IA64 machine check
Summary: ATI Mach8/32/64 probe in ATI driver makes IA64 machine check
Status: RESOLVED FIXED
Alias: None
Product: xorg
Classification: Unclassified
Component: Driver/mach64 (show other bugs)
Version: 6.8.1
Hardware: PA-RISC (HP) Linux (All)
: high normal
Assignee: Egbert Eich
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on:
Blocks: 7509
  Show dependency treegraph
 
Reported: 2004-11-30 00:08 UTC by Egbert Eich
Modified: 2006-08-25 08:01 UTC (History)
4 users (show)

See Also:
i915 platform:
i915 features:


Attachments
[FIXED_X11R68x] Fix (3.13 KB, patch)
2004-11-30 00:12 UTC, Egbert Eich
roland.mainz: 6.8-branch+
Details | Splinter Review
Proposed patch (1.72 KB, patch)
2005-04-25 13:39 UTC, Kevin E. Martin
no flags Details | Splinter Review

Description Egbert Eich 2004-11-30 00:08:46 UTC
The ATi driver pokes around certain PIO registers to detect Mach8/32/64 non-PCI
hardware - among those 8514 registers. This makes certain legacy free IA64 (for
example HPs zx1) machine check.
These probes should be avoided.
Comment 1 Egbert Eich 2004-11-30 00:12:06 UTC
Created attachment 1423 [details] [review]
[FIXED_X11R68x] Fix

The patch sets a compile time define so that PIO access is avoided.
Doing this ISA Mach8/32/64 hw is no longer usable with this driver.
No IA64 comes with ISA slots anyway.
Comment 2 Egbert Eich 2004-11-30 00:45:22 UTC
Fixed in CVS head.
Comment 3 Roland Mainz 2004-12-08 05:58:15 UTC
Comment on attachment 1423 [details] [review]
[FIXED_X11R68x] Fix

Approved in the 2004-12-06 release-wranglers conference call.
Please do not commit it yourself, I'll do that later today...
Comment 4 Roland Mainz 2004-12-15 00:47:46 UTC
Comment on attachment 1423 [details] [review]
[FIXED_X11R68x] Fix

Patch checked-in into X11R6.8.x stable branch:

/cvs/xorg/xc/ChangeLog,v  <--  ChangeLog
new revision: 1.365.2.86; previous revision: 1.365.2.85
cvs commit: Using deprecated info format strings.  Convert your scripts to use
the new argument format and remove '1's from your info file format strings.
/cvs/xorg/xc/programs/Xserver/hw/xfree86/drivers/ati/Imakefile,v  <-- 
Imakefile
new revision: 1.6.2.1; previous revision: 1.6
/cvs/xorg/xc/programs/Xserver/hw/xfree86/drivers/ati/atiprobe.c,v  <-- 
atiprobe.c
new revision: 1.3.2.2; previous revision: 1.3.2.1
cvs commit: Using deprecated info format strings.  Convert your scripts to use
the new argument format and remove '1's from your info file format strings.
Mailing the commit message to xorg-commit@lists.freedesktop.org...
Comment 5 Kevin E. Martin 2005-04-25 13:35:36 UTC
Unfortunately, this patch is causing quite a few ia64 systems to fail.

Egbert, is it possible to just disable non-PCI card probing and solve the
problem that you're seeing, but still allow MMIO access to older Mach64 PCI cards?

I'll attach a suggested patch next.
Comment 6 Kevin E. Martin 2005-04-25 13:38:37 UTC
(In reply to comment #5)
> Unfortunately, this patch is causing quite a few ia64 systems to fail.
> 
> Egbert, is it possible to just disable non-PCI card probing and solve the
> problem that you're seeing, but still allow MMIO access to older Mach64 PCI cards?

Oops, I meant to say: ... but still allow PIO access on older Mach64 PCI cards
for ia64?
Comment 7 Kevin E. Martin 2005-04-25 13:39:09 UTC
Created attachment 2545 [details] [review]
Proposed patch
Comment 8 Mike A. Harris 2005-04-26 09:07:45 UTC
Whenever this gets committed to CVS, whoever does it, please don't close the
bug report until it is committed both to HEAD and to stable branch.
Comment 9 George - 2006-08-04 18:33:50 UTC
egbert suggestion: AVOID_CPIO=yes
kevin suggestion:  AVOID_CPIO=no, AVOID_NON_PCI=yes

The modular driver does egbert's suggestion. Any updates on how ia64 should be
handled ?

note: the modular driver always does AVOID_NON_PCI=yes because of a typo I think.
Comment 10 George - 2006-08-25 08:01:54 UTC
Closing.

commit 1e06212f792d1ed1f7545e16824b86fbd34fa284 dropped the coin the kevin side.


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.