Bug 374201 - anaconda fails to start installation
anaconda fails to start installation
Status: CLOSED DUPLICATE of bug 370761
Product: Fedora
Classification: Fedora
Component: rhpxl (Show other bugs)
powerpc Linux
low Severity urgent
: ---
: ---
Assigned To: Chris Lumens
Fedora Extras Quality Assurance
: 389301 (view as bug list)
Depends On:
  Show dependency treegraph
Reported: 2007-11-09 20:35 EST by Stefan Seefeld
Modified: 2008-01-03 19:32 EST (History)
4 users (show)

See Also:
Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of:
Last Closed: 2007-11-11 17:57:01 EST
Type: ---
Regression: ---
Mount Type: ---
Documentation: ---
Verified Versions:
Category: ---
oVirt Team: ---
RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: ---

Attachments (Terms of Use)

  None (edit)
Description Stefan Seefeld 2007-11-09 20:35:18 EST
Description of problem:
anaconda raises a python exception during display configuration

Version-Release number of selected component (if applicable):

How reproducible:
Run installation on PS3

Steps to Reproduce:
1. Enter F8 installation on PS3
Actual results:

Anaconda starts, then stops with a python traceback

(around rhpxl/videocard.py, line 149, in __init__
  elif os.uname()[4].count("ppc") > 0 and card.pcidom > 0:
AttributeError: device instance has no attribute 'pcidom'

Expected results:

installation to proceed normally

Additional info:

(I tried installing in text-only mode but couldn't figure out how.)
Comment 1 David Woodhouse 2007-11-09 22:15:48 EST
/me looks suspiciously at

AFAICT both hunks of that are broken -- the first because of the trace observed,
and the second because it'll change the fallback case on PowerPC machines to
'vesa' instead of fbdev.
Comment 2 David Woodhouse 2007-11-10 16:12:30 EST
Hm, I think it's even more broken than it seems at first glance. The assumption
that devices in domains > 0 will be mapped above 4GiB is completely bogus in the
first place -- we support machines where that isn't the case, on which X works fine.

If you want to avoid using 'real' X drivers on cards where the resources are
mapped above 4GiB, then perhaps the best way to achieve that is to try checking
where the resources are mapped, rather than making up broken heuristics?

Besides -- if X maps the resources correctly through sysfs, why does it care
about the physical address of them anyway?
Comment 3 Stefan Seefeld 2007-11-11 11:32:34 EST
For the record: I have reverted the above change, and dropped the modified
module into a RHupdates directory. With this, I was able to get F8 installed on
my PS3.

Thanks a lot for the quick help !
Comment 4 David Woodhouse 2007-11-11 17:57:01 EST

*** This bug has been marked as a duplicate of 370761 ***
Comment 5 Greg Morgan 2007-11-15 18:27:05 EST
*** Bug 381541 has been marked as a duplicate of this bug. ***
Comment 6 David Woodhouse 2007-12-06 20:50:05 EST
Any chance of getting this fixed, please?
Comment 7 David Woodhouse 2008-01-03 12:20:04 EST
Comment 8 David Woodhouse 2008-01-03 19:32:22 EST
*** Bug 389301 has been marked as a duplicate of this bug. ***

Note You need to log in before you can comment on or make changes to this bug.