Red Hat Bugzilla – Bug 180016
PCIID not recognised
Last modified: 2007-11-30 17:11:23 EST
Description of problem:
Radeon X550 PCIe card not recognised by "system-config-display --reconfig"
but works with manual ChipId/Driver settings
Version-Release number of selected component (if applicable):
Name : xorg-x11-drv-ati
Arch : x86_64
various log and conf files attached ...
Created attachment 124176 [details]
My everyday working Xorg config with manual driver/chipid
Created attachment 124177 [details]
auto generated config file
Produced a new config file
# system-config-display --reconfig --verbose --noui
Trying with card: ATI Technologies Inc RV370 secondary [ATI Sapphire X550
Backing up /etc/X11/xorg.conf as /etc/X11/xorg.conf.backup
Writing configuration to /etc/X11/xorg.conf
Removing old /etc/X11/X
Creating /etc/X11/X symlink
Created attachment 124178 [details]
Log file from starting x with auto generated config
Thanks for testing.
I've reviewed your attachments, and your config file is configured to use
the "vesa" driver, and the log file is a log from using the vesa driver.
What is needed, is for you to run system-config-display --reconfig and then
examine the xorg.conf. If it has the "vesa" driver present, you need to
manually change it to "radeon" and do not add or remove or modify anything
else in the config. Then start the X server and see if the card works with
the "radeon" driver. If it does, without adding any other options, then
your card is supported by the radeon driver, and I need you to attach the
X server log file and config file from the new working session. That
shows the PCI ID, and confirms that the driver supports the hardware
Once I have that info, I can update the radeon.xinf to map your card to the
Radeon driver, and it should autodetect it properly then.
Thanks in advance.
OK, I've been a away a day or two and still got to dash now, so no time to
perform the test on latest driver until later today or tomorrow, but if you
didn't change the PCIIDs in the driver since the X11R7RC4 then I can confirm
that simply changing the driver isn't enough, I had to also add the ChipId to
force it to believe it would handle it ...
alias pcivideo:v00001002d00005B63sv*sd*bc*sc*i* radeon
This chip is already listed. If it doesn't autodetect properly, file a bug
Created attachment 124426 [details]
xorg startup with radeon driver forced
just tried again before I saw comment #5 so will post those results anyway
I upgraded packages to latest rawhide
did the system-config-display --reconfig which detected vesa, changed only the
vesa to radeon, restarted X which failed with no screens, log file attached.
Mid-Air collision avoided and freedesktop.org bugzilla id# corrected!
yes the "v00001002d00005B63sv*sd*" looks correct to match the card
# lspci -vn
01:00.0 0300: 1002:5b63
01:00.1 0380: 1002:5b73
# lspci -v
01:00.0 VGA compatible controller: ATI Technologies Inc RV370 [ATI Sapphire X550
Silent] (prog-if 00 [VGA])
Subsystem: PC Partner Limited Unknown device 1490
01:00.1 Display controller: ATI Technologies Inc RV370 secondary [ATI Sapphire X
Subsystem: PC Partner Limited Unknown device 1491
Perhaps that is where system-config-display gets the message
"Trying with card: ATI Technologies Inc RV370 secondary [ATI Sapphire X550
Given that when I force the radeon driver it fails, but it works when I also add
the chipid doesn't that still indicate a driver issue? (just to be clear the
actual card is 0x5B63 but I have to force it to 0x5B62) if there is *also* a
kudzu issue I'm happy to file a bug for it too ...
*** This bug has been marked as a duplicate of 176978 ***
OK, over on BZ#176978, notting has fixed kudzu and/or s-c-display so that
# system-config-display --reconfig --verbose --noui
now *does* decide to use the radeon driver for this card, instead of the vesa
one, however X still refuse to start until I add "ChipId 0x5B62"
Revisiting my old bugs ...
Up to yesterday my system was still running rawhide from just pre-FC5 with post
FC5-updates, so I flattened it and installed FC5.
Booted from DVD, when anaconda tried to start X it saw the X550 card, but failed
to start X using radeon driver, it then also failed to start using the VESA
driver (don't know why that should fail), so I had to do my installation using
vnc, which went OK.
After installation X still failed to start until I added my old faithful "ChipId
0x5b62" to xorg.conf which was necessary all through the FC5 test series.
I know this bug was bounced between X and kudzu, but somewhere along the way it
still doesn't work in FC5, it's not a big problem for me as I know how to work
around it, but perhaps it could get some attention for FC6?
OK, if we're picking radeon by default, but it's not working, that's a driver
issue. Assigning there.
Hmm, from the last few comments it would sound as if the PCI ID list inside
the driver may have a typo or missing entry in one location.
------- Additional Comments From email@example.com 2006-06-10 05:56 EST -------
Ok... it turns out that the driver does not actually support this chip
at all currently, which is why it is not working for you until you force
the chip ID.
Forcing the ChipID tells the driver to ignore what your hardware really
is, and pretend to be another chip, so the driver then gladly pretends
you have a 1002:5B62 which /is/ supported. If the 5B63 chip works close
enough to the 5B62, then the result is working video, however this is not
always the case.
It appears that it works ok for you, so we can update the radeon driver's
internal PCI ID lists to detect this chip automatically now, and treat
it like the 5B62 internally. If there are any differences in the
chips however, you may experience problems. If that happens, you'll need
to file a bug report to X.Org at http://bugs.freedesktop.org for
In the mean time though, I think it is safe to update the driver to
handle this chip so long as you seem to be having no problems with it
as things stand currently.
------- Additional Comments From firstname.lastname@example.org 2006-06-10 06:10 EST -------
* Sat Jun 10 2006 Mike A. Harris <email@example.com> 220.127.116.11-4.FC5.0
- Added ati-radeon-pci-id-update.patch, which adds support for Sapphire X550,
1002:5B63 for bug (#180016)
When we do an ati driver update for FC5, this fix will be included.
------- Additional Comments From firstname.lastname@example.org 2006-06-10 06:11 EST -------
The chip is already supported in rawhide, so closing as "RAWHIDE", however
as indicated above, it'll also be released in a future FC5 update as well.
------- Additional Comments From email@example.com 2006-06-10 14:13 EST -------
(In reply to comment #14)
> > In the mean time though, I think it is safe to update the driver to
> > handle this chip so long as you seem to be having no problems with it
> > as things stand currently.
After forcing the ChipID the only issue I have had was when under a xen kernel,
where some slight garbling happened, this was "usually" repeatable for me and
confirmed as also affecting Daniel Veillard