Bug 27225 - Radeon KMS fails on Xpress 200M
Summary: Radeon KMS fails on Xpress 200M
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Radeon (show other bugs)
Version: unspecified
Hardware: Other All
: medium normal
Assignee: Default DRI bug account
QA Contact:
URL: https://bugs.launchpad.net/bugs/557611
Depends on:
Reported: 2010-03-21 08:28 UTC by Maarten Fonville
Modified: 2010-04-07 13:38 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:

Normal failure (2.49 KB, text/plain)
2010-03-21 08:29 UTC, Maarten Fonville
no flags Details
Failure with forcing agpmode=-1 (5.80 KB, text/plain)
2010-03-21 08:29 UTC, Maarten Fonville
no flags Details
fix (1.22 KB, patch)
2010-03-21 08:55 UTC, Alex Deucher
no flags Details | Splinter Review

Description Maarten Fonville 2010-03-21 08:28:29 UTC
On an Ubuntu Lucid installation (at the moment of writing still beta) KMS should be enabled on the Radeon Xpress 200M.
But intialization of KMS does fail with the lines in attachment kms_fail.txt

Apparently the driver doesn't like the card's memory, which I believe is an IGP card without memory itself.

Next thing I tried was passing 'radeon agpmode=-1' to the module, to force it into PCI mode.
(even though the card does say in lshw: 
description: VGA compatible controller
                product: RC410 [Radeon Xpress 200M]
                vendor: ATI Technologies Inc
                physical id: 5
                bus info: pci@0000:01:05.0
                version: 00
                width: 32 bits
                clock: 66MHz
                capabilities: agp agp-3.0 pm msi cap_list
                configuration: latency=64 mingnt=8
                resources: memory:c0000000-cfffffff(prefetchable) ioport:9800(size=256) memory:fe1f0000-fe1fffff memory:fe1c0000fe1dffff(prefetchable)

and so I think is an AGP card and not a PCI(E) card)

anyhow, this does seem to improve the situation, because now can KMS be loaded. But a new error appears: it says it fails the r100_ring_test with 0xCAFEDEAD. I don't know why it fails, I found multiple people with this problem on the internet, but nobody with a fix.
The log of this one I will attach as kms_forceagp-1_fail.txt

Tested: kernel 2.6.32(from lucid with DRM code), 2.6.33 and 2.6.34-rc1
Tested: DRM 2.4.18 (from lucid), 2.4.19+git20100318 (xorg-edgers repository)
I load radeon in /etc/modules to force the loading of the driver in time to assure that the firmware does load in time (recommended by developers)

Already thanks for your time, if you need more information/testing I am willing to of course.
Comment 1 Maarten Fonville 2010-03-21 08:29:04 UTC
Created attachment 34296 [details]
Normal failure
Comment 2 Maarten Fonville 2010-03-21 08:29:24 UTC
Created attachment 34297 [details]
Failure with forcing agpmode=-1
Comment 3 Alex Deucher 2010-03-21 08:55:03 UTC
Created attachment 34298 [details] [review]

RS4xx+ IGP chips use an internal gart, however,
some of them have the agp cap bits set in their pci
configs.  Make sure to clear the AGP flag as AGP will
not work with them.
Comment 4 Maarten Fonville 2010-03-21 09:14:28 UTC
I will try this fix immediately

and *wow* you are fast :) and (hopefully) a hero ;)
Comment 5 Maarten Fonville 2010-03-21 15:46:57 UTC
I can confirm the patch of Alex does work. So if this patch can gets commited soon into the mainline code it is 'fixed'
Comment 6 Alex Deucher 2010-03-21 16:45:03 UTC
corrected patch sent to Dave.

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.