Bug 34212 - radeon: 3650 AGP lockup on radeon.ko load
Summary: radeon: 3650 AGP lockup on radeon.ko load
Status: RESOLVED OBSOLETE
Alias: None
Product: Drivers
Classification: Unclassified
Component: Video(DRI - non Intel) (show other bugs)
Hardware: All Linux
: P1 normal
Assignee: drivers_video-dri
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2011-05-02 00:48 UTC by Jason Detring
Modified: 2012-08-23 13:47 UTC (History)
1 user (show)

See Also:
Kernel Version: 2.6.39-rc5 + drm-next 1794d25
Subsystem:
Regression: No
Bisected commit-id:


Attachments
Netconsole output, kernel drm-next 1794d25 (6.03 KB, text/plain)
2011-05-02 00:50 UTC, Jason Detring
Details

Description Jason Detring 2011-05-02 00:48:56 UTC
Hardware:
2x 1.7GHz Xeon (Foster) on a SuperMicro P4DC6 workstation board
RadeonHD 3650 AGP (built by VisionTek)

Radeon board has 1x VGA, 1x DVI, and 1x S-Video ports.  A TV is attached to the S-Video port through a generic S-Video multiplexer, the others are not connected.  TV is a US/NTSC standard definition CRT.

System boots into a VGA text console with no problem.  When attempting "modprobe radeon", the system switches to graphics mode, then prints a backtrace on the framebuffer and hangs (no keyboard response, no ping, ssh dies).  A mostly-solid stripe of green pixel noise can be seen across the bottom 2 or 3 scanlines of the TV.

The setup works when using the fglrx driver.  An NV31+nouveau arrangement also seems to work, so I don't think the motherboard is at fault.  This isn't really a "regression" per-se, as it has never worked in any vendor- or self-compiled kernel I've tried (since about July 2010).
Comment 1 Jason Detring 2011-05-02 00:50:49 UTC
Created attachment 56132 [details]
Netconsole output, kernel drm-next 1794d25
Comment 2 Alan 2012-08-23 13:47:59 UTC
If this bug is still seen with modern kernels please update/reopen

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