Index of /radeon-woes

Icon  Name                    Last modified      Size  Description
[PARENTDIR] Parent Directory                             -   
[   ] Xorg.log-3.1.10         2012-01-31 06:00   38K  
[   ] Xorg.log-3.2.2          2012-01-31 06:03   60K  
[   ] dmesg-3.1.10            2012-01-31 05:59   51K  
[   ] dmesg-3.2.2             2012-01-31 06:03   48K  

All 3.2.x kernels seem to panic (but no flashing keyboard leds) in early 
boot.  The kernel starts booting, and I see udev started inside initramfs,
and as soon as the drm/kms stuff loads (firmware for this card is also
present in my initramfs), the screen goes blank and the machine remains
unresponsive from then on (no net or some such later).  This does not 
occur on 3.1.10 (or any 3.1.x or 3.0.x kernels) - it is new since 3.2.0
(or some point during development thereof).


First, here's the graphics chipset:
  VGA compatible controller: ATI Technologies Inc RS690 [Radeon X1200 Series]

Second, this is all running on libdrm-2.4.29, mesa-7.11.2, xorg-server-1.11.4,
and xf86-video-ati-6.14.3 with udev-180 (though the problem existed on 178 and
179 as well).  


Here's what happens on 3.1.10:

  # lsmod | grep radeon
  radeon                963980  2 
  ttm                    53329  1 radeon
  drm_kms_helper         25312  1 radeon
  drm                   184837  4 drm_kms_helper,ttm,radeon
  hwmon                   1425  4 radeon,asus_atk0110,thermal_sys,k8temp
  i2c_algo_bit            5191  1 radeon
  i2c_core               19909  5 i2c_algo_bit,drm,drm_kms_helper,radeon,i2c_piix4

  See dmesg-3.1.10 and Xorg.log-3.1.10 in this same directory if interested.

Since the crash or whatever was occurring early (immediately after the drm
and kms and radeon modules load in the initramfs) when the display changes
from the "standard" vga mode to the kms console, I thought it was worth a
try to disable modesetting, so here's the results from that:

Kernel 3.2.2 (booted with "options radeon modeset=0" in initrd)

  # lsmod | grep radeon
  radeon                754156  0 
  ttm                    53457  1 radeon
  drm_kms_helper         25728  1 radeon
  drm                   187598  3 drm_kms_helper,ttm,radeon
  hwmon                   1329  4 radeon,asus_atk0110,k8temp,thermal_sys
  i2c_algo_bit            5223  1 radeon
  i2c_core               19941  5 i2c_algo_bit,drm,drm_kms_helper,radeon,i2c_piix4

  It booted just fine with no crash, and I'll admit to being surprised when 
  X started up just fine -- I had forgotten that kms in xf86-video-ati was
  still optional...

  See dmesg-3.2.2 and Xorg.log-3.2.2 in this same directory if interested.

Feedback from IRC - this is just so I don't forget (I'm not at home now):

22:13 < rworkman> Not sure whose bug this is, but would like some feedback and/or suggestions on how to proceed from here: 
                  http://rlworkman.net/radeon-woes/
22:13 < rworkman> airlied: ^
22:21 < RAOF> rworkman: I'd guess that the most useful next step would be to boot with drm.debug=0xe and see if dmesg survives in 
              /var/log/kern.log or somewhere.
22:21 < RAOF> rworkman: Then, bugzilla time!
22:22 < RAOF> That would be with modesetting on, obviously.