[Bug 36598] New: [RS880] Slow resume from suspend generates kernel oops

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=36598

           Summary: [RS880] Slow resume from suspend generates kernel oops
           Product: DRI
           Version: XOrg CVS
          Platform: x86-64 (AMD64)
        OS/Version: Linux (All)
            Status: NEW
          Severity: normal
          Priority: medium
         Component: DRM/Radeon
        AssignedTo: dri-devel at lists.freedesktop.org
        ReportedBy: steve at stevemurphy.org


Created an attachment (id=46075)
 --> (https://bugs.freedesktop.org/attachment.cgi?id=46075)
dmesg output for boot and suspend, showing the issue around 78.3

I'm forwarding this bug from ubuntu -
https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-ati/+bug/764788

Problem]
Resume from suspend triggers a kernel oops due to the radeon kms module taking
more than 10 seconds to resume

[Discussion]
When resuming from the suspend state, a popup box is displayed indicating an
error has occured and "Your system might become unstable now and might need to
be restarted.". In the dmesg output, the following lines are present before the
kernel oops, showing that the radeon driver took 11408.919 msecs to resume and
the kernel oops is due to resume taking longer than 10 seconds total (11430
msecs).

Interesting lines in the dmesg output when the error occurs are below (with
kernel.drm=0x04 set, adding a bit of verbosity when the radeon driver is
restoring):

[ 69.792851] PM: resume of drv:scsi_disk dev:2:0:0:0 complete after 2644.379
msecs
[ 70.122093] [drm:output_poll_execute], [CONNECTOR:15:VGA-1] status updated
from 2 to 2
[ 70.124058] [drm:radeon_atombios_connected_scratch_regs], DFP1 disconnected
[ 70.124062] [drm:output_poll_execute], [CONNECTOR:17:HDMI-A-1] status updated
from 2 to 2
[ 71.960104] [drm:radeon_atom_encoder_dpms], encoder dpms 21 to mode 3, devices
00000001, active_devices 00000000
[ 71.960113] [drm:radeon_atom_encoder_dpms], encoder dpms 30 to mode 3, devices
00000008, active_devices 00000000
[ 71.960136] [drm:radeon_compute_pll_avivo], 7187, pll dividers - fb: 130.5
ref: 2, post 13
[ 71.980174] [drm:drm_crtc_helper_set_mode], [ENCODER:12:LVDS-12] set
[MODE:30:1366x768]
[ 76.950136] [drm:avivo_crtc_load_lut], 0
[ 76.950200] [drm:radeon_atom_encoder_dpms], encoder dpms 31 to mode 0, devices
00000002, active_devices 00000002
[ 76.960133] [drm:radeon_atom_encoder_dpms], encoder dpms 31 to mode 3, devices
00000002, active_devices 00000002
[ 78.303704] [drm:radeon_atom_encoder_dpms], encoder dpms 21 to mode 3, devices
00000001, active_devices 00000000
[ 78.303713] [drm:radeon_atom_encoder_dpms], encoder dpms 30 to mode 3, devices
00000008, active_devices 00000000
[ 78.303734] [drm:avivo_crtc_load_lut], 0
[ 78.303790] [drm:radeon_atom_encoder_dpms], encoder dpms 31 to mode 0, devices
00000002, active_devices 00000002
[ 78.310132] PM: resume of drv:radeon dev:0000:01:05.0 complete after 11408.919
msecs
[ 78.310148] PM: resume of drv:drm dev:controlD64 complete after 8513.702 msecs
[ 78.310173] PM: resume of devices complete after 11409.422 msecs
[ 78.330262] PM: resume devices took 11.430 seconds
[ 78.330265] ------------[ cut here ]------------
[ 78.330272] WARNING: at
/build/buildd/linux-2.6.38/kernel/power/suspend_test.c:53
suspend_test_finish+0x86/0x90()
[ 78.330274] Hardware name: 1215T
[ 78.330277] Component: resume devices, time: 11430

Upon investigating further, it looks like portions of the the following
workarounds were discovered that allow resume from suspend to work as expected,
by avoid the radeon driver:

[Workarounds]
1. Add "nomodeset" to the linux command line in grub
2. Use the proprietary FGLRX driver.

[System Environment]
ASUS 1215T running a fresh install of 11.04 beta or 11.04 beta2.

-- 
Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.