Bug 211475

Description of problem:
The latest kernel update (kernel-2.6.18-1.2200.fc5) has made suspend unreliable
on my dell inspiron 8500. Previous kernels (e.g. 2.6.17) have been working fine.
It will suspend and resume several times before eventually refusing to suspend
with the following messages in the logs;

Oct 19 17:49:32 echo kernel: Stopping tasks:
========================================================================================================
Oct 19 17:49:32 echo kernel:  stopping tasks timed out after 20 seconds (10
tasks remaining):
Oct 19 17:49:32 echo kernel:   khubd
Oct 19 17:49:32 echo kernel:   kseriod
Oct 19 17:49:32 echo kernel:   pdflush
Oct 19 17:49:32 echo kernel:   pdflush
Oct 19 17:49:32 echo kernel:   kswapd0
Oct 19 17:49:32 echo kernel:   pccardd
Oct 19 17:49:32 echo kernel:   kjournald
Oct 19 17:49:32 echo kernel:   kauditd
Oct 19 17:49:32 echo kernel:   knodemgrd_0
Oct 19 17:49:32 echo kernel:   hald-addon-stor
Oct 19 17:49:32 echo kernel: Restarting tasks...<6> Strange, khubd not stopped
Oct 19 17:49:32 echo kernel:  Strange, kseriod not stopped
Oct 19 17:49:32 echo kernel:  Strange, pdflush not stopped
Oct 19 17:49:32 echo kernel:  Strange, pdflush not stopped
Oct 19 17:49:32 echo kernel:  Strange, kswapd0 not stopped
Oct 19 17:49:32 echo kernel:  Strange, pccardd not stopped
Oct 19 17:49:32 echo kernel:  Strange, kjournald not stopped
Oct 19 17:49:32 echo kernel:  Strange, kauditd not stopped
Oct 19 17:49:32 echo kernel:  Strange, knodemgrd_0 not stopped
Oct 19 17:49:32 echo kernel:  Strange, hald-addon-stor not stopped
Oct 19 17:49:32 echo kernel:  done
Oct 19 17:49:32 echo gnome-power-manager: (sg) Resuming computer

Possibly related is that often on resume I now get the following backtrace,
although my machine continues to operate normally;

Oct 19 17:45:43 echo kernel: Stopping tasks:
=======================================================================================================================|
Oct 19 17:45:43 echo kernel: eth1: Going into suspend...
Oct 19 17:45:43 echo kernel: NVRM: RmPowerManagement: 3
Oct 19 17:45:43 echo kernel: Intel machine check architecture supported.
Oct 19 17:45:43 echo kernel: Intel machine check reporting enabled on CPU#0.
Oct 19 17:45:43 echo kernel: CPU0: Intel P4/Xeon Extended MCE MSRs (12) available
Oct 19 17:45:43 echo kernel: BUG: sleeping function called from invalid context
at kernel/rwsem.c:20
Oct 19 17:45:43 echo kernel: in_atomic():0, irqs_disabled():1
Oct 19 17:45:43 echo kernel:  [<c0403f10>] dump_trace+0x69/0x1af
Oct 19 17:45:43 echo smartd[1525]: Device: /dev/hda, SMART Usage Attribute: 194
Temperature_Celsius changed from 127 to 177
Oct 19 17:45:43 echo kernel:  [<c040406e>] show_trace_log_lvl+0x18/0x2c
Oct 19 17:45:43 echo NetworkManager: <information>      Deactivating device eth1.
Oct 19 17:45:43 echo kernel:  [<c04045e9>] show_trace+0xf/0x11
Oct 19 17:45:43 echo kernel:  [<c0404673>] dump_stack+0x15/0x17
Oct 19 17:45:43 echo kernel:  [<c042ef78>] down_read+0x12/0x1f
Oct 19 17:45:43 echo kernel:  [<c0427923>] blocking_notifier_call_chain+0xe/0x29
Oct 19 17:45:43 echo kernel:  [<c05970df>] cpufreq_resume+0x118/0x13f
Oct 19 17:45:43 echo kernel:  [<c053f3d4>] __sysdev_resume+0x20/0x53
Oct 19 17:45:43 echo kernel:  [<c053f515>] sysdev_resume+0x16/0x47
Oct 19 17:45:43 echo kernel:  [<c05435a1>] device_power_up+0x5/0xa
Oct 19 17:45:43 echo kernel:  [<c0436861>] suspend_enter+0x3b/0x44
Oct 19 17:45:43 echo kernel:  [<c0436997>] enter_state+0x12d/0x14f
Oct 19 17:45:43 echo kernel:  [<c0436a3e>] state_store+0x85/0x99
Oct 19 17:45:43 echo kernel:  [<c04969ea>] subsys_attr_store+0x1e/0x22
Oct 19 17:45:43 echo kernel:  [<c0496adc>] sysfs_write_file+0xa6/0xcc
Oct 19 17:45:43 echo kernel:  [<c0461292>] vfs_write+0xa8/0x159
Oct 19 17:45:43 echo kernel:  [<c04617d8>] sys_write+0x41/0x67
Oct 19 17:45:43 echo kernel:  [<c0402d9b>] syscall_call+0x7/0xb
Oct 19 17:45:43 echo kernel: DWARF2 unwinder stuck at syscall_call+0x7/0xb
Oct 19 17:45:43 echo kernel: Leftover inexact backtrace:
Oct 19 17:45:43 echo kernel:  =======================

I have also noticed that this kernel will also sometimes "beep" on resume which
never happened before.

Version-Release number of selected component (if applicable):
kernel-2.6.18-1.2200.fc5

How reproducible:
Fairly often.

Steps to Reproduce:
1. Upgrade to new kernel
2. Suspend/Resume a few times
3. 
  
Actual results:
Entering suspend does not always work.

Expected results:
Entering suspend always works

Additional info:
I have also reported the stacktrace under bug #187293 when attempting to use the
ondemand scaling governor. I do however have pm hook scripts to switch the
governor to performance during the suspend/resume process.