Andrew Morton <akpm@xxxxxxxx> uttered the following thing:
> On Sun, 13 Aug 2006 13:45:35 +0200
> Maciej Rutecki <maciej.rutecki@xxxxxxxxx> wrote:
> > Andrew Morton napisa??(a):
> > > ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.18-rc4/2.6.18-rc4-mm1/
> > I have problem with my keyboard. I have no error in dmesg and syslog,
> > but it doesn't work. I read google and try "i8042.nomux", but it didn't
> > help.
> > I enclose dmesg with "i8042.debug=1" option and my config.
> > Maybe I forgot something in config?
> Could be i8042-get-rid-of-polling-timer-v4.patch. Please try the below
> reversion patch, on top of rc4-mm1, thanks.
Acking the same issue. Applied the revert patch and my keyboard now
works. Also, it turns out that my keyboard is now the only thing that
failed to resume from S3 on my HP Nc6400, but adding "irqpoll" has fixed
that for now.
Also, to two other things I spotted with cpufreq. Running
speedstep-centrino on a Core Duo T2400 (1.83GHz) I see the
"cpuinfo_max_freq" is 1833000, but "scaling_max_freq" is fixed at 1333000,
regardless of the governor, and I cannot change it.
Also, whenever I echo anything to "scaling_governor", I get the
following kernel message:
[ 734.156000] BUG: warning at kernel/cpu.c:38/lock_cpu_hotplug()
[ 734.156000] [<c013c3ec>] lock_cpu_hotplug+0x7c/0x90
[ 734.156000] [<c01327f4>] __create_workqueue+0x44/0x140
[ 734.156000] [<c02dcf7b>] mutex_lock+0xb/0x20
[ 734.156000] [<e01f2665>] cpufreq_governor_dbs+0x2b5/0x310
[ 734.156000] [<c012f6a5>] notifier_call_chain+0x25/0x40
[ 734.156000] [<c0274b06>] __cpufreq_governor+0x46/0xe0
[ 734.156000] [<c0274c84>] __cpufreq_set_policy+0xe4/0x130
[ 734.156000] [<c0275ae4>] store_scaling_governor+0xd4/0x210
[ 734.160000] [<c02756a0>] handle_update+0x0/0x10
[ 734.160000] [<c01e2000>] kobject_get+0x0/0x20
[ 734.160000] [<c0275a10>] store_scaling_governor+0x0/0x210
[ 734.160000] [<c027521d>] store+0x3d/0x60
[ 734.160000] [<c01a2e0c>] sysfs_write_file+0x9c/0xf0
[ 734.160000] [<c016a8c6>] vfs_write+0xa6/0x160
[ 734.160000] [<c01a2d70>] sysfs_write_file+0x0/0xf0
[ 734.160000] [<c016b001>] sys_write+0x41/0x70
[ 734.160000] [<c0103115>] sysenter_past_esp+0x56/0x79
It seems that scaling still works, but this message is a bit unnerving.