Re: 2.6.0-test7-mm1

[Andrew Morton <akpm@osdl.org>]

Bradley Chapman <kakadu_croc@yahoo.com> wrote:
>
> Mr. Morton,
> 
> After backing out these two patches, recompiling and rebooting, I get these
> two unrelated Oopses:
> 
> PID hash table entries: 2048 (order 11: 16384 bytes)
> Debug: sleeping function called from invalid context at include/linux/rwsem.h:66
> in_atomic():1, irqs_disabled():1
> Call Trace:
>  [<c0123fc6>] __might_sleep+0xa0/0xc2
>  [<c02e9727>] cpufreq_register_notifier+0x30/0x96
>  [<c0107000>] rest_init+0x0/0xf4
>  [<c0493af9>] init_tsc+0x54/0xf2
>  [<c0107000>] rest_init+0x0/0xf4
>  [<c011a141>] select_timer+0x26/0x55
>  [<c048f4d5>] time_init+0x38/0x49
>  [<c048c66b>] start_kernel+0x124/0x22e
>  [<c048c41b>] unknown_bootoption+0x0/0xff

It's not an oops, it's a warning.  cpufreq_register_notifier() is being
called super-early but is taking a semaphore.  I've asked the cpufreq guys
if this can be moved to an initcall but received one of those stupid "your
message is awaiting moderator approval" replies from the mailing list.  I
gave up at that point.