svn commit: r310336 - head/sys/kern

[Julian Elischer]

On 21/12/2016 3:44 AM, John Baldwin wrote:
> Author: jhb
> Date: Tue Dec 20 19:44:44 2016
> New Revision: 310336
> URL: https://svnweb.freebsd.org/changeset/base/310336
>
> Log:
>    Don't spin in pause() during early boot for kthreads other than thread0.
>    
>    pause() uses a spin loop to simulate a sleep during early boot.  However,
>    we only need this for thread0 to get far enough in the boot process to
>    enable timers (at which point pause() can sleep).  For other kthreads,
>    sleeping in pause() is ok as the callout will be scheduled and will
>    eventually fire once thread0 initializes timers.

maybe we just need to depend on a variable "have_timers"
that is independent of the cold/hot/whatever  settings.
first code to enable a timer sets it.

>    
>    Tested by: 	Steven Kargl
>    Sleuthing by:	markj
>    MFC after:	1 week
>    Sponsored by:	Netflix
>
> Modified:
>    head/sys/kern/kern_synch.c
>
> Modified: head/sys/kern/kern_synch.c
> ==============================================================================
> --- head/sys/kern/kern_synch.c	Tue Dec 20 19:30:21 2016	(r310335)
> +++ head/sys/kern/kern_synch.c	Tue Dec 20 19:44:44 2016	(r310336)
> @@ -321,7 +321,8 @@ pause_sbt(const char *wmesg, sbintime_t
>   	if (sbt == 0)
>   		sbt = tick_sbt;
>   
> -	if (cold || kdb_active || SCHEDULER_STOPPED()) {
> +	if ((cold && curthread == &thread0) || kdb_active ||
> +	    SCHEDULER_STOPPED()) {
>   		/*
>   		 * We delay one second at a time to avoid overflowing the
>   		 * system specific DELAY() function(s):
>
>