[UPDATE] new pid allocation patch

[Andre Oppermann]

Jun Su wrote:
> Hi All,
> 
> John Baldwin's proc lock commit today made my patch broken, so I have
> revised the patch, and uploaded it to the same position:
> http://www.arbornet.org/~junsu/pid.diff
> 
> The revised patch has been improved based on the feedback. Basically,
> 
> 1. Added a new sysctl value kern.pidmax to control the current pid range.
> This value can influence the reuse pid period. The initial value is 4096.

This sysctl name looks very misleading to me.  Unless the highest PID
I can get here is actually 4096.

> 2. Change the meaning of PID_MAX back to what it was, and sainty-check
> whether the pid is 5 digit based on PID_MAX. (I think it is impossible that
> pid expand PID_MAX for the algorithm nature, though.)
> 
> 3. The initial pid table size is bumped from 1<<5 to 1<<7.

-- 
Andre