>From: Darren Reed <darrenr@reed.wattle.id.au>
>To: mrg@eterna.com.au (matthew green)
>CC: abs@netbsd.org, tech-kern@netbsd.org
>Subject: Re: Increasing SHMMAXPGS
>Date: Fri, 17 May 2002 20:16:29 +1000 (EST)
>
>In some email I received from matthew green, sie wrote:
> >
> >    	We default to SHMMAXPGS=1024, which seems rather small given this
> >    	is the maximum limit and changes to it shouldn't affect the amount
> >    	of memory in use unless an application takes advantage of a higher
> >    	limit.
> >
> >    	Ogle (to give an example) likes a SHMMAXPGS of 16384.
> >
> >    	Would anyone object to increasing the default to 16384 for all
> >    	platforms?
> >
> >
> > perhaps, on a per-port basis, it could become a dynamic based on
> > the amount of ram available?
>
>why can't this be changable with sysctl ?

Does NetBSD use the traditional UNIX/BSD way of using fixed-size tables for 
resource management? How would this be turned into something changeable a 
run-time? realloc()? How are other changeable prameters handled?

Is it may be possible to have more SysV shared memory In THE SYSTEM than a 
single process could map into its address space? I.e. can several processes 
allocate shared memory and the sum total may be more than any one cold map 
individually? Is SHMMAXPGS the limit per process or for the entire system?

_________________________________________________________________
Send and receive Hotmail on your mobile device: http://mobile.msn.com