Date: Tue, 10 Sep 1996 03:22:57 -0700 From: David Greenman <dg@root.com> To: Michael Smith <msmith@atrad.adelaide.edu.au> Cc: lars@elbe.desy.de (Lars Gerhard Kuehl), Duncan.Barclay@pa-consulting.com, hackers@FreeBSD.org Subject: Re: undocumented kernel priority changing Message-ID: <199609101022.DAA17937@root.com> In-Reply-To: Your message of "Tue, 10 Sep 1996 19:22:28 %2B0930." <199609100952.TAA05559@genesis.atrad.adelaide.edu.au>
next in thread | previous in thread | raw e-mail | index | archive | help
>Lars Gerhard Kuehl stands accused of saying:
>>
>> For sites with only a few active processes that isn't a problem at all.
>> (No competition.) If there are many active processes 'renicing' is highly
>> desirable. The limit is usually reached only by jobs submitted
>> with 'interactive' nice levels. It might be an advantage to increase
>> the limit if someone likes to use xemacs on a 16MHz 386SX.
>> (10 minutes cpu time even on a 100 MHz 586 is pretty a lot ;)
>
>It's peanuts for long-lived processes in any sort of 'embedded' application:
>
>mstradar:/home/radar>uptime
>11:49AM up 23:35, 3 users, load averages: 0.19, 0.27, 0.25
>mstradar:/home/radar>ps ax
>...
> 8651 ?? SN 36:01.06 /home/radar/rd12/libexec/FreeBSD/exptd -f /home/radar
> 3303 p0- SN 83:44.55 /usr/local/rsi/idl//bin/bin.linux/idl analysis_init
>
>As you can see, it's been up less than a day, and the current load is pretty
>low. Depending on configuration, with just these two running the system
>will push a load average of 1.8 or more nonstop. (These two also normally
>start out of /etc/rc.local, so they've been restarted some time after
>the system booted.)
>
>Your point about only having a few processes is quite valid though - there's
>no problem with either responsiveness or overall performance there.
FreeBSD already has a sophisticated mechanism for controlling process
priorities (not nice value) for CPU hungry processes. The code in mi_switch()
looks like a total hack to me and should be removed in my opinion. Nothing
except the user or superuser should change the 'base scheduling priority'
("nice" value) of a process - certainly not automagically in the kernel.
-DG
David Greenman
Core-team/Principal Architect, The FreeBSD Project
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199609101022.DAA17937>