Messages in this thread | | | Date | Wed, 11 Nov 1998 12:04:13 +0100 | From | Kurt Garloff <> | Subject | Re: Schedule idle |
| |
On Tue, Nov 10, 1998 at 07:28:39PM -0700, yodaiken@chelm.cs.nmt.edu wrote: > Decoupled enough for what? If you want RC5 jobs to not even use this tiny > amount of cpu while the batch jobs are running, why not > > start batch_job > signal rc5 process to shut down > if( !fork()) > exec batch job > else wait() > restart rc5 > > > Making the kernel do what can be done better by a shell script is not good.
No, you just workaround a missing feature of the kernel: The ability to have things run only if nothing important (interactive jobs ...) can be done.
With your shell script logic, you should start to try to figure out, what the names of the processes, which were meant for IDLE scheduling, and sent them SIGSTOP if you don't want to be disturbed. At the end you wake them again with SIGCONT. You are not even allowed to do that for other user's jobs on a multiuser machine.
This is totally broken by design. You will have to guess which jobs were meant for IDLE scheduling. You will use some heuristic and guess, that all niced'19 jobs with names rc5des, cracker, ...
It's the one who starts an IDLE job who should decide about a job being idle or not, not somebody who want to run a nice'd job (which is something completely different!) and wants not to compete with IDLE jobs.
You need scheduler support for this, face it!
Nobody claims, that this makes Linux a full RT system, which a simple scheduler patch won't do and which is probably not possible without changing some major design features of Linux. I do maintain, however, that this feature is very useful for a lot of people, most of them just using rc5des or cracker.
Regards, -- Kurt Garloff <K.Garloff@ping.de> (Dortmund, FRG) PGP key on http://student.physik.uni-dortmund.de/homepages/garloff Unix IS user friendly - it's just selective about who its friends are!
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |