O(1) Scheduler from Ingo vs. O(1) Scheduler from Robert

O(1) Scheduler from Ingo vs. O(1) Scheduler from Robert

[Marc-Christian Petersen]

Hi there,

say, can anyone explain me why $subject patches are so different?
What exactly are the important differences, what patch should we use?

-- 
Kind regards
        Marc-Christian Petersen

http://sourceforge.net/projects/wolk

PGP/GnuPG Key: 1024D/569DE2E3DB441A16
Fingerprint: 3469 0CF8 CA7E 0042 7824 080A 569D E2E3 DB44 1A16
Key available at www.keyserver.net. Encrypted e-mail preferred.

Re: O(1) Scheduler from Ingo vs. O(1) Scheduler from Robert

[Ingo Molnar]

On Fri, 4 Oct 2002, Marc-Christian Petersen wrote:

> say, can anyone explain me why $subject patches are so different? What
> exactly are the important differences, what patch should we use?

well as far as i can tell Robert has put other stuff into his patch, which
isnt really part of the O(1) scheduler. So i'd call it "the O(1) scheduler
plus stuff".

	Ingo

Re: O(1) Scheduler from Ingo vs. O(1) Scheduler from Robert

[Robert Love]

On Fri, 2002-10-04 at 11:17, Ingo Molnar wrote:

> On Fri, 4 Oct 2002, Marc-Christian Petersen wrote:
> 
> > say, can anyone explain me why $subject patches are so different? What
> > exactly are the important differences, what patch should we use?
> 
> well as far as i can tell Robert has put other stuff into his patch, which
> isnt really part of the O(1) scheduler. So i'd call it "the O(1) scheduler
> plus stuff".

There should _not_ be other things in the patch aside from the
scheduler.  Those patches are based on Ingo's original 2.4 patches with
back-ported fixes from 2.4-ac and 2.5.  Unfortunately, at the moment the
patch is a bit out of sync.  The only 2.4 version of the scheduler I
have been able to keep up-to-date is 2.4-ac... but the patch is not too
bad.

I think the reason my patches differ from Ingo's is that Ingo includes
code that is not yet in mainline 2.5.  For example, last I checked his
patches had the SCHED_BATCH stuff, which is good, but I only want to put
code that is in 2.5 already and tested.

	Robert Love

Re: O(1) Scheduler from Ingo vs. O(1) Scheduler from Robert

[Ingo Molnar]

On 4 Oct 2002, Robert Love wrote:

> There should _not_ be other things in the patch aside from the
> scheduler. [...]

hm, i remember there was some 'set max RT priority in .config' stuff in
it, isnt that the case anymore?

> I think the reason my patches differ from Ingo's is that Ingo includes
> code that is not yet in mainline 2.5.  For example, last I checked his
> patches had the SCHED_BATCH stuff, [...]

hm, i forgot about that. Well, it's pretty harmless.

	Ingo

Re: O(1) Scheduler from Ingo vs. O(1) Scheduler from Robert

[Robert Love]

On Fri, 2002-10-04 at 11:28, Ingo Molnar wrote:

> On 4 Oct 2002, Robert Love wrote:
> 
> > There should _not_ be other things in the patch aside from the
> > scheduler. [...]
> 
> hm, i remember there was some 'set max RT priority in .config' stuff
> in it, isnt that the case anymore?

Oh, yes, indeed.  I forgot about that.

Since we cleaned up the whole MAX_PRIO thing the only main difference in
my tree is that fact it is exported as a CONFIG setting (and the logic
to implement a new ffs if BITMAP_SIZE changes).

Otherwise the trees aim to be in-sync.

	Robert Love