0(1)-patch, where did it go?

0(1)-patch, where did it go?

[listmail]

Ok,
Maybe I am just lost and confused, but I can't see anywhere what has
happened to this patch....

Did it get merged when I wasn't looking?  I can't seem to find a record of
it in the change logs?  It just seems to stop getting updated just before
2.4.18 release, but I don't see a message about it getting merged....

Could someone please update me?

Is it in 2.4.19 someplace, or  2.5.x?

-Haplo

Re: 0(1)-patch, where did it go?

[J Sloan]

listmail@majere.epithna.com wrote:

>Ok,
>Maybe I am just lost and confused, but I can't see anywhere what has
>happened to this patch....
>
>Did it get merged when I wasn't looking?  I can't seem to find a record of
>it in the change logs?  It just seems to stop getting updated just before
>2.4.18 release, but I don't see a message about it getting merged....
>
>Could someone please update me?
>
>Is it in 2.4.19 someplace, or  2.5.x?
>
It's in 2.5, and also in the 2.4 -ac tree

Joe

Re: 0(1)-patch, where did it go?

[J.A. Magallon]

On 2002.04.09 listmail@majere.epithna.com wrote:
>Ok,
>Maybe I am just lost and confused, but I can't see anywhere what has
>happened to this patch....
>
>Did it get merged when I wasn't looking?  I can't seem to find a record of
>it in the change logs?  It just seems to stop getting updated just before
>2.4.18 release, but I don't see a message about it getting merged....
>
>Could someone please update me?
>

You can get an up-to-date version in the 2.4.19-pre6-jam1 patcset in

http://giga.cps.unizar.es/~magallon/linux/kernel/


-- 
J.A. Magallon                           #  Let the source be with you...        
mailto:jamagallon@able.es
Mandrake Linux release 8.3 (Cooker) for i586
Linux werewolf 2.4.19-pre6-jam1 #1 SMP Sun Apr 7 00:50:05 CEST 2002 i686

Re: 0(1)-patch, where did it go?

[Martin J. Bligh]

> You can get an up-to-date version in the 2.4.19-pre6-jam1 patcset in
> 
> http://giga.cps.unizar.es/~magallon/linux/kernel/

Do you have a version of this broken out as a seperate patch 
anywhere? There seem to have been updates to the O(1) scheduler 
in 2.5 since the K3 version of the scheduler, but I've not seen
any new version of the 2.4 version of the patch ...

Thanks,

M.

Re: 0(1)-patch, where did it go?

[J.A. Magallon]

On 2002.04.10 Martin J. Bligh wrote:
>> You can get an up-to-date version in the 2.4.19-pre6-jam1 patcset in
>> 
>> http://giga.cps.unizar.es/~magallon/linux/kernel/
>
>Do you have a version of this broken out as a seperate patch 
>anywhere? There seem to have been updates to the O(1) scheduler 
>in 2.5 since the K3 version of the scheduler, but I've not seen
>any new version of the 2.4 version of the patch ...
>

The patch in that location is split in pieces, but it patches on top of
all the previous patches in the set. I could make it patch cleanly on
plain 2.4.19-pre6, but if there have been changes in the O1 scheduler
it would be better that Ingo backported them to 2.4 if he has the
time.

-- 
J.A. Magallon                           #  Let the source be with you...        
mailto:jamagallon@able.es
Mandrake Linux release 8.3 (Cooker) for i586
Linux werewolf 2.4.19-pre6-jam1 #1 SMP Sun Apr 7 00:50:05 CEST 2002 i686

Re: 0(1)-patch, where did it go?

[Dieter Nützel]

> You can get an up-to-date version in the 2.4.19-pre6-jam1 patcset in
> 
> http://giga.cps.unizar.es/~magallon/linux/kernel/

Didn't you noticed any of the reports about very bad numbers for latency since 
Ingo's latest 2.4.17-K3 version?
Even Alan's tree show the same (latest I've checked was 2.4.19-pre2-ac2).

We do need some words from Ingo first.
He haven't answered my posts since February ;-(
But yaybe he didn't got them 'cause I send them to mingo@elte.hu ???

If you run without O(1) latency is like before.
You'll get better numbers with preemption+lock-break (ongoing merge of 
Andrew's lowlatency patches).

But I see some kernel hangs with preemption on UP.
It happens only during "make bzlilo" (the linking stage). Robert?
Apart from that it works well.

-Dieter

-- 
Dieter Nützel
Graduate Student, Computer Science

University of Hamburg
Department of Computer Science
@home: Dieter.Nuetzel@hamburg.de

Re: 0(1)-patch, where did it go?

[Robert Love]

On Wed, 2002-04-10 at 23:27, Dieter Nützel wrote:

> But I see some kernel hangs with preemption on UP.
> It happens only during "make bzlilo" (the linking stage). Robert?
> Apart from that it works well.

It is probably lock-break, not preempt.  I don't have lock-break patches
for 2.4.19-pre yet.  Lock-break/low-latency and the more general lock
breaking / explicit schedule work is very reliant on the version of the
kernel they were designed against.  This is why this approach is not a
proper long-term solution ...

	Robert Love

Re: 0(1)-patch, where did it go?

[Dieter Nützel]

On Donnerstag, 11. April 2002 :30, Robert Love wrote:
> On Wed, 2002-04-10 at 23:27, Dieter Nützel wrote:
> > But I see some kernel hangs with preemption on UP.
> > It happens only during "make bzlilo" (the linking stage). Robert?
> > Apart from that it works well.
>
> It is probably lock-break, not preempt.  I don't have lock-break patches
> for 2.4.19-pre yet.  Lock-break/low-latency and the more general lock
> breaking / explicit schedule work is very reliant on the version of the
> kernel they were designed against.  This is why this approach is not a
> proper long-term solution ...

OK, thanks Robert will try without it after some sleep.

But preemption without lock-break on 2.4 is like running without preemption.
The general latency problem with O(1) for 2.4 still stands.
Do you have similar observations with the current -ac tree?
You should have my numbers.

I only would bring your focus somewhat back to 2.4 'cause 2.6 is so far...

Thanks,
	Dieter

Re: 0(1)-patch, where did it go?

[KELEMEN Peter]

* Dieter Nützel (dieter.nuetzel@hamburg.de) [20020411 05:27]:

> We do need some words from Ingo first.  He haven't answered my
> posts since February ;-( But yaybe he didn't got them 'cause I
> send them to mingo@elte.hu ???

mingo@elte.hu is accepting mails just fine.

Peter

-- 
    .+'''+.         .+'''+.         .+'''+.         .+'''+.         .+''
 Kelemen Péter     /       \       /       \       /      fuji@elte.hu
.+'         `+...+'         `+...+'         `+...+'         `+...+'