2.4 O(1) scheduler

2.4 O(1) scheduler

[anton wilson]

	

I'm working on a project that uses the O(1) scheduler and I am forced to use 
a 2.4 kernel for time issues. Will the O(1) patches for 2.4 kernels be 
updated once the 2.5.26+ patch becomes stable?

Anton

Re: 2.4 O(1) scheduler

[J Sloan]

Use 2.4-aa, 2.4-ac or 2.4-redhat kernel
and you get the O(1) secheduler at
no extra cost -

Joe

anton wilson wrote:

>	
>
>I'm working on a project that uses the O(1) scheduler and I am forced to use 
>a 2.4 kernel for time issues. Will the O(1) patches for 2.4 kernels be 
>updated once the 2.5.26+ patch becomes stable?
>
>Anton
>-
>To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
>the body of a message to majordomo@vger.kernel.org
>More majordomo info at  http://vger.kernel.org/majordomo-info.html
>Please read the FAQ at  http://www.tux.org/lkml/
>
>  
>

Re: 2.4 O(1) scheduler

[anton wilson]

On Friday 19 July 2002 03:54 pm, J Sloan wrote:
> Use 2.4-aa, 2.4-ac or 2.4-redhat kernel
> and you get the O(1) secheduler at
> no extra cost -
>


> Joe


I'm actually worried not about just the O(1) scheduler but if these patches 
will be incorporating the O(1) bug fixes such as the serious one in 
balance_load where curr->next was used instead of current->prev. Also, I need 
to use a patch that won't tamper with the usb implementation because I'd have 
to update our current usb driver to fit into the new system, and I'm getting 
flack about wasting time trying to update that thing already . . . So if you 
tell me no, I can go tell my boss I have to update the usb driver.


Anton

Re: 2.4 O(1) scheduler

[Austin Gonyou]

Been using 2.4.19-rc1-aa2 on a fairly large x86 box for about 16 days
now. Nothing but love. Using the 0/1 scheduler, though I'm not sure if I
could tweak it for "better" performance.(hint: like to see some
"best-practices" type of doc, for 4/8-way SMP boxen)

Aside from that, I've got 8GB ram, a ~750GB Oracle instance running, and
4GB SHMMAX attatched to some Copper FC1 disks and using QLA2200's. 

It's been very happy since rc1. Anyway, it's worth a shot. As soon as
2.4.19 is "done" this box will go into production as soon as this RC2 VM
stuff is cleared up.


On Fri, 2002-07-19 at 15:17, anton wilson wrote:
> On Friday 19 July 2002 03:54 pm, J Sloan wrote:
> > Use 2.4-aa, 2.4-ac or 2.4-redhat kernel
> > and you get the O(1) secheduler at
> > no extra cost -
> >
> 
> 
> > Joe
> 
> 
> I'm actually worried not about just the O(1) scheduler but if these patches 
> will be incorporating the O(1) bug fixes such as the serious one in 
> balance_load where curr->next was used instead of current->prev. Also, I need 
> to use a patch that won't tamper with the usb implementation because I'd have 
> to update our current usb driver to fit into the new system, and I'm getting 
> flack about wasting time trying to update that thing already . . . So if you 
> tell me no, I can go tell my boss I have to update the usb driver.
> 
> 
> Anton
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
-- 
Austin Gonyou <austin@digitalroadkill.net>

Re: 2.4 O(1) scheduler

[J Sloan]

Well, I can't speak for Alan, Andrea or
Red Hat, but their respective kernels
are all likely to be kept pretty well up
to date on the current O(1) - that's
certainly the case AFAICT -

Joe

anton wilson wrote:

>On Friday 19 July 2002 03:54 pm, J Sloan wrote:
>  
>
>>Use 2.4-aa, 2.4-ac or 2.4-redhat kernel
>>and you get the O(1) secheduler at
>>no extra cost -
>>
>>    
>>
>
>
>  
>
>>Joe
>>    
>>
>
>
>I'm actually worried not about just the O(1) scheduler but if these patches 
>will be incorporating the O(1) bug fixes such as the serious one in 
>balance_load where curr->next was used instead of current->prev. Also, I need 
>to use a patch that won't tamper with the usb implementation because I'd have 
>to update our current usb driver to fit into the new system, and I'm getting 
>flack about wasting time trying to update that thing already . . . So if you 
>tell me no, I can go tell my boss I have to update the usb driver.
>
>
>Anton
>
>  
>

Re: 2.4 O(1) scheduler

[Ingo Molnar]

On Fri, 19 Jul 2002, anton wilson wrote:

> I'm actually worried not about just the O(1) scheduler but if these
> patches will be incorporating the O(1) bug fixes such as the serious one
> in balance_load where curr->next was used instead of current->prev.

It's a harmless bug, somewhat reducing the amount of balancing we can do
on SMP, but the balancer was still pretty much intact (we'd have noticed
it earlier if it wasnt). The bug was found by Scott Rhine and myself not
because the scheduler behaved badly, but via code review, because the
comments did not match the code :-)

	Ingo

Re: 2.4 O(1) scheduler

[Ingo Molnar]

On Fri, 19 Jul 2002, anton wilson wrote:

> I'm working on a project that uses the O(1) scheduler and I am forced to
> use a 2.4 kernel for time issues. Will the O(1) patches for 2.4 kernels
> be updated once the 2.5.26+ patch becomes stable?

you can find the latest 2.4 based O(1) scheduler patch at:

  http://redhat.com/~mingo/O(1)-scheduler/sched-2.4.19-rc2-A4

this patch includes the load_balance() fixes as well. (generally you
should check this directory regularly, sometimes i put out patches without
announcing them - like this one.)

	Ingo