mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Ingo Molnar <mingo@elte.hu>
To: Zwane Mwaikambo <zwane@linux.realnet.co.sz>
Cc: Robert Love <rml@mvista.com>, Alan Cox <alan@lxorguk.ukuu.org.uk>,
	"David S. Miller" <davem@redhat.com>,
	<linux-kernel@vger.kernel.org>
Subject: Re: [patch] 2.4.19-pre10-ac2: O(1) scheduler merge, -A3.
Date: Mon, 17 Jun 2002 11:00:26 +0200 (CEST)	[thread overview]
Message-ID: <Pine.LNX.4.44.0206171050050.9574-100000@e2> (raw)
In-Reply-To: <Pine.LNX.4.44.0206171012270.1263-100000@netfinity.realnet.co.sz>


On Mon, 17 Jun 2002, Zwane Mwaikambo wrote:

> > > Can we add a config time option for irqbalance? I consider it extra
> > > overhead for setups which can do the interrupt distribution via hardware
> > > properly, [...]
> > 
> > What x86 hardware do you have in mind?
> 
> ye olde generic x86 SMP box, the interrupt handling imbalance came about
> with the P4 and their newer APIC setup did it not? Although i am aware
> that some x86 SMP boxes don't do the distribution properly too, thats
> why i reckon config option would be best.

even generic x86 SMP boxes benefit from irqbalance due to better irq
affinity. Actually one could hardly find a worse way to distribute
interrupts than the IO-APIC + lowest priority delivery mode does... [in
fact there is one, the P4's do it ;-) ]

> > > [...] also irqbalance breaks NUMAQ horribly seeing as it assumes a
> > > number of things like addressing modes.
> > 
> > exactly what does it assume that breaks NUMAQ?
> 
> <Disclaimer>
> I am not a NUMAQ expert and do not even have access to one for testing
> </Disclaimer>
> 
> The addressing mode, irq_balance assumes that the addressing mode is
> logical mode (when programming the IOREDTBL entries), whilst NUMAQ uses
> a completely different addressing architecture. [...]

irqbalance uses the set_ioapic_affinity() method to set affinity. The
clustered APIC code is broken if it doesnt handle this properly. (i dont
have such hardware so i cant tell, but it indeed doesnt appear to handle
this case properly.) By wrapping around at node boundary the irqbalance
code will work just fine.

> [...] Also another thing is consider this situation;
> 
> irqbalance programs IOAPIC#0 on node0 to deliver to CPU#6 on node1
> 
> Will that interrupt get delivered?

i agree that this could be a problem, but set_ioapic_affinity() can be
made dependent on the actual NUMA setup that is used. This is absolutely
needed anyway for a proper /proc/irq/*/smp_affinity feature.

	Ingo


  reply	other threads:[~2002-06-17  9:02 UTC|newest]

Thread overview: 36+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2002-06-13 19:21 [PATCH] 2.4-ac: sparc64 support for O(1) scheduler Robert Love
2002-06-14  4:25 ` David S. Miller
2002-06-14 17:32   ` Robert Love
2002-06-15 13:22     ` David S. Miller
2002-06-20 19:42       ` Alan Cox
2002-06-16 15:19     ` Ingo Molnar
2002-06-16 17:00       ` [patch] 2.4.19-pre10-ac2: O(1) scheduler merge, -A3 Ingo Molnar
2002-06-16 23:57         ` Robert Love
2002-06-17  0:13           ` J.A. Magallon
2002-06-17  4:28             ` Ingo Molnar
2002-06-17  0:15           ` Robert Love
2002-06-17  3:49             ` Ingo Molnar
2002-06-17  3:57               ` Robert Love
2002-06-17  4:07                 ` Ingo Molnar
2002-06-17  4:02               ` Robert Love
2002-06-17  4:26                 ` Ingo Molnar
2002-06-17  4:49                 ` [patch] 2.5.22 current scheduler bits #1 Ingo Molnar
2002-06-17  3:24           ` [patch] 2.4.19-pre10-ac2: O(1) scheduler merge, -A3 Ingo Molnar
2002-06-17  3:35             ` Robert Love
2002-06-17  4:01               ` Ingo Molnar
2002-06-17  7:50             ` Zwane Mwaikambo
2002-06-17  8:32               ` Ingo Molnar
2002-06-17  8:23                 ` Zwane Mwaikambo
2002-06-17  9:00                   ` Ingo Molnar [this message]
2002-06-17  9:34                     ` Zwane Mwaikambo
2002-06-18  7:16                     ` William Lee Irwin III
2002-06-19  1:05                       ` Matthew Dobson
2002-06-20 20:22                         ` Andrew Theurer
2002-06-24  0:16                       ` Martin J. Bligh
2002-06-17 16:26             ` Rusty Russell
2002-06-17  4:51           ` Toshiba PCToPIC97 PC Card freeze in 2.4.18 Stephen Satchell
2002-06-16 23:45       ` [PATCH] 2.4-ac: sparc64 support for O(1) scheduler Robert Love
2002-06-17  5:28         ` David S. Miller
2002-06-17 21:18           ` Robert Love
2002-06-14 22:00   ` Thomas Duffy
2002-06-15 13:35     ` David S. Miller

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=Pine.LNX.4.44.0206171050050.9574-100000@e2 \
    --to=mingo@elte.hu \
    --cc=alan@lxorguk.ukuu.org.uk \
    --cc=davem@redhat.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=rml@mvista.com \
    --cc=zwane@linux.realnet.co.sz \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®