public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Mike Travis <travis@sgi.com>
To: Bian LuLu <helianthus.lu@gmail.com>
Cc: Bjorn Helgaas <bhelgaas@google.com>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	"linux-pci@vger.kernel.org" <linux-pci@vger.kernel.org>
Subject: Re: [RFC] The meaning of local_cpulist and local_cpus
Date: Thu, 25 Apr 2013 07:17:52 -0700	[thread overview]
Message-ID: <51793B10.7060603@sgi.com> (raw)
In-Reply-To: <CAG82n_kBYQpCRDmY7ghivXY60hBMF_5HHZUAT1d-xWD9OZvdrg@mail.gmail.com>



On 4/25/2013 6:57 AM, Bian LuLu wrote:
> Thankyou .
> But I still can not understand what is the meaning of cuplist.
> my linux version is ubuntu 12.04, and
> my system:/$ cat /sys/devices/pci0000:00/0000:00:1a.0/local_cpulist
>  the result is '0-7',

Ahh, sorry, I misinterpreted what you are asking.  On NUMA systems,
I/O can be spread over many nodes and DMA/interrupts are faster from the
device to memory and cpus local to that node.  The numa_node shows
which node the device is on, and the cpulist shows which cpus are local
to that same node.  In general you also want the IRQ affinity to be
the local cpus as well.

> 
> I just don't konw what are they represent .Could  you give me more
> detailed information about local_cpulist and local_cpus,and their
> relations.
> thanks in advance :)
> 
> On Thu, Apr 25, 2013 at 12:56 AM, Mike Travis <travis@sgi.com> wrote:
>>
>>
>> On 4/24/2013 9:48 AM, Bjorn Helgaas wrote:
>>> [+cc linux-pci, Mike]
>>>
>>> On Wed, Apr 24, 2013 at 4:06 AM, Bian LuLu <helianthus.lu@gmail.com> wrote:
>>>> Hi all,
>>>>
>>>> Recently, i read some codes of PCI portions. I think
>>>> local_cpulist is a list about one kind of CPU and
>>>> local_cpus is a mask of CPU. But i am not sure when
>>>> and how i should use these two parameters.
>>>>
>>>> See http://lxr.linux.no/linux+v3.5.4/drivers/pci/pci-sysfs.c#L390 for
>>>> details.
>>>>
>>>> Would anyone please give me some suggestions?
>>>> Thanks in advance ;-)
>>>
>>> I don't know off-hand, but maybe Mike or somebody on linux-pci does.
>>> It looks like Mike added local_cpulist with 39106dcf85.
>>>
>>
>> It primarily comes into play when you have a large # of cpus.
> I can not understand the symbol of '#' ,what is the meaning of it?
>> Here's the difference on a system that has 1024 cpu threads:
>>
>> harp31-sys:/sys/devices/system/node/node20 # cat cpulist
>> 160-167,672-679
>> harp31-sys:/sys/devices/system/node/node20 # cat cpumap
>> 00000000,00000000,00000000,00000000,00000000,00000000,00000000,
>> 00000000,00000000,00000000,000000ff,00000000,00000000,00000000,
>> 00000000,00000000,00000000,00000000,00000000,00000000,00000000,
>> 00000000,00000000,00000000,00000000,00000000,000000ff,00000000,
>> 00000000,00000000,00000000,00000000
>>
>> Which is easier to interpret?  :)
>>
>> But there are some older user side utilities that still
>> use the mask format.

      reply	other threads:[~2013-04-25 14:17 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-04-24 10:06 [RFC] The meaning of local_cpulist and local_cpus Bian LuLu
2013-04-24 16:48 ` Bjorn Helgaas
2013-04-24 16:56   ` Mike Travis
2013-04-25 13:57     ` Bian LuLu
2013-04-25 14:17       ` Mike Travis [this message]

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=51793B10.7060603@sgi.com \
    --to=travis@sgi.com \
    --cc=bhelgaas@google.com \
    --cc=helianthus.lu@gmail.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-pci@vger.kernel.org \
    /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