linux-crypto.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: "Michael S. Tsirkin" <mst@redhat.com>
To: Caleb Raitto <caleb.raitto@gmail.com>
Cc: herbert@gondor.apana.org.au, davem@davemloft.net,
	arei.gonglei@huawei.com, jasowang@redhat.com,
	netdev@vger.kernel.org, linux-crypto@vger.kernel.org,
	Caleb Raitto <caraitto@google.com>
Subject: Re: [PATCH v2 net-next 0/2] virtio_net: Expand affinity to arbitrary numbers of cpu and vq
Date: Sun, 12 Aug 2018 11:44:53 +0300	[thread overview]
Message-ID: <20180812114424-mutt-send-email-mst@kernel.org> (raw)
In-Reply-To: <20180810011828.199888-1-caleb.raitto@gmail.com>

On Thu, Aug 09, 2018 at 06:18:27PM -0700, Caleb Raitto wrote:
> From: Caleb Raitto <caraitto@google.com>
> 
> Virtio-net tries to pin each virtual queue rx and tx interrupt to a cpu if
> there are as many queues as cpus.
> 
> Expand this heuristic to configure a reasonable affinity setting also
> when the number of cpus != the number of virtual queues.
> 
> Patch 1 allows vqs to take an affinity mask with more than 1 cpu.
> Patch 2 generalizes the algorithm in virtnet_set_affinity beyond
> the case where #cpus == #vqs.

Acked-by: Michael S. Tsirkin <mst@redhat.com>

Let's see where does it take us.

> v2 changes:
> Renamed "virtio_net: Make vp_set_vq_affinity() take a mask." to
> "virtio: Make vp_set_vq_affinity() take a mask."
> 
> Tested:
> 
> # 16 vCPU, 16 queue pairs, Debian 9 recent net-next kernel, GCE
> 
> # Disable GCE scripts setting affinities during startup.
> #
> # Add the following to
> # /etc/default/instance_configs.cfg.template and reboot:
> [InstanceSetup]
> set_multiqueue = false
> 
> $ cd /proc/irq
> $ for i in `seq 24 60` ; do sudo grep ".*" $i/smp_affinity_list;  done
> 0-15
> 0
> 0
> 1
> 1
> 2
> 2
> 3
> 3
> 4
> 4
> 5
> 5
> 6
> 6
> 7
> 7
> 8
> 8
> 9
> 9
> 10
> 10
> 11
> 11
> 12
> 12
> 13
> 13
> 14
> 14
> 15
> 15
> 0-15
> 0-15
> 0-15
> 0-15
> 
> $ cd /sys/class/net/eth0/queues/
> $ for i in `seq 0 15` ; do sudo grep ".*" tx-$i/xps_cpus; done
> 0001
> 0002
> 0004
> 0008
> 0010
> 0020
> 0040
> 0080
> 0100
> 0200
> 0400
> 0800
> 1000
> 2000
> 4000
> 8000
> 
> # 16 vCPU, 15 queue pairs
> $ sudo ethtool -L eth0 combined 15
> 
> $ cd /proc/irq
> $ for i in `seq 24 60` ; do sudo grep ".*" $i/smp_affinity_list;  done
> 0-15
> 0-1
> 0-1
> 2
> 2
> 3
> 3
> 4
> 4
> 5
> 5
> 6
> 6
> 7
> 7
> 8
> 8
> 9
> 9
> 10
> 10
> 11
> 11
> 12
> 12
> 13
> 13
> 14
> 14
> 15
> 15
> 15
> 15
> 0-15
> 0-15
> 0-15
> 0-15
> 
> $ cd /sys/class/net/eth0/queues/
> $ for i in `seq 0 14` ; do sudo grep ".*" tx-$i/xps_cpus; done
> 0003
> 0004
> 0008
> 0010
> 0020
> 0040
> 0080
> 0100
> 0200
> 0400
> 0800
> 1000
> 2000
> 4000
> 8000
> 
> # 16 vCPU, 8 queue pairs
> $ sudo ethtool -L eth0 combined 8
> 
> $ cd /proc/irq
> $ for i in `seq 24 60` ; do sudo grep ".*" $i/smp_affinity_list;  done
> 0-15
> 0-1
> 0-1
> 2-3
> 2-3
> 4-5
> 4-5
> 6-7
> 6-7
> 8-9
> 8-9
> 10-11
> 10-11
> 12-13
> 12-13
> 14-15
> 14-15
> 9
> 9
> 10
> 10
> 11
> 11
> 12
> 12
> 13
> 13
> 14
> 14
> 15
> 15
> 15
> 15
> 0-15
> 0-15
> 0-15
> 0-15
> 
> $ cd /sys/class/net/eth0/queues/
> $ for i in `seq 0 7` ; do sudo grep ".*" tx-$i/xps_cpus; done
> 0003
> 000c
> 0030
> 00c0
> 0300
> 0c00
> 3000
> c000
> 
> # 15 vCPU, 16 queue pairs
> $ sudo ethtool -L eth0 combined 16
> $ sudo sh -c "echo 0 > /sys/devices/system/cpu/cpu15/online"
> 
> $ cd /proc/irq
> $ for i in `seq 24 60` ; do sudo grep ".*" $i/smp_affinity_list;  done
> 0-15
> 0
> 0
> 1
> 1
> 2
> 2
> 3
> 3
> 4
> 4
> 5
> 5
> 6
> 6
> 7
> 7
> 8
> 8
> 9
> 9
> 10
> 10
> 11
> 11
> 12
> 12
> 13
> 13
> 14
> 14
> 0
> 0
> 0-15
> 0-15
> 0-15
> 0-15
> 
> $ cd /sys/class/net/eth0/queues/
> $ for i in `seq 0 15` ; do sudo grep ".*" tx-$i/xps_cpus; done
> 0001
> 0002
> 0004
> 0008
> 0010
> 0020
> 0040
> 0080
> 0100
> 0200
> 0400
> 0800
> 1000
> 2000
> 4000
> 0001
> 
> # 8 vCPU, 16 queue pairs
> $ for i in `seq 8 15`; \
> do sudo sh -c "echo 0 > /sys/devices/system/cpu/cpu$i/online"; done
> 
> $ cd /proc/irq
> $ for i in `seq 24 60` ; do sudo grep ".*" $i/smp_affinity_list;  done
> 0-15
> 0
> 0
> 1
> 1
> 2
> 2
> 3
> 3
> 4
> 4
> 5
> 5
> 6
> 6
> 7
> 7
> 0
> 0
> 1
> 1
> 2
> 2
> 3
> 3
> 4
> 4
> 5
> 5
> 6
> 6
> 7
> 7
> 0-15
> 0-15
> 0-15
> 0-15
> 
> $ cd /sys/class/net/eth0/queues/
> $ for i in `seq 0 15` ; do sudo grep ".*" tx-$i/xps_cpus; done
> 0001
> 0002
> 0004
> 0008
> 0010
> 0020
> 0040
> 0080
> 0001
> 0002
> 0004
> 0008
> 0010
> 0020
> 0040
> 0080
> 
> Caleb Raitto (2):
>   virtio: Make vp_set_vq_affinity() take a mask.
>   virtio_net: Stripe queue affinities across cores.
> 
>  drivers/crypto/virtio/virtio_crypto_core.c |  4 +-
>  drivers/net/virtio_net.c                   | 46 ++++++++++++++--------
>  drivers/virtio/virtio_pci_common.c         |  7 ++--
>  drivers/virtio/virtio_pci_common.h         |  2 +-
>  include/linux/virtio_config.h              |  7 ++--
>  5 files changed, 39 insertions(+), 27 deletions(-)
> 
> -- 
> 2.18.0.597.ga71716f1ad-goog

      parent reply	other threads:[~2018-08-12  8:44 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-08-10  1:18 [PATCH v2 net-next 0/2] virtio_net: Expand affinity to arbitrary numbers of cpu and vq Caleb Raitto
2018-08-10  1:18 ` [PATCH net-next 1/2] virtio: Make vp_set_vq_affinity() take a mask Caleb Raitto
2018-08-10  1:38   ` Gonglei (Arei)
2018-08-11 19:03 ` [PATCH v2 net-next 0/2] virtio_net: Expand affinity to arbitrary numbers of cpu and vq David Miller
2018-08-12  8:44 ` Michael S. Tsirkin [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=20180812114424-mutt-send-email-mst@kernel.org \
    --to=mst@redhat.com \
    --cc=arei.gonglei@huawei.com \
    --cc=caleb.raitto@gmail.com \
    --cc=caraitto@google.com \
    --cc=davem@davemloft.net \
    --cc=herbert@gondor.apana.org.au \
    --cc=jasowang@redhat.com \
    --cc=linux-crypto@vger.kernel.org \
    --cc=netdev@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;
as well as URLs for NNTP newsgroup(s).