From: "Liu, Yong" <yong.liu@intel.com>
To: "xu,huilong" <huilongx.xu@intel.com>, dts@dpdk.org
Cc: lei yao <lei.a.yao@intel.com>
Subject: Re: [dts] [PATCH V2] framework: add server mode and multi queue support for qemu framework
Date: Tue, 20 Dec 2016 08:25:57 +0800 [thread overview]
Message-ID: <58587A95.90603@intel.com> (raw)
In-Reply-To: <1482132845-6802-1-git-send-email-huilongx.xu@intel.com>
Thanks, applied.
On 12/19/2016 03:34 PM, xu,huilong wrote:
> From: lei yao <lei.a.yao@intel.com>
>
> 1. Server mode In vhost reconnection test, we need set qemu
> work in server mode(default is client mode) The sample command is :
> "-chardev socket,id=char0,path=./vhost-net1,server"
> 2. Multiqueue
> We can set multiqueue connection between vhost and virtio. Add this parameter for the
> multi-queue performance testing. The sample command is: "-netdev
> type=vhost-user,id=mynet1,chardev=char0,vhostforce,queues=4"
> 3. Change the VM launch waiting time.Some times,we find that lauching vm need more time.
> So change count=20 to count 40
>
> Signed-off-by: lei yao <lei.a.yao@intel.com>
> ---
> conf/vhost_sample.cfg | 59 +++++++++++++++++++++++++++++++++++++++++++++++++++
> framework/qemu_kvm.py | 24 +++++++++++++++------
> 2 files changed, 76 insertions(+), 7 deletions(-)
>
> diff --git a/conf/vhost_sample.cfg b/conf/vhost_sample.cfg
> index ce34af8..8bdc20e 100644
> --- a/conf/vhost_sample.cfg
> +++ b/conf/vhost_sample.cfg
> @@ -1,4 +1,63 @@
> +# QEMU options
> +# name
> +# name: vm0
> +#
> +# enable_kvm
> +# enable: [yes | no]
> +#
> +# cpu
> +# model: [host | core2duo | ...]
> +# usage:
> +# choose model value from the command
> +# qemu-system-x86_64 -cpu help
> +# number: '4' #number of vcpus
> +# cpupin: '3 4 5 6' # host cpu list
> +#
> +# mem
> +# size: 1024
> +#
> +# disk
> +# file: /path/to/image/test.img
> +#
> +# char
> +# opt_path: define the file path to vhost-net socket file
> +# opt_server: define the qemu socket connection work at server mode or client mdoe. Default is client
> +#
> +# net
> +# type: [vhost-user | ...]
> +# vhost-user
> +# opt_queue: queue=n, used for multi-queue connection between vhost and virtio
> +#
> +# device
> +# driver: [ virtio-net-pci | ...]
> +# virtio-net-pci
> +# prop_netdev: netdev0
> +# opt_mac: 00:00:00:00:01:03
> +#
> +# monitor
> +# port: 6061
> +# note: if adding monitor to vm, need to specicy
> +# this port, else it will get a free port
> +# on the host machine.
> +#
> +# qga
> +# enable: [yes | no]
> +#
> +# serial_port
> +# enable: [yes | no]
> +#
> +# vnc
> +# displayNum: 1
> +# note: you can choose a number not used on the host.
> +#
> +# daemon
> +# enable: 'yes'
> +# note:
> +# By default VM will start with the daemonize status.
> +# Not support starting it on the stdin now.
> +
> # vm configuration for vhost sample case
> +
> [vm0]
> cpu =
> model=host,number=2,cpupin=24 25;
> diff --git a/framework/qemu_kvm.py b/framework/qemu_kvm.py
> index 6ced54a..e9d29cc 100644
> --- a/framework/qemu_kvm.py
> +++ b/framework/qemu_kvm.py
> @@ -606,21 +606,31 @@ class QEMUKvm(VirtBase):
> if 'opt_path' in options.keys() and options['opt_path']:
> dev_boot_line = '-chardev socket'
> char_id = 'char%d' % self.char_idx
> - dev_boot_line += separator + 'id=%s' % char_id + separator + 'path=%s' % options['opt_path']
> - self.char_idx += 1
> - self.__add_boot_line(dev_boot_line)
> + if 'opt_server' in options.keys() and options['opt_server']:
> + dev_boot_line += separator + 'id=%s' % char_id + separator + 'path=%s' %options['opt_path'] + separator + '%s' % options['opt_server']
> + self.char_idx += 1
> + self.__add_boot_line(dev_boot_line)
> + else:
> + dev_boot_line += separator + 'id=%s' % char_id + separator + 'path=%s' %options['opt_path']
> + self.char_idx += 1
> + self.__add_boot_line(dev_boot_line)
> # netdev parameter
> netdev_id = 'netdev%d' % self.netdev_idx
> self.netdev_idx += 1
> - dev_boot_line = '-netdev type=vhost-user,id=%s,chardev=%s,vhostforce' % (netdev_id, char_id)
> + if 'opt_queue' in options.keys() and options['opt_queue']:
> + queue_num=options['opt_queue']
> + dev_boot_line = '-netdev type=vhost-user,id=%s,chardev=%s,vhostforce,queues=%s' % (netdev_id, char_id,queue_num)
> + else:
> + dev_boot_line = '-netdev type=vhost-user,id=%s,chardev=%s,vhostforce' % (netdev_id, char_id)
> self.__add_boot_line(dev_boot_line)
> # device parameter
> opts = {'opt_netdev': '%s' % netdev_id}
> if 'opt_mac' in options.keys() and \
> options['opt_mac']:
> opts['opt_mac'] = options['opt_mac']
> -
> - self.__add_vm_virtio_net_pci(**opts)
> + if 'opt_settings' in options.keys() and options['opt_settings']:
> + opts['opt_settings'] = options['opt_settings']
> + self.__add_vm_virtio_net_pci(**opts)
>
> def __add_vm_virtio_cuse_pci(self, **options):
> """
> @@ -939,7 +949,7 @@ class QEMUKvm(VirtBase):
> wait for 120 seconds for vm net ready
> 10.0.2.* is the default ip address allocated by qemu
> """
> - count = 20
> + count = 40
> while count:
> out = self.__control_session('ifconfig')
> if "10.0.2" in out:
prev parent reply other threads:[~2016-12-20 0:28 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-12-19 7:34 xu,huilong
2016-12-20 0:25 ` Liu, Yong [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=58587A95.90603@intel.com \
--to=yong.liu@intel.com \
--cc=dts@dpdk.org \
--cc=huilongx.xu@intel.com \
--cc=lei.a.yao@intel.com \
/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).