Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 28 Jun 2014 14:07:52 +0100
From:      Carlos Ferreira <carlosmf.pt@gmail.com>
To:        Luigi Rizzo <rizzo@iet.unipi.it>
Cc:        "freebsd-net@freebsd.org" <freebsd-net@freebsd.org>, Prashant Upadhyaya <praupadhyaya@gmail.com>
Subject:   Re: netmap
Message-ID:  <CAJpYY6VBtynBeSQkVprz%2B6BY=TF9RcGEV9hWWKnVQzUvfpKh%2BQ@mail.gmail.com>
In-Reply-To: <CA%2BhQ2%2Bgi9SiZuM4Nvw4CjJG9sX3cNZ4KNfXSkDcFfcPvW4A22Q@mail.gmail.com>
References:  <CAJpYY6Wj6NkYVdPzitswLQ2oVoX9JAv5tCnzbP6icsgJTdfm8Q@mail.gmail.com> <CA%2BhQ2%2BhHz2FKUkCRnHiGAVxZqxY5WHPPo87mVgSnSmPzovNLWw@mail.gmail.com> <CAJpYY6V6Ewc%2BXQYBi3BQbFtWgLiFKN_qmBLjAx1HyWhB9Q0vcg@mail.gmail.com> <CAJpYY6VJ9ih1ON1s1mnzrc-P7Xa_vsVS3UC49gnii0cLX%2B=TNA@mail.gmail.com> <20140614101523.GD22187@onelab2.iet.unipi.it> <CAJpYY6Usfv98OEsgXedP5MCqdJsd_rz7N4EhgRmP2NcxZ9eQ9g@mail.gmail.com> <CA%2BhQ2%2Bj4RrCxjSu79Zt9BVnKUTcRxO615msGH5_uoLrEinbcPQ@mail.gmail.com> <CAJpYY6XBg8=g7sjVbu%2B6%2BEG1-vW1dymsvD8pL-EB08B=_z=usA@mail.gmail.com> <CAPBAu3U9_VTUCPk_61maSAyxM_VHgSui8kbT65=b-jsXNzqJcA@mail.gmail.com> <CA%2BhQ2%2Bgi9SiZuM4Nvw4CjJG9sX3cNZ4KNfXSkDcFfcPvW4A22Q@mail.gmail.com>

next in thread | previous in thread | raw e-mail | index | archive | help
Hello to all.
Unfortunately, I have been unable to do the test that Luigi requested due
to the lack of hardware. All the Gateworks Cambria SBC that I have are
currently allocated to another project. Only in about one week, will I have
the opportunity to test it.
As soon as I have news, I will post them. :)


On 28 June 2014 09:24, Luigi Rizzo <rizzo@iet.unipi.it> wrote:

>
>
> On Saturday, June 28, 2014, Prashant Upadhyaya <praupadhyaya@gmail.com>
> wrote:
>
>> Hi,
>>
>> Any further news ?
>>
>> Professor Luigi, one question regarding the tx with netmap.
>> Whenever, I write a packet from user space into netmap rings, and if I
>> want netmap to send this out immediately, do I necessarily have to do
>> a ioctl(fd, NIOCTXSYNC, NULL) ?
>>
>
> Yes it is up to the application to decide when to push packets out with a
> txsync or select() or poll(), and unfortunately there is a tradeoff betwe=
en
> efficiency and latency
>
> Cheers
> Luigi
>
>
>>
>> I have an application which receives packets, does some processing and
>> then sends them out. If I keep doing ioctl's on every packet send, then
>> there will be too may system calls hitting performance, the application
>> can't afford to block it has to return back to polling for the receipt o=
f
>> next packet.
>>
>> On the receive side, I see that I don't have a problem because I can pol=
l
>> the ring without initiating an RXSYNC and whenever in user space I find
>> that there is nothing on the ring (probably half way down the ring size)=
, I
>> do an RXSYNC to get more packets thus saving system calls.
>>
>> But on tx side, I have noticed that unless I do a TXSYNC, the packet doe=
s
>> not go out, please let me know if I am missing something.
>>
>> Regards
>> -Prashant
>>
>>
>>
>> On Tue, Jun 17, 2014 at 7:33 PM, Carlos Ferreira <carlosmf.pt@gmail.com>
>> wrote:
>>
>>> Great! :)
>>> I will give you the results as soon as I can get them :)
>>>
>>>
>>>
>>> On 17 June 2014 12:55, Luigi Rizzo <rizzo@iet.unipi.it> wrote:
>>>
>>> > On Mon, Jun 16, 2014 at 5:30 PM, Carlos Ferreira <
>>> carlosmf.pt@gmail.com>
>>> > wrote:
>>> >
>>> >> Ok, thanks for the enlightenment regarding the loss of performance.
>>> >>
>>> >> One question, just to be sure. Does the kernel module contains the
>>> VALA
>>> >> switch code? Or do I need to compile extra code to have the switch
>>> working?
>>> >> Also, where can I find the documentation to use the Vala Switch?
>>> >>
>>> >
>>> > =E2=80=8BVALE is part of the netmap kernel module, the only thing you=
 need
>>> > to know to use it is port names:
>>> > you can have multiple switch instances with multiple ports each,
>>> >
>>> > valeX:Y means port Y on switch X, X and Y are arbitrary strings
>>> > with the constraint that the whole name must fit 15 characters.
>>> >
>>> > Details in the netmap manpage
>>> >
>>> > cheers
>>> > luigi
>>> >
>>> >
>>>
>>>
>>> --
>>>
>>> Carlos Miguel Ferreira
>>> Researcher at Telecommunications Institute
>>> Aveiro - Portugal
>>> Work E-mail - cmf@av.it.pt
>>> Skype & GTalk -> carlosmf.pt@gmail.com
>>> LinkedIn -> http://www.linkedin.com/in/carlosmferreira
>>> _______________________________________________
>>> freebsd-net@freebsd.org mailing list
>>> http://lists.freebsd.org/mailman/listinfo/freebsd-net
>>> To unsubscribe, send any mail to "freebsd-net-unsubscribe@freebsd.org"
>>>
>>
>>
>
> --
> -----------------------------------------+-------------------------------
>  Prof. Luigi RIZZO, rizzo@iet.unipi.it  . Dip. di Ing. dell'Informazione
>  http://www.iet.unipi.it/~luigi/        . Universita` di Pisa
>  TEL      +39-050-2211611               . via Diotisalvi 2
>  Mobile   +39-338-6809875               . 56122 PISA (Italy)
> -----------------------------------------+-------------------------------
>
>


--=20

Carlos Miguel Ferreira
Researcher at Telecommunications Institute
Aveiro - Portugal
Work E-mail - cmf@av.it.pt
Skype & GTalk -> carlosmf.pt@gmail.com
LinkedIn -> http://www.linkedin.com/in/carlosmferreira



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAJpYY6VBtynBeSQkVprz%2B6BY=TF9RcGEV9hWWKnVQzUvfpKh%2BQ>