All of lore.kernel.org
 help / color / mirror / Atom feed
From: Anthony Liguori <anthony@codemonkey.ws>
To: qemu-devel@nongnu.org
Cc: Juergen Lock <nox@jelal.kn-bremen.de>,
	Jung-uk Kim <jkim@FreeBSD.org>,
	Sebastian Herbszt <herbszt@gmx.de>
Subject: Re: [Qemu-devel] Re: [PATCH] networking using libpcap
Date: Fri, 25 Jul 2008 15:51:12 -0500	[thread overview]
Message-ID: <488A3CC0.60802@codemonkey.ws> (raw)
In-Reply-To: <200807181639.49136.jkim@FreeBSD.org>

Jung-uk Kim wrote:
> On Friday 18 July 2008 12:12 pm, Sebastian Herbszt wrote:
>
> I simplified FreeBSD ports patch with the above patch:
>
> http://people.freebsd.org/~jkim/qemu-devel-20080620_1-pcap.diff
>   

I spent some time with this patch looking to merge it.  It seems to have 
decent performance and is very easy to use.  Unfortunately, I ran into 
the following:

1) If I specify -net pcap,ifname=eth0 -net nic,model=rtl8139, I get an 
error, whereas if I specify -net nic,model=rtl8139 -net pcap,ifname=eth0 
it works.  The reason for this is that in net_pcap_init(), you search 
the vlan to discover what the IP address is of the NIC on the vlan is.

This is a big no-no as you can have one or more NICs on a single vlan.  
Moreover, NICs can be added and removed from a vlan long after it's 
created.  The patch needs a fair bit more work to handle this proper 
(presumably building a new filter rule as NICs are added and removed 
from the vlan).

2) When using ifname=eth0, traffic doesn't work from host=>guest nor 
from guest=>host.  This seems like a major short-coming to me.  Having 
another networking option that is easy to use but doesn't work in some 
fundamental way that a user would expect seems like it's just going to 
lead to even further confusion.  This patch isn't attractive to me 
unless it works just like you were bridging to a physical interface (and 
I don't think there's anything fundamental preventing that).

If you can fix these issues, I'd be happy to merge this feature.

Regards,

Anthony Liguori

> Thanks!
>
> Jung-uk Kim
>
>
>   

  parent reply	other threads:[~2008-07-25 20:51 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-07-17 22:12 [Qemu-devel] Re: [PATCH] networking using libpcap Jung-uk Kim
2008-07-18 16:12 ` Sebastian Herbszt
2008-07-18 20:39   ` Jung-uk Kim
2008-07-18 23:07     ` Jung-uk Kim
2008-07-21 15:35     ` Sebastian Herbszt
2008-07-25 20:51     ` Anthony Liguori [this message]
2008-07-26 17:30       ` Sebastian Herbszt
2008-07-27  0:28         ` Anthony Liguori
2008-07-27 12:55       ` Paul Brook
2008-07-23  1:55 ` Anthony Liguori
  -- strict thread matches above, loose matches on Subject: below --
2008-07-02 15:02 [Qemu-devel] " Ulrich Hecht
2008-07-02 23:39 ` [Qemu-devel] " Sebastian Herbszt

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=488A3CC0.60802@codemonkey.ws \
    --to=anthony@codemonkey.ws \
    --cc=herbszt@gmx.de \
    --cc=jkim@FreeBSD.org \
    --cc=nox@jelal.kn-bremen.de \
    --cc=qemu-devel@nongnu.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.