From crobinso at redhat.com Tue May 17 08:33:14 2016
Content-Type: multipart/mixed; boundary="===============2540155458390397085=="
MIME-Version: 1.0
From: Cole Robinson <crobinso at redhat.com>
To: devel at lists.libvirt.org
Subject: Re: [libvirt] [PATCH] lxc: support <interface type='ethernet'>
Date: Tue, 17 May 2016 08:33:12 -0400
Message-ID: <dd90dc48-ed7b-0308-77fc-ea3f2e472fc2@redhat.com>
In-Reply-To: CACaajQvkhDeFx+26gxgFWWmJCxYU24TdpCkKYnr=fGEf_Fpv1w@mail.gmail.com

--===============2540155458390397085==
Content-Type: text/plain; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: quoted-printable

On 05/17/2016 07:48 AM, Vasiliy Tolstov wrote:
> 2016-05-17 1:53 GMT+03:00 Cole Robinson <crobinso(a)redhat.com>:
>> I suggest explicitly rejecting the <script> option then, since at least =
the
>> formatdomain docs make it sound like it's the only valid type=3Dethernet=
 config.
>> virLXCProcessSetupInterfaceTap seems like the only shared place for it t=
hough
>>
>> Also, this is https://bugzilla.redhat.com/show_bug.cgi?id=3D1325687
>>
>> Patch looks fine otherwise, ACK if it's simple to reject the script bit,=
 up to you
> =

> =

> I think that script may be useful.... May be simple utilize logic to
> run script like in qemu?
> =


I think Laine was saying 'I personally don't care about ->script so I'm not
planning on implementing it', so if it isn't implemented we should error
explicitly rather than silently ignore it.

That said it's probably not hard to get ->script to work since we already h=
ave
the plumbing as you mention, so 'patches welcome' :)

- Cole


--===============2540155458390397085==--