3:01 p.m.
Hi,
I am trying to write a configuration file for the following setup, but
unsure if it is supported...
I have a physical interface eth0, on which 2 vlans, say eth0.1 &
eth0.2, are manually created before starting connman.
I need to assign a static link-layer address on eth0.1 and use dhcp on
eth0.2. I know we assign a configuration to a specific interface using
the MAC address, but as the vlan's interfaces have the same MAC address
as the actual physical interface, is it possible to match a
configuration to a specific vlan interface?
Thanks in advance for your support!
--
Maxime Langlois
9:05 a.m.
Hi Maxime,
On Tue, Apr 02, 2019 at 01:01:46PM +0000, Langlois, Maxime wrote:
I need to assign a static link-layer address on eth0.1 and use dhcp
on
eth0.2. I know we assign a configuration to a specific interface using
the MAC address, but as the vlan's interfaces have the same MAC address
as the actual physical interface, is it possible to match a
configuration to a specific vlan interface?
You are right. Currently the only way ConnMan matches the config
entries against the interfaces is by using the MAC Address. Adding
another way to reliable match them isn't that hard to add. I just
don't know if there is a good/reliable way to identify the VLAN
interfaces.
Maybe with SIOCGIFPFLAGS could help here. Just an idea.
Thanks,
Daniel
3:05 p.m.
Hi Daniel,
couldn't we use the interface name for matching? That seems a reliable
way to me.
I am developing an embedded software and when deploying the software in
production, the only thing that I know for sure will match is the
interfaces name, as the MAC addresses will obviously be different from
board to board.
Currently, if I wanted to release my software with connman, I would
have to add a script running before connman starts in order to tweak
the config file to put the board specific MAC addresses in there, so I
end up with the expected config. And I would get the correct MAC using
the interfaces name!
I guess I am missing something because I don't understand why it is not
already supported!
Thanks for your answer,
Maxime
8:47 p.m.
Hi Maxime,
On Thu, Apr 04, 2019 at 01:05:34PM +0000, Langlois, Maxime wrote:
couldn't we use the interface name for matching? That seems a
reliable
way to me.
In the past some argued that only the MAC address is kind of reliable
to match. Obviously, this is not really true. Changing a MAC address
is as simple has changing an interface (udev rules).
I am developing an embedded software and when deploying the software
in
production, the only thing that I know for sure will match is the
interfaces name, as the MAC addresses will obviously be different from
board to board.
This seems to be a common use case for embedded systems. So yes, I
agree we should just add an additional match for the interface names.
Care to send a patch?
Thanks,
Daniel
10:25 p.m.
Hi Daniel,
would love to contribute! However I haven't planned that extra time for
diving into the connman source code... and deadlines being what they
are, I might have to move to an alternative :(
Decision is not yet made though, as I still have other things to
evaluate. Our networking requirements are quite complex and I'm still
trying to figure every bits and pieces!
If it ever occurs that I retain the connman solution, I certainly might
submit the patches.
Thanks,
Maxime
**********************************************************************
DISCLAIMER:
Privileged and/or Confidential information may be contained in this message. If you are
not the addressee of this message, you may not copy, use or deliver this message to
anyone. In such event, you should destroy the message and kindly notify the sender by
reply e-mail. It is understood that opinions or conclusions that do not relate to the
official business of the company are neither given nor endorsed by the company. Thank
You.
8:27 a.m.
Hi Maxime,
On 4/4/19 10:25 PM, Langlois, Maxime wrote:
would love to contribute! However I haven't planned that extra
time for
diving into the connman source code... and deadlines being what they
are, I might have to move to an alternative :(
Fair enough. Anyway, you should plan some time to fix/maintain any open
source project you choose. There is no free lunch.
The change isn't too hard to implement and test. I'll do it in the
coming days.
Decision is not yet made though, as I still have other things to
evaluate. Our networking requirements are quite complex and I'm still
trying to figure every bits and pieces!
Sure :)
If it ever occurs that I retain the connman solution, I certainly
might
submit the patches.
That is good to hear!
Thanks,
Daniel
8:51 a.m.
On Mon, Apr 15, 2019 at 11:27 PM Daniel Wagner <wagi(a)monom.org> wrote:
Hi Maxime,
On 4/4/19 10:25 PM, Langlois, Maxime wrote:
> would love to contribute! However I haven't planned that extra time for
> diving into the connman source code... and deadlines being what they
> are, I might have to move to an alternative :(
Fair enough. Anyway, you should plan some time to fix/maintain any open
source project you choose. There is no free lunch.
The change isn't too hard to implement and test. I'll do it in the
coming days.
> Decision is not yet made though, as I still have other things to
> evaluate. Our networking requirements are quite complex and I'm still
> trying to figure every bits and pieces!
Sure :)
> If it ever occurs that I retain the connman solution, I certainly might
> submit the patches.
That is good to hear!
Thanks,
Daniel
VLAN support was added a few years ago
(https://git.kernel.org/pub/scm/network/connman/connman.git/commit/?id=f48...).
At the time, I also posted a patch that would have Connman also
manager the creation and deletion of VLAN interfaces.
-Justin
11:16 a.m.
Hi Justin,
On 29.04.19 08:51, Justin Maggard wrote:
VLAN support was added a few years ago
(https://git.kernel.org/pub/scm/network/connman/connman.git/commit/?id=f48...).
At the time, I also posted a patch that would have Connman also
manager the creation and deletion of VLAN interfaces.
Thanks for the pointer. The missing functionality is in the provisiong
code. Currently you can't write a provisioning file which matches on
device names, only on MAC addresses. This evening I should have time to
write the patch for this.
Thanks,
Daniel
7:41 p.m.
On Mon, 2019-04-29 at 21:30 +0200, Daniel Wagner wrote:
Just posted a patch which adds support to match on device name.
Please
if possible give them a spin and report back.
Hi Daniel,
sure I can give it a go as soon as I have a chance! Will keep you posted...
Thanks!
- Maxime
**********************************************************************
DISCLAIMER:
Privileged and/or Confidential information may be contained in this message. If you are
not the addressee of this message, you may not copy, use or deliver this message to
anyone. In such event, you should destroy the message and kindly notify the sender by
reply e-mail. It is understood that opinions or conclusions that do not relate to the
official business of the company are neither given nor endorsed by the company. Thank
You.
545
days inactive
573
days old
9 comments
3 participants
participants (3)
-
Daniel Wagner -
Justin Maggard -
Langlois, Maxime