[P4-dev] Questions concerning BMv2

hemant at mnkcg.com hemant at mnkcg.com
Thu May 7 09:32:38 EDT 2020



Regarding item 2, please read the documentation for bmv2 here:


Search on the page for “multicast” and you will see the reference to
v1model.p4 which restricts the multicast id to 16 bits.  I think, v1model.p4
is frozen for any change.  If you play with p4c and the bmv2 switch in your
private sandbox, you will have to change both p4c and the behavioral-model.
See p4c/p4include/psa.p4 which uses 32 bits for multicast id.  However, the
simple_switch for PSA is work in progress and may not support all features
that v1model.p4 simple_switch supports.




From: P4-dev <p4-dev-bounces at lists.p4.org> On Behalf Of Michael Höflmaier
Sent: Thursday, May 07, 2020 5:23 AM
To: p4-dev at lists.p4.org
Subject: [P4-dev] Questions concerning BMv2


Dear p4-devs,

I have two questions concerning BMv2:


1.	What types of interfaces can be used with BMv2? I already asked this
on BMv2-GitHub, but no one had any experience on this question. I already
experimented by passing an VCAN-interfaces to the simple-switch -i argument
and this succeeded. Nevertheless, is it possible to pass any kind of
interface-type within     the Linux-kernel to BMv2 or are there any
restriction? Is there any documentation on this topic at all, because I
found none.
2.	Is it possible to extend the number of multicast-groups beyond 16
bit? If yes: Where can I adjust this? I already tried to find evidence on
this cap, but found none.

I would be very grateful if your experience could help me resolve my

Best regards
Michael Höflmaier 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.p4.org/pipermail/p4-dev_lists.p4.org/attachments/20200507/8d3d942f/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 1594 bytes
Desc: not available
URL: <http://lists.p4.org/pipermail/p4-dev_lists.p4.org/attachments/20200507/8d3d942f/attachment.p7s>

More information about the P4-dev mailing list