I scheduled a meeting on the fdio calendar for 7-8am Thursday Pacific 11/3.
Webex will be forthcoming.
This is per action item from last week's meeting.
Cheers
Keith.
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp
Hi all,
regarding the discussed changes, here's a patch:
https://gerrit.fd.io/r/#/c/3641/
please do check out:
make test-help
to get a list of knobs to turn with a brief explanation. I got rid of
all the different one-letter switches (kept only the familiar
V=) plus added a stepping feature re
We cannot get NDR at any rate in multi-threads mode for VPP 16.09. The
following is our config.
(1)Version : VPP 16.09; multi-threads mode
(2)CPU config. in vpp startup.conf
cpu {
main-core 0 corelist-workers 1-2
}
(3)Show output
vpp# show thread
ID NameT
Hello
I have a question about IO thread mode in VPP. is there any strong reason(s)
why IO threads are deprecated in 16.09? This mode may be useful in some
scenarios.
Thanks
Ming
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailm
Folks,
I've pushed an API reorg draft to the gerrit server. I'd be glad to add anyone
as a reviewer who's interested in looking at it.
It's a draft to avoid wasting Jenkins cycles. The jvpp build doesn't currently
work.
I'd appreciate it if someone who understands the jvpp build would review t
Dear Ole,
Yes, the reorganization I have in mind will yield a single API composed from N
.api / .c files. As I wrote, the reorganization is about reducing patch merge /
rebase pain; to a lesser extent about api placement within the vpp source tree.
My plan is to yank apart vpe.api / api.c, for
Currently there are no semantic dependencies:
-lisp_adjacency is used just by lisp
-counter types by vnet_interface_counters,
But in the future it would be beneficial to have for example 'string' or
'ip-address' type.
It would make Java/Python and perhaps other language binding
The reorganization that I described has to do with reducing merge conflicts on
.api files, and to placing API definitions and their message handlers near the
feature code that they control.
None of this has to do with semantic dependency of any kind. Except between
foo_t and foo_reply_t - which
Hi,
Great idea! It would make Java/Python APIs more manageable as well => we could
group related APIs in packages/namespaces.
But have you considered dependencies between the groups?
It could be useful for example to be able to define a custom type that is used
in two different groups.
Regards
> It’s an SMT [simple matter of typing] to create a number xxx.api files,
> related message handler implementation files, and to place them next to
> whatever the indicate API collection controls.
>
> The current setup is a historical artifact.
I'll have a go at the MAP APIs ASAP.
(Been on my
Thanks Keith, I reviewed VPP Init but could not find what I want. I mean
what vnet/vnet/config.h api provides?!
On Sun, Oct 30, 2016 at 11:35 PM, Keith Burns wrote:
> Check out the session on VPP Init from https://wiki.fd.io/view/
> Events/Training-2016-04-Content
>
> On Oct 30, 2016 7:13 AM, "m
11 matches
Mail list logo
