https://docs.fd.io/csit/master/trending/trending/ip4-2n-clx-xxv710.html#t1c
native AVF driver 22.5 Mpps vs DPDK i40e 17.5Mpps on the same hardware.
Beside what Benoit listed we have following native drivers:
- virtio
- vhost-user
- tap
- vmxnet
- memif
- af_packet
- af_xdp
Ben was right
Hi Magesh,
That code is pretty old, could you try with master latest? There have been at
least a couple of fixes to the bt since then (see [1] and [2]). And, as per
assert that follows, the bts should be valid and therefore the bug is most
probably in a different place.
On a separate note, th
We got a crash at the following lines of code in src/vnet/tcp/tcp_bt.c .
The code base is
vpp_1908 branch.
354 /* Find original tx sample and cache flags in case the sample
355 * is freed or the pool moves */
356 bts = bt_lookup_seq (bt, start);
357 bts_flags = bts->flags;
when bt_look
Folks,
We'd like to deprecate the GBP plugin and the support for VXLAN-GBP:
https://gerrit.fd.io/r/c/vpp/+/28817
Does anyone has an objection?
Thanks
ben
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#20528): https://lists.fd.io/g/vpp-dev/message
I’d say that that’s quite pointless. If your vector size is small it means VPP
has a lot of headroom in your environment. Maybe take a look at config options
- there is one which catches my eye ‘unix {poll-sleep-usec}’ which might help
you get what you want at the cost of probably not being able
Hi Ben,
Good to see you are looking at MAP. Always interested in deployment details if
you want to share. Feel free to ping me offline.
The code is written this way, considering the IPv6 tunnel endpoint as a
functional address. I.e. not being used to address an interface, but to address
the M
