====================
#fdio-vpp: fd.io vpp
====================


Meeting started by DaveBarach at 16:00:56 UTC.  The full logs are
available at
http://ircbot.wl.linuxfoundation.org/meetings/fdio-vpp/2018/fdio-vpp.2018-01-09-16.00.log.html
.



Meeting summary
---------------

* agenda additions?  (DaveBarach, 16:03:18)

* CSIT (Peter Mikus reporting)  (DaveBarach, 16:03:58)
  * expect measurement data tomorrow  (DaveBarach, 16:04:10)

* NSH/SFC  (DaveBarach, 16:05:50)
  * PTL election in progress, hope to hear soon  (DaveBarach, 16:05:58)

* ONE/LISP (Florin)  (DaveBarach, 16:06:11)
  * maintenance mode for the moment  (DaveBarach, 16:06:18)
  * ACTION: dbarach remove ONE/LISP from weekly agenda  (DaveBarach,
    16:06:42)

* host stack (Florin)  (DaveBarach, 16:06:59)
  * binary API / vlibmemory work on socket transport refactoring
    (DaveBarach, 16:07:26)
  * can now connect via socket and bootstrap a pairwise shared-memory
    connection to vpp  (DaveBarach, 16:07:46)
  * big patch pending to refactor vlibmemory in simple ways
    (DaveBarach, 16:08:45)
  * better client/server isolation  (DaveBarach, 16:08:58)
  * separate vlib API code from (memory,socket) transport  (DaveBarach,
    16:09:39)
  * LINK: https://gerrit.fd.io/r/#/c/9976/  (florinc, 16:10:07)
  * Dave Wallace on LD_PRELOAD work  (DaveBarach, 16:10:59)
  * elog events added by Keith  (DaveBarach, 16:11:20)
  * Dave Wallace refactoring LD_PRELOAD: make the library stateless
    (DaveBarach, 16:11:45)
  * 4KLOC removed  (DaveBarach, 16:12:13)
  * "no more hash tables!"  (DaveBarach, 16:12:34)
  * mixed vpp/kernel select, epoll: split bitmaps, etc.  (DaveBarach,
    16:13:59)
  * better tracing so people will find out that "X is unimplemented" in
    an obvious way  (DaveBarach, 16:15:17)
  * Keith will be updating wiki with elog info  (DaveBarach, 16:15:49)
  * demo at Kubecon - thanks to Ed, Jan - native integration with Envoy
    in progress  (DaveBarach, 16:16:20)
  * working on native VCL integration into apps  (DaveBarach, 16:16:36)
  * discussion around Spectre and Meltdown perf implications
    (DaveBarach, 16:19:40)

* arm / aarch64 (Tina T reporting)  (DaveBarach, 16:20:43)
  * CSIT added support for aarch64  (DaveBarach, 16:21:26)
  * next two weeks, continue w/ testing, perf benchmarking  (DaveBarach,
    16:21:55)

* 18.01 release (Dave Wallace)  (DaveBarach, 16:22:43)
  * RC1 branch cut last week  (DaveBarach, 16:22:51)
  * OpenSUSE artifacts not being published yet  (DaveBarach, 16:23:36)
  * request to Marco: don't build old branches since they won't work
    (DaveBarach, 16:24:17)

* Coverity SA currently broken  (DaveBarach, 16:26:45)
  * seems like disk breakage knock-on effects  (DaveBarach, 16:27:26)
  * not in bad shape, we'll see if we can "get to zero" before final
    release  (DaveBarach, 16:27:59)

* fd.io events (Ed, Ray)  (DaveBarach, 16:29:44)
  * Ray mentions 4 vpp talks at FOSDEM (Feb, Brussels)  (DaveBarach,
    16:30:07)
  * ONS 2018 CFP opaque. Ed working to construct an fd.io minisummit
    (DaveBarach, 16:31:19)
  * Kubecon Europe (May) CFP closes Jan 12  (DaveBarach, 16:32:07)
  * ACTION: Ed to add Kubecon Europe details to wiki  (DaveBarach,
    16:32:21)

* Damjan on native driver for Macchiato  (DaveBarach, 16:34:39)
  * use Marvell user-mode SDK. Driver working, need to resolve
    shared-memory base VA issue  (DaveBarach, 16:35:26)
  * vpp overusing prefetch slots. vpp 3.5 mpps vs. sample app 9 mpps
    (DaveBarach, 16:36:42)

* Damjan on tap v2 driver  (DaveBarach, 16:45:34)
  * ring size tuning may be required. Default 256 is intended to keep
    older kernels happy  (DaveBarach, 16:46:02)



Meeting ended at 16:46:52 UTC.



People present (lines said)
---------------------------

* DaveBarach (47)
* collabot (3)
* florinc (1)



Generated by `MeetBot`_ 0.1.4