[dpdk-dev] 19.11.9 patches review and test

[Christian Ehrhardt]

On Thu, Jul 1, 2021 at 5:50 PM Ali Alnubani <alialnu at nvidia.com> wrote:
>
> Hi,
>
> > -----Original Message-----
> > From: Christian Ehrhardt <christian.ehrhardt at canonical.com>
> > Sent: Thursday, June 17, 2021 9:38 AM
> > To: stable at dpdk.org
> > Cc: dev at dpdk.org; Abhishek Marathe <Abhishek.Marathe at microsoft.com>;
> > Akhil Goyal <akhil.goyal at nxp.com>; Ali Alnubani <alialnu at nvidia.com>;
> > benjamin.walker at intel.com; David Christensen <drc at linux.vnet.ibm.com>;
> > hariprasad.govindharajan at intel.com; Hemant Agrawal
> > <hemant.agrawal at nxp.com>; Ian Stokes <ian.stokes at intel.com>; Jerin
> > Jacob <jerinj at marvell.com>; John McNamara <john.mcnamara at intel.com>;
> > Ju-Hyoung Lee <juhlee at microsoft.com>; Kevin Traynor
> > <ktraynor at redhat.com>; Luca Boccassi <bluca at debian.org>; Pei Zhang
> > <pezhang at redhat.com>; pingx.yu at intel.com; qian.q.xu at intel.com; Raslan
> > Darawsheh <rasland at nvidia.com>; NBU-Contact-Thomas Monjalon
> > <thomas at monjalon.net>; yuan.peng at intel.com; zhaoyan.chen at intel.com
> > Subject: 19.11.9 patches review and test
> >
> > Hi all,
> >
> > Here is a list of patches targeted for stable release 19.11.9.
> >
> > The planned date for the final release is 2nd July.
> >
> > Please help with testing and validation of your use cases and report any
> > issues/results with reply-all to this mail. For the final release the fixes and
> > reported validations will be added to the release notes.
> >
> > A renewed release candidate tarball can be found at:
> >
> >     https://dpdk.org/browse/dpdk-stable/tag/?id=v19.11.9-rc3
> >
> > These patches are located at branch 19.11 of dpdk-stable repo:
> >     https://dpdk.org/browse/dpdk-stable/
> >
> > Thanks.
> >
> > Christian Ehrhardt <christian.ehrhardt at canonical.com>
> >
>
> The following covers the functional tests that we ran on Mellanox hardware for this release:
> - Basic functionality:
>   Send and receive multiple types of traffic.
> - testpmd xstats counter test.
> - testpmd timestamp test.
> - Changing/checking link status through testpmd.
> - RTE flow tests:
>   Items: eth / vlan / ipv4 / ipv6 / tcp / udp / icmp / gre / nvgre / vxlan / ip in ip / mplsoudp / mplsogre
>   Actions: drop / queue / rss / mark / flag / jump / count / raw_encap / raw_decap / vxlan_encap / vxlan_decap / NAT / dec_ttl
> - Some RSS tests.
> - VLAN filtering, stripping and insertion tests.
> - Checksum and TSO tests.
> - ptype tests.
> - link_status_interrupt example application tests.
> - l3fwd-power example application tests.
> - Multi-process example applications tests.
>
> Functional tests ran on:
> - NIC: ConnectX-4 Lx / OS: Ubuntu 20.04 LTS / Driver: MLNX_OFED_LINUX-5.3-1.0.0.1 / Firmware: 14.30.1004
> - NIC: ConnectX-5 / OS: Ubuntu 20.04 LTS / Driver: MLNX_OFED_LINUX-5.3-1.0.0.1 / Firmware: 16.30.1004
>
> We discovered 2 new issues due to environment changes:
> - can't create some rules with count action, fixed by https://inbox.dpdk.org/stable/20210621145105.963179-1-lmargalit@nvidia.com/T/#u.

As discussed in the release meeting yesterday, that is a regression
and needs to be added.
While I feel slightly tired of new respins this is what we do the
validation for.
So thanks you (all!) for the testing!

I have now applied the referred fix to 19.11.9

FYI: I see this came up ~5 days into the -rc3 testing, but I missed it
so I checked why.
It was breaking the usual subject syntax (double colon) and not
flagged for 19.11.9 (as requested in backport mails and the common
style) in subject or body.

> - rte_flow hit counter doesn't increment, still discussing this internally.

For this issue please tell me until Monday what you expect.
Does it look close to being resolved (then I'll hold -rc4 back a bit
longer) or should I go on planning to release 19.11.9 without it (then
I'd tag -rc4 on Monday).

Generally on -rc4 I think it would be sufficient if you would re-run
your testing as your tests cover the only two things we have touched
in -rc4 (clang 12 and the flow issue).
I'm not "against" more tests, just saying that those should be
sufficient in case no others come back then.
Because other than the issues reported in here all other tests already
came back fully green for -rc3.

FYI the usual non-official WIP repo on
https://github.com/cpaelzer/dpdk-stable-queue/commits/19.11 holds what
19.11.9-rc4 would become if you want to pre-test anything.

> Compilation tests with multiple configurations in the following OS/driver combinations are also passing:
> - Ubuntu 20.04.2 with MLNX_OFED_LINUX-5.3-1.0.0.1.
> - Ubuntu 20.04.2 with rdma-core master (a66e2a5).
> - Ubuntu 20.04.2 with rdma-core v28.0.
> - Ubuntu 18.04.5 with rdma-core v17.1.
> - Ubuntu 18.04.5 with rdma-core master (a66e2a5) (i386).
> - Ubuntu 16.04.7 with rdma-core v22.7.
> - Fedora 34 with rdma-core v35.0 (only with gcc).
> - Fedora 35 (Rawhide) with rdma-core v35.0 (only with gcc).
> - CentOS 7 7.9.2009 with rdma-core master (a66e2a5).
> - CentOS 7 7.9.2009 with MLNX_OFED_LINUX-5.3-1.0.0.1.
> - CentOS 8 8.3.2011 with rdma-core master (7f2d460).
> - OpenSUSE Leap 15.3 with rdma-core v31.0.
>
> Note that clang builds are failing in Fedora 34 and 35 due to the following issues:
> - https://bugs.dpdk.org/show_bug.cgi?id=745

This is still totally unknown and not specific to 19.11 series.
I guess this clang issue has to wait for the next stable release then.

> - https://bugs.dpdk.org/show_bug.cgi?id=733

We've identified the patch that needs to be backported for this one.
I have done so and built across all platforms - LGTM.
If your tests could also confirm this once we have -rc4 tagged that
would be great.

> We don't see any other critical issues blocking this release.
>
> Thanks,
> Ali



--
Christian Ehrhardt
Staff Engineer, Ubuntu Server
Canonical Ltd