16.11.9 (LTS) patches review and test

Hi all,

Here is a list of patches targeted for LTS release 16.11.9. Please help review and test. The planned date for the final release is December the 10th.
Before that, please shout if anyone has objections with these patches being applied.

Also for the companies committed to running regression tests, please run the tests and report any issue before the release date by replying to this email.

This will be the last 16.11 point release. As with all good things, it must come to an end!

A release candidate tarball can be found at:

    https://dpdk.org/browse/dpdk-stable/tag/?id=v16.11.9-rc1

These patches are located at branch 16.11 of dpdk-stable repo:
    https://dpdk.org/browse/dpdk-stable/

Thanks.

Luca Boccassi

---
Ajit Khaparde (3):
      net/bnxt: set MAC filtering as outer for non tunnel frames net/bnxt: set a VNIC as default only once net/bnxt: remove excess log messages

Alejandro Lucero (3):
      net/nfp: fix live MAC changes not supported
      net/nfp: fix mbuf flags with checksum good
      net/nfp: fix RSS

Anatoly Burakov (3):
      vfio: fix build on old kernel
      vfio: fix build
      usertools: check for lspci dependency

Beilei Xing (2):
      net/i40e: update Tx offload mask net/i40e: fix X710 Rx after reading some registers

Brian Archbold (1):
      app/testpmd: fix duplicate exit

Chas Williams (1):
      net/bonding: fix Rx slave fairness

Dan Gora (1):
      kni: fix build on CentOS 7.4

Dharmik Thakkar (1):
      test/hash: fix build

Ferruh Yigit (8):
      app/testpmd: fix displaying RSS hash functions
      kni: fix build on Linux 4.19
      fix global variable issues
      eal: fix build with gcc 9.0
      igb_uio: remove device reset in open
      igb_uio: remove device reset in release test/reorder: fix out of bound access bus/pci: fix allocation of device path

Fiona Trahe (1):
      test/crypto: fix number of queue pairs

Haiyue Wang (1):
      net/i40e: enable loopback function for X722 MAC

Herakliusz Lipiec (1):
      examples/ipv4_multicast: fix leak of cloned packets

Honnappa Nagarahalli (2):
      build: enable ARM NEON flag when __aarch64__ defined
      hash: fix key store element alignment

Huaibin Wang (1):
      net/i40e: keep promiscuous on if allmulticast is enabled

Hyong Youb Kim (3):
      net/enic: add devarg to specify ingress VLAN rewrite mode net/enic: do not use non-standard integer types net/enic: set Rx VLAN offload flag for non-stripped packets

Igor Romanov (2):
      net/bonding: use evenly distributed default RSS RETA net/bonding: do not ignore RSS key on device config

Ilya Maximets (1):
      vhost-user: drop connection on message handling failures

Jeff Guo (1):
      igb_uio: fix unexpected removal for hot-unplug

Jerin Jacob (3):
      app/testpmd: fix csum parse-tunnel command invocation
      eal: introduce rte version of fls
      eal: fix build

Jiayu Hu (1):
      vhost: fix corner case for enqueue operation

Joyce Kong (1):
      config: enable more than 128 cores for arm64

Keith Wiles (1):
      net/bonding: reduce slave starvation on Rx poll

Konstantin Ananyev (3):
      acl: forbid rule with priority zero
      ip_frag: check fragment length of incoming packet
      ip_frag: use key length for key comparison

Li Han (1):
      ip_frag: fix overflow in key comparison

Luca Boccassi (5):
      net/e1000: do not error out if Rx drop enable is set net/virtio: register/unregister intr handler on start/stop eal/linux: handle UIO read failure in interrupt handler net/enic: add dependency on librte_kvargs
      vhost-user: fix false negative in handling user messages

Marko Kovacevic (1):
      doc: fix formatting in IP reassembly app guide

Matthew Smith (1):
      net/ixgbe: wait longer for link after fiber MAC setup

Maxime Coquelin (1):
      vhost: fix payload size of reply

Naga Suresh Somarowthu (1):
      test/kni: check module dependency

Nirmoy Das (2):
      kni: fix build on Suse 12 SP3
      kni: fix SLE version detection

Paul M Stillwell Jr (1):
      ethdev: fix doxygen comment to be with structure

Phil Yang (2):
      app/testpmd: optimize mbuf pool allocation
      kni: fix kernel FIFO synchronization

Qi Zhang (5):
      net/i40e: fix send admin queue command before init net/i40e/base: fix partition id calculation for X722 net/i40e/base: properly clean resources net/i40e/base: gracefully clean the resources net/i40e/base: correct global reset timeout calculation

Radu Nicolau (2):
      net/bonding: stop and deactivate slaves on stop net/bonding: fix crash when stopping mode 4 port

Rafal Kozik (1):
      doc: add VFIO in ENA guide

Rami Rosen (2):
      doc: fix wrong usage of bind command
      examples/vhost: remove unnecessary constant

Rasesh Mody (5):
      net/bnx2x: fix logging to include device name net/bnx2x: fix to disable further interrupts net/bnx2x: fix call to link handling periodic function net/bnx2x: fix to add PHY lock
      net/bnx2x: fix VF link state update

Raslan Darawsheh (1):
      app/testpmd: fix L4 length for UDP checksum

Shahaf Shuler (1):
      net: fix build with pedantic

Shijith Thotton (1):
      igb_uio: issue FLR during open and release of device file

Somnath Kotur (1):
      net/bnxt: fix uninitialized pointer access in Tx

Stewart Allen (1):
      net/ena: fix passing RSS hash to mbuf

Subrahmanyam Nilla (1):
      net/thunderx: fix Tx desc corruption in scatter-gather mode

Thomas Monjalon (7):
      kvargs: fix processing a null list
      kni: fix build on Linux < 3.14
      pci: fix parsing of address without function number net/mlx5: fix build on PPC64
      eal: fix build with -O1
      kni: fix possible uninitialized variable fix dpdk.org URLs

Tiwei Bie (5):
      net/virtio: add missing supported features net/vhost: fix parameters string net/virtio-user: do not reset owner when driver resets net/virtio-user: fix typo in error message
      vhost: remove unneeded null pointer check

Vivek Sharma (1):
      eal: use correct data type for bitmap slab operations

Wenzhuo Lu (1):
      ethdev: fix invalid configuration after failure

Xiaolong Ye (1):
      net/ixgbe: fix maximum wait time in comment

Xiaoxin Peng (1):
      net/bnxt: reduce polling interval for valid bit

Yanglong Wu (2):
      net/ixgbevf: fix link state
      net/ixgbe: fix TDH register write

Yipeng Wang (2):
      test/hash: fix bucket size in perf test
      hash: remove unnecessary pause

Yong Wang (2):
      doc: fix typo in testpmd guide
      doc: fix NUMA library name in Linux guide

Yongseok Koh (1):
      mk: disable gcc AVX512F support

Yunjian Wang (2):
      net/bonding: support matching QinQ ethertype net/e1000/base: fix uninitialized variable

Zhihong Wang (1):
      config: make AVX and AVX512 configurable

Zhirun Yan (2):
      net/ixgbe: update Tx offload mask
      net/igb: update Tx offload mask

Ziye Yang (1):
      eal/linux: fix memory leak of logid

Hi,

Unfortunately due to the conference and other commitments it was not possible to have the test results of 16.11.9-rc1 yet, so the final release will have to be postponed.
Tentative date is a week from now, on December the 17th.

Apologies if this delay causes any issue.

Kind regards,
Luca Boccassi

Hi,

For the same reasons, unfortunately we have to delay the 16.11.9 release further. The tentative release date is now Friday the 21st.

Apologies if this delay causes any issue.

Kind regards,
Luca Boccassi

Hi,

I have now tagged 16.11.9-rc2 to fix a regression that was introduced back in 16.11.7:

Yongseok Koh (1):
      ethdev: fix queue start and stop

The tarball can be found at:

https://dpdk.org/browse/dpdk-stable/tag/?id=v16.11.9-rc2

Kind regards,
Luca Boccassi

Hi,

Unfortunately test results are yet not ready, so I have to postpone again. I will send a further update when a tentative date becomes clearer.

Kind regards,
Luca Boccassi

Hi,

Due to conferences, holidays and other higher priority deliverables the regression tests on 16.11.9 has still to be ran. The newest estimate is by the end of February - with the alternative being releasing now without the testing.
During the release meeting we agreed that, unless there are objections and urgent requirements to have 16.11.9 release in January, we will wait for the regressions and postpone to end of February. Please let us know if there are any objections.

Again, apologies for this delay.

Kind regards,
Luca Boccassi

Hi,

Testing results came back, there are known issues with bonding and GRE checksum offloads (i40e) that were also affecting previous releases, no new regressions have been found.

Therefore, I'll proceed to tag 16.11.9 later today.

Kind regards,
Luca Boccassi

Hi,

Regression tests from AT&T came back green, no new issue found.

Kind regards,
Luca Boccassi