[dpdk-dev] 18.11.9 (LTS) patches review and test

Kevin Traynor ktraynor at redhat.com
Tue Jun 30 19:03:28 CEST 2020


On 30/06/2020 17:32, Stokes, Ian wrote:
> 
> 
>> -----Original Message-----
>> From: Kevin Traynor <ktraynor at redhat.com>
>> Sent: Tuesday, June 30, 2020 5:21 PM
>> To: Govindharajan, Hariprasad <hariprasad.govindharajan at intel.com>;
>> stable at dpdk.org; Stokes, Ian <ian.stokes at intel.com>; Loftus, Ciara
>> <ciara.loftus at intel.com>
>> Cc: dev at dpdk.org; Abhishek Marathe <Abhishek.Marathe at microsoft.com>;
>> Akhil Goyal <akhil.goyal at nxp.com>; Ali Alnubani <alialnu at mellanox.com>;
>> Walker, Benjamin <benjamin.walker at intel.com>; David Christensen
>> <drc at linux.vnet.ibm.com>; Hemant Agrawal <hemant.agrawal at nxp.com>;
>> Jerin Jacob <jerinj at marvell.com>; Mcnamara, John
>> <john.mcnamara at intel.com>; Ju-Hyoung Lee <juhlee at microsoft.com>; Luca
>> Boccassi <bluca at debian.org>; Pei Zhang <pezhang at redhat.com>; Yu, PingX
>> <pingx.yu at intel.com>; Xu, Qian Q <qian.q.xu at intel.com>; Raslan Darawsheh
>> <rasland at mellanox.com>; Thomas Monjalon <thomas at monjalon.net>; Peng,
>> Yuan <yuan.peng at intel.com>; Chen, Zhaoyan <zhaoyan.chen at intel.com>
>> Subject: Re: [dpdk-dev] 18.11.9 (LTS) patches review and test
>>
>> On 30/06/2020 16:38, Govindharajan, Hariprasad wrote:
>>>
>>>
>>>> -----Original Message-----
>>>> From: dev <dev-bounces at dpdk.org> On Behalf Of Kevin Traynor
>>>> Sent: Friday, June 26, 2020 1:53 PM
>>>> 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 mellanox.com>; Walker, Benjamin <benjamin.walker at intel.com>;
>>>> David Christensen <drc at linux.vnet.ibm.com>; Hemant Agrawal
>>>> <hemant.agrawal at nxp.com>; Stokes, Ian <ian.stokes at intel.com>; Jerin
>>>> Jacob <jerinj at marvell.com>; Mcnamara, John <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>; Yu, PingX <pingx.yu at intel.com>; Xu, Qian Q
>>>> <qian.q.xu at intel.com>; Raslan Darawsheh <rasland at mellanox.com>;
>>>> Thomas Monjalon <thomas at monjalon.net>; Peng, Yuan
>>>> <yuan.peng at intel.com>; Chen, Zhaoyan <zhaoyan.chen at intel.com>
>>>> Subject: [dpdk-dev] 18.11.9 (LTS) patches review and test
>>>>
>>>> Hi all,
>>>>
>>>> Here is a list of patches targeted for LTS release 18.11.9.
>>>>
>>>> The planned date for the final release is 3rd 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 release candidate tarball can be found at:
>>>>
>>>>     https://dpdk.org/browse/dpdk-stable/tag/?id=v18.11.9-rc2
>>>>
>>>> These patches are located at branch 18.11 of dpdk-stable repo:
>>>>     https://dpdk.org/browse/dpdk-stable/
>>>>
>>>> rc2:
>>>> Creating this to include important VF fix.
>>>> http://inbox.dpdk.org/dev/20200625035046.19820-1-
>>>> haiyue.wang at intel.com/
>>>>
>>>> New patches in rc2:
>>>> 145bb0e780 bus/pci: fix VF memory access
>>>> 420a437247 net/tap: fix mbuf double free when writev fails
>>>>
>>>> Thanks.
>>>>
>>>> Kevin.
>>> [Govindharajan, Hariprasad] Hi Kevin,
>>>
>>
>> Hi G Hariprasad,
>>
>> Thanks for testing.
>>
>>> The following performance and functional tests carried out with ixgbe,
>>> i40e and vhost devices
>>>
>>> DPDK 18.11.9 RC2 with OvS 2.11 and OvS 2.12
>>>
>>> P2p throughput
>>> P2p multi queue
>>> PVP rss
>>> vHost reconnect
>>> Flow control
>>>
>>> Zero  copy feature could not be tested.
>>> From ovs, zero copy is used only vhost user client. So this particular patch
>> 0309429b198800f4d17d578aad650672476ba054 is breaking the zero copy
>> support in ovs.
>>> so for ovs to move to support 18.11.9, it should probably remove support for
>> zero-copy in the ovs code base?
>>>
>>
>> The code says that zero copy is incompatible with vhost client mode.
>> It seems that the check for client/server was incorrect, and now it is fixed and
>> correctly detects this incompatible combination and returns an error to OVS.
>>
>> I don't think we can call that breaking. It was previously incompatible also, it just
>> didn't return an error as there was a bug.
>>
>> I don't see an issue with having this patch and the new behaviour.
>>
>> Ian/Ciara, let me know if I'm missing something.
> 
> I think OVS only supports zero copy for vhostuserclient devices. In this case it looks like that’s no longer supported. I guess the real question becomes should the support in OVS be removed now? It is not broken as you said, just no longer supported in the OVS use case.
> 

ah ok, it's not the commit above then, it's this one is the issue

commit 81e025d7ed6a802845909df6fb90505508dc0fbf
Author: Xuan Ding <xuan.ding at intel.com>
Date:   Wed Apr 29 02:59:46 2020 +0000

    vhost: prevent zero-copy with incompatible client mode

    [ upstream commit 715070ea10e6da1169deef2a3ea77ae934b4c333 ]

    In server mode, virtio-user inits under the assumption that vhost-user
    supports a list of features. However, this could be problematic when
    in_order feature is negotiated but not supported by vhost-user when
    enables dequeue_zero_copy later.

    Add handling when vhost-user enables dequeue_zero_copy as client.
--

It's unfortunate that it is both fixing a bug and removing functionality
at the same time :/

> I don’t think it needs to block the release of this, but it's a follow up for the OVS. I believe you'd see the same behavior for OVS master and 2.13.
> 

Yes, these patches are already in 19.11.3.

> BR
> Ian
> 
>>
>> thanks,
>> Kevin.
>>
>>> Regards
>>> G Hariprasad
>>>>
>>>> ---
>>>> Adam Dybkowski (2):
>>>>       app/crypto-perf: fix display of sample test vector
>>>>       cryptodev: fix SHA-1 digest enum comment
>>>>
>>>> Ajit Khaparde (3):
>>>>       net/bnxt: fix FW version query
>>>>       net/bnxt: fix error log for command timeout
>>>>       net/bnxt: fix using RSS config struct
>>>>
>>>> Alex Kiselev (1):
>>>>       lpm6: fix size of tbl8 group
>>>>
>>>> Alexander Kozyrev (1):
>>>>       net/mlx5: fix Tx queue release debug log timing
>>>>
>>>> Alvin Zhang (1):
>>>>       net/e1000: fix port hotplug for multi-process
>>>>
>>>> Amit Gupta (1):
>>>>       net/octeontx: fix meson build for disabled drivers
>>>>
>>>> Andrew Rybchenko (1):
>>>>       net/sfc: fix reported promiscuous/multicast mode
>>>>
>>>> Apeksha Gupta (1):
>>>>       bus/fslmc: fix dereferencing null pointer
>>>>
>>>> Arek Kusztal (1):
>>>>       crypto/qat: fix cipher descriptor for ZUC and SNOW
>>>>
>>>> Bing Zhao (1):
>>>>       mem: fix overflow on allocation
>>>>
>>>> Bruce Richardson (1):
>>>>       drivers: fix log type variables for -fno-common
>>>>
>>>> Cheng Peng (1):
>>>>       net/iavf: fix stats query error code
>>>>
>>>> Darek Stojaczyk (1):
>>>>       pci: accept 32-bit domain numbers
>>>>
>>>> David Christensen (2):
>>>>       pci: fix build on ppc
>>>>       eal/ppc: fix build with gcc 9.3
>>>>
>>>> David Marchand (4):
>>>>       mem: mark pages as not accessed when reserving VA
>>>>       eal: fix typo in endian conversion macros
>>>>       remove references to private PCI probe function
>>>>       doc: prefer https when pointing to dpdk.org
>>>>
>>>> Dekel Peled (2):
>>>>       net/mlx5: fix mask used for IPv6 item validation
>>>>       net/mlx5: fix CVLAN tag set in IP item translation
>>>>
>>>> Eugeny Parshutin (1):
>>>>       ethdev: fix build when vtune profiling is on
>>>>
>>>> Fan Zhang (1):
>>>>       vhost/crypto: add missing user protocol flag
>>>>
>>>> Ferruh Yigit (3):
>>>>       net/null: fix secondary burst function selection
>>>>       net/null: remove redundant check
>>>>       examples/kni: fix crash during MTU set
>>>>
>>>> Gaetan Rivet (4):
>>>>       doc: fix number of failsafe sub-devices
>>>>       net/ring: fix device pointer on allocation
>>>>       pci: reject negative values in PCI id
>>>>       kvargs: fix strcmp helper documentation
>>>>
>>>> Gavin Hu (1):
>>>>       net/i40e: relax barrier in Tx for NEON
>>>>
>>>> Guinan Sun (2):
>>>>       net/ixgbe: fix statistics in flow control mode
>>>>       net/ixgbe: check driver type in MACsec API
>>>>
>>>> Haifeng Lin (1):
>>>>       eal/arm64: fix precise TSC
>>>>
>>>> Haiyue Wang (1):
>>>>       bus/pci: fix VF memory access
>>>>
>>>> Harry van Haaren (3):
>>>>       service: fix crash on exit
>>>>       examples/eventdev: fix crash on exit
>>>>       test/flow_classify: enable multi-sockets system
>>>>
>>>> Hemant Agrawal (3):
>>>>       drivers: add crypto as dependency for event drivers
>>>>       bus/fslmc: fix size of qman fq descriptor
>>>>       mempool/dpaa2: install missing header with meson
>>>>
>>>> Honnappa Nagarahalli (2):
>>>>       service: fix race condition for MT unsafe service
>>>>       service: fix identification of service running on other lcore
>>>>
>>>> Igor Chauskin (2):
>>>>       net/ena/base: make allocation macros thread-safe
>>>>       net/ena/base: prevent allocation of zero sized memory
>>>>
>>>> Igor Romanov (7):
>>>>       net/sfc: fix initialization error path
>>>>       net/sfc: fix Rx queue start failure path
>>>>       net/sfc/base: reduce filter priorities to implemented only
>>>>       net/sfc/base: reject automatic filter creation by users
>>>>       net/sfc/base: refactor filter lookup loop in EF10
>>>>       net/sfc/base: handle manual and auto filter clashes in EF10
>>>>       net/sfc/base: fix manual filter delete in EF10
>>>>
>>>> Itsuro Oda (2):
>>>>       net/vhost: fix potential memory leak on close
>>>>       vhost: make IOTLB cache name unique among processes
>>>>
>>>> Ivan Dyukov (3):
>>>>       net/virtio-user: fix devargs parsing
>>>>       app: remove extra new line after link duplex
>>>>       examples: remove extra new line after link duplex
>>>>
>>>> Jasvinder Singh (3):
>>>>       net/softnic: fix memory leak for thread
>>>>       net/softnic: fix resource leak for pipeline
>>>>       examples/ip_pipeline: remove check of null response
>>>>
>>>> Jeff Guo (2):
>>>>       net/i40e: fix setting L2TAG
>>>>       net/iavf: fix setting L2TAG
>>>>
>>>> Jim Harris (1):
>>>>       contigmem: cleanup properly when load fails
>>>>
>>>> Junyu Jiang (2):
>>>>       examples/vmdq: fix output of pools/queues
>>>>       examples/vmdq: fix RSS configuration
>>>>
>>>> Kalesh AP (2):
>>>>       net/bnxt: fix VLAN add when port is stopped
>>>>       app/testpmd: fix memory failure handling for i40e DDP
>>>>
>>>> Kevin Traynor (14):
>>>>       kni: fix ethtool build with kernel 5.6
>>>>       net/avp: fix gcc 10 maybe-uninitialized warning
>>>>       eal/x86: ignore gcc 10 stringop-overflow warnings
>>>>       crypto/kasumi: fix extern declaration
>>>>       build: disable gcc 10 zero-length-bounds warning
>>>>       kni: fix ethtool dev_open build error
>>>>       kni: fix ethtool pointer type build error
>>>>       drivers/crypto: disable gcc 10 fno-common errors
>>>>       kni: fix gcc 10 ethtool build error
>>>>       kni: fix ethtool maybe-uninitialized warnings
>>>>       crypto/octeontx: fix gcc 10 -fno-common build errors
>>>>       app/testpmd: disable gcc 10 -fno-common build errors
>>>>       examples/vm_power: fix build because of missing include
>>>>       version: 18.11.9-rc1
>>>>
>>>> Konstantin Ananyev (1):
>>>>       security: fix crash at accessing non-implemented ops
>>>>
>>>> Li Feng (1):
>>>>       mem: mark pages as not accessed when freeing memory
>>>>
>>>> Louise Kilheeney (1):
>>>>       examples/l2fwd-keepalive: fix mbuf pool size
>>>>
>>>> Luca Boccassi (2):
>>>>       fix various typos found by Lintian
>>>>       usertools: check for pci.ids in /usr/share/misc
>>>>
>>>> Lukasz Wojciechowski (5):
>>>>       drivers/crypto: fix log type variables for -fno-common
>>>>       security: fix verification of parameters
>>>>       security: fix return types in documentation
>>>>       security: fix session counter
>>>>       test: remove redundant macro
>>>>
>>>> Mattias Rönnblom (3):
>>>>       event/dsw: remove redundant control ring poll
>>>>       event/dsw: remove unnecessary read barrier
>>>>       event/dsw: avoid reusing previously recorded events
>>>>
>>>> Michael Baum (1):
>>>>       net/mlx4: fix drop queue error handling
>>>>
>>>> Michael Haeuptle (1):
>>>>       vfio: fix race condition with sysfs
>>>>
>>>> Michal Krawczyk (5):
>>>>       net/ena/base: fix testing for supported hash function
>>>>       net/ena/base: fix documentation of functions
>>>>       net/ena/base: fix indentation in CQ polling
>>>>       net/ena/base: fix indentation of multiple defines
>>>>       net/ena: set IO ring size to valid value
>>>>
>>>> Muhammad Bilal (2):
>>>>       fix same typo in multiple places
>>>>       doc: fix typo in contributors guide
>>>>
>>>> Nicolas Chautru (1):
>>>>       bbdev: fix doxygen comments
>>>>
>>>> Nithin Dabilpuram (1):
>>>>       devtools: fix symbol map change check
>>>>
>>>> Olivier Matz (3):
>>>>       test/kvargs: fix to consider empty elements as valid
>>>>       test/kvargs: fix invalid cases check
>>>>       kvargs: fix invalid token parsing on FreeBSD
>>>>
>>>> Pablo de Lara (3):
>>>>       test/crypto: fix flag check
>>>>       crypto/openssl: fix out-of-place encryption
>>>>       doc: add NASM installation steps
>>>>
>>>> Pavan Nikhilesh (2):
>>>>       eventdev: fix probe and remove for secondary process
>>>>       app/eventdev: check Tx adapter service ID
>>>>
>>>> Phil Yang (1):
>>>>       service: remove rte prefix from static functions
>>>>
>>>> Qiming Yang (3):
>>>>       net/i40e: fix X722 performance
>>>>       doc: fix multicast filter feature announcement
>>>>       net/i40e: fix queue related exception handling
>>>>
>>>> Rasesh Mody (3):
>>>>       net/qede: fix link state configuration
>>>>       net/qede: fix port reconfiguration
>>>>       examples/kni: fix MTU change to setup Tx queue
>>>>
>>>> Raslan Darawsheh (3):
>>>>       net/mlx5: fix validation of VXLAN/VXLAN-GPE specs
>>>>       app/testpmd: add parsing for QinQ VLAN headers
>>>>       net/mlx5: fix matching for UDP tunnels with Verbs
>>>>
>>>> Rohit Raj (1):
>>>>       net/dpaa2: fix 10G port negotiation
>>>>
>>>> Roland Qi (1):
>>>>       vhost: fix peer close check
>>>>
>>>> Ruifeng Wang (1):
>>>>       test: skip some subtests in no-huge mode
>>>>
>>>> Sarosh Arif (1):
>>>>       doc: fix typo in contributors guide
>>>>
>>>> Somnath Kotur (1):
>>>>       bus/pci: fix devargs on probing again
>>>>
>>>> Stephen Hemminger (21):
>>>>       ethdev: fix spelling
>>>>       net/mvneta: do not use PMD log type
>>>>       net/virtio: do not use PMD log type
>>>>       net/tap: do not use PMD log type
>>>>       net/bnxt: do not use PMD log type
>>>>       net/dpaa: use dynamic log type
>>>>       net/thunderx: use dynamic log type
>>>>       net/netvsc: propagate descriptor limits from VF
>>>>       net/netvsc: handle Rx packets during multi-channel setup
>>>>       net/netvsc: split send buffers from Tx descriptors
>>>>       net/netvsc: fix memory free on device close
>>>>       net/netvsc: remove process event optimization
>>>>       net/netvsc: handle Tx completions based on burst size
>>>>       net/netvsc: avoid possible live lock
>>>>       lpm6: fix comments spelling
>>>>       eal: fix comments spelling
>>>>       net/netvsc: fix comment spelling
>>>>       bus/vmbus: fix comment spelling
>>>>       net/tap: fix crash in flow destroy
>>>>       eal: fix C++17 compilation
>>>>       net/vmxnet3: handle bad host framing
>>>>
>>>> Sunil Kumar Kori (1):
>>>>       eal: fix log message print for regex
>>>>
>>>> Thomas Monjalon (11):
>>>>       doc: fix sphinx compatibility
>>>>       log: fix level picked with globbing on type register
>>>>       doc: fix matrix CSS for recent sphinx
>>>>       net/mlx4: fix build with -fno-common
>>>>       app: fix usage help of options separated by dashes
>>>>       net/mvpp2: fix build with gcc 10
>>>>       examples/vm_power: drop Unix path limit redefinition
>>>>       doc: fix build with doxygen 1.8.18
>>>>       doc: fix API index
>>>>       mlx5: fix build with -fno-common
>>>>       examples/vm_power: fix build with -fno-common
>>>>
>>>> Timothy Redaelli (3):
>>>>       app/pipeline: fix build with gcc 10
>>>>       examples/eventdev: fix build with gcc 10
>>>>       crypto/octeontx: fix build with gcc 10
>>>>
>>>> Ting Xu (1):
>>>>       app/testpmd: fix DCB set
>>>>
>>>> Vijaya Mohan Guvva (1):
>>>>       bus/pci: fix UIO resource access from secondary process
>>>>
>>>> Wei Hu (Xavier) (2):
>>>>       vfio: fix use after free with multiprocess
>>>>       app/testpmd: fix statistics after reset
>>>>
>>>> Wei Zhao (4):
>>>>       net/i40e: fix flow director for ARP packets
>>>>       doc: add i40e limitation for flow director
>>>>       net/i40e: fix flush of flow director filter
>>>>       net/i40e: fix wild pointer
>>>>
>>>> Xiaolong Ye (3):
>>>>       net/virtio: fix outdated comment
>>>>       vhost: remove unused variable
>>>>       doc: fix log level example in Linux guide
>>>>
>>>> Xiaoyu Min (1):
>>>>       net/mlx5: fix RSS enablement
>>>>
>>>> Xiaoyun Li (2):
>>>>       net/ixgbe/base: update copyright
>>>>       net/i40e/base: update copyright
>>>>
>>>> Xuan Ding (2):
>>>>       vhost: prevent zero-copy with incompatible client mode
>>>>       vhost: fix zero-copy server mode
>>>>
>>>> Yunjian Wang (14):
>>>>       kvargs: fix buffer overflow when parsing list
>>>>       net/tap: remove unused assert
>>>>       net/tap: fix mbuf and mem leak during queue release
>>>>       net/tap: fix check for mbuf number of segment
>>>>       net/tap: fix file close on remove
>>>>       net/tap: fix fd leak on creation failure
>>>>       net/tap: fix unexpected link handler
>>>>       net/tap: fix queues fd check before close
>>>>       net/octeontx: fix dangling pointer on init failure
>>>>       crypto/ccp: fix fd leak on probe failure
>>>>       net/failsafe: fix fd leak
>>>>       crypto/caam_jr: fix check of file descriptors
>>>>       crypto/caam_jr: fix IRQ functions return type
>>>>       net/tap: fix mbuf double free when writev fails
>>>>
>>>> Yuri Chipchev (1):
>>>>       event/dsw: fix enqueue burst return value
>>>>
>>>> Zhihong Peng (1):
>>>>       net/ixgbe: fix link status synchronization on BSD
>>>
> 



More information about the dev mailing list