22.11.3 patches review and test

22.11.3 patches review and test

[Xueming Li]

Hi all,

Here is a list of patches targeted for stable release 22.11.3.

The planned date for the final release is 31th August.

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=v22.11.3-rc1

These patches are located at branch 22.11 of dpdk-stable repo:
    https://dpdk.org/browse/dpdk-stable/

Thanks.

Xueming Li <xuemingl@nvidia.com>

---
Aakash Sasidharan (1):
      test/crypto: fix PDCP-SDAP test vectors

Akhil Goyal (1):
      doc: fix auth algos in cryptoperf app

Alexander Kozyrev (4):
      net/mlx5: forbid MPRQ restart
      net/mlx5: fix MPRQ stride size for headroom
      net/mlx5: fix handle validation for meter mark
      app/testpmd: fix meter mark handle update

Ali Alnubani (2):
      doc: fix typos and wording in flow API guide
      doc: improve wording of cuda guide

Alvaro Karsz (1):
      dma/dpaa2: set VFA bit for route-by-port with VF

Anoob Joseph (1):
      cryptodev: clarify error codes for symmetric session

Artemii Morozov (1):
      common/sfc_efx/base: fix Rx queue without RSS hash prefix

Ashwin Sekhar T K (2):
      mempool/cnxk: avoid hang when counting batch allocs
      doc: fix typo in graph guide

Bing Zhao (12):
      app/testpmd: fix flow rule number parsing
      net/mlx5: fix error set in Tx representor tagging
      net/mlx5: fix return value of vport action
      net/mlx5: fix error set in control tables create
      net/mlx5: fix error set for age pool initialization
      net/mlx5: fix error in VLAN actions creation
      common/mlx5: fix obtaining IB device in LAG mode
      net/mlx5: fix profile check of meter mark
      net/mlx5: reduce counter pool name length
      net/mlx5: fix drop action memory leak
      net/mlx5: fix validation for conntrack indirect action
      doc: add flow template API requirements for mlx5

Boleslav Stankevich (2):
      net/virtio: propagate interrupt configuration error values
      net/virtio: fix initialization to return negative errno

Bruce Richardson (4):
      kernel/freebsd: fix function parameter list
      telemetry: fix autotest on Alpine
      eal: avoid calling cleanup twice
      test/bonding: fix include of standard header

Chaoyong He (14):
      net/nfp: fix VLAN push flow action
      net/nfp: fix IPv6 flow item
      net/nfp: fix TOS of IPv6 VXLAN encap flow action
      net/nfp: fix TOS of IPv6 GENEVE encap flow action
      net/nfp: fix TOS of IPv6 NVGRE encap flow action
      net/nfp: fix TP flow action for UDP
      net/nfp: fix IPv6 address for set flow action
      net/nfp: fix unneeded endian conversion
      net/nfp: fix endian conversion for tunnel decap action
      net/nfp: fix VNI of VXLAN encap action
      net/nfp: fix VNI of IPv4 NVGRE encap action
      net/nfp: fix VNI of IPv6 NVGRE encap action
      net/nfp: fix Tx descriptor free logic of NFD3
      net/nfp: fix offloading flows

Chengwen Feng (4):
      net/hns3: fix Rx multiple firmware reset interrupts
      net/hns3: fix mbuf leakage when RxQ started during reset
      net/hns3: fix mbuf leakage when RxQ started after reset
      net/hns3: fix device start return value

Ciara Power (6):
      crypto/qat: fix stack buffer overflow in SGL loop
      crypto/scheduler: fix last element for valid args
      crypto/ipsec_mb: optimize allocation in session
      crypto/qat: fix null algorithm digest placement
      cryptodev: fix device socket ID type
      app/crypto-perf: fix socket ID default value

Dariusz Sosnowski (1):
      net/mlx5: fix drop action attribute validation

David Christensen (1):
      net/tap: set locally administered bit for fixed MAC address

David Marchand (9):
      ci: fix build for Arm cross compilation in GHA
      net/virtio-user: fix leak when initialisation fails
      net/mlx5: enhance error log for tunnel offloading
      examples/l2fwd-cat: fix external build
      examples/fips_validation: fix external build
      test: add graph tests
      mbuf: fix Doxygen comment of distributor metadata
      ci: fix libabigail cache in GHA
      doc: fix link to flow capabilities from bnxt guide

Dengdui Huang (3):
      net/hns3: fix variable type mismatch
      net/hns3: fix inaccurate log
      net/hns3: fix redundant line break in log

Denis Pryazhennikov (3):
      ethdev: update documentation for API to set FEC
      ethdev: check that at least one FEC mode is specified
      ethdev: update documentation for API to get FEC

Devendra Singh Rawat (1):
      net/qede: fix RSS indirection table initialization

Didier Pallard (2):
      crypto/openssl: fix memory leak in auth processing
      crypto/openssl: skip workaround at compilation time

Dmitry Kozlyuk (1):
      member: fix PRNG seed reset in NitroSketch mode

Dongdong Liu (3):
      doc: fix kernel patch link in hns3 guide
      doc: fix syntax in hns3 guide
      doc: fix number of leading spaces in hns3 guide

Elena Agostini (1):
      net: fix return type of IPv4 L4 packet checksum

Erez Ferber (1):
      common/mlx5: adjust fork call with new kernel API

Erik Gabriel Carrillo (1):
      eventdev/timer: fix buffer flush

Feifei Wang (1):
      ring: fix dequeue parameter name

Fengnan Chang (2):
      eal/linux: fix legacy mem init with many segments
      mem: fix memsegs exhausted message

Ferruh Yigit (3):
      kni: fix build with Linux 6.3
      app/testpmd: revert primary process polling all queues fix
      kni: fix build with Linux 6.5

Gregory Etelson (1):
      net/mlx5: fix flow workspace destruction

Heng Jiang (1):
      net/mlx5: fix LRO TCP checksum

Hernan Vargas (2):
      baseband/fpga_5gnr_fec: fix possible division by zero
      baseband/fpga_5gnr_fec: fix starting unconfigured queue

Huisong Li (10):
      ethdev: fix MAC address occupies two entries
      net/hns3: fix never set MAC flow control
      net/hns3: fix RTC time on initialization
      net/hns3: fix RTC time after reset
      net/hns3: uninitialize PTP
      net/hns3: extract PTP to its own header file
      net/hns3: fix IMP reset trigger
      net/hns3: delete duplicate macro definition
      net/hns3: fix non-zero weight for disabled TC
      net/hns3: fix index to look up table in NEON Rx

Ivan Malov (2):
      net/sfc: stop misuse of Rx ingress m-port metadata on EF100
      net/sfc: invalidate dangling MAE flow action FW resource IDs

Jerin Jacob (4):
      examples/ip_pipeline: fix build with GCC 13
      examples/ntb: fix build with GCC 13
      net/hns3: fix build warning
      doc: remove warning with Doxygen 1.9.7

Jiawen Wu (11):
      net/txgbe/base: fix Tx with fiber hotplug
      net/txgbe: fix interrupt enable mask
      net/txgbe: adapt to MNG veto bit setting
      net/txgbe: fix to set autoneg for 1G speed
      net/txgbe: fix extended statistics
      net/ngbe: fix extended statistics
      net/ngbe: adapt to MNG veto bit setting
      net/ngbe: fix link status in no LSC mode
      net/ngbe: remove redundant codes
      net/txgbe: fix blocking system events
      net/ngbe: fix RSS offload capability

Jie Hai (6):
      net/hns3: fix FEC mode for 200G ports
      net/hns3: fix FEC mode check
      net/hns3: fix missing FEC capability
      net/hns3: fix uninitialized variable
      ethdev: fix calloc arguments
      app/testpmd: fix primary process not polling all queues

Jieqiang Wang (1):
      net/i40e: fix comments

Junfeng Guo (4):
      net/ice: fix protocol agnostic offloading with big packets
      net/iavf: fix protocol agnostic offloading with big packets
      raw/ntb: avoid disabling interrupt twice
      doc: update BIOS settings and supported HW for NTB

Kai Ji (1):
      examples/ipsec-secgw: fix socket ID default value

Kaijun Zeng (1):
      net/vmxnet3: fix return code in initializing

Kaiwen Deng (3):
      net/ice: fix DCF RSS initialization
      net/iavf: fix virtchnl command called in interrupt
      net/ice: fix MAC type of E822 and E823

Kaiyu Zhang (1):
      ethdev: fix potential leak in PCI probing helper

Kamalakannan R (1):
      pipeline: fix double free for table stats

Ke Xu (1):
      net/iavf: fix tunnel TSO path selection

Leyi Rong (1):
      eal/x86: improve multiple of 64 bytes memcpy performance

Long Li (7):
      net/mana: use datapath logging
      net/mana: avoid unnecessary assignments in data path
      net/mana: optimize completion queue by batch processing
      net/mana: return probing failure if no device found
      net/mana: fix counter overflow for posted WQE
      net/mana: fix WQE count for ringing RQ doorbell
      net/mana: fix Tx queue statistics

Long Wu (5):
      net/nfp: fix control mempool creation
      net/nfp: fix representor creation
      net/nfp: fix flow hash table creation
      net/bonding: fix destroy dedicated queues flow
      net/nfp: fix representor name too long

Maayan Kashani (1):
      net/mlx5: fix RSS expansion inner buffer overflow

Mattias Rönnblom (1):
      event/dsw: free rings on close

Maxime Coquelin (3):
      vhost: fix notification stats for packed ring
      vhost: fix invalid call FD handling
      crypto/qat: fix sym device prototype

Michael Baum (1):
      doc: fix format in flow API guide

Michal Schmidt (1):
      net/iavf: fix VLAN offload with AVX512

Min Zhou (1):
      net/ixgbe: add proper memory barriers in Rx

Mingjin Ye (3):
      net/ice: fix statistics
      net/ice: fix DCF control thread crash
      net/iavf: fix abnormal disable HW interrupt

Mohammad Iqbal Ahmad (1):
      examples/l3fwd: fix duplicate expression for default nexthop

Nithin Dabilpuram (2):
      common/cnxk: fix inline device VF identification
      net/cnxk: fix cookies check with security offload

Ori Kam (1):
      net/mlx5: fix query for NIC flow capability

Pavan Nikhilesh (2):
      doc: fix event timer adapter guide
      event/cnxk: fix nanoseconds to ticks conversion

Peng Zhang (1):
      net/nfp: fix address always related with PF ID 0

Philip Prindeville (1):
      vfio: fix include with musl runtime

Qi Zhang (2):
      common/idpf: remove device stop flag
      net/ice: initialize parser for double VLAN

Qiming Yang (7):
      net/ice/base: fix incorrect defines for DCBx
      net/ice/base: remove unreachable code
      net/e1000: fix queue number initialization
      net/ice: fix VLAN mode parser
      net/ixgbe: fix Rx and Tx queue status
      net/igc: fix Rx and Tx queue status
      net/e1000: fix Rx and Tx queue status

Qin Ke (1):
      net/nfp: fix disabling promiscuous mode

Rahul Bhansali (6):
      examples/ipsec-secgw: fix zero address in ethernet header
      common/cnxk: fix IPsec IPv6 tunnel address byte swap
      event/cnxk: fix Tx adapter data pointer
      event/cnxk: fix mempool cookies check
      test/security: fix event inline IPsec reassembly tests
      common/cnxk: fix CPT backpressure disable on LBK

Rakesh Kudurumalla (1):
      common/cnxk: fix receive queue with multiple mask

Raslan Darawsheh (2):
      app/testpmd: fix GTP L2 length in checksum engine
      devtools: fix bashism in mailmap check

Ronak Doshi (1):
      net/vmxnet3: fix drop of empty segments in Tx

Rongwei Liu (4):
      net/mlx5: fix matcher layout size calculation
      net/mlx5: fix VXLAN matching with zero value
      net/mlx5: forbid duplicated tag index in pattern template
      net/mlx5: fix duplicated tag index matching in SWS

Ruifeng Wang (5):
      test/malloc: fix missing free
      test/malloc: fix statistics checks
      build: fix warning when getting NUMA nodes
      test/mbuf: fix crash in a forked process
      net/mlx5: fix risk in NEON Rx descriptor read

Samina Arshad (1):
      examples/fips_validation: fix digest length in AES-GCM

Saoirse O'Donovan (3):
      test/crypto: fix return value for SNOW3G
      crypto/openssl: fix memory free
      crypto/ipsec_mb: fix enqueue counter for SNOW3G

Satha Rao (1):
      net/cnxk: flush SQ before configuring MTU

Satheesh Paul (3):
      common/cnxk: fix setting channel mask for SDP interfaces
      common/cnxk: fix uninitialized pointer read
      net/cnxk: fix flow queue index validation

Sean Zhang (1):
      net/mlx5: fix flow dump for modify field

Shijith Thotton (2):
      eventdev/timer: fix timeout event wait behavior
      event/cnxk: fix setting attributes in empty get work

Shiyang He (4):
      net/iavf: release large VF when closing device
      net/ice: fix tunnel packet Tx descriptor
      net/i40e: fix tunnel packet Tx descriptor
      app/testpmd: fix checksum engine with GTP on 32-bit

Stephen Hemminger (2):
      net/netvsc: fix sizeof calculation
      doc: fix description of runtime directories

Suanming Mou (1):
      ethdev: fix indirect action conversion

Tejasree Kondoj (3):
      test/crypto: fix session creation check
      crypto/cnxk: fix IPsec CCM capabilities
      test/crypto: fix IPsec AES CCM vector

Thierry Herbelot (1):
      doc: fix typo in cnxk platform guide

Thomas Monjalon (3):
      pci: fix comment referencing renamed function
      cryptodev: fix comments of modular operation parameters
      devtools: fix mailmap check for parentheses

Tianli Lai (1):
      net/dpaa2: fix checksum good flags

Viacheslav Ovsiienko (2):
      net/mlx5: fix device removal event handling
      ipc: fix file descriptor leakage with unhandled messages

Vikash Poddar (2):
      common/qat: fix command parameter corruption
      common/qat: detach crypto from compress build

Vladimir Medvedkin (3):
      fib: fix adding default route
      hash: fix reading unaligned bits in Toeplitz hash
      examples/ipsec-secgw: fix TAP default MAC address

Wenjing Qiao (7):
      common/idpf/base: fix control queue send and receive
      common/idpf/base: fix ITR register definitions for AVF
      common/idpf/base: fix parameters when send msg to cp
      common/idpf/base: fix memory leaks on control queue
      net/ice: adjust timestamp mbuf register
      net/ice: fix timestamp enabling
      common/iavf: fix MAC type for 710 NIC

Wenjun Wu (5):
      net/i40e: fix Rx data buffer size
      net/ice: fix Rx data buffer size
      net/iavf: fix Rx data buffer size
      net/idpf: fix Rx data buffer size
      common/idpf: fix memory leak on AVX512 Tx queue close

Wenzhuo Lu (1):
      net/iavf: fix VLAN insertion in vector path

Xiao Liang (1):
      ipsec: fix NAT-T header length

Yiding Zhou (1):
      net/ice: fix 32-bit build

Yunjian Wang (2):
      ring: fix use after free
      net/txgbe: fix use-after-free on remove

Yuying Zhang (1):
      net/ice: fix RSS hash key generation

Zerun Fu (1):
      net/bonding: fix startup when NUMA is not supported

Zhichao Zeng (2):
      net/ice: fix outer UDP checksum offload
      net/iavf: fix stop ordering

Zhihong Wang (1):
      eal/linux: fix secondary process crash for mp hotplug requests

RE: 22.11.3 patches review and test

[Ali Alnubani]

> -----Original Message-----
> From: Xueming(Steven) Li <xuemingl@nvidia.com>
> Sent: Thursday, August 17, 2023 9:14 AM
> To: stable@dpdk.org
> Cc: xuemingl@nvdia.com; dev@dpdk.org; Abhishek Marathe
> <Abhishek.Marathe@microsoft.com>; Ali Alnubani <alialnu@nvidia.com>;
> benjamin.walker@intel.com; David Christensen <drc@linux.vnet.ibm.com>;
> Hemant Agrawal <hemant.agrawal@nxp.com>; Ian Stokes
> <ian.stokes@intel.com>; Jerin Jacob <jerinj@marvell.com>; John McNamara
> <john.mcnamara@intel.com>; Ju-Hyoung Lee <juhlee@microsoft.com>; Kevin
> Traynor <ktraynor@redhat.com>; Luca Boccassi <bluca@debian.org>; Pei
> Zhang <pezhang@redhat.com>; qian.q.xu@intel.com; Raslan Darawsheh
> <rasland@nvidia.com>; NBU-Contact-Thomas Monjalon (EXTERNAL)
> <thomas@monjalon.net>; Yanghang Liu <yanghliu@redhat.com>;
> yuan.peng@intel.com; zhaoyan.chen@intel.com
> Subject: 22.11.3 patches review and test
> 
> Hi all,
> 
> Here is a list of patches targeted for stable release 22.11.3.
> 
> The planned date for the final release is 31th August.
> 
> 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=v22.11.3-rc1
> 
> These patches are located at branch 22.11 of dpdk-stable repo:
>     https://dpdk.org/browse/dpdk-stable/
> 
> Thanks.
> 
> Xueming Li <xuemingl@nvidia.com>
> 
> ---

Hello,

We ran the following functional tests with Nvidia hardware on v22.11.3-rc1:
- 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.
- 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.
- Hardware LRO tests.
- Regex application tests.
- Buffer Split tests.
- Tx scheduling tests.

Functional tests ran on:
- NIC: ConnectX-6 Dx / OS: Ubuntu 20.04 / Driver: MLNX_OFED_LINUX-23.07-0.5.0.0 / Firmware: 22.38.1002
- NIC: ConnectX-7 / OS: Ubuntu 20.04 / Driver: MLNX_OFED_LINUX-23.07-0.5.0.0 / Firmware: 28.38.1002
- DPU: BlueField-2 / DOCA SW version: 2.2.0 / Firmware: 24.38.1002

Build tests (all passed):
- Ubuntu 20.04.6 with MLNX_OFED_LINUX-23.04-1.1.3.0.
- Ubuntu 20.04.6 with rdma-core master (4cce53f).
- Ubuntu 20.04.6 with rdma-core v28.0.
- Ubuntu 18.04.6 with rdma-core master (4cce53f) (i386).
- Fedora 38 with rdma-core v44.0.
- Fedora 39 (Rawhide) with rdma-core v46.0.
- OpenSUSE Leap 15.5 with rdma-core v42.0.
- Windows Server 2019 with Clang 11.0.0.

We don't see new issues caused by the changes in this release.

Thanks,
Ali

RE: 22.11.3 patches review and test

[Xueming(Steven) Li]

Hi Ali,

Thanks for testing and report!

Regards,
Xueming

> -----Original Message-----
> From: Ali Alnubani <alialnu@nvidia.com>
> Sent: 8/25/2023 0:02
> To: Xueming(Steven) Li <xuemingl@nvidia.com>; stable@dpdk.org
> Cc: xuemingl@nvdia.com; dev@dpdk.org; Abhishek Marathe
> <Abhishek.Marathe@microsoft.com>; benjamin.walker@intel.com; David
> Christensen <drc@linux.vnet.ibm.com>; Hemant Agrawal
> <hemant.agrawal@nxp.com>; Ian Stokes <ian.stokes@intel.com>; Jerin Jacob
> <jerinj@marvell.com>; John McNamara <john.mcnamara@intel.com>; Ju-
> Hyoung Lee <juhlee@microsoft.com>; Kevin Traynor <ktraynor@redhat.com>;
> Luca Boccassi <bluca@debian.org>; Pei Zhang <pezhang@redhat.com>;
> qian.q.xu@intel.com; Raslan Darawsheh <rasland@nvidia.com>; NBU-Contact-
> Thomas Monjalon (EXTERNAL) <thomas@monjalon.net>; Yanghang Liu
> <yanghliu@redhat.com>; yuan.peng@intel.com; zhaoyan.chen@intel.com;
> Wael Abualrub <waela@nvidia.com>
> Subject: RE: 22.11.3 patches review and test
> 
> > -----Original Message-----
> > From: Xueming(Steven) Li <xuemingl@nvidia.com>
> > Sent: Thursday, August 17, 2023 9:14 AM
> > To: stable@dpdk.org
> > Cc: xuemingl@nvdia.com; dev@dpdk.org; Abhishek Marathe
> > <Abhishek.Marathe@microsoft.com>; Ali Alnubani <alialnu@nvidia.com>;
> > benjamin.walker@intel.com; David Christensen <drc@linux.vnet.ibm.com>;
> > Hemant Agrawal <hemant.agrawal@nxp.com>; Ian Stokes
> > <ian.stokes@intel.com>; Jerin Jacob <jerinj@marvell.com>; John
> > McNamara <john.mcnamara@intel.com>; Ju-Hyoung Lee
> > <juhlee@microsoft.com>; Kevin Traynor <ktraynor@redhat.com>; Luca
> > Boccassi <bluca@debian.org>; Pei Zhang <pezhang@redhat.com>;
> > qian.q.xu@intel.com; Raslan Darawsheh <rasland@nvidia.com>;
> > NBU-Contact-Thomas Monjalon (EXTERNAL) <thomas@monjalon.net>;
> Yanghang
> > Liu <yanghliu@redhat.com>; yuan.peng@intel.com;
> zhaoyan.chen@intel.com
> > Subject: 22.11.3 patches review and test
> >
> > Hi all,
> >
> > Here is a list of patches targeted for stable release 22.11.3.
> >
> > The planned date for the final release is 31th August.
> >
> > 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=v22.11.3-rc1
> >
> > These patches are located at branch 22.11 of dpdk-stable repo:
> >     https://dpdk.org/browse/dpdk-stable/
> >
> > Thanks.
> >
> > Xueming Li <xuemingl@nvidia.com>
> >
> > ---
> 
> Hello,
> 
> We ran the following functional tests with Nvidia hardware on v22.11.3-rc1:
> - 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.
> - 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.
> - Hardware LRO tests.
> - Regex application tests.
> - Buffer Split tests.
> - Tx scheduling tests.
> 
> Functional tests ran on:
> - NIC: ConnectX-6 Dx / OS: Ubuntu 20.04 / Driver: MLNX_OFED_LINUX-23.07-
> 0.5.0.0 / Firmware: 22.38.1002
> - NIC: ConnectX-7 / OS: Ubuntu 20.04 / Driver: MLNX_OFED_LINUX-23.07-
> 0.5.0.0 / Firmware: 28.38.1002
> - DPU: BlueField-2 / DOCA SW version: 2.2.0 / Firmware: 24.38.1002
> 
> Build tests (all passed):
> - Ubuntu 20.04.6 with MLNX_OFED_LINUX-23.04-1.1.3.0.
> - Ubuntu 20.04.6 with rdma-core master (4cce53f).
> - Ubuntu 20.04.6 with rdma-core v28.0.
> - Ubuntu 18.04.6 with rdma-core master (4cce53f) (i386).
> - Fedora 38 with rdma-core v44.0.
> - Fedora 39 (Rawhide) with rdma-core v46.0.
> - OpenSUSE Leap 15.5 with rdma-core v42.0.
> - Windows Server 2019 with Clang 11.0.0.
> 
> We don't see new issues caused by the changes in this release.
> 
> Thanks,
> Ali

RE: 22.11.3 patches review and test

[Xu, HailinX]

> -----Original Message-----
> From: Xueming Li <xuemingl@nvidia.com>
> Sent: Thursday, August 17, 2023 2:14 PM
> To: stable@dpdk.org
> Cc: xuemingl@nvdia.com; dev@dpdk.org; Abhishek Marathe
> <Abhishek.Marathe@microsoft.com>; Ali Alnubani <alialnu@nvidia.com>;
> Walker, Benjamin <benjamin.walker@intel.com>; David Christensen
> <drc@linux.vnet.ibm.com>; Hemant Agrawal <hemant.agrawal@nxp.com>;
> Stokes, Ian <ian.stokes@intel.com>; Jerin Jacob <jerinj@marvell.com>;
> Mcnamara, John <john.mcnamara@intel.com>; Ju-Hyoung Lee
> <juhlee@microsoft.com>; Kevin Traynor <ktraynor@redhat.com>; Luca Boccassi
> <bluca@debian.org>; Pei Zhang <pezhang@redhat.com>; Xu, Qian Q
> <qian.q.xu@intel.com>; Raslan Darawsheh <rasland@nvidia.com>; Thomas
> Monjalon <thomas@monjalon.net>; Yanghang Liu <yanghliu@redhat.com>;
> Peng, Yuan <yuan.peng@intel.com>; Chen, Zhaoyan <zhaoyan.chen@intel.com>
> Subject: 22.11.3 patches review and test
> 
> Hi all,
> 
> Here is a list of patches targeted for stable release 22.11.3.
> 
> The planned date for the final release is 31th August.
> 
> 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=v22.11.3-rc1
> 
> These patches are located at branch 22.11 of dpdk-stable repo:
>     https://dpdk.org/browse/dpdk-stable/
> 
> Thanks.

We are conducting DPDK testing and have found two issues.

1. The startup speed of testpmd is significantly slower in the os of SUSE
  This issue fix patch has been merged into main, But it has not backported to 22.11.3.
  Fix patch commit id on DPDK main: 7e7b6762eac292e78c77ad37ec0973c0c944b845

2. The SCTP tunnel packet of iavf cannot be forwarded in avx512 mode
  commit 9b7215f150d0bfc5cb00fce68ff08e5217c7f2b3 on v22.11.3-rc1.
  This commit is for the new feature (avx512 checksum offload) in DPDK 23.03, which should not be backported to the LTS version since avx512 checksum offload is not supported in v22.11.3 LTS.

Regards,
Xu, Hailin

RE: 22.11.3 patches review and test

[Xueming(Steven) Li]

> -----Original Message-----
> From: Xu, HailinX <hailinx.xu@intel.com>
> Sent: 8/29/2023 16:22
> To: Xueming(Steven) Li <xuemingl@nvidia.com>; stable@dpdk.org
> Cc: xuemingl@nvdia.com; dev@dpdk.org; Abhishek Marathe
> <Abhishek.Marathe@microsoft.com>; Ali Alnubani <alialnu@nvidia.com>;
> Walker, Benjamin <benjamin.walker@intel.com>; David Christensen
> <drc@linux.vnet.ibm.com>; Hemant Agrawal <hemant.agrawal@nxp.com>;
> Stokes, Ian <ian.stokes@intel.com>; Jerin Jacob <jerinj@marvell.com>;
> Mcnamara, John <john.mcnamara@intel.com>; Ju-Hyoung Lee
> <juhlee@microsoft.com>; Kevin Traynor <ktraynor@redhat.com>; Luca
> Boccassi <bluca@debian.org>; Pei Zhang <pezhang@redhat.com>; Xu, Qian Q
> <qian.q.xu@intel.com>; Raslan Darawsheh <rasland@nvidia.com>; NBU-
> Contact-Thomas Monjalon (EXTERNAL) <thomas@monjalon.net>; Yanghang
> Liu <yanghliu@redhat.com>; Peng, Yuan <yuan.peng@intel.com>; Chen,
> Zhaoyan <zhaoyan.chen@intel.com>
> Subject: RE: 22.11.3 patches review and test
> 
> > -----Original Message-----
> > From: Xueming Li <xuemingl@nvidia.com>
> > Sent: Thursday, August 17, 2023 2:14 PM
> > To: stable@dpdk.org
> > Cc: xuemingl@nvdia.com; dev@dpdk.org; Abhishek Marathe
> > <Abhishek.Marathe@microsoft.com>; Ali Alnubani <alialnu@nvidia.com>;
> > Walker, Benjamin <benjamin.walker@intel.com>; David Christensen
> > <drc@linux.vnet.ibm.com>; Hemant Agrawal <hemant.agrawal@nxp.com>;
> > Stokes, Ian <ian.stokes@intel.com>; Jerin Jacob <jerinj@marvell.com>;
> > Mcnamara, John <john.mcnamara@intel.com>; Ju-Hyoung Lee
> > <juhlee@microsoft.com>; Kevin Traynor <ktraynor@redhat.com>; Luca
> > Boccassi <bluca@debian.org>; Pei Zhang <pezhang@redhat.com>; Xu, Qian
> > Q <qian.q.xu@intel.com>; Raslan Darawsheh <rasland@nvidia.com>;
> Thomas
> > Monjalon <thomas@monjalon.net>; Yanghang Liu <yanghliu@redhat.com>;
> > Peng, Yuan <yuan.peng@intel.com>; Chen, Zhaoyan
> > <zhaoyan.chen@intel.com>
> > Subject: 22.11.3 patches review and test
> >
> > Hi all,
> >
> > Here is a list of patches targeted for stable release 22.11.3.
> >
> > The planned date for the final release is 31th August.
> >
> > 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=v22.11.3-rc1
> >
> > These patches are located at branch 22.11 of dpdk-stable repo:
> >     https://dpdk.org/browse/dpdk-stable/
> >
> > Thanks.
> 
> We are conducting DPDK testing and have found two issues.
> 
> 1. The startup speed of testpmd is significantly slower in the os of SUSE
>   This issue fix patch has been merged into main, But it has not backported to
> 22.11.3.
>   Fix patch commit id on DPDK main:
> 7e7b6762eac292e78c77ad37ec0973c0c944b845

Hailin, thanks for the verification and feedback!
From original patch discussion, the origin decision is no backport, that's why
"cc:stable@dpdk.org" get removed.

> 
> 2. The SCTP tunnel packet of iavf cannot be forwarded in avx512 mode
>   commit 9b7215f150d0bfc5cb00fce68ff08e5217c7f2b3 on v22.11.3-rc1.
>   This commit is for the new feature (avx512 checksum offload) in DPDK 23.03,
> which should not be backported to the LTS version since avx512 checksum
> offload is not supported in v22.11.3 LTS.

Okay, I'll revert this patch. Please note the patch has been backported to 21.11,
Do we want to fix it? @Kevin Traynor

> 
> Regards,
> Xu, Hailin

Re: 22.11.3 patches review and test

[YangHang Liu]

[-- Attachment #1: Type: text/plain, Size: 18102 bytes --]

Hi, Xueming

RedHat QE did not find any new issues about 22.11.3 rc1 dpdk during the
tests.

I tested below 18 scenarios and all got PASS on RHEL9:

   - Guest with device assignment(PF) throughput testing(1G hugepage size):
   PASS
   - Guest with device assignment(PF) throughput testing(2M hugepage size)
   : PASS
   - Guest with device assignment(VF) throughput testing: PASS
   - PVP (host dpdk testpmd as vswitch) 1Q: throughput testing: PASS
   - PVP vhost-user 2Q throughput testing: PASS
   - PVP vhost-user 1Q - cross numa node throughput testing: PASS
   - Guest with vhost-user 2 queues throughput testing: PASS
   - vhost-user reconnect with dpdk-client, qemu-server qemu reconnect: PASS
   - vhost-user reconnect with dpdk-client, qemu-server ovs reconnect: PASS
   - PVP  reconnect with dpdk-client, qemu-server: PASS
   - PVP 1Q live migration testing: PASS
   - PVP 1Q cross numa node live migration testing: PASS
   - Guest with ovs+dpdk+vhost-user 1Q live migration testing: PASS
   - Guest with ovs+dpdk+vhost-user 1Q live migration testing (2M): PASS
   - Guest with ovs+dpdk+vhost-user 2Q live migration testing: PASS
   - Guest with ovs+dpdk+vhost-user 4Q live migration testing: PASS
   - Host PF + DPDK testing: PASS
   - Host VF + DPDK testing: PASS

Test Versions:

   - qemu-kvm-6.2.0 + qemu-kvm-7.2.0
   - kernel 5.14
   - git log
        commit 5e87ad0cf8f2f90b18558913f3c71e7552aa0af5
        Author: Xueming Li <xuemingl@nvidia.com>
        Date:   Thu Aug 17 14:04:17 2023 +0800
        version: 22.11.3-rc1
        Signed-off-by: Xueming Li <xuemingl@nvidia.com>


   - Test device : X540-AT2 NIC(ixgbe, 10G)


Best Regards,
YangHang Liu



On Thu, Aug 17, 2023 at 2:14 PM Xueming Li <xuemingl@nvidia.com> wrote:

> Hi all,
>
> Here is a list of patches targeted for stable release 22.11.3.
>
> The planned date for the final release is 31th August.
>
> 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=v22.11.3-rc1
>
> These patches are located at branch 22.11 of dpdk-stable repo:
>     https://dpdk.org/browse/dpdk-stable/
>
> Thanks.
>
> Xueming Li <xuemingl@nvidia.com>
>
> ---
> Aakash Sasidharan (1):
>       test/crypto: fix PDCP-SDAP test vectors
>
> Akhil Goyal (1):
>       doc: fix auth algos in cryptoperf app
>
> Alexander Kozyrev (4):
>       net/mlx5: forbid MPRQ restart
>       net/mlx5: fix MPRQ stride size for headroom
>       net/mlx5: fix handle validation for meter mark
>       app/testpmd: fix meter mark handle update
>
> Ali Alnubani (2):
>       doc: fix typos and wording in flow API guide
>       doc: improve wording of cuda guide
>
> Alvaro Karsz (1):
>       dma/dpaa2: set VFA bit for route-by-port with VF
>
> Anoob Joseph (1):
>       cryptodev: clarify error codes for symmetric session
>
> Artemii Morozov (1):
>       common/sfc_efx/base: fix Rx queue without RSS hash prefix
>
> Ashwin Sekhar T K (2):
>       mempool/cnxk: avoid hang when counting batch allocs
>       doc: fix typo in graph guide
>
> Bing Zhao (12):
>       app/testpmd: fix flow rule number parsing
>       net/mlx5: fix error set in Tx representor tagging
>       net/mlx5: fix return value of vport action
>       net/mlx5: fix error set in control tables create
>       net/mlx5: fix error set for age pool initialization
>       net/mlx5: fix error in VLAN actions creation
>       common/mlx5: fix obtaining IB device in LAG mode
>       net/mlx5: fix profile check of meter mark
>       net/mlx5: reduce counter pool name length
>       net/mlx5: fix drop action memory leak
>       net/mlx5: fix validation for conntrack indirect action
>       doc: add flow template API requirements for mlx5
>
> Boleslav Stankevich (2):
>       net/virtio: propagate interrupt configuration error values
>       net/virtio: fix initialization to return negative errno
>
> Bruce Richardson (4):
>       kernel/freebsd: fix function parameter list
>       telemetry: fix autotest on Alpine
>       eal: avoid calling cleanup twice
>       test/bonding: fix include of standard header
>
> Chaoyong He (14):
>       net/nfp: fix VLAN push flow action
>       net/nfp: fix IPv6 flow item
>       net/nfp: fix TOS of IPv6 VXLAN encap flow action
>       net/nfp: fix TOS of IPv6 GENEVE encap flow action
>       net/nfp: fix TOS of IPv6 NVGRE encap flow action
>       net/nfp: fix TP flow action for UDP
>       net/nfp: fix IPv6 address for set flow action
>       net/nfp: fix unneeded endian conversion
>       net/nfp: fix endian conversion for tunnel decap action
>       net/nfp: fix VNI of VXLAN encap action
>       net/nfp: fix VNI of IPv4 NVGRE encap action
>       net/nfp: fix VNI of IPv6 NVGRE encap action
>       net/nfp: fix Tx descriptor free logic of NFD3
>       net/nfp: fix offloading flows
>
> Chengwen Feng (4):
>       net/hns3: fix Rx multiple firmware reset interrupts
>       net/hns3: fix mbuf leakage when RxQ started during reset
>       net/hns3: fix mbuf leakage when RxQ started after reset
>       net/hns3: fix device start return value
>
> Ciara Power (6):
>       crypto/qat: fix stack buffer overflow in SGL loop
>       crypto/scheduler: fix last element for valid args
>       crypto/ipsec_mb: optimize allocation in session
>       crypto/qat: fix null algorithm digest placement
>       cryptodev: fix device socket ID type
>       app/crypto-perf: fix socket ID default value
>
> Dariusz Sosnowski (1):
>       net/mlx5: fix drop action attribute validation
>
> David Christensen (1):
>       net/tap: set locally administered bit for fixed MAC address
>
> David Marchand (9):
>       ci: fix build for Arm cross compilation in GHA
>       net/virtio-user: fix leak when initialisation fails
>       net/mlx5: enhance error log for tunnel offloading
>       examples/l2fwd-cat: fix external build
>       examples/fips_validation: fix external build
>       test: add graph tests
>       mbuf: fix Doxygen comment of distributor metadata
>       ci: fix libabigail cache in GHA
>       doc: fix link to flow capabilities from bnxt guide
>
> Dengdui Huang (3):
>       net/hns3: fix variable type mismatch
>       net/hns3: fix inaccurate log
>       net/hns3: fix redundant line break in log
>
> Denis Pryazhennikov (3):
>       ethdev: update documentation for API to set FEC
>       ethdev: check that at least one FEC mode is specified
>       ethdev: update documentation for API to get FEC
>
> Devendra Singh Rawat (1):
>       net/qede: fix RSS indirection table initialization
>
> Didier Pallard (2):
>       crypto/openssl: fix memory leak in auth processing
>       crypto/openssl: skip workaround at compilation time
>
> Dmitry Kozlyuk (1):
>       member: fix PRNG seed reset in NitroSketch mode
>
> Dongdong Liu (3):
>       doc: fix kernel patch link in hns3 guide
>       doc: fix syntax in hns3 guide
>       doc: fix number of leading spaces in hns3 guide
>
> Elena Agostini (1):
>       net: fix return type of IPv4 L4 packet checksum
>
> Erez Ferber (1):
>       common/mlx5: adjust fork call with new kernel API
>
> Erik Gabriel Carrillo (1):
>       eventdev/timer: fix buffer flush
>
> Feifei Wang (1):
>       ring: fix dequeue parameter name
>
> Fengnan Chang (2):
>       eal/linux: fix legacy mem init with many segments
>       mem: fix memsegs exhausted message
>
> Ferruh Yigit (3):
>       kni: fix build with Linux 6.3
>       app/testpmd: revert primary process polling all queues fix
>       kni: fix build with Linux 6.5
>
> Gregory Etelson (1):
>       net/mlx5: fix flow workspace destruction
>
> Heng Jiang (1):
>       net/mlx5: fix LRO TCP checksum
>
> Hernan Vargas (2):
>       baseband/fpga_5gnr_fec: fix possible division by zero
>       baseband/fpga_5gnr_fec: fix starting unconfigured queue
>
> Huisong Li (10):
>       ethdev: fix MAC address occupies two entries
>       net/hns3: fix never set MAC flow control
>       net/hns3: fix RTC time on initialization
>       net/hns3: fix RTC time after reset
>       net/hns3: uninitialize PTP
>       net/hns3: extract PTP to its own header file
>       net/hns3: fix IMP reset trigger
>       net/hns3: delete duplicate macro definition
>       net/hns3: fix non-zero weight for disabled TC
>       net/hns3: fix index to look up table in NEON Rx
>
> Ivan Malov (2):
>       net/sfc: stop misuse of Rx ingress m-port metadata on EF100
>       net/sfc: invalidate dangling MAE flow action FW resource IDs
>
> Jerin Jacob (4):
>       examples/ip_pipeline: fix build with GCC 13
>       examples/ntb: fix build with GCC 13
>       net/hns3: fix build warning
>       doc: remove warning with Doxygen 1.9.7
>
> Jiawen Wu (11):
>       net/txgbe/base: fix Tx with fiber hotplug
>       net/txgbe: fix interrupt enable mask
>       net/txgbe: adapt to MNG veto bit setting
>       net/txgbe: fix to set autoneg for 1G speed
>       net/txgbe: fix extended statistics
>       net/ngbe: fix extended statistics
>       net/ngbe: adapt to MNG veto bit setting
>       net/ngbe: fix link status in no LSC mode
>       net/ngbe: remove redundant codes
>       net/txgbe: fix blocking system events
>       net/ngbe: fix RSS offload capability
>
> Jie Hai (6):
>       net/hns3: fix FEC mode for 200G ports
>       net/hns3: fix FEC mode check
>       net/hns3: fix missing FEC capability
>       net/hns3: fix uninitialized variable
>       ethdev: fix calloc arguments
>       app/testpmd: fix primary process not polling all queues
>
> Jieqiang Wang (1):
>       net/i40e: fix comments
>
> Junfeng Guo (4):
>       net/ice: fix protocol agnostic offloading with big packets
>       net/iavf: fix protocol agnostic offloading with big packets
>       raw/ntb: avoid disabling interrupt twice
>       doc: update BIOS settings and supported HW for NTB
>
> Kai Ji (1):
>       examples/ipsec-secgw: fix socket ID default value
>
> Kaijun Zeng (1):
>       net/vmxnet3: fix return code in initializing
>
> Kaiwen Deng (3):
>       net/ice: fix DCF RSS initialization
>       net/iavf: fix virtchnl command called in interrupt
>       net/ice: fix MAC type of E822 and E823
>
> Kaiyu Zhang (1):
>       ethdev: fix potential leak in PCI probing helper
>
> Kamalakannan R (1):
>       pipeline: fix double free for table stats
>
> Ke Xu (1):
>       net/iavf: fix tunnel TSO path selection
>
> Leyi Rong (1):
>       eal/x86: improve multiple of 64 bytes memcpy performance
>
> Long Li (7):
>       net/mana: use datapath logging
>       net/mana: avoid unnecessary assignments in data path
>       net/mana: optimize completion queue by batch processing
>       net/mana: return probing failure if no device found
>       net/mana: fix counter overflow for posted WQE
>       net/mana: fix WQE count for ringing RQ doorbell
>       net/mana: fix Tx queue statistics
>
> Long Wu (5):
>       net/nfp: fix control mempool creation
>       net/nfp: fix representor creation
>       net/nfp: fix flow hash table creation
>       net/bonding: fix destroy dedicated queues flow
>       net/nfp: fix representor name too long
>
> Maayan Kashani (1):
>       net/mlx5: fix RSS expansion inner buffer overflow
>
> Mattias Rönnblom (1):
>       event/dsw: free rings on close
>
> Maxime Coquelin (3):
>       vhost: fix notification stats for packed ring
>       vhost: fix invalid call FD handling
>       crypto/qat: fix sym device prototype
>
> Michael Baum (1):
>       doc: fix format in flow API guide
>
> Michal Schmidt (1):
>       net/iavf: fix VLAN offload with AVX512
>
> Min Zhou (1):
>       net/ixgbe: add proper memory barriers in Rx
>
> Mingjin Ye (3):
>       net/ice: fix statistics
>       net/ice: fix DCF control thread crash
>       net/iavf: fix abnormal disable HW interrupt
>
> Mohammad Iqbal Ahmad (1):
>       examples/l3fwd: fix duplicate expression for default nexthop
>
> Nithin Dabilpuram (2):
>       common/cnxk: fix inline device VF identification
>       net/cnxk: fix cookies check with security offload
>
> Ori Kam (1):
>       net/mlx5: fix query for NIC flow capability
>
> Pavan Nikhilesh (2):
>       doc: fix event timer adapter guide
>       event/cnxk: fix nanoseconds to ticks conversion
>
> Peng Zhang (1):
>       net/nfp: fix address always related with PF ID 0
>
> Philip Prindeville (1):
>       vfio: fix include with musl runtime
>
> Qi Zhang (2):
>       common/idpf: remove device stop flag
>       net/ice: initialize parser for double VLAN
>
> Qiming Yang (7):
>       net/ice/base: fix incorrect defines for DCBx
>       net/ice/base: remove unreachable code
>       net/e1000: fix queue number initialization
>       net/ice: fix VLAN mode parser
>       net/ixgbe: fix Rx and Tx queue status
>       net/igc: fix Rx and Tx queue status
>       net/e1000: fix Rx and Tx queue status
>
> Qin Ke (1):
>       net/nfp: fix disabling promiscuous mode
>
> Rahul Bhansali (6):
>       examples/ipsec-secgw: fix zero address in ethernet header
>       common/cnxk: fix IPsec IPv6 tunnel address byte swap
>       event/cnxk: fix Tx adapter data pointer
>       event/cnxk: fix mempool cookies check
>       test/security: fix event inline IPsec reassembly tests
>       common/cnxk: fix CPT backpressure disable on LBK
>
> Rakesh Kudurumalla (1):
>       common/cnxk: fix receive queue with multiple mask
>
> Raslan Darawsheh (2):
>       app/testpmd: fix GTP L2 length in checksum engine
>       devtools: fix bashism in mailmap check
>
> Ronak Doshi (1):
>       net/vmxnet3: fix drop of empty segments in Tx
>
> Rongwei Liu (4):
>       net/mlx5: fix matcher layout size calculation
>       net/mlx5: fix VXLAN matching with zero value
>       net/mlx5: forbid duplicated tag index in pattern template
>       net/mlx5: fix duplicated tag index matching in SWS
>
> Ruifeng Wang (5):
>       test/malloc: fix missing free
>       test/malloc: fix statistics checks
>       build: fix warning when getting NUMA nodes
>       test/mbuf: fix crash in a forked process
>       net/mlx5: fix risk in NEON Rx descriptor read
>
> Samina Arshad (1):
>       examples/fips_validation: fix digest length in AES-GCM
>
> Saoirse O'Donovan (3):
>       test/crypto: fix return value for SNOW3G
>       crypto/openssl: fix memory free
>       crypto/ipsec_mb: fix enqueue counter for SNOW3G
>
> Satha Rao (1):
>       net/cnxk: flush SQ before configuring MTU
>
> Satheesh Paul (3):
>       common/cnxk: fix setting channel mask for SDP interfaces
>       common/cnxk: fix uninitialized pointer read
>       net/cnxk: fix flow queue index validation
>
> Sean Zhang (1):
>       net/mlx5: fix flow dump for modify field
>
> Shijith Thotton (2):
>       eventdev/timer: fix timeout event wait behavior
>       event/cnxk: fix setting attributes in empty get work
>
> Shiyang He (4):
>       net/iavf: release large VF when closing device
>       net/ice: fix tunnel packet Tx descriptor
>       net/i40e: fix tunnel packet Tx descriptor
>       app/testpmd: fix checksum engine with GTP on 32-bit
>
> Stephen Hemminger (2):
>       net/netvsc: fix sizeof calculation
>       doc: fix description of runtime directories
>
> Suanming Mou (1):
>       ethdev: fix indirect action conversion
>
> Tejasree Kondoj (3):
>       test/crypto: fix session creation check
>       crypto/cnxk: fix IPsec CCM capabilities
>       test/crypto: fix IPsec AES CCM vector
>
> Thierry Herbelot (1):
>       doc: fix typo in cnxk platform guide
>
> Thomas Monjalon (3):
>       pci: fix comment referencing renamed function
>       cryptodev: fix comments of modular operation parameters
>       devtools: fix mailmap check for parentheses
>
> Tianli Lai (1):
>       net/dpaa2: fix checksum good flags
>
> Viacheslav Ovsiienko (2):
>       net/mlx5: fix device removal event handling
>       ipc: fix file descriptor leakage with unhandled messages
>
> Vikash Poddar (2):
>       common/qat: fix command parameter corruption
>       common/qat: detach crypto from compress build
>
> Vladimir Medvedkin (3):
>       fib: fix adding default route
>       hash: fix reading unaligned bits in Toeplitz hash
>       examples/ipsec-secgw: fix TAP default MAC address
>
> Wenjing Qiao (7):
>       common/idpf/base: fix control queue send and receive
>       common/idpf/base: fix ITR register definitions for AVF
>       common/idpf/base: fix parameters when send msg to cp
>       common/idpf/base: fix memory leaks on control queue
>       net/ice: adjust timestamp mbuf register
>       net/ice: fix timestamp enabling
>       common/iavf: fix MAC type for 710 NIC
>
> Wenjun Wu (5):
>       net/i40e: fix Rx data buffer size
>       net/ice: fix Rx data buffer size
>       net/iavf: fix Rx data buffer size
>       net/idpf: fix Rx data buffer size
>       common/idpf: fix memory leak on AVX512 Tx queue close
>
> Wenzhuo Lu (1):
>       net/iavf: fix VLAN insertion in vector path
>
> Xiao Liang (1):
>       ipsec: fix NAT-T header length
>
> Yiding Zhou (1):
>       net/ice: fix 32-bit build
>
> Yunjian Wang (2):
>       ring: fix use after free
>       net/txgbe: fix use-after-free on remove
>
> Yuying Zhang (1):
>       net/ice: fix RSS hash key generation
>
> Zerun Fu (1):
>       net/bonding: fix startup when NUMA is not supported
>
> Zhichao Zeng (2):
>       net/ice: fix outer UDP checksum offload
>       net/iavf: fix stop ordering
>
> Zhihong Wang (1):
>       eal/linux: fix secondary process crash for mp hotplug requests
>
>

[-- Attachment #2: Type: text/html, Size: 22609 bytes --]

RE: 22.11.3 patches review and test

[Xueming(Steven) Li]

Hi Yanghang,

Thanks for the verification and report, will include in release notes.

Regards,
Xueming Li

From: YangHang Liu <yanghliu@redhat.com> 
Sent: 8/30/2023 0:03
To: Xueming(Steven) Li <xuemingl@nvidia.com>
Cc: stable@dpdk.org; xuemingl@nvdia.com; dev@dpdk.org; Abhishek Marathe <Abhishek.Marathe@microsoft.com>; Ali Alnubani <alialnu@nvidia.com>; benjamin.walker@intel.com; David Christensen <drc@linux.vnet.ibm.com>; Hemant Agrawal <hemant.agrawal@nxp.com>; Ian Stokes <ian.stokes@intel.com>; Jerin Jacob <jerinj@marvell.com>; John McNamara <john.mcnamara@intel.com>; Ju-Hyoung Lee <juhlee@microsoft.com>; Kevin Traynor <ktraynor@redhat.com>; Luca Boccassi <bluca@debian.org>; Pei Zhang <pezhang@redhat.com>; qian.q.xu@intel.com; Raslan Darawsheh <rasland@nvidia.com>; NBU-Contact-Thomas Monjalon (EXTERNAL) <thomas@monjalon.net>; yuan.peng@intel.com; zhaoyan.chen@intel.com
Subject: Re: 22.11.3 patches review and test

Hi, Xueming

RedHat QE did not find any new issues about 22.11.3 rc1 dpdk during the tests.

I tested below 18 scenarios and all got PASS on RHEL9:
• Guest with device assignment(PF) throughput testing(1G hugepage size): PASS
• Guest with device assignment(PF) throughput testing(2M hugepage size) : PASS
• Guest with device assignment(VF) throughput testing: PASS
• PVP (host dpdk testpmd as vswitch) 1Q: throughput testing: PASS
• PVP vhost-user 2Q throughput testing: PASS
• PVP vhost-user 1Q - cross numa node throughput testing: PASS
• Guest with vhost-user 2 queues throughput testing: PASS
• vhost-user reconnect with dpdk-client, qemu-server qemu reconnect: PASS
• vhost-user reconnect with dpdk-client, qemu-server ovs reconnect: PASS
• PVP  reconnect with dpdk-client, qemu-server: PASS
• PVP 1Q live migration testing: PASS
• PVP 1Q cross numa node live migration testing: PASS
• Guest with ovs+dpdk+vhost-user 1Q live migration testing: PASS
• Guest with ovs+dpdk+vhost-user 1Q live migration testing (2M): PASS
• Guest with ovs+dpdk+vhost-user 2Q live migration testing: PASS
• Guest with ovs+dpdk+vhost-user 4Q live migration testing: PASS
• Host PF + DPDK testing: PASS
• Host VF + DPDK testing: PASS
Test Versions:
• qemu-kvm-6.2.0 + qemu-kvm-7.2.0
• kernel 5.14 
• git log
     commit 5e87ad0cf8f2f90b18558913f3c71e7552aa0af5
     Author: Xueming Li <mailto:xuemingl@nvidia.com>
     Date:   Thu Aug 17 14:04:17 2023 +0800
     version: 22.11.3-rc1
     Signed-off-by: Xueming Li <mailto:xuemingl@nvidia.com>
• Test device : X540-AT2 NIC(ixgbe, 10G)

Best Regards,
YangHang Liu



On Thu, Aug 17, 2023 at 2:14 PM Xueming Li <mailto:xuemingl@nvidia.com> wrote:
Hi all,

Here is a list of patches targeted for stable release 22.11.3.

The planned date for the final release is 31th August.

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=v22.11.3-rc1

These patches are located at branch 22.11 of dpdk-stable repo:
    https://dpdk.org/browse/dpdk-stable/

Thanks.

Xueming Li <mailto:xuemingl@nvidia.com>

---
Aakash Sasidharan (1):
      test/crypto: fix PDCP-SDAP test vectors

Akhil Goyal (1):
      doc: fix auth algos in cryptoperf app

Alexander Kozyrev (4):
      net/mlx5: forbid MPRQ restart
      net/mlx5: fix MPRQ stride size for headroom
      net/mlx5: fix handle validation for meter mark
      app/testpmd: fix meter mark handle update

Ali Alnubani (2):
      doc: fix typos and wording in flow API guide
      doc: improve wording of cuda guide

Alvaro Karsz (1):
      dma/dpaa2: set VFA bit for route-by-port with VF

Anoob Joseph (1):
      cryptodev: clarify error codes for symmetric session

Artemii Morozov (1):
      common/sfc_efx/base: fix Rx queue without RSS hash prefix

Ashwin Sekhar T K (2):
      mempool/cnxk: avoid hang when counting batch allocs
      doc: fix typo in graph guide

Bing Zhao (12):
      app/testpmd: fix flow rule number parsing
      net/mlx5: fix error set in Tx representor tagging
      net/mlx5: fix return value of vport action
      net/mlx5: fix error set in control tables create
      net/mlx5: fix error set for age pool initialization
      net/mlx5: fix error in VLAN actions creation
      common/mlx5: fix obtaining IB device in LAG mode
      net/mlx5: fix profile check of meter mark
      net/mlx5: reduce counter pool name length
      net/mlx5: fix drop action memory leak
      net/mlx5: fix validation for conntrack indirect action
      doc: add flow template API requirements for mlx5

Boleslav Stankevich (2):
      net/virtio: propagate interrupt configuration error values
      net/virtio: fix initialization to return negative errno

Bruce Richardson (4):
      kernel/freebsd: fix function parameter list
      telemetry: fix autotest on Alpine
      eal: avoid calling cleanup twice
      test/bonding: fix include of standard header

Chaoyong He (14):
      net/nfp: fix VLAN push flow action
      net/nfp: fix IPv6 flow item
      net/nfp: fix TOS of IPv6 VXLAN encap flow action
      net/nfp: fix TOS of IPv6 GENEVE encap flow action
      net/nfp: fix TOS of IPv6 NVGRE encap flow action
      net/nfp: fix TP flow action for UDP
      net/nfp: fix IPv6 address for set flow action
      net/nfp: fix unneeded endian conversion
      net/nfp: fix endian conversion for tunnel decap action
      net/nfp: fix VNI of VXLAN encap action
      net/nfp: fix VNI of IPv4 NVGRE encap action
      net/nfp: fix VNI of IPv6 NVGRE encap action
      net/nfp: fix Tx descriptor free logic of NFD3
      net/nfp: fix offloading flows

Chengwen Feng (4):
      net/hns3: fix Rx multiple firmware reset interrupts
      net/hns3: fix mbuf leakage when RxQ started during reset
      net/hns3: fix mbuf leakage when RxQ started after reset
      net/hns3: fix device start return value

Ciara Power (6):
      crypto/qat: fix stack buffer overflow in SGL loop
      crypto/scheduler: fix last element for valid args
      crypto/ipsec_mb: optimize allocation in session
      crypto/qat: fix null algorithm digest placement
      cryptodev: fix device socket ID type
      app/crypto-perf: fix socket ID default value

Dariusz Sosnowski (1):
      net/mlx5: fix drop action attribute validation

David Christensen (1):
      net/tap: set locally administered bit for fixed MAC address

David Marchand (9):
      ci: fix build for Arm cross compilation in GHA
      net/virtio-user: fix leak when initialisation fails
      net/mlx5: enhance error log for tunnel offloading
      examples/l2fwd-cat: fix external build
      examples/fips_validation: fix external build
      test: add graph tests
      mbuf: fix Doxygen comment of distributor metadata
      ci: fix libabigail cache in GHA
      doc: fix link to flow capabilities from bnxt guide

Dengdui Huang (3):
      net/hns3: fix variable type mismatch
      net/hns3: fix inaccurate log
      net/hns3: fix redundant line break in log

Denis Pryazhennikov (3):
      ethdev: update documentation for API to set FEC
      ethdev: check that at least one FEC mode is specified
      ethdev: update documentation for API to get FEC

Devendra Singh Rawat (1):
      net/qede: fix RSS indirection table initialization

Didier Pallard (2):
      crypto/openssl: fix memory leak in auth processing
      crypto/openssl: skip workaround at compilation time

Dmitry Kozlyuk (1):
      member: fix PRNG seed reset in NitroSketch mode

Dongdong Liu (3):
      doc: fix kernel patch link in hns3 guide
      doc: fix syntax in hns3 guide
      doc: fix number of leading spaces in hns3 guide

Elena Agostini (1):
      net: fix return type of IPv4 L4 packet checksum

Erez Ferber (1):
      common/mlx5: adjust fork call with new kernel API

Erik Gabriel Carrillo (1):
      eventdev/timer: fix buffer flush

Feifei Wang (1):
      ring: fix dequeue parameter name

Fengnan Chang (2):
      eal/linux: fix legacy mem init with many segments
      mem: fix memsegs exhausted message

Ferruh Yigit (3):
      kni: fix build with Linux 6.3
      app/testpmd: revert primary process polling all queues fix
      kni: fix build with Linux 6.5

Gregory Etelson (1):
      net/mlx5: fix flow workspace destruction

Heng Jiang (1):
      net/mlx5: fix LRO TCP checksum

Hernan Vargas (2):
      baseband/fpga_5gnr_fec: fix possible division by zero
      baseband/fpga_5gnr_fec: fix starting unconfigured queue

Huisong Li (10):
      ethdev: fix MAC address occupies two entries
      net/hns3: fix never set MAC flow control
      net/hns3: fix RTC time on initialization
      net/hns3: fix RTC time after reset
      net/hns3: uninitialize PTP
      net/hns3: extract PTP to its own header file
      net/hns3: fix IMP reset trigger
      net/hns3: delete duplicate macro definition
      net/hns3: fix non-zero weight for disabled TC
      net/hns3: fix index to look up table in NEON Rx

Ivan Malov (2):
      net/sfc: stop misuse of Rx ingress m-port metadata on EF100
      net/sfc: invalidate dangling MAE flow action FW resource IDs

Jerin Jacob (4):
      examples/ip_pipeline: fix build with GCC 13
      examples/ntb: fix build with GCC 13
      net/hns3: fix build warning
      doc: remove warning with Doxygen 1.9.7

Jiawen Wu (11):
      net/txgbe/base: fix Tx with fiber hotplug
      net/txgbe: fix interrupt enable mask
      net/txgbe: adapt to MNG veto bit setting
      net/txgbe: fix to set autoneg for 1G speed
      net/txgbe: fix extended statistics
      net/ngbe: fix extended statistics
      net/ngbe: adapt to MNG veto bit setting
      net/ngbe: fix link status in no LSC mode
      net/ngbe: remove redundant codes
      net/txgbe: fix blocking system events
      net/ngbe: fix RSS offload capability

Jie Hai (6):
      net/hns3: fix FEC mode for 200G ports
      net/hns3: fix FEC mode check
      net/hns3: fix missing FEC capability
      net/hns3: fix uninitialized variable
      ethdev: fix calloc arguments
      app/testpmd: fix primary process not polling all queues

Jieqiang Wang (1):
      net/i40e: fix comments

Junfeng Guo (4):
      net/ice: fix protocol agnostic offloading with big packets
      net/iavf: fix protocol agnostic offloading with big packets
      raw/ntb: avoid disabling interrupt twice
      doc: update BIOS settings and supported HW for NTB

Kai Ji (1):
      examples/ipsec-secgw: fix socket ID default value

Kaijun Zeng (1):
      net/vmxnet3: fix return code in initializing

Kaiwen Deng (3):
      net/ice: fix DCF RSS initialization
      net/iavf: fix virtchnl command called in interrupt
      net/ice: fix MAC type of E822 and E823

Kaiyu Zhang (1):
      ethdev: fix potential leak in PCI probing helper

Kamalakannan R (1):
      pipeline: fix double free for table stats

Ke Xu (1):
      net/iavf: fix tunnel TSO path selection

Leyi Rong (1):
      eal/x86: improve multiple of 64 bytes memcpy performance

Long Li (7):
      net/mana: use datapath logging
      net/mana: avoid unnecessary assignments in data path
      net/mana: optimize completion queue by batch processing
      net/mana: return probing failure if no device found
      net/mana: fix counter overflow for posted WQE
      net/mana: fix WQE count for ringing RQ doorbell
      net/mana: fix Tx queue statistics

Long Wu (5):
      net/nfp: fix control mempool creation
      net/nfp: fix representor creation
      net/nfp: fix flow hash table creation
      net/bonding: fix destroy dedicated queues flow
      net/nfp: fix representor name too long

Maayan Kashani (1):
      net/mlx5: fix RSS expansion inner buffer overflow

Mattias Rönnblom (1):
      event/dsw: free rings on close

Maxime Coquelin (3):
      vhost: fix notification stats for packed ring
      vhost: fix invalid call FD handling
      crypto/qat: fix sym device prototype

Michael Baum (1):
      doc: fix format in flow API guide

Michal Schmidt (1):
      net/iavf: fix VLAN offload with AVX512

Min Zhou (1):
      net/ixgbe: add proper memory barriers in Rx

Mingjin Ye (3):
      net/ice: fix statistics
      net/ice: fix DCF control thread crash
      net/iavf: fix abnormal disable HW interrupt

Mohammad Iqbal Ahmad (1):
      examples/l3fwd: fix duplicate expression for default nexthop

Nithin Dabilpuram (2):
      common/cnxk: fix inline device VF identification
      net/cnxk: fix cookies check with security offload

Ori Kam (1):
      net/mlx5: fix query for NIC flow capability

Pavan Nikhilesh (2):
      doc: fix event timer adapter guide
      event/cnxk: fix nanoseconds to ticks conversion

Peng Zhang (1):
      net/nfp: fix address always related with PF ID 0

Philip Prindeville (1):
      vfio: fix include with musl runtime

Qi Zhang (2):
      common/idpf: remove device stop flag
      net/ice: initialize parser for double VLAN

Qiming Yang (7):
      net/ice/base: fix incorrect defines for DCBx
      net/ice/base: remove unreachable code
      net/e1000: fix queue number initialization
      net/ice: fix VLAN mode parser
      net/ixgbe: fix Rx and Tx queue status
      net/igc: fix Rx and Tx queue status
      net/e1000: fix Rx and Tx queue status

Qin Ke (1):
      net/nfp: fix disabling promiscuous mode

Rahul Bhansali (6):
      examples/ipsec-secgw: fix zero address in ethernet header
      common/cnxk: fix IPsec IPv6 tunnel address byte swap
      event/cnxk: fix Tx adapter data pointer
      event/cnxk: fix mempool cookies check
      test/security: fix event inline IPsec reassembly tests
      common/cnxk: fix CPT backpressure disable on LBK

Rakesh Kudurumalla (1):
      common/cnxk: fix receive queue with multiple mask

Raslan Darawsheh (2):
      app/testpmd: fix GTP L2 length in checksum engine
      devtools: fix bashism in mailmap check

Ronak Doshi (1):
      net/vmxnet3: fix drop of empty segments in Tx

Rongwei Liu (4):
      net/mlx5: fix matcher layout size calculation
      net/mlx5: fix VXLAN matching with zero value
      net/mlx5: forbid duplicated tag index in pattern template
      net/mlx5: fix duplicated tag index matching in SWS

Ruifeng Wang (5):
      test/malloc: fix missing free
      test/malloc: fix statistics checks
      build: fix warning when getting NUMA nodes
      test/mbuf: fix crash in a forked process
      net/mlx5: fix risk in NEON Rx descriptor read

Samina Arshad (1):
      examples/fips_validation: fix digest length in AES-GCM

Saoirse O'Donovan (3):
      test/crypto: fix return value for SNOW3G
      crypto/openssl: fix memory free
      crypto/ipsec_mb: fix enqueue counter for SNOW3G

Satha Rao (1):
      net/cnxk: flush SQ before configuring MTU

Satheesh Paul (3):
      common/cnxk: fix setting channel mask for SDP interfaces
      common/cnxk: fix uninitialized pointer read
      net/cnxk: fix flow queue index validation

Sean Zhang (1):
      net/mlx5: fix flow dump for modify field

Shijith Thotton (2):
      eventdev/timer: fix timeout event wait behavior
      event/cnxk: fix setting attributes in empty get work

Shiyang He (4):
      net/iavf: release large VF when closing device
      net/ice: fix tunnel packet Tx descriptor
      net/i40e: fix tunnel packet Tx descriptor
      app/testpmd: fix checksum engine with GTP on 32-bit

Stephen Hemminger (2):
      net/netvsc: fix sizeof calculation
      doc: fix description of runtime directories

Suanming Mou (1):
      ethdev: fix indirect action conversion

Tejasree Kondoj (3):
      test/crypto: fix session creation check
      crypto/cnxk: fix IPsec CCM capabilities
      test/crypto: fix IPsec AES CCM vector

Thierry Herbelot (1):
      doc: fix typo in cnxk platform guide

Thomas Monjalon (3):
      pci: fix comment referencing renamed function
      cryptodev: fix comments of modular operation parameters
      devtools: fix mailmap check for parentheses

Tianli Lai (1):
      net/dpaa2: fix checksum good flags

Viacheslav Ovsiienko (2):
      net/mlx5: fix device removal event handling
      ipc: fix file descriptor leakage with unhandled messages

Vikash Poddar (2):
      common/qat: fix command parameter corruption
      common/qat: detach crypto from compress build

Vladimir Medvedkin (3):
      fib: fix adding default route
      hash: fix reading unaligned bits in Toeplitz hash
      examples/ipsec-secgw: fix TAP default MAC address

Wenjing Qiao (7):
      common/idpf/base: fix control queue send and receive
      common/idpf/base: fix ITR register definitions for AVF
      common/idpf/base: fix parameters when send msg to cp
      common/idpf/base: fix memory leaks on control queue
      net/ice: adjust timestamp mbuf register
      net/ice: fix timestamp enabling
      common/iavf: fix MAC type for 710 NIC

Wenjun Wu (5):
      net/i40e: fix Rx data buffer size
      net/ice: fix Rx data buffer size
      net/iavf: fix Rx data buffer size
      net/idpf: fix Rx data buffer size
      common/idpf: fix memory leak on AVX512 Tx queue close

Wenzhuo Lu (1):
      net/iavf: fix VLAN insertion in vector path

Xiao Liang (1):
      ipsec: fix NAT-T header length

Yiding Zhou (1):
      net/ice: fix 32-bit build

Yunjian Wang (2):
      ring: fix use after free
      net/txgbe: fix use-after-free on remove

Yuying Zhang (1):
      net/ice: fix RSS hash key generation

Zerun Fu (1):
      net/bonding: fix startup when NUMA is not supported

Zhichao Zeng (2):
      net/ice: fix outer UDP checksum offload
      net/iavf: fix stop ordering

Zhihong Wang (1):
      eal/linux: fix secondary process crash for mp hotplug requests

Re: 22.11.3 patches review and test

[Kevin Traynor]

On 29/08/2023 09:22, Xu, HailinX wrote:
>> -----Original Message-----
>> From: Xueming Li <xuemingl@nvidia.com>
>> Sent: Thursday, August 17, 2023 2:14 PM
>> To: stable@dpdk.org
>> Cc: xuemingl@nvdia.com; dev@dpdk.org; Abhishek Marathe
>> <Abhishek.Marathe@microsoft.com>; Ali Alnubani <alialnu@nvidia.com>;
>> Walker, Benjamin <benjamin.walker@intel.com>; David Christensen
>> <drc@linux.vnet.ibm.com>; Hemant Agrawal <hemant.agrawal@nxp.com>;
>> Stokes, Ian <ian.stokes@intel.com>; Jerin Jacob <jerinj@marvell.com>;
>> Mcnamara, John <john.mcnamara@intel.com>; Ju-Hyoung Lee
>> <juhlee@microsoft.com>; Kevin Traynor <ktraynor@redhat.com>; Luca Boccassi
>> <bluca@debian.org>; Pei Zhang <pezhang@redhat.com>; Xu, Qian Q
>> <qian.q.xu@intel.com>; Raslan Darawsheh <rasland@nvidia.com>; Thomas
>> Monjalon <thomas@monjalon.net>; Yanghang Liu <yanghliu@redhat.com>;
>> Peng, Yuan <yuan.peng@intel.com>; Chen, Zhaoyan <zhaoyan.chen@intel.com>
>> Subject: 22.11.3 patches review and test
>>
>> Hi all,
>>
>> Here is a list of patches targeted for stable release 22.11.3.
>>
>> The planned date for the final release is 31th August.
>>
>> 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=v22.11.3-rc1
>>
>> These patches are located at branch 22.11 of dpdk-stable repo:
>>      https://dpdk.org/browse/dpdk-stable/
>>
>> Thanks.
> 
> We are conducting DPDK testing and have found two issues.
> 
> 1. The startup speed of testpmd is significantly slower in the os of SUSE
>    This issue fix patch has been merged into main, But it has not backported to 22.11.3.
>    Fix patch commit id on DPDK main: 7e7b6762eac292e78c77ad37ec0973c0c944b845
> 
> 2. The SCTP tunnel packet of iavf cannot be forwarded in avx512 mode
>    commit 9b7215f150d0bfc5cb00fce68ff08e5217c7f2b3 on v22.11.3-rc1.
>    This commit is for the new feature (avx512 checksum offload) in DPDK 23.03, which should not be backported to the LTS version since avx512 checksum offload is not supported in v22.11.3 LTS.
> 

Thanks for flagging Xueming.

The issue is that it was listed as fixing 059f18ae2aec ("net/iavf: add 
offload path for Tx AVX512") which goes back to 21.05.

This could have been avoided if the 'Fixes:' tag was correct, or if the 
authors replied to the email about queued backports :/

Requesting iavf/next-net-intel maintainers to check Fixes: tags are 
correct before merging.

DPDK 21.11.5 is already released with this patch. Any idea why it did 
not show up in validation for 21.11.5 ? Is it an issue for 21.11.5 ? How 
critical is it ?

I can revert it on the 21.11 branch, but it will need to wait until 
21.11.6 in December before it will be reverted in a released version.

thanks,
Kevin.

> Regards,
> Xu, Hailin
> 

RE: 22.11.3 patches review and test

[Xu, HailinX]

> -----Original Message-----
> From: Xueming(Steven) Li <xuemingl@nvidia.com>
> Sent: Tuesday, August 29, 2023 4:56 PM
> To: Xu, HailinX <hailinx.xu@intel.com>; stable@dpdk.org; Kevin Traynor
> <ktraynor@redhat.com>
> Cc: xuemingl@nvdia.com; dev@dpdk.org; Abhishek Marathe
> <Abhishek.Marathe@microsoft.com>; Ali Alnubani <alialnu@nvidia.com>;
> Walker, Benjamin <benjamin.walker@intel.com>; David Christensen
> <drc@linux.vnet.ibm.com>; Hemant Agrawal <hemant.agrawal@nxp.com>;
> Stokes, Ian <ian.stokes@intel.com>; Jerin Jacob <jerinj@marvell.com>;
> Mcnamara, John <john.mcnamara@intel.com>; Ju-Hyoung Lee
> <juhlee@microsoft.com>; Luca Boccassi <bluca@debian.org>; Pei Zhang
> <pezhang@redhat.com>; Xu, Qian Q <qian.q.xu@intel.com>; Raslan Darawsheh
> <rasland@nvidia.com>; NBU-Contact-Thomas Monjalon (EXTERNAL)
> <thomas@monjalon.net>; Yanghang Liu <yanghliu@redhat.com>; Peng, Yuan
> <yuan.peng@intel.com>; Chen, Zhaoyan <zhaoyan.chen@intel.com>
> Subject: RE: 22.11.3 patches review and test
> 
> 
> 
> > -----Original Message-----
> > From: Xu, HailinX <hailinx.xu@intel.com>
> > Sent: 8/29/2023 16:22
> > To: Xueming(Steven) Li <xuemingl@nvidia.com>; stable@dpdk.org
> > Cc: xuemingl@nvdia.com; dev@dpdk.org; Abhishek Marathe
> > <Abhishek.Marathe@microsoft.com>; Ali Alnubani <alialnu@nvidia.com>;
> > Walker, Benjamin <benjamin.walker@intel.com>; David Christensen
> > <drc@linux.vnet.ibm.com>; Hemant Agrawal <hemant.agrawal@nxp.com>;
> > Stokes, Ian <ian.stokes@intel.com>; Jerin Jacob <jerinj@marvell.com>;
> > Mcnamara, John <john.mcnamara@intel.com>; Ju-Hyoung Lee
> > <juhlee@microsoft.com>; Kevin Traynor <ktraynor@redhat.com>; Luca
> > Boccassi <bluca@debian.org>; Pei Zhang <pezhang@redhat.com>; Xu, Qian
> > Q <qian.q.xu@intel.com>; Raslan Darawsheh <rasland@nvidia.com>; NBU-
> > Contact-Thomas Monjalon (EXTERNAL) <thomas@monjalon.net>; Yanghang Liu
> > <yanghliu@redhat.com>; Peng, Yuan <yuan.peng@intel.com>; Chen, Zhaoyan
> > <zhaoyan.chen@intel.com>
> > Subject: RE: 22.11.3 patches review and test
> >
> > > -----Original Message-----
> > > From: Xueming Li <xuemingl@nvidia.com>
> > > Sent: Thursday, August 17, 2023 2:14 PM
> > > To: stable@dpdk.org
> > > Cc: xuemingl@nvdia.com; dev@dpdk.org; Abhishek Marathe
> > > <Abhishek.Marathe@microsoft.com>; Ali Alnubani <alialnu@nvidia.com>;
> > > Walker, Benjamin <benjamin.walker@intel.com>; David Christensen
> > > <drc@linux.vnet.ibm.com>; Hemant Agrawal <hemant.agrawal@nxp.com>;
> > > Stokes, Ian <ian.stokes@intel.com>; Jerin Jacob
> > > <jerinj@marvell.com>; Mcnamara, John <john.mcnamara@intel.com>;
> > > Ju-Hyoung Lee <juhlee@microsoft.com>; Kevin Traynor
> > > <ktraynor@redhat.com>; Luca Boccassi <bluca@debian.org>; Pei Zhang
> > > <pezhang@redhat.com>; Xu, Qian Q <qian.q.xu@intel.com>; Raslan
> > > Darawsheh <rasland@nvidia.com>;
> > Thomas
> > > Monjalon <thomas@monjalon.net>; Yanghang Liu <yanghliu@redhat.com>;
> > > Peng, Yuan <yuan.peng@intel.com>; Chen, Zhaoyan
> > > <zhaoyan.chen@intel.com>
> > > Subject: 22.11.3 patches review and test
> > >
> > > Hi all,
> > >
> > > Here is a list of patches targeted for stable release 22.11.3.
> > >
> > > The planned date for the final release is 31th August.
> > >
> > > 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=v22.11.3-rc1
> > >
> > > These patches are located at branch 22.11 of dpdk-stable repo:
> > >     https://dpdk.org/browse/dpdk-stable/
> > >
> > > Thanks.
> >
> > We are conducting DPDK testing and have found two issues.
> >
> > 1. The startup speed of testpmd is significantly slower in the os of SUSE
> >   This issue fix patch has been merged into main, But it has not
> > backported to 22.11.3.
> >   Fix patch commit id on DPDK main:
> > 7e7b6762eac292e78c77ad37ec0973c0c944b845
> 
> Hailin, thanks for the verification and feedback!
> From original patch discussion, the origin decision is no backport, that's why
> "cc:stable@dpdk.org" get removed.
> 
> >
> > 2. The SCTP tunnel packet of iavf cannot be forwarded in avx512 mode
> >   commit 9b7215f150d0bfc5cb00fce68ff08e5217c7f2b3 on v22.11.3-rc1.
> >   This commit is for the new feature (avx512 checksum offload) in DPDK
> > 23.03, which should not be backported to the LTS version since avx512
> > checksum offload is not supported in v22.11.3 LTS.
> 
> Okay, I'll revert this patch. Please note the patch has been backported to 21.11,
> Do we want to fix it? @Kevin Traynor
> 
> >
> > Regards,
> > Xu, Hailin
Update the test status for Intel part. dpdk22.11.3-rc1 all validation tests done. Found 1 crypto bug.

# Basic Intel(R) NIC testing
* Build & CFLAG compile: cover the build test combination with latest GCC/Clang version and the popular OS revision such as
  Ubuntu20.04, Ubuntu22.04, Fedora38, RHEL8.7, RHEL9.2, FreeBSD13.1, SUSE15, CentOS7.9, openEuler22.03-SP1，OpenAnolis8.8 etc.
  - All test done. No new dpdk issue is found.
* PF(i40e, ixgbe): test scenarios including RTE_FLOW/TSO/Jumboframe/checksum offload/VLAN/VXLAN, etc. 
  - All test done. No new dpdk issue is found.
* VF(i40e, ixgbe): test scenarios including VF-RTE_FLOW/TSO/Jumboframe/checksum offload/VLAN/VXLAN, etc.		
  - All test done. No new dpdk issue is found.
* PF/VF(ice): test scenarios including Switch features/Package Management/Flow Director/Advanced Tx/Advanced RSS/ACL/DCF/Flexible Descriptor, etc.
  - All test done. Here have an issue. Need revert this patch.
     The SCTP tunnel packet of iavf cannot be forwarded in avx512 mode.
      commit 9b7215f150d0bfc5cb00fce68ff08e5217c7f2b3 on v22.11.3-rc1.
      This commit is for the new feature (avx512 checksum offload) in DPDK 23.03, which should not be backported to the LTS version since avx512 checksum offload is not supported in v22.11.3 LTS.
* Intel NIC single core/NIC performance: test scenarios including PF/VF single core performance test, etc.
  - All test done. No new dpdk issue is found.
* IPsec: test scenarios including ipsec/ipsec-gw/ipsec library basic test - QAT&SW/FIB library, etc.
  - All test done. No new dpdk issue is found.
		
# Basic cryptodev and virtio testing
* Virtio: both function and performance test are covered. Such as PVP/Virtio_loopback/virtio-user loopback/virtio-net VM2VM perf testing/VMAWARE ESXI 8.0, etc.
  - All test done. No new dpdk issue is found.
* Cryptodev: 
  *Function test: test scenarios including Cryptodev API testing/CompressDev ISA-L/QAT/ZLIB PMD Testing/FIPS, etc.
    - All test done. No new dpdk issue is found.
  *Performance test: test scenarios including Thoughput Performance/Cryptodev Latency, etc.
    - All test done. Found new bug.
        crypto: Performance drop for 1c1t scenario -> Intel dev is under investigating.

Regards,
Xu, Hailin

RE: 22.11.3 patches review and test

[Xueming(Steven) Li]

> -----Original Message-----
> From: Xu, HailinX <hailinx.xu@intel.com>
> Sent: 8/31/2023 9:50
> To: Xueming(Steven) Li <xuemingl@nvidia.com>; stable@dpdk.org
> Cc: xuemingl@nvdia.com; dev@dpdk.org; Abhishek Marathe
> <Abhishek.Marathe@microsoft.com>; Ali Alnubani <alialnu@nvidia.com>;
> Walker, Benjamin <benjamin.walker@intel.com>; David Christensen
> <drc@linux.vnet.ibm.com>; Hemant Agrawal <hemant.agrawal@nxp.com>;
> Stokes, Ian <ian.stokes@intel.com>; Jerin Jacob <jerinj@marvell.com>;
> Mcnamara, John <john.mcnamara@intel.com>; Ju-Hyoung Lee
> <juhlee@microsoft.com>; Luca Boccassi <bluca@debian.org>; Pei Zhang
> <pezhang@redhat.com>; Xu, Qian Q <qian.q.xu@intel.com>; Raslan
> Darawsheh <rasland@nvidia.com>; NBU-Contact-Thomas Monjalon
> (EXTERNAL) <thomas@monjalon.net>; Yanghang Liu <yanghliu@redhat.com>;
> Peng, Yuan <yuan.peng@intel.com>; Chen, Zhaoyan
> <zhaoyan.chen@intel.com>
> Subject: RE: 22.11.3 patches review and test
> 
> > -----Original Message-----
> > From: Xueming(Steven) Li <xuemingl@nvidia.com>
> > Sent: Tuesday, August 29, 2023 4:56 PM
> > To: Xu, HailinX <hailinx.xu@intel.com>; stable@dpdk.org; Kevin Traynor
> > <ktraynor@redhat.com>
> > Cc: xuemingl@nvdia.com; dev@dpdk.org; Abhishek Marathe
> > <Abhishek.Marathe@microsoft.com>; Ali Alnubani <alialnu@nvidia.com>;
> > Walker, Benjamin <benjamin.walker@intel.com>; David Christensen
> > <drc@linux.vnet.ibm.com>; Hemant Agrawal <hemant.agrawal@nxp.com>;
> > Stokes, Ian <ian.stokes@intel.com>; Jerin Jacob <jerinj@marvell.com>;
> > Mcnamara, John <john.mcnamara@intel.com>; Ju-Hyoung Lee
> > <juhlee@microsoft.com>; Luca Boccassi <bluca@debian.org>; Pei Zhang
> > <pezhang@redhat.com>; Xu, Qian Q <qian.q.xu@intel.com>; Raslan
> > Darawsheh <rasland@nvidia.com>; NBU-Contact-Thomas Monjalon
> (EXTERNAL)
> > <thomas@monjalon.net>; Yanghang Liu <yanghliu@redhat.com>; Peng,
> Yuan
> > <yuan.peng@intel.com>; Chen, Zhaoyan <zhaoyan.chen@intel.com>
> > Subject: RE: 22.11.3 patches review and test
> >
> >
> >
> > > -----Original Message-----
> > > From: Xu, HailinX <hailinx.xu@intel.com>
> > > Sent: 8/29/2023 16:22
> > > To: Xueming(Steven) Li <xuemingl@nvidia.com>; stable@dpdk.org
> > > Cc: xuemingl@nvdia.com; dev@dpdk.org; Abhishek Marathe
> > > <Abhishek.Marathe@microsoft.com>; Ali Alnubani <alialnu@nvidia.com>;
> > > Walker, Benjamin <benjamin.walker@intel.com>; David Christensen
> > > <drc@linux.vnet.ibm.com>; Hemant Agrawal
> <hemant.agrawal@nxp.com>;
> > > Stokes, Ian <ian.stokes@intel.com>; Jerin Jacob
> > > <jerinj@marvell.com>; Mcnamara, John <john.mcnamara@intel.com>;
> > > Ju-Hyoung Lee <juhlee@microsoft.com>; Kevin Traynor
> > > <ktraynor@redhat.com>; Luca Boccassi <bluca@debian.org>; Pei Zhang
> > > <pezhang@redhat.com>; Xu, Qian Q <qian.q.xu@intel.com>; Raslan
> > > Darawsheh <rasland@nvidia.com>; NBU- Contact-Thomas Monjalon
> > > (EXTERNAL) <thomas@monjalon.net>; Yanghang Liu
> > > <yanghliu@redhat.com>; Peng, Yuan <yuan.peng@intel.com>; Chen,
> > > Zhaoyan <zhaoyan.chen@intel.com>
> > > Subject: RE: 22.11.3 patches review and test
> > >
> > > > -----Original Message-----
> > > > From: Xueming Li <xuemingl@nvidia.com>
> > > > Sent: Thursday, August 17, 2023 2:14 PM
> > > > To: stable@dpdk.org
> > > > Cc: xuemingl@nvdia.com; dev@dpdk.org; Abhishek Marathe
> > > > <Abhishek.Marathe@microsoft.com>; Ali Alnubani
> > > > <alialnu@nvidia.com>; Walker, Benjamin
> > > > <benjamin.walker@intel.com>; David Christensen
> > > > <drc@linux.vnet.ibm.com>; Hemant Agrawal
> <hemant.agrawal@nxp.com>;
> > > > Stokes, Ian <ian.stokes@intel.com>; Jerin Jacob
> > > > <jerinj@marvell.com>; Mcnamara, John <john.mcnamara@intel.com>;
> > > > Ju-Hyoung Lee <juhlee@microsoft.com>; Kevin Traynor
> > > > <ktraynor@redhat.com>; Luca Boccassi <bluca@debian.org>; Pei Zhang
> > > > <pezhang@redhat.com>; Xu, Qian Q <qian.q.xu@intel.com>; Raslan
> > > > Darawsheh <rasland@nvidia.com>;
> > > Thomas
> > > > Monjalon <thomas@monjalon.net>; Yanghang Liu
> > > > <yanghliu@redhat.com>; Peng, Yuan <yuan.peng@intel.com>; Chen,
> > > > Zhaoyan <zhaoyan.chen@intel.com>
> > > > Subject: 22.11.3 patches review and test
> > > >
> > > > Hi all,
> > > >
> > > > Here is a list of patches targeted for stable release 22.11.3.
> > > >
> > > > The planned date for the final release is 31th August.
> > > >
> > > > 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=v22.11.3-rc1
> > > >
> > > > These patches are located at branch 22.11 of dpdk-stable repo:
> > > >     https://dpdk.org/browse/dpdk-stable/
> > > >
> > > > Thanks.
> > >
> > > We are conducting DPDK testing and have found two issues.
> > >
> > > 1. The startup speed of testpmd is significantly slower in the os of SUSE
> > >   This issue fix patch has been merged into main, But it has not
> > > backported to 22.11.3.
> > >   Fix patch commit id on DPDK main:
> > > 7e7b6762eac292e78c77ad37ec0973c0c944b845
> >
> > Hailin, thanks for the verification and feedback!
> > From original patch discussion, the origin decision is no backport,
> > that's why "cc:stable@dpdk.org" get removed.
> >
> > >
> > > 2. The SCTP tunnel packet of iavf cannot be forwarded in avx512 mode
> > >   commit 9b7215f150d0bfc5cb00fce68ff08e5217c7f2b3 on v22.11.3-rc1.
> > >   This commit is for the new feature (avx512 checksum offload) in
> > > DPDK 23.03, which should not be backported to the LTS version since
> > > avx512 checksum offload is not supported in v22.11.3 LTS.
> >
> > Okay, I'll revert this patch. Please note the patch has been
> > backported to 21.11, Do we want to fix it? @Kevin Traynor
> >
> > >
> > > Regards,
> > > Xu, Hailin
> Update the test status for Intel part. dpdk22.11.3-rc1 all validation tests done.
> Found 1 crypto bug.
> 
> # Basic Intel(R) NIC testing
> * Build & CFLAG compile: cover the build test combination with latest
> GCC/Clang version and the popular OS revision such as
>   Ubuntu20.04, Ubuntu22.04, Fedora38, RHEL8.7, RHEL9.2, FreeBSD13.1,
> SUSE15, CentOS7.9, openEuler22.03-SP1，OpenAnolis8.8 etc.
>   - All test done. No new dpdk issue is found.
> * PF(i40e, ixgbe): test scenarios including
> RTE_FLOW/TSO/Jumboframe/checksum offload/VLAN/VXLAN, etc.
>   - All test done. No new dpdk issue is found.
> * VF(i40e, ixgbe): test scenarios including VF-
> RTE_FLOW/TSO/Jumboframe/checksum offload/VLAN/VXLAN, etc.
> 
>   - All test done. No new dpdk issue is found.
> * PF/VF(ice): test scenarios including Switch features/Package
> Management/Flow Director/Advanced Tx/Advanced RSS/ACL/DCF/Flexible
> Descriptor, etc.
>   - All test done. Here have an issue. Need revert this patch.
>      The SCTP tunnel packet of iavf cannot be forwarded in avx512 mode.
>       commit 9b7215f150d0bfc5cb00fce68ff08e5217c7f2b3 on v22.11.3-rc1.
>       This commit is for the new feature (avx512 checksum offload) in DPDK
> 23.03, which should not be backported to the LTS version since avx512
> checksum offload is not supported in v22.11.3 LTS.
> * Intel NIC single core/NIC performance: test scenarios including PF/VF single
> core performance test, etc.
>   - All test done. No new dpdk issue is found.
> * IPsec: test scenarios including ipsec/ipsec-gw/ipsec library basic test -
> QAT&SW/FIB library, etc.
>   - All test done. No new dpdk issue is found.
> 
> # Basic cryptodev and virtio testing
> * Virtio: both function and performance test are covered. Such as
> PVP/Virtio_loopback/virtio-user loopback/virtio-net VM2VM perf
> testing/VMAWARE ESXI 8.0, etc.
>   - All test done. No new dpdk issue is found.
> * Cryptodev:
>   *Function test: test scenarios including Cryptodev API testing/CompressDev
> ISA-L/QAT/ZLIB PMD Testing/FIPS, etc.
>     - All test done. No new dpdk issue is found.
>   *Performance test: test scenarios including Thoughput
> Performance/Cryptodev Latency, etc.
>     - All test done. Found new bug.
>         crypto: Performance drop for 1c1t scenario -> Intel dev is under
> investigating.

Hi Hailin,

Thanks for the verification and report, I'll hold the release, please kindly let me know if any progress of the crypto bug.

BTW, I reverted "net/iavf: fix tunnel TSO path selection" and pushed to stable/ 22.11-staging, it will
be included in official 22.11.3 TLS if no objection.

> 
> Regards,
> Xu, Hailin

Re: 22.11.3 patches review and test

[Kevin Traynor]

On 30/08/2023 17:25, Kevin Traynor wrote:
> On 29/08/2023 09:22, Xu, HailinX wrote:
>>> -----Original Message-----
>>> From: Xueming Li <xuemingl@nvidia.com>
>>> Sent: Thursday, August 17, 2023 2:14 PM
>>> To: stable@dpdk.org
>>> Cc: xuemingl@nvdia.com; dev@dpdk.org; Abhishek Marathe
>>> <Abhishek.Marathe@microsoft.com>; Ali Alnubani <alialnu@nvidia.com>;
>>> Walker, Benjamin <benjamin.walker@intel.com>; David Christensen
>>> <drc@linux.vnet.ibm.com>; Hemant Agrawal <hemant.agrawal@nxp.com>;
>>> Stokes, Ian <ian.stokes@intel.com>; Jerin Jacob <jerinj@marvell.com>;
>>> Mcnamara, John <john.mcnamara@intel.com>; Ju-Hyoung Lee
>>> <juhlee@microsoft.com>; Kevin Traynor <ktraynor@redhat.com>; Luca Boccassi
>>> <bluca@debian.org>; Pei Zhang <pezhang@redhat.com>; Xu, Qian Q
>>> <qian.q.xu@intel.com>; Raslan Darawsheh <rasland@nvidia.com>; Thomas
>>> Monjalon <thomas@monjalon.net>; Yanghang Liu <yanghliu@redhat.com>;
>>> Peng, Yuan <yuan.peng@intel.com>; Chen, Zhaoyan <zhaoyan.chen@intel.com>
>>> Subject: 22.11.3 patches review and test
>>>
>>> Hi all,
>>>
>>> Here is a list of patches targeted for stable release 22.11.3.
>>>
>>> The planned date for the final release is 31th August.
>>>
>>> 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=v22.11.3-rc1
>>>
>>> These patches are located at branch 22.11 of dpdk-stable repo:
>>>       https://dpdk.org/browse/dpdk-stable/
>>>
>>> Thanks.
>>
>> We are conducting DPDK testing and have found two issues.
>>
>> 1. The startup speed of testpmd is significantly slower in the os of SUSE
>>     This issue fix patch has been merged into main, But it has not backported to 22.11.3.
>>     Fix patch commit id on DPDK main: 7e7b6762eac292e78c77ad37ec0973c0c944b845
>>
>> 2. The SCTP tunnel packet of iavf cannot be forwarded in avx512 mode

Need to clarify this sentence. It looks like it is not a functional bug 
where avx512 mode is selected and then an SCTP tunnel packet cannot be 
sent. Instead, it is a possible performance issue that avx512 mode will 
not be selected where it might have been due to unneeded additions 
(RTE_ETH_TX_OFFLOAD_*_TNL_TSO) to IAVF_TX_NO_VECTOR_FLAGS.

@IAVF maintainers - please confirm my analysis is correct ?

In that case, as it is a possible performance issue in a specific case 
for a single driver I think it is non-critical for 21.11 and we can just 
revert the patch on the branch and wait for 21.11.6 release in December.

thanks,
Kevin.

>>     commit 9b7215f150d0bfc5cb00fce68ff08e5217c7f2b3 on v22.11.3-rc1.
>>     This commit is for the new feature (avx512 checksum offload) in DPDK 23.03, which should not be backported to the LTS version since avx512 checksum offload is not supported in v22.11.3 LTS.
>>
> 
> Thanks for flagging Xueming.
> 
> The issue is that it was listed as fixing 059f18ae2aec ("net/iavf: add
> offload path for Tx AVX512") which goes back to 21.05.
> 
> This could have been avoided if the 'Fixes:' tag was correct, or if the
> authors replied to the email about queued backports :/
> 
> Requesting iavf/next-net-intel maintainers to check Fixes: tags are
> correct before merging.
> 
> DPDK 21.11.5 is already released with this patch. Any idea why it did
> not show up in validation for 21.11.5 ? Is it an issue for 21.11.5 ? How
> critical is it ?
> 
> I can revert it on the 21.11 branch, but it will need to wait until
> 21.11.6 in December before it will be reverted in a released version.
> 
> thanks,
> Kevin.
> 
>> Regards,
>> Xu, Hailin
>>
> 

RE: 22.11.3 patches review and test

[Zeng, ZhichaoX]

> -----Original Message-----
> From: Kevin Traynor <ktraynor@redhat.com>
> Sent: Thursday, August 31, 2023 8:18 PM
> To: Xu, HailinX <hailinx.xu@intel.com>; Xueming Li <xuemingl@nvidia.com>;
> stable@dpdk.org; Wu, Jingjing <jingjing.wu@intel.com>; Xing, Beilei
> <beilei.xing@intel.com>; Xu, Ke1 <ke1.xu@intel.com>; Zeng, ZhichaoX
> <zhichaox.zeng@intel.com>; Zhang, Qi Z <qi.z.zhang@intel.com>
> Cc: xuemingl@nvdia.com; dev@dpdk.org; Stokes, Ian <ian.stokes@intel.com>;
> Mcnamara, John <john.mcnamara@intel.com>; Luca Boccassi
> <bluca@debian.org>; Xu, Qian Q <qian.q.xu@intel.com>; Thomas Monjalon
> <thomas@monjalon.net>; Peng, Yuan <yuan.peng@intel.com>; Chen,
> Zhaoyan <zhaoyan.chen@intel.com>
> Subject: Re: 22.11.3 patches review and test
> 
> On 30/08/2023 17:25, Kevin Traynor wrote:
> > On 29/08/2023 09:22, Xu, HailinX wrote:
> >>> -----Original Message-----
> >>> From: Xueming Li <xuemingl@nvidia.com>
> >>> Sent: Thursday, August 17, 2023 2:14 PM
> >>> To: stable@dpdk.org
> >>> Cc: xuemingl@nvdia.com; dev@dpdk.org; Abhishek Marathe
> >>> <Abhishek.Marathe@microsoft.com>; Ali Alnubani <alialnu@nvidia.com>;
> >>> Walker, Benjamin <benjamin.walker@intel.com>; David Christensen
> >>> <drc@linux.vnet.ibm.com>; Hemant Agrawal
> <hemant.agrawal@nxp.com>;
> >>> Stokes, Ian <ian.stokes@intel.com>; Jerin Jacob
> >>> <jerinj@marvell.com>; Mcnamara, John <john.mcnamara@intel.com>;
> >>> Ju-Hyoung Lee <juhlee@microsoft.com>; Kevin Traynor
> >>> <ktraynor@redhat.com>; Luca Boccassi <bluca@debian.org>; Pei Zhang
> >>> <pezhang@redhat.com>; Xu, Qian Q <qian.q.xu@intel.com>; Raslan
> >>> Darawsheh <rasland@nvidia.com>; Thomas Monjalon
> >>> <thomas@monjalon.net>; Yanghang Liu <yanghliu@redhat.com>; Peng,
> >>> Yuan <yuan.peng@intel.com>; Chen, Zhaoyan
> <zhaoyan.chen@intel.com>
> >>> Subject: 22.11.3 patches review and test
> >>>
> >>> Hi all,
> >>>
> >>> Here is a list of patches targeted for stable release 22.11.3.
> >>>
> >>> The planned date for the final release is 31th August.
> >>>
> >>> 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=v22.11.3-rc1
> >>>
> >>> These patches are located at branch 22.11 of dpdk-stable repo:
> >>>       https://dpdk.org/browse/dpdk-stable/
> >>>
> >>> Thanks.
> >>
> >> We are conducting DPDK testing and have found two issues.
> >>
> >> 1. The startup speed of testpmd is significantly slower in the os of SUSE
> >>     This issue fix patch has been merged into main, But it has not backported
> to 22.11.3.
> >>     Fix patch commit id on DPDK main:
> >> 7e7b6762eac292e78c77ad37ec0973c0c944b845
> >>
> >> 2. The SCTP tunnel packet of iavf cannot be forwarded in avx512 mode
> 
> Need to clarify this sentence. It looks like it is not a functional bug where
> avx512 mode is selected and then an SCTP tunnel packet cannot be sent.
> Instead, it is a possible performance issue that avx512 mode will not be
> selected where it might have been due to unneeded additions
> (RTE_ETH_TX_OFFLOAD_*_TNL_TSO) to IAVF_TX_NO_VECTOR_FLAGS.
> 
> @IAVF maintainers - please confirm my analysis is correct ?
> 
> In that case, as it is a possible performance issue in a specific case for a single
> driver I think it is non-critical for 21.11 and we can just revert the patch on the
> branch and wait for 21.11.6 release in December.

Hi Kevin,

Since the LTS version of the IAVF driver does not support avx512 checksum offload,
the scalar path should be selected, but this patch makes it incorrectly select the
avx512 path, and the SCTP tunnel packets can't be forwarded properly.

Yes, we can revert this commit for 21.11.6 release, thanks.

Regards
Zhichao

> thanks,
> Kevin.
> 
> >>     commit 9b7215f150d0bfc5cb00fce68ff08e5217c7f2b3 on v22.11.3-
> rc1.
> >>     This commit is for the new feature (avx512 checksum offload) in DPDK
> 23.03, which should not be backported to the LTS version since avx512
> checksum offload is not supported in v22.11.3 LTS.
> >>
> >
> > Thanks for flagging Xueming.
> >
> > The issue is that it was listed as fixing 059f18ae2aec ("net/iavf: add
> > offload path for Tx AVX512") which goes back to 21.05.
> >
> > This could have been avoided if the 'Fixes:' tag was correct, or if
> > the authors replied to the email about queued backports :/
> >
> > Requesting iavf/next-net-intel maintainers to check Fixes: tags are
> > correct before merging.
> >
> > DPDK 21.11.5 is already released with this patch. Any idea why it did
> > not show up in validation for 21.11.5 ? Is it an issue for 21.11.5 ?
> > How critical is it ?
> >
> > I can revert it on the 21.11 branch, but it will need to wait until
> > 21.11.6 in December before it will be reverted in a released version.
> >
> > thanks,
> > Kevin.
> >
> >> Regards,
> >> Xu, Hailin
> >>
> >

Re: 22.11.3 patches review and test

[Kevin Traynor]

On 01/09/2023 04:23, Zeng, ZhichaoX wrote:
>> -----Original Message-----
>> From: Kevin Traynor <ktraynor@redhat.com>
>> Sent: Thursday, August 31, 2023 8:18 PM
>> To: Xu, HailinX <hailinx.xu@intel.com>; Xueming Li <xuemingl@nvidia.com>;
>> stable@dpdk.org; Wu, Jingjing <jingjing.wu@intel.com>; Xing, Beilei
>> <beilei.xing@intel.com>; Xu, Ke1 <ke1.xu@intel.com>; Zeng, ZhichaoX
>> <zhichaox.zeng@intel.com>; Zhang, Qi Z <qi.z.zhang@intel.com>
>> Cc: xuemingl@nvdia.com; dev@dpdk.org; Stokes, Ian <ian.stokes@intel.com>;
>> Mcnamara, John <john.mcnamara@intel.com>; Luca Boccassi
>> <bluca@debian.org>; Xu, Qian Q <qian.q.xu@intel.com>; Thomas Monjalon
>> <thomas@monjalon.net>; Peng, Yuan <yuan.peng@intel.com>; Chen,
>> Zhaoyan <zhaoyan.chen@intel.com>
>> Subject: Re: 22.11.3 patches review and test
>>
>> On 30/08/2023 17:25, Kevin Traynor wrote:
>>> On 29/08/2023 09:22, Xu, HailinX wrote:
>>>>> -----Original Message-----
>>>>> From: Xueming Li <xuemingl@nvidia.com>
>>>>> Sent: Thursday, August 17, 2023 2:14 PM
>>>>> To: stable@dpdk.org
>>>>> Cc: xuemingl@nvdia.com; dev@dpdk.org; Abhishek Marathe
>>>>> <Abhishek.Marathe@microsoft.com>; Ali Alnubani <alialnu@nvidia.com>;
>>>>> Walker, Benjamin <benjamin.walker@intel.com>; David Christensen
>>>>> <drc@linux.vnet.ibm.com>; Hemant Agrawal
>> <hemant.agrawal@nxp.com>;
>>>>> Stokes, Ian <ian.stokes@intel.com>; Jerin Jacob
>>>>> <jerinj@marvell.com>; Mcnamara, John <john.mcnamara@intel.com>;
>>>>> Ju-Hyoung Lee <juhlee@microsoft.com>; Kevin Traynor
>>>>> <ktraynor@redhat.com>; Luca Boccassi <bluca@debian.org>; Pei Zhang
>>>>> <pezhang@redhat.com>; Xu, Qian Q <qian.q.xu@intel.com>; Raslan
>>>>> Darawsheh <rasland@nvidia.com>; Thomas Monjalon
>>>>> <thomas@monjalon.net>; Yanghang Liu <yanghliu@redhat.com>; Peng,
>>>>> Yuan <yuan.peng@intel.com>; Chen, Zhaoyan
>> <zhaoyan.chen@intel.com>
>>>>> Subject: 22.11.3 patches review and test
>>>>>
>>>>> Hi all,
>>>>>
>>>>> Here is a list of patches targeted for stable release 22.11.3.
>>>>>
>>>>> The planned date for the final release is 31th August.
>>>>>
>>>>> 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=v22.11.3-rc1
>>>>>
>>>>> These patches are located at branch 22.11 of dpdk-stable repo:
>>>>>        https://dpdk.org/browse/dpdk-stable/
>>>>>
>>>>> Thanks.
>>>>
>>>> We are conducting DPDK testing and have found two issues.
>>>>
>>>> 1. The startup speed of testpmd is significantly slower in the os of SUSE
>>>>      This issue fix patch has been merged into main, But it has not backported
>> to 22.11.3.
>>>>      Fix patch commit id on DPDK main:
>>>> 7e7b6762eac292e78c77ad37ec0973c0c944b845
>>>>
>>>> 2. The SCTP tunnel packet of iavf cannot be forwarded in avx512 mode
>>
>> Need to clarify this sentence. It looks like it is not a functional bug where
>> avx512 mode is selected and then an SCTP tunnel packet cannot be sent.
>> Instead, it is a possible performance issue that avx512 mode will not be
>> selected where it might have been due to unneeded additions
>> (RTE_ETH_TX_OFFLOAD_*_TNL_TSO) to IAVF_TX_NO_VECTOR_FLAGS.
>>
>> @IAVF maintainers - please confirm my analysis is correct ?
>>
>> In that case, as it is a possible performance issue in a specific case for a single
>> driver I think it is non-critical for 21.11 and we can just revert the patch on the
>> branch and wait for 21.11.6 release in December.
> 
> Hi Kevin,
> 
> Since the LTS version of the IAVF driver does not support avx512 checksum offload,
> the scalar path should be selected, but this patch makes it incorrectly select the
> avx512 path, and the SCTP tunnel packets can't be forwarded properly.
> 

ok, let's take a look at the patch and usage.

diff --git a/drivers/net/iavf/iavf_rxtx.h b/drivers/net/iavf/iavf_rxtx.h
index 8d4a77271a..605ea3f824 100644
--- a/drivers/net/iavf/iavf_rxtx.h
+++ b/drivers/net/iavf/iavf_rxtx.h
@@ -32,4 +32,8 @@
                 RTE_ETH_TX_OFFLOAD_MULTI_SEGS |          \
                 RTE_ETH_TX_OFFLOAD_TCP_TSO |             \
+               RTE_ETH_TX_OFFLOAD_VXLAN_TNL_TSO |       \
+               RTE_ETH_TX_OFFLOAD_GRE_TNL_TSO |         \
+               RTE_ETH_TX_OFFLOAD_IPIP_TNL_TSO |        \
+               RTE_ETH_TX_OFFLOAD_GENEVE_TNL_TSO |      \
                 RTE_ETH_TX_OFFLOAD_SECURITY)

<snip>

So we now have:
#define IAVF_TX_NO_VECTOR_FLAGS (				 \
		RTE_ETH_TX_OFFLOAD_VLAN_INSERT |		 \
		RTE_ETH_TX_OFFLOAD_QINQ_INSERT |		 \
		RTE_ETH_TX_OFFLOAD_MULTI_SEGS |		 \
		RTE_ETH_TX_OFFLOAD_TCP_TSO |		 \
		RTE_ETH_TX_OFFLOAD_VXLAN_TNL_TSO |	 \
		RTE_ETH_TX_OFFLOAD_GRE_TNL_TSO |	 \
		RTE_ETH_TX_OFFLOAD_IPIP_TNL_TSO |	 \
		RTE_ETH_TX_OFFLOAD_GENEVE_TNL_TSO |	 \
		RTE_ETH_TX_OFFLOAD_SECURITY)

<snip>

static inline int
iavf_tx_vec_queue_default(struct iavf_tx_queue *txq)
{
	if (!txq)
		return -1;

	if (txq->rs_thresh < IAVF_VPMD_TX_MAX_BURST ||
	    txq->rs_thresh > IAVF_VPMD_TX_MAX_FREE_BUF)
		return -1;

	if (txq->offloads & IAVF_TX_NO_VECTOR_FLAGS)
		return -1;
             ^^^ Adding the extra bits to IAVF_TX_NO_VECTOR_FLAGS gives 
*more* reasons for why this statement might not be true, so returning -1 
indicating that vector cannot be used for tx queue


<snip>

static inline bool
check_tx_vec_allow(struct iavf_tx_queue *txq)
{
	if (!(txq->offloads & IAVF_TX_NO_VECTOR_FLAGS) &&

             ^^^ Adding the extra bits to IAVF_TX_NO_VECTOR_FLAGS gives 
*more* reason for this statement will be false and then false returned 
indicating that vector cannot be used.

	    txq->rs_thresh >= IAVF_VPMD_TX_MAX_BURST &&
	    txq->rs_thresh <= IAVF_VPMD_TX_MAX_FREE_BUF) {
		PMD_INIT_LOG(DEBUG, "Vector tx can be enabled on this txq.");
		return true;
	}
	PMD_INIT_LOG(DEBUG, "Vector Tx cannot be enabled on this txq.");
	return false;
}

--

It looks like that adding the extra bits gives it less reasons to select 
vector mode. However, you are saying that this patch means there is a 
case where it now selects vector where it should not, meaning additional 
reason to select vector mode. So maybe I miss something ?

> Yes, we can revert this commit for 21.11.6 release, thanks.
> 
> Regards
> Zhichao
> 
>> thanks,
>> Kevin.
>>
>>>>      commit 9b7215f150d0bfc5cb00fce68ff08e5217c7f2b3 on v22.11.3-
>> rc1.
>>>>      This commit is for the new feature (avx512 checksum offload) in DPDK
>> 23.03, which should not be backported to the LTS version since avx512
>> checksum offload is not supported in v22.11.3 LTS.
>>>>
>>>
>>> Thanks for flagging Xueming.
>>>
>>> The issue is that it was listed as fixing 059f18ae2aec ("net/iavf: add
>>> offload path for Tx AVX512") which goes back to 21.05.
>>>
>>> This could have been avoided if the 'Fixes:' tag was correct, or if
>>> the authors replied to the email about queued backports :/
>>>
>>> Requesting iavf/next-net-intel maintainers to check Fixes: tags are
>>> correct before merging.
>>>
>>> DPDK 21.11.5 is already released with this patch. Any idea why it did
>>> not show up in validation for 21.11.5 ? Is it an issue for 21.11.5 ?
>>> How critical is it ?
>>>
>>> I can revert it on the 21.11 branch, but it will need to wait until
>>> 21.11.6 in December before it will be reverted in a released version.
>>>
>>> thanks,
>>> Kevin.
>>>
>>>> Regards,
>>>> Xu, Hailin
>>>>
>>>
> 

Re: 22.11.3 patches review and test

[Kevin Traynor]

On 04/09/2023 10:32, Kevin Traynor wrote:
> On 01/09/2023 04:23, Zeng, ZhichaoX wrote:
>>> -----Original Message-----
>>> From: Kevin Traynor <ktraynor@redhat.com>
>>> Sent: Thursday, August 31, 2023 8:18 PM
>>> To: Xu, HailinX <hailinx.xu@intel.com>; Xueming Li <xuemingl@nvidia.com>;
>>> stable@dpdk.org; Wu, Jingjing <jingjing.wu@intel.com>; Xing, Beilei
>>> <beilei.xing@intel.com>; Xu, Ke1 <ke1.xu@intel.com>; Zeng, ZhichaoX
>>> <zhichaox.zeng@intel.com>; Zhang, Qi Z <qi.z.zhang@intel.com>
>>> Cc: xuemingl@nvdia.com; dev@dpdk.org; Stokes, Ian <ian.stokes@intel.com>;
>>> Mcnamara, John <john.mcnamara@intel.com>; Luca Boccassi
>>> <bluca@debian.org>; Xu, Qian Q <qian.q.xu@intel.com>; Thomas Monjalon
>>> <thomas@monjalon.net>; Peng, Yuan <yuan.peng@intel.com>; Chen,
>>> Zhaoyan <zhaoyan.chen@intel.com>
>>> Subject: Re: 22.11.3 patches review and test
>>>
>>> On 30/08/2023 17:25, Kevin Traynor wrote:
>>>> On 29/08/2023 09:22, Xu, HailinX wrote:
>>>>>> -----Original Message-----
>>>>>> From: Xueming Li <xuemingl@nvidia.com>
>>>>>> Sent: Thursday, August 17, 2023 2:14 PM
>>>>>> To: stable@dpdk.org
>>>>>> Cc: xuemingl@nvdia.com; dev@dpdk.org; Abhishek Marathe
>>>>>> <Abhishek.Marathe@microsoft.com>; Ali Alnubani <alialnu@nvidia.com>;
>>>>>> Walker, Benjamin <benjamin.walker@intel.com>; David Christensen
>>>>>> <drc@linux.vnet.ibm.com>; Hemant Agrawal
>>> <hemant.agrawal@nxp.com>;
>>>>>> Stokes, Ian <ian.stokes@intel.com>; Jerin Jacob
>>>>>> <jerinj@marvell.com>; Mcnamara, John <john.mcnamara@intel.com>;
>>>>>> Ju-Hyoung Lee <juhlee@microsoft.com>; Kevin Traynor
>>>>>> <ktraynor@redhat.com>; Luca Boccassi <bluca@debian.org>; Pei Zhang
>>>>>> <pezhang@redhat.com>; Xu, Qian Q <qian.q.xu@intel.com>; Raslan
>>>>>> Darawsheh <rasland@nvidia.com>; Thomas Monjalon
>>>>>> <thomas@monjalon.net>; Yanghang Liu <yanghliu@redhat.com>; Peng,
>>>>>> Yuan <yuan.peng@intel.com>; Chen, Zhaoyan
>>> <zhaoyan.chen@intel.com>
>>>>>> Subject: 22.11.3 patches review and test
>>>>>>
>>>>>> Hi all,
>>>>>>
>>>>>> Here is a list of patches targeted for stable release 22.11.3.
>>>>>>
>>>>>> The planned date for the final release is 31th August.
>>>>>>
>>>>>> 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=v22.11.3-rc1
>>>>>>
>>>>>> These patches are located at branch 22.11 of dpdk-stable repo:
>>>>>>         https://dpdk.org/browse/dpdk-stable/
>>>>>>
>>>>>> Thanks.
>>>>>
>>>>> We are conducting DPDK testing and have found two issues.
>>>>>
>>>>> 1. The startup speed of testpmd is significantly slower in the os of SUSE
>>>>>       This issue fix patch has been merged into main, But it has not backported
>>> to 22.11.3.
>>>>>       Fix patch commit id on DPDK main:
>>>>> 7e7b6762eac292e78c77ad37ec0973c0c944b845
>>>>>
>>>>> 2. The SCTP tunnel packet of iavf cannot be forwarded in avx512 mode
>>>
>>> Need to clarify this sentence. It looks like it is not a functional bug where
>>> avx512 mode is selected and then an SCTP tunnel packet cannot be sent.
>>> Instead, it is a possible performance issue that avx512 mode will not be
>>> selected where it might have been due to unneeded additions
>>> (RTE_ETH_TX_OFFLOAD_*_TNL_TSO) to IAVF_TX_NO_VECTOR_FLAGS.
>>>
>>> @IAVF maintainers - please confirm my analysis is correct ?
>>>
>>> In that case, as it is a possible performance issue in a specific case for a single
>>> driver I think it is non-critical for 21.11 and we can just revert the patch on the
>>> branch and wait for 21.11.6 release in December.
>>
>> Hi Kevin,
>>
>> Since the LTS version of the IAVF driver does not support avx512 checksum offload,
>> the scalar path should be selected, but this patch makes it incorrectly select the
>> avx512 path, and the SCTP tunnel packets can't be forwarded properly.
>>
> 
> ok, let's take a look at the patch and usage.
> 
> diff --git a/drivers/net/iavf/iavf_rxtx.h b/drivers/net/iavf/iavf_rxtx.h
> index 8d4a77271a..605ea3f824 100644
> --- a/drivers/net/iavf/iavf_rxtx.h
> +++ b/drivers/net/iavf/iavf_rxtx.h
> @@ -32,4 +32,8 @@
>                   RTE_ETH_TX_OFFLOAD_MULTI_SEGS |          \
>                   RTE_ETH_TX_OFFLOAD_TCP_TSO |             \
> +               RTE_ETH_TX_OFFLOAD_VXLAN_TNL_TSO |       \
> +               RTE_ETH_TX_OFFLOAD_GRE_TNL_TSO |         \
> +               RTE_ETH_TX_OFFLOAD_IPIP_TNL_TSO |        \
> +               RTE_ETH_TX_OFFLOAD_GENEVE_TNL_TSO |      \
>                   RTE_ETH_TX_OFFLOAD_SECURITY)
> 
> <snip>
> 
> So we now have:
> #define IAVF_TX_NO_VECTOR_FLAGS (				 \
> 		RTE_ETH_TX_OFFLOAD_VLAN_INSERT |		 \
> 		RTE_ETH_TX_OFFLOAD_QINQ_INSERT |		 \
> 		RTE_ETH_TX_OFFLOAD_MULTI_SEGS |		 \
> 		RTE_ETH_TX_OFFLOAD_TCP_TSO |		 \
> 		RTE_ETH_TX_OFFLOAD_VXLAN_TNL_TSO |	 \
> 		RTE_ETH_TX_OFFLOAD_GRE_TNL_TSO |	 \
> 		RTE_ETH_TX_OFFLOAD_IPIP_TNL_TSO |	 \
> 		RTE_ETH_TX_OFFLOAD_GENEVE_TNL_TSO |	 \
> 		RTE_ETH_TX_OFFLOAD_SECURITY)
> 
> <snip>
> 
> static inline int
> iavf_tx_vec_queue_default(struct iavf_tx_queue *txq)
> {
> 	if (!txq)
> 		return -1;
> 
> 	if (txq->rs_thresh < IAVF_VPMD_TX_MAX_BURST ||
> 	    txq->rs_thresh > IAVF_VPMD_TX_MAX_FREE_BUF)
> 		return -1;
> 
> 	if (txq->offloads & IAVF_TX_NO_VECTOR_FLAGS)
> 		return -1;
>               ^^^ Adding the extra bits to IAVF_TX_NO_VECTOR_FLAGS gives
> *more* reasons for why this statement might not be true, so returning -1
> indicating that vector cannot be used for tx queue
> 

typo here - just to clarify, the new flags give more reasons for this 
statement to be true, so returning -1.

> 
> <snip>
> 
> static inline bool
> check_tx_vec_allow(struct iavf_tx_queue *txq)
> {
> 	if (!(txq->offloads & IAVF_TX_NO_VECTOR_FLAGS) &&
> 
>               ^^^ Adding the extra bits to IAVF_TX_NO_VECTOR_FLAGS gives
> *more* reason for this statement will be false and then false returned
> indicating that vector cannot be used.
> 
> 	    txq->rs_thresh >= IAVF_VPMD_TX_MAX_BURST &&
> 	    txq->rs_thresh <= IAVF_VPMD_TX_MAX_FREE_BUF) {
> 		PMD_INIT_LOG(DEBUG, "Vector tx can be enabled on this txq.");
> 		return true;
> 	}
> 	PMD_INIT_LOG(DEBUG, "Vector Tx cannot be enabled on this txq.");
> 	return false;
> }
> 
> --
> 
> It looks like that adding the extra bits gives it less reasons to select
> vector mode. However, you are saying that this patch means there is a
> case where it now selects vector where it should not, meaning additional
> reason to select vector mode. So maybe I miss something ?
> 
>> Yes, we can revert this commit for 21.11.6 release, thanks.
>>
>> Regards
>> Zhichao
>>
>>> thanks,
>>> Kevin.
>>>
>>>>>       commit 9b7215f150d0bfc5cb00fce68ff08e5217c7f2b3 on v22.11.3-
>>> rc1.
>>>>>       This commit is for the new feature (avx512 checksum offload) in DPDK
>>> 23.03, which should not be backported to the LTS version since avx512
>>> checksum offload is not supported in v22.11.3 LTS.
>>>>>
>>>>
>>>> Thanks for flagging Xueming.
>>>>
>>>> The issue is that it was listed as fixing 059f18ae2aec ("net/iavf: add
>>>> offload path for Tx AVX512") which goes back to 21.05.
>>>>
>>>> This could have been avoided if the 'Fixes:' tag was correct, or if
>>>> the authors replied to the email about queued backports :/
>>>>
>>>> Requesting iavf/next-net-intel maintainers to check Fixes: tags are
>>>> correct before merging.
>>>>
>>>> DPDK 21.11.5 is already released with this patch. Any idea why it did
>>>> not show up in validation for 21.11.5 ? Is it an issue for 21.11.5 ?
>>>> How critical is it ?
>>>>
>>>> I can revert it on the 21.11 branch, but it will need to wait until
>>>> 21.11.6 in December before it will be reverted in a released version.
>>>>
>>>> thanks,
>>>> Kevin.
>>>>
>>>>> Regards,
>>>>> Xu, Hailin
>>>>>
>>>>
>>
> 

RE: 22.11.3 patches review and test

[Zeng, ZhichaoX]

Regards
Zhichao
> -----Original Message-----
> From: Kevin Traynor <ktraynor@redhat.com>
> Sent: Monday, September 4, 2023 10:15 PM
> To: Zeng, ZhichaoX <zhichaox.zeng@intel.com>; Xu, HailinX
> <hailinx.xu@intel.com>; Xueming Li <xuemingl@nvidia.com>;
> stable@dpdk.org; Wu, Jingjing <jingjing.wu@intel.com>; Xing, Beilei
> <beilei.xing@intel.com>; Xu, Ke1 <ke1.xu@intel.com>; Zhang, Qi Z
> <qi.z.zhang@intel.com>
> Cc: xuemingl@nvdia.com; dev@dpdk.org; Stokes, Ian <ian.stokes@intel.com>;
> Mcnamara, John <john.mcnamara@intel.com>; Luca Boccassi
> <bluca@debian.org>; Xu, Qian Q <qian.q.xu@intel.com>; Thomas Monjalon
> <thomas@monjalon.net>; Peng, Yuan <yuan.peng@intel.com>; Chen,
> Zhaoyan <zhaoyan.chen@intel.com>
> Subject: Re: 22.11.3 patches review and test
> 
> On 04/09/2023 10:32, Kevin Traynor wrote:
> > On 01/09/2023 04:23, Zeng, ZhichaoX wrote:
> >>> -----Original Message-----
> >>> From: Kevin Traynor <ktraynor@redhat.com>
> >>> Sent: Thursday, August 31, 2023 8:18 PM
> >>> To: Xu, HailinX <hailinx.xu@intel.com>; Xueming Li
> >>> <xuemingl@nvidia.com>; stable@dpdk.org; Wu, Jingjing
> >>> <jingjing.wu@intel.com>; Xing, Beilei <beilei.xing@intel.com>; Xu,
> >>> Ke1 <ke1.xu@intel.com>; Zeng, ZhichaoX <zhichaox.zeng@intel.com>;
> >>> Zhang, Qi Z <qi.z.zhang@intel.com>
> >>> Cc: xuemingl@nvdia.com; dev@dpdk.org; Stokes, Ian
> >>> <ian.stokes@intel.com>; Mcnamara, John <john.mcnamara@intel.com>;
> >>> Luca Boccassi <bluca@debian.org>; Xu, Qian Q <qian.q.xu@intel.com>;
> >>> Thomas Monjalon <thomas@monjalon.net>; Peng, Yuan
> >>> <yuan.peng@intel.com>; Chen, Zhaoyan <zhaoyan.chen@intel.com>
> >>> Subject: Re: 22.11.3 patches review and test
> >>>
> >>> On 30/08/2023 17:25, Kevin Traynor wrote:
> >>>> On 29/08/2023 09:22, Xu, HailinX wrote:
> >>>>>> -----Original Message-----
> >>>>>> From: Xueming Li <xuemingl@nvidia.com>
> >>>>>> Sent: Thursday, August 17, 2023 2:14 PM
> >>>>>> To: stable@dpdk.org
> >>>>>> Cc: xuemingl@nvdia.com; dev@dpdk.org; Abhishek Marathe
> >>>>>> <Abhishek.Marathe@microsoft.com>; Ali Alnubani
> >>>>>> <alialnu@nvidia.com>; Walker, Benjamin
> >>>>>> <benjamin.walker@intel.com>; David Christensen
> >>>>>> <drc@linux.vnet.ibm.com>; Hemant Agrawal
> >>> <hemant.agrawal@nxp.com>;
> >>>>>> Stokes, Ian <ian.stokes@intel.com>; Jerin Jacob
> >>>>>> <jerinj@marvell.com>; Mcnamara, John
> <john.mcnamara@intel.com>;
> >>>>>> Ju-Hyoung Lee <juhlee@microsoft.com>; Kevin Traynor
> >>>>>> <ktraynor@redhat.com>; Luca Boccassi <bluca@debian.org>; Pei
> >>>>>> Zhang <pezhang@redhat.com>; Xu, Qian Q <qian.q.xu@intel.com>;
> >>>>>> Raslan Darawsheh <rasland@nvidia.com>; Thomas Monjalon
> >>>>>> <thomas@monjalon.net>; Yanghang Liu <yanghliu@redhat.com>;
> Peng,
> >>>>>> Yuan <yuan.peng@intel.com>; Chen, Zhaoyan
> >>> <zhaoyan.chen@intel.com>
> >>>>>> Subject: 22.11.3 patches review and test
> >>>>>>
> >>>>>> Hi all,
> >>>>>>
> >>>>>> Here is a list of patches targeted for stable release 22.11.3.
> >>>>>>
> >>>>>> The planned date for the final release is 31th August.
> >>>>>>
> >>>>>> 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=v22.11.3-rc1
> >>>>>>
> >>>>>> These patches are located at branch 22.11 of dpdk-stable repo:
> >>>>>>         https://dpdk.org/browse/dpdk-stable/
> >>>>>>
> >>>>>> Thanks.
> >>>>>
> >>>>> We are conducting DPDK testing and have found two issues.
> >>>>>
> >>>>> 1. The startup speed of testpmd is significantly slower in the os of SUSE
> >>>>>       This issue fix patch has been merged into main, But it has
> >>>>> not backported
> >>> to 22.11.3.
> >>>>>       Fix patch commit id on DPDK main:
> >>>>> 7e7b6762eac292e78c77ad37ec0973c0c944b845
> >>>>>
> >>>>> 2. The SCTP tunnel packet of iavf cannot be forwarded in avx512
> >>>>> mode
> >>>
> >>> Need to clarify this sentence. It looks like it is not a functional
> >>> bug where
> >>> avx512 mode is selected and then an SCTP tunnel packet cannot be sent.
> >>> Instead, it is a possible performance issue that avx512 mode will
> >>> not be selected where it might have been due to unneeded additions
> >>> (RTE_ETH_TX_OFFLOAD_*_TNL_TSO) to IAVF_TX_NO_VECTOR_FLAGS.
> >>>
> >>> @IAVF maintainers - please confirm my analysis is correct ?
> >>>
> >>> In that case, as it is a possible performance issue in a specific
> >>> case for a single driver I think it is non-critical for 21.11 and we
> >>> can just revert the patch on the branch and wait for 21.11.6 release in
> December.
> >>
> >> Hi Kevin,
> >>
> >> Since the LTS version of the IAVF driver does not support avx512
> >> checksum offload, the scalar path should be selected, but this patch
> >> makes it incorrectly select the
> >> avx512 path, and the SCTP tunnel packets can't be forwarded properly.
> >>
> >
> > ok, let's take a look at the patch and usage.
> >
> > diff --git a/drivers/net/iavf/iavf_rxtx.h
> > b/drivers/net/iavf/iavf_rxtx.h index 8d4a77271a..605ea3f824 100644
> > --- a/drivers/net/iavf/iavf_rxtx.h
> > +++ b/drivers/net/iavf/iavf_rxtx.h
> > @@ -32,4 +32,8 @@
> >                   RTE_ETH_TX_OFFLOAD_MULTI_SEGS |          \
> >                   RTE_ETH_TX_OFFLOAD_TCP_TSO |             \
> > +               RTE_ETH_TX_OFFLOAD_VXLAN_TNL_TSO |       \
> > +               RTE_ETH_TX_OFFLOAD_GRE_TNL_TSO |         \
> > +               RTE_ETH_TX_OFFLOAD_IPIP_TNL_TSO |        \
> > +               RTE_ETH_TX_OFFLOAD_GENEVE_TNL_TSO |      \
> >                   RTE_ETH_TX_OFFLOAD_SECURITY)
> >
> > <snip>
> >
> > So we now have:
> > #define IAVF_TX_NO_VECTOR_FLAGS (				 \
> > 		RTE_ETH_TX_OFFLOAD_VLAN_INSERT |		 \
> > 		RTE_ETH_TX_OFFLOAD_QINQ_INSERT |		 \
> > 		RTE_ETH_TX_OFFLOAD_MULTI_SEGS |		 \
> > 		RTE_ETH_TX_OFFLOAD_TCP_TSO |		 \
> > 		RTE_ETH_TX_OFFLOAD_VXLAN_TNL_TSO |	 \
> > 		RTE_ETH_TX_OFFLOAD_GRE_TNL_TSO |	 \
> > 		RTE_ETH_TX_OFFLOAD_IPIP_TNL_TSO |	 \
> > 		RTE_ETH_TX_OFFLOAD_GENEVE_TNL_TSO |	 \
> > 		RTE_ETH_TX_OFFLOAD_SECURITY)
> >
> > <snip>
> >

Hi Kevin,

This patch also removes two flags from IAVF_TX_NO_VECTOR_FLAGS, RTE_ETH_TX_OFFLOAD_OUTER_IPV4_CKSUM
and RTE_ETH_TX_OFFLOAD_OUTER_UDP_CKSUM.

> > static inline int
> > iavf_tx_vec_queue_default(struct iavf_tx_queue *txq) {
> > 	if (!txq)
> > 		return -1;
> >
> > 	if (txq->rs_thresh < IAVF_VPMD_TX_MAX_BURST ||
> > 	    txq->rs_thresh > IAVF_VPMD_TX_MAX_FREE_BUF)
> > 		return -1;
> >
> > 	if (txq->offloads & IAVF_TX_NO_VECTOR_FLAGS)
> > 		return -1;
> >               ^^^ Adding the extra bits to IAVF_TX_NO_VECTOR_FLAGS
> > gives
> > *more* reasons for why this statement might not be true, so returning
> > -1 indicating that vector cannot be used for tx queue
> >
> 
> typo here - just to clarify, the new flags give more reasons for this statement
> to be true, so returning -1.
> 
> >
> > <snip>
> >
> > static inline bool
> > check_tx_vec_allow(struct iavf_tx_queue *txq) {
> > 	if (!(txq->offloads & IAVF_TX_NO_VECTOR_FLAGS) &&
> >
> >               ^^^ Adding the extra bits to IAVF_TX_NO_VECTOR_FLAGS
> > gives
> > *more* reason for this statement will be false and then false returned
> > indicating that vector cannot be used.
> >
> > 	    txq->rs_thresh >= IAVF_VPMD_TX_MAX_BURST &&
> > 	    txq->rs_thresh <= IAVF_VPMD_TX_MAX_FREE_BUF) {
> > 		PMD_INIT_LOG(DEBUG, "Vector tx can be enabled on this
> txq.");
> > 		return true;
> > 	}
> > 	PMD_INIT_LOG(DEBUG, "Vector Tx cannot be enabled on this txq.");
> > 	return false;
> > }
> >
> > --
> >
> > It looks like that adding the extra bits gives it less reasons to
> > select vector mode. However, you are saying that this patch means
> > there is a case where it now selects vector where it should not,
> > meaning additional reason to select vector mode. So maybe I miss
> something ?

Originally, the vector path would not be selected after configuring outer checksum offload,
but it will be selected after removing the two flags.
But the vector path doesn't support outer checksum offload on stable DPDK, so there will be a problem.

The key issue is that these two flags are removed, RTE_ETH_TX_OFFLOAD_OUTER_IPV4_CKSUM
and RTE_ETH_TX_OFFLOAD_OUTER_UDP_CKSUM.

Regards
Zhichao

> >
> >> Yes, we can revert this commit for 21.11.6 release, thanks.
> >>
> >> Regards
> >> Zhichao
> >>
> >>> thanks,
> >>> Kevin.
> >>>
> >>>>>       commit 9b7215f150d0bfc5cb00fce68ff08e5217c7f2b3 on
> v22.11.3-
> >>> rc1.
> >>>>>       This commit is for the new feature (avx512 checksum offload)
> >>>>> in DPDK
> >>> 23.03, which should not be backported to the LTS version since
> >>> avx512 checksum offload is not supported in v22.11.3 LTS.
> >>>>>
> >>>>
> >>>> Thanks for flagging Xueming.
> >>>>
> >>>> The issue is that it was listed as fixing 059f18ae2aec ("net/iavf:
> >>>> add offload path for Tx AVX512") which goes back to 21.05.
> >>>>
> >>>> This could have been avoided if the 'Fixes:' tag was correct, or if
> >>>> the authors replied to the email about queued backports :/
> >>>>
> >>>> Requesting iavf/next-net-intel maintainers to check Fixes: tags are
> >>>> correct before merging.
> >>>>
> >>>> DPDK 21.11.5 is already released with this patch. Any idea why it
> >>>> did not show up in validation for 21.11.5 ? Is it an issue for 21.11.5 ?
> >>>> How critical is it ?
> >>>>
> >>>> I can revert it on the 21.11 branch, but it will need to wait until
> >>>> 21.11.6 in December before it will be reverted in a released version.
> >>>>
> >>>> thanks,
> >>>> Kevin.
> >>>>
> >>>>> Regards,
> >>>>> Xu, Hailin
> >>>>>
> >>>>
> >>
> >

RE: 22.11.3 patches review and test

[Zeng, ZhichaoX]

> -----Original Message-----
> From: Kevin Traynor <ktraynor@redhat.com>
> Sent: Monday, September 4, 2023 10:15 PM
> To: Zeng, ZhichaoX <zhichaox.zeng@intel.com>; Xu, HailinX
> <hailinx.xu@intel.com>; Xueming Li <xuemingl@nvidia.com>;
> stable@dpdk.org; Wu, Jingjing <jingjing.wu@intel.com>; Xing, Beilei
> <beilei.xing@intel.com>; Xu, Ke1 <ke1.xu@intel.com>; Zhang, Qi Z
> <qi.z.zhang@intel.com>
> Cc: xuemingl@nvdia.com; dev@dpdk.org; Stokes, Ian <ian.stokes@intel.com>;
> Mcnamara, John <john.mcnamara@intel.com>; Luca Boccassi
> <bluca@debian.org>; Xu, Qian Q <qian.q.xu@intel.com>; Thomas Monjalon
> <thomas@monjalon.net>; Peng, Yuan <yuan.peng@intel.com>; Chen,
> Zhaoyan <zhaoyan.chen@intel.com>
> Subject: Re: 22.11.3 patches review and test
> 
> On 04/09/2023 10:32, Kevin Traynor wrote:
> > On 01/09/2023 04:23, Zeng, ZhichaoX wrote:
> >>> -----Original Message-----
> >>> From: Kevin Traynor <ktraynor@redhat.com>
> >>> Sent: Thursday, August 31, 2023 8:18 PM
> >>> To: Xu, HailinX <hailinx.xu@intel.com>; Xueming Li
> >>> <xuemingl@nvidia.com>; stable@dpdk.org; Wu, Jingjing
> >>> <jingjing.wu@intel.com>; Xing, Beilei <beilei.xing@intel.com>; Xu,
> >>> Ke1 <ke1.xu@intel.com>; Zeng, ZhichaoX <zhichaox.zeng@intel.com>;
> >>> Zhang, Qi Z <qi.z.zhang@intel.com>
> >>> Cc: xuemingl@nvdia.com; dev@dpdk.org; Stokes, Ian
> >>> <ian.stokes@intel.com>; Mcnamara, John <john.mcnamara@intel.com>;
> >>> Luca Boccassi <bluca@debian.org>; Xu, Qian Q <qian.q.xu@intel.com>;
> >>> Thomas Monjalon <thomas@monjalon.net>; Peng, Yuan
> >>> <yuan.peng@intel.com>; Chen, Zhaoyan <zhaoyan.chen@intel.com>
> >>> Subject: Re: 22.11.3 patches review and test
> >>>
> >>> On 30/08/2023 17:25, Kevin Traynor wrote:
> >>>> On 29/08/2023 09:22, Xu, HailinX wrote:
> >>>>>> -----Original Message-----
> >>>>>> From: Xueming Li <xuemingl@nvidia.com>
> >>>>>> Sent: Thursday, August 17, 2023 2:14 PM
> >>>>>> To: stable@dpdk.org
> >>>>>> Cc: xuemingl@nvdia.com; dev@dpdk.org; Abhishek Marathe
> >>>>>> <Abhishek.Marathe@microsoft.com>; Ali Alnubani
> >>>>>> <alialnu@nvidia.com>; Walker, Benjamin
> >>>>>> <benjamin.walker@intel.com>; David Christensen
> >>>>>> <drc@linux.vnet.ibm.com>; Hemant Agrawal
> >>> <hemant.agrawal@nxp.com>;
> >>>>>> Stokes, Ian <ian.stokes@intel.com>; Jerin Jacob
> >>>>>> <jerinj@marvell.com>; Mcnamara, John
> <john.mcnamara@intel.com>;
> >>>>>> Ju-Hyoung Lee <juhlee@microsoft.com>; Kevin Traynor
> >>>>>> <ktraynor@redhat.com>; Luca Boccassi <bluca@debian.org>; Pei
> >>>>>> Zhang <pezhang@redhat.com>; Xu, Qian Q <qian.q.xu@intel.com>;
> >>>>>> Raslan Darawsheh <rasland@nvidia.com>; Thomas Monjalon
> >>>>>> <thomas@monjalon.net>; Yanghang Liu <yanghliu@redhat.com>;
> Peng,
> >>>>>> Yuan <yuan.peng@intel.com>; Chen, Zhaoyan
> >>> <zhaoyan.chen@intel.com>
> >>>>>> Subject: 22.11.3 patches review and test
> >>>>>>
> >>>>>> Hi all,
> >>>>>>
> >>>>>> Here is a list of patches targeted for stable release 22.11.3.
> >>>>>>
> >>>>>> The planned date for the final release is 31th August.
> >>>>>>
> >>>>>> 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=v22.11.3-rc1
> >>>>>>
> >>>>>> These patches are located at branch 22.11 of dpdk-stable repo:
> >>>>>>         https://dpdk.org/browse/dpdk-stable/
> >>>>>>
> >>>>>> Thanks.
> >>>>>
> >>>>> We are conducting DPDK testing and have found two issues.
> >>>>>
> >>>>> 1. The startup speed of testpmd is significantly slower in the os of SUSE
> >>>>>       This issue fix patch has been merged into main, But it has
> >>>>> not backported
> >>> to 22.11.3.
> >>>>>       Fix patch commit id on DPDK main:
> >>>>> 7e7b6762eac292e78c77ad37ec0973c0c944b845
> >>>>>
> >>>>> 2. The SCTP tunnel packet of iavf cannot be forwarded in avx512
> >>>>> mode
> >>>
> >>> Need to clarify this sentence. It looks like it is not a functional
> >>> bug where
> >>> avx512 mode is selected and then an SCTP tunnel packet cannot be sent.
> >>> Instead, it is a possible performance issue that avx512 mode will
> >>> not be selected where it might have been due to unneeded additions
> >>> (RTE_ETH_TX_OFFLOAD_*_TNL_TSO) to IAVF_TX_NO_VECTOR_FLAGS.
> >>>
> >>> @IAVF maintainers - please confirm my analysis is correct ?
> >>>
> >>> In that case, as it is a possible performance issue in a specific
> >>> case for a single driver I think it is non-critical for 21.11 and we
> >>> can just revert the patch on the branch and wait for 21.11.6 release in
> December.
> >>
> >> Hi Kevin,
> >>
> >> Since the LTS version of the IAVF driver does not support avx512
> >> checksum offload, the scalar path should be selected, but this patch
> >> makes it incorrectly select the
> >> avx512 path, and the SCTP tunnel packets can't be forwarded properly.
> >>
> >
> > ok, let's take a look at the patch and usage.
> >
> > diff --git a/drivers/net/iavf/iavf_rxtx.h
> > b/drivers/net/iavf/iavf_rxtx.h index 8d4a77271a..605ea3f824 100644
> > --- a/drivers/net/iavf/iavf_rxtx.h
> > +++ b/drivers/net/iavf/iavf_rxtx.h
> > @@ -32,4 +32,8 @@
> >                   RTE_ETH_TX_OFFLOAD_MULTI_SEGS |          \
> >                   RTE_ETH_TX_OFFLOAD_TCP_TSO |             \
> > +               RTE_ETH_TX_OFFLOAD_VXLAN_TNL_TSO |       \
> > +               RTE_ETH_TX_OFFLOAD_GRE_TNL_TSO |         \
> > +               RTE_ETH_TX_OFFLOAD_IPIP_TNL_TSO |        \
> > +               RTE_ETH_TX_OFFLOAD_GENEVE_TNL_TSO |      \
> >                   RTE_ETH_TX_OFFLOAD_SECURITY)
> >
> > <snip>
> >
> > So we now have:
> > #define IAVF_TX_NO_VECTOR_FLAGS (				 \
> > 		RTE_ETH_TX_OFFLOAD_VLAN_INSERT |		 \
> > 		RTE_ETH_TX_OFFLOAD_QINQ_INSERT |		 \
> > 		RTE_ETH_TX_OFFLOAD_MULTI_SEGS |		 \
> > 		RTE_ETH_TX_OFFLOAD_TCP_TSO |		 \
> > 		RTE_ETH_TX_OFFLOAD_VXLAN_TNL_TSO |	 \
> > 		RTE_ETH_TX_OFFLOAD_GRE_TNL_TSO |	 \
> > 		RTE_ETH_TX_OFFLOAD_IPIP_TNL_TSO |	 \
> > 		RTE_ETH_TX_OFFLOAD_GENEVE_TNL_TSO |	 \
> > 		RTE_ETH_TX_OFFLOAD_SECURITY)
> >
> > <snip>
> >

Hi Kevin,

This patch also removes two flags from IAVF_TX_NO_VECTOR_FLAGS, RTE_ETH_TX_OFFLOAD_OUTER_IPV4_CKSUM
and RTE_ETH_TX_OFFLOAD_OUTER_UDP_CKSUM.

> > static inline int
> > iavf_tx_vec_queue_default(struct iavf_tx_queue *txq) {
> > 	if (!txq)
> > 		return -1;
> >
> > 	if (txq->rs_thresh < IAVF_VPMD_TX_MAX_BURST ||
> > 	    txq->rs_thresh > IAVF_VPMD_TX_MAX_FREE_BUF)
> > 		return -1;
> >
> > 	if (txq->offloads & IAVF_TX_NO_VECTOR_FLAGS)
> > 		return -1;
> >               ^^^ Adding the extra bits to IAVF_TX_NO_VECTOR_FLAGS
> > gives
> > *more* reasons for why this statement might not be true, so returning
> > -1 indicating that vector cannot be used for tx queue
> >
> 
> typo here - just to clarify, the new flags give more reasons for this statement
> to be true, so returning -1.
> 
> >
> > <snip>
> >
> > static inline bool
> > check_tx_vec_allow(struct iavf_tx_queue *txq) {
> > 	if (!(txq->offloads & IAVF_TX_NO_VECTOR_FLAGS) &&
> >
> >               ^^^ Adding the extra bits to IAVF_TX_NO_VECTOR_FLAGS
> > gives
> > *more* reason for this statement will be false and then false returned
> > indicating that vector cannot be used.
> >
> > 	    txq->rs_thresh >= IAVF_VPMD_TX_MAX_BURST &&
> > 	    txq->rs_thresh <= IAVF_VPMD_TX_MAX_FREE_BUF) {
> > 		PMD_INIT_LOG(DEBUG, "Vector tx can be enabled on this
> txq.");
> > 		return true;
> > 	}
> > 	PMD_INIT_LOG(DEBUG, "Vector Tx cannot be enabled on this txq.");
> > 	return false;
> > }
> >
> > --
> >
> > It looks like that adding the extra bits gives it less reasons to
> > select vector mode. However, you are saying that this patch means
> > there is a case where it now selects vector where it should not,
> > meaning additional reason to select vector mode. So maybe I miss
> something ?

Originally, the vector path would not be selected after configuring outer checksum offload,
but it will be selected after removing the two flags.
But the vector path doesn't support outer checksum offload on stable DPDK, so there will be a problem.

The key issue is that these two flags are removed, RTE_ETH_TX_OFFLOAD_OUTER_IPV4_CKSUM
and RTE_ETH_TX_OFFLOAD_OUTER_UDP_CKSUM.

Regards
Zhichao

> >
> >> Yes, we can revert this commit for 21.11.6 release, thanks.
> >>
> >> Regards
> >> Zhichao
> >>
> >>> thanks,
> >>> Kevin.
> >>>
> >>>>>       commit 9b7215f150d0bfc5cb00fce68ff08e5217c7f2b3 on
> v22.11.3-
> >>> rc1.
> >>>>>       This commit is for the new feature (avx512 checksum offload)
> >>>>> in DPDK
> >>> 23.03, which should not be backported to the LTS version since
> >>> avx512 checksum offload is not supported in v22.11.3 LTS.
> >>>>>
> >>>>
> >>>> Thanks for flagging Xueming.
> >>>>
> >>>> The issue is that it was listed as fixing 059f18ae2aec ("net/iavf:
> >>>> add offload path for Tx AVX512") which goes back to 21.05.
> >>>>
> >>>> This could have been avoided if the 'Fixes:' tag was correct, or if
> >>>> the authors replied to the email about queued backports :/
> >>>>
> >>>> Requesting iavf/next-net-intel maintainers to check Fixes: tags are
> >>>> correct before merging.
> >>>>
> >>>> DPDK 21.11.5 is already released with this patch. Any idea why it
> >>>> did not show up in validation for 21.11.5 ? Is it an issue for 21.11.5 ?
> >>>> How critical is it ?
> >>>>
> >>>> I can revert it on the 21.11 branch, but it will need to wait until
> >>>> 21.11.6 in December before it will be reverted in a released version.
> >>>>
> >>>> thanks,
> >>>> Kevin.
> >>>>
> >>>>> Regards,
> >>>>> Xu, Hailin
> >>>>>
> >>>>
> >>
> >

Re: 22.11.3 patches review and test

[Kevin Traynor]

On 05/09/2023 02:51, Zeng, ZhichaoX wrote:
> 
> 
> Regards
> Zhichao
>> -----Original Message-----
>> From: Kevin Traynor <ktraynor@redhat.com>
>> Sent: Monday, September 4, 2023 10:15 PM
>> To: Zeng, ZhichaoX <zhichaox.zeng@intel.com>; Xu, HailinX
>> <hailinx.xu@intel.com>; Xueming Li <xuemingl@nvidia.com>;
>> stable@dpdk.org; Wu, Jingjing <jingjing.wu@intel.com>; Xing, Beilei
>> <beilei.xing@intel.com>; Xu, Ke1 <ke1.xu@intel.com>; Zhang, Qi Z
>> <qi.z.zhang@intel.com>
>> Cc: xuemingl@nvdia.com; dev@dpdk.org; Stokes, Ian <ian.stokes@intel.com>;
>> Mcnamara, John <john.mcnamara@intel.com>; Luca Boccassi
>> <bluca@debian.org>; Xu, Qian Q <qian.q.xu@intel.com>; Thomas Monjalon
>> <thomas@monjalon.net>; Peng, Yuan <yuan.peng@intel.com>; Chen,
>> Zhaoyan <zhaoyan.chen@intel.com>
>> Subject: Re: 22.11.3 patches review and test
>>
>> On 04/09/2023 10:32, Kevin Traynor wrote:
>>> On 01/09/2023 04:23, Zeng, ZhichaoX wrote:
>>>>> -----Original Message-----
>>>>> From: Kevin Traynor <ktraynor@redhat.com>
>>>>> Sent: Thursday, August 31, 2023 8:18 PM
>>>>> To: Xu, HailinX <hailinx.xu@intel.com>; Xueming Li
>>>>> <xuemingl@nvidia.com>; stable@dpdk.org; Wu, Jingjing
>>>>> <jingjing.wu@intel.com>; Xing, Beilei <beilei.xing@intel.com>; Xu,
>>>>> Ke1 <ke1.xu@intel.com>; Zeng, ZhichaoX <zhichaox.zeng@intel.com>;
>>>>> Zhang, Qi Z <qi.z.zhang@intel.com>
>>>>> Cc: xuemingl@nvdia.com; dev@dpdk.org; Stokes, Ian
>>>>> <ian.stokes@intel.com>; Mcnamara, John <john.mcnamara@intel.com>;
>>>>> Luca Boccassi <bluca@debian.org>; Xu, Qian Q <qian.q.xu@intel.com>;
>>>>> Thomas Monjalon <thomas@monjalon.net>; Peng, Yuan
>>>>> <yuan.peng@intel.com>; Chen, Zhaoyan <zhaoyan.chen@intel.com>
>>>>> Subject: Re: 22.11.3 patches review and test
>>>>>
>>>>> On 30/08/2023 17:25, Kevin Traynor wrote:
>>>>>> On 29/08/2023 09:22, Xu, HailinX wrote:
>>>>>>>> -----Original Message-----
>>>>>>>> From: Xueming Li <xuemingl@nvidia.com>
>>>>>>>> Sent: Thursday, August 17, 2023 2:14 PM
>>>>>>>> To: stable@dpdk.org
>>>>>>>> Cc: xuemingl@nvdia.com; dev@dpdk.org; Abhishek Marathe
>>>>>>>> <Abhishek.Marathe@microsoft.com>; Ali Alnubani
>>>>>>>> <alialnu@nvidia.com>; Walker, Benjamin
>>>>>>>> <benjamin.walker@intel.com>; David Christensen
>>>>>>>> <drc@linux.vnet.ibm.com>; Hemant Agrawal
>>>>> <hemant.agrawal@nxp.com>;
>>>>>>>> Stokes, Ian <ian.stokes@intel.com>; Jerin Jacob
>>>>>>>> <jerinj@marvell.com>; Mcnamara, John
>> <john.mcnamara@intel.com>;
>>>>>>>> Ju-Hyoung Lee <juhlee@microsoft.com>; Kevin Traynor
>>>>>>>> <ktraynor@redhat.com>; Luca Boccassi <bluca@debian.org>; Pei
>>>>>>>> Zhang <pezhang@redhat.com>; Xu, Qian Q <qian.q.xu@intel.com>;
>>>>>>>> Raslan Darawsheh <rasland@nvidia.com>; Thomas Monjalon
>>>>>>>> <thomas@monjalon.net>; Yanghang Liu <yanghliu@redhat.com>;
>> Peng,
>>>>>>>> Yuan <yuan.peng@intel.com>; Chen, Zhaoyan
>>>>> <zhaoyan.chen@intel.com>
>>>>>>>> Subject: 22.11.3 patches review and test
>>>>>>>>
>>>>>>>> Hi all,
>>>>>>>>
>>>>>>>> Here is a list of patches targeted for stable release 22.11.3.
>>>>>>>>
>>>>>>>> The planned date for the final release is 31th August.
>>>>>>>>
>>>>>>>> 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=v22.11.3-rc1
>>>>>>>>
>>>>>>>> These patches are located at branch 22.11 of dpdk-stable repo:
>>>>>>>>          https://dpdk.org/browse/dpdk-stable/
>>>>>>>>
>>>>>>>> Thanks.
>>>>>>>
>>>>>>> We are conducting DPDK testing and have found two issues.
>>>>>>>
>>>>>>> 1. The startup speed of testpmd is significantly slower in the os of SUSE
>>>>>>>        This issue fix patch has been merged into main, But it has
>>>>>>> not backported
>>>>> to 22.11.3.
>>>>>>>        Fix patch commit id on DPDK main:
>>>>>>> 7e7b6762eac292e78c77ad37ec0973c0c944b845
>>>>>>>
>>>>>>> 2. The SCTP tunnel packet of iavf cannot be forwarded in avx512
>>>>>>> mode
>>>>>
>>>>> Need to clarify this sentence. It looks like it is not a functional
>>>>> bug where
>>>>> avx512 mode is selected and then an SCTP tunnel packet cannot be sent.
>>>>> Instead, it is a possible performance issue that avx512 mode will
>>>>> not be selected where it might have been due to unneeded additions
>>>>> (RTE_ETH_TX_OFFLOAD_*_TNL_TSO) to IAVF_TX_NO_VECTOR_FLAGS.
>>>>>
>>>>> @IAVF maintainers - please confirm my analysis is correct ?
>>>>>
>>>>> In that case, as it is a possible performance issue in a specific
>>>>> case for a single driver I think it is non-critical for 21.11 and we
>>>>> can just revert the patch on the branch and wait for 21.11.6 release in
>> December.
>>>>
>>>> Hi Kevin,
>>>>
>>>> Since the LTS version of the IAVF driver does not support avx512
>>>> checksum offload, the scalar path should be selected, but this patch
>>>> makes it incorrectly select the
>>>> avx512 path, and the SCTP tunnel packets can't be forwarded properly.
>>>>
>>>
>>> ok, let's take a look at the patch and usage.
>>>
>>> diff --git a/drivers/net/iavf/iavf_rxtx.h
>>> b/drivers/net/iavf/iavf_rxtx.h index 8d4a77271a..605ea3f824 100644
>>> --- a/drivers/net/iavf/iavf_rxtx.h
>>> +++ b/drivers/net/iavf/iavf_rxtx.h
>>> @@ -32,4 +32,8 @@
>>>                    RTE_ETH_TX_OFFLOAD_MULTI_SEGS |          \
>>>                    RTE_ETH_TX_OFFLOAD_TCP_TSO |             \
>>> +               RTE_ETH_TX_OFFLOAD_VXLAN_TNL_TSO |       \
>>> +               RTE_ETH_TX_OFFLOAD_GRE_TNL_TSO |         \
>>> +               RTE_ETH_TX_OFFLOAD_IPIP_TNL_TSO |        \
>>> +               RTE_ETH_TX_OFFLOAD_GENEVE_TNL_TSO |      \
>>>                    RTE_ETH_TX_OFFLOAD_SECURITY)
>>>
>>> <snip>
>>>
>>> So we now have:
>>> #define IAVF_TX_NO_VECTOR_FLAGS (				 \
>>> 		RTE_ETH_TX_OFFLOAD_VLAN_INSERT |		 \
>>> 		RTE_ETH_TX_OFFLOAD_QINQ_INSERT |		 \
>>> 		RTE_ETH_TX_OFFLOAD_MULTI_SEGS |		 \
>>> 		RTE_ETH_TX_OFFLOAD_TCP_TSO |		 \
>>> 		RTE_ETH_TX_OFFLOAD_VXLAN_TNL_TSO |	 \
>>> 		RTE_ETH_TX_OFFLOAD_GRE_TNL_TSO |	 \
>>> 		RTE_ETH_TX_OFFLOAD_IPIP_TNL_TSO |	 \
>>> 		RTE_ETH_TX_OFFLOAD_GENEVE_TNL_TSO |	 \
>>> 		RTE_ETH_TX_OFFLOAD_SECURITY)
>>>
>>> <snip>
>>>
> 
> Hi Kevin,
> 
> This patch also removes two flags from IAVF_TX_NO_VECTOR_FLAGS, RTE_ETH_TX_OFFLOAD_OUTER_IPV4_CKSUM
> and RTE_ETH_TX_OFFLOAD_OUTER_UDP_CKSUM.
> 
>>> static inline int
>>> iavf_tx_vec_queue_default(struct iavf_tx_queue *txq) {
>>> 	if (!txq)
>>> 		return -1;
>>>
>>> 	if (txq->rs_thresh < IAVF_VPMD_TX_MAX_BURST ||
>>> 	    txq->rs_thresh > IAVF_VPMD_TX_MAX_FREE_BUF)
>>> 		return -1;
>>>
>>> 	if (txq->offloads & IAVF_TX_NO_VECTOR_FLAGS)
>>> 		return -1;
>>>                ^^^ Adding the extra bits to IAVF_TX_NO_VECTOR_FLAGS
>>> gives
>>> *more* reasons for why this statement might not be true, so returning
>>> -1 indicating that vector cannot be used for tx queue
>>>
>>
>> typo here - just to clarify, the new flags give more reasons for this statement
>> to be true, so returning -1.
>>
>>>
>>> <snip>
>>>
>>> static inline bool
>>> check_tx_vec_allow(struct iavf_tx_queue *txq) {
>>> 	if (!(txq->offloads & IAVF_TX_NO_VECTOR_FLAGS) &&
>>>
>>>                ^^^ Adding the extra bits to IAVF_TX_NO_VECTOR_FLAGS
>>> gives
>>> *more* reason for this statement will be false and then false returned
>>> indicating that vector cannot be used.
>>>
>>> 	    txq->rs_thresh >= IAVF_VPMD_TX_MAX_BURST &&
>>> 	    txq->rs_thresh <= IAVF_VPMD_TX_MAX_FREE_BUF) {
>>> 		PMD_INIT_LOG(DEBUG, "Vector tx can be enabled on this
>> txq.");
>>> 		return true;
>>> 	}
>>> 	PMD_INIT_LOG(DEBUG, "Vector Tx cannot be enabled on this txq.");
>>> 	return false;
>>> }
>>>
>>> --
>>>
>>> It looks like that adding the extra bits gives it less reasons to
>>> select vector mode. However, you are saying that this patch means
>>> there is a case where it now selects vector where it should not,
>>> meaning additional reason to select vector mode. So maybe I miss
>> something ?
> 
> Originally, the vector path would not be selected after configuring outer checksum offload,
> but it will be selected after removing the two flags.
> But the vector path doesn't support outer checksum offload on stable DPDK, so there will be a problem.
> 
> The key issue is that these two flags are removed, RTE_ETH_TX_OFFLOAD_OUTER_IPV4_CKSUM
> and RTE_ETH_TX_OFFLOAD_OUTER_UDP_CKSUM.
> 

ok, i got it now. In the main branch (ca34627be5d6) and the 21.11 
backport (3eb4ad8ed694) those flags are not removed. Those flags are 
only removed in the 22.11 branch (9b7215f150d0).

So this is not an issue for 21.11. Thanks for helping clear it up.

Kevin.

> Regards
> Zhichao
> 
>>>
>>>> Yes, we can revert this commit for 21.11.6 release, thanks.
>>>>
>>>> Regards
>>>> Zhichao
>>>>
>>>>> thanks,
>>>>> Kevin.
>>>>>
>>>>>>>        commit 9b7215f150d0bfc5cb00fce68ff08e5217c7f2b3 on
>> v22.11.3-
>>>>> rc1.
>>>>>>>        This commit is for the new feature (avx512 checksum offload)
>>>>>>> in DPDK
>>>>> 23.03, which should not be backported to the LTS version since
>>>>> avx512 checksum offload is not supported in v22.11.3 LTS.
>>>>>>>
>>>>>>
>>>>>> Thanks for flagging Xueming.
>>>>>>
>>>>>> The issue is that it was listed as fixing 059f18ae2aec ("net/iavf:
>>>>>> add offload path for Tx AVX512") which goes back to 21.05.
>>>>>>
>>>>>> This could have been avoided if the 'Fixes:' tag was correct, or if
>>>>>> the authors replied to the email about queued backports :/
>>>>>>
>>>>>> Requesting iavf/next-net-intel maintainers to check Fixes: tags are
>>>>>> correct before merging.
>>>>>>
>>>>>> DPDK 21.11.5 is already released with this patch. Any idea why it
>>>>>> did not show up in validation for 21.11.5 ? Is it an issue for 21.11.5 ?
>>>>>> How critical is it ?
>>>>>>
>>>>>> I can revert it on the 21.11 branch, but it will need to wait until
>>>>>> 21.11.6 in December before it will be reverted in a released version.
>>>>>>
>>>>>> thanks,
>>>>>> Kevin.
>>>>>>
>>>>>>> Regards,
>>>>>>> Xu, Hailin
>>>>>>>
>>>>>>
>>>>
>>>
> 

RE: 22.11.3 patches review and test

[Xueming(Steven) Li]

> -----Original Message-----
> From: Kevin Traynor <ktraynor@redhat.com>
> Sent: 9/5/2023 16:50
> To: Zeng, ZhichaoX <zhichaox.zeng@intel.com>; Xu, HailinX
> <hailinx.xu@intel.com>; Xueming(Steven) Li <xuemingl@nvidia.com>;
> stable@dpdk.org; Wu, Jingjing <jingjing.wu@intel.com>; Xing, Beilei
> <beilei.xing@intel.com>; Xu, Ke1 <ke1.xu@intel.com>; Zhang, Qi Z
> <qi.z.zhang@intel.com>
> Cc: xuemingl@nvdia.com; dev@dpdk.org; Stokes, Ian <ian.stokes@intel.com>;
> Mcnamara, John <john.mcnamara@intel.com>; Luca Boccassi
> <bluca@debian.org>; Xu, Qian Q <qian.q.xu@intel.com>; NBU-Contact-
> Thomas Monjalon (EXTERNAL) <thomas@monjalon.net>; Peng, Yuan
> <yuan.peng@intel.com>; Chen, Zhaoyan <zhaoyan.chen@intel.com>; Yang,
> Qiming <qiming.yang@intel.com>; Zhou, YidingX <yidingx.zhou@intel.com>;
> Cui, KaixinX <kaixinx.cui@intel.com>
> Subject: Re: 22.11.3 patches review and test
> 
> On 05/09/2023 02:51, Zeng, ZhichaoX wrote:
> >
> >
> > Regards
> > Zhichao
> >> -----Original Message-----
> >> From: Kevin Traynor <ktraynor@redhat.com>
> >> Sent: Monday, September 4, 2023 10:15 PM
> >> To: Zeng, ZhichaoX <zhichaox.zeng@intel.com>; Xu, HailinX
> >> <hailinx.xu@intel.com>; Xueming Li <xuemingl@nvidia.com>;
> >> stable@dpdk.org; Wu, Jingjing <jingjing.wu@intel.com>; Xing, Beilei
> >> <beilei.xing@intel.com>; Xu, Ke1 <ke1.xu@intel.com>; Zhang, Qi Z
> >> <qi.z.zhang@intel.com>
> >> Cc: xuemingl@nvdia.com; dev@dpdk.org; Stokes, Ian
> >> <ian.stokes@intel.com>; Mcnamara, John <john.mcnamara@intel.com>;
> >> Luca Boccassi <bluca@debian.org>; Xu, Qian Q <qian.q.xu@intel.com>;
> >> Thomas Monjalon <thomas@monjalon.net>; Peng, Yuan
> >> <yuan.peng@intel.com>; Chen, Zhaoyan <zhaoyan.chen@intel.com>
> >> Subject: Re: 22.11.3 patches review and test
> >>
> >> On 04/09/2023 10:32, Kevin Traynor wrote:
> >>> On 01/09/2023 04:23, Zeng, ZhichaoX wrote:
> >>>>> -----Original Message-----
> >>>>> From: Kevin Traynor <ktraynor@redhat.com>
> >>>>> Sent: Thursday, August 31, 2023 8:18 PM
> >>>>> To: Xu, HailinX <hailinx.xu@intel.com>; Xueming Li
> >>>>> <xuemingl@nvidia.com>; stable@dpdk.org; Wu, Jingjing
> >>>>> <jingjing.wu@intel.com>; Xing, Beilei <beilei.xing@intel.com>; Xu,
> >>>>> Ke1 <ke1.xu@intel.com>; Zeng, ZhichaoX <zhichaox.zeng@intel.com>;
> >>>>> Zhang, Qi Z <qi.z.zhang@intel.com>
> >>>>> Cc: xuemingl@nvdia.com; dev@dpdk.org; Stokes, Ian
> >>>>> <ian.stokes@intel.com>; Mcnamara, John
> <john.mcnamara@intel.com>;
> >>>>> Luca Boccassi <bluca@debian.org>; Xu, Qian Q
> >>>>> <qian.q.xu@intel.com>; Thomas Monjalon <thomas@monjalon.net>;
> >>>>> Peng, Yuan <yuan.peng@intel.com>; Chen, Zhaoyan
> >>>>> <zhaoyan.chen@intel.com>
> >>>>> Subject: Re: 22.11.3 patches review and test
> >>>>>
> >>>>> On 30/08/2023 17:25, Kevin Traynor wrote:
> >>>>>> On 29/08/2023 09:22, Xu, HailinX wrote:
> >>>>>>>> -----Original Message-----
> >>>>>>>> From: Xueming Li <xuemingl@nvidia.com>
> >>>>>>>> Sent: Thursday, August 17, 2023 2:14 PM
> >>>>>>>> To: stable@dpdk.org
> >>>>>>>> Cc: xuemingl@nvdia.com; dev@dpdk.org; Abhishek Marathe
> >>>>>>>> <Abhishek.Marathe@microsoft.com>; Ali Alnubani
> >>>>>>>> <alialnu@nvidia.com>; Walker, Benjamin
> >>>>>>>> <benjamin.walker@intel.com>; David Christensen
> >>>>>>>> <drc@linux.vnet.ibm.com>; Hemant Agrawal
> >>>>> <hemant.agrawal@nxp.com>;
> >>>>>>>> Stokes, Ian <ian.stokes@intel.com>; Jerin Jacob
> >>>>>>>> <jerinj@marvell.com>; Mcnamara, John
> >> <john.mcnamara@intel.com>;
> >>>>>>>> Ju-Hyoung Lee <juhlee@microsoft.com>; Kevin Traynor
> >>>>>>>> <ktraynor@redhat.com>; Luca Boccassi <bluca@debian.org>; Pei
> >>>>>>>> Zhang <pezhang@redhat.com>; Xu, Qian Q <qian.q.xu@intel.com>;
> >>>>>>>> Raslan Darawsheh <rasland@nvidia.com>; Thomas Monjalon
> >>>>>>>> <thomas@monjalon.net>; Yanghang Liu <yanghliu@redhat.com>;
> >> Peng,
> >>>>>>>> Yuan <yuan.peng@intel.com>; Chen, Zhaoyan
> >>>>> <zhaoyan.chen@intel.com>
> >>>>>>>> Subject: 22.11.3 patches review and test
> >>>>>>>>
> >>>>>>>> Hi all,
> >>>>>>>>
> >>>>>>>> Here is a list of patches targeted for stable release 22.11.3.
> >>>>>>>>
> >>>>>>>> The planned date for the final release is 31th August.
> >>>>>>>>
> >>>>>>>> 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=v22.11.3-rc1
> >>>>>>>>
> >>>>>>>> These patches are located at branch 22.11 of dpdk-stable repo:
> >>>>>>>>          https://dpdk.org/browse/dpdk-stable/
> >>>>>>>>
> >>>>>>>> Thanks.
> >>>>>>>
> >>>>>>> We are conducting DPDK testing and have found two issues.
> >>>>>>>
> >>>>>>> 1. The startup speed of testpmd is significantly slower in the os of
> SUSE
> >>>>>>>        This issue fix patch has been merged into main, But it
> >>>>>>> has not backported
> >>>>> to 22.11.3.
> >>>>>>>        Fix patch commit id on DPDK main:
> >>>>>>> 7e7b6762eac292e78c77ad37ec0973c0c944b845
> >>>>>>>
> >>>>>>> 2. The SCTP tunnel packet of iavf cannot be forwarded in avx512
> >>>>>>> mode
> >>>>>
> >>>>> Need to clarify this sentence. It looks like it is not a
> >>>>> functional bug where
> >>>>> avx512 mode is selected and then an SCTP tunnel packet cannot be
> sent.
> >>>>> Instead, it is a possible performance issue that avx512 mode will
> >>>>> not be selected where it might have been due to unneeded additions
> >>>>> (RTE_ETH_TX_OFFLOAD_*_TNL_TSO) to IAVF_TX_NO_VECTOR_FLAGS.
> >>>>>
> >>>>> @IAVF maintainers - please confirm my analysis is correct ?
> >>>>>
> >>>>> In that case, as it is a possible performance issue in a specific
> >>>>> case for a single driver I think it is non-critical for 21.11 and
> >>>>> we can just revert the patch on the branch and wait for 21.11.6
> >>>>> release in
> >> December.
> >>>>
> >>>> Hi Kevin,
> >>>>
> >>>> Since the LTS version of the IAVF driver does not support avx512
> >>>> checksum offload, the scalar path should be selected, but this
> >>>> patch makes it incorrectly select the
> >>>> avx512 path, and the SCTP tunnel packets can't be forwarded properly.
> >>>>
> >>>
> >>> ok, let's take a look at the patch and usage.
> >>>
> >>> diff --git a/drivers/net/iavf/iavf_rxtx.h
> >>> b/drivers/net/iavf/iavf_rxtx.h index 8d4a77271a..605ea3f824 100644
> >>> --- a/drivers/net/iavf/iavf_rxtx.h
> >>> +++ b/drivers/net/iavf/iavf_rxtx.h
> >>> @@ -32,4 +32,8 @@
> >>>                    RTE_ETH_TX_OFFLOAD_MULTI_SEGS |          \
> >>>                    RTE_ETH_TX_OFFLOAD_TCP_TSO |             \
> >>> +               RTE_ETH_TX_OFFLOAD_VXLAN_TNL_TSO |       \
> >>> +               RTE_ETH_TX_OFFLOAD_GRE_TNL_TSO |         \
> >>> +               RTE_ETH_TX_OFFLOAD_IPIP_TNL_TSO |        \
> >>> +               RTE_ETH_TX_OFFLOAD_GENEVE_TNL_TSO |      \
> >>>                    RTE_ETH_TX_OFFLOAD_SECURITY)
> >>>
> >>> <snip>
> >>>
> >>> So we now have:
> >>> #define IAVF_TX_NO_VECTOR_FLAGS (				 \
> >>> 		RTE_ETH_TX_OFFLOAD_VLAN_INSERT |		 \
> >>> 		RTE_ETH_TX_OFFLOAD_QINQ_INSERT |		 \
> >>> 		RTE_ETH_TX_OFFLOAD_MULTI_SEGS |		 \
> >>> 		RTE_ETH_TX_OFFLOAD_TCP_TSO |		 \
> >>> 		RTE_ETH_TX_OFFLOAD_VXLAN_TNL_TSO |	 \
> >>> 		RTE_ETH_TX_OFFLOAD_GRE_TNL_TSO |	 \
> >>> 		RTE_ETH_TX_OFFLOAD_IPIP_TNL_TSO |	 \
> >>> 		RTE_ETH_TX_OFFLOAD_GENEVE_TNL_TSO |	 \
> >>> 		RTE_ETH_TX_OFFLOAD_SECURITY)
> >>>
> >>> <snip>
> >>>
> >
> > Hi Kevin,
> >
> > This patch also removes two flags from IAVF_TX_NO_VECTOR_FLAGS,
> > RTE_ETH_TX_OFFLOAD_OUTER_IPV4_CKSUM
> > and RTE_ETH_TX_OFFLOAD_OUTER_UDP_CKSUM.
> >
> >>> static inline int
> >>> iavf_tx_vec_queue_default(struct iavf_tx_queue *txq) {
> >>> 	if (!txq)
> >>> 		return -1;
> >>>
> >>> 	if (txq->rs_thresh < IAVF_VPMD_TX_MAX_BURST ||
> >>> 	    txq->rs_thresh > IAVF_VPMD_TX_MAX_FREE_BUF)
> >>> 		return -1;
> >>>
> >>> 	if (txq->offloads & IAVF_TX_NO_VECTOR_FLAGS)
> >>> 		return -1;
> >>>                ^^^ Adding the extra bits to IAVF_TX_NO_VECTOR_FLAGS
> >>> gives
> >>> *more* reasons for why this statement might not be true, so
> >>> returning
> >>> -1 indicating that vector cannot be used for tx queue
> >>>
> >>
> >> typo here - just to clarify, the new flags give more reasons for this
> >> statement to be true, so returning -1.
> >>
> >>>
> >>> <snip>
> >>>
> >>> static inline bool
> >>> check_tx_vec_allow(struct iavf_tx_queue *txq) {
> >>> 	if (!(txq->offloads & IAVF_TX_NO_VECTOR_FLAGS) &&
> >>>
> >>>                ^^^ Adding the extra bits to IAVF_TX_NO_VECTOR_FLAGS
> >>> gives
> >>> *more* reason for this statement will be false and then false
> >>> returned indicating that vector cannot be used.
> >>>
> >>> 	    txq->rs_thresh >= IAVF_VPMD_TX_MAX_BURST &&
> >>> 	    txq->rs_thresh <= IAVF_VPMD_TX_MAX_FREE_BUF) {
> >>> 		PMD_INIT_LOG(DEBUG, "Vector tx can be enabled on this
> >> txq.");
> >>> 		return true;
> >>> 	}
> >>> 	PMD_INIT_LOG(DEBUG, "Vector Tx cannot be enabled on this txq.");
> >>> 	return false;
> >>> }
> >>>
> >>> --
> >>>
> >>> It looks like that adding the extra bits gives it less reasons to
> >>> select vector mode. However, you are saying that this patch means
> >>> there is a case where it now selects vector where it should not,
> >>> meaning additional reason to select vector mode. So maybe I miss
> >> something ?
> >
> > Originally, the vector path would not be selected after configuring
> > outer checksum offload, but it will be selected after removing the two flags.
> > But the vector path doesn't support outer checksum offload on stable DPDK,
> so there will be a problem.
> >
> > The key issue is that these two flags are removed,
> > RTE_ETH_TX_OFFLOAD_OUTER_IPV4_CKSUM
> > and RTE_ETH_TX_OFFLOAD_OUTER_UDP_CKSUM.
> >
> 
> ok, i got it now. In the main branch (ca34627be5d6) and the 21.11 backport
> (3eb4ad8ed694) those flags are not removed. Those flags are only removed in
> the 22.11 branch (9b7215f150d0).
> 
> So this is not an issue for 21.11. Thanks for helping clear it up.

Hi Guys, thanks for the clarification, the patch has been reverted in 22.11 candidate branch.

Thanks,
Xueming

> 
> Kevin.
> 
> > Regards
> > Zhichao
> >
> >>>
> >>>> Yes, we can revert this commit for 21.11.6 release, thanks.
> >>>>
> >>>> Regards
> >>>> Zhichao
> >>>>
> >>>>> thanks,
> >>>>> Kevin.
> >>>>>
> >>>>>>>        commit 9b7215f150d0bfc5cb00fce68ff08e5217c7f2b3 on
> >> v22.11.3-
> >>>>> rc1.
> >>>>>>>        This commit is for the new feature (avx512 checksum
> >>>>>>> offload) in DPDK
> >>>>> 23.03, which should not be backported to the LTS version since
> >>>>> avx512 checksum offload is not supported in v22.11.3 LTS.
> >>>>>>>
> >>>>>>
> >>>>>> Thanks for flagging Xueming.
> >>>>>>
> >>>>>> The issue is that it was listed as fixing 059f18ae2aec ("net/iavf:
> >>>>>> add offload path for Tx AVX512") which goes back to 21.05.
> >>>>>>
> >>>>>> This could have been avoided if the 'Fixes:' tag was correct, or
> >>>>>> if the authors replied to the email about queued backports :/
> >>>>>>
> >>>>>> Requesting iavf/next-net-intel maintainers to check Fixes: tags
> >>>>>> are correct before merging.
> >>>>>>
> >>>>>> DPDK 21.11.5 is already released with this patch. Any idea why it
> >>>>>> did not show up in validation for 21.11.5 ? Is it an issue for 21.11.5 ?
> >>>>>> How critical is it ?
> >>>>>>
> >>>>>> I can revert it on the 21.11 branch, but it will need to wait
> >>>>>> until
> >>>>>> 21.11.6 in December before it will be reverted in a released version.
> >>>>>>
> >>>>>> thanks,
> >>>>>> Kevin.
> >>>>>>
> >>>>>>> Regards,
> >>>>>>> Xu, Hailin
> >>>>>>>
> >>>>>>
> >>>>
> >>>
> >