On 21/09/2020 11:59, Pei Zhang wrote: > Hello Kevin, > > Testing with dpdk 18.11.10 from Red Hat looks good. > > We cover below 15 scenarios and and all get PASS on RHEL8 testing: > > (1)Guest with device assignment(PF) throughput testing(1G hugepage size): PASS > (2)Guest with device assignment(PF) throughput testing(2M hugepage size) : > PASS > (3)Guest with device assignment(VF) throughput testing: PASS > (4)PVP (host dpdk testpmd as vswitch) 1Q: throughput testing: PASS > (5)PVP vhost-user 2Q throughput testing: PASS > (6)PVP vhost-user 1Q - cross numa node throughput testing: PASS > (7)Guest with vhost-user 2 queues throughput testing: PASS > (8)vhost-user reconnect with dpdk-client, qemu-server: qemu reconnect: PASS > (9)PVP 1Q live migration testing: PASS > (10)PVP 1Q cross numa node live migration testing: PASS > (11)Guest with ovs+dpdk+vhost-user 1Q live migration testing: PASS > (12)Guest with ovs+dpdk+vhost-user 1Q live migration testing (2M): PASS > (13)Guest with ovs+dpdk+vhost-user 2Q live migration testing: PASS > (14)Host PF + DPDK testing: PASS > (15)Host VF + DPDK testing: PASS > > Versions: > > kernel 4.18 > qemu 5.1 > dpdk: git://dpdk.org/dpdk-stable remotes/origin/18.11 > # git log -1 > commit 1f212ec29c3d1be9b49e1d647be4509d6c35d6c6 (HEAD, origin/18.11) > Author: Anatoly Burakov <anatoly.bura...@intel.com> > Date: Wed Sep 16 12:26:15 2020 +0100 > > vfio: map contiguous areas in one go > > [ upstream commit d1c7c0cdf7bac5eb40d3a2a690453aefeee5887b ] > > Currently, when we are creating DMA mappings for memory that's > either external or is backed by hugepages in IOVA as PA mode, we > assume that each page is necessarily discontiguous. This may not > actually be the case, especially for external memory, where the > user is able to create their own IOVA table and make it > contiguous. This is a problem because VFIO has a limited number > of DMA mappings, and it does not appear to concatenate them and > treats each mapping as separate, even when they cover adjacent > areas. > > Fix this so that we always map contiguous memory in a single > chunk, as opposed to mapping each segment separately. > > Signed-off-by: Anatoly Burakov <anatoly.bura...@intel.com> > > > NICs: X540-AT2 NIC(ixgbe, 10G) > > Best regards, > > Pei >
Great, thanks for your testing Pei. Kevin. > > > ----- Original Message ----- >> From: "Kevin Traynor" <ktray...@redhat.com> >> To: sta...@dpdk.org >> Cc: dev@dpdk.org, "Abhishek Marathe" <abhishek.mara...@microsoft.com>, >> "Akhil Goyal" <akhil.go...@nxp.com>, "Ali >> Alnubani" <alia...@mellanox.com>, "benjamin walker" >> <benjamin.wal...@intel.com>, "David Christensen" >> <d...@linux.vnet.ibm.com>, "Hemant Agrawal" <hemant.agra...@nxp.com>, "Ian >> Stokes" <ian.sto...@intel.com>, "Jerin >> Jacob" <jer...@marvell.com>, "John McNamara" <john.mcnam...@intel.com>, >> "Ju-Hyoung Lee" <juh...@microsoft.com>, >> "Kevin Traynor" <ktray...@redhat.com>, "Luca Boccassi" <bl...@debian.org>, >> "Pei Zhang" <pezh...@redhat.com>, "pingx >> yu" <pingx...@intel.com>, "qian q xu" <qian.q...@intel.com>, "Raslan >> Darawsheh" <rasl...@mellanox.com>, "Thomas >> Monjalon" <tho...@monjalon.net>, "yuan peng" <yuan.p...@intel.com>, "zhaoyan >> chen" <zhaoyan.c...@intel.com> >> Sent: Monday, September 7, 2020 6:12:24 PM >> Subject: 18.11.10 (LTS) patches review and test >> >> Hi all, >> >> Here is a list of patches targeted for LTS release 18.11.10. >> >> The planned date for the final release is 21 September. >> >> Please help with testing and validation of your use cases and report >> any issues/results with reply-all to this mail. For the final release >> the fixes and reported validations will be added to the release notes. >> >> A release candidate tarball can be found at: >> >> https://dpdk.org/browse/dpdk-stable/tag/?id=v18.11.10-rc1 >> >> These patches are located at branch 18.11 of dpdk-stable repo: >> https://dpdk.org/browse/dpdk-stable/ >> >> Thanks. >> >> Kevin. >> >> --- >> Adam Dybkowski (2): >> test/crypto: fix asymmetric session mempool creation >> common/qat: fix uninitialized variable >> >> Akhil Goyal (1): >> crypto/dpaax_sec: fix inline query for descriptors >> >> Alan Dewar (1): >> sched: fix port time rounding >> >> Alexander Kozyrev (1): >> net/mlx5: fix vectorized Rx burst termination >> >> Ankur Dwivedi (1): >> common/cpt: fix encryption offset >> >> Apeksha Gupta (2): >> app/eventdev: fix capability check in pipeline ATQ test >> event/dpaa2: add all-types queue capability flag >> >> Archana Muniganti (1): >> examples/fips_validation: fix count overwrite for TDES >> >> Ayuj Verma (1): >> examples/fips_validation: fix parsing of TDES vectors >> >> Bruce Richardson (4): >> mem: fix 32-bit init config with meson >> rawdev: allow getting info for unknown device >> rawdev: fill NUMA socket ID in info >> rawdev: export dump function in map file >> >> Chas Williams (1): >> net/netvsc: fix crash during Tx >> >> Chenbo Xia (1): >> net/virtio-user: check tap system call setting >> >> Chenmin Sun (1): >> net/i40e: fix flow director Rx writeback packet >> >> David Marchand (4): >> net/mvpp2: fix non-EAL thread support >> devtools: fix path in forbidden token check >> lib: remind experimental status in headers >> net/bonding: fix socket ID check >> >> Devendra Singh Rawat (1): >> net/qede: fix multicast drop in promiscuous mode >> >> Dongyang Pan (1): >> net/bonding: delete redundant code >> >> Ferruh Yigit (4): >> net/iavf: fix uninitialized variable >> net/ixgbe/base: remove dead code >> ethdev: fix log type for some error messages >> drivers/net: fix exposing internal headers >> >> Gaurav Singh (1): >> net/i40e: remove duplicate tunnel type check >> >> Georgiy Levashov (1): >> examples: add flush after stats printing >> >> Guinan Sun (5): >> net/ixgbe/base: fix host interface shadow RAM read >> net/ixgbe/base: fix x550em 10G NIC link status >> net/ixgbe/base: fix infinite recursion on PCIe link down >> net/ixgbe: fix MAC control frame forward >> net/ixgbe: fix flow control status >> >> Haiyue Wang (1): >> eal: fix uuid header dependencies >> >> Harman Kalra (1): >> eal: fix parentheses in alignment macros >> >> Harry van Haaren (1): >> examples/eventdev: fix 32-bit coremask >> >> Heinrich Kuhn (1): >> net/nfp: fix RSS hash configuration reporting >> >> Hemant Agrawal (1): >> rawdev: remove remaining experimental tags >> >> Hongzhi Guo (3): >> net: fix IPv4 checksum >> net: fix unneeded replacement of TCP checksum 0 >> net: fix checksum on big endian CPUs >> >> Honnappa Nagarahalli (4): >> eal/armv8: fix timer frequency calibration with PMU >> app/testpmd: use clock time in throughput calculation >> app/testpmd: fix burst percentage calculation >> doc: fix some typos in Linux guide >> >> Ian Dolzhansky (1): >> net/failsafe: fix RSS RETA size info >> >> Jecky Pei (1): >> net/kni: set packet input port in Rx >> >> Jeff Guo (1): >> net/e1000: fix crash on Tx done clean up >> >> Kalesh AP (1): >> net/bnxt: remove unused enum declaration >> >> Kevin Traynor (1): >> kni: fix ethtool build error on kernel 5.7 >> >> Konstantin Ananyev (1): >> bpf: fix add/sub min/max estimations >> >> Levend Sayar (1): >> service: fix C++ linkage >> >> Long Li (3): >> bus/vmbus: fix ring buffer mapping >> net/netvsc: fix underflow when Rx external mbuf >> net/netvsc: fix chimney index >> >> Lukasz Wojciechowski (1): >> service: fix core mapping reset >> >> Mao Jiang (2): >> net/i40e: fix binding interrupt without MSI-X vector >> net/i40e: fix flow director MSI-X resource allocation >> >> Matan Azrad (1): >> vhost: fix features definition location >> >> Maxime Coquelin (1): >> vhost: fix virtio ready flag check >> >> Michael Baum (4): >> net/mlx5: fix unreachable MPLS error path >> net/mlx5: fix iterator type in Rx queue management >> net/mlx5: fix crash in NVGRE item translation >> net/mlx4: optimize stack memory size in probe >> >> Muhammad Bilal (1): >> bus/vdev: fix a typo in doxygen comment >> >> Nipun Gupta (2): >> bus/fslmc: fix getting FD error >> net/dpaa: fix FD offset data type >> >> Pablo de Lara (1): >> drivers/crypto: add missing OOP feature flag >> >> Patrick Fu (1): >> vhost: fix double-free with zero-copy >> >> Pavel Ivashchenko (1): >> net/i40e: fix getting EEPROM information >> >> Rahul Lakkireddy (3): >> net/cxgbe: fix CLIP leak in filter error path >> net/cxgbe: fix double MPS alloc by flow validate and create >> net/cxgbe: fix L2T leak in filter error and free path >> >> Raslan Darawsheh (2): >> net: fix pedantic build >> net/mlx5: fix flow items size calculation >> >> Ruifeng Wang (6): >> net/i40e: enable NEON Rx/Tx in meson >> eal/arm: add vcopyq intrinsic for aarch32 >> net/ixgbe: fix include of vector header file >> test/cycles: restore default delay callback >> crypto/armv8: remove debug option >> crypto/armv8: use dedicated log type >> >> Sarosh Arif (3): >> doc: fix typo in bbdev test guide >> examples/packet_ordering: use proper exit method >> doc: fix ethtool app path >> >> Stephen Hemminger (4): >> net/netvsc: do not spin forever waiting for reply >> cfgfile: fix stack buffer underflow >> kni: fix reference to master/slave process >> doc: fix reference to master process >> >> Tal Shnaiderman (1): >> pci: fix address domain format size >> >> Timothy Redaelli (1): >> test: fix rpath for drivers with meson >> >> Wei Hu (Xavier) (5): >> ethdev: fix data room size verification in Rx queue setup >> net/bonding: fix error code on device creation >> net/bonding: fix MAC address when switching active port >> net/bonding: fix MAC address when one port resets >> app/testpmd: fix stats error message >> >> Wei Zhao (1): >> net/ixgbe: report 10Mbps link speed for x553 >> >> Xiao Wang (1): >> net/virtio-user: fix status management >> >> Yicai Lu (1): >> net/bonding: fix LACP negotiation >> >> Yunjian Wang (8): >> app/testpmd: fix memory leak on error path >> vfio: remove unused variable >> net/af_packet: fix check of file descriptors >> net/af_packet: fix memory leak on init failure >> net/af_packet: fix munmap on init failure >> hash: fix out-of-memory handling in hash creation >> event/dpaa: remove dead code >> net/qede: remove dead code >> >> Zhiguang He (1): >> net/bonding: fix dead loop on RSS RETA update >> >> Zhihong Peng (2): >> net/i40e: report VLAN filter capability >> net/e1000: report VLAN extend capability >>
