[ovs-build] Canceled: williamtu/ovs-travis#223 (afxdp-v18 - e51592d)
Travis CI
builds at travis-ci.org
Thu Jul 18 20:21:53 UTC 2019
Build Update for williamtu/ovs-travis
-------------------------------------
Build: #223
Status: Canceled
Duration: 56 mins and 19 secs
Commit: e51592d (afxdp-v18)
Author: William Tu
Message: netdev-afxdp: add new netdev type for AF_XDP.
The patch introduces experimental AF_XDP support for OVS netdev.
AF_XDP, the Address Family of the eXpress Data Path, is a new Linux socket
type built upon the eBPF and XDP technology. It is aims to have comparable
performance to DPDK but cooperate better with existing kernel's networking
stack. An AF_XDP socket receives and sends packets from an eBPF/XDP program
attached to the netdev, by-passing a couple of Linux kernel's subsystems
As a result, AF_XDP socket shows much better performance than AF_PACKET
For more details about AF_XDP, please see linux kernel's
Documentation/networking/af_xdp.rst. Note that by default, this feature is
not compiled in.
Signed-off-by: William Tu <u9012063 at gmail.com>
---
v15:
* address review feedback from Ilya
https://patchwork.ozlabs.org/patch/1125476/
* skip TCP related test cases
* reclaim all CONS_NUM_DESC at complete tx
* add retries to kick_tx
* increase memory pool size
* remove redundant xdp flag and bind flag
* remove unused rx_dropped var
* make tx_dropped counter atomic
* refactor dp_packet_init_afxdp using dp_packet_init__
* rebase to ovs master, test with latest bpf-next kernel commit b14a260e33ddb4
Ilya's kernel patches are required
commit 455302d1c9ae ("xdp: fix hang while unregistering device bound to xdp socket")
commit 162c820ed896 ("xdp: hold device for umem regardless of zero-copy mode")
Possible issues:
* still lots of afxdp_cq_skip (ovs-appctl coverage/show)
afxdp_cq_skip 44325273.6/sec 34362312.683/sec 572705.2114/sec total: 2106010377
* TODO:
'make check-afxdp' still not all pass
IP fragmentation expiry test not fix yet, need to implement
deferral memory free, s.t like dpdk_mp_sweep. Currently hit
some missing umem descs when reclaiming.
NSH test case still failed (not due to afxdp)
v16:
* address feedbacks from Ilya
* add deferral memory free
* add afxdp testsuites files to gitignore
v17:
* address feedbacks from Ilya and Ben
https://patchwork.ozlabs.org/patch/1131547/
* ovs_spin_lock: add pthread_spin_lock checks, fix typo
* update NEWS
* add pthread_spin_lock check at OVS_CHECK_LINUX_AF_XDP
* fix bug in xmalloc_size_align
* rename xdpsock to netdev-afxdp-pool
* remove struct umem_elem, use void *
* fix style and comments
* fix afxdp.rst
* rebase to OVS master, tested on kernel 5.2.0-rc6
Note: I still leave the last_tsc in pmd_perf_stats the same as v16
v18:
* address feedbacks from Ilya and Ben
https://patchwork.ozlabs.org/patch/1133416/
* update document about tcp and reconfiguration
* fix leak in tx spin locks
* refactor __umem_pool alloc and assert
* refactor macro and defines used in netdev-afxdp[-pool]
* refactor the xpool->array to remove using type casting
* add empty netdev_afxdp_rxq_destruct to avoid closing
afxdp socket
View the changeset: https://github.com/williamtu/ovs-travis/compare/516ab1325749^...e51592d21109
View the full build log and details: https://travis-ci.org/williamtu/ovs-travis/builds/560643004?utm_medium=notification&utm_source=email
--
You can unsubscribe from build emails from the williamtu/ovs-travis repository going to https://travis-ci.org/account/preferences/unsubscribe?repository=6871381&utm_medium=notification&utm_source=email.
Or unsubscribe from *all* email updating your settings at https://travis-ci.org/account/preferences/unsubscribe?utm_medium=notification&utm_source=email.
Or configure specific recipients for build notifications in your .travis.yml file. See https://docs.travis-ci.com/user/notifications.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openvswitch.org/pipermail/ovs-build/attachments/20190718/c65a2af8/attachment-0001.html>
More information about the build
mailing list