Currently, in the fast tests suite, there are 2 tests failing and preventing the environment from being put into production: interrupt_autotest and memory_autotest. This test was on DPDK main, commit a95bbb72623c310df4d0c8ad45c2ee06f538e01b. interrupt_autotest fails quite quickly: stdout: RTE>>interrupt_autotest Check unknown valid interrupt full path callback has not been called failure occurred during checking unknown valid interrupt full path Clearing for interrupt tests Test Failed memory_autotest fails after dumping memory: stdout: RTE>>memory_autotest Dump memory layout ... Error getting segment fd's Test Failed From discussions on the CI mailing list, it has been reproduced and needs further debugging to resolve the issues. Bug here is for tracking the fixes within DPDK, https://mails.dpdk.org/archives/ci/2021-July/001267.html Once a patch is available, please let the CI team know.
It should be fixed with: https://git.dpdk.org/dpdk/commit/?id=086d426406bd Sorry, I forgot to reference the bz before pushing.
The interrupt test issue is not addressed yet.
David, Do you plan to take care of the interrupt test issue? Thanks
No. Parking to Bruce as FreeBSD maintainer.
The interrupt test issue is still blocking the deployment of FreeBSD 13.1 unit testing for CI. Should we expect this issue to be fixed soon?
I have no time available at present to look at this, so it's unlikely to be fixed in the short term.
Hi. Just nothing that we are still monitoring this ticket for bringing unit tests online for our freebsd 13.2 VM. We are also adding a freebsd 14 VM to the lab, so if helpful we can run unit tests across that machine when ready and post here whether unit tests succeed in that environment.
Hi, unit tests are also failing for FreeBSD 14.0. Our /boot/loader.conf contains: kern.geom.label.disk_ident.enable="0" kern.geom.label.gptid.enable="0" cryptodev_load="YES" zfs_load="YES" hw.contigmem.num_buffers=4 hw.contigmem.buffer_size=1073741824 contigmem_load="YES" nic_uio_load="YES" And our /etc/fstab contains the line: proc /proc procfs rw 0 0 We are building DPDK as root with `meson build` and `ninja -C build install` and running unit tests with `meson test --suite fast-tests`. I noticed that most of the errors are related to EAL flag errors so I was wondering if there were any configurations that I can set to fix this. The Meson Unit Test Log is 100KB too large to attach here but I can share it with people if needed. Thanks.