On 08/02/2017 05:13 PM, Jim Fehlig wrote:
FYI, I've noticed that this patch causes build failures of 3.6.0 on s390x and ppc64
[ 189s] 100) node-name-detect(iscsi) ... [ 189s] In '/home/abuild/rpmbuild/BUILD/libvirt-3.6.0/tests/qemumonitorjsondata/qemumonitorjson-nodename-iscsi.result':
[ 189s] Offset 6 [ 189s] Expect [scsi0-0-1 [ 189s] filename : 'json:{"lun": "0", "portal": "example.com:3260", "driver": "iscsi", "transport": "tcp", "target": "iqn.2016-09.com.example:server"}'
As you can see the device order is swapped in the "Actual" output. It is odd that I don't see the failure on other arches. I have little experience with the monitor test code. Any ideas on what might cause that?
Known issue with endianness affecting the mocked hashing; patch posted here: https://www.redhat.com/archives/libvir-list/2017-August/msg00091.html -- Eric Blake, Principal Software Engineer Red Hat, Inc. +1-919-301-3266 Virtualization: qemu.org | libvirt.org