Discussion:
4209 Populate zfstest with the remainder of the STF tests
Christopher Siden
2013-10-14 22:16:59 UTC
Permalink
http://cr.illumos.org/~webrev/csiden/illumos-zfs-tests/

Work by John Kennedy. This ports the rest of the ZFS test suite from
STF (in a different repository on github) to the testrunner framework
in illumos-gate.

Chris
Yuri Pankov
2013-10-15 12:24:39 UTC
Permalink
Post by Christopher Siden
http://cr.illumos.org/~webrev/csiden/illumos-zfs-tests/
Work by John Kennedy. This ports the rest of the ZFS test suite from
STF (in a different repository on github) to the testrunner framework
in illumos-gate.
Looks good, though I haven't looked through all added tests, only at the
infrastructure changes (thanks for fixing my typo).

One question - are we really interested in *physical* CPUs in is_mp()
function? The tests where it's used don't give any hints why.
John Kennedy
2013-10-15 20:07:25 UTC
Permalink
Post by Yuri Pankov
Post by Christopher Siden
http://cr.illumos.org/~webrev/csiden/illumos-zfs-tests/
Work by John Kennedy. This ports the rest of the ZFS test suite from
STF (in a different repository on github) to the testrunner framework
in illumos-gate.
Looks good, though I haven't looked through all added tests, only at the
infrastructure changes (thanks for fixing my typo).
One question - are we really interested in *physical* CPUs in is_mp()
function? The tests where it's used don't give any hints why.
No, not really. I simply copied the usage that existed in several other
places to a function without really scrutinizing it. I've filed a follow on
bug to address this.
--
John Wren Kennedy

275 Middlefield Road, Suite 50
Menlo Park, CA 94025
http://www.delphix.com/
Yuri Pankov
2013-10-23 14:42:09 UTC
Permalink
Post by John Kennedy
Post by Yuri Pankov
Post by Christopher Siden
http://cr.illumos.org/~webrev/csiden/illumos-zfs-tests/
Work by John Kennedy. This ports the rest of the ZFS test suite from
STF (in a different repository on github) to the testrunner framework
in illumos-gate.
Looks good, though I haven't looked through all added tests, only at the
infrastructure changes (thanks for fixing my typo).
One question - are we really interested in *physical* CPUs in is_mp()
function? The tests where it's used don't give any hints why.
No, not really. I simply copied the usage that existed in several other
places to a function without really scrutinizing it. I've filed a follow on
bug to address this.
Thanks. Given that it's not something critical, can we get the tests in?
Loading...