diff options
author | Kevin Wolf <kwolf@redhat.com> | 2014-09-30 13:27:09 +0200 |
---|---|---|
committer | Stefan Hajnoczi <stefanha@redhat.com> | 2014-10-03 10:30:33 +0100 |
commit | d9323e9b20b1a74720f17e81387cffe013d9cf0b (patch) | |
tree | 35cb637dbfc5e917a5aa858e42ac65b7a82ac8e6 /COPYING | |
parent | 20d6cd47d0459bdc6a8b64e7abe8f713e0d4718e (diff) |
make check-block: Use default cache modes
When qemu-iotests only gave a choice between cache=none and
cache=writethrough, we picked cache=none because it was the option that
would complete the test in finite time. Some tests could only work for
one of the two options and would be skipped with cache=none, but that
was an acceptable trade-off at the time.
Today, however, qemu-iotests is a bit more flexible than that and you
can specify any of the cache modes supported by qemu. The default is
writeback, like in qemu, which is fast and (unlike cache=none) compatible
with any host filesystem. Test cases that have specific requirements for
the cache mode can also specify a different default.
In order to get a fast test run that works everywhere and doesn't skip
tests that need a different cache mode, not specifying any cache mode
and instead relying on the default is the best we can do today.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Message-id: 1412076430-11623-2-git-send-email-kwolf@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Diffstat (limited to 'COPYING')
0 files changed, 0 insertions, 0 deletions