On Tue, Jul 25, 2017 at 04:45:46PM +0100, Stefan Hajnoczi wrote: > On Mon, Jul 24, 2017 at 02:44:13PM +0800, Jing Liu wrote: > > On 2017/7/21 上午11:47, Cleber Rosa wrote: > > > One of the "sub-"tests of test 087 requires CONFIG_LINUX_AIO. > > > > > > As a PoC/RFC, this goes the easy route and skips the test as a whole > > > when that feature is missing. Other approaches include splitting > > > the test and adding extra filtering. > > > > > > Signed-off-by: Cleber Rosa <cr...@redhat.com> > > > --- > > > tests/qemu-iotests/087 | 1 + > > > 1 file changed, 1 insertion(+) > > > > > > diff --git a/tests/qemu-iotests/087 b/tests/qemu-iotests/087 > > > index f8e4903..a2fb7de 100755 > > > --- a/tests/qemu-iotests/087 > > > +++ b/tests/qemu-iotests/087 > > > @@ -34,6 +34,7 @@ status=1 # failure is the default! > > > _supported_fmt qcow2 > > > _supported_proto file > > > _supported_os Linux > > > +_require_feature CONFIG_LINUX_AIO > > I tested that CONFIG_NETTLE_KDF is also a necessary for 087. > > > > +_require_feature CONFIG_NETTLE_KDF > > Are you sure? Looks like either nettle or gcrypt is needed:
Correct, it works with either. > crypto/Makefile.objs:crypto-obj-$(CONFIG_NETTLE_KDF) += pbkdf-nettle.o > crypto/Makefile.objs:crypto-obj-$(if > $(CONFIG_NETTLE_KDF),n,$(CONFIG_GCRYPT_KDF)) += pbkdf-gcrypt.o > > But this shows why the compile-time testing of features is ugly: > > 1. It duplicates build dependency logic into the test cases. > > 2. The test cases don't care about nettle vs gcrypt and they shouldn't > have to know about it. They just care whether LUKS is available or > not. Yeah that knowledge is messy and fragile wrt future changes. Regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|