On Mon, Dec 09, 2013 at 10:16:38AM +0100, Markus Armbruster wrote: > Amos Kong <ak...@redhat.com> writes: > > > The type of "strict" should be bool. > > > > Signed-off-by: Amos Kong <ak...@redhat.com> > > --- > > vl.c | 2 +- > > 1 files changed, 1 insertions(+), 1 deletions(-) > > > > diff --git a/vl.c b/vl.c > > index b0399de..12fe1f7 100644 > > --- a/vl.c > > +++ b/vl.c > > @@ -461,7 +461,7 @@ static QemuOptsList qemu_boot_opts = { > > .type = QEMU_OPT_STRING, > > }, { > > .name = "strict", > > - .type = QEMU_OPT_STRING, > > + .type = QEMU_OPT_BOOL, > > }, > > { /*End of list */ } > > }, > > No, this isn't just a typo fix, it's an interface change. It's not > immediately obvious whether it's okay. You are correct, it's not a typo, menu's type was also changed from STRING to BOOL. We can just process strict as menu.
> -boot parameter 'strict' appeared in commit c8a6ae8. It was declared > QEMU_OPT_STRING, but the code using the value rejected anything but 'on' > and 'off'. This shipped in 1.5.0. > > Commit 6ef4716 accidentally dropped the code using the value of > 'strict'. Since then, any parameter value is accepted and silently > ignored. Shipped in 1.6.0. > > Your change restores the "reject anything but 'on' and 'off'" part (the > "use the value" part gets restored in the next patch). Okay, because > the values erroneously accepted in the meantime didn't do anything. Thanks your explanation :) -- Amos.