On 5 December 2014 at 15:33, Greg Bellows <greg.bell...@linaro.org> wrote: > > > On 5 December 2014 at 09:18, Peter Maydell <peter.mayd...@linaro.org> wrote: >> >> On 3 December 2014 at 20:05, Greg Bellows <greg.bell...@linaro.org> wrote: >> > Added 'secure' qemu boolean option to qemu_machine_opts[]. >> > >> > Signed-off-by: Greg Bellows <greg.bell...@linaro.org> >> > --- >> > vl.c | 4 ++++ >> > 1 file changed, 4 insertions(+) >> > >> > diff --git a/vl.c b/vl.c >> > index eb89d62..5d640f7 100644 >> > --- a/vl.c >> > +++ b/vl.c >> > @@ -387,6 +387,10 @@ static QemuOptsList qemu_machine_opts = { >> > .name = "iommu", >> > .type = QEMU_OPT_BOOL, >> > .help = "Set on/off to enable/disable Intel IOMMU (VT-d)", >> > + },{ >> > + .name = "secure", >> > + .type = QEMU_OPT_BOOL, >> > + .help = "Set on/off to enable/disable secure state", >> > }, >> >> If patch 5 adds 'secure' as a machine property for only those >> boards where it makes sense, why do we need this global switch? >> > > That is what I thought as well, but this is apparently needed as we get an > invalid machine property otherwise. Below is the error, I'll look again, > but it appeared all machine properties were defined here. > > qemu-system-aarch64: -machine type=vexpress-a15,secure=off: Invalid > parameter 'secure'
That would seem to defeat the point of the machine opts design, so it looks a bit strange. Marcel: how is this supposed to work for board-specific -machine options? thanks -- PMM