On 08/27/2011 01:07 PM, Edgar E. Iglesias wrote:
> >+
> >+static ISADeviceInfo testdev_info = {
> >+ .qdev.name = "testdev",
> >+ .qdev.size = sizeof(struct testdev),
> >+ .init = init_test_device,
> >+ .qdev.props = (Property[]) {
> >+ DEFINE_PROP_CHR("chardev", struct testdev, chr),
> >+ DEFINE_PROP_END_OF_LIST(),
> >+ },
> >+};
>
> Should this use MemoryRegion?
Yes. And what is the reason for using IO ports?
Mostly for ease of use. The tests were originally run under a separate
kvm userspace that didn't emulate a full machine.
There are archs that dont have ioport connections out from the CPU.
If we are adding virtual devices for tests, they should preferably work for
all archs.
I think all the functionallity here (apart from that which Anthony
pointed out is available by other means) is x86 specific.
--
I have a truly marvellous patch that fixes the bug which this
signature is too narrow to contain.