Re: [Qemu-devel] [PATCH v1 14/14] scripts/qemu.py: allow arches use KVM for their 32bit cousins

["Alex Bennée" <alex.bennee@linaro.org>]

Good point. I should fall through.

On Sat, 26 Jan 2019, 03:37 Eduardo Habkost <ehabkost@redhat.com wrote:

> On Fri, Jan 25, 2019 at 02:00:17PM +0000, Alex Bennée wrote:
> > A lot of architectures can run their 32 bit cousins on KVM so the
> > kvm_available function needs to be a little less restricting when
> > deciding if KVM is available.
> >
> > Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
> > ---
> >  scripts/qemu.py | 11 ++++++++++-
> >  1 file changed, 10 insertions(+), 1 deletion(-)
> >
> > diff --git a/scripts/qemu.py b/scripts/qemu.py
> > index 0a5e02eb56..2934ee12c5 100644
> > --- a/scripts/qemu.py
> > +++ b/scripts/qemu.py
> > @@ -25,9 +25,18 @@ import tempfile
> >
> >  LOG = logging.getLogger(__name__)
> >
> > +# Mapping host architecture to any additional architectures it can
> > +# support which often includes its 32 bit cousin.
> > +ADDITIONAL_ARCHES = {
> > +    "x86_64" : "i386",
> > +    "aarch64" : "armhf"
> > +}
> >
> >  def kvm_available(target_arch=None):
> > -    if target_arch and target_arch != os.uname()[4]:
> > +    host_arch = os.uname()[4]
> > +    if target_arch and target_arch != host_arch:
> > +        if target_arch == ADDITIONAL_ARCHES[host_arch]:
>
> This will crash host_arch isn't "x86_64" or "aarch64".  I suggest
> ADDITIONAL_ARCHES.get(host_arch)
>
> > +            return True
>
> I don't think we should skip the /dev/kvm check here.
>
>
> >          return False
> >      return os.access("/dev/kvm", os.R_OK | os.W_OK)
> >
> > --
> > 2.17.1
> >
>
> --
> Eduardo
>