[Libosinfo] [PATCH libosinfo v2] test-isodetect: add Debian's arm flavours

[Guido Günther]

On Thu, Jul 13, 2017 at 04:35:43PM +0100, Daniel P. Berrange wrote:
> On Thu, Jul 13, 2017 at 02:39:47PM +0200, Guido Günther wrote:
> > to avoid "Unknown arch" debug messages
> > ---
> >  tests/test-isodetect.c | 8 +++++++-
> >  1 file changed, 7 insertions(+), 1 deletion(-)
> > 
> > diff --git a/tests/test-isodetect.c b/tests/test-isodetect.c
> > index c72b8fc..9306be4 100644
> > --- a/tests/test-isodetect.c
> > +++ b/tests/test-isodetect.c
> > @@ -128,8 +128,14 @@ static struct ISOInfo *load_iso(GFile *file, const gchar *shortid, const gchar *
> >          arch = "ia64";
> >      else if (strstr(name, "s390x"))
> >          arch = "s390x";
> > -    else if (strstr(name, "aarch64"))
> > +    else if (strstr(name, "aarch64") ||
> > +             strstr(name, "arm64"))
> >          arch = "aarch64";
> > +    else if (strstr(name, "armv7l"))
> > +        arch = "armv7l";
> > +    else if (strstr(name, "armhf") ||
> > +             strstr(name, "armv7hl"))
> > +        arch = "armv7hl";
> 
> These should all just be normalizing to 'armv7l' - ie arm7 little endian. No
> need for a hard vs soft float distinction here, that's just a CPU sub-feature
> not a different architecture.

That makes sense if armv7l mandates VFP. I couldn't figure that out
reliably yesterday but looking at [1] and [2] it seems so. It would be
be similar to mips running o32, n32 or n64 and all ABIs are supported by
the same CPU. v3 forthcoming.
Thanks for the clarification,
 -- Guido

[1]: https://en.wikipedia.org/wiki/Comparison_of_ARMv7-A_cores
[2]: http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.dui0471k/pge1358786964554.html