Re: [PATCH v6 0/9] target/arm/kvm: enable SVE in guests

[Peter Maydell]

On Mon, 21 Oct 2019 at 17:12, Andrew Jones <address@hidden> wrote:
>
> On Mon, Oct 21, 2019 at 04:43:22PM +0100, Peter Maydell wrote:
> > On Mon, 21 Oct 2019 at 15:23, Andrew Jones <address@hidden> wrote:
> > > Peter, would you mind running your test on the kvm32 machine with this
> > > change before I send a v7?
> >
> > Still fails:
> >
> > pm215@pm-ct:~/qemu/build/arm$
> > QTEST_QEMU_BINARY=arm-softmmu/qemu-system-arm tests/arm-cpu-features
> > /arm/arm/query-cpu-model-expansion: OK
> > /arm/arm/kvm/query-cpu-model-expansion: **
> > ERROR:/home/pm215/qemu/tests/arm-cpu-features.c:498:test_query_cpu_model_expansion_kvm:
> > assertion failed: (resp_has_props(_resp))
> > Aborted
> >
> > This is asserting on the line:
> > 498             assert_has_not_feature(qts, "host", "sve");
> >
>
> Oh, I see. It's not failing the specific absence of 'sve', but the test
> code (assert_has_not_feature()) is assuming at least one property is
> present. This isn't the case for kvm32 'host' cpus. They apparently
> have none. We need this patch too, then
>
> diff --git a/tests/arm-cpu-features.c b/tests/arm-cpu-features.c
> index 14100ebd8521..9aa728ed8469 100644
> --- a/tests/arm-cpu-features.c
> +++ b/tests/arm-cpu-features.c
> @@ -136,8 +136,8 @@ static bool resp_get_feature(QDict *resp, const char 
> *feature)
>  ({                                                                     \
>      QDict *_resp = do_query_no_props(qts, cpu_type);                   \
>      g_assert(_resp);                                                   \
> -    g_assert(resp_has_props(_resp));                                   \
> -    g_assert(!qdict_get(resp_get_props(_resp), feature));              \
> +    g_assert(!resp_has_props(_resp) ||                                 \
> +             !qdict_get(resp_get_props(_resp), feature));              \
>      qobject_unref(_resp);                                              \
>  })

Yep, with that extra the test passes. I'm just rerunning the
full 'make check'...

thanks
-- PMM