Peter Maydell <peter.mayd...@linaro.org> 于2020年6月25日周四 下午8:22写道： > > On Thu, 25 Jun 2020 at 10:18, P J P <ppan...@redhat.com> wrote: > > > > +-- On Thu, 25 Jun 2020, Philippe Mathieu-Daudé wrote --+ > > | > @@ -196,7 +196,22 @@ static bool tz_ppc_dummy_accepts(void *opaque, > > hwaddr addr, > > | > g_assert_not_reached(); > > | > > | This is a shame we now have to fill the read/write handlers for > > | unreachable code :( > > | > > | > +static uint64_t tz_ppc_dummy_read(void *opaque, hwaddr addr, unsigned > > size) > > | > > | I'd instead use a clearer: > > | g_assert_not_reached(); > > | > > | > +static void tz_ppc_dummy_write(void *opaque, hwaddr addr, > > | > > | Ditto: > > | g_assert_not_reached(); > > > > This will likely be called in tz_ppc_dummy_accepts() above. Do we still want > > to revise this patch? considering read/write callbacks are unreachable. > > The point of g_assert_not_reached() is that it documents and > asserts that the code is not reachable. If the read and write > callbacks are unreachable (which they are) then having their > bodies just be a call to g_assert_not_reached() is the best > way to document that.
I agree with this. Thanks, Li Qiang > > thanks > -- PMM
