Re: [RFC v19 10/15] i386: move TCG btp_helper into softmmu/

From: Paolo Bonzini <pbonzini@redhat.com>
To: Claudio Fontana <cfontana@suse.de>
Cc: qemu-devel <qemu-devel@nongnu.org>
Subject: Re: [RFC v19 10/15] i386: move TCG btp_helper into softmmu/
Date: Tue, 16 Feb 2021 18:36:02 +0100	[thread overview]
Message-ID: <CABgObfYxzZ2h9+z8YnbVOKc6irVirE7JODz92ANJ7XYt2pROTw@mail.gmail.com> (raw)
In-Reply-To: <633d0100-73aa-7129-e636-31174140ac62@suse.de>

> >  {
> >      if (s->flags & HF_IOBPT_MASK) {
> > +#ifdef CONFIG_USER_ONLY
> > +        /* user-mode cpu should not be in IOBPT mode */
> > +        g_assert_not_reached();
> > +#else
> >          TCGv_i32 t_size = tcg_const_i32(1 << ot);
> >          TCGv t_next = tcg_const_tl(s->pc - s->cs_base);
> >
> >          gen_helper_bpt_io(cpu_env, t_port, t_size, t_next);
> >          tcg_temp_free_i32(t_size);
> >          tcg_temp_free(t_next);
> > +#endif /* CONFIG_USER_ONLY */
> >      }
> >  }
>
> Just highlighting this for your convenience Paolo,
> as I am just flailing in the dark here looking for similar patterns.
>
> Is it the case that we should never be in the HF_IOBPT mode when building for CONFIG_USER_ONLY ?

Yes, pretty much. Maybe worth adding a qemu_assert_softmmu() (return
true for softmmu, abort for user-only), like

if ((s->flags & HF_IOBPT_MASK) && qemu_assert_softmmu()) {
}

Slightly weird API, maybe, but something along those lines.

Paolo