Re: [PATCH v5] xen/consoled: clean up console handling for PV shim

[Denis Mukhin <dmkhn@xxxxxxxxx>]

On Thursday, March 6th, 2025 at 8:35 AM, Jan Beulich <jbeulich@xxxxxxxx> wrote:

> 
> 
> On 06.03.2025 08:59, dmkhn@xxxxxxxxx wrote:> --- a/xen/include/xen/consoled.h
> 
> > +++ b/xen/include/xen/consoled.h
> > @@ -1,12 +1,23 @@
> > +/* SPDX-License-Identifier: GPL-2.0-only */
> > #ifndef XEN_CONSOLED_H
> > #define XEN_CONSOLED_H
> > 
> > #include <public/io/console.h>
> > 
> > +int consoled_guest_tx(char c);
> > +
> > +#ifdef CONFIG_PV_SHIM
> > +
> > void consoled_set_ring_addr(struct xencons_interface *ring);
> > struct xencons_interface consoled_get_ring_addr(void);
> > -void consoled_guest_rx(void);
> > -void consoled_guest_tx(char c);
> > +int consoled_guest_rx(void);
> > +bool consoled_is_enabled(void);
> > +
> > +#else
> > +
> > +#define consoled_is_enabled() (false)
> > +
> > +#endif / CONFIG_PV_SHIM */
> 
> 
> I'm sorry to be picky, but why did you move just consoled_guest_tx() out
> of the conditional? Once something needs moving out, imo everything that
> doesn't strictly need to be there wants to move out as well. Which would
> leave just consoled_is_enabled() there. Then
> Acked-by: Jan Beulich jbeulich@xxxxxxxx
> 
> I can certainly make the adjustment while committing, if no other need
> for a v6 arises.

I will appreciate help!
Thanks

> 
> Jan