Re: [PATCH v2 3/7] Mini-OS: add support for runtime mounts

[Samuel Thibault <samuel.thibault@ens-lyon.org>]

Juergen Gross, le ven. 10 févr. 2023 13:14:20 +0100, a ecrit:
> On 10.02.23 12:43, Andrew Cooper wrote:
> > On 10/02/2023 10:46 am, Juergen Gross wrote:
> > > diff --git a/lib/sys.c b/lib/sys.c
> > > index 2f33c937..dc8a8c69 100644
> > > --- a/lib/sys.c
> > > +++ b/lib/sys.c
> > > @@ -339,7 +339,14 @@ static int open_mem(struct mount_point *mnt, const char *pathname, int flags,
> > >       return fd;
> > >   }
> > > -static struct mount_point mount_points[] = {
> > > +#ifdef CONFIG_CONSFRONT
> > > +#define STATIC_MNTS   4
> > > +#else
> > > +#define STATIC_MNTS   2
> > > +#endif
> > 
> > This ought to be
> > 
> > #define STATIC_MNTS (2 + (IS_ENABLED(CONFIG_CONSFRONT) * 2))
> > 
> > because it shows where the parts come from, and is much cleaner to add a
> > 3rd one to in due course.
> 
> In principle fine, but I'm not sure it is worth the effort to add
> IS_ENABLED() support to Mini-OS.
> 
> > That said, it would be simpler to just have a total mounts set at 16 or
> > so?  Does a difference of two dynamic mounts depending on CONSFRONT
> > actually matter?
> 
> Probably not. Samuel, any thoughts on that?

We can probably as well just have 16 mount entries indeed.

Samuel