Re: [PATCH v1] acpi/arm: relax MADT GICC entry length check to support newer ACPI revisions

[Julien Grall <julien@xxxxxxx>]

Hi Oleksii,

On 22/12/2025 11:30, Oleksii Kurochko wrote:
> > > diff --git a/xen/arch/arm/gic.c b/xen/arch/arm/gic.c
> > > index ee75258fc3..a0ccda14bf 100644
> > > --- a/xen/arch/arm/gic.c
> > > +++ b/xen/arch/arm/gic.c
> > > @@ -418,8 +418,18 @@ unsigned long gic_get_hwdom_madt_size(const 
> > > struct domain *d)
> > >   {
> > >       unsigned long madt_size;
> > >   +    struct acpi_subtable_header *header;
> > > +    struct acpi_madt_generic_interrupt *host_gicc;
> > For both variables, you don't seem to modify the content. So I think 
> > they should be const.
> > > +
> > > +    header = 
> > > acpi_table_get_entry_madt(ACPI_MADT_TYPE_GENERIC_INTERRUPT, 0);
> > > +    if ( !header )
> > > +        panic("Can't get GICC entry");
> > I would feel happier to use panic() in this function if 
> > gic_get_hwdom_madt_size() is __init (its only caller is during boot)).
> I am okay with making gic_get_hwdom_madt_size() as __init.
>
> > An alternative is to stash the GICC size in a global variable.
> We also could to return 0 in the case of ( !header ) and then check if 
> the size is zero or not in estimate_acpi_efi_size(),
> and if it is zero then just return -EINVAL. Does it make sense?
This would also work for me. I will let you chose between the two 
approaches.
Cheers,

--
Julien Grall