On 28.10.2024 10:18, Andrew Cooper wrote: > There's no need for microcode_init_cache() to be called exactly where it is in > __start_xen(). All that matters is it must be after xmalloc() is available > and before APs start up. > > As a consequence, microcode_init_cache() runs a little later on boot now. > > No functional change. > > Signed-off-by: Andrew Cooper <[email protected]>
Acked-by: Jan Beulich <[email protected]> > --- a/xen/arch/x86/cpu/microcode/core.c > +++ b/xen/arch/x86/cpu/microcode/core.c > @@ -819,8 +819,9 @@ static int __init early_update_cache(const void *data, > size_t len) > return rc; > } > > -int __init microcode_init_cache(struct boot_info *bi) > +static int __init cf_check microcode_init_cache(void) > { > + struct boot_info *bi = &xen_boot_info; I take it that this not becoming pointer-to-const is related to your reply to Roger elsewhere, where he asked for adding const, while you expect the const to need dropping again in the course of Daniel's work? Jan
