> > > +/*
> > > + * Get the number of page-tables needed to replace a block with a
> > > + * fully populated tree, up to the PTE level, at particular level.
> > > + */
> > > +static inline int stage2_block_get_nr_page_tables(u32 level)
> > > +{
> > > + if (WARN_ON_ONCE(level < KVM_PGTABLE_MIN_BLOCK_LEVEL ||
> > > + level >= KVM_PGTABLE_MAX_LEVELS))
> > > + return -EINVAL;
> > > +
> > > + switch (level) {
> > > + case 1:
> > > + return PTRS_PER_PTE + 1;
> > > + case 2:
> > > + return 1;
> > > + case 3:
> > > + return 0;
> > > + default:
> > > + return -EINVAL;
> > > + };
> > > +}
> >
> > Shaoqin,
> >
> > "Is the level 3 check really needed?"
> >
> > Regarding your question about the need for "case 3". You are right,
> > it's not actually needed in this particular case (when called from
> > stage2_split_walker()). However, it would be nice to reuse this
> > function and so it should cover all functions.
> >
> > Thanks,
> > Ricardo
> >
>
> Hi Ricardo,
>
> Thanks for your explanation. And I'm wondering since
> stage2_block_get_nr_page_tables() is aimed to get how many pages needed to
> populate a new pgtable and replace the block mapping, and level 3 is the
> PTE, it can't be split. So is it still ok to put level 3 in this function,
> although put it here has no effect.
>
> Shaoqin
Hi,
The "level 3" check is not used by kvm_pgtable_stage2_split(), but
stage2_block_get_nr_page_tables(level=3) is not an invalid (EINVAL) call
either.
Sorry for the late reply.
Thanks,
Ricardo
