On 29.05.2020 13:11, Hongyan Xia wrote:
> From: Wei Liu <wei.l...@citrix.com>
>
> Avoid repetitive mapping of l2_ro_mpt by keeping it across loops, and
> only unmap and map it when crossing 1G boundaries.
>
> Signed-off-by: Wei Liu <wei.l...@citrix.com>
> Signed-off-by: Hongyan Xia <hongy...@amazon.com>
Reviewed-by: Jan Beulich <jbeul...@suse.com>
I do think, however, that ...
> @@ -438,32 +443,29 @@ static int setup_m2p_table(struct mem_hotadd_info *info)
>
> ASSERT(!(l3e_get_flags(l3_ro_mpt[l3_table_offset(va)]) &
> _PAGE_PSE));
> - if ( l3e_get_flags(l3_ro_mpt[l3_table_offset(va)]) &
> - _PAGE_PRESENT )
> - l2_ro_mpt = l3e_to_l2e(l3_ro_mpt[l3_table_offset(va)]) +
> - l2_table_offset(va);
> - else
> + if ( (l3e_get_flags(l3_ro_mpt[l3_table_offset(va)]) &
> + _PAGE_PRESENT) && !l2_ro_mpt)
> + l2_ro_mpt = map_l2t_from_l3e(l3_ro_mpt[l3_table_offset(va)]);
> + else if ( !l2_ro_mpt )
> {
... this would be slightly neater as
if ( l2_ro_mpt )
/* nothing */;
else if ( l3e_get_flags(l3_ro_mpt[l3_table_offset(va)]) &
_PAGE_PRESENT )
l2_ro_mpt = map_l2t_from_l3e(l3_ro_mpt[l3_table_offset(va)]);
else
{
...
My R-b holds if you would change it like this.
Jan
