On Tue, 2021-05-11 at 13:05 +0300, Mike Rapoport wrote:
> From: Mike Rapoport <rppt@linux.ibm.com>
> 
> The struct pages representing a reserved memory region are initialized
> using reserve_bootmem_range() function. This function is called for each
> reserved region just before the memory is freed from memblock to the buddy
> page allocator.
> 
> The struct pages for MEMBLOCK_NOMAP regions are kept with the default
> values set by the memory map initialization which makes it necessary to
> have a special treatment for such pages in pfn_valid() and
> pfn_valid_within().
> 
> Split out initialization of the reserved pages to a function with a
> meaningful name and treat the MEMBLOCK_NOMAP regions the same way as the
> reserved regions and mark struct pages for the NOMAP regions as
> PageReserved.
> 
> Signed-off-by: Mike Rapoport <rppt@linux.ibm.com>
> Reviewed-by: David Hildenbrand <david@redhat.com>
> Reviewed-by: Anshuman Khandual <anshuman.khandual@arm.com>
> ---
>  include/linux/memblock.h |  4 +++-
>  mm/memblock.c            | 28 ++++++++++++++++++++++++++--
>  2 files changed, 29 insertions(+), 3 deletions(-)
> 
> diff --git a/include/linux/memblock.h b/include/linux/memblock.h
> index 5984fff3f175..1b4c97c151ae 100644
> --- a/include/linux/memblock.h
> +++ b/include/linux/memblock.h
> @@ -30,7 +30,9 @@ extern unsigned long long max_possible_pfn;
>   * @MEMBLOCK_NONE: no special request
>   * @MEMBLOCK_HOTPLUG: hotpluggable region
>   * @MEMBLOCK_MIRROR: mirrored region
> - * @MEMBLOCK_NOMAP: don't add to kernel direct mapping
> + * @MEMBLOCK_NOMAP: don't add to kernel direct mapping and treat as
> + * reserved in the memory map; refer to memblock_mark_nomap() description
> + * for further details
>   */
>  enum memblock_flags {
>  	MEMBLOCK_NONE		= 0x0,	/* No special request */
> diff --git a/mm/memblock.c b/mm/memblock.c
> index afaefa8fc6ab..3abf2c3fea7f 100644
> --- a/mm/memblock.c
> +++ b/mm/memblock.c
> @@ -906,6 +906,11 @@ int __init_memblock memblock_mark_mirror(phys_addr_t base, phys_addr_t size)
>   * @base: the base phys addr of the region
>   * @size: the size of the region
>   *
> + * The memory regions marked with %MEMBLOCK_NOMAP will not be added to the
> + * direct mapping of the physical memory. These regions will still be
> + * covered by the memory map. The struct page representing NOMAP memory
> + * frames in the memory map will be PageReserved()
> + *
>   * Return: 0 on success, -errno on failure.
>   */
>  int __init_memblock memblock_mark_nomap(phys_addr_t base, phys_addr_t size)
> @@ -2002,6 +2007,26 @@ static unsigned long __init __free_memory_core(phys_addr_t start,
>  	return end_pfn - start_pfn;
>  }
>  
> +static void __init memmap_init_reserved_pages(void)
> +{
> +	struct memblock_region *region;
> +	phys_addr_t start, end;
> +	u64 i;
> +
> +	/* initialize struct pages for the reserved regions */
> +	for_each_reserved_mem_range(i, &start, &end)
> +		reserve_bootmem_region(start, end);
> +
> +	/* and also treat struct pages for the NOMAP regions as PageReserved */
> +	for_each_mem_region(region) {
> +		if (memblock_is_nomap(region)) {
> +			start = region->base;
> +			end = start + region->size;
> +			reserve_bootmem_region(start, end);
> +		}
> +	}
> +}
> +

In some cases, that whole call to reserve_bootmem_region() may be a no-
op because pfn_valid() is not true for *any* address in that range.

But reserve_bootmem_region() spends a long time iterating of them all,
and eventually doing nothing:

void __meminit reserve_bootmem_region(phys_addr_t start,
                                      phys_addr_t end, int nid)
{
        unsigned long start_pfn = PFN_DOWN(start);
        unsigned long end_pfn = PFN_UP(end);

        for (; start_pfn < end_pfn; start_pfn++) {
                if (pfn_valid(start_pfn)) {
                        struct page *page = pfn_to_page(start_pfn);

                        init_reserved_page(start_pfn, nid);

                        /*
                         * no need for atomic set_bit because the struct
                         * page is not visible yet so nobody should
                         * access it yet.
                         */
                        __SetPageReserved(page);
                }
        }
}

On platforms with large NOMAP regions (e.g. which are actually reserved
for guest memory to keep it out of the Linux address map and allow for
kexec-based live update of the hypervisor), this pointless loop ends up
taking a significant amount of time which is visible as guest steal
time during the live update.

Can reserve_bootmem_region() skip the loop *completely* if no PFN in
the range from start to end is valid? Or tweak the loop itself to have
an 'else' case which skips to the next valid PFN? Something like

 for(...) {
    if (pfn_valid(start_pfn)) {
       ...
    } else {
       start_pfn = next_valid_pfn(start_pfn);
    }
 }