Re: [PATCH 2/2] mm/hugetlb: optimize bootmem hugepage tail vmemmap init

["Li Zhe" <lizhe.67@bytedance.com>]

On 11/9/26 11:01 AM, Li Zhe wrote:
>
> On 11/6/26 6:23 PM, Balbir Singh wrote:
>> On Thu, Nov 05, 2026 at 08:06:15PM +0800, Li Zhe wrote:
>>> Boot-time HugeTLB allocation has to initialize tail struct pages itself
>>> when HugeTLB vmemmap optimization (HVO) is unavailable or fails.  For
>>> gigantic pages, this can cover a large number of tail pages. Most state
>>> installed by __init_single_page() and prep_compound_tail() is identical
>>> for every tail page in the same huge page; only PFN-dependent fields
>>> need refreshing.
>>>
>>> Initialize the first tail page normally, copy it into a local template,
>>> then initialize the remaining tails from that template with
>>> memmap_init_page_from_template().  This avoids repeatedly doing the 
>>> full
>>> __init_single_page() and prep_compound_tail() sequence for state 
>>> that is
>>> identical across tail pages in the same huge page.
>>>
>>> On an arm64 QEMU/KVM guest using 4K pages and 64 boot-time 1G hugepages
>>> (HVO unavailable on arm64 in v7.4-rc1), the measured fallback tail
>>> initialization time improved by 36.43%, measured with a temporary
>>> jiffies-based printk modeled after the existing timing in
>>> memmap_init_zone_device():
>>>
>>>    base:             318.4 ms
>>>    with this patch:  202.4 ms
>>>
>>> Signed-off-by: Li Zhe <lizhe.67@bytedance.com>
>>> ---
>>>   mm/hugetlb.c | 23 ++++++++++++++++++-----
>>>   1 file changed, 18 insertions(+), 5 deletions(-)
>>>
>>> diff --git a/mm/hugetlb.c b/mm/hugetlb.c
>>> index da5d04edf7c1..987f16a5ea27 100644
>>> --- a/mm/hugetlb.c
>>> +++ b/mm/hugetlb.c
>>> @@ -3219,19 +3219,32 @@ static void __init 
>>> hugetlb_folio_init_tail_vmemmap(struct folio *folio,
>>>       enum zone_type zone = folio_zonenum(folio);
>>>       int nid = folio_nid(folio);
>>>       struct page *page = folio_page(folio, start_page_number);
>>> +    struct page template;
>>>       unsigned long head_pfn = folio_pfn(folio);
>>>       unsigned long pfn, end_pfn = head_pfn + end_page_number;
>>>       unsigned int order = huge_page_order(h);
>>>   +    if (start_page_number >= end_page_number)
>>> +        return;
>>> +
>> Can this actually happen?
>
>
> No, current callers always pass a non-empty range. I kept this to
> preserve the old for-loop no-op semantics for an empty range, but it is
> not needed here. I will remove it in v2.


I need to correct this. A local Sashiko run pointed out that the bootmem
fallback caller can pass an empty range on SuperH.

SuperH supports a 64M HugeTLB page size, and with 64K base pages
pages_per_huge_page(h) is 1024. HUGETLB_VMEMMAP_RESERVE_PAGES is
PAGE_SIZE / sizeof(struct page), which can also be 1024 or larger on a
32-bit SuperH configuration.

In that case the old for loop is a no-op. So I will keep the empty-range
check in v2, before creating the seed template.

Thanks,
Zhe

>
>>
>>>       /*
>>>        * As we marked all tail pages with 
>>> memblock_reserved_mark_noinit(),
>>>        * we must initialize them ourselves here.
>>>        */
>>> -    for (pfn = head_pfn + start_page_number; pfn < end_pfn; page++, 
>>> pfn++) {
>>> -        __init_single_page(page, pfn, zone, nid);
>>> -        prep_compound_tail(page, &folio->page, order);
>>> -        set_page_count(page, 0);
>>> -    }
>>> +    pfn = head_pfn + start_page_number;
>>> +
>>> +    /*
>>> +     * Seed a template from the first real tail page and refresh only
>>> +     * PFN-dependent fields before copying it.
>>> +     */
>>> +    __init_single_page(page, pfn, zone, nid);
>>> +    prep_compound_tail(page, &folio->page, order);
>>> +    set_page_count(page, 0);
>>> +    memcpy(&template, page, sizeof(template));
>> Do we need a separate template and page?
> Yes. page is the first tail page that has already been initialized
> normally. A separate template is needed so that later iterations can
> refresh the PFN-dependent fields before memcpy_nontemporal() copies the
> final state to each destination page.
>
> Thanks for the review and the Ack.
>
> Thanks,
> Zhe
>
>>
>>> +
>>> +    for (pfn++; pfn < end_pfn; pfn++)
>>> +        memmap_init_page_from_template(pfn_to_page(pfn), pfn, zone,
>>> +                           &template);
>>>   }
>>>     static void __init hugetlb_folio_init_vmemmap(struct folio *folio,
>> Acked-by: Balbir Singh <balbirs@nvidia.com>