[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <24ed8a1f-2134-6082-7bc9-f8662dc723bd@bytedance.com>
Date: Wed, 2 Aug 2023 11:06:49 +0100
From: Usama Arif <usama.arif@...edance.com>
To: Muchun Song <muchun.song@...ux.dev>
Cc: linux-kernel@...r.kernel.org, fam.zheng@...edance.com,
liangma@...ngbit.com, simon.evans@...edance.com,
punit.agrawal@...edance.com, linux-mm@...ck.org,
mike.kravetz@...cle.com, rppt@...nel.org
Subject: Re: [External] Re: [v2 1/6] mm: hugetlb: Skip prep of tail pages when
HVO is enabled
On 01/08/2023 03:04, Muchun Song wrote:
>
>
> On 2023/7/30 23:16, Usama Arif wrote:
>> When vmemmap is optimizable, it will free all the duplicated tail
>> pages in hugetlb_vmemmap_optimize while preparing the new hugepage.
>> Hence, there is no need to prepare them.
>>
>> For 1G x86 hugepages, it avoids preparing
>> 262144 - 64 = 262080 struct pages per hugepage.
>>
>> The indirection of using __prep_compound_gigantic_folio is also removed,
>> as it just creates extra functions to indicate demote which can be done
>> with the argument.
>>
>> Signed-off-by: Usama Arif <usama.arif@...edance.com>
>> ---
>> mm/hugetlb.c | 32 ++++++++++++++------------------
>> mm/hugetlb_vmemmap.c | 2 +-
>> mm/hugetlb_vmemmap.h | 15 +++++++++++----
>> 3 files changed, 26 insertions(+), 23 deletions(-)
>>
>> diff --git a/mm/hugetlb.c b/mm/hugetlb.c
>> index 64a3239b6407..541c07b6d60f 100644
>> --- a/mm/hugetlb.c
>> +++ b/mm/hugetlb.c
>> @@ -1942,14 +1942,23 @@ static void prep_new_hugetlb_folio(struct
>> hstate *h, struct folio *folio, int ni
>> spin_unlock_irq(&hugetlb_lock);
>> }
>> -static bool __prep_compound_gigantic_folio(struct folio *folio,
>> - unsigned int order, bool demote)
>> +static bool prep_compound_gigantic_folio(struct folio *folio, struct
>> hstate *h, bool demote)
>> {
>> int i, j;
>> + int order = huge_page_order(h);
>> int nr_pages = 1 << order;
>> struct page *p;
>> __folio_clear_reserved(folio);
>> +
>> + /*
>> + * No need to prep pages that will be freed later by
>> hugetlb_vmemmap_optimize.
>> + * Hence, reduce nr_pages to the pages that will be kept.
>> + */
>> + if (IS_ENABLED(CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP) &&
>> + vmemmap_should_optimize(h, &folio->page))
>> + nr_pages = HUGETLB_VMEMMAP_RESERVE_SIZE / sizeof(struct page);
>
> We need to initialize the refcount to zero of tail pages (see the big
> comment below in this function), given a situation that someone (maybe
> GUP) could get a ref on the tail pages when the vmemmap is optimizing,
> what prevent this from happening?
>
> Thanks.
>
Thanks for pointing this out, will limit to boot time for solving this
in next version.
Powered by blists - more mailing lists