linux-kernel - Re: [PATCH v3 07/14] mm/ksm: use folio in write_protect

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date: Mon, 8 Apr 2024 14:58:27 +0800
From: Alex Shi <seakeel@...il.com>
To: David Hildenbrand <david@...hat.com>, alexs@...nel.org,
 Matthew Wilcox <willy@...radead.org>, Andrea Arcangeli
 <aarcange@...hat.com>, Izik Eidus <izik.eidus@...ellosystems.com>,
 Andrew Morton <akpm@...ux-foundation.org>, linux-mm@...ck.org,
 linux-kernel@...r.kernel.org, ryncsn@...il.com
Cc: Hugh Dickins <hughd@...gle.com>, Chris Wright <chrisw@...s-sol.org>
Subject: Re: [PATCH v3 07/14] mm/ksm: use folio in write_protect_page



On 4/5/24 3:23 PM, David Hildenbrand wrote:
> On 25.03.24 13:48, alexs@...nel.org wrote:
>> From: "Alex Shi (tencent)" <alexs@...nel.org>
>>
>> Compound page is checked and skipped before write_protect_page() called,
>> use folio to save a few compound_head checking.
>>
>> Signed-off-by: Alex Shi (tencent) <alexs@...nel.org>
>> Cc: Izik Eidus <izik.eidus@...ellosystems.com>
>> Cc: Matthew Wilcox <willy@...radead.org>
>> Cc: Andrea Arcangeli <aarcange@...hat.com>
>> Cc: Hugh Dickins <hughd@...gle.com>
>> Cc: Chris Wright <chrisw@...s-sol.org>
>> ---
>>   mm/ksm.c | 22 +++++++++++-----------
>>   1 file changed, 11 insertions(+), 11 deletions(-)
>>
>> diff --git a/mm/ksm.c b/mm/ksm.c
>> index 95a487a21eed..5d1f62e7462a 100644
>> --- a/mm/ksm.c
>> +++ b/mm/ksm.c
>> @@ -1289,22 +1289,22 @@ static u32 calc_checksum(struct page *page)
>>       return checksum;
>>   }
>>   -static int write_protect_page(struct vm_area_struct *vma, struct page *page,
>> +static int write_protect_page(struct vm_area_struct *vma, struct folio *folio,
>>                     pte_t *orig_pte)
>>   {
>>       struct mm_struct *mm = vma->vm_mm;
>> -    DEFINE_PAGE_VMA_WALK(pvmw, page, vma, 0, 0);
>> +    DEFINE_FOLIO_VMA_WALK(pvmw, folio, vma, 0, 0);
>>       int swapped;
>>       int err = -EFAULT;
>>       struct mmu_notifier_range range;
>>       bool anon_exclusive;
>>       pte_t entry;
>>   -    pvmw.address = page_address_in_vma(page, vma);
>> +    pvmw.address = page_address_in_vma(&folio->page, vma);
>>       if (pvmw.address == -EFAULT)
>>           goto out;
>>   -    BUG_ON(PageTransCompound(page));
>> +    VM_BUG_ON(folio_test_large(folio));
> 
> I suggest
> 
> if (WARN_ON_ONCE(folio_test_large(folio)))
>     return err;
> 
> before the page_address_in_vma() call.
> 

Thanks for the suggestion, will take it.

> 
> Reviewed-by: David Hildenbrand <david@...hat.com>
>