linux-kernel - Re: [PATCH] mm/hugetlb: fix copy_hugetlb_page_range() to use ->pt_share

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] [day] [month] [year] [list]

Message-ID: <eb6e18c0-533f-4e77-a56f-60ab8cacc369@redhat.com>
Date: Fri, 12 Sep 2025 09:31:04 +0200
From: David Hildenbrand <david@...hat.com>
To: jane.chu@...cle.com, harry.yoo@...cle.com, osalvador@...e.de,
 liushixin2@...wei.com, muchun.song@...ux.dev, akpm@...ux-foundation.org,
 linux-mm@...ck.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] mm/hugetlb: fix copy_hugetlb_page_range() to use
 ->pt_share_count

On 11.09.25 21:54, jane.chu@...cle.com wrote:
> 
> On 9/9/2025 11:45 PM, David Hildenbrand wrote:
> [..]
>>> -        /*
>>> -         * If the pagetables are shared don't copy or take references.
>>> -         *
>>> -         * dst_pte == src_pte is the common case of src/dest sharing.
>>> -         * However, src could have 'unshared' and dst shares with
>>> -         * another vma. So page_count of ptep page is checked instead
>>> -         * to reliably determine whether pte is shared.
>>> -         */
>>> -        if (page_count(virt_to_page(dst_pte)) > 1) {
>>> +#ifdef CONFIG_HUGETLB_PMD_PAGE_TABLE_SHARING
>>> +        /* If the pagetables are shared don't copy or take
>>> references. */
>>
>> Why remove so much of the original comment?
> 
> Because, this part of checking has already advanced from the "dst_pte ==
> src_pte" to "page_count() > 1" to ->pt_share_count > 0, it seems cleaner
> to just keep an one liner comment.
> That said, if you feel the comments should be kept, I'd be happy to
> restore them with a bit revision.

Well, the comment explains why checking the pte pointers is insufficient 
and why there is a corner case where the pointers differ but we still 
want to unshare. :)

But yeah, I agree that reading the code it's clear: if dst is already 
shared, just don't do anything.

I would probably rephrase the comment to something simpler like

"/* If the pagetables are shared, there is nothing to do. */

If you resend, please add a comment to the patch description like "While 
at it, simplify the comment, the details are not actually relevant anymore".

-- 
Cheers

David / dhildenb