linux-kernel - Re: [PATCH v2 08/11] hugetlb: batch freeing of vmemmap pages

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]

Message-Id: <9F499C49-3046-4EF2-8C2A-3458A954B2DE@linux.dev>
Date:   Thu, 7 Sep 2023 14:19:45 +0800
From:   Muchun Song <muchun.song@...ux.dev>
To:     Mike Kravetz <mike.kravetz@...cle.com>
Cc:     Linux-MM <linux-mm@...ck.org>, LKML <linux-kernel@...r.kernel.org>,
        Muchun Song <songmuchun@...edance.com>,
        Joao Martins <joao.m.martins@...cle.com>,
        Oscar Salvador <osalvador@...e.de>,
        David Hildenbrand <david@...hat.com>,
        Miaohe Lin <linmiaohe@...wei.com>,
        David Rientjes <rientjes@...gle.com>,
        Anshuman Khandual <anshuman.khandual@....com>,
        Naoya Horiguchi <naoya.horiguchi@...ux.dev>,
        Michal Hocko <mhocko@...e.com>,
        Matthew Wilcox <willy@...radead.org>,
        Xiongchun Duan <duanxiongchun@...edance.com>,
        Andrew Morton <akpm@...ux-foundation.org>
Subject: Re: [PATCH v2 08/11] hugetlb: batch freeing of vmemmap pages



> On Sep 7, 2023, at 05:38, Mike Kravetz <mike.kravetz@...cle.com> wrote:
> 
> On 09/06/23 15:38, Muchun Song wrote:
>> 
>> 
>> On 2023/9/6 05:44, Mike Kravetz wrote:
>>> Now that batching of hugetlb vmemmap optimization processing is possible,
>>> batch the freeing of vmemmap pages.  When freeing vmemmap pages for a
>>> hugetlb page, we add them to a list that is freed after the entire batch
>>> has been processed.
>>> 
>>> This enhances the ability to return contiguous ranges of memory to the
>>> low level allocators.
>>> 
>>> Signed-off-by: Mike Kravetz <mike.kravetz@...cle.com>
>>> ---
>>>  mm/hugetlb_vmemmap.c | 60 ++++++++++++++++++++++++++++----------------
>>>  1 file changed, 38 insertions(+), 22 deletions(-)
>>> 
>>> diff --git a/mm/hugetlb_vmemmap.c b/mm/hugetlb_vmemmap.c
>>> index 79de984919ef..a715712df831 100644
>>> --- a/mm/hugetlb_vmemmap.c
>>> +++ b/mm/hugetlb_vmemmap.c
>>> @@ -306,18 +306,21 @@ static void vmemmap_restore_pte(pte_t *pte, unsigned long addr,
>>>   * @end: end address of the vmemmap virtual address range that we want to
>>>   * remap.
>>>   * @reuse: reuse address.
>>> + * @vmemmap_pages: list to deposit vmemmap pages to be freed.  It is callers
>>> + * responsibility to free pages.
>>>   *
>>>   * Return: %0 on success, negative error code otherwise.
>>>   */
>>>  static int vmemmap_remap_free(unsigned long start, unsigned long end,
>>> -       unsigned long reuse)
>>> +       unsigned long reuse,
>>> +       struct list_head *vmemmap_pages)
>>>  {
>>>   int ret;
>>> - LIST_HEAD(vmemmap_pages);
>>> + LIST_HEAD(freed_pages);
>> 
>> IIUC, we could reuse the parameter of @vmemmap_pages directly instead of
>> a temporary variable, could it be dropped?
>> 
> 
> I was concerned about the error case where we call vmemmap_remap_range a
> second time.  In the first call to vmemmap_remap_range with vmemmap_remap_pte,
> vmemmap pages to be freed are added to the end of the list (list_add_tail).
> In the call to vmemmap_remap_range after error with vmemmap_restore_pte,
> pages are taken off the head of the list (list_first_entry).  So, it seems
> that it would be possible to use a different set of pages in the restore

Yes.

> operation.  This would be an issue if pages had different characteristics such
> as being on different nodes.  Is that a real concern?

A good point. Now I see your concern, it is better to keep the same node
as before when error occurs.

> 
> I suppose we could change vmemmap_remap_pte to add pages to the head of
> the list?  I do not recall the reasoning behind adding to tail.

I think we could do this, the code will be a little simple. Actually, there
is no reason behind adding to tail (BTW, the first commit is introduced by
me, no secret here :-)).

Thanks.

> -- 
> Mike Kravetz