lkml.org 
[lkml]   [2021]   [Dec]   [22]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
SubjectRe: [PATCH v2] mm/damon: Add access checking for hugetlb pages
From


On 12/22/2021 5:10 PM, SeongJae Park wrote:
> Hi Baolin,
>
>
> Basically, the code looks ok to me. I left so trivial cosmetic nitpicks below,
> though.
>
> On Thu, 16 Dec 2021 18:38:03 +0800 Baolin Wang <baolin.wang@linux.alibaba.com> wrote:
>
>> The process's VMAs can be mapped by hugetlb page, but now the DAMON
>> did not implement the access checking for hugetlb pte, so we can not
>> get the actual access count like below if a process VMAs were mapped
>> by hugetlb.
>>
>> damon_aggregated: target_id=18446614368406014464
>> nr_regions=12 4194304-5476352: 0 545
>> damon_aggregated: target_id=18446614368406014464
>> nr_regions=12 140662370467840-140662372970496: 0 545
>> damon_aggregated: target_id=18446614368406014464
>> nr_regions=12 140662372970496-140662375460864: 0 545
>> damon_aggregated: target_id=18446614368406014464
>> nr_regions=12 140662375460864-140662377951232: 0 545
>> damon_aggregated: target_id=18446614368406014464
>> nr_regions=12 140662377951232-140662380449792: 0 545
>> damon_aggregated: target_id=18446614368406014464
>> nr_regions=12 140662380449792-140662382944256: 0 545
>> ......
>
> I'd prefer indenting the program output with 4 spaces and not wrapping it.
> e.g.,
>
> damon_aggregated: target_id=18446614368406014464 nr_regions=12 4194304-5476352: 0 545
> damon_aggregated: target_id=18446614368406014464 nr_regions=12 140662370467840-140662372970496: 0 545

Sure.

>>
>> Thus this patch adds hugetlb access checking support, with this patch
>> we can see below VMA mapped by hugetlb access count.
>>
>> damon_aggregated: target_id=18446613056935405824
>> nr_regions=12 140296486649856-140296489914368: 1 3
>> damon_aggregated: target_id=18446613056935405824
>> nr_regions=12 140296489914368-140296492978176: 1 3
>> damon_aggregated: target_id=18446613056935405824
>> nr_regions=12 140296492978176-140296495439872: 1 3
>> damon_aggregated: target_id=18446613056935405824
>> nr_regions=12 140296495439872-140296498311168: 1 3
>> damon_aggregated: target_id=18446613056935405824
>> nr_regions=12 140296498311168-140296501198848: 1 3
>> damon_aggregated: target_id=18446613056935405824
>> nr_regions=12 140296501198848-140296504320000: 1 3
>> damon_aggregated: target_id=18446613056935405824
>> nr_regions=12 140296504320000-140296507568128: 1 2
>> ......
>
> ditto.

Sure.

>> +static int damon_mkold_hugetlb_entry(pte_t *pte, unsigned long hmask,
>> + unsigned long addr, unsigned long end,
>> + struct mm_walk *walk)
>> +{
>> + struct hstate *h = hstate_vma(walk->vma);
>> + spinlock_t *ptl;
>> + pte_t entry;
>> +
>> + ptl = huge_pte_lock(h, walk->mm, pte);
>> + entry = huge_ptep_get(pte);
>> + if (!pte_present(entry))
>> + goto out;
>> +
>> + damon_hugetlb_mkold(pte, walk->mm, walk->vma, addr);
>> +
>> +out:
>> + spin_unlock(ptl);
>> + return 0;
>> +}
>> +#else
>> +#define damon_mkold_hugetlb_entry NULL
>> +#endif
>
> Could we append a comment saying this #endif is for #ifdef CONFIG_HUGETLB_PAGE,
> like below?
>
> #endif /* CONFIG_HUGETLB_PAGE */

Sure.

>> +#ifdef CONFIG_HUGETLB_PAGE
>> +static int damon_young_hugetlb_entry(pte_t *pte, unsigned long hmask,
>> + unsigned long addr, unsigned long end,
>> + struct mm_walk *walk)
>> +{
>> + struct damon_young_walk_private *priv = walk->private;
>> + struct hstate *h = hstate_vma(walk->vma);
>> + struct page *page;
>> + spinlock_t *ptl;
>> + pte_t entry;
>> +
>> + ptl = huge_pte_lock(h, walk->mm, pte);
>> + entry = huge_ptep_get(pte);
>> + if (!pte_present(entry))
>> + goto out;
>> +
>> + page = pte_page(entry);
>> + if (!page)
>> + goto out;
>> +
>> + get_page(page);
>> +
>> + if (pte_young(entry) || !page_is_idle(page) ||
>> + mmu_notifier_test_young(walk->mm, addr)) {
>> + *priv->page_sz = huge_page_size(h);
>> + priv->young = true;
>> + }
>> +
>> + put_page(page);
>> +
>> +out:
>> + spin_unlock(ptl);
>> + return 0;
>> +}
>> +#else
>> +#define damon_young_hugetlb_entry NULL
>> +#endif
>
> ditto.

Sure.

But I saw Andrew had applied this version into his branch.

Andrew, would you like me to send a new version? or an increment patch
to fix the coding style issue? Thanks.

\
 
 \ /
  Last update: 2021-12-22 10:32    [W:0.045 / U:0.408 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site