lkml.org 
[lkml]   [2015]   [Mar]   [29]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
From
SubjectRe: [PATCHv4 12/24] thp: PMD splitting without splitting compound page
Date
"Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> writes:

> Current split_huge_page() combines two operations: splitting PMDs into
> tables of PTEs and splitting underlying compound page. This patch
> changes split_huge_pmd() implementation to split the given PMD without
> splitting other PMDs this page mapped with or underlying compound page.
>
> In order to do this we have to get rid of tail page refcounting, which
> uses _mapcount of tail pages. Tail page refcounting is needed to be able
> to split THP page at any point: we always know which of tail pages is
> pinned (i.e. by get_user_pages()) and can distribute page count
> correctly.
>
> We can avoid this by allowing split_huge_page() to fail if the compound
> page is pinned. This patch removes all infrastructure for tail page
> refcounting and make split_huge_page() to always return -EBUSY. All
> split_huge_page() users already know how to handle its fail. Proper
> implementation will be added later.
>
> Without tail page refcounting, implementation of split_huge_pmd() is
> pretty straight-forward.
>
> Memory cgroup is not yet ready for new refcouting. Let's disable it on
> Kconfig level.
>
.....
......

>
> spin_lock(ptl);
> if (page)
> - put_user_huge_page(page);
> + put_page(page);
> if (unlikely(!pmd_same(*pmd, orig_pmd))) {
> spin_unlock(ptl);
> mem_cgroup_cancel_charge(new_page, memcg);
> @@ -1662,51 +1631,78 @@ static void __split_huge_zero_page_pmd(struct vm_area_struct *vma,
> put_huge_zero_page();
> }
>
> -void __split_huge_page_pmd(struct vm_area_struct *vma, unsigned long address,
> - pmd_t *pmd)
> +
> +static void __split_huge_pmd_locked(struct vm_area_struct *vma,
> + pmd_t *pmd, unsigned long address)
> {
> - spinlock_t *ptl;
> + unsigned long haddr = address & HPAGE_PMD_MASK;
> struct page *page;
> struct mm_struct *mm = vma->vm_mm;
> - unsigned long haddr = address & HPAGE_PMD_MASK;
> - unsigned long mmun_start; /* For mmu_notifiers */
> - unsigned long mmun_end; /* For mmu_notifiers */
> + pgtable_t pgtable;
> + pmd_t _pmd;
> + bool young, write;
> + int i;
>
> - BUG_ON(vma->vm_start > haddr || vma->vm_end < haddr + HPAGE_PMD_SIZE);
> + VM_BUG_ON_VMA(vma->vm_start > haddr, vma);
> + VM_BUG_ON_VMA(vma->vm_end < haddr + HPAGE_PMD_SIZE, vma);
> +
> + if (is_huge_zero_pmd(*pmd))
> + return __split_huge_zero_page_pmd(vma, haddr, pmd);
> +
> + page = pmd_page(*pmd);
> + VM_BUG_ON_PAGE(!page_count(page), page);
> + atomic_add(HPAGE_PMD_NR - 1, &page->_count);
> +
> + write = pmd_write(*pmd);
> + young = pmd_young(*pmd);
> +
> + /* leave pmd empty until pte is filled */
> + pmdp_clear_flush_notify(vma, haddr, pmd);
> +

So we now mark pmd none, while we go ahead and split the pmd. But then what
happens to a parallel fault ? We don't hold mmap_sem here right ?

> + pgtable = pgtable_trans_huge_withdraw(mm, pmd);
> + pmd_populate(mm, &_pmd, pgtable);
> +
> + for (i = 0; i < HPAGE_PMD_NR; i++, haddr += PAGE_SIZE) {
> + pte_t entry, *pte;
> + /*
> + * Note that NUMA hinting access restrictions are not
> + * transferred to avoid any possibility of altering
> + * permissions across VMAs.
> + */
> + entry = mk_pte(page + i, vma->vm_page_prot);
> + entry = maybe_mkwrite(pte_mkdirty(entry), vma);
> + if (!write)
> + entry = pte_wrprotect(entry);
> + if (!young)
> + entry = pte_mkold(entry);
> + pte = pte_offset_map(&_pmd, haddr);
> + BUG_ON(!pte_none(*pte));
> + atomic_inc(&page[i]._mapcount);
> + set_pte_at(mm, haddr, pte, entry);
> + pte_unmap(pte);
> + }
> + smp_wmb(); /* make pte visible before pmd */
> + pmd_populate(mm, pmd, pgtable);
> + atomic_dec(compound_mapcount_ptr(page));
> +}
> +

-aneesh



\
 
 \ /
  Last update: 2015-03-29 19:01    [W:0.539 / U:0.040 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site