Re: [PATCH 1/2] Revert "mm: make faultaround produce old ptes"

[Michal Hocko]

On Tue 14-06-16 11:42:29, Kirill A. Shutemov wrote:
> This reverts commit 5c0a85fad949212b3e059692deecdeed74ae7ec7.
> 
> The commit causes ~6% regression in unixbench.

Is the regression fully explained? My understanding from the email
thread is that this is suspiciously too high. It is not like I would
be against the revert but having an explanation would be really
appreciated.

> Let's revert it for now and consider other solution for reclaim problem
> later.
> 
> Signed-off-by: Kirill A. Shutemov <kirill.shutemov@xxxxxxxxxxxxxxx>
> Reported-by: "Huang, Ying" <ying.huang@xxxxxxxxx>
> ---
>  include/linux/mm.h |  2 +-
>  mm/filemap.c       |  2 +-
>  mm/memory.c        | 23 +++++------------------
>  3 files changed, 7 insertions(+), 20 deletions(-)
> 
> diff --git a/include/linux/mm.h b/include/linux/mm.h
> index 5df5feb49575..ece042dfe23c 100644
> --- a/include/linux/mm.h
> +++ b/include/linux/mm.h
> @@ -602,7 +602,7 @@ static inline pte_t maybe_mkwrite(pte_t pte, struct vm_area_struct *vma)
>  }
>  
>  void do_set_pte(struct vm_area_struct *vma, unsigned long address,
> -		struct page *page, pte_t *pte, bool write, bool anon, bool old);
> +		struct page *page, pte_t *pte, bool write, bool anon);
>  #endif
>  
>  /*
> diff --git a/mm/filemap.c b/mm/filemap.c
> index 00ae878b2a38..20f3b1f33f0e 100644
> --- a/mm/filemap.c
> +++ b/mm/filemap.c
> @@ -2186,7 +2186,7 @@ repeat:
>  		if (file->f_ra.mmap_miss > 0)
>  			file->f_ra.mmap_miss--;
>  		addr = address + (page->index - vmf->pgoff) * PAGE_SIZE;
> -		do_set_pte(vma, addr, page, pte, false, false, true);
> +		do_set_pte(vma, addr, page, pte, false, false);
>  		unlock_page(page);
>  		goto next;
>  unlock:
> diff --git a/mm/memory.c b/mm/memory.c
> index 15322b73636b..61fe7e7b56bf 100644
> --- a/mm/memory.c
> +++ b/mm/memory.c
> @@ -2877,7 +2877,7 @@ static int __do_fault(struct vm_area_struct *vma, unsigned long address,
>   * vm_ops->map_pages.
>   */
>  void do_set_pte(struct vm_area_struct *vma, unsigned long address,
> -		struct page *page, pte_t *pte, bool write, bool anon, bool old)
> +		struct page *page, pte_t *pte, bool write, bool anon)
>  {
>  	pte_t entry;
>  
> @@ -2885,8 +2885,6 @@ void do_set_pte(struct vm_area_struct *vma, unsigned long address,
>  	entry = mk_pte(page, vma->vm_page_prot);
>  	if (write)
>  		entry = maybe_mkwrite(pte_mkdirty(entry), vma);
> -	if (old)
> -		entry = pte_mkold(entry);
>  	if (anon) {
>  		inc_mm_counter_fast(vma->vm_mm, MM_ANONPAGES);
>  		page_add_new_anon_rmap(page, vma, address, false);
> @@ -3032,20 +3030,9 @@ static int do_read_fault(struct mm_struct *mm, struct vm_area_struct *vma,
>  	 */
>  	if (vma->vm_ops->map_pages && fault_around_bytes >> PAGE_SHIFT > 1) {
>  		pte = pte_offset_map_lock(mm, pmd, address, &ptl);
> -		if (!pte_same(*pte, orig_pte))
> -			goto unlock_out;
>  		do_fault_around(vma, address, pte, pgoff, flags);
> -		/* Check if the fault is handled by faultaround */
> -		if (!pte_same(*pte, orig_pte)) {
> -			/*
> -			 * Faultaround produce old pte, but the pte we've
> -			 * handler fault for should be young.
> -			 */
> -			pte_t entry = pte_mkyoung(*pte);
> -			if (ptep_set_access_flags(vma, address, pte, entry, 0))
> -				update_mmu_cache(vma, address, pte);
> +		if (!pte_same(*pte, orig_pte))
>  			goto unlock_out;
> -		}
>  		pte_unmap_unlock(pte, ptl);
>  	}
>  
> @@ -3060,7 +3047,7 @@ static int do_read_fault(struct mm_struct *mm, struct vm_area_struct *vma,
>  		put_page(fault_page);
>  		return ret;
>  	}
> -	do_set_pte(vma, address, fault_page, pte, false, false, false);
> +	do_set_pte(vma, address, fault_page, pte, false, false);
>  	unlock_page(fault_page);
>  unlock_out:
>  	pte_unmap_unlock(pte, ptl);
> @@ -3111,7 +3098,7 @@ static int do_cow_fault(struct mm_struct *mm, struct vm_area_struct *vma,
>  		}
>  		goto uncharge_out;
>  	}
> -	do_set_pte(vma, address, new_page, pte, true, true, false);
> +	do_set_pte(vma, address, new_page, pte, true, true);
>  	mem_cgroup_commit_charge(new_page, memcg, false, false);
>  	lru_cache_add_active_or_unevictable(new_page, vma);
>  	pte_unmap_unlock(pte, ptl);
> @@ -3164,7 +3151,7 @@ static int do_shared_fault(struct mm_struct *mm, struct vm_area_struct *vma,
>  		put_page(fault_page);
>  		return ret;
>  	}
> -	do_set_pte(vma, address, fault_page, pte, true, false, false);
> +	do_set_pte(vma, address, fault_page, pte, true, false);
>  	pte_unmap_unlock(pte, ptl);
>  
>  	if (set_page_dirty(fault_page))
> -- 
> 2.8.1
> 

-- 
Michal Hocko
SUSE Labs