Re: [PATCH -V3] mm: Add PageLayzyFree() helper functions for MADV_FREE

[David Hildenbrand]

On 09.03.20 10:15, Huang, Ying wrote:
> David Hildenbrand <david@xxxxxxxxxx> writes:
> 
>> On 09.03.20 03:17, Huang, Ying wrote:
>>> From: Huang Ying <ying.huang@xxxxxxxxx>
>>>
>>> Now PageSwapBacked() is used as the helper function to check whether
>>> pages have been freed lazily via MADV_FREE.  This isn't very obvious.
>>> So Dave suggested to add PageLazyFree() family helper functions to
>>> improve the code readability.
>>>
>>> Signed-off-by: "Huang, Ying" <ying.huang@xxxxxxxxx>
>>> Suggested-by: Dave Hansen <dave.hansen@xxxxxxxxxxxxxxx>
>>> Cc: David Hildenbrand <david@xxxxxxxxxx>
>>> Cc: Mel Gorman <mgorman@xxxxxxx>
>>> Cc: Vlastimil Babka <vbabka@xxxxxxx>
>>> Cc: Zi Yan <ziy@xxxxxxxxxx>
>>> Cc: Michal Hocko <mhocko@xxxxxxxxxx>
>>> Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
>>> Cc: Minchan Kim <minchan@xxxxxxxxxx>
>>> Cc: Johannes Weiner <hannes@xxxxxxxxxxx>
>>> Cc: Hugh Dickins <hughd@xxxxxxxxxx>
>>> ---
>>> Changelog:
>>>
>>> v3:
>>>
>>> - Improved comments, Thanks David Hildenbrand!
>>>
>>> - Use the helper function in /proc/PID/smaps lazyfree reporting.
>>>
>>> v2:
>>>
>>> - Avoid code bloat via removing VM_BUG_ON_PAGE(), which doesn't exist
>>>   in the original code.  Now there is no any text/data/bss size
>>>   change.
>>>
>>> - Fix one wrong replacement in try_to_unmap_one().
>>>
>>> ---
>>>  fs/proc/task_mmu.c         |  2 +-
>>>  include/linux/page-flags.h | 25 +++++++++++++++++++++++++
>>>  mm/rmap.c                  |  4 ++--
>>>  mm/swap.c                  | 11 +++--------
>>>  mm/vmscan.c                |  7 +++----
>>>  5 files changed, 34 insertions(+), 15 deletions(-)
>>>
>>> diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c
>>> index 3ba9ae83bff5..3458d5711e57 100644
>>> --- a/fs/proc/task_mmu.c
>>> +++ b/fs/proc/task_mmu.c
>>> @@ -471,7 +471,7 @@ static void smaps_account(struct mem_size_stats *mss, struct page *page,
>>>  	 */
>>>  	if (PageAnon(page)) {
>>>  		mss->anonymous += size;
>>> -		if (!PageSwapBacked(page) && !dirty && !PageDirty(page))
>>> +		if (__PageLazyFree(page) && !dirty && !PageDirty(page))
>>>  			mss->lazyfree += size;
>>>  	}
>>>  
>>> diff --git a/include/linux/page-flags.h b/include/linux/page-flags.h
>>> index 49c2697046b9..bb26f74cbe8e 100644
>>> --- a/include/linux/page-flags.h
>>> +++ b/include/linux/page-flags.h
>>> @@ -498,6 +498,31 @@ static __always_inline int PageKsm(struct page *page)
>>>  TESTPAGEFLAG_FALSE(Ksm)
>>>  #endif
>>>  
>>> +/*
>>> + * For pages freed lazily via MADV_FREE.  Lazyfree pages are clean
>>> + * anonymous pages.  They don't have PG_swapbacked set, to distinguish
>>> + * them from normal anonymous pages.
>>> + */
>>> +static __always_inline int __PageLazyFree(struct page *page)
>>> +{
>>> +	return !PageSwapBacked(page);
>>> +}
>>> +
>>> +static __always_inline int PageLazyFree(struct page *page)
>>> +{
>>> +	return PageAnon(page) && __PageLazyFree(page);
>>> +}
>>> +
>>> +static __always_inline void SetPageLazyFree(struct page *page)
>>> +{
>>> +	ClearPageSwapBacked(page);
>>> +}
>>> +
>>> +static __always_inline void ClearPageLazyFree(struct page *page)
>>> +{
>>> +	SetPageSwapBacked(page);
>>> +}
>>> +
>>>  u64 stable_page_flags(struct page *page);
>>>  
>>>  static inline int PageUptodate(struct page *page)
>>> diff --git a/mm/rmap.c b/mm/rmap.c
>>> index 1c02adaa233e..6ec96c8e7826 100644
>>> --- a/mm/rmap.c
>>> +++ b/mm/rmap.c
>>> @@ -1609,7 +1609,7 @@ static bool try_to_unmap_one(struct page *page, struct vm_area_struct *vma,
>>>  			}
>>>  
>>>  			/* MADV_FREE page check */
>>> -			if (!PageSwapBacked(page)) {
>>> +			if (__PageLazyFree(page)) {
>>>  				if (!PageDirty(page)) {
>>>  					/* Invalidate as we cleared the pte */
>>>  					mmu_notifier_invalidate_range(mm,
>>> @@ -1623,7 +1623,7 @@ static bool try_to_unmap_one(struct page *page, struct vm_area_struct *vma,
>>>  				 * discarded. Remap the page to page table.
>>>  				 */
>>>  				set_pte_at(mm, address, pvmw.pte, pteval);
>>> -				SetPageSwapBacked(page);
>>> +				ClearPageLazyFree(page);
>>>  				ret = false;
>>>  				page_vma_mapped_walk_done(&pvmw);
>>>  				break;
>>> diff --git a/mm/swap.c b/mm/swap.c
>>> index c1d3ca80ea10..d83f2cd4cdb8 100644
>>> --- a/mm/swap.c
>>> +++ b/mm/swap.c
>>> @@ -563,7 +563,7 @@ static void lru_deactivate_fn(struct page *page, struct lruvec *lruvec,
>>>  static void lru_lazyfree_fn(struct page *page, struct lruvec *lruvec,
>>>  			    void *arg)
>>>  {
>>> -	if (PageLRU(page) && PageAnon(page) && PageSwapBacked(page) &&
>>> +	if (PageLRU(page) && PageAnon(page) && !__PageLazyFree(page) &&
>>>  	    !PageSwapCache(page) && !PageUnevictable(page)) {
>>>  		bool active = PageActive(page);
>>>  
>>> @@ -571,12 +571,7 @@ static void lru_lazyfree_fn(struct page *page, struct lruvec *lruvec,
>>>  				       LRU_INACTIVE_ANON + active);
>>>  		ClearPageActive(page);
>>>  		ClearPageReferenced(page);
>>> -		/*
>>> -		 * lazyfree pages are clean anonymous pages. They have
>>> -		 * SwapBacked flag cleared to distinguish normal anonymous
>>> -		 * pages
>>> -		 */
>>> -		ClearPageSwapBacked(page);
>>> +		SetPageLazyFree(page);
>>>  		add_page_to_lru_list(page, lruvec, LRU_INACTIVE_FILE);
>>>  
>>>  		__count_vm_events(PGLAZYFREE, hpage_nr_pages(page));
>>> @@ -678,7 +673,7 @@ void deactivate_page(struct page *page)
>>>   */
>>>  void mark_page_lazyfree(struct page *page)
>>>  {
>>> -	if (PageLRU(page) && PageAnon(page) && PageSwapBacked(page) &&
>>> +	if (PageLRU(page) && PageAnon(page) && !__PageLazyFree(page) &&
>>>  	    !PageSwapCache(page) && !PageUnevictable(page)) {
>>>  		struct pagevec *pvec = &get_cpu_var(lru_lazyfree_pvecs);
>>>  
>>> diff --git a/mm/vmscan.c b/mm/vmscan.c
>>> index eca49a1c2f68..40bb41ada2d2 100644
>>> --- a/mm/vmscan.c
>>> +++ b/mm/vmscan.c
>>> @@ -1043,8 +1043,7 @@ static void page_check_dirty_writeback(struct page *page,
>>>  	 * Anonymous pages are not handled by flushers and must be written
>>>  	 * from reclaim context. Do not stall reclaim based on them
>>>  	 */
>>> -	if (!page_is_file_cache(page) ||
>>> -	    (PageAnon(page) && !PageSwapBacked(page))) {
>>> +	if (!page_is_file_cache(page) || PageLazyFree(page)) {
>>>  		*dirty = false;
>>>  		*writeback = false;
>>>  		return;
>>> @@ -1235,7 +1234,7 @@ static unsigned long shrink_page_list(struct list_head *page_list,
>>>  		 * Try to allocate it some swap space here.
>>>  		 * Lazyfree page could be freed directly
>>>  		 */
>>> -		if (PageAnon(page) && PageSwapBacked(page)) {
>>> +		if (PageAnon(page) && !__PageLazyFree(page)) {
>>>  			if (!PageSwapCache(page)) {
>>>  				if (!(sc->gfp_mask & __GFP_IO))
>>>  					goto keep_locked;
>>> @@ -1411,7 +1410,7 @@ static unsigned long shrink_page_list(struct list_head *page_list,
>>>  			}
>>>  		}
>>>  
>>> -		if (PageAnon(page) && !PageSwapBacked(page)) {
>>> +		if (PageLazyFree(page)) {
>>>  			/* follow __remove_mapping for reference */
>>>  			if (!page_ref_freeze(page, 1))
>>>  				goto keep_locked;
>>>
>>
>> I still prefer something like
>>
>> diff --git a/include/linux/page-flags.h b/include/linux/page-flags.h
>> index fd6d4670ccc3..7538501230bd 100644
>> --- a/include/linux/page-flags.h
>> +++ b/include/linux/page-flags.h
>> @@ -63,6 +63,10 @@
>>   * page_waitqueue(page) is a wait queue of all tasks waiting for the page
>>   * to become unlocked.
>>   *
>> + * PG_swapbacked used with anonymous pages (PageAnon()) indicates that a
>> + * page is backed by swap. Anonymous pages without PG_swapbacked are
>> + * pages that can be lazily freed (e.g., MADV_FREE) on demand.
>> + *
>>   * PG_uptodate tells whether the page's contents is valid.  When a read
>>   * completes, the page becomes uptodate, unless a disk I/O error happened.
>>   *
> 
> Why not just send a formal patch?  So Andrew can just pick anything he
> likes.  I am totally OK with that.

Because you're working on cleaning this up.

> 
>> and really don't like the use of !__PageLazyFree() instead of PageSwapBacked().
> 
> If adopted, !__PageLazyFree() should only be used in the context where
> we really want to check whether pages are freed lazily.  Otherwise,
> PageSwapBacked() should be used.
> 

Yeah, and once again, personally, I don't like this approach. E.g.,
ClearPageLazyFree() sets PG_swapbacked. You already have to be aware
that this is a single flag being used in the background and what the
implications are. IMHO, in no way better than the current approach. I
prefer better documentation instead.

But I am just a reviewer and not a maintainer, so it's just my 2 cents.

-- 
Thanks,

David / dhildenb