Re: [PATCH v4 34/36] rmap: add folio_add_file_rmap_range()

[Ryan Roberts]

On 16/03/2023 16:27, Yin, Fengwei wrote:
> Hi Matthew,
> 
> On 3/16/2023 12:08 AM, Ryan Roberts wrote:
>> On 15/03/2023 13:34, Ryan Roberts wrote:
>>> On 15/03/2023 05:14, Matthew Wilcox (Oracle) wrote:
>>>> From: Yin Fengwei <fengwei.yin@xxxxxxxxx>
>>>>
>>>> folio_add_file_rmap_range() allows to add pte mapping to a specific
>>>> range of file folio. Comparing to page_add_file_rmap(), it batched
>>>> updates __lruvec_stat for large folio.
>>>>
>>>> Signed-off-by: Yin Fengwei <fengwei.yin@xxxxxxxxx>
>>>> Signed-off-by: Matthew Wilcox (Oracle) <willy@xxxxxxxxxxxxx>
>>>> ---
>>>>  include/linux/rmap.h |  2 ++
>>>>  mm/rmap.c            | 60 +++++++++++++++++++++++++++++++++-----------
>>>>  2 files changed, 48 insertions(+), 14 deletions(-)
>>>>
>>>> diff --git a/include/linux/rmap.h b/include/linux/rmap.h
>>>> index b87d01660412..a3825ce81102 100644
>>>> --- a/include/linux/rmap.h
>>>> +++ b/include/linux/rmap.h
>>>> @@ -198,6 +198,8 @@ void folio_add_new_anon_rmap(struct folio *, struct vm_area_struct *,
>>>>  		unsigned long address);
>>>>  void page_add_file_rmap(struct page *, struct vm_area_struct *,
>>>>  		bool compound);
>>>> +void folio_add_file_rmap_range(struct folio *, struct page *, unsigned int nr,
>>>> +		struct vm_area_struct *, bool compound);
>>>>  void page_remove_rmap(struct page *, struct vm_area_struct *,
>>>>  		bool compound);
>>>>  
>>>> diff --git a/mm/rmap.c b/mm/rmap.c
>>>> index 4898e10c569a..a91906b28835 100644
>>>> --- a/mm/rmap.c
>>>> +++ b/mm/rmap.c
>>>> @@ -1301,31 +1301,39 @@ void folio_add_new_anon_rmap(struct folio *folio, struct vm_area_struct *vma,
>>>>  }
>>>>  
>>>>  /**
>>>> - * page_add_file_rmap - add pte mapping to a file page
>>>> - * @page:	the page to add the mapping to
>>>> + * folio_add_file_rmap_range - add pte mapping to page range of a folio
>>>> + * @folio:	The folio to add the mapping to
>>>> + * @page:	The first page to add
>>>> + * @nr_pages:	The number of pages which will be mapped
>>>>   * @vma:	the vm area in which the mapping is added
>>>>   * @compound:	charge the page as compound or small page
>>>>   *
>>>> + * The page range of folio is defined by [first_page, first_page + nr_pages)
>>>> + *
>>>>   * The caller needs to hold the pte lock.
>>>>   */
>>>> -void page_add_file_rmap(struct page *page, struct vm_area_struct *vma,
>>>> -		bool compound)
>>>> +void folio_add_file_rmap_range(struct folio *folio, struct page *page,
>>>> +			unsigned int nr_pages, struct vm_area_struct *vma,
>>>> +			bool compound)
>>>>  {
>>>> -	struct folio *folio = page_folio(page);
>>>>  	atomic_t *mapped = &folio->_nr_pages_mapped;
>>>> -	int nr = 0, nr_pmdmapped = 0;
>>>> -	bool first;
>>>> +	unsigned int nr_pmdmapped = 0, first;
>>>> +	int nr = 0;
>>>>  
>>>> -	VM_BUG_ON_PAGE(compound && !PageTransHuge(page), page);
>>>> +	VM_WARN_ON_FOLIO(compound && !folio_test_pmd_mappable(folio), folio);
>>>>  
>>>>  	/* Is page being mapped by PTE? Is this its first map to be added? */
>>>>  	if (likely(!compound)) {
>>>> -		first = atomic_inc_and_test(&page->_mapcount);
>>>> -		nr = first;
>>>> -		if (first && folio_test_large(folio)) {
>>>> -			nr = atomic_inc_return_relaxed(mapped);
>>>> -			nr = (nr < COMPOUND_MAPPED);
>>>> -		}
>>>> +		do {
>>>> +			first = atomic_inc_and_test(&page->_mapcount);
>>>> +			if (first && folio_test_large(folio)) {
>>>> +				first = atomic_inc_return_relaxed(mapped);
>>>> +				first = (nr < COMPOUND_MAPPED);
>>>
>>> This still contains the typo that Yin Fengwei spotted in the previous version:
>>> https://lore.kernel.org/linux-mm/20230228213738.272178-1-willy@xxxxxxxxxxxxx/T/#m84673899e25bc31356093a1177941f2cc35e5da8
>>>
>>> FYI, I'm seeing a perf regression of about 1% when compiling the kernel on
>>> Ampere Altra (arm64) with this whole series on top of v6.3-rc1 (In a VM using
>>> ext4 filesystem). Looks like instruction aborts are taking much longer and a
>>> selection of syscalls are a bit slower. Still hunting down the root cause. Will
>>> report once I have conclusive diagnosis.
>>
>> I'm sorry - I'm struggling to find the exact cause. But its spending over 2x the
>> amount of time in the instruction abort handling code once patches 32-36 are
>> included. Everything in the flame graph is just taking longer. Perhaps we are
>> getting more instruction aborts somehow? I have the flamegraphs if anyone wants
>> them - just shout and I'll email them separately.
> Thanks a lot to Ryan for sharing the flamegraphs to me. I found the __do_fault()
> is called with patch 32-36 while no __do_fault() just with first 31 patches. I 
> suspect the folio_add_file_rmap_range() missed some PTEs population. Please give
> me few days to find the root cause and fix. Sorry for this.

You're welcome. Give me a shout once you have a re-spin and I'll rerun the tests.

> 
> 
> Regards
> Yin, Fengwei
> 
>>
>>>
>>> Thanks,
>>> Ryan
>>>
>>>
>>>> +			}
>>>> +
>>>> +			if (first)
>>>> +				nr++;
>>>> +		} while (page++, --nr_pages > 0);
>>>>  	} else if (folio_test_pmd_mappable(folio)) {
>>>>  		/* That test is redundant: it's for safety or to optimize out */
>>>>  
>>>> @@ -1354,6 +1362,30 @@ void page_add_file_rmap(struct page *page, struct vm_area_struct *vma,
>>>>  	mlock_vma_folio(folio, vma, compound);
>>>>  }
>>>>  
>>>> +/**
>>>> + * page_add_file_rmap - add pte mapping to a file page
>>>> + * @page:	the page to add the mapping to
>>>> + * @vma:	the vm area in which the mapping is added
>>>> + * @compound:	charge the page as compound or small page
>>>> + *
>>>> + * The caller needs to hold the pte lock.
>>>> + */
>>>> +void page_add_file_rmap(struct page *page, struct vm_area_struct *vma,
>>>> +		bool compound)
>>>> +{
>>>> +	struct folio *folio = page_folio(page);
>>>> +	unsigned int nr_pages;
>>>> +
>>>> +	VM_WARN_ON_ONCE_PAGE(compound && !PageTransHuge(page), page);
>>>> +
>>>> +	if (likely(!compound))
>>>> +		nr_pages = 1;
>>>> +	else
>>>> +		nr_pages = folio_nr_pages(folio);
>>>> +
>>>> +	folio_add_file_rmap_range(folio, page, nr_pages, vma, compound);
>>>> +}
>>>> +
>>>>  /**
>>>>   * page_remove_rmap - take down pte mapping from a page
>>>>   * @page:	page to remove mapping from
>>>
>>