[RFC] f2fs: Convert f2fs_write_cache_pages() to use filemap_get_folios_tag()

  Converted the function to use a folio_batch instead of pagevec. This is in
preparation for the removal of find_get_pages_range_tag().

Also modified f2fs_all_cluster_page_ready to take in a folio_batch instead
of pagevec. This does NOT support large folios. The function currently
only utilizes folios of size 1 so this shouldn't cause any issues right
now.

This version of the patch limits the number of pages fetched to
F2FS_ONSTACK_PAGES. If that ever happens, update the start index here
since filemap_get_folios_tag() updates the index to be after the last
found folio, not necessarily the last used page.

Signed-off-by: Vishal Moola (Oracle) <vishal.moola@gmail.com>
---

Let me know if you prefer this version and I'll include it in v5
of the patch series when I rebase it after the merge window.

---
 fs/f2fs/data.c | 86 ++++++++++++++++++++++++++++++++++----------------
 1 file changed, 59 insertions(+), 27 deletions(-)
  

On 2022/12/13 3:13, Vishal Moola (Oracle) wrote:
> Converted the function to use a folio_batch instead of pagevec. This is in
> preparation for the removal of find_get_pages_range_tag().
> 
> Also modified f2fs_all_cluster_page_ready to take in a folio_batch instead
> of pagevec. This does NOT support large folios. The function currently
> only utilizes folios of size 1 so this shouldn't cause any issues right
> now.
> 
> This version of the patch limits the number of pages fetched to
> F2FS_ONSTACK_PAGES. If that ever happens, update the start index here
> since filemap_get_folios_tag() updates the index to be after the last
> found folio, not necessarily the last used page.
> 
> Signed-off-by: Vishal Moola (Oracle) <vishal.moola@gmail.com>
> ---
> 
> Let me know if you prefer this version and I'll include it in v5
> of the patch series when I rebase it after the merge window.
> 
> ---
>   fs/f2fs/data.c | 86 ++++++++++++++++++++++++++++++++++----------------
>   1 file changed, 59 insertions(+), 27 deletions(-)
> 
> diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c
> index a71e818cd67b..1703e353f0e0 100644
> --- a/fs/f2fs/data.c
> +++ b/fs/f2fs/data.c
> @@ -2939,6 +2939,7 @@ static int f2fs_write_cache_pages(struct address_space *mapping,
>   	int ret = 0;
>   	int done = 0, retry = 0;
>   	struct page *pages[F2FS_ONSTACK_PAGES];
> +	struct folio_batch fbatch;
>   	struct f2fs_sb_info *sbi = F2FS_M_SB(mapping);
>   	struct bio *bio = NULL;
>   	sector_t last_block;
> @@ -2959,6 +2960,7 @@ static int f2fs_write_cache_pages(struct address_space *mapping,
>   		.private = NULL,
>   	};
>   #endif
> +	int nr_folios, p, idx;
>   	int nr_pages;
>   	pgoff_t index;
>   	pgoff_t end;		/* Inclusive */
> @@ -2969,6 +2971,8 @@ static int f2fs_write_cache_pages(struct address_space *mapping,
>   	int submitted = 0;
>   	int i;
>   
> +	folio_batch_init(&fbatch);
> +
>   	if (get_dirty_pages(mapping->host) <=
>   				SM_I(F2FS_M_SB(mapping))->min_hot_blocks)
>   		set_inode_flag(mapping->host, FI_HOT_DATA);
> @@ -2994,13 +2998,38 @@ static int f2fs_write_cache_pages(struct address_space *mapping,
>   		tag_pages_for_writeback(mapping, index, end);
>   	done_index = index;
>   	while (!done && !retry && (index <= end)) {
> -		nr_pages = find_get_pages_range_tag(mapping, &index, end,
> -				tag, F2FS_ONSTACK_PAGES, pages);
> -		if (nr_pages == 0)
> +		nr_pages = 0;
> +again:
> +		nr_folios = filemap_get_folios_tag(mapping, &index, end,
> +				tag, &fbatch);
> +		if (nr_folios == 0) {
> +			if (nr_pages)
> +				goto write;
>   			break;
> +		}
>   
> +		for (i = 0; i < nr_folios; i++) {
> +			struct folio* folio = fbatch.folios[i];
> +
> +			idx = 0;
> +			p = folio_nr_pages(folio);
> +add_more:
> +			pages[nr_pages] = folio_page(folio,idx);
> +			folio_ref_inc(folio);

It looks if CONFIG_LRU_GEN is not set, folio_ref_inc() does nothing. For those
folios recorded in pages array, we need to call folio_get() here to add one more
reference on each of them?

> +			if (++nr_pages == F2FS_ONSTACK_PAGES) {
> +				index = folio->index + idx + 1;
> +				folio_batch_release(&fbatch);

Otherwise after folio_batch_release(), it may cause use-after-free issue
when accessing pages array? Or am I missing something?

> +				goto write;
> +			}
> +			if (++idx < p)
> +				goto add_more;
> +		}
> +		folio_batch_release(&fbatch);
> +		goto again;
> +write:
>   		for (i = 0; i < nr_pages; i++) {
>   			struct page *page = pages[i];
> +			struct folio *folio = page_folio(page);
>   			bool need_readd;
>   readd:
>   			need_readd = false;
> @@ -3017,7 +3046,7 @@ static int f2fs_write_cache_pages(struct address_space *mapping,
>   				}
>   
>   				if (!f2fs_cluster_can_merge_page(&cc,
> -								page->index)) {
> +								folio->index)) {
>   					ret = f2fs_write_multi_pages(&cc,
>   						&submitted, wbc, io_type);
>   					if (!ret)
> @@ -3026,27 +3055,28 @@ static int f2fs_write_cache_pages(struct address_space *mapping,
>   				}
>   
>   				if (unlikely(f2fs_cp_error(sbi)))
> -					goto lock_page;
> +					goto lock_folio;
>   
>   				if (!f2fs_cluster_is_empty(&cc))
> -					goto lock_page;
> +					goto lock_folio;
>   
>   				if (f2fs_all_cluster_page_ready(&cc,
>   					pages, i, nr_pages, true))
> -					goto lock_page;
> +					goto lock_folio;
>   
>   				ret2 = f2fs_prepare_compress_overwrite(
>   							inode, &pagep,
> -							page->index, &fsdata);
> +							folio->index, &fsdata);
>   				if (ret2 < 0) {
>   					ret = ret2;
>   					done = 1;
>   					break;
>   				} else if (ret2 &&
>   					(!f2fs_compress_write_end(inode,
> -						fsdata, page->index, 1) ||
> +						fsdata, folio->index, 1) ||
>   					 !f2fs_all_cluster_page_ready(&cc,
> -						pages, i, nr_pages, false))) {
> +						pages, i, nr_pages,
> +						false))) {
>   					retry = 1;
>   					break;
>   				}
> @@ -3059,46 +3089,47 @@ static int f2fs_write_cache_pages(struct address_space *mapping,
>   				break;
>   			}
>   #ifdef CONFIG_F2FS_FS_COMPRESSION
> -lock_page:
> +lock_folio:
>   #endif
> -			done_index = page->index;
> +			done_index = folio->index;
>   retry_write:
> -			lock_page(page);
> +			folio_lock(folio);
>   
> -			if (unlikely(page->mapping != mapping)) {
> +			if (unlikely(folio->mapping != mapping)) {
>   continue_unlock:
> -				unlock_page(page);
> +				folio_unlock(folio);
>   				continue;
>   			}
>   
> -			if (!PageDirty(page)) {
> +			if (!folio_test_dirty(folio)) {
>   				/* someone wrote it for us */
>   				goto continue_unlock;
>   			}
>   
> -			if (PageWriteback(page)) {
> +			if (folio_test_writeback(folio)) {
>   				if (wbc->sync_mode != WB_SYNC_NONE)
> -					f2fs_wait_on_page_writeback(page,
> +					f2fs_wait_on_page_writeback(
> +							&folio->page,
>   							DATA, true, true);
>   				else
>   					goto continue_unlock;
>   			}
>   
> -			if (!clear_page_dirty_for_io(page))
> +			if (!folio_clear_dirty_for_io(folio))
>   				goto continue_unlock;
>   
>   #ifdef CONFIG_F2FS_FS_COMPRESSION
>   			if (f2fs_compressed_file(inode)) {
> -				get_page(page);
> -				f2fs_compress_ctx_add_page(&cc, page);
> +				folio_get(folio);
> +				f2fs_compress_ctx_add_page(&cc, &folio->page);
>   				continue;
>   			}
>   #endif
> -			ret = f2fs_write_single_data_page(page, &submitted,
> -					&bio, &last_block, wbc, io_type,
> -					0, true);
> +			ret = f2fs_write_single_data_page(&folio->page,
> +					&submitted, &bio, &last_block,
> +					wbc, io_type, 0, true);
>   			if (ret == AOP_WRITEPAGE_ACTIVATE)
> -				unlock_page(page);
> +				folio_unlock(folio);
>   #ifdef CONFIG_F2FS_FS_COMPRESSION
>   result:
>   #endif
> @@ -3122,7 +3153,8 @@ static int f2fs_write_cache_pages(struct address_space *mapping,
>   					}
>   					goto next;
>   				}
> -				done_index = page->index + 1;
> +				done_index = folio->index +
> +					folio_nr_pages(folio);
>   				done = 1;
>   				break;
>   			}
> @@ -3136,7 +3168,7 @@ static int f2fs_write_cache_pages(struct address_space *mapping,
>   			if (need_readd)
>   				goto readd;
>   		}
> -		release_pages(pages, nr_pages);
> +		release_pages(pages,nr_pages);

No need to change?

Thanks,

>   		cond_resched();
>   	}
>   #ifdef CONFIG_F2FS_FS_COMPRESSION
  

On Thu, Dec 15, 2022 at 09:48:41AM +0800, Chao Yu wrote:
> On 2022/12/13 3:13, Vishal Moola (Oracle) wrote:
> > +add_more:
> > +			pages[nr_pages] = folio_page(folio,idx);
> > +			folio_ref_inc(folio);
> 
> It looks if CONFIG_LRU_GEN is not set, folio_ref_inc() does nothing. For those
> folios recorded in pages array, we need to call folio_get() here to add one more
> reference on each of them?

static inline void folio_get(struct folio *folio)
{
        VM_BUG_ON_FOLIO(folio_ref_zero_or_close_to_overflow(folio), folio);
        folio_ref_inc(folio);
}

That said, folio_ref_inct() is very much MM-internal and filesystems
should be using folio_get(), so please make that modification in the
next revision, Vishal.
  

On 12/12, Vishal Moola (Oracle) wrote:
> Converted the function to use a folio_batch instead of pagevec. This is in
> preparation for the removal of find_get_pages_range_tag().
> 
> Also modified f2fs_all_cluster_page_ready to take in a folio_batch instead
> of pagevec. This does NOT support large folios. The function currently
> only utilizes folios of size 1 so this shouldn't cause any issues right
> now.
> 
> This version of the patch limits the number of pages fetched to
> F2FS_ONSTACK_PAGES. If that ever happens, update the start index here
> since filemap_get_folios_tag() updates the index to be after the last
> found folio, not necessarily the last used page.
> 
> Signed-off-by: Vishal Moola (Oracle) <vishal.moola@gmail.com>
> ---
> 
> Let me know if you prefer this version and I'll include it in v5
> of the patch series when I rebase it after the merge window.
> 
> ---
>  fs/f2fs/data.c | 86 ++++++++++++++++++++++++++++++++++----------------
>  1 file changed, 59 insertions(+), 27 deletions(-)
> 
> diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c
> index a71e818cd67b..1703e353f0e0 100644
> --- a/fs/f2fs/data.c
> +++ b/fs/f2fs/data.c
> @@ -2939,6 +2939,7 @@ static int f2fs_write_cache_pages(struct address_space *mapping,
>  	int ret = 0;
>  	int done = 0, retry = 0;
>  	struct page *pages[F2FS_ONSTACK_PAGES];
> +	struct folio_batch fbatch;
>  	struct f2fs_sb_info *sbi = F2FS_M_SB(mapping);
>  	struct bio *bio = NULL;
>  	sector_t last_block;
> @@ -2959,6 +2960,7 @@ static int f2fs_write_cache_pages(struct address_space *mapping,
>  		.private = NULL,
>  	};
>  #endif
> +	int nr_folios, p, idx;
>  	int nr_pages;
>  	pgoff_t index;
>  	pgoff_t end;		/* Inclusive */
> @@ -2969,6 +2971,8 @@ static int f2fs_write_cache_pages(struct address_space *mapping,
>  	int submitted = 0;
>  	int i;
>  
> +	folio_batch_init(&fbatch);
> +
>  	if (get_dirty_pages(mapping->host) <=
>  				SM_I(F2FS_M_SB(mapping))->min_hot_blocks)
>  		set_inode_flag(mapping->host, FI_HOT_DATA);
> @@ -2994,13 +2998,38 @@ static int f2fs_write_cache_pages(struct address_space *mapping,
>  		tag_pages_for_writeback(mapping, index, end);
>  	done_index = index;
>  	while (!done && !retry && (index <= end)) {
> -		nr_pages = find_get_pages_range_tag(mapping, &index, end,
> -				tag, F2FS_ONSTACK_PAGES, pages);
> -		if (nr_pages == 0)
> +		nr_pages = 0;
> +again:
> +		nr_folios = filemap_get_folios_tag(mapping, &index, end,
> +				tag, &fbatch);

Can't folio handle this internally with F2FS_ONSTACK_PAGES and pages?

> +		if (nr_folios == 0) {
> +			if (nr_pages)
> +				goto write;
>  			break;
> +		}
>  
> +		for (i = 0; i < nr_folios; i++) {
> +			struct folio* folio = fbatch.folios[i];
> +
> +			idx = 0;
> +			p = folio_nr_pages(folio);
> +add_more:
> +			pages[nr_pages] = folio_page(folio,idx);
> +			folio_ref_inc(folio);
> +			if (++nr_pages == F2FS_ONSTACK_PAGES) {
> +				index = folio->index + idx + 1;
> +				folio_batch_release(&fbatch);
> +				goto write;
> +			}
> +			if (++idx < p)
> +				goto add_more;
> +		}
> +		folio_batch_release(&fbatch);
> +		goto again;
> +write:
>  		for (i = 0; i < nr_pages; i++) {
>  			struct page *page = pages[i];
> +			struct folio *folio = page_folio(page);
>  			bool need_readd;
>  readd:
>  			need_readd = false;
> @@ -3017,7 +3046,7 @@ static int f2fs_write_cache_pages(struct address_space *mapping,
>  				}
>  
>  				if (!f2fs_cluster_can_merge_page(&cc,
> -								page->index)) {
> +								folio->index)) {
>  					ret = f2fs_write_multi_pages(&cc,
>  						&submitted, wbc, io_type);
>  					if (!ret)
> @@ -3026,27 +3055,28 @@ static int f2fs_write_cache_pages(struct address_space *mapping,
>  				}
>  
>  				if (unlikely(f2fs_cp_error(sbi)))
> -					goto lock_page;
> +					goto lock_folio;
>  
>  				if (!f2fs_cluster_is_empty(&cc))
> -					goto lock_page;
> +					goto lock_folio;
>  
>  				if (f2fs_all_cluster_page_ready(&cc,
>  					pages, i, nr_pages, true))
> -					goto lock_page;
> +					goto lock_folio;
>  
>  				ret2 = f2fs_prepare_compress_overwrite(
>  							inode, &pagep,
> -							page->index, &fsdata);
> +							folio->index, &fsdata);
>  				if (ret2 < 0) {
>  					ret = ret2;
>  					done = 1;
>  					break;
>  				} else if (ret2 &&
>  					(!f2fs_compress_write_end(inode,
> -						fsdata, page->index, 1) ||
> +						fsdata, folio->index, 1) ||
>  					 !f2fs_all_cluster_page_ready(&cc,
> -						pages, i, nr_pages, false))) {
> +						pages, i, nr_pages,
> +						false))) {
>  					retry = 1;
>  					break;
>  				}
> @@ -3059,46 +3089,47 @@ static int f2fs_write_cache_pages(struct address_space *mapping,
>  				break;
>  			}
>  #ifdef CONFIG_F2FS_FS_COMPRESSION
> -lock_page:
> +lock_folio:
>  #endif
> -			done_index = page->index;
> +			done_index = folio->index;
>  retry_write:
> -			lock_page(page);
> +			folio_lock(folio);
>  
> -			if (unlikely(page->mapping != mapping)) {
> +			if (unlikely(folio->mapping != mapping)) {
>  continue_unlock:
> -				unlock_page(page);
> +				folio_unlock(folio);
>  				continue;
>  			}
>  
> -			if (!PageDirty(page)) {
> +			if (!folio_test_dirty(folio)) {
>  				/* someone wrote it for us */
>  				goto continue_unlock;
>  			}
>  
> -			if (PageWriteback(page)) {
> +			if (folio_test_writeback(folio)) {
>  				if (wbc->sync_mode != WB_SYNC_NONE)
> -					f2fs_wait_on_page_writeback(page,
> +					f2fs_wait_on_page_writeback(
> +							&folio->page,
>  							DATA, true, true);
>  				else
>  					goto continue_unlock;
>  			}
>  
> -			if (!clear_page_dirty_for_io(page))
> +			if (!folio_clear_dirty_for_io(folio))
>  				goto continue_unlock;
>  
>  #ifdef CONFIG_F2FS_FS_COMPRESSION
>  			if (f2fs_compressed_file(inode)) {
> -				get_page(page);
> -				f2fs_compress_ctx_add_page(&cc, page);
> +				folio_get(folio);
> +				f2fs_compress_ctx_add_page(&cc, &folio->page);
>  				continue;
>  			}
>  #endif
> -			ret = f2fs_write_single_data_page(page, &submitted,
> -					&bio, &last_block, wbc, io_type,
> -					0, true);
> +			ret = f2fs_write_single_data_page(&folio->page,
> +					&submitted, &bio, &last_block,
> +					wbc, io_type, 0, true);
>  			if (ret == AOP_WRITEPAGE_ACTIVATE)
> -				unlock_page(page);
> +				folio_unlock(folio);
>  #ifdef CONFIG_F2FS_FS_COMPRESSION
>  result:
>  #endif
> @@ -3122,7 +3153,8 @@ static int f2fs_write_cache_pages(struct address_space *mapping,
>  					}
>  					goto next;
>  				}
> -				done_index = page->index + 1;
> +				done_index = folio->index +
> +					folio_nr_pages(folio);
>  				done = 1;
>  				break;
>  			}
> @@ -3136,7 +3168,7 @@ static int f2fs_write_cache_pages(struct address_space *mapping,
>  			if (need_readd)
>  				goto readd;
>  		}
> -		release_pages(pages, nr_pages);
> +		release_pages(pages,nr_pages);
>  		cond_resched();
>  	}
>  #ifdef CONFIG_F2FS_FS_COMPRESSION
> -- 
> 2.38.1
  

On Thu, Dec 15, 2022 at 10:45 AM Matthew Wilcox <willy@infradead.org> wrote:
>
> On Thu, Dec 15, 2022 at 09:48:41AM +0800, Chao Yu wrote:
> > On 2022/12/13 3:13, Vishal Moola (Oracle) wrote:
> > > +add_more:
> > > +                   pages[nr_pages] = folio_page(folio,idx);
> > > +                   folio_ref_inc(folio);
> >
> > It looks if CONFIG_LRU_GEN is not set, folio_ref_inc() does nothing. For those
> > folios recorded in pages array, we need to call folio_get() here to add one more
> > reference on each of them?
>
> static inline void folio_get(struct folio *folio)
> {
>         VM_BUG_ON_FOLIO(folio_ref_zero_or_close_to_overflow(folio), folio);
>         folio_ref_inc(folio);
> }
>
> That said, folio_ref_inct() is very much MM-internal and filesystems
> should be using folio_get(), so please make that modification in the
> next revision, Vishal.

Ok, I'll go through and fix all of those in the next version.
  

On Wed, Dec 21, 2022 at 09:17:30AM -0800, Vishal Moola wrote:
> > That said, folio_ref_inct() is very much MM-internal and filesystems
> > should be using folio_get(), so please make that modification in the
> > next revision, Vishal.
> 
> Ok, I'll go through and fix all of those in the next version.

Btw, something a lot more productive in this area would be to figure out
how we could convert all these copy and paste versions of
write_cache_pages to use common code.  This might need changes to the
common code, but the amount of duplicate and poorly maintained versions
of this loop is a bit alarming:

 - btree_write_cache_pages
 - extent_write_cache_pages
 - f2fs_write_cache_pages
 - gfs2_write_cache_jdata
  

On Thu, Dec 15, 2022 at 11:02:24AM -0800, Jaegeuk Kim wrote:
> On 12/12, Vishal Moola (Oracle) wrote:
> > @@ -2994,13 +2998,38 @@ static int f2fs_write_cache_pages(struct address_space *mapping,
> >  		tag_pages_for_writeback(mapping, index, end);
> >  	done_index = index;
> >  	while (!done && !retry && (index <= end)) {
> > -		nr_pages = find_get_pages_range_tag(mapping, &index, end,
> > -				tag, F2FS_ONSTACK_PAGES, pages);
> > -		if (nr_pages == 0)
> > +		nr_pages = 0;
> > +again:
> > +		nr_folios = filemap_get_folios_tag(mapping, &index, end,
> > +				tag, &fbatch);
> 
> Can't folio handle this internally with F2FS_ONSTACK_PAGES and pages?

I really want to discourage filesystems from doing this kind of thing.
The folio_batch is the natural size for doing batches of work, and
having the consistency across all these APIs of passing in a folio_batch
is quite valuable.  I understand f2fs wants to get more memory in a
single batch, but the right way to do that is to use larger folios.