Re: [f2fs-dev] [RFC PATCH v2 01/10] f2fs: extend folio state for large folio write path

[Chao Yu via Linux-f2fs-devel <linux-f2fs-devel@lists.sourceforge.net>]

On 6/23/26 00:08, Nanzhe Zhao wrote:
> Large folio write path needs a subpage status bitmap and write
> pages pending counter, while keeping compatible with f2fs private
> flags.
> 
> Move struct f2fs_folio_state to f2fs.h, add private_flags and
> subpage state bitmap, and change PAGE_PRIVATE functions to be
> compatible with f2fs_folio_state. Allocate f2fs_folio_state via kzalloc
> instead of kmem_cache, since the state size depends on the folio order.
> 
> Note: Now if a path wants to use f2fs_folio_state, it must call
> `folio_has_ffs` instead of `folio_test_large`` to make check.
> 
> Signed-off-by: Nanzhe Zhao <zhaonanzhe@xiaomi.com>
> ---
>  fs/f2fs/compress.c |  2 ++
>  fs/f2fs/data.c     | 46 ++++++++++++++++----------------
>  fs/f2fs/f2fs.h     | 66 ++++++++++++++++++++++++++++++++++++++--------
>  3 files changed, 80 insertions(+), 34 deletions(-)
> 
> diff --git a/fs/f2fs/compress.c b/fs/f2fs/compress.c
> index 91855d91bbdd..de1305fcc6c1 100644
> --- a/fs/f2fs/compress.c
> +++ b/fs/f2fs/compress.c
> @@ -76,6 +76,8 @@ bool f2fs_is_compressed_page(struct folio *folio)
>  {
>  	if (!folio->private)
>  		return false;
> +	if (folio_has_ffs(folio))
> +		return false;
>  	if (folio_test_f2fs_nonpointer(folio))
>  		return false;

private is a pointer for compressed page, so it will be better to
check folio_test_f2fs_nonpointer() first, and then folio_has_ffs().

if (folio_test_f2fs_nonpointer(folio))
	return false;
if (folio_has_ffs(folio))
	return false;

>  
> diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c
> index ac1cf4de3d62..23758c00758d 100644
> --- a/fs/f2fs/data.c
> +++ b/fs/f2fs/data.c
> @@ -32,20 +32,13 @@
>  
>  static struct kmem_cache *bio_post_read_ctx_cache;
>  static struct kmem_cache *bio_entry_slab;
> -static struct kmem_cache *ffs_entry_slab;
>  static mempool_t *bio_post_read_ctx_pool;
>  static struct bio_set f2fs_bioset;
>  
> -struct f2fs_folio_state {
> -	spinlock_t		state_lock;
> -	unsigned int		read_pages_pending;
> -};
> -
>  struct f2fs_bio {
>  	struct work_struct work;
>  	struct bio bio;
>  };
> -
>  #define	F2FS_BIO_POOL_SIZE	NR_CURSEG_TYPE
>  
>  int __init f2fs_init_bioset(void)
> @@ -2514,15 +2507,30 @@ int f2fs_read_multi_pages(struct compress_ctx *cc, struct bio **bio_ret,
>  
>  static struct f2fs_folio_state *ffs_find_or_alloc(struct folio *folio)
>  {
> -	struct f2fs_folio_state *ffs = folio->private;
> +	struct f2fs_folio_state *ffs;
> +	unsigned int nr_subpages = folio_nr_pages(folio);
> +	unsigned long private_flags = 0;
>  
> -	if (ffs)
> -		return ffs;
> +	f2fs_bug_on(F2FS_F_SB(folio), !folio_test_large(folio));
>  
> -	ffs = f2fs_kmem_cache_alloc(ffs_entry_slab,
> -			GFP_NOIO | __GFP_ZERO, true, NULL);
> +	if (folio_has_ffs(folio))
> +		return (struct f2fs_folio_state *)folio->private;
> +
> +	if (folio_test_private(folio) && folio_test_f2fs_nonpointer(folio))

Why this can happen? allocating large folio from other paths?

> +		private_flags = (unsigned long)folio->private;
> +
> +	ffs = kzalloc(struct_size(ffs, state, BITS_TO_LONGS(2 * nr_subpages)),
> +			GFP_NOIO | __GFP_NOFAIL);
>  
>  	spin_lock_init(&ffs->state_lock);
> +	ffs->private_flags = private_flags;
> +	if (folio_test_uptodate(folio))
> +		bitmap_set(ffs->state, 0, nr_subpages);
> +	if (folio_test_dirty(folio))
> +		bitmap_set(ffs->state, nr_subpages, nr_subpages);

Can you describe ffs->state layout in its definition?

> +
> +	if (folio_test_private(folio))
> +		folio_detach_private(folio);
>  	folio_attach_private(folio, ffs);
>  	return ffs;
>  }
> @@ -2531,7 +2539,7 @@ static void ffs_detach_free(struct folio *folio)
>  {
>  	struct f2fs_folio_state *ffs;
>  
> -	if (!folio_test_large(folio)) {
> +	if (!folio_has_ffs(folio)) {
>  		folio_detach_private(folio);
>  		return;
>  	}
> @@ -2541,7 +2549,8 @@ static void ffs_detach_free(struct folio *folio)
>  		return;
>  
>  	WARN_ON_ONCE(ffs->read_pages_pending != 0);
> -	kmem_cache_free(ffs_entry_slab, ffs);
> +	WARN_ON_ONCE(atomic_read(&ffs->write_pages_pending));
> +	kfree(ffs);
>  }
>  
>  static int f2fs_read_data_large_folio(struct inode *inode,
> @@ -4571,21 +4580,12 @@ int __init f2fs_init_bio_entry_cache(void)
>  	if (!bio_entry_slab)
>  		return -ENOMEM;
>  
> -	ffs_entry_slab = f2fs_kmem_cache_create("f2fs_ffs_slab",
> -			sizeof(struct f2fs_folio_state));
> -
> -	if (!ffs_entry_slab) {
> -		kmem_cache_destroy(bio_entry_slab);
> -		return -ENOMEM;
> -	}
> -
>  	return 0;
>  }
>  
>  void f2fs_destroy_bio_entry_cache(void)
>  {
>  	kmem_cache_destroy(bio_entry_slab);
> -	kmem_cache_destroy(ffs_entry_slab);
>  }
>  
>  static int f2fs_iomap_begin(struct inode *inode, loff_t offset, loff_t length,
> diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
> index f1774d4e18d2..e4778c17394e 100644
> --- a/fs/f2fs/f2fs.h
> +++ b/fs/f2fs/f2fs.h
> @@ -1620,6 +1620,15 @@ static inline void f2fs_clear_bit(unsigned int nr, char *addr);
>   * Layout B: lowest bit should be 0
>   * page.private is a wrapped pointer.
>   */
> +
> +struct f2fs_folio_state {
> +	spinlock_t		state_lock;
> +	unsigned int		read_pages_pending;
> +	atomic_t		write_pages_pending;
> +	unsigned long		private_flags;
> +	unsigned long		state[];
> +};
> +
>  enum {
>  	PAGE_PRIVATE_NOT_POINTER,		/* private contains non-pointer data */
>  	PAGE_PRIVATE_ONGOING_MIGRATION,		/* data page which is on-going migrating */
> @@ -1629,6 +1638,14 @@ enum {
>  	PAGE_PRIVATE_MAX
>  };
>  
> +static inline bool folio_has_ffs(const struct folio *folio)
> +{
> +	unsigned long private = (unsigned long)folio->private;
> +
> +	return folio_test_large(folio) && private &&
> +		!(private & BIT(PAGE_PRIVATE_NOT_POINTER));
> +}
> +
>  /* For compression */
>  enum compress_algorithm_type {
>  	COMPRESS_LZO,
> @@ -2638,9 +2655,15 @@ static inline int inc_valid_block_count(struct f2fs_sb_info *sbi,
>  #define PAGE_PRIVATE_GET_FUNC(name, flagname) \
>  static inline bool folio_test_f2fs_##name(const struct folio *folio)	\

Seems there are redundant codes below, let's have a try to wrap them w/ a macro for cleanup?

>  {									\
> -	unsigned long priv = (unsigned long)folio->private;		\
> +	unsigned long priv;						\
>  	unsigned long v = (1UL << PAGE_PRIVATE_NOT_POINTER) |		\
>  			     (1UL << PAGE_PRIVATE_##flagname);		\
> +	if (folio_has_ffs(folio)) {					\
> +		struct f2fs_folio_state *ffs = folio->private;		\
> +		priv = ffs->private_flags;				\
> +	} else {							\
> +		priv = (unsigned long)folio->private;			\
> +	}								\
>  	return (priv & v) == v;						\
>  }									\
>  static inline bool page_private_##name(struct page *page) \
> @@ -2655,7 +2678,10 @@ static inline void folio_set_f2fs_##name(struct folio *folio)		\
>  {									\
>  	unsigned long v = (1UL << PAGE_PRIVATE_NOT_POINTER) |		\
>  			     (1UL << PAGE_PRIVATE_##flagname);		\
> -	if (!folio->private)						\
> +	if (folio_has_ffs(folio)) {					\
> +		struct f2fs_folio_state *ffs = folio->private;		\
> +		ffs->private_flags |= v;				\
> +	} else if (!folio->private)					\
>  		folio_attach_private(folio, (void *)v);			\

Ditto?

>  	else {								\
>  		v |= (unsigned long)folio->private;			\
> @@ -2673,13 +2699,18 @@ static inline void set_page_private_##name(struct page *page) \
>  #define PAGE_PRIVATE_CLEAR_FUNC(name, flagname) \
>  static inline void folio_clear_f2fs_##name(struct folio *folio)		\
>  {									\
> -	unsigned long v = (unsigned long)folio->private;		\
> +	if (folio_has_ffs(folio)) {					\
> +		struct f2fs_folio_state *ffs = folio->private;		\
> +		ffs->private_flags &= ~(1UL << PAGE_PRIVATE_##flagname); \
> +	} else {							\
> +		unsigned long v = (unsigned long)folio->private;	\
>  									\
> -	v &= ~(1UL << PAGE_PRIVATE_##flagname);				\
> -	if (v == (1UL << PAGE_PRIVATE_NOT_POINTER))			\
> -		folio_detach_private(folio);				\
> -	else								\
> -		folio->private = (void *)v;				\
> +		v &= ~(1UL << PAGE_PRIVATE_##flagname);		\
> +		if (v == (1UL << PAGE_PRIVATE_NOT_POINTER))	\
> +			folio_detach_private(folio);			\
> +		else							\
> +			folio->private = (void *)v;			\
> +	}								\
>  }									\
>  static inline void clear_page_private_##name(struct page *page) \
>  { \
> @@ -2705,7 +2736,15 @@ PAGE_PRIVATE_CLEAR_FUNC(atomic, ATOMIC_WRITE);
>  
>  static inline unsigned long folio_get_f2fs_data(struct folio *folio)
>  {
> -	unsigned long data = (unsigned long)folio->private;
> +	unsigned long data;
> +
> +	if (folio_has_ffs(folio)) {
> +		struct f2fs_folio_state *ffs = folio->private;
> +
> +		data = ffs->private_flags;
> +	} else {
> +		data = (unsigned long)folio->private;
> +	}

Ditto?

>  
>  	if (!test_bit(PAGE_PRIVATE_NOT_POINTER, &data))
>  		return 0;
> @@ -2716,10 +2755,15 @@ static inline void folio_set_f2fs_data(struct folio *folio, unsigned long data)
>  {
>  	data = (1UL << PAGE_PRIVATE_NOT_POINTER) | (data << PAGE_PRIVATE_MAX);
>  
> -	if (!folio_test_private(folio))
> +	if (folio_has_ffs(folio)) {
> +		struct f2fs_folio_state *ffs = folio->private;
> +
> +		ffs->private_flags |= data;
> +	} else if (!folio_test_private(folio)) {
>  		folio_attach_private(folio, (void *)data);

Ditto?

Thanks,

> -	else
> +	} else {
>  		folio->private = (void *)((unsigned long)folio->private | data);
> +	}
>  }
>  
>  static inline void dec_valid_block_count(struct f2fs_sb_info *sbi,



_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel