Re: [PATCH 7/8] mm/swap: remove SWP_FS_OPS

[Baoquan He <baoquan.he@linux.dev>]

On 06/01/26 at 01:34pm, Christoph Hellwig wrote:
...snip...
> diff --git a/mm/swap.h b/mm/swap.h
> index 5119feff0e93..edb512e619ee 100644
> --- a/mm/swap.h
> +++ b/mm/swap.h
> @@ -84,7 +84,11 @@ struct swap_io_ctx {
>  	struct swap_info_struct	*sis;
>  };
>  
> +#define SWAP_OPS_F_NOFS		(1U << 0)
> +

The name of this flag and its usage are a little weird, surely the old
one SWP_FS_OPS is similar. Do you think if we need add some document
as below to make it clearer? Personal opinion and just for reference.

diff --git a/mm/swap.h b/mm/swap.h
index edb512e619ee..3d978382a7d5 100644
--- a/mm/swap.h
+++ b/mm/swap.h
@@ -84,6 +84,18 @@ struct swap_io_ctx {
 	struct swap_info_struct	*sis;
 };
 
+/*
+ * SWAP_OPS_F_NOFS: swap operations that may recurse into the filesystem
+ * (e.g. NFS, CIFS).  When set, callers (specifically reclaim) must operate
+ * under GFP_NOFS / !__GFP_FS to avoid deadlock.  When clear (block-device
+ * swap), the swap I/O path does not touch the filesystem, so reclaim may
+ * safely enter the filesystem even with only __GFP_IO.
+ *
+ * In may_enter_fs() this appears as a double negative (!SWAP_OPS_F_NOFS)
+ * which reads as "not a nofs swap => may enter fs".  This is intentional:
+ * it is bdev swap (the common case) that allows filesystem re-entry, while
+ * filesystem-based swap explicitly forbids it.
+ */
 #define SWAP_OPS_F_NOFS		(1U << 0)
 
 struct swap_ops {
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 2d44ebfebdea..a92956de5e10 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -1039,7 +1039,9 @@ static bool may_enter_fs(struct folio *folio, gfp_t gfp_mask)
 		return true;
 	/*
 	 * We can "enter_fs" for swap-cache with only __GFP_IO unless backed by
-	 * a swapfile that requires GFP_NOFS I/O.
+	 * a swapfile that requires GFP_NOFS I/O (i.e. filesystem-based swap
+	 * like NFS/CIFS, which sets SWAP_OPS_F_NOFS).  Block-device swap
+	 * never touches the filesystem, so it is safe to enter fs here.
 	 */
 	if (folio_test_swapcache(folio) && (gfp_mask & __GFP_IO) &&
 	    !(__swap_entry_to_info(folio->swap)->ops->flags & SWAP_OPS_F_NOFS))

>  struct swap_ops {
> +	unsigned int		flags;
> +
>  	bool (*can_merge)(struct folio *folio, struct folio *prev_folio,
>  			size_t prev_folio_size, int rw);
>  	void (*submit_write)(struct swap_io_ctx *ctx);
> @@ -335,11 +339,6 @@ struct folio *swapin_sync(swp_entry_t entry, gfp_t flag, unsigned long orders,
>  void swap_update_readahead(struct folio *folio, struct vm_area_struct *vma,
>  			   unsigned long addr);
>  
> -static inline unsigned int folio_swap_flags(struct folio *folio)
> -{
> -	return __swap_entry_to_info(folio->swap)->flags;
> -}
> -
>  #else /* CONFIG_SWAP */
>  static inline struct swap_cluster_info *swap_cluster_lock(
>  	struct swap_info_struct *si, pgoff_t offset, bool irq)
> @@ -470,16 +469,9 @@ static inline void __swap_cache_replace_folio(struct swap_cluster_info *ci,
>  		struct folio *old, struct folio *new)
>  {
>  }
> -
> -static inline unsigned int folio_swap_flags(struct folio *folio)
> -{
> -	return 0;
> -}
> -
>  #endif /* CONFIG_SWAP */
>  
>  extern const struct swap_ops swap_bdev_ops;
> -extern const struct swap_ops swap_fs_ops;
>  
>  int shmem_writeout(struct swap_io_ctx *ctx, struct folio *folio,
>  		struct list_head *folio_list);
> diff --git a/mm/swapfile.c b/mm/swapfile.c
> index a670635e0fbe..284eebc40a70 100644
> --- a/mm/swapfile.c
> +++ b/mm/swapfile.c
> @@ -2858,8 +2858,6 @@ static int setup_swap_extents(struct swap_info_struct *sis,
>  		ret = mapping->a_ops->swap_activate(sis, swap_file, span);
>  		if (ret < 0)
>  			return ret;
> -		if (sis->flags & SWP_FS_OPS)
> -			sis->ops = &swap_fs_ops;
>  		sis->flags |= SWP_ACTIVATED;
>  		return ret;
>  	}
> diff --git a/mm/vmscan.c b/mm/vmscan.c
> index c43177a8e4dd..2d44ebfebdea 100644
> --- a/mm/vmscan.c
> +++ b/mm/vmscan.c
> @@ -1037,16 +1037,14 @@ static bool may_enter_fs(struct folio *folio, gfp_t gfp_mask)
>  {
>  	if (gfp_mask & __GFP_FS)
>  		return true;
> -	if (!folio_test_swapcache(folio) || !(gfp_mask & __GFP_IO))
> -		return false;
>  	/*
> -	 * We can "enter_fs" for swap-cache with only __GFP_IO
> -	 * providing this isn't SWP_FS_OPS.
> -	 * ->flags can be updated non-atomically,
> -	 * but that will never affect SWP_FS_OPS, so the data_race
> -	 * is safe.
> +	 * We can "enter_fs" for swap-cache with only __GFP_IO unless backed by
> +	 * a swapfile that requires GFP_NOFS I/O.
>  	 */
> -	return !data_race(folio_swap_flags(folio) & SWP_FS_OPS);
> +	if (folio_test_swapcache(folio) && (gfp_mask & __GFP_IO) &&
> +	    !(__swap_entry_to_info(folio->swap)->ops->flags & SWAP_OPS_F_NOFS))
> +		return true;
> +	return false;
>  }
>  
>  /*
> -- 
> 2.53.0
>