[PATCH 0/5] Cleanups for swap

[PATCH 0/5] Cleanups for swap

[Miaohe Lin]

Hi all,
This series contains just cleanups to remove some unused variables, delete
meaningless forward declarations and so on. More details can be found in
the respective changelogs. Thanks!

Miaohe Lin (5):
  mm/swapfile: move get_swap_page_of_type() under CONFIG_HIBERNATION
  mm/swap: remove unused local variable nr_shadows
  mm/swap_slots.c: delete meaningless forward declarations
  mm/swap: remove unused global variable nr_swapper_spaces
  mm/swap: simplify the code of find_get_incore_page()

 mm/swap_slots.c |  2 --
 mm/swap_state.c | 15 +++------------
 mm/swapfile.c   | 48 ++++++++++++++++++++++++------------------------
 3 files changed, 27 insertions(+), 38 deletions(-)

-- 
2.23.0

[PATCH 1/5] mm/swapfile: move get_swap_page_of_type() under CONFIG_HIBERNATION

[Miaohe Lin]

We should move get_swap_page_of_type() under CONFIG_HIBERNATION since the
only caller of this function is now suspend routine.

Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
---
 mm/swapfile.c | 48 ++++++++++++++++++++++++------------------------
 1 file changed, 24 insertions(+), 24 deletions(-)

diff --git a/mm/swapfile.c b/mm/swapfile.c
index 2aad85751991..4e7628b82e63 100644
--- a/mm/swapfile.c
+++ b/mm/swapfile.c
@@ -1137,30 +1137,6 @@ int get_swap_pages(int n_goal, swp_entry_t swp_entries[], int entry_size)
 	return n_ret;
 }
 
-/* The only caller of this function is now suspend routine */
-swp_entry_t get_swap_page_of_type(int type)
-{
-	struct swap_info_struct *si = swap_type_to_swap_info(type);
-	pgoff_t offset;
-
-	if (!si)
-		goto fail;
-
-	spin_lock(&si->lock);
-	if (si->flags & SWP_WRITEOK) {
-		/* This is called for allocating swap entry, not cache */
-		offset = scan_swap_map(si, 1);
-		if (offset) {
-			atomic_long_dec(&nr_swap_pages);
-			spin_unlock(&si->lock);
-			return swp_entry(type, offset);
-		}
-	}
-	spin_unlock(&si->lock);
-fail:
-	return (swp_entry_t) {0};
-}
-
 static struct swap_info_struct *__swap_info_get(swp_entry_t entry)
 {
 	struct swap_info_struct *p;
@@ -1812,6 +1788,30 @@ int free_swap_and_cache(swp_entry_t entry)
 }
 
 #ifdef CONFIG_HIBERNATION
+
+swp_entry_t get_swap_page_of_type(int type)
+{
+	struct swap_info_struct *si = swap_type_to_swap_info(type);
+	pgoff_t offset;
+
+	if (!si)
+		goto fail;
+
+	spin_lock(&si->lock);
+	if (si->flags & SWP_WRITEOK) {
+		/* This is called for allocating swap entry, not cache */
+		offset = scan_swap_map(si, 1);
+		if (offset) {
+			atomic_long_dec(&nr_swap_pages);
+			spin_unlock(&si->lock);
+			return swp_entry(type, offset);
+		}
+	}
+	spin_unlock(&si->lock);
+fail:
+	return (swp_entry_t) {0};
+}
+
 /*
  * Find the swap type that corresponds to given device (if any).
  *
-- 
2.23.0

[PATCH 2/5] mm/swap: remove unused local variable nr_shadows

[Miaohe Lin]

Since commit 55c653b71e8c ("mm: stop accounting shadow entries"),
nr_shadows is not used anymore.

Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
---
 mm/swap_state.c | 4 ----
 1 file changed, 4 deletions(-)

diff --git a/mm/swap_state.c b/mm/swap_state.c
index df5405384520..bab386c7b655 100644
--- a/mm/swap_state.c
+++ b/mm/swap_state.c
@@ -114,7 +114,6 @@ int add_to_swap_cache(struct page *page, swp_entry_t entry,
 	SetPageSwapCache(page);
 
 	do {
-		unsigned long nr_shadows = 0;
 
 		xas_lock_irq(&xas);
 		xas_create_range(&xas);
@@ -124,7 +123,6 @@ int add_to_swap_cache(struct page *page, swp_entry_t entry,
 			VM_BUG_ON_PAGE(xas.xa_index != idx + i, page);
 			old = xas_load(&xas);
 			if (xa_is_value(old)) {
-				nr_shadows++;
 				if (shadowp)
 					*shadowp = old;
 			}
@@ -260,7 +258,6 @@ void clear_shadow_from_swap_cache(int type, unsigned long begin,
 	void *old;
 
 	for (;;) {
-		unsigned long nr_shadows = 0;
 		swp_entry_t entry = swp_entry(type, curr);
 		struct address_space *address_space = swap_address_space(entry);
 		XA_STATE(xas, &address_space->i_pages, curr);
@@ -270,7 +267,6 @@ void clear_shadow_from_swap_cache(int type, unsigned long begin,
 			if (!xa_is_value(old))
 				continue;
 			xas_store(&xas, NULL);
-			nr_shadows++;
 		}
 		xa_unlock_irq(&address_space->i_pages);
 
-- 
2.23.0

Re: [PATCH 2/5] mm/swap: remove unused local variable nr_shadows

[Matthew Wilcox]

On Tue, May 18, 2021 at 09:53:49PM +0800, Miaohe Lin wrote:
> Since commit 55c653b71e8c ("mm: stop accounting shadow entries"),
> nr_shadows is not used anymore.
> 
> Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>

Reviewed-by: Matthew Wilcox (Oracle) <willy@infradead.org>

> +++ b/mm/swap_state.c
> @@ -114,7 +114,6 @@ int add_to_swap_cache(struct page *page, swp_entry_t entry,
>  	SetPageSwapCache(page);
>  
>  	do {
> -		unsigned long nr_shadows = 0;
>  

Perhaps delete the blank line as well?

Re: [PATCH 2/5] mm/swap: remove unused local variable nr_shadows

[Miaohe Lin]

On 2021/5/18 22:37, Matthew Wilcox wrote:
> On Tue, May 18, 2021 at 09:53:49PM +0800, Miaohe Lin wrote:
>> Since commit 55c653b71e8c ("mm: stop accounting shadow entries"),
>> nr_shadows is not used anymore.
>>
>> Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
> 
> Reviewed-by: Matthew Wilcox (Oracle) <willy@infradead.org>
> 
>> +++ b/mm/swap_state.c
>> @@ -114,7 +114,6 @@ int add_to_swap_cache(struct page *page, swp_entry_t entry,
>>  	SetPageSwapCache(page);
>>  
>>  	do {
>> -		unsigned long nr_shadows = 0;
>>  
> 
> Perhaps delete the blank line as well?

Will do. Many thanks for the review.

> 
> .
> 

[PATCH 3/5] mm/swap_slots.c: delete meaningless forward declarations

[Miaohe Lin]

deactivate_swap_slots_cache() and reactivate_swap_slots_cache() are only
called below their implementations. So these forward declarations are
meaningless and should be removed.

Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
---
 mm/swap_slots.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/mm/swap_slots.c b/mm/swap_slots.c
index 6248d1030a9b..a66f3e0ec973 100644
--- a/mm/swap_slots.c
+++ b/mm/swap_slots.c
@@ -43,8 +43,6 @@ static DEFINE_MUTEX(swap_slots_cache_mutex);
 static DEFINE_MUTEX(swap_slots_cache_enable_mutex);
 
 static void __drain_swap_slots_cache(unsigned int type);
-static void deactivate_swap_slots_cache(void);
-static void reactivate_swap_slots_cache(void);
 
 #define use_swap_slot_cache (swap_slot_cache_active && swap_slot_cache_enabled)
 #define SLOTS_CACHE 0x1
-- 
2.23.0

[PATCH 4/5] mm/swap: remove unused global variable nr_swapper_spaces

[Miaohe Lin]

Since commit b6038942480e ("mm: memcg: add swapcache stat for memcg v2"),
the last user of nr_swapper_spaces is gone. So we remove this unused one.

Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
---
 mm/swap_state.c | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/mm/swap_state.c b/mm/swap_state.c
index bab386c7b655..a55b7b74b0e6 100644
--- a/mm/swap_state.c
+++ b/mm/swap_state.c
@@ -37,7 +37,6 @@ static const struct address_space_operations swap_aops = {
 };
 
 struct address_space *swapper_spaces[MAX_SWAPFILES] __read_mostly;
-static unsigned int nr_swapper_spaces[MAX_SWAPFILES] __read_mostly;
 static bool enable_vma_readahead __read_mostly = true;
 
 #define SWAP_RA_WIN_SHIFT	(PAGE_SHIFT / 2)
@@ -686,7 +685,6 @@ int init_swap_address_space(unsigned int type, unsigned long nr_pages)
 		/* swap cache doesn't use writeback related tags */
 		mapping_set_no_writeback_tags(space);
 	}
-	nr_swapper_spaces[type] = nr;
 	swapper_spaces[type] = spaces;
 
 	return 0;
@@ -695,7 +693,6 @@ int init_swap_address_space(unsigned int type, unsigned long nr_pages)
 void exit_swap_address_space(unsigned int type)
 {
 	kvfree(swapper_spaces[type]);
-	nr_swapper_spaces[type] = 0;
 	swapper_spaces[type] = NULL;
 }
 
-- 
2.23.0

[PATCH 5/5] mm/swap: simplify the code of find_get_incore_page()

[Miaohe Lin]

pagecache_get_page() can do find_subpage() for us if we do not specify
FGP_HEAD. No functional change intended.

Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
---
 mm/swap_state.c | 8 +++-----
 1 file changed, 3 insertions(+), 5 deletions(-)

diff --git a/mm/swap_state.c b/mm/swap_state.c
index a55b7b74b0e6..bc92e4893fec 100644
--- a/mm/swap_state.c
+++ b/mm/swap_state.c
@@ -394,13 +394,11 @@ struct page *find_get_incore_page(struct address_space *mapping, pgoff_t index)
 {
 	swp_entry_t swp;
 	struct swap_info_struct *si;
-	struct page *page = pagecache_get_page(mapping, index,
-						FGP_ENTRY | FGP_HEAD, 0);
+	struct page *page = pagecache_get_page(mapping, index, FGP_ENTRY, 0);
 
-	if (!page)
+	if (!page || !xa_is_value(page))
 		return page;
-	if (!xa_is_value(page))
-		return find_subpage(page, index);
+
 	if (!shmem_mapping(mapping))
 		return NULL;
 
-- 
2.23.0

Re: [PATCH 5/5] mm/swap: simplify the code of find_get_incore_page()

[Matthew Wilcox]

On Tue, May 18, 2021 at 09:53:52PM +0800, Miaohe Lin wrote:
> pagecache_get_page() can do find_subpage() for us if we do not specify
> FGP_HEAD. No functional change intended.

Please, no.  This interferes with the folio work.

Re: [PATCH 5/5] mm/swap: simplify the code of find_get_incore_page()

[Andrew Morton]

On Tue, 18 May 2021 15:33:54 +0100 Matthew Wilcox <willy@infradead.org> wrote:

> On Tue, May 18, 2021 at 09:53:52PM +0800, Miaohe Lin wrote:
> > pagecache_get_page() can do find_subpage() for us if we do not specify
> > FGP_HEAD. No functional change intended.
> 
> Please, no.  This interferes with the folio work.

In what way?  Can't the folio patches simply revert this or suitably
alter it?

Re: [PATCH 5/5] mm/swap: simplify the code of find_get_incore_page()

[Matthew Wilcox]

On Tue, May 18, 2021 at 02:13:04PM -0700, Andrew Morton wrote:
> On Tue, 18 May 2021 15:33:54 +0100 Matthew Wilcox <willy@infradead.org> wrote:
> 
> > On Tue, May 18, 2021 at 09:53:52PM +0800, Miaohe Lin wrote:
> > > pagecache_get_page() can do find_subpage() for us if we do not specify
> > > FGP_HEAD. No functional change intended.
> > 
> > Please, no.  This interferes with the folio work.
> 
> In what way?  Can't the folio patches simply revert this or suitably
> alter it?

Of course, it's just software.  Anything can just be modified.  I don't
see the point of putting in a cleanup patch that creates a conflict with
important work.

Re: [PATCH 5/5] mm/swap: simplify the code of find_get_incore_page()

[Miaohe Lin]

On 2021/5/19 5:34, Matthew Wilcox wrote:
> On Tue, May 18, 2021 at 02:13:04PM -0700, Andrew Morton wrote:
>> On Tue, 18 May 2021 15:33:54 +0100 Matthew Wilcox <willy@infradead.org> wrote:
>>
>>> On Tue, May 18, 2021 at 09:53:52PM +0800, Miaohe Lin wrote:
>>>> pagecache_get_page() can do find_subpage() for us if we do not specify
>>>> FGP_HEAD. No functional change intended.
>>>
>>> Please, no.  This interferes with the folio work.
>>
>> In what way?  Can't the folio patches simply revert this or suitably
>> alter it?
> 
> Of course, it's just software.  Anything can just be modified.  I don't
> see the point of putting in a cleanup patch that creates a conflict with
> important work.

I could hold this patch out until folio work is done if you ask for it.
Thanks.

> .
>