Skip to contentSkip to conversation

mm: swap: small fixes and comment cleanups

PATCH Applied Youngjun Park 4 participants 22 messages Oct 31, 2025 – Nov 20, 2025

Subsystems (2)
lore
Share & tools
lore

Youngjun Park

author reviewer
#0 lore raw

v2 cover mm: swap: small fixes and comment cleanups

Hello,
This series provides a few small fixes and cleanups for the swap code.

The first patch fixes a memory leak in an error path that was recently introduced. The subsequent patches include minor logic adjustments and the removal of redundant comments.

This version addresses feedback from Kairui Song. Thank you for the review.

---
v1->v2:
- Reworked the error path in setup_clusters() to use a single exit label as suggested by Kairui.
- Added a Fixes: tag to the first patch to identify the commit that introduced.
- Added Kairui's Reviewed-by tag.

Link to v1:
https://lore.kernel.org/linux-mm/[email protected]/
---

Youngjun Park (5):
mm, swap: Fix memory leak in setup_clusters() error path mm, swap: Use SWP_SOLIDSTATE to determine if swap is rotational mm, swap: Remove redundant comment for read_swap_cache_async
mm: swap: change swap_alloc_slow() to void
mm: swap: remove scan_swap_map_slots() references from comments

 mm/swap_state.c |  4 ----
 mm/swapfile.c   | 34 +++++++++++++++-------------------
 2 files changed, 15 insertions(+), 23 deletions(-)

base-commit: df412f2f12d492234da858e4dbaeabed9ecff3c4 (mm-new)

Signature · 1 line
2.34.1
author reviewer
#1 lore raw

PATCH v2 1/5 mm, swap: Fix memory leak in setup_clusters() error path

The setup_clusters() function could leak 'cluster_info' memory if an error occurred on a path that did not jump to the 'err_free' label.

This patch simplifies the error handling by removing the goto label and instead calling free_cluster_info() on all error exit paths.

The new logic is safe, as free_cluster_info() already handles NULL pointer inputs.

Fixes 07adc4cf1ecd ("mm, swap: implement dynamic allocation of swap table")
mm/swapfile.c +2 -3
index c35bb8593f50..ed2c8e4edb71 100644
@@ -3324,7 +3324,7 @@ static struct swap_cluster_info *setup_clusters(struct swap_info_struct *si,
 		si->global_cluster = kmalloc(sizeof(*si->global_cluster),
 				     GFP_KERNEL);
 		if (!si->global_cluster)
-			goto err_free;
+			goto err;
 		for (i = 0; i < SWAP_NR_ORDERS; i++)
 			si->global_cluster->next[i] = SWAP_ENTRY_INVALID;
 		spin_lock_init(&si->global_cluster_lock);
@@ -3377,9 +3377,8 @@ static struct swap_cluster_info *setup_clusters(struct swap_info_struct *si,
 	}
 
 	return cluster_info;
-err_free:
-	free_cluster_info(cluster_info, maxpages);
 err:
+	free_cluster_info(cluster_info, maxpages);
 	return ERR_PTR(err);
 }
 
Signature · 1 line
2.34.1
Reviewed-by on v2 1/5
#2 lore raw

Re: [PATCH v2 1/5] mm, swap: Fix memory leak in setup_clusters() error path

mm/swapfile.c

Quoted 17 lines · Youngjun Park

The setup_clusters() function could leak 'cluster_info' memory if an error occurred on a path that did not jump to the 'err_free' label.

This patch simplifies the error handling by removing the goto label and instead calling free_cluster_info() on all error exit paths.

The new logic is safe, as free_cluster_info() already handles NULL pointer inputs.

Fixes 07adc4cf1ecd ("mm, swap: implement dynamic allocation of swap table")
mm/swapfile.c +0 -0
index c35bb8593f50..ed2c8e4edb71 100644

LGTM,

Quoted 23 lines
mm/swapfile.c +2 -3
@@ -3324,7 +3324,7 @@ static struct swap_cluster_info *setup_clusters(struct swap_info_struct *si,
 		si->global_cluster = kmalloc(sizeof(*si->global_cluster),
 				     GFP_KERNEL);
 		if (!si->global_cluster)
-			goto err_free;
+			goto err;
 		for (i = 0; i < SWAP_NR_ORDERS; i++)
 			si->global_cluster->next[i] = SWAP_ENTRY_INVALID;
 		spin_lock_init(&si->global_cluster_lock);
@@ -3377,9 +3377,8 @@ static struct swap_cluster_info *setup_clusters(struct swap_info_struct *si,
 	}
 
 	return cluster_info;
-err_free:
-	free_cluster_info(cluster_info, maxpages);
 err:
+	free_cluster_info(cluster_info, maxpages);
 	return ERR_PTR(err);
 }
 
Signature · 1 line
2.34.1
maintainer Acked-by on v2 1/5
#3 lore raw

Re: [PATCH v2 1/5] mm, swap: Fix memory leak in setup_clusters() error path

In reply to Youngjun Park

Sorry for the late reply. Has been super busy.

Chris

Quoted 41 lines · Youngjun Park

The setup_clusters() function could leak 'cluster_info' memory if an error occurred on a path that did not jump to the 'err_free' label.

This patch simplifies the error handling by removing the goto label and instead calling free_cluster_info() on all error exit paths.

The new logic is safe, as free_cluster_info() already handles NULL pointer inputs.

Fixes 07adc4cf1ecd ("mm, swap: implement dynamic allocation of swap table")
mm/swapfile.c +2 -4
index c35bb8593f50..ed2c8e4edb71 100644
@@ -3324,7 +3324,7 @@ static struct swap_cluster_info *setup_clusters(struct swap_info_struct *si,
                si->global_cluster = kmalloc(sizeof(*si->global_cluster),
                                     GFP_KERNEL);
                if (!si->global_cluster)
-                       goto err_free;
+                       goto err;
                for (i = 0; i < SWAP_NR_ORDERS; i++)
                        si->global_cluster->next[i] = SWAP_ENTRY_INVALID;
                spin_lock_init(&si->global_cluster_lock);
@@ -3377,9 +3377,8 @@ static struct swap_cluster_info *setup_clusters(struct swap_info_struct *si,
        }
 
        return cluster_info;
-err_free:
-       free_cluster_info(cluster_info, maxpages);
 err:
+       free_cluster_info(cluster_info, maxpages);
        return ERR_PTR(err);
 }
 
--

2.34.1

author reviewer
#4 lore raw

PATCH v2 2/5 mm, swap: Use SWP_SOLIDSTATE to determine if swap is rotational

In reply to Youngjun Park

The current non rotational check is unreliable
as the device's rotational status can be changed by a user via sysfs.

Use the more reliable SWP_SOLIDSTATE flag which is set at swapon time, to ensure the nr_rotate_swap count remains consistent. Plus, it is easy to read and simple.

Fixes 81a0298bdfab ("mm, swap: don't use VMA based swap readahead if HDD is used as swap")
mm/swapfile.c +1 -1
index ed2c8e4edb71..b7cde8ef6742 100644
@@ -2913,7 +2913,7 @@ SYSCALL_DEFINE1(swapoff, const char __user *, specialfile)
 	if (p->flags & SWP_CONTINUED)
 		free_swap_count_continuations(p);
 
-	if (!p->bdev || !bdev_nonrot(p->bdev))
+	if (!(p->flags & SWP_SOLIDSTATE))
 		atomic_dec(&nr_rotate_swap);
 
 	mutex_lock(&swapon_mutex);
Signature · 1 line
2.34.1
Reviewed-by on v2 2/5
#5 lore raw

Re: [PATCH v2 2/5] mm, swap: Use SWP_SOLIDSTATE to determine if swap is rotational

Quoted 15 lines · Youngjun Park

The current non rotational check is unreliable
as the device's rotational status can be changed by a user via sysfs.

Use the more reliable SWP_SOLIDSTATE flag which is set at swapon time, to ensure the nr_rotate_swap count remains consistent. Plus, it is easy to read and simple.

Fixes 81a0298bdfab ("mm, swap: don't use VMA based swap readahead if HDD is used as swap")
mm/swapfile.c +0 -0
index ed2c8e4edb71..b7cde8ef6742 100644

mm/swapfile.c SYSCALL_DEFINE1(swapoff, const char __user *, specialfile) · line 2917

Quoted 6 lines
mm/swapfile.c +1 -1
@@ -2913,7 +2913,7 @@ SYSCALL_DEFINE1(swapoff, const char __user *, specialfile)
 	if (p->flags & SWP_CONTINUED)
 		free_swap_count_continuations(p);
 
-	if (!p->bdev || !bdev_nonrot(p->bdev))
+	if (!(p->flags & SWP_SOLIDSTATE))
 		atomic_dec(&nr_rotate_swap);
Quoted 5 lines
mm/swapfile.c +0 -0
 	mutex_lock(&swapon_mutex);
Signature · 1 line
2.34.1
maintainer Acked-by on v2 2/5
#6 lore raw

Re: [PATCH v2 2/5] mm, swap: Use SWP_SOLIDSTATE to determine if swap is rotational

In reply to Youngjun Park

Chris

Quoted 27 lines · Youngjun Park

The current non rotational check is unreliable
as the device's rotational status can be changed by a user via sysfs.

Use the more reliable SWP_SOLIDSTATE flag which is set at swapon time, to ensure the nr_rotate_swap count remains consistent. Plus, it is easy to read and simple.

Fixes 81a0298bdfab ("mm, swap: don't use VMA based swap readahead if HDD is used as swap")
mm/swapfile.c +1 -2
index ed2c8e4edb71..b7cde8ef6742 100644
@@ -2913,7 +2913,7 @@ SYSCALL_DEFINE1(swapoff, const char __user *, specialfile)
        if (p->flags & SWP_CONTINUED)
                free_swap_count_continuations(p);
 
-       if (!p->bdev || !bdev_nonrot(p->bdev))
+       if (!(p->flags & SWP_SOLIDSTATE))
                atomic_dec(&nr_rotate_swap);
 
        mutex_lock(&swapon_mutex);
--

2.34.1

author reviewer
#7 lore raw

PATCH v2 3/5 mm, swap: Remove redundant comment for read_swap_cache_async

In reply to Youngjun Park

The function now manages get/put_swap_device() internally, making the comment explaining this behavior to callers unnecessary.

mm/swap_state.c +0 -4
index b13e9c4baa90..d20d238109f9 100644
@@ -509,10 +509,6 @@ struct folio *__read_swap_cache_async(swp_entry_t entry, gfp_t gfp_mask,
  * and reading the disk if it is not already cached.
  * A failure return means that either the page allocation failed or that
  * the swap entry is no longer in use.
- *
- * get/put_swap_device() aren't needed to call this function, because
- * __read_swap_cache_async() call them and swap_read_folio() holds the
- * swap cache folio lock.
  */
 struct folio *read_swap_cache_async(swp_entry_t entry, gfp_t gfp_mask,
 		struct vm_area_struct *vma, unsigned long addr,
Signature · 1 line
2.34.1
Reviewed-by on v2 3/5
#8 lore raw

Re: [PATCH v2 3/5] mm, swap: Remove redundant comment for read_swap_cache_async

Quoted 10 lines · Youngjun Park

The function now manages get/put_swap_device() internally, making the comment explaining this behavior to callers unnecessary.

mm/swap_state.c +0 -0
index b13e9c4baa90..d20d238109f9 100644

mm/swap_state.c struct folio *__read_swap_cache_async(swp_entry_t entry, gfp_t gfp_mask, · line 514

Quoted 10 lines
mm/swap_state.c +0 -4
@@ -509,10 +509,6 @@ struct folio *__read_swap_cache_async(swp_entry_t entry, gfp_t gfp_mask,
  * and reading the disk if it is not already cached.
  * A failure return means that either the page allocation failed or that
  * the swap entry is no longer in use.
- *
- * get/put_swap_device() aren't needed to call this function, because
- * __read_swap_cache_async() call them and swap_read_folio() holds the
- * swap cache folio lock.
  */
 struct folio *read_swap_cache_async(swp_entry_t entry, gfp_t gfp_mask,
 		struct vm_area_struct *vma, unsigned long addr,
Quoted 3 lines
Signature · 1 line
2.34.1
maintainer
#9 lore raw

Re: [PATCH v2 3/5] mm, swap: Remove redundant comment for read_swap_cache_async

In reply to Youngjun Park
Quoted 3 lines · Youngjun Park

The function now manages get/put_swap_device() internally, making the comment explaining this behavior to callers unnecessary.

This commit message is not precise. The previous comment explaining this behavior is not just unnecessary. The comment is plain wrong. It does not match the code. __read_swap_cache_async() does not internally call get/put_swap_device() any more. So this comment regarding get/put_swap_device should move to __read_swap_cache_async(), the caller of __read_swap_cache_async() needs to manage the get/put_swap_device() or by other means not to race with swap off.

Quoted 8 lines
mm/swap_state.c +0 -0
index b13e9c4baa90..d20d238109f9 100644

mm/swap_state.c struct folio *__read_swap_cache_async(swp_entry_t entry, gfp_t gfp_mask, · line 515

Quoted 7 lines
mm/swap_state.c +0 -4
@@ -509,10 +509,6 @@ struct folio *__read_swap_cache_async(swp_entry_t entry, gfp_t gfp_mask,
  * and reading the disk if it is not already cached.
  * A failure return means that either the page allocation failed or that
  * the swap entry is no longer in use.
- *
- * get/put_swap_device() aren't needed to call this function, because
- * __read_swap_cache_async() call them and swap_read_folio() holds the
- * swap cache folio lock.

I don't think we should simply delete this comment. We can delete from this function but please add a comment to __read_swap_cache_async() regarding the caller's responsibility.

Chris

maintainer
#10 lore raw

Re: [PATCH v2 3/5] mm, swap: Remove redundant comment for read_swap_cache_async

Actually this patch can be dropped. The swap table phase II 1/19 already addresses those comments and fits my description of what needs to be done for __read_swap_cache_async().

Also, the comment adjustment is much better done within a series that actually overhaul the code rather than out of series pure comment clean up.

Hi Youngjun and Andrew,

Do you mind putting this 5 patches series AFTER the swap table phase II? It is the comment adjustment anyway. If there is a conflict with phase II, I think phase II should take priority.

Chris

Quoted 36 lines · Chris Li
Quoted 3 lines · Youngjun Park

The function now manages get/put_swap_device() internally, making the comment explaining this behavior to callers unnecessary.

This commit message is not precise. The previous comment explaining this behavior is not just unnecessary. The comment is plain wrong. It does not match the code. __read_swap_cache_async() does not internally call get/put_swap_device() any more. So this comment regarding get/put_swap_device should move to __read_swap_cache_async(), the caller of __read_swap_cache_async() needs to manage the get/put_swap_device() or by other means not to race with swap off.

Quoted 8 lines
mm/swap_state.c +0 -0
index b13e9c4baa90..d20d238109f9 100644

mm/swap_state.c struct folio *__read_swap_cache_async(swp_entry_t entry, gfp_t gfp_mask, · line 515

Quoted 7 lines
mm/swap_state.c +0 -4
@@ -509,10 +509,6 @@ struct folio *__read_swap_cache_async(swp_entry_t entry, gfp_t gfp_mask,
  * and reading the disk if it is not already cached.
  * A failure return means that either the page allocation failed or that
  * the swap entry is no longer in use.
- *
- * get/put_swap_device() aren't needed to call this function, because
- * __read_swap_cache_async() call them and swap_read_folio() holds the
- * swap cache folio lock.

I don't think we should simply delete this comment. We can delete from this function but please add a comment to __read_swap_cache_async() regarding the caller's responsibility.

Chris

author reviewer
#11 lore raw

Re: [PATCH v2 3/5] mm, swap: Remove redundant comment for read_swap_cache_async

Quoted 8 lines · Chris Li

Actually this patch can be dropped. The swap table phase II 1/19 already addresses those comments and fits my description of what needs to be done for __read_swap_cache_async().

Also, the comment adjustment is much better done within a series that actually overhaul the code rather than out of series pure comment clean up.

Quoted 7 lines

Hi Youngjun and Andrew,

Do you mind putting this 5 patches series AFTER the swap table phase II? It is the comment adjustment anyway. If there is a conflict with phase II, I think phase II should take priority.

Chris

Hi Chris,

Sounds good to me. Happy to have these queued with adjustment after phase II.

Thank you,
Youngjun

maintainer
#12 lore raw

Re: [PATCH v2 3/5] mm, swap: Remove redundant comment for read_swap_cache_async

In reply to Chris Li
Quoted 3 lines · On Tue, 4 Nov 2025 23:58:17 -0800 Chris Li <[email protected]> wrote:

Do you mind putting this 5 patches series AFTER the swap table phase II? It is the comment adjustment anyway. If there is a conflict with phase II, I think phase II should take priority.

OK, I'll drop this series "mm: swap: small fixes and comment cleanups", v2.

maintainer
#13 lore raw

Re: [PATCH v2 3/5] mm, swap: Remove redundant comment for read_swap_cache_async

Quoted 8 lines · Andrew Morton
Quoted 3 lines · On Tue, 4 Nov 2025 23:58:17 -0800 Chris Li <[email protected]> wrote:

Do you mind putting this 5 patches series AFTER the swap table phase II? It is the comment adjustment anyway. If there is a conflict with phase II, I think phase II should take priority.

OK, I'll drop this series "mm: swap: small fixes and comment cleanups", v2.

Hi Andrew,

I discuss with Kairui, his swap table phase II series will need to refresh a V3 anyway.
Please re-apply this cleanup 5 patch series from YongJun, it is already reviewed. Kairui's phase II V3 will use it as base so we don't hold up this cleanup series for too long.

Thanks

Chris

maintainer
#14 lore raw

Re: [PATCH v2 3/5] mm, swap: Remove redundant comment for read_swap_cache_async

Quoted 10 lines · On Wed, 19 Nov 2025 21:00:50 -0800 Chris Li <[email protected]> wrote:
Quoted 8 lines · Andrew Morton
Quoted 3 lines · On Tue, 4 Nov 2025 23:58:17 -0800 Chris Li <[email protected]> wrote:

Do you mind putting this 5 patches series AFTER the swap table phase II? It is the comment adjustment anyway. If there is a conflict with phase II, I think phase II should take priority.

OK, I'll drop this series "mm: swap: small fixes and comment cleanups", v2.

Quoted 7 lines

Hi Andrew,

I discuss with Kairui, his swap table phase II series will need to refresh a V3 anyway.
Please re-apply this cleanup 5 patch series from YongJun, it is already reviewed. Kairui's phase II V3 will use it as base so we don't hold up this cleanup series for too long.

OK...

maintainer
#15 lore raw

Re: [PATCH v2 3/5] mm, swap: Remove redundant comment for read_swap_cache_async

Quoted 9 lines · Andrew Morton
Quoted 7 lines

Hi Andrew,

I discuss with Kairui, his swap table phase II series will need to refresh a V3 anyway.
Please re-apply this cleanup 5 patch series from YongJun, it is already reviewed. Kairui's phase II V3 will use it as base so we don't hold up this cleanup series for too long.

OK...

Thank you so much and sorry for the back and forth, the phase II review takes longer than expected.

Chris

author reviewer
#16 lore raw

PATCH v2 4/5 mm: swap: change swap_alloc_slow() to void

In reply to Youngjun Park

swap_alloc_slow() does not need to return a bool, as all callers handle allocation results via the entry parameter. Update the function signature and remove return statements accordingly.

mm/swapfile.c +3 -4
index b7cde8ef6742..4f8a1843262e 100644
@@ -1339,7 +1339,7 @@ static bool swap_alloc_fast(swp_entry_t *entry,
 }
 
 /* Rotate the device and switch to a new cluster */
-static bool swap_alloc_slow(swp_entry_t *entry,
+static void swap_alloc_slow(swp_entry_t *entry,
 			    int order)
 {
 	unsigned long offset;
@@ -1356,10 +1356,10 @@ static bool swap_alloc_slow(swp_entry_t *entry,
 			put_swap_device(si);
 			if (offset) {
 				*entry = swp_entry(si->type, offset);
-				return true;
+				return;
 			}
 			if (order)
-				return false;
+				return;
 		}
 
 		spin_lock(&swap_avail_lock);
@@ -1378,7 +1378,6 @@ static bool swap_alloc_slow(swp_entry_t *entry,
 			goto start_over;
 	}
 	spin_unlock(&swap_avail_lock);
-	return false;
 }
 
 /*
Signature · 1 line
2.34.1
Reviewed-by on v2 4/5
#17 lore raw

Re: [PATCH v2 4/5] mm: swap: change swap_alloc_slow() to void

mm/swapfile.c static bool swap_alloc_fast(swp_entry_t *entry,

Quoted 12 lines · Youngjun Park

swap_alloc_slow() does not need to return a bool, as all callers handle allocation results via the entry parameter. Update the function signature and remove return statements accordingly.

mm/swapfile.c +0 -0
index b7cde8ef6742..4f8a1843262e 100644
@@ -1339,7 +1339,7 @@ static bool swap_alloc_fast(swp_entry_t *entry,
Quoted 32 lines
mm/swapfile.c +3 -4
 }
 
 /* Rotate the device and switch to a new cluster */
-static bool swap_alloc_slow(swp_entry_t *entry,
+static void swap_alloc_slow(swp_entry_t *entry,
 			    int order)
 {
 	unsigned long offset;
@@ -1356,10 +1356,10 @@ static bool swap_alloc_slow(swp_entry_t *entry,
 			put_swap_device(si);
 			if (offset) {
 				*entry = swp_entry(si->type, offset);
-				return true;
+				return;
 			}
 			if (order)
-				return false;
+				return;
 		}
 
 		spin_lock(&swap_avail_lock);
@@ -1378,7 +1378,6 @@ static bool swap_alloc_slow(swp_entry_t *entry,
 			goto start_over;
 	}
 	spin_unlock(&swap_avail_lock);
-	return false;
 }
 
 /*
Signature · 1 line
2.34.1
maintainer Acked-by on v2 4/5
#18 lore raw

Re: [PATCH v2 4/5] mm: swap: change swap_alloc_slow() to void

In reply to Youngjun Park

Chris

Quoted 45 lines · Youngjun Park

swap_alloc_slow() does not need to return a bool, as all callers handle allocation results via the entry parameter. Update the function signature and remove return statements accordingly.

mm/swapfile.c +3 -5
index b7cde8ef6742..4f8a1843262e 100644
@@ -1339,7 +1339,7 @@ static bool swap_alloc_fast(swp_entry_t *entry,
 }
 
 /* Rotate the device and switch to a new cluster */
-static bool swap_alloc_slow(swp_entry_t *entry,
+static void swap_alloc_slow(swp_entry_t *entry,
                            int order)
 {
        unsigned long offset;
@@ -1356,10 +1356,10 @@ static bool swap_alloc_slow(swp_entry_t *entry,
                        put_swap_device(si);
                        if (offset) {
                                *entry = swp_entry(si->type, offset);
-                               return true;
+                               return;
                        }
                        if (order)
-                               return false;
+                               return;
                }
 
                spin_lock(&swap_avail_lock);
@@ -1378,7 +1378,6 @@ static bool swap_alloc_slow(swp_entry_t *entry,
                        goto start_over;
        }
        spin_unlock(&swap_avail_lock);
-       return false;
 }
 
 /*
--

2.34.1

author reviewer
#19 lore raw

PATCH v2 5/5 mm: swap: remove scan_swap_map_slots() references from comments

In reply to Youngjun Park

The scan_swap_map_slots() helper has been removed, but several comments still referred to it in swap allocation and reclaim paths. This patch cleans up those outdated references and reflows the affected comment blocks to match kernel coding style.

mm/swapfile.c +9 -11
index 4f8a1843262e..543f303f101d 100644
@@ -236,11 +236,10 @@ static int __try_to_reclaim_swap(struct swap_info_struct *si,
 	ret = -nr_pages;
 
 	/*
-	 * When this function is called from scan_swap_map_slots() and it's
-	 * called by vmscan.c at reclaiming folios. So we hold a folio lock
-	 * here. We have to use trylock for avoiding deadlock. This is a special
-	 * case and you should use folio_free_swap() with explicit folio_lock()
-	 * in usual operations.
+	 * We hold a folio lock here. We have to use trylock for
+	 * avoiding deadlock. This is a special case and you should
+	 * use folio_free_swap() with explicit folio_lock() in usual
+	 * operations.
 	 */
 	if (!folio_trylock(folio))
 		goto out;
@@ -1365,14 +1364,13 @@ static void swap_alloc_slow(swp_entry_t *entry,
 		spin_lock(&swap_avail_lock);
 		/*
 		 * if we got here, it's likely that si was almost full before,
-		 * and since scan_swap_map_slots() can drop the si->lock,
 		 * multiple callers probably all tried to get a page from the
 		 * same si and it filled up before we could get one; or, the si
-		 * filled up between us dropping swap_avail_lock and taking
-		 * si->lock. Since we dropped the swap_avail_lock, the
-		 * swap_avail_head list may have been modified; so if next is
-		 * still in the swap_avail_head list then try it, otherwise
-		 * start over if we have not gotten any slots.
+		 * filled up between us dropping swap_avail_lock.
+		 * Since we dropped the swap_avail_lock, the swap_avail_list
+		 * may have been modified; so if next is still in the
+		 * swap_avail_head list then try it, otherwise start over if we
+		 * have not gotten any slots.
 		 */
 		if (plist_node_empty(&si->avail_list))
 			goto start_over;
Signature · 1 line
2.34.1
Reviewed-by on v2 5/5
#20 lore raw

Re: [PATCH v2 5/5] mm: swap: remove scan_swap_map_slots() references from comments

mm/swapfile.c

Quoted 11 lines · Youngjun Park

The scan_swap_map_slots() helper has been removed, but several comments still referred to it in swap allocation and reclaim paths. This patch cleans up those outdated references and reflows the affected comment blocks to match kernel coding style.

mm/swapfile.c +0 -0
index 4f8a1843262e..543f303f101d 100644
Quoted 39 lines
mm/swapfile.c +9 -11
@@ -236,11 +236,10 @@ static int __try_to_reclaim_swap(struct swap_info_struct *si,
 	ret = -nr_pages;
 
 	/*
-	 * When this function is called from scan_swap_map_slots() and it's
-	 * called by vmscan.c at reclaiming folios. So we hold a folio lock
-	 * here. We have to use trylock for avoiding deadlock. This is a special
-	 * case and you should use folio_free_swap() with explicit folio_lock()
-	 * in usual operations.
+	 * We hold a folio lock here. We have to use trylock for
+	 * avoiding deadlock. This is a special case and you should
+	 * use folio_free_swap() with explicit folio_lock() in usual
+	 * operations.
 	 */
 	if (!folio_trylock(folio))
 		goto out;
@@ -1365,14 +1364,13 @@ static void swap_alloc_slow(swp_entry_t *entry,
 		spin_lock(&swap_avail_lock);
 		/*
 		 * if we got here, it's likely that si was almost full before,
-		 * and since scan_swap_map_slots() can drop the si->lock,
 		 * multiple callers probably all tried to get a page from the
 		 * same si and it filled up before we could get one; or, the si
-		 * filled up between us dropping swap_avail_lock and taking
-		 * si->lock. Since we dropped the swap_avail_lock, the
-		 * swap_avail_head list may have been modified; so if next is
-		 * still in the swap_avail_head list then try it, otherwise
-		 * start over if we have not gotten any slots.
+		 * filled up between us dropping swap_avail_lock.
+		 * Since we dropped the swap_avail_lock, the swap_avail_list
+		 * may have been modified; so if next is still in the
+		 * swap_avail_head list then try it, otherwise start over if we
+		 * have not gotten any slots.
 		 */
 		if (plist_node_empty(&si->avail_list))
 			goto start_over;
Signature · 1 line
2.34.1
maintainer Acked-by on v2 5/5
#21 lore raw

Re: [PATCH v2 5/5] mm: swap: remove scan_swap_map_slots() references from comments

In reply to Youngjun Park

Chris

Quoted 51 lines · Youngjun Park

The scan_swap_map_slots() helper has been removed, but several comments still referred to it in swap allocation and reclaim paths. This patch cleans up those outdated references and reflows the affected comment blocks to match kernel coding style.

mm/swapfile.c +9 -12
index 4f8a1843262e..543f303f101d 100644
@@ -236,11 +236,10 @@ static int __try_to_reclaim_swap(struct swap_info_struct *si,
        ret = -nr_pages;
 
        /*
-        * When this function is called from scan_swap_map_slots() and it's
-        * called by vmscan.c at reclaiming folios. So we hold a folio lock
-        * here. We have to use trylock for avoiding deadlock. This is a special
-        * case and you should use folio_free_swap() with explicit folio_lock()
-        * in usual operations.
+        * We hold a folio lock here. We have to use trylock for
+        * avoiding deadlock. This is a special case and you should
+        * use folio_free_swap() with explicit folio_lock() in usual
+        * operations.
         */
        if (!folio_trylock(folio))
                goto out;
@@ -1365,14 +1364,13 @@ static void swap_alloc_slow(swp_entry_t *entry,
                spin_lock(&swap_avail_lock);
                /*
                 * if we got here, it's likely that si was almost full before,
-                * and since scan_swap_map_slots() can drop the si->lock,
                 * multiple callers probably all tried to get a page from the
                 * same si and it filled up before we could get one; or, the si
-                * filled up between us dropping swap_avail_lock and taking
-                * si->lock. Since we dropped the swap_avail_lock, the
-                * swap_avail_head list may have been modified; so if next is
-                * still in the swap_avail_head list then try it, otherwise
-                * start over if we have not gotten any slots.
+                * filled up between us dropping swap_avail_lock.
+                * Since we dropped the swap_avail_lock, the swap_avail_list
+                * may have been modified; so if next is still in the
+                * swap_avail_head list then try it, otherwise start over if we
+                * have not gotten any slots.
                 */
                if (plist_node_empty(&si->avail_list))
                        goto start_over;
--

2.34.1