[PATCH 0/4] mm/zswap: misc cleanup of code and documentations

[PATCH 0/4] mm/zswap: misc cleanup of code and documentations

[SeongJae Park]

Cleanup an unnecessary local variable write in incompressible pages
handling, typos (s/zwap/zswap/) and outdated comments/documentations
about the zswap's red-black tree, which is replaced by xarray.

SeongJae Park (4):
  mm/zswap: remove unnecessary dlen writes for incompressible pages
  mm/zswap: fix typos: s/zwap/zswap/
  mm/zswap: s/red-black tree/xarray/
  Docs/admin-guide/mm/zswap: s/red-black tree/xarray/

 Documentation/admin-guide/mm/zswap.rst | 10 +++++-----
 mm/memcontrol.c                        |  2 +-
 mm/zswap.c                             |  7 +++----
 3 files changed, 9 insertions(+), 10 deletions(-)


base-commit: 2e943d9cc1fdd399a2f95d1c1d6b865d04bfa239
-- 
2.39.5

[PATCH 1/4] mm/zswap: remove unnecessary dlen writes for incompressible pages

[SeongJae Park]

Incompressible pages handling logic in zswap_compress() is setting
'dlen' as PAGE_SIZE twice.  Once before deciding whether to save the
content as is, and once again after it is decided to save it as is.
But the value of 'dlen' is used only if it is decided to save the
content as is, so the first write is unnecessary.  It is not causing
real user issues, but making code confusing to read.  Remove the
unnecessary write operation.

Signed-off-by: SeongJae Park <sj@kernel.org>
---
 mm/zswap.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/mm/zswap.c b/mm/zswap.c
index c1af782e54ec..80619c8589a7 100644
--- a/mm/zswap.c
+++ b/mm/zswap.c
@@ -894,7 +894,6 @@ static bool zswap_compress(struct page *page, struct zswap_entry *entry,
 	 * to the active LRU list in the case.
 	 */
 	if (comp_ret || !dlen || dlen >= PAGE_SIZE) {
-		dlen = PAGE_SIZE;
 		if (!mem_cgroup_zswap_writeback_enabled(
 					folio_memcg(page_folio(page)))) {
 			comp_ret = comp_ret ? comp_ret : -EINVAL;
-- 
2.39.5

[PATCH 2/4] mm/zswap: fix typos: s/zwap/zswap/

[SeongJae Park]

As the subject says.

Signed-off-by: SeongJae Park <sj@kernel.org>
---
 mm/memcontrol.c | 2 +-
 mm/zswap.c      | 4 ++--
 2 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index 69c970554e85..74b1bc2252b6 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -5421,7 +5421,7 @@ bool obj_cgroup_may_zswap(struct obj_cgroup *objcg)
  * @size: size of compressed object
  *
  * This forces the charge after obj_cgroup_may_zswap() allowed
- * compression and storage in zwap for this cgroup to go ahead.
+ * compression and storage in zswap for this cgroup to go ahead.
  */
 void obj_cgroup_charge_zswap(struct obj_cgroup *objcg, size_t size)
 {
diff --git a/mm/zswap.c b/mm/zswap.c
index 80619c8589a7..f6b1c8832a4f 100644
--- a/mm/zswap.c
+++ b/mm/zswap.c
@@ -879,7 +879,7 @@ static bool zswap_compress(struct page *page, struct zswap_entry *entry,
 	 * acomp instance, then get those requests done simultaneously. but in this
 	 * case, zswap actually does store and load page by page, there is no
 	 * existing method to send the second page before the first page is done
-	 * in one thread doing zwap.
+	 * in one thread doing zswap.
 	 * but in different threads running on different cpu, we have different
 	 * acomp instance, so multiple threads can do (de)compression in parallel.
 	 */
@@ -1128,7 +1128,7 @@ static enum lru_status shrink_memcg_cb(struct list_head *item, struct list_lru_o
 	 *
 	 * 1. We extract the swp_entry_t to the stack, allowing
 	 *    zswap_writeback_entry() to pin the swap entry and
-	 *    then validate the zwap entry against that swap entry's
+	 *    then validate the zswap entry against that swap entry's
 	 *    tree using pointer value comparison. Only when that
 	 *    is successful can the entry be dereferenced.
 	 *
-- 
2.39.5

[PATCH 3/4] mm/zswap: s/red-black tree/xarray/

[SeongJae Park]

Changes made by commit 796c2c23e14e ("zswap: replace RB tree with
xarray") is not reflected on a comment.  Update the comment.

Signed-off-by: SeongJae Park <sj@kernel.org>
---
 mm/zswap.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/mm/zswap.c b/mm/zswap.c
index f6b1c8832a4f..5d0f8b13a958 100644
--- a/mm/zswap.c
+++ b/mm/zswap.c
@@ -175,7 +175,7 @@ static struct shrinker *zswap_shrinker;
  * This structure contains the metadata for tracking a single compressed
  * page within zswap.
  *
- * swpentry - associated swap entry, the offset indexes into the red-black tree
+ * swpentry - associated swap entry, the offset indexes into the xarray
  * length - the length in bytes of the compressed page data.  Needed during
  *          decompression.
  * referenced - true if the entry recently entered the zswap pool. Unset by the
-- 
2.39.5

[PATCH 4/4] Docs/admin-guide/mm/zswap: s/red-black tree/xarray/

[SeongJae Park]

The change from commit 796c2c23e14e ("zswap: replace RB tree with
xarray") is not reflected on the document.  Update the document.

Signed-off-by: SeongJae Park <sj@kernel.org>
---
 Documentation/admin-guide/mm/zswap.rst | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/Documentation/admin-guide/mm/zswap.rst b/Documentation/admin-guide/mm/zswap.rst
index 283d77217c6f..2464425c783d 100644
--- a/Documentation/admin-guide/mm/zswap.rst
+++ b/Documentation/admin-guide/mm/zswap.rst
@@ -59,11 +59,11 @@ returned by the allocation routine and that handle must be mapped before being
 accessed.  The compressed memory pool grows on demand and shrinks as compressed
 pages are freed.  The pool is not preallocated.
 
-When a swap page is passed from swapout to zswap, zswap maintains a mapping
-of the swap entry, a combination of the swap type and swap offset, to the
-zsmalloc handle that references that compressed swap page.  This mapping is
-achieved with a red-black tree per swap type.  The swap offset is the search
-key for the tree nodes.
+When a swap page is passed from swapout to zswap, zswap maintains a mapping of
+the swap entry, a combination of the swap type and swap offset, to the zsmalloc
+handle that references that compressed swap page.  This mapping is achieved
+with an xarray per swap type.  The swap offset is the search key for the xarray
+nodes.
 
 During a page fault on a PTE that is a swap entry, the swapin code calls the
 zswap load function to decompress the page into the page allocated by the page
-- 
2.39.5

Re: [PATCH 1/4] mm/zswap: remove unnecessary dlen writes for incompressible pages

[Yosry Ahmed]

On Fri, Oct 03, 2025 at 01:38:48PM -0700, SeongJae Park wrote:
> Incompressible pages handling logic in zswap_compress() is setting
> 'dlen' as PAGE_SIZE twice.  Once before deciding whether to save the
> content as is, and once again after it is decided to save it as is.
> But the value of 'dlen' is used only if it is decided to save the
> content as is, so the first write is unnecessary.  It is not causing
> real user issues, but making code confusing to read.  Remove the
> unnecessary write operation.
> 
> Signed-off-by: SeongJae Park <sj@kernel.org>

Acked-by: Yosry Ahmed <yosry.ahmed@linux.dev>

> ---
>  mm/zswap.c | 1 -
>  1 file changed, 1 deletion(-)
> 
> diff --git a/mm/zswap.c b/mm/zswap.c
> index c1af782e54ec..80619c8589a7 100644
> --- a/mm/zswap.c
> +++ b/mm/zswap.c
> @@ -894,7 +894,6 @@ static bool zswap_compress(struct page *page, struct zswap_entry *entry,
>  	 * to the active LRU list in the case.
>  	 */
>  	if (comp_ret || !dlen || dlen >= PAGE_SIZE) {
> -		dlen = PAGE_SIZE;
>  		if (!mem_cgroup_zswap_writeback_enabled(
>  					folio_memcg(page_folio(page)))) {
>  			comp_ret = comp_ret ? comp_ret : -EINVAL;
> -- 
> 2.39.5

Re: [PATCH 2/4] mm/zswap: fix typos: s/zwap/zswap/

[Yosry Ahmed]

On Fri, Oct 03, 2025 at 01:38:49PM -0700, SeongJae Park wrote:
> As the subject says.
> 
> Signed-off-by: SeongJae Park <sj@kernel.org>

Acked-by: Yosry Ahmed <yosry.ahmed@linux.dev>

> ---
>  mm/memcontrol.c | 2 +-
>  mm/zswap.c      | 4 ++--
>  2 files changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/mm/memcontrol.c b/mm/memcontrol.c
> index 69c970554e85..74b1bc2252b6 100644
> --- a/mm/memcontrol.c
> +++ b/mm/memcontrol.c
> @@ -5421,7 +5421,7 @@ bool obj_cgroup_may_zswap(struct obj_cgroup *objcg)
>   * @size: size of compressed object
>   *
>   * This forces the charge after obj_cgroup_may_zswap() allowed
> - * compression and storage in zwap for this cgroup to go ahead.
> + * compression and storage in zswap for this cgroup to go ahead.
>   */
>  void obj_cgroup_charge_zswap(struct obj_cgroup *objcg, size_t size)
>  {
> diff --git a/mm/zswap.c b/mm/zswap.c
> index 80619c8589a7..f6b1c8832a4f 100644
> --- a/mm/zswap.c
> +++ b/mm/zswap.c
> @@ -879,7 +879,7 @@ static bool zswap_compress(struct page *page, struct zswap_entry *entry,
>  	 * acomp instance, then get those requests done simultaneously. but in this
>  	 * case, zswap actually does store and load page by page, there is no
>  	 * existing method to send the second page before the first page is done
> -	 * in one thread doing zwap.
> +	 * in one thread doing zswap.
>  	 * but in different threads running on different cpu, we have different
>  	 * acomp instance, so multiple threads can do (de)compression in parallel.
>  	 */
> @@ -1128,7 +1128,7 @@ static enum lru_status shrink_memcg_cb(struct list_head *item, struct list_lru_o
>  	 *
>  	 * 1. We extract the swp_entry_t to the stack, allowing
>  	 *    zswap_writeback_entry() to pin the swap entry and
> -	 *    then validate the zwap entry against that swap entry's
> +	 *    then validate the zswap entry against that swap entry's
>  	 *    tree using pointer value comparison. Only when that
>  	 *    is successful can the entry be dereferenced.
>  	 *
> -- 
> 2.39.5

Re: [PATCH 3/4] mm/zswap: s/red-black tree/xarray/

[Yosry Ahmed]

On Fri, Oct 03, 2025 at 01:38:50PM -0700, SeongJae Park wrote:
> Changes made by commit 796c2c23e14e ("zswap: replace RB tree with
> xarray") is not reflected on a comment.  Update the comment.
> 
> Signed-off-by: SeongJae Park <sj@kernel.org>

Acked-by: Yosry Ahmed <yosry.ahmed@linux.dev>

> ---
>  mm/zswap.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/mm/zswap.c b/mm/zswap.c
> index f6b1c8832a4f..5d0f8b13a958 100644
> --- a/mm/zswap.c
> +++ b/mm/zswap.c
> @@ -175,7 +175,7 @@ static struct shrinker *zswap_shrinker;
>   * This structure contains the metadata for tracking a single compressed
>   * page within zswap.
>   *
> - * swpentry - associated swap entry, the offset indexes into the red-black tree
> + * swpentry - associated swap entry, the offset indexes into the xarray
>   * length - the length in bytes of the compressed page data.  Needed during
>   *          decompression.
>   * referenced - true if the entry recently entered the zswap pool. Unset by the
> -- 
> 2.39.5

Re: [PATCH 4/4] Docs/admin-guide/mm/zswap: s/red-black tree/xarray/

[Yosry Ahmed]

On Fri, Oct 03, 2025 at 01:38:51PM -0700, SeongJae Park wrote:
> The change from commit 796c2c23e14e ("zswap: replace RB tree with
> xarray") is not reflected on the document.  Update the document.
> 
> Signed-off-by: SeongJae Park <sj@kernel.org>

Acked-by: Yosry Ahmed <yosry.ahmed@linux.dev>

> ---
>  Documentation/admin-guide/mm/zswap.rst | 10 +++++-----
>  1 file changed, 5 insertions(+), 5 deletions(-)
> 
> diff --git a/Documentation/admin-guide/mm/zswap.rst b/Documentation/admin-guide/mm/zswap.rst
> index 283d77217c6f..2464425c783d 100644
> --- a/Documentation/admin-guide/mm/zswap.rst
> +++ b/Documentation/admin-guide/mm/zswap.rst
> @@ -59,11 +59,11 @@ returned by the allocation routine and that handle must be mapped before being
>  accessed.  The compressed memory pool grows on demand and shrinks as compressed
>  pages are freed.  The pool is not preallocated.
>  
> -When a swap page is passed from swapout to zswap, zswap maintains a mapping
> -of the swap entry, a combination of the swap type and swap offset, to the
> -zsmalloc handle that references that compressed swap page.  This mapping is
> -achieved with a red-black tree per swap type.  The swap offset is the search
> -key for the tree nodes.
> +When a swap page is passed from swapout to zswap, zswap maintains a mapping of
> +the swap entry, a combination of the swap type and swap offset, to the zsmalloc
> +handle that references that compressed swap page.  This mapping is achieved
> +with an xarray per swap type.  The swap offset is the search key for the xarray
> +nodes.
>  
>  During a page fault on a PTE that is a swap entry, the swapin code calls the
>  zswap load function to decompress the page into the page allocated by the page
> -- 
> 2.39.5

Re: [PATCH 1/4] mm/zswap: remove unnecessary dlen writes for incompressible pages

[Nhat Pham]

On Fri, Oct 3, 2025 at 1:38 PM SeongJae Park <sj@kernel.org> wrote:
>
> Incompressible pages handling logic in zswap_compress() is setting
> 'dlen' as PAGE_SIZE twice.  Once before deciding whether to save the
> content as is, and once again after it is decided to save it as is.
> But the value of 'dlen' is used only if it is decided to save the
> content as is, so the first write is unnecessary.  It is not causing
> real user issues, but making code confusing to read.  Remove the
> unnecessary write operation.
>
> Signed-off-by: SeongJae Park <sj@kernel.org>

Acked-by: Nhat Pham <nphamcs@gmail.com>

Re: [PATCH 2/4] mm/zswap: fix typos: s/zwap/zswap/

[Nhat Pham]

On Fri, Oct 3, 2025 at 1:38 PM SeongJae Park <sj@kernel.org> wrote:
>
> As the subject says.
>
> Signed-off-by: SeongJae Park <sj@kernel.org>

Acked-by: Nhat Pham <nphamcs@gmail.com>

Re: [PATCH 3/4] mm/zswap: s/red-black tree/xarray/

[Nhat Pham]

On Fri, Oct 3, 2025 at 1:38 PM SeongJae Park <sj@kernel.org> wrote:
>
> Changes made by commit 796c2c23e14e ("zswap: replace RB tree with
> xarray") is not reflected on a comment.  Update the comment.
>
> Signed-off-by: SeongJae Park <sj@kernel.org>

Acked-by: Nhat Pham <nphamcs@gmail.com>

Re: [PATCH 4/4] Docs/admin-guide/mm/zswap: s/red-black tree/xarray/

[Nhat Pham]

On Fri, Oct 3, 2025 at 1:38 PM SeongJae Park <sj@kernel.org> wrote:
>
> The change from commit 796c2c23e14e ("zswap: replace RB tree with
> xarray") is not reflected on the document.  Update the document.
>
> Signed-off-by: SeongJae Park <sj@kernel.org>
> ---
>  Documentation/admin-guide/mm/zswap.rst | 10 +++++-----
>  1 file changed, 5 insertions(+), 5 deletions(-)
>
> diff --git a/Documentation/admin-guide/mm/zswap.rst b/Documentation/admin-guide/mm/zswap.rst
> index 283d77217c6f..2464425c783d 100644
> --- a/Documentation/admin-guide/mm/zswap.rst
> +++ b/Documentation/admin-guide/mm/zswap.rst
> @@ -59,11 +59,11 @@ returned by the allocation routine and that handle must be mapped before being
>  accessed.  The compressed memory pool grows on demand and shrinks as compressed
>  pages are freed.  The pool is not preallocated.
>
> -When a swap page is passed from swapout to zswap, zswap maintains a mapping
> -of the swap entry, a combination of the swap type and swap offset, to the
> -zsmalloc handle that references that compressed swap page.  This mapping is
> -achieved with a red-black tree per swap type.  The swap offset is the search
> -key for the tree nodes.
> +When a swap page is passed from swapout to zswap, zswap maintains a mapping of
> +the swap entry, a combination of the swap type and swap offset, to the zsmalloc
> +handle that references that compressed swap page.  This mapping is achieved
> +with an xarray per swap type.  The swap offset is the search key for the xarray
> +nodes.
>

Acked-by: Nhat Pham <nphamcs@gmail.com>

Re: [PATCH 1/4] mm/zswap: remove unnecessary dlen writes for incompressible pages

[Chengming Zhou]

On 2025/10/4 04:38, SeongJae Park wrote:
> Incompressible pages handling logic in zswap_compress() is setting
> 'dlen' as PAGE_SIZE twice.  Once before deciding whether to save the
> content as is, and once again after it is decided to save it as is.
> But the value of 'dlen' is used only if it is decided to save the
> content as is, so the first write is unnecessary.  It is not causing
> real user issues, but making code confusing to read.  Remove the
> unnecessary write operation.
> 
> Signed-off-by: SeongJae Park <sj@kernel.org>

Reviewed-by: Chengming Zhou <chengming.zhou@linux.dev>

> ---
>   mm/zswap.c | 1 -
>   1 file changed, 1 deletion(-)
> 
> diff --git a/mm/zswap.c b/mm/zswap.c
> index c1af782e54ec..80619c8589a7 100644
> --- a/mm/zswap.c
> +++ b/mm/zswap.c
> @@ -894,7 +894,6 @@ static bool zswap_compress(struct page *page, struct zswap_entry *entry,
>   	 * to the active LRU list in the case.
>   	 */
>   	if (comp_ret || !dlen || dlen >= PAGE_SIZE) {
> -		dlen = PAGE_SIZE;
>   		if (!mem_cgroup_zswap_writeback_enabled(
>   					folio_memcg(page_folio(page)))) {
>   			comp_ret = comp_ret ? comp_ret : -EINVAL;

Re: [PATCH 2/4] mm/zswap: fix typos: s/zwap/zswap/

[Chengming Zhou]

On 2025/10/4 04:38, SeongJae Park wrote:
> As the subject says.
> 
> Signed-off-by: SeongJae Park <sj@kernel.org>

Reviewed-by: Chengming Zhou <chengming.zhou@linux.dev>

> ---
>   mm/memcontrol.c | 2 +-
>   mm/zswap.c      | 4 ++--
>   2 files changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/mm/memcontrol.c b/mm/memcontrol.c
> index 69c970554e85..74b1bc2252b6 100644
> --- a/mm/memcontrol.c
> +++ b/mm/memcontrol.c
> @@ -5421,7 +5421,7 @@ bool obj_cgroup_may_zswap(struct obj_cgroup *objcg)
>    * @size: size of compressed object
>    *
>    * This forces the charge after obj_cgroup_may_zswap() allowed
> - * compression and storage in zwap for this cgroup to go ahead.
> + * compression and storage in zswap for this cgroup to go ahead.
>    */
>   void obj_cgroup_charge_zswap(struct obj_cgroup *objcg, size_t size)
>   {
> diff --git a/mm/zswap.c b/mm/zswap.c
> index 80619c8589a7..f6b1c8832a4f 100644
> --- a/mm/zswap.c
> +++ b/mm/zswap.c
> @@ -879,7 +879,7 @@ static bool zswap_compress(struct page *page, struct zswap_entry *entry,
>   	 * acomp instance, then get those requests done simultaneously. but in this
>   	 * case, zswap actually does store and load page by page, there is no
>   	 * existing method to send the second page before the first page is done
> -	 * in one thread doing zwap.
> +	 * in one thread doing zswap.
>   	 * but in different threads running on different cpu, we have different
>   	 * acomp instance, so multiple threads can do (de)compression in parallel.
>   	 */
> @@ -1128,7 +1128,7 @@ static enum lru_status shrink_memcg_cb(struct list_head *item, struct list_lru_o
>   	 *
>   	 * 1. We extract the swp_entry_t to the stack, allowing
>   	 *    zswap_writeback_entry() to pin the swap entry and
> -	 *    then validate the zwap entry against that swap entry's
> +	 *    then validate the zswap entry against that swap entry's
>   	 *    tree using pointer value comparison. Only when that
>   	 *    is successful can the entry be dereferenced.
>   	 *

Re: [PATCH 3/4] mm/zswap: s/red-black tree/xarray/

[Chengming Zhou]

On 2025/10/4 04:38, SeongJae Park wrote:
> Changes made by commit 796c2c23e14e ("zswap: replace RB tree with
> xarray") is not reflected on a comment.  Update the comment.
> 
> Signed-off-by: SeongJae Park <sj@kernel.org>

Reviewed-by: Chengming Zhou <chengming.zhou@linux.dev>

> ---
>   mm/zswap.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/mm/zswap.c b/mm/zswap.c
> index f6b1c8832a4f..5d0f8b13a958 100644
> --- a/mm/zswap.c
> +++ b/mm/zswap.c
> @@ -175,7 +175,7 @@ static struct shrinker *zswap_shrinker;
>    * This structure contains the metadata for tracking a single compressed
>    * page within zswap.
>    *
> - * swpentry - associated swap entry, the offset indexes into the red-black tree
> + * swpentry - associated swap entry, the offset indexes into the xarray
>    * length - the length in bytes of the compressed page data.  Needed during
>    *          decompression.
>    * referenced - true if the entry recently entered the zswap pool. Unset by the

Re: [PATCH 4/4] Docs/admin-guide/mm/zswap: s/red-black tree/xarray/

[Chengming Zhou]

On 2025/10/4 04:38, SeongJae Park wrote:
> The change from commit 796c2c23e14e ("zswap: replace RB tree with
> xarray") is not reflected on the document.  Update the document.
> 
> Signed-off-by: SeongJae Park <sj@kernel.org>

Reviewed-by: Chengming Zhou <chengming.zhou@linux.dev>

> ---
>   Documentation/admin-guide/mm/zswap.rst | 10 +++++-----
>   1 file changed, 5 insertions(+), 5 deletions(-)
> 
> diff --git a/Documentation/admin-guide/mm/zswap.rst b/Documentation/admin-guide/mm/zswap.rst
> index 283d77217c6f..2464425c783d 100644
> --- a/Documentation/admin-guide/mm/zswap.rst
> +++ b/Documentation/admin-guide/mm/zswap.rst
> @@ -59,11 +59,11 @@ returned by the allocation routine and that handle must be mapped before being
>   accessed.  The compressed memory pool grows on demand and shrinks as compressed
>   pages are freed.  The pool is not preallocated.
>   
> -When a swap page is passed from swapout to zswap, zswap maintains a mapping
> -of the swap entry, a combination of the swap type and swap offset, to the
> -zsmalloc handle that references that compressed swap page.  This mapping is
> -achieved with a red-black tree per swap type.  The swap offset is the search
> -key for the tree nodes.
> +When a swap page is passed from swapout to zswap, zswap maintains a mapping of
> +the swap entry, a combination of the swap type and swap offset, to the zsmalloc
> +handle that references that compressed swap page.  This mapping is achieved
> +with an xarray per swap type.  The swap offset is the search key for the xarray
> +nodes.
>   
>   During a page fault on a PTE that is a swap entry, the swapin code calls the
>   zswap load function to decompress the page into the page allocated by the page