[PATCH] fs/drop_caches: skip filesystems without fs writeback

[PATCH] fs/drop_caches: skip filesystems without fs writeback

[Huang Shijie]

Filesystems backed by the noop backing device (procfs, sysfs, kernfs,
tmpfs, ramfs, etc.) have no fs page-cache writeback.  Skip them in
drop_pagecache_sb() instead of walking their inode lists.

This also skips tmpfs/ramfs. tmpfs has clean folios which are
zero-filled pages from reading holes, which are trivially recreatable.

Suggested-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Huang Shijie <huangsj@hygon.cn>
---
The RFC mails:
   https://lore.kernel.org/all/arIVBpMZk5-O2AZc@infradead.org/
---
 fs/drop_caches.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/fs/drop_caches.c b/fs/drop_caches.c
index 49f56a598ecb..28c5fb9c7357 100644
--- a/fs/drop_caches.c
+++ b/fs/drop_caches.c
@@ -7,6 +7,7 @@
 #include <linux/kernel.h>
 #include <linux/mm.h>
 #include <linux/fs.h>
+#include <linux/backing-dev.h>
 #include <linux/writeback.h>
 #include <linux/sysctl.h>
 #include <linux/gfp.h>
@@ -20,6 +21,9 @@ static void drop_pagecache_sb(struct super_block *sb, void *unused)
 {
 	struct inode *inode, *toput_inode = NULL;
 
+	if (sb->s_bdi == &noop_backing_dev_info)
+		return;
+
 	spin_lock(&sb->s_inode_list_lock);
 	list_for_each_entry(inode, &sb->s_inodes, i_sb_list) {
 		spin_lock(&inode->i_lock);
-- 
2.53.0

Re: [PATCH] fs/drop_caches: skip filesystems without fs writeback

[Jan Kara]

On Wed 23-09-26 15:01:16, Huang Shijie wrote:
> Filesystems backed by the noop backing device (procfs, sysfs, kernfs,
> tmpfs, ramfs, etc.) have no fs page-cache writeback.  Skip them in
> drop_pagecache_sb() instead of walking their inode lists.
> 
> This also skips tmpfs/ramfs. tmpfs has clean folios which are
> zero-filled pages from reading holes, which are trivially recreatable.
> 
> Suggested-by: Christoph Hellwig <hch@lst.de>
> Signed-off-by: Huang Shijie <huangsj@hygon.cn>

This looks indeed nice & simple. I'm slightly wondering if there isn't some
exotic read-only fs somewhere which doesn't set sb->s_bdi but so far I'm
not aware of anything (even MTD filesystems do properly setup bdi) so feel
free to add:

Reviewed-by: Jan Kara <jack@suse.cz>

								Honza

> ---
> The RFC mails:
>    https://lore.kernel.org/all/arIVBpMZk5-O2AZc@infradead.org/
> ---
>  fs/drop_caches.c | 4 ++++
>  1 file changed, 4 insertions(+)
> 
> diff --git a/fs/drop_caches.c b/fs/drop_caches.c
> index 49f56a598ecb..28c5fb9c7357 100644
> --- a/fs/drop_caches.c
> +++ b/fs/drop_caches.c
> @@ -7,6 +7,7 @@
>  #include <linux/kernel.h>
>  #include <linux/mm.h>
>  #include <linux/fs.h>
> +#include <linux/backing-dev.h>
>  #include <linux/writeback.h>
>  #include <linux/sysctl.h>
>  #include <linux/gfp.h>
> @@ -20,6 +21,9 @@ static void drop_pagecache_sb(struct super_block *sb, void *unused)
>  {
>  	struct inode *inode, *toput_inode = NULL;
>  
> +	if (sb->s_bdi == &noop_backing_dev_info)
> +		return;
> +
>  	spin_lock(&sb->s_inode_list_lock);
>  	list_for_each_entry(inode, &sb->s_inodes, i_sb_list) {
>  		spin_lock(&inode->i_lock);
> -- 
> 2.53.0
> 
> 
-- 
Jan Kara <jack@suse.com>
SUSE Labs, CR

Re: [PATCH] fs/drop_caches: skip filesystems without fs writeback

[Christoph Hellwig]

Looks good:

Reviewed-by: Christoph Hellwig <hch@lst.de>