1. Patchew
  2. linux
  3. [PATCH v2 0/6] Extend freeze support to suspend and hibernate
  4. View patch

[PATCH v2 1/6] super: remove pointless s_root checks

Christian Brauner posted 6 patches 8 months, 3 weeks ago
[PATCH v2 1/6] super: remove pointless s_root checks
Posted by Christian Brauner 8 months, 3 weeks ago 
The locking guarantees that the superblock is alive and sb->s_root is
still set. Remove the pointless check.

Signed-off-by: Christian Brauner <brauner@kernel.org>
---
 fs/super.c | 19 ++++++-------------
 1 file changed, 6 insertions(+), 13 deletions(-)

diff --git a/fs/super.c b/fs/super.c
index 97a17f9d9023..dc14f4bf73a6 100644
--- a/fs/super.c
+++ b/fs/super.c
@@ -930,8 +930,7 @@ void iterate_supers(void (*f)(struct super_block *, void *), void *arg)
 
 		locked = super_lock_shared(sb);
 		if (locked) {
-			if (sb->s_root)
-				f(sb, arg);
+			f(sb, arg);
 			super_unlock_shared(sb);
 		}
 
@@ -967,11 +966,8 @@ void iterate_supers_type(struct file_system_type *type,
 		spin_unlock(&sb_lock);
 
 		locked = super_lock_shared(sb);
-		if (locked) {
-			if (sb->s_root)
-				f(sb, arg);
-			super_unlock_shared(sb);
-		}
+		if (locked)
+			f(sb, arg);
 
 		spin_lock(&sb_lock);
 		if (p)
@@ -991,18 +987,15 @@ struct super_block *user_get_super(dev_t dev, bool excl)
 
 	spin_lock(&sb_lock);
 	list_for_each_entry(sb, &super_blocks, s_list) {
-		if (sb->s_dev ==  dev) {
+		if (sb->s_dev == dev) {
 			bool locked;
 
 			sb->s_count++;
 			spin_unlock(&sb_lock);
 			/* still alive? */
 			locked = super_lock(sb, excl);
-			if (locked) {
-				if (sb->s_root)
-					return sb;
-				super_unlock(sb, excl);
-			}
+			if (locked)
+				return sb; /* caller will drop */
 			/* nope, got unmounted */
 			spin_lock(&sb_lock);
 			__put_super(sb);

-- 
2.47.2
Re: [PATCH v2 1/6] super: remove pointless s_root checks
Posted by Darrick J. Wong 6 months, 1 week ago 
On Sat, Mar 29, 2025 at 09:42:14AM +0100, Christian Brauner wrote:
> The locking guarantees that the superblock is alive and sb->s_root is
> still set. Remove the pointless check.
> 
> Signed-off-by: Christian Brauner <brauner@kernel.org>
> ---
>  fs/super.c | 19 ++++++-------------
>  1 file changed, 6 insertions(+), 13 deletions(-)
> 
> diff --git a/fs/super.c b/fs/super.c
> index 97a17f9d9023..dc14f4bf73a6 100644
> --- a/fs/super.c
> +++ b/fs/super.c
> @@ -930,8 +930,7 @@ void iterate_supers(void (*f)(struct super_block *, void *), void *arg)
>  
>  		locked = super_lock_shared(sb);
>  		if (locked) {
> -			if (sb->s_root)
> -				f(sb, arg);
> +			f(sb, arg);
>  			super_unlock_shared(sb);
>  		}
>  
> @@ -967,11 +966,8 @@ void iterate_supers_type(struct file_system_type *type,
>  		spin_unlock(&sb_lock);
>  
>  		locked = super_lock_shared(sb);
> -		if (locked) {
> -			if (sb->s_root)
> -				f(sb, arg);
> -			super_unlock_shared(sb);
> -		}
> +		if (locked)
> +			f(sb, arg);

Hey Christian,

I might be trying to be the second(?) user of iterate_supers_type[1]. :)

This change removes the call to super_unlock_shared, which means that
iterate_supers_type returns with the super_lock(s) still held.  I'm
guessing that this is a bug and not an intentional change to require the
callback to call super_unlock_shared, right?

--D

[1] https://git.kernel.org/pub/scm/linux/kernel/git/djwong/xfs-linux.git/commit/?h=health-monitoring&id=3ae9b1d43dcdeaa38e93dc400d1871872ba0e27f

>  
>  		spin_lock(&sb_lock);
>  		if (p)
> @@ -991,18 +987,15 @@ struct super_block *user_get_super(dev_t dev, bool excl)
>  
>  	spin_lock(&sb_lock);
>  	list_for_each_entry(sb, &super_blocks, s_list) {
> -		if (sb->s_dev ==  dev) {
> +		if (sb->s_dev == dev) {
>  			bool locked;
>  
>  			sb->s_count++;
>  			spin_unlock(&sb_lock);
>  			/* still alive? */
>  			locked = super_lock(sb, excl);
> -			if (locked) {
> -				if (sb->s_root)
> -					return sb;
> -				super_unlock(sb, excl);
> -			}
> +			if (locked)
> +				return sb; /* caller will drop */
>  			/* nope, got unmounted */
>  			spin_lock(&sb_lock);
>  			__put_super(sb);
> 
> -- 
> 2.47.2
> 
>
Re: [PATCH v2 1/6] super: remove pointless s_root checks
Posted by Christian Brauner 6 months, 1 week ago 
On Wed, Jun 11, 2025 at 09:26:29AM -0700, Darrick J. Wong wrote:
> On Sat, Mar 29, 2025 at 09:42:14AM +0100, Christian Brauner wrote:
> > The locking guarantees that the superblock is alive and sb->s_root is
> > still set. Remove the pointless check.
> > 
> > Signed-off-by: Christian Brauner <brauner@kernel.org>
> > ---
> >  fs/super.c | 19 ++++++-------------
> >  1 file changed, 6 insertions(+), 13 deletions(-)
> > 
> > diff --git a/fs/super.c b/fs/super.c
> > index 97a17f9d9023..dc14f4bf73a6 100644
> > --- a/fs/super.c
> > +++ b/fs/super.c
> > @@ -930,8 +930,7 @@ void iterate_supers(void (*f)(struct super_block *, void *), void *arg)
> >  
> >  		locked = super_lock_shared(sb);
> >  		if (locked) {
> > -			if (sb->s_root)
> > -				f(sb, arg);
> > +			f(sb, arg);
> >  			super_unlock_shared(sb);
> >  		}
> >  
> > @@ -967,11 +966,8 @@ void iterate_supers_type(struct file_system_type *type,
> >  		spin_unlock(&sb_lock);
> >  
> >  		locked = super_lock_shared(sb);
> > -		if (locked) {
> > -			if (sb->s_root)
> > -				f(sb, arg);
> > -			super_unlock_shared(sb);
> > -		}
> > +		if (locked)
> > +			f(sb, arg);
> 
> Hey Christian,
> 
> I might be trying to be the second(?) user of iterate_supers_type[1]. :)
> 
> This change removes the call to super_unlock_shared, which means that
> iterate_supers_type returns with the super_lock(s) still held.  I'm
> guessing that this is a bug and not an intentional change to require the
> callback to call super_unlock_shared, right?
> 
> --D
> 
> [1] https://git.kernel.org/pub/scm/linux/kernel/git/djwong/xfs-linux.git/commit/?h=health-monitoring&id=3ae9b1d43dcdeaa38e93dc400d1871872ba0e27f

Yes, that's a bug. Can you send me a fix, please?
Re: [PATCH v2 1/6] super: remove pointless s_root checks
Posted by Jan Kara 8 months, 3 weeks ago 
On Sat 29-03-25 09:42:14, Christian Brauner wrote:
> The locking guarantees that the superblock is alive and sb->s_root is
> still set. Remove the pointless check.
> 
> Signed-off-by: Christian Brauner <brauner@kernel.org>

Looks good. In fact most sb->s_root checks in fs/super.c look pointless
these days since AFAICT if you have SB_BORN && !SB_DYING superblock (as
super_lock_*() ascertains), then sb->s_root != NULL. Anyway feel free to
add:

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

								Honza

> ---
>  fs/super.c | 19 ++++++-------------
>  1 file changed, 6 insertions(+), 13 deletions(-)
> 
> diff --git a/fs/super.c b/fs/super.c
> index 97a17f9d9023..dc14f4bf73a6 100644
> --- a/fs/super.c
> +++ b/fs/super.c
> @@ -930,8 +930,7 @@ void iterate_supers(void (*f)(struct super_block *, void *), void *arg)
>  
>  		locked = super_lock_shared(sb);
>  		if (locked) {
> -			if (sb->s_root)
> -				f(sb, arg);
> +			f(sb, arg);
>  			super_unlock_shared(sb);
>  		}
>  
> @@ -967,11 +966,8 @@ void iterate_supers_type(struct file_system_type *type,
>  		spin_unlock(&sb_lock);
>  
>  		locked = super_lock_shared(sb);
> -		if (locked) {
> -			if (sb->s_root)
> -				f(sb, arg);
> -			super_unlock_shared(sb);
> -		}
> +		if (locked)
> +			f(sb, arg);
>  
>  		spin_lock(&sb_lock);
>  		if (p)
> @@ -991,18 +987,15 @@ struct super_block *user_get_super(dev_t dev, bool excl)
>  
>  	spin_lock(&sb_lock);
>  	list_for_each_entry(sb, &super_blocks, s_list) {
> -		if (sb->s_dev ==  dev) {
> +		if (sb->s_dev == dev) {
>  			bool locked;
>  
>  			sb->s_count++;
>  			spin_unlock(&sb_lock);
>  			/* still alive? */
>  			locked = super_lock(sb, excl);
> -			if (locked) {
> -				if (sb->s_root)
> -					return sb;
> -				super_unlock(sb, excl);
> -			}
> +			if (locked)
> +				return sb; /* caller will drop */
>  			/* nope, got unmounted */
>  			spin_lock(&sb_lock);
>  			__put_super(sb);
> 
> -- 
> 2.47.2
> 
-- 
Jan Kara <jack@suse.com>
SUSE Labs, CR

Patchew 2.1-devel-b67e4c2

© 2016 - 2025 Red Hat, Inc.