[PATCH v2 1/4] include/linux/fs.h: add inode_lock_killable()

[PATCH v2 1/4] include/linux/fs.h: add inode_lock_killable()

[Max Kellermann]

Prepare for making inode operations killable while they're waiting for
the lock.

Signed-off-by: Max Kellermann <max.kellermann@ionos.com>
---
 include/linux/fs.h | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/include/linux/fs.h b/include/linux/fs.h
index 016b0fe1536e..5e4ac873228d 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -867,6 +867,11 @@ static inline void inode_lock(struct inode *inode)
 	down_write(&inode->i_rwsem);
 }
 
+static inline __must_check int inode_lock_killable(struct inode *inode)
+{
+	return down_write_killable(&inode->i_rwsem);
+}
+
 static inline void inode_unlock(struct inode *inode)
 {
 	up_write(&inode->i_rwsem);
@@ -877,6 +882,11 @@ static inline void inode_lock_shared(struct inode *inode)
 	down_read(&inode->i_rwsem);
 }
 
+static inline __must_check int inode_lock_shared_killable(struct inode *inode)
+{
+	return down_read_killable(&inode->i_rwsem);
+}
+
 static inline void inode_unlock_shared(struct inode *inode)
 {
 	up_read(&inode->i_rwsem);
-- 
2.47.2

[PATCH v2 2/4] fs/open: make chmod_common() and chown_common() killable

[Max Kellermann]

Allows killing processes that are waiting for the inode lock.

Signed-off-by: Max Kellermann <max.kellermann@ionos.com>
---
v2: split into separate patches

This part was reviewed by Christian Brauner here:
 https://lore.kernel.org/linux-fsdevel/20250512-unrat-kapital-2122d3777c5d@brauner/

Signed-off-by: Max Kellermann <max.kellermann@ionos.com>
---
 fs/open.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

diff --git a/fs/open.c b/fs/open.c
index a9063cca9911..d2f2df52c458 100644
--- a/fs/open.c
+++ b/fs/open.c
@@ -635,7 +635,9 @@ int chmod_common(const struct path *path, umode_t mode)
 	if (error)
 		return error;
 retry_deleg:
-	inode_lock(inode);
+	error = inode_lock_killable(inode);
+	if (error)
+		goto out_mnt_unlock;
 	error = security_path_chmod(path, mode);
 	if (error)
 		goto out_unlock;
@@ -650,6 +652,7 @@ int chmod_common(const struct path *path, umode_t mode)
 		if (!error)
 			goto retry_deleg;
 	}
+out_mnt_unlock:
 	mnt_drop_write(path->mnt);
 	return error;
 }
@@ -769,7 +772,9 @@ int chown_common(const struct path *path, uid_t user, gid_t group)
 		return -EINVAL;
 	if ((group != (gid_t)-1) && !setattr_vfsgid(&newattrs, gid))
 		return -EINVAL;
-	inode_lock(inode);
+	error = inode_lock_killable(inode);
+	if (error)
+		return error;
 	if (!S_ISDIR(inode->i_mode))
 		newattrs.ia_valid |= ATTR_KILL_SUID | ATTR_KILL_PRIV |
 				     setattr_should_drop_sgid(idmap, inode);
-- 
2.47.2

[PATCH v2 3/4] fs/open: make do_truncate() killable

[Max Kellermann]

Allows killing processes that are waiting for the inode lock.

Signed-off-by: Max Kellermann <max.kellermann@ionos.com>
---
v2: split into separate patches

Review here (though nothing about do_truncate()):
 https://lore.kernel.org/linux-fsdevel/20250512-unrat-kapital-2122d3777c5d@brauner/

Signed-off-by: Max Kellermann <max.kellermann@ionos.com>
---
 fs/open.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/fs/open.c b/fs/open.c
index d2f2df52c458..7828234a7caa 100644
--- a/fs/open.c
+++ b/fs/open.c
@@ -60,7 +60,10 @@ int do_truncate(struct mnt_idmap *idmap, struct dentry *dentry,
 	if (ret)
 		newattrs.ia_valid |= ret | ATTR_FORCE;
 
-	inode_lock(dentry->d_inode);
+	ret = inode_lock_killable(dentry->d_inode);
+	if (ret)
+		return ret;
+
 	/* Note any delegations or leases have already been broken: */
 	ret = notify_change(idmap, dentry, &newattrs, NULL);
 	inode_unlock(dentry->d_inode);
-- 
2.47.2

[PATCH v2 4/4] fs/read_write: make default_llseek() killable

[Max Kellermann]

Allows killing processes that are waiting for the inode lock.

Signed-off-by: Max Kellermann <max.kellermann@ionos.com>
---
v2: split into separate patches

TODO: review whether all callers can handle EINTR; see
 https://lore.kernel.org/linux-fsdevel/20250512-unrat-kapital-2122d3777c5d@brauner/
and
 https://lore.kernel.org/linux-fsdevel/hzrj5b7x3rvtxt4qgjxdihhi5vjoc5gw3i35pbyopa7ccucizo@q5c42kjlkly3/

Signed-off-by: Max Kellermann <max.kellermann@ionos.com>
---
 fs/read_write.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/fs/read_write.c b/fs/read_write.c
index bb0ed26a0b3a..0ef70e128c4a 100644
--- a/fs/read_write.c
+++ b/fs/read_write.c
@@ -332,7 +332,9 @@ loff_t default_llseek(struct file *file, loff_t offset, int whence)
 	struct inode *inode = file_inode(file);
 	loff_t retval;
 
-	inode_lock(inode);
+	retval = inode_lock_killable(inode);
+	if (retval)
+		return retval;
 	switch (whence) {
 		case SEEK_END:
 			offset += i_size_read(inode);
-- 
2.47.2

Re: [PATCH v2 1/4] include/linux/fs.h: add inode_lock_killable()

[Christian Brauner]

On Tue, 13 May 2025 17:03:24 +0200, Max Kellermann wrote:
> Prepare for making inode operations killable while they're waiting for
> the lock.
> 
> 

Applied to the vfs-6.16.misc branch of the vfs/vfs.git tree.
Patches in the vfs-6.16.misc branch should appear in linux-next soon.

Please report any outstanding bugs that were missed during review in a
new review to the original patch series allowing us to drop it.

It's encouraged to provide Acked-bys and Reviewed-bys even though the
patch has now been applied. If possible patch trailers will be updated.

Note that commit hashes shown below are subject to change due to rebase,
trailer updates or similar. If in doubt, please check the listed branch.

tree:   https://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs.git
branch: vfs-6.16.misc

[1/4] include/linux/fs.h: add inode_lock_killable()
      https://git.kernel.org/vfs/vfs/c/d8c5507cd140
[2/4] fs/open: make chmod_common() and chown_common() killable
      https://git.kernel.org/vfs/vfs/c/28a3f6ab2fe0
[3/4] fs/open: make do_truncate() killable
      https://git.kernel.org/vfs/vfs/c/d68687564280
[4/4] fs/read_write: make default_llseek() killable
      https://git.kernel.org/vfs/vfs/c/2e1a8fbff51b

Re: [PATCH v2 1/4] include/linux/fs.h: add inode_lock_killable()

[Jan Kara]

On Tue 13-05-25 17:03:24, Max Kellermann wrote:
> Prepare for making inode operations killable while they're waiting for
> the lock.
> 
> Signed-off-by: Max Kellermann <max.kellermann@ionos.com>

Looks good. Feel free to add:

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

								Honza

> ---
>  include/linux/fs.h | 10 ++++++++++
>  1 file changed, 10 insertions(+)
> 
> diff --git a/include/linux/fs.h b/include/linux/fs.h
> index 016b0fe1536e..5e4ac873228d 100644
> --- a/include/linux/fs.h
> +++ b/include/linux/fs.h
> @@ -867,6 +867,11 @@ static inline void inode_lock(struct inode *inode)
>  	down_write(&inode->i_rwsem);
>  }
>  
> +static inline __must_check int inode_lock_killable(struct inode *inode)
> +{
> +	return down_write_killable(&inode->i_rwsem);
> +}
> +
>  static inline void inode_unlock(struct inode *inode)
>  {
>  	up_write(&inode->i_rwsem);
> @@ -877,6 +882,11 @@ static inline void inode_lock_shared(struct inode *inode)
>  	down_read(&inode->i_rwsem);
>  }
>  
> +static inline __must_check int inode_lock_shared_killable(struct inode *inode)
> +{
> +	return down_read_killable(&inode->i_rwsem);
> +}
> +
>  static inline void inode_unlock_shared(struct inode *inode)
>  {
>  	up_read(&inode->i_rwsem);
> -- 
> 2.47.2
> 
-- 
Jan Kara <jack@suse.com>
SUSE Labs, CR

Re: [PATCH v2 2/4] fs/open: make chmod_common() and chown_common() killable

[Jan Kara]

On Tue 13-05-25 17:03:25, Max Kellermann wrote:
> Allows killing processes that are waiting for the inode lock.
> 
> Signed-off-by: Max Kellermann <max.kellermann@ionos.com>

Looks good. Feel free to add:

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

								Honza

> ---
> v2: split into separate patches
> 
> This part was reviewed by Christian Brauner here:
>  https://lore.kernel.org/linux-fsdevel/20250512-unrat-kapital-2122d3777c5d@brauner/
> 
> Signed-off-by: Max Kellermann <max.kellermann@ionos.com>
> ---
>  fs/open.c | 9 +++++++--
>  1 file changed, 7 insertions(+), 2 deletions(-)
> 
> diff --git a/fs/open.c b/fs/open.c
> index a9063cca9911..d2f2df52c458 100644
> --- a/fs/open.c
> +++ b/fs/open.c
> @@ -635,7 +635,9 @@ int chmod_common(const struct path *path, umode_t mode)
>  	if (error)
>  		return error;
>  retry_deleg:
> -	inode_lock(inode);
> +	error = inode_lock_killable(inode);
> +	if (error)
> +		goto out_mnt_unlock;
>  	error = security_path_chmod(path, mode);
>  	if (error)
>  		goto out_unlock;
> @@ -650,6 +652,7 @@ int chmod_common(const struct path *path, umode_t mode)
>  		if (!error)
>  			goto retry_deleg;
>  	}
> +out_mnt_unlock:
>  	mnt_drop_write(path->mnt);
>  	return error;
>  }
> @@ -769,7 +772,9 @@ int chown_common(const struct path *path, uid_t user, gid_t group)
>  		return -EINVAL;
>  	if ((group != (gid_t)-1) && !setattr_vfsgid(&newattrs, gid))
>  		return -EINVAL;
> -	inode_lock(inode);
> +	error = inode_lock_killable(inode);
> +	if (error)
> +		return error;
>  	if (!S_ISDIR(inode->i_mode))
>  		newattrs.ia_valid |= ATTR_KILL_SUID | ATTR_KILL_PRIV |
>  				     setattr_should_drop_sgid(idmap, inode);
> -- 
> 2.47.2
> 
-- 
Jan Kara <jack@suse.com>
SUSE Labs, CR

Re: [PATCH v2 3/4] fs/open: make do_truncate() killable

[Jan Kara]

On Tue 13-05-25 17:03:26, Max Kellermann wrote:
> Allows killing processes that are waiting for the inode lock.
> 
> Signed-off-by: Max Kellermann <max.kellermann@ionos.com>

Looks good. Feel free to add:

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

								Honza

> ---
> v2: split into separate patches
> 
> Review here (though nothing about do_truncate()):
>  https://lore.kernel.org/linux-fsdevel/20250512-unrat-kapital-2122d3777c5d@brauner/
> 
> Signed-off-by: Max Kellermann <max.kellermann@ionos.com>
> ---
>  fs/open.c | 5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
> 
> diff --git a/fs/open.c b/fs/open.c
> index d2f2df52c458..7828234a7caa 100644
> --- a/fs/open.c
> +++ b/fs/open.c
> @@ -60,7 +60,10 @@ int do_truncate(struct mnt_idmap *idmap, struct dentry *dentry,
>  	if (ret)
>  		newattrs.ia_valid |= ret | ATTR_FORCE;
>  
> -	inode_lock(dentry->d_inode);
> +	ret = inode_lock_killable(dentry->d_inode);
> +	if (ret)
> +		return ret;
> +
>  	/* Note any delegations or leases have already been broken: */
>  	ret = notify_change(idmap, dentry, &newattrs, NULL);
>  	inode_unlock(dentry->d_inode);
> -- 
> 2.47.2
> 
-- 
Jan Kara <jack@suse.com>
SUSE Labs, CR

Re: [PATCH v2 4/4] fs/read_write: make default_llseek() killable

[Jan Kara]

On Tue 13-05-25 17:03:27, Max Kellermann wrote:
> Allows killing processes that are waiting for the inode lock.
> 
> Signed-off-by: Max Kellermann <max.kellermann@ionos.com>

Looks good. Feel free to add:

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

> v2: split into separate patches
> 
> TODO: review whether all callers can handle EINTR; see

I did a quick audit and everything seems OK AFAICT.

								Honza

>  https://lore.kernel.org/linux-fsdevel/20250512-unrat-kapital-2122d3777c5d@brauner/
> and
>  https://lore.kernel.org/linux-fsdevel/hzrj5b7x3rvtxt4qgjxdihhi5vjoc5gw3i35pbyopa7ccucizo@q5c42kjlkly3/
> 
> Signed-off-by: Max Kellermann <max.kellermann@ionos.com>
> ---
>  fs/read_write.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/fs/read_write.c b/fs/read_write.c
> index bb0ed26a0b3a..0ef70e128c4a 100644
> --- a/fs/read_write.c
> +++ b/fs/read_write.c
> @@ -332,7 +332,9 @@ loff_t default_llseek(struct file *file, loff_t offset, int whence)
>  	struct inode *inode = file_inode(file);
>  	loff_t retval;
>  
> -	inode_lock(inode);
> +	retval = inode_lock_killable(inode);
> +	if (retval)
> +		return retval;
>  	switch (whence) {
>  		case SEEK_END:
>  			offset += i_size_read(inode);
> -- 
> 2.47.2
> 
-- 
Jan Kara <jack@suse.com>
SUSE Labs, CR