Re: [PATCH][RFC][CFT] kill vfs_submount(), already

[Christian Brauner <brauner@kernel.org>]

On Sat, May 03, 2025 at 10:29:25PM +0100, Al Viro wrote:
> The last remaining user of vfs_submount() (tracefs) is easy to convert
> to fs_context_for_submount(); do that and bury that thing, along with
> SB_SUBMOUNT
> 
> If nobody objects, I'm going to throw that into the mount-related pile;
> alternatively, that could be split into kernel/trace.c part (in invariant
> branch, to be pulled by tracefs folks and into the mount pile before
> the rest of the patch).  Preferences?
> 
> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
> ---

Looks good,
Reviewed-by: Christian Brauner <brauner@kernel.org>

> diff --git a/fs/namespace.c b/fs/namespace.c
> index 07f636036b86..293e6f925eff 100644
> --- a/fs/namespace.c
> +++ b/fs/namespace.c
> @@ -1297,21 +1297,6 @@ struct vfsmount *vfs_kern_mount(struct file_system_type *type,
>  }
>  EXPORT_SYMBOL_GPL(vfs_kern_mount);
>  
> -struct vfsmount *
> -vfs_submount(const struct dentry *mountpoint, struct file_system_type *type,
> -	     const char *name, void *data)
> -{
> -	/* Until it is worked out how to pass the user namespace
> -	 * through from the parent mount to the submount don't support
> -	 * unprivileged mounts with submounts.
> -	 */
> -	if (mountpoint->d_sb->s_user_ns != &init_user_ns)
> -		return ERR_PTR(-EPERM);
> -
> -	return vfs_kern_mount(type, SB_SUBMOUNT, name, data);
> -}
> -EXPORT_SYMBOL_GPL(vfs_submount);
> -
>  static struct mount *clone_mnt(struct mount *old, struct dentry *root,
>  					int flag)
>  {
> diff --git a/fs/super.c b/fs/super.c
> index 97a17f9d9023..1886e4c930e0 100644
> --- a/fs/super.c
> +++ b/fs/super.c
> @@ -823,13 +823,6 @@ struct super_block *sget(struct file_system_type *type,
>  	struct super_block *old;
>  	int err;
>  
> -	/* We don't yet pass the user namespace of the parent
> -	 * mount through to here so always use &init_user_ns
> -	 * until that changes.
> -	 */
> -	if (flags & SB_SUBMOUNT)
> -		user_ns = &init_user_ns;

Oh thank god this disgusting hack is finally gone.

> -
>  retry:
>  	spin_lock(&sb_lock);
>  	if (test) {
> @@ -849,7 +842,7 @@ struct super_block *sget(struct file_system_type *type,
>  	}
>  	if (!s) {
>  		spin_unlock(&sb_lock);
> -		s = alloc_super(type, (flags & ~SB_SUBMOUNT), user_ns);
> +		s = alloc_super(type, flags, user_ns);
>  		if (!s)
>  			return ERR_PTR(-ENOMEM);
>  		goto retry;
> diff --git a/include/linux/fs.h b/include/linux/fs.h
> index 016b0fe1536e..515e702d98ae 100644
> --- a/include/linux/fs.h
> +++ b/include/linux/fs.h
> @@ -1240,7 +1240,6 @@ extern int send_sigurg(struct file *file);
>  /* These sb flags are internal to the kernel */
>  #define SB_DEAD         BIT(21)
>  #define SB_DYING        BIT(24)
> -#define SB_SUBMOUNT     BIT(26)
>  #define SB_FORCE        BIT(27)
>  #define SB_NOSEC        BIT(28)
>  #define SB_BORN         BIT(29)
> diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
> index fa488721019f..7b6248ba4428 100644
> --- a/kernel/trace/trace.c
> +++ b/kernel/trace/trace.c
> @@ -51,6 +51,7 @@
>  #include <linux/workqueue.h>
>  #include <linux/sort.h>
>  #include <linux/io.h> /* vmap_page_range() */
> +#include <linux/fs_context.h>
>  
>  #include <asm/setup.h> /* COMMAND_LINE_SIZE */
>  
> @@ -10072,6 +10073,8 @@ static struct vfsmount *trace_automount(struct dentry *mntpt, void *ingore)
>  {
>  	struct vfsmount *mnt;
>  	struct file_system_type *type;
> +	struct fs_context *fc;
> +	int ret;
>  
>  	/*
>  	 * To maintain backward compatibility for tools that mount
> @@ -10081,10 +10084,20 @@ static struct vfsmount *trace_automount(struct dentry *mntpt, void *ingore)
>  	type = get_fs_type("tracefs");
>  	if (!type)
>  		return NULL;
> -	mnt = vfs_submount(mntpt, type, "tracefs", NULL);
> +
> +	fc = fs_context_for_submount(type, mntpt);
> +	if (IS_ERR(fc))
> +		return ERR_CAST(fc);
> +
> +	ret = vfs_parse_fs_string(fc, "source",
> +				  "tracefs", strlen("tracefs"));
> +	if (!ret)
> +		mnt = fc_mount(fc);
> +	else
> +		mnt = ERR_PTR(ret);
> +
> +	put_fs_context(fc);
>  	put_filesystem(type);
> -	if (IS_ERR(mnt))
> -		return NULL;
>  	return mnt;
>  }
>