Re: [PATCH v2] vfs: replace ints with enum component_type for LAST_XXX

[Amir Goldstein]

On Sun, Apr 19, 2026 at 6:16 PM Jori Koolstra <jkoolstra@xxxxxxxxx> wrote:
>
> Several functions in namei.c take an "int *type" parameter, such as
> filename_parentat(). To know what values this can take you have to find
> the anonymous struct that defines the LAST_XXX values. I would argue
> that the readability of the code is improved by making this an explicit
> type.
>
> Signed-off-by: Jori Koolstra <jkoolstra@xxxxxxxxx>
> Reviewed-by: Jan Kara <jack@xxxxxxx>
>
> ---
>
> v2: move back to LAST_XXX and change int to component_type in
> fs/smb/server/vfs.c.
>
> ---
>  fs/namei.c            | 41 ++++++++++++++++++++++-------------------
>  fs/smb/server/vfs.c   |  5 +++--
>  include/linux/namei.h |  4 ++--
>  3 files changed, 27 insertions(+), 23 deletions(-)
>
> diff --git a/fs/namei.c b/fs/namei.c
> index 9e5500dad14f..a880454a6415 100644
> --- a/fs/namei.c
> +++ b/fs/namei.c
> @@ -721,15 +721,15 @@ EXPORT_SYMBOL(path_put);
>
>  #define EMBEDDED_LEVELS 2
>  struct nameidata {
> -       struct path     path;
> -       struct qstr     last;
> -       struct path     root;
> -       struct inode    *inode; /* path.dentry.d_inode */
> -       unsigned int    flags, state;
> -       unsigned        seq, next_seq, m_seq, r_seq;
> -       int             last_type;

Please update documentation entry for last_type in
Documentation/filesystems/path-lookup.rst

> -       unsigned        depth;
> -       int             total_link_count;
> +       struct path             path;
> +       struct qstr             last;
> +       struct path             root;
> +       struct inode            *inode; /* path.dentry.d_inode */
> +       unsigned int            flags, state;
> +       unsigned                seq, next_seq, m_seq, r_seq;
> +       enum component_type     last_type;

I really dislike all this white space churn just because the enum name
is so long. If it was long for a good reason - increasing readability -
I may have conceded, but I don't think that is the case.

I do not think that 'component_type' is more clear than 'last_type'
for a developer passing by reading the code.
If anything, the opposite, because 'last_type' or 'type' variables always
appear in the same function with 'last' variable, so it is more clear WHICH
object the type is referring to and matches the LAST_ enum prefix.

> +       unsigned                depth;
> +       int                     total_link_count;
>         struct saved {
>                 struct path link;
>                 struct delayed_call done;
> @@ -2221,7 +2221,7 @@ static struct dentry *follow_dotdot(struct nameidata *nd)
>         return dget(nd->path.dentry);
>  }
>
> -static const char *handle_dots(struct nameidata *nd, int type)
> +static const char *handle_dots(struct nameidata *nd, enum component_type type)
>  {
>         if (type == LAST_DOTDOT) {
>                 const char *error = NULL;
> @@ -2869,7 +2869,7 @@ static int path_parentat(struct nameidata *nd, unsigned flags,
>  /* Note: this does not consume "name" */
>  static int __filename_parentat(int dfd, struct filename *name,
>                                unsigned int flags, struct path *parent,
> -                              struct qstr *last, int *type,
> +                              struct qstr *last, enum component_type *type,
>                                const struct path *root)
>  {
>         int retval;
> @@ -2894,7 +2894,7 @@ static int __filename_parentat(int dfd, struct filename *name,
>
>  static int filename_parentat(int dfd, struct filename *name,
>                              unsigned int flags, struct path *parent,
> -                            struct qstr *last, int *type)
> +                            struct qstr *last, enum component_type *type)
>  {
>         return __filename_parentat(dfd, name, flags, parent, last, type, NULL);
>  }
> @@ -2963,7 +2963,8 @@ static struct dentry *__start_removing_path(int dfd, struct filename *name,
>         struct path parent_path __free(path_put) = {};
>         struct dentry *d;
>         struct qstr last;
> -       int type, error;
> +       enum component_type type;
> +       int error;
>
>         error = filename_parentat(dfd, name, 0, &parent_path, &last, &type);
>         if (error)
> @@ -3009,7 +3010,8 @@ struct dentry *kern_path_parent(const char *name, struct path *path)
>         CLASS(filename_kernel, filename)(name);
>         struct dentry *d;
>         struct qstr last;
> -       int type, error;
> +       enum component_type type;
> +       int error;
>
>         error = filename_parentat(AT_FDCWD, filename, 0, &parent_path, &last, &type);
>         if (error)
> @@ -3057,7 +3059,7 @@ EXPORT_SYMBOL(kern_path);
>   * @root: pointer to struct path of the base directory
>   */
>  int vfs_path_parent_lookup(struct filename *filename, unsigned int flags,
> -                          struct path *parent, struct qstr *last, int *type,
> +                          struct path *parent, struct qstr *last, enum component_type *type,
>                            const struct path *root)
>  {
>         return  __filename_parentat(AT_FDCWD, filename, flags, parent, last,
> @@ -4903,7 +4905,7 @@ static struct dentry *filename_create(int dfd, struct filename *name,
>         bool want_dir = lookup_flags & LOOKUP_DIRECTORY;
>         unsigned int reval_flag = lookup_flags & LOOKUP_REVAL;
>         unsigned int create_flags = LOOKUP_CREATE | LOOKUP_EXCL;
> -       int type;
> +       enum component_type type;
>         int error;
>
>         error = filename_parentat(dfd, name, reval_flag, path, &last, &type);
> @@ -5365,7 +5367,7 @@ int filename_rmdir(int dfd, struct filename *name)
>         struct dentry *dentry;
>         struct path path;
>         struct qstr last;
> -       int type;
> +       enum component_type type;
>         unsigned int lookup_flags = 0;
>         struct delegated_inode delegated_inode = { };
>  retry:
> @@ -5383,6 +5385,7 @@ int filename_rmdir(int dfd, struct filename *name)
>         case LAST_ROOT:
>                 error = -EBUSY;
>                 goto exit2;
> +       case LAST_NORM: ; // OK

I was not aware that the compiler warns about missing cases for enums -
cool feature.
but that needs to be a break; rather than fallthough the end
and as a matter of taste, I would put the valid case first - not critical.

>         }
>
>         error = mnt_want_write(path.mnt);
> @@ -5507,7 +5510,7 @@ int filename_unlinkat(int dfd, struct filename *name)
>         struct dentry *dentry;
>         struct path path;
>         struct qstr last;
> -       int type;
> +       enum component_type type;
>         struct inode *inode;
>         struct delegated_inode delegated_inode = { };
>         unsigned int lookup_flags = 0;
> @@ -6074,7 +6077,7 @@ int filename_renameat2(int olddfd, struct filename *from,
>         struct renamedata rd;
>         struct path old_path, new_path;
>         struct qstr old_last, new_last;
> -       int old_type, new_type;
> +       enum component_type old_type, new_type;
>         struct delegated_inode delegated_inode = { };
>         unsigned int lookup_flags = 0;
>         bool should_retry = false;
> diff --git a/fs/smb/server/vfs.c b/fs/smb/server/vfs.c
> index d08973b288e5..b12d481f5ba9 100644
> --- a/fs/smb/server/vfs.c
> +++ b/fs/smb/server/vfs.c
> @@ -56,7 +56,8 @@ static int ksmbd_vfs_path_lookup(struct ksmbd_share_config *share_conf,
>  {
>         struct qstr last;
>         const struct path *root_share_path = &share_conf->vfs_path;
> -       int err, type;
> +       int err;
> +       enum component_type type;
>         struct dentry *d;
>
>         if (pathname[0] == '\0') {
> @@ -668,7 +669,7 @@ int ksmbd_vfs_rename(struct ksmbd_work *work, const struct path *old_path,
>         struct renamedata rd;
>         struct ksmbd_share_config *share_conf = work->tcon->share_conf;
>         struct ksmbd_file *parent_fp;
> -       int new_type;
> +       enum component_type new_type;
>         int err, lookup_flags = LOOKUP_NO_SYMLINKS;
>
>         if (ksmbd_override_fsids(work))
> diff --git a/include/linux/namei.h b/include/linux/namei.h
> index 58600cf234bc..fa3ae87762b7 100644
> --- a/include/linux/namei.h
> +++ b/include/linux/namei.h
> @@ -16,7 +16,7 @@ enum { MAX_NESTED_LINKS = 8 };
>  /*
>   * Type of the last component on LOOKUP_PARENT
>   */
> -enum {LAST_NORM, LAST_ROOT, LAST_DOT, LAST_DOTDOT};
> +enum component_type {LAST_NORM, LAST_ROOT, LAST_DOT, LAST_DOTDOT};

Please use last_type to match the LAST_ prefix.

TBH, the thing that bothers me most is why is this enum exposed
outside of namei.c
in the first place?

>
>  /* pathwalk mode */
>  #define LOOKUP_FOLLOW          BIT(0)  /* follow links at the end */
> @@ -70,7 +70,7 @@ static inline void end_removing_path(const struct path *path , struct dentry *de
>         end_creating_path(path, dentry);
>  }
>  int vfs_path_parent_lookup(struct filename *filename, unsigned int flags,
> -                          struct path *parent, struct qstr *last, int *type,
> +                          struct path *parent, struct qstr *last, enum component_type *type,
>                            const struct path *root);

As your patch demonstrates, the only use of enum last_type outside of namei.c is
in ksmbd calls to vfs_path_parent_lookup().

This is an "internal" lookup helper that was exposed by commit
74d7970febf7e ("ksmbd: fix racy issue from using ->d_parent and ->d_name")

This commit appears to be using the helper inconsistently in
its two call sites, because ksmbd_vfs_rename(), is not checking the
unlikely(type != LAST_NORM) case.

Perhaps it would be wiser not to expose enum last_type at all and move
the unlikely(type != LAST_NORM) check into the helper itself in namei.c.

WDYT?

Thanks,
Amir.