[PATCH v2 0/2] io_uring: add mkdirat support

[PATCH v2 0/2] io_uring: add mkdirat support

[Dmitry Kadashev]

This adds mkdirat support to io_uring and is heavily based on recently
added renameat() / unlinkat() support.

The first patch is preparation with no functional changes, makes
do_mkdirat accept struct filename pointer rather than the user string.

The second one leverages that to implement mkdirat in io_uring.

Based on for-5.11/io_uring.

Changes since v1:
- do not mess with struct filename's refcount in do_mkdirat, instead add
  and use __filename_create() that does not drop the name on success;

Dmitry Kadashev (2):
  fs: make do_mkdirat() take struct filename
  io_uring: add support for IORING_OP_MKDIRAT

 fs/internal.h                 |  1 +
 fs/io_uring.c                 | 58 +++++++++++++++++++++++++++++++++++
 fs/namei.c                    | 25 +++++++++++----
 include/uapi/linux/io_uring.h |  1 +
 4 files changed, 79 insertions(+), 6 deletions(-)

-- 
2.30.0

[PATCH v2 1/2] fs: make do_mkdirat() take struct filename

[Dmitry Kadashev]

Pass in the struct filename pointers instead of the user string, and
update the three callers to do the same. This is heavily based on
commit dbea8d345177 ("fs: make do_renameat2() take struct filename").

This behaves like do_unlinkat() and do_renameat2().

Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Dmitry Kadashev <dkadashev@gmail.com>
---
 fs/internal.h |  1 +
 fs/namei.c    | 25 +++++++++++++++++++------
 2 files changed, 20 insertions(+), 6 deletions(-)

diff --git a/fs/internal.h b/fs/internal.h
index c6c85f6ad598..b10005dfaa48 100644
--- a/fs/internal.h
+++ b/fs/internal.h
@@ -76,6 +76,7 @@ long do_unlinkat(int dfd, struct filename *name);
 int may_linkat(struct path *link);
 int do_renameat2(int olddfd, struct filename *oldname, int newdfd,
 		 struct filename *newname, unsigned int flags);
+long do_mkdirat(int dfd, struct filename *name, umode_t mode);
 
 /*
  * namespace.c
diff --git a/fs/namei.c b/fs/namei.c
index 4cae88733a5c..3657bdf1aafc 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -3431,7 +3431,7 @@ struct file *do_file_open_root(struct dentry *dentry, struct vfsmount *mnt,
 	return file;
 }
 
-static struct dentry *filename_create(int dfd, struct filename *name,
+static struct dentry *__filename_create(int dfd, struct filename *name,
 				struct path *path, unsigned int lookup_flags)
 {
 	struct dentry *dentry = ERR_PTR(-EEXIST);
@@ -3487,7 +3487,6 @@ static struct dentry *filename_create(int dfd, struct filename *name,
 		error = err2;
 		goto fail;
 	}
-	putname(name);
 	return dentry;
 fail:
 	dput(dentry);
@@ -3502,6 +3501,16 @@ static struct dentry *filename_create(int dfd, struct filename *name,
 	return dentry;
 }
 
+static inline struct dentry *filename_create(int dfd, struct filename *name,
+				struct path *path, unsigned int lookup_flags)
+{
+	struct dentry *res = __filename_create(dfd, name, path, lookup_flags);
+
+	if (!IS_ERR(res))
+		putname(name);
+	return res;
+}
+
 struct dentry *kern_path_create(int dfd, const char *pathname,
 				struct path *path, unsigned int lookup_flags)
 {
@@ -3654,15 +3663,18 @@ int vfs_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode)
 }
 EXPORT_SYMBOL(vfs_mkdir);
 
-static long do_mkdirat(int dfd, const char __user *pathname, umode_t mode)
+long do_mkdirat(int dfd, struct filename *name, umode_t mode)
 {
 	struct dentry *dentry;
 	struct path path;
 	int error;
 	unsigned int lookup_flags = LOOKUP_DIRECTORY;
 
+	if (IS_ERR(name))
+		return PTR_ERR(name);
+
 retry:
-	dentry = user_path_create(dfd, pathname, &path, lookup_flags);
+	dentry = __filename_create(dfd, name, &path, lookup_flags);
 	if (IS_ERR(dentry))
 		return PTR_ERR(dentry);
 
@@ -3676,17 +3688,18 @@ static long do_mkdirat(int dfd, const char __user *pathname, umode_t mode)
 		lookup_flags |= LOOKUP_REVAL;
 		goto retry;
 	}
+	putname(name);
 	return error;
 }
 
 SYSCALL_DEFINE3(mkdirat, int, dfd, const char __user *, pathname, umode_t, mode)
 {
-	return do_mkdirat(dfd, pathname, mode);
+	return do_mkdirat(dfd, getname(pathname), mode);
 }
 
 SYSCALL_DEFINE2(mkdir, const char __user *, pathname, umode_t, mode)
 {
-	return do_mkdirat(AT_FDCWD, pathname, mode);
+	return do_mkdirat(AT_FDCWD, getname(pathname), mode);
 }
 
 int vfs_rmdir(struct inode *dir, struct dentry *dentry)
-- 
2.30.0

Re: [PATCH v2 1/2] fs: make do_mkdirat() take struct filename

[Dmitry Kadashev]

On Tue, Feb 2, 2021 at 3:25 PM Dmitry Kadashev <dkadashev@gmail.com> wrote:
>
> Pass in the struct filename pointers instead of the user string, and
> update the three callers to do the same. This is heavily based on
> commit dbea8d345177 ("fs: make do_renameat2() take struct filename").
>
> This behaves like do_unlinkat() and do_renameat2().
>
> Cc: Al Viro <viro@zeniv.linux.org.uk>
> Signed-off-by: Dmitry Kadashev <dkadashev@gmail.com>
> ---
>  fs/internal.h |  1 +
>  fs/namei.c    | 25 +++++++++++++++++++------
>  2 files changed, 20 insertions(+), 6 deletions(-)
>
> diff --git a/fs/internal.h b/fs/internal.h
> index c6c85f6ad598..b10005dfaa48 100644
> --- a/fs/internal.h
> +++ b/fs/internal.h
> @@ -76,6 +76,7 @@ long do_unlinkat(int dfd, struct filename *name);
>  int may_linkat(struct path *link);
>  int do_renameat2(int olddfd, struct filename *oldname, int newdfd,
>                  struct filename *newname, unsigned int flags);
> +long do_mkdirat(int dfd, struct filename *name, umode_t mode);
>
>  /*
>   * namespace.c
> diff --git a/fs/namei.c b/fs/namei.c
> index 4cae88733a5c..3657bdf1aafc 100644
> --- a/fs/namei.c
> +++ b/fs/namei.c
> @@ -3431,7 +3431,7 @@ struct file *do_file_open_root(struct dentry *dentry, struct vfsmount *mnt,
>         return file;
>  }
>
> -static struct dentry *filename_create(int dfd, struct filename *name,
> +static struct dentry *__filename_create(int dfd, struct filename *name,
>                                 struct path *path, unsigned int lookup_flags)
>  {
>         struct dentry *dentry = ERR_PTR(-EEXIST);
> @@ -3487,7 +3487,6 @@ static struct dentry *filename_create(int dfd, struct filename *name,
>                 error = err2;
>                 goto fail;
>         }
> -       putname(name);
>         return dentry;
>  fail:
>         dput(dentry);
> @@ -3502,6 +3501,16 @@ static struct dentry *filename_create(int dfd, struct filename *name,
>         return dentry;
>  }
>
> +static inline struct dentry *filename_create(int dfd, struct filename *name,
> +                               struct path *path, unsigned int lookup_flags)
> +{
> +       struct dentry *res = __filename_create(dfd, name, path, lookup_flags);
> +
> +       if (!IS_ERR(res))
> +               putname(name);
> +       return res;
> +}
> +
>  struct dentry *kern_path_create(int dfd, const char *pathname,
>                                 struct path *path, unsigned int lookup_flags)
>  {
> @@ -3654,15 +3663,18 @@ int vfs_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode)
>  }
>  EXPORT_SYMBOL(vfs_mkdir);
>
> -static long do_mkdirat(int dfd, const char __user *pathname, umode_t mode)
> +long do_mkdirat(int dfd, struct filename *name, umode_t mode)
>  {
>         struct dentry *dentry;
>         struct path path;
>         int error;
>         unsigned int lookup_flags = LOOKUP_DIRECTORY;
>
> +       if (IS_ERR(name))
> +               return PTR_ERR(name);
> +
>  retry:
> -       dentry = user_path_create(dfd, pathname, &path, lookup_flags);
> +       dentry = __filename_create(dfd, name, &path, lookup_flags);
>         if (IS_ERR(dentry))
>                 return PTR_ERR(dentry);
>
> @@ -3676,17 +3688,18 @@ static long do_mkdirat(int dfd, const char __user *pathname, umode_t mode)
>                 lookup_flags |= LOOKUP_REVAL;
>                 goto retry;
>         }
> +       putname(name);
>         return error;
>  }
>
>  SYSCALL_DEFINE3(mkdirat, int, dfd, const char __user *, pathname, umode_t, mode)
>  {
> -       return do_mkdirat(dfd, pathname, mode);
> +       return do_mkdirat(dfd, getname(pathname), mode);
>  }
>
>  SYSCALL_DEFINE2(mkdir, const char __user *, pathname, umode_t, mode)
>  {
> -       return do_mkdirat(AT_FDCWD, pathname, mode);
> +       return do_mkdirat(AT_FDCWD, getname(pathname), mode);
>  }
>
>  int vfs_rmdir(struct inode *dir, struct dentry *dentry)
> --
> 2.30.0

Hi Al,

Are you OK with this version?

-- 
Dmitry Kadashev

[PATCH v2 2/2] io_uring: add support for IORING_OP_MKDIRAT

[Dmitry Kadashev]

IORING_OP_MKDIRAT behaves like mkdirat(2) and takes the same flags
and arguments.

Signed-off-by: Dmitry Kadashev <dkadashev@gmail.com>
---
 fs/io_uring.c                 | 58 +++++++++++++++++++++++++++++++++++
 include/uapi/linux/io_uring.h |  1 +
 2 files changed, 59 insertions(+)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index 24ad36d71289..000d7dce5902 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -582,6 +582,13 @@ struct io_unlink {
 	struct filename			*filename;
 };
 
+struct io_mkdir {
+	struct file			*file;
+	int				dfd;
+	umode_t				mode;
+	struct filename			*filename;
+};
+
 struct io_completion {
 	struct file			*file;
 	struct list_head		list;
@@ -712,6 +719,7 @@ struct io_kiocb {
 		struct io_shutdown	shutdown;
 		struct io_rename	rename;
 		struct io_unlink	unlink;
+		struct io_mkdir		mkdir;
 		/* use only after cleaning per-op data, see io_clean_op() */
 		struct io_completion	compl;
 	};
@@ -996,6 +1004,10 @@ static const struct io_op_def io_op_defs[] = {
 		.work_flags		= IO_WQ_WORK_MM | IO_WQ_WORK_FILES |
 						IO_WQ_WORK_FS | IO_WQ_WORK_BLKCG,
 	},
+	[IORING_OP_MKDIRAT] = {
+		.work_flags		= IO_WQ_WORK_MM | IO_WQ_WORK_FILES |
+						IO_WQ_WORK_FS | IO_WQ_WORK_BLKCG,
+	},
 };
 
 enum io_mem_account {
@@ -3805,6 +3817,44 @@ static int io_unlinkat(struct io_kiocb *req, bool force_nonblock)
 	return 0;
 }
 
+static int io_mkdirat_prep(struct io_kiocb *req,
+			    const struct io_uring_sqe *sqe)
+{
+	struct io_mkdir *mkd = &req->mkdir;
+	const char __user *fname;
+
+	if (unlikely(req->flags & REQ_F_FIXED_FILE))
+		return -EBADF;
+
+	mkd->dfd = READ_ONCE(sqe->fd);
+	mkd->mode = READ_ONCE(sqe->len);
+
+	fname = u64_to_user_ptr(READ_ONCE(sqe->addr));
+	mkd->filename = getname(fname);
+	if (IS_ERR(mkd->filename))
+		return PTR_ERR(mkd->filename);
+
+	req->flags |= REQ_F_NEED_CLEANUP;
+	return 0;
+}
+
+static int io_mkdirat(struct io_kiocb *req, bool force_nonblock)
+{
+	struct io_mkdir *mkd = &req->mkdir;
+	int ret;
+
+	if (force_nonblock)
+		return -EAGAIN;
+
+	ret = do_mkdirat(mkd->dfd, mkd->filename, mkd->mode);
+
+	req->flags &= ~REQ_F_NEED_CLEANUP;
+	if (ret < 0)
+		req_set_fail_links(req);
+	io_req_complete(req, ret);
+	return 0;
+}
+
 static int io_shutdown_prep(struct io_kiocb *req,
 			    const struct io_uring_sqe *sqe)
 {
@@ -6101,6 +6151,8 @@ static int io_req_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe)
 		return io_renameat_prep(req, sqe);
 	case IORING_OP_UNLINKAT:
 		return io_unlinkat_prep(req, sqe);
+	case IORING_OP_MKDIRAT:
+		return io_mkdirat_prep(req, sqe);
 	}
 
 	printk_once(KERN_WARNING "io_uring: unhandled opcode %d\n",
@@ -6228,6 +6280,9 @@ static void __io_clean_op(struct io_kiocb *req)
 		case IORING_OP_UNLINKAT:
 			putname(req->unlink.filename);
 			break;
+		case IORING_OP_MKDIRAT:
+			putname(req->mkdir.filename);
+			break;
 		}
 		req->flags &= ~REQ_F_NEED_CLEANUP;
 	}
@@ -6340,6 +6395,9 @@ static int io_issue_sqe(struct io_kiocb *req, bool force_nonblock,
 	case IORING_OP_UNLINKAT:
 		ret = io_unlinkat(req, force_nonblock);
 		break;
+	case IORING_OP_MKDIRAT:
+		ret = io_mkdirat(req, force_nonblock);
+		break;
 	default:
 		ret = -EINVAL;
 		break;
diff --git a/include/uapi/linux/io_uring.h b/include/uapi/linux/io_uring.h
index ac4e1738a9af..890edd850a9e 100644
--- a/include/uapi/linux/io_uring.h
+++ b/include/uapi/linux/io_uring.h
@@ -137,6 +137,7 @@ enum {
 	IORING_OP_SHUTDOWN,
 	IORING_OP_RENAMEAT,
 	IORING_OP_UNLINKAT,
+	IORING_OP_MKDIRAT,
 
 	/* this goes last, obviously */
 	IORING_OP_LAST,
-- 
2.30.0

Re: [PATCH v2 0/2] io_uring: add mkdirat support

[Dmitry Kadashev]

On Tue, Feb 2, 2021 at 3:24 PM Dmitry Kadashev <dkadashev@gmail.com> wrote:
>
> This adds mkdirat support to io_uring and is heavily based on recently
> added renameat() / unlinkat() support.
>
> The first patch is preparation with no functional changes, makes
> do_mkdirat accept struct filename pointer rather than the user string.
>
> The second one leverages that to implement mkdirat in io_uring.
>
> Based on for-5.11/io_uring.
>
> Changes since v1:
> - do not mess with struct filename's refcount in do_mkdirat, instead add
>   and use __filename_create() that does not drop the name on success;
>
> Dmitry Kadashev (2):
>   fs: make do_mkdirat() take struct filename
>   io_uring: add support for IORING_OP_MKDIRAT
>
>  fs/internal.h                 |  1 +
>  fs/io_uring.c                 | 58 +++++++++++++++++++++++++++++++++++
>  fs/namei.c                    | 25 +++++++++++----
>  include/uapi/linux/io_uring.h |  1 +
>  4 files changed, 79 insertions(+), 6 deletions(-)

Hi Jens,

Ping. I've tried reaching out to Al wrt the first patch, but that did not seem
to work. Is there a chance to get this into 5.12 at this point?

-- 
Dmitry Kadashev

Re: [PATCH v2 0/2] io_uring: add mkdirat support

[Jens Axboe]

On 2/2/21 1:23 AM, Dmitry Kadashev wrote:
> This adds mkdirat support to io_uring and is heavily based on recently
> added renameat() / unlinkat() support.
> 
> The first patch is preparation with no functional changes, makes
> do_mkdirat accept struct filename pointer rather than the user string.
> 
> The second one leverages that to implement mkdirat in io_uring.
> 
> Based on for-5.11/io_uring.

Can you check if it still applies against for-5.13/io_uring? Both the
vfs and io_uring bits.

It'd be nice to get this moving forward, there's no reason why this
should keep getting stalled.

-- 
Jens Axboe

Re: [PATCH v2 0/2] io_uring: add mkdirat support

[Dmitry Kadashev]

On Tue, Mar 30, 2021 at 3:37 AM Jens Axboe <axboe@kernel.dk> wrote:
> On 2/2/21 1:23 AM, Dmitry Kadashev wrote:
> > Based on for-5.11/io_uring.

Actually this was a typo (copy-n-paste error), it was on top of
for-5.12/io_uring. Doesn't really matter though.

> Can you check if it still applies against for-5.13/io_uring? Both the
> vfs and io_uring bits.

It does not (the io_uring bits), I'll send v3 soon.

-- 
Dmitry Kadashev