[PATCH v8 00/11] io_uring: add mkdir and [sym]linkat support

[PATCH v8 00/11] io_uring: add mkdir and [sym]linkat support

[Dmitry Kadashev]

This started out as an attempt to add mkdirat support to io_uring which
is heavily based on renameat() / unlinkat() support.

During the review process more operations were added (linkat, symlinkat,
mknodat) mainly to keep things uniform internally (in namei.c), and
with things changed in namei.c adding support for these operations to
io_uring is trivial, so that was done too (except for mknodat). See
https://lore.kernel.org/io-uring/20210514145259.wtl4xcsp52woi6ab@wittgenstein/

The first patch makes putname() ignore IS_ERR_OR_NULL names and converts
a couple of places where the check was already implemented in the
callers. No functional changes.

The second one splits filename_lookup() that used to consume the passed
struct filename * on error but not on the success (returning it) into
two: filename_lookup(), that always consumes the name and
__filename_lookup() that never does. This is a preparation change to
enable the subsequent changes to filename_create and filename_lookup. No
functional changes.

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

The fourth one leverages that to implement mkdirat in io_uring.

5-8 just convert other similar do_* functions in namei.c to accept
struct filename, for uniformity with do_mkdirat, do_renameat and
do_unlinkat. No functional changes there.

9 changes do_* helpers in namei.c to return ints rather than some of
them returning ints and some longs.

10-11 add symlinkat and linkat support to io_uring correspondingly.

Based on for-5.14/io_uring.

v8:
- update filename_parentat() calling conventions to be uniform with the
  ones followed by (changed in subsequent patches) filename_create()
  and filename_lookup()

v7:
- rebase
- make putname() ignore IS_ERR_OR_NULL names, remove conditional calls
  to it from the callers

v6:

- rebase
- add safety checks for IOPOLL mode
- add safety checks for unused sqe parts
- drop mknodat support from io_uring as requested by Jens
- add Christian's Acked-by

v5:
- rebase
- add symlinkat, linkat and mknodat support to io_uring

v4:
- update do_mknodat, do_symlinkat and do_linkat to accept struct
  filename for uniformity with do_mkdirat, do_renameat and do_unlinkat;

v3:
- rebase;

v2:
- 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 (11):
  namei: ignore ERR/NULL names in putname()
  namei: change filename_parentat() calling conventions
  fs: make do_mkdirat() take struct filename
  io_uring: add support for IORING_OP_MKDIRAT
  fs: make do_mknodat() take struct filename
  fs: make do_symlinkat() take struct filename
  namei: add getname_uflags()
  fs: make do_linkat() take struct filename
  fs: update do_*() helpers to return ints
  io_uring: add support for IORING_OP_SYMLINKAT
  io_uring: add support for IORING_OP_LINKAT

 fs/exec.c                     |   8 +-
 fs/internal.h                 |   8 +-
 fs/io_uring.c                 | 196 ++++++++++++++++++++++++++++
 fs/namei.c                    | 239 +++++++++++++++++++---------------
 include/linux/fs.h            |   1 +
 include/uapi/linux/io_uring.h |   4 +
 6 files changed, 346 insertions(+), 110 deletions(-)

-- 
2.30.2

[PATCH v8 01/11] namei: ignore ERR/NULL names in putname()

[Dmitry Kadashev]

Supporting ERR/NULL names in putname() makes callers code cleaner, and
is what some other path walking functions already support for the same
reason.

This also removes a few existing IS_ERR checks before putname().

Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Link: https://lore.kernel.org/io-uring/CAHk-=wgCac9hBsYzKMpHk0EbLgQaXR=OUAjHaBtaY+G8A9KhFg@mail.gmail.com/
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Christian Brauner <christian.brauner@ubuntu.com>
Signed-off-by: Dmitry Kadashev <dkadashev@gmail.com>
---
 fs/namei.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

diff --git a/fs/namei.c b/fs/namei.c
index 79b0ff9b151e..70caf4ef1134 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -247,6 +247,9 @@ getname_kernel(const char * filename)
 
 void putname(struct filename *name)
 {
+	if (IS_ERR_OR_NULL(name))
+		return;
+
 	BUG_ON(name->refcnt <= 0);
 
 	if (--name->refcnt > 0)
@@ -4718,11 +4721,9 @@ int do_renameat2(int olddfd, struct filename *from, int newdfd,
 		goto retry;
 	}
 put_both:
-	if (!IS_ERR(from))
-		putname(from);
+	putname(from);
 put_new:
-	if (!IS_ERR(to))
-		putname(to);
+	putname(to);
 	return error;
 }
 
-- 
2.30.2

[PATCH v8 02/11] namei: change filename_parentat() calling conventions

[Dmitry Kadashev]

Since commit 5c31b6cedb675 ("namei: saner calling conventions for
filename_parentat()") filename_parentat() had the following behavior WRT
the passed in struct filename *:

* On error the name is consumed (putname() is called on it);
* On success the name is returned back as the return value;

Now there is a need for filename_create() and filename_lookup() variants
that do not consume the passed filename, and following the same "consume
the name only on error" semantics is proven to be hard to reason about
and result in confusing code.

Hence this preparation change splits filename_parentat() into two: one
that always consumes the name and another that never consumes the name.
This will allow to implement two filename_create() variants in the same
way, and is a consistent and hopefully easier to reason about approach.

Link: https://lore.kernel.org/io-uring/CAOKbgA7MiqZAq3t-HDCpSGUFfco4hMA9ArAE-74fTpU+EkvKPw@mail.gmail.com/
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Christian Brauner <christian.brauner@ubuntu.com>
Signed-off-by: Dmitry Kadashev <dkadashev@gmail.com>
---
 fs/namei.c | 108 ++++++++++++++++++++++++++---------------------------
 1 file changed, 53 insertions(+), 55 deletions(-)

diff --git a/fs/namei.c b/fs/namei.c
index 70caf4ef1134..2995b3695724 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -2485,7 +2485,7 @@ static int path_parentat(struct nameidata *nd, unsigned flags,
 	return err;
 }
 
-static struct filename *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)
 {
@@ -2493,7 +2493,7 @@ static struct filename *filename_parentat(int dfd, struct filename *name,
 	struct nameidata nd;
 
 	if (IS_ERR(name))
-		return name;
+		return PTR_ERR(name);
 	set_nameidata(&nd, dfd, name);
 	retval = path_parentat(&nd, flags | LOOKUP_RCU, parent);
 	if (unlikely(retval == -ECHILD))
@@ -2504,29 +2504,34 @@ static struct filename *filename_parentat(int dfd, struct filename *name,
 		*last = nd.last;
 		*type = nd.last_type;
 		audit_inode(name, parent->dentry, AUDIT_INODE_PARENT);
-	} else {
-		putname(name);
-		name = ERR_PTR(retval);
 	}
 	restore_nameidata();
-	return name;
+	return retval;
+}
+
+static int filename_parentat(int dfd, struct filename *name,
+				unsigned int flags, struct path *parent,
+				struct qstr *last, int *type)
+{
+	int retval = __filename_parentat(dfd, name, flags, parent, last, type);
+
+	putname(name);
+	return retval;
 }
 
 /* does lookup, returns the object with parent locked */
 struct dentry *kern_path_locked(const char *name, struct path *path)
 {
-	struct filename *filename;
 	struct dentry *d;
 	struct qstr last;
-	int type;
+	int type, error;
 
-	filename = filename_parentat(AT_FDCWD, getname_kernel(name), 0, path,
+	error = filename_parentat(AT_FDCWD, getname_kernel(name), 0, path,
 				    &last, &type);
-	if (IS_ERR(filename))
-		return ERR_CAST(filename);
+	if (error)
+		return ERR_PTR(error);
 	if (unlikely(type != LAST_NORM)) {
 		path_put(path);
-		putname(filename);
 		return ERR_PTR(-EINVAL);
 	}
 	inode_lock_nested(path->dentry->d_inode, I_MUTEX_PARENT);
@@ -2535,7 +2540,6 @@ struct dentry *kern_path_locked(const char *name, struct path *path)
 		inode_unlock(path->dentry->d_inode);
 		path_put(path);
 	}
-	putname(filename);
 	return d;
 }
 
@@ -3575,9 +3579,9 @@ static struct dentry *filename_create(int dfd, struct filename *name,
 	 */
 	lookup_flags &= LOOKUP_REVAL;
 
-	name = filename_parentat(dfd, name, lookup_flags, path, &last, &type);
-	if (IS_ERR(name))
-		return ERR_CAST(name);
+	error = filename_parentat(dfd, name, lookup_flags, path, &last, &type);
+	if (error)
+		return ERR_PTR(error);
 
 	/*
 	 * Yucky last component or no last component at all?
@@ -3615,7 +3619,6 @@ static struct dentry *filename_create(int dfd, struct filename *name,
 		error = err2;
 		goto fail;
 	}
-	putname(name);
 	return dentry;
 fail:
 	dput(dentry);
@@ -3626,7 +3629,6 @@ static struct dentry *filename_create(int dfd, struct filename *name,
 		mnt_drop_write(path->mnt);
 out:
 	path_put(path);
-	putname(name);
 	return dentry;
 }
 
@@ -3917,59 +3919,59 @@ EXPORT_SYMBOL(vfs_rmdir);
 long do_rmdir(int dfd, struct filename *name)
 {
 	struct user_namespace *mnt_userns;
-	int error = 0;
+	int error;
 	struct dentry *dentry;
 	struct path path;
 	struct qstr last;
 	int type;
 	unsigned int lookup_flags = 0;
 retry:
-	name = filename_parentat(dfd, name, lookup_flags,
-				&path, &last, &type);
-	if (IS_ERR(name))
-		return PTR_ERR(name);
+	error = __filename_parentat(dfd, name, lookup_flags, &path, &last, &type);
+	if (error)
+		goto exit1;
 
 	switch (type) {
 	case LAST_DOTDOT:
 		error = -ENOTEMPTY;
-		goto exit1;
+		goto exit2;
 	case LAST_DOT:
 		error = -EINVAL;
-		goto exit1;
+		goto exit2;
 	case LAST_ROOT:
 		error = -EBUSY;
-		goto exit1;
+		goto exit2;
 	}
 
 	error = mnt_want_write(path.mnt);
 	if (error)
-		goto exit1;
+		goto exit2;
 
 	inode_lock_nested(path.dentry->d_inode, I_MUTEX_PARENT);
 	dentry = __lookup_hash(&last, path.dentry, lookup_flags);
 	error = PTR_ERR(dentry);
 	if (IS_ERR(dentry))
-		goto exit2;
+		goto exit3;
 	if (!dentry->d_inode) {
 		error = -ENOENT;
-		goto exit3;
+		goto exit4;
 	}
 	error = security_path_rmdir(&path, dentry);
 	if (error)
-		goto exit3;
+		goto exit4;
 	mnt_userns = mnt_user_ns(path.mnt);
 	error = vfs_rmdir(mnt_userns, path.dentry->d_inode, dentry);
-exit3:
+exit4:
 	dput(dentry);
-exit2:
+exit3:
 	inode_unlock(path.dentry->d_inode);
 	mnt_drop_write(path.mnt);
-exit1:
+exit2:
 	path_put(&path);
 	if (retry_estale(error, lookup_flags)) {
 		lookup_flags |= LOOKUP_REVAL;
 		goto retry;
 	}
+exit1:
 	putname(name);
 	return error;
 }
@@ -4063,17 +4065,17 @@ long do_unlinkat(int dfd, struct filename *name)
 	struct inode *delegated_inode = NULL;
 	unsigned int lookup_flags = 0;
 retry:
-	name = filename_parentat(dfd, name, lookup_flags, &path, &last, &type);
-	if (IS_ERR(name))
-		return PTR_ERR(name);
+	error = __filename_parentat(dfd, name, lookup_flags, &path, &last, &type);
+	if (error)
+		goto exit1;
 
 	error = -EISDIR;
 	if (type != LAST_NORM)
-		goto exit1;
+		goto exit2;
 
 	error = mnt_want_write(path.mnt);
 	if (error)
-		goto exit1;
+		goto exit2;
 retry_deleg:
 	inode_lock_nested(path.dentry->d_inode, I_MUTEX_PARENT);
 	dentry = __lookup_hash(&last, path.dentry, lookup_flags);
@@ -4090,11 +4092,11 @@ long do_unlinkat(int dfd, struct filename *name)
 		ihold(inode);
 		error = security_path_unlink(&path, dentry);
 		if (error)
-			goto exit2;
+			goto exit3;
 		mnt_userns = mnt_user_ns(path.mnt);
 		error = vfs_unlink(mnt_userns, path.dentry->d_inode, dentry,
 				   &delegated_inode);
-exit2:
+exit3:
 		dput(dentry);
 	}
 	inode_unlock(path.dentry->d_inode);
@@ -4107,13 +4109,14 @@ long do_unlinkat(int dfd, struct filename *name)
 			goto retry_deleg;
 	}
 	mnt_drop_write(path.mnt);
-exit1:
+exit2:
 	path_put(&path);
 	if (retry_estale(error, lookup_flags)) {
 		lookup_flags |= LOOKUP_REVAL;
 		inode = NULL;
 		goto retry;
 	}
+exit1:
 	putname(name);
 	return error;
 
@@ -4124,7 +4127,7 @@ long do_unlinkat(int dfd, struct filename *name)
 		error = -EISDIR;
 	else
 		error = -ENOTDIR;
-	goto exit2;
+	goto exit3;
 }
 
 SYSCALL_DEFINE3(unlinkat, int, dfd, const char __user *, pathname, int, flag)
@@ -4595,29 +4598,25 @@ int do_renameat2(int olddfd, struct filename *from, int newdfd,
 	int error = -EINVAL;
 
 	if (flags & ~(RENAME_NOREPLACE | RENAME_EXCHANGE | RENAME_WHITEOUT))
-		goto put_both;
+		goto put_names;
 
 	if ((flags & (RENAME_NOREPLACE | RENAME_WHITEOUT)) &&
 	    (flags & RENAME_EXCHANGE))
-		goto put_both;
+		goto put_names;
 
 	if (flags & RENAME_EXCHANGE)
 		target_flags = 0;
 
 retry:
-	from = filename_parentat(olddfd, from, lookup_flags, &old_path,
+	error = __filename_parentat(olddfd, from, lookup_flags, &old_path,
 					&old_last, &old_type);
-	if (IS_ERR(from)) {
-		error = PTR_ERR(from);
-		goto put_new;
-	}
+	if (error)
+		goto put_names;
 
-	to = filename_parentat(newdfd, to, lookup_flags, &new_path, &new_last,
+	error = __filename_parentat(newdfd, to, lookup_flags, &new_path, &new_last,
 				&new_type);
-	if (IS_ERR(to)) {
-		error = PTR_ERR(to);
+	if (error)
 		goto exit1;
-	}
 
 	error = -EXDEV;
 	if (old_path.mnt != new_path.mnt)
@@ -4720,9 +4719,8 @@ int do_renameat2(int olddfd, struct filename *from, int newdfd,
 		lookup_flags |= LOOKUP_REVAL;
 		goto retry;
 	}
-put_both:
+put_names:
 	putname(from);
-put_new:
 	putname(to);
 	return error;
 }
-- 
2.30.2

Re: [PATCH v8 02/11] namei: change filename_parentat() calling conventions

[Linus Torvalds]

On Wed, Jul 7, 2021 at 5:28 AM Dmitry Kadashev <dkadashev@gmail.com> wrote:
>
> Hence this preparation change splits filename_parentat() into two: one
> that always consumes the name and another that never consumes the name.
> This will allow to implement two filename_create() variants in the same
> way, and is a consistent and hopefully easier to reason about approach.

I like it.

The patch itself is a bit hard to read, but the end result seems to make sense.

My main reaction is that this could have probably done a bit more
cleanup by avoiding some of the "goto exit1" kind of things.

Just as an example, now the rule is that "do_rmdir()" always does that

>         putname(name);
>         return error;

at the end, and I think this means that this whole function could be
split into a few trivial helper functions instead, and we'd have

   long do_rmdir(int dfd, struct filename *name)
  {
        int error;

        error = rmdir_helper(...)
        if (!retry_estale(error, lookup_flags)) {
                lookup_flags |= LOOKUP_REVAL;
                error = rmdir_helper(...);
        }
        putname(name);
        return error;
  }

which gets rid of both the "goto retry" and the "goto exit1".

With the meat of "do_rmdir()" done in that "rmdir_helper()" function.

I think the same is basically true of "do_unlinkat()" too.

But I wouldn't mind that cleanup as a separate patch. My point is that
I think this new rule for when the name is consumed is better and can
result in further cleanups.

(NOTE! This is from just reading the patch, I might have missed some case).

            Linus

[PATCH v8 03/11] 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>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Dmitry Kadashev <dkadashev@gmail.com>
Acked-by: Christian Brauner <christian.brauner@ubuntu.com>
---
 fs/internal.h |  1 +
 fs/namei.c    | 26 +++++++++++++++++++-------
 2 files changed, 20 insertions(+), 7 deletions(-)

diff --git a/fs/internal.h b/fs/internal.h
index 6aeae7ef3380..848e165ef0f1 100644
--- a/fs/internal.h
+++ b/fs/internal.h
@@ -77,6 +77,7 @@ long do_unlinkat(int dfd, struct filename *name);
 int may_linkat(struct user_namespace *mnt_userns, 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 2995b3695724..54d5f19ee1ce 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -3563,7 +3563,7 @@ 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);
@@ -3579,7 +3579,7 @@ static struct dentry *filename_create(int dfd, struct filename *name,
 	 */
 	lookup_flags &= LOOKUP_REVAL;
 
-	error = filename_parentat(dfd, name, lookup_flags, path, &last, &type);
+	error = __filename_parentat(dfd, name, lookup_flags, path, &last, &type);
 	if (error)
 		return ERR_PTR(error);
 
@@ -3632,6 +3632,15 @@ 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);
+
+	putname(name);
+	return res;
+}
+
 struct dentry *kern_path_create(int dfd, const char *pathname,
 				struct path *path, unsigned int lookup_flags)
 {
@@ -3822,7 +3831,7 @@ int vfs_mkdir(struct user_namespace *mnt_userns, struct inode *dir,
 }
 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;
@@ -3830,9 +3839,10 @@ static long do_mkdirat(int dfd, const char __user *pathname, umode_t mode)
 	unsigned int lookup_flags = LOOKUP_DIRECTORY;
 
 retry:
-	dentry = user_path_create(dfd, pathname, &path, lookup_flags);
+	dentry = __filename_create(dfd, name, &path, lookup_flags);
+	error = PTR_ERR(dentry);
 	if (IS_ERR(dentry))
-		return PTR_ERR(dentry);
+		goto out_putname;
 
 	if (!IS_POSIXACL(path.dentry->d_inode))
 		mode &= ~current_umask();
@@ -3848,17 +3858,19 @@ static long do_mkdirat(int dfd, const char __user *pathname, umode_t mode)
 		lookup_flags |= LOOKUP_REVAL;
 		goto retry;
 	}
+out_putname:
+	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);
 }
 
 /**
-- 
2.30.2

Re: [PATCH v8 03/11] fs: make do_mkdirat() take struct filename

[Linus Torvalds]

On Wed, Jul 7, 2021 at 5:28 AM 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().

.. and now I like this version too. And that do_mkdirat() could have
the same pattern with a "mkdirat_helper()" and avoiding the "goto
retry" and "goto out_putname" things, and be more understandable that
way.

Again, that could - and maybe should - be a separate cleanup.

             Linus

[PATCH v8 04/11] io_uring: add support for IORING_OP_MKDIRAT

[Dmitry Kadashev]

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

Cc: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Dmitry Kadashev <dkadashev@gmail.com>
Acked-by: Christian Brauner <christian.brauner@ubuntu.com>
---
 fs/io_uring.c                 | 59 +++++++++++++++++++++++++++++++++++
 include/uapi/linux/io_uring.h |  1 +
 2 files changed, 60 insertions(+)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index 5b840bb1e8ec..42d54f9bbbb2 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -674,6 +674,13 @@ 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;
@@ -831,6 +838,7 @@ 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;
 	};
@@ -1042,6 +1050,7 @@ static const struct io_op_def io_op_defs[] = {
 	},
 	[IORING_OP_RENAMEAT] = {},
 	[IORING_OP_UNLINKAT] = {},
+	[IORING_OP_MKDIRAT] = {},
 };
 
 static bool io_disarm_next(struct io_kiocb *req);
@@ -3545,6 +3554,48 @@ static int io_unlinkat(struct io_kiocb *req, unsigned int issue_flags)
 	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->ctx->flags & IORING_SETUP_IOPOLL))
+		return -EINVAL;
+	if (sqe->ioprio || sqe->off || sqe->rw_flags || sqe->buf_index)
+		return -EINVAL;
+	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, int issue_flags)
+{
+	struct io_mkdir *mkd = &req->mkdir;
+	int ret;
+
+	if (issue_flags & IO_URING_F_NONBLOCK)
+		return -EAGAIN;
+
+	ret = do_mkdirat(mkd->dfd, mkd->filename, mkd->mode);
+
+	req->flags &= ~REQ_F_NEED_CLEANUP;
+	if (ret < 0)
+		req_set_fail(req);
+	io_req_complete(req, ret);
+	return 0;
+}
+
 static int io_shutdown_prep(struct io_kiocb *req,
 			    const struct io_uring_sqe *sqe)
 {
@@ -5953,6 +6004,8 @@ 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",
@@ -6114,6 +6167,9 @@ 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;
 		}
 	}
 	if ((req->flags & REQ_F_POLLED) && req->apoll) {
@@ -6242,6 +6298,9 @@ static int io_issue_sqe(struct io_kiocb *req, unsigned int issue_flags)
 	case IORING_OP_UNLINKAT:
 		ret = io_unlinkat(req, issue_flags);
 		break;
+	case IORING_OP_MKDIRAT:
+		ret = io_mkdirat(req, issue_flags);
+		break;
 	default:
 		ret = -EINVAL;
 		break;
diff --git a/include/uapi/linux/io_uring.h b/include/uapi/linux/io_uring.h
index 79126d5cd289..a926407c230e 100644
--- a/include/uapi/linux/io_uring.h
+++ b/include/uapi/linux/io_uring.h
@@ -133,6 +133,7 @@ enum {
 	IORING_OP_SHUTDOWN,
 	IORING_OP_RENAMEAT,
 	IORING_OP_UNLINKAT,
+	IORING_OP_MKDIRAT,
 
 	/* this goes last, obviously */
 	IORING_OP_LAST,
-- 
2.30.2

[PATCH v8 05/11] fs: make do_mknodat() take struct filename

[Dmitry Kadashev]

Pass in the struct filename pointers instead of the user string, for
uniformity with the recently converted do_unlinkat(), do_renameat(),
do_mkdirat().

Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Christian Brauner <christian.brauner@ubuntu.com>
Link: https://lore.kernel.org/io-uring/20210330071700.kpjoyp5zlni7uejm@wittgenstein/
Signed-off-by: Dmitry Kadashev <dkadashev@gmail.com>
Acked-by: Christian Brauner <christian.brauner@ubuntu.com>
---
 fs/namei.c | 19 +++++++++++--------
 1 file changed, 11 insertions(+), 8 deletions(-)

diff --git a/fs/namei.c b/fs/namei.c
index 54d5f19ee1ce..0bc8ff637934 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -3729,7 +3729,7 @@ static int may_mknod(umode_t mode)
 	}
 }
 
-static long do_mknodat(int dfd, const char __user *filename, umode_t mode,
+static long do_mknodat(int dfd, struct filename *name, umode_t mode,
 		unsigned int dev)
 {
 	struct user_namespace *mnt_userns;
@@ -3740,17 +3740,18 @@ static long do_mknodat(int dfd, const char __user *filename, umode_t mode,
 
 	error = may_mknod(mode);
 	if (error)
-		return error;
+		goto out1;
 retry:
-	dentry = user_path_create(dfd, filename, &path, lookup_flags);
+	dentry = __filename_create(dfd, name, &path, lookup_flags);
+	error = PTR_ERR(dentry);
 	if (IS_ERR(dentry))
-		return PTR_ERR(dentry);
+		goto out1;
 
 	if (!IS_POSIXACL(path.dentry->d_inode))
 		mode &= ~current_umask();
 	error = security_path_mknod(&path, dentry, mode, dev);
 	if (error)
-		goto out;
+		goto out2;
 
 	mnt_userns = mnt_user_ns(path.mnt);
 	switch (mode & S_IFMT) {
@@ -3769,24 +3770,26 @@ static long do_mknodat(int dfd, const char __user *filename, umode_t mode,
 					  dentry, mode, 0);
 			break;
 	}
-out:
+out2:
 	done_path_create(&path, dentry);
 	if (retry_estale(error, lookup_flags)) {
 		lookup_flags |= LOOKUP_REVAL;
 		goto retry;
 	}
+out1:
+	putname(name);
 	return error;
 }
 
 SYSCALL_DEFINE4(mknodat, int, dfd, const char __user *, filename, umode_t, mode,
 		unsigned int, dev)
 {
-	return do_mknodat(dfd, filename, mode, dev);
+	return do_mknodat(dfd, getname(filename), mode, dev);
 }
 
 SYSCALL_DEFINE3(mknod, const char __user *, filename, umode_t, mode, unsigned, dev)
 {
-	return do_mknodat(AT_FDCWD, filename, mode, dev);
+	return do_mknodat(AT_FDCWD, getname(filename), mode, dev);
 }
 
 /**
-- 
2.30.2

Re: [PATCH v8 05/11] fs: make do_mknodat() take struct filename

[Linus Torvalds]

On Wed, Jul 7, 2021 at 5:28 AM Dmitry Kadashev <dkadashev@gmail.com> wrote:
>
> Pass in the struct filename pointers instead of the user string, for
> uniformity with the recently converted do_unlinkat(), do_renameat(),
> do_mkdirat().

Yup. And the uniformity continues with that "we could avoid the goto
retry/out1 with a mknodat_helper() function for the inner meat of the
function.

               Linus

[PATCH v8 06/11] fs: make do_symlinkat() take struct filename

[Dmitry Kadashev]

Pass in the struct filename pointers instead of the user string, for
uniformity with the recently converted do_mkdnodat(), do_unlinkat(),
do_renameat(), do_mkdirat().

Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Christian Brauner <christian.brauner@ubuntu.com>
Link: https://lore.kernel.org/io-uring/20210330071700.kpjoyp5zlni7uejm@wittgenstein/
Signed-off-by: Dmitry Kadashev <dkadashev@gmail.com>
Acked-by: Christian Brauner <christian.brauner@ubuntu.com>
---
 fs/namei.c | 23 ++++++++++++-----------
 1 file changed, 12 insertions(+), 11 deletions(-)

diff --git a/fs/namei.c b/fs/namei.c
index 0bc8ff637934..add984e4bfd0 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -4197,23 +4197,23 @@ int vfs_symlink(struct user_namespace *mnt_userns, struct inode *dir,
 }
 EXPORT_SYMBOL(vfs_symlink);
 
-static long do_symlinkat(const char __user *oldname, int newdfd,
-		  const char __user *newname)
+static long do_symlinkat(struct filename *from, int newdfd,
+		  struct filename *to)
 {
 	int error;
-	struct filename *from;
 	struct dentry *dentry;
 	struct path path;
 	unsigned int lookup_flags = 0;
 
-	from = getname(oldname);
-	if (IS_ERR(from))
-		return PTR_ERR(from);
+	if (IS_ERR(from)) {
+		error = PTR_ERR(from);
+		goto out_putnames;
+	}
 retry:
-	dentry = user_path_create(newdfd, newname, &path, lookup_flags);
+	dentry = __filename_create(newdfd, to, &path, lookup_flags);
 	error = PTR_ERR(dentry);
 	if (IS_ERR(dentry))
-		goto out_putname;
+		goto out_putnames;
 
 	error = security_path_symlink(&path, dentry, from->name);
 	if (!error) {
@@ -4228,7 +4228,8 @@ static long do_symlinkat(const char __user *oldname, int newdfd,
 		lookup_flags |= LOOKUP_REVAL;
 		goto retry;
 	}
-out_putname:
+out_putnames:
+	putname(to);
 	putname(from);
 	return error;
 }
@@ -4236,12 +4237,12 @@ static long do_symlinkat(const char __user *oldname, int newdfd,
 SYSCALL_DEFINE3(symlinkat, const char __user *, oldname,
 		int, newdfd, const char __user *, newname)
 {
-	return do_symlinkat(oldname, newdfd, newname);
+	return do_symlinkat(getname(oldname), newdfd, getname(newname));
 }
 
 SYSCALL_DEFINE2(symlink, const char __user *, oldname, const char __user *, newname)
 {
-	return do_symlinkat(oldname, AT_FDCWD, newname);
+	return do_symlinkat(getname(oldname), AT_FDCWD, getname(newname));
 }
 
 /**
-- 
2.30.2

[PATCH v8 07/11] namei: add getname_uflags()

[Dmitry Kadashev]

There are a couple of places where we already open-code the (flags &
AT_EMPTY_PATH) check and io_uring will likely add another one in the
future.  Let's just add a simple helper getname_uflags() that handles
this directly and use it.

Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Christian Brauner <christian.brauner@ubuntu.com>
Link: https://lore.kernel.org/io-uring/20210415100815.edrn4a7cy26wkowe@wittgenstein/
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
Signed-off-by: Dmitry Kadashev <dkadashev@gmail.com>
Acked-by: Christian Brauner <christian.brauner@ubuntu.com>
---
 fs/exec.c          | 8 ++------
 fs/namei.c         | 8 ++++++++
 include/linux/fs.h | 1 +
 3 files changed, 11 insertions(+), 6 deletions(-)

diff --git a/fs/exec.c b/fs/exec.c
index 18594f11c31f..df33ecaf2111 100644
--- a/fs/exec.c
+++ b/fs/exec.c
@@ -2069,10 +2069,8 @@ SYSCALL_DEFINE5(execveat,
 		const char __user *const __user *, envp,
 		int, flags)
 {
-	int lookup_flags = (flags & AT_EMPTY_PATH) ? LOOKUP_EMPTY : 0;
-
 	return do_execveat(fd,
-			   getname_flags(filename, lookup_flags, NULL),
+			   getname_uflags(filename, flags),
 			   argv, envp, flags);
 }
 
@@ -2090,10 +2088,8 @@ COMPAT_SYSCALL_DEFINE5(execveat, int, fd,
 		       const compat_uptr_t __user *, envp,
 		       int,  flags)
 {
-	int lookup_flags = (flags & AT_EMPTY_PATH) ? LOOKUP_EMPTY : 0;
-
 	return compat_do_execveat(fd,
-				  getname_flags(filename, lookup_flags, NULL),
+				  getname_uflags(filename, flags),
 				  argv, envp, flags);
 }
 #endif
diff --git a/fs/namei.c b/fs/namei.c
index add984e4bfd0..ab7979f9daaa 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -203,6 +203,14 @@ getname_flags(const char __user *filename, int flags, int *empty)
 	return result;
 }
 
+struct filename *
+getname_uflags(const char __user *filename, int uflags)
+{
+	int flags = (uflags & AT_EMPTY_PATH) ? LOOKUP_EMPTY : 0;
+
+	return getname_flags(filename, flags, NULL);
+}
+
 struct filename *
 getname(const char __user * filename)
 {
diff --git a/include/linux/fs.h b/include/linux/fs.h
index c3c88fdb9b2a..5885a68d2c12 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -2781,6 +2781,7 @@ static inline struct file *file_clone_open(struct file *file)
 extern int filp_close(struct file *, fl_owner_t id);
 
 extern struct filename *getname_flags(const char __user *, int, int *);
+extern struct filename *getname_uflags(const char __user *, int);
 extern struct filename *getname(const char __user *);
 extern struct filename *getname_kernel(const char *);
 extern void putname(struct filename *name);
-- 
2.30.2

[PATCH v8 08/11] fs: make do_linkat() take struct filename

[Dmitry Kadashev]

Pass in the struct filename pointers instead of the user string, for
uniformity with do_renameat2, do_unlinkat, do_mknodat, etc.

Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Christian Brauner <christian.brauner@ubuntu.com>
Link: https://lore.kernel.org/io-uring/20210330071700.kpjoyp5zlni7uejm@wittgenstein/
Signed-off-by: Dmitry Kadashev <dkadashev@gmail.com>
Acked-by: Christian Brauner <christian.brauner@ubuntu.com>
---
 fs/namei.c | 45 +++++++++++++++++++++++++++++----------------
 1 file changed, 29 insertions(+), 16 deletions(-)

diff --git a/fs/namei.c b/fs/namei.c
index ab7979f9daaa..c4e13bd8652f 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -2450,7 +2450,7 @@ static int path_lookupat(struct nameidata *nd, unsigned flags, struct path *path
 	return err;
 }
 
-int filename_lookup(int dfd, struct filename *name, unsigned flags,
+static int __filename_lookup(int dfd, struct filename *name, unsigned flags,
 		    struct path *path, struct path *root)
 {
 	int retval;
@@ -2472,6 +2472,14 @@ int filename_lookup(int dfd, struct filename *name, unsigned flags,
 		audit_inode(name, path->dentry,
 			    flags & LOOKUP_MOUNTPOINT ? AUDIT_INODE_NOEVAL : 0);
 	restore_nameidata();
+	return retval;
+}
+
+int filename_lookup(int dfd, struct filename *name, unsigned flags,
+		    struct path *path, struct path *root)
+{
+	int retval = __filename_lookup(dfd, name, flags, path, root);
+
 	putname(name);
 	return retval;
 }
@@ -4351,8 +4359,8 @@ EXPORT_SYMBOL(vfs_link);
  * with linux 2.0, and to avoid hard-linking to directories
  * and other special files.  --ADM
  */
-static int do_linkat(int olddfd, const char __user *oldname, int newdfd,
-	      const char __user *newname, int flags)
+static int do_linkat(int olddfd, struct filename *old, int newdfd,
+	      struct filename *new, int flags)
 {
 	struct user_namespace *mnt_userns;
 	struct dentry *new_dentry;
@@ -4361,31 +4369,32 @@ static int do_linkat(int olddfd, const char __user *oldname, int newdfd,
 	int how = 0;
 	int error;
 
-	if ((flags & ~(AT_SYMLINK_FOLLOW | AT_EMPTY_PATH)) != 0)
-		return -EINVAL;
+	if ((flags & ~(AT_SYMLINK_FOLLOW | AT_EMPTY_PATH)) != 0) {
+		error = -EINVAL;
+		goto out_putnames;
+	}
 	/*
 	 * To use null names we require CAP_DAC_READ_SEARCH
 	 * This ensures that not everyone will be able to create
 	 * handlink using the passed filedescriptor.
 	 */
-	if (flags & AT_EMPTY_PATH) {
-		if (!capable(CAP_DAC_READ_SEARCH))
-			return -ENOENT;
-		how = LOOKUP_EMPTY;
+	if (flags & AT_EMPTY_PATH && !capable(CAP_DAC_READ_SEARCH)) {
+		error = -ENOENT;
+		goto out_putnames;
 	}
 
 	if (flags & AT_SYMLINK_FOLLOW)
 		how |= LOOKUP_FOLLOW;
 retry:
-	error = user_path_at(olddfd, oldname, how, &old_path);
+	error = __filename_lookup(olddfd, old, how, &old_path, NULL);
 	if (error)
-		return error;
+		goto out_putnames;
 
-	new_dentry = user_path_create(newdfd, newname, &new_path,
+	new_dentry = __filename_create(newdfd, new, &new_path,
 					(how & LOOKUP_REVAL));
 	error = PTR_ERR(new_dentry);
 	if (IS_ERR(new_dentry))
-		goto out;
+		goto out_putpath;
 
 	error = -EXDEV;
 	if (old_path.mnt != new_path.mnt)
@@ -4413,8 +4422,11 @@ static int do_linkat(int olddfd, const char __user *oldname, int newdfd,
 		how |= LOOKUP_REVAL;
 		goto retry;
 	}
-out:
+out_putpath:
 	path_put(&old_path);
+out_putnames:
+	putname(old);
+	putname(new);
 
 	return error;
 }
@@ -4422,12 +4434,13 @@ static int do_linkat(int olddfd, const char __user *oldname, int newdfd,
 SYSCALL_DEFINE5(linkat, int, olddfd, const char __user *, oldname,
 		int, newdfd, const char __user *, newname, int, flags)
 {
-	return do_linkat(olddfd, oldname, newdfd, newname, flags);
+	return do_linkat(olddfd, getname_uflags(oldname, flags),
+		newdfd, getname(newname), flags);
 }
 
 SYSCALL_DEFINE2(link, const char __user *, oldname, const char __user *, newname)
 {
-	return do_linkat(AT_FDCWD, oldname, AT_FDCWD, newname, 0);
+	return do_linkat(AT_FDCWD, getname(oldname), AT_FDCWD, getname(newname), 0);
 }
 
 /**
-- 
2.30.2

[PATCH v8 09/11] fs: update do_*() helpers to return ints

[Dmitry Kadashev]

Update the following to return int rather than long, for uniformity with
the rest of the do_* helpers in namei.c:

* do_rmdir()
* do_unlinkat()
* do_mkdirat()
* do_mknodat()
* do_symlinkat()

Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Christian Brauner <christian.brauner@ubuntu.com>
Link: https://lore.kernel.org/io-uring/20210514143202.dmzfcgz5hnauy7ze@wittgenstein/
Signed-off-by: Dmitry Kadashev <dkadashev@gmail.com>
Acked-by: Christian Brauner <christian.brauner@ubuntu.com>
---
 fs/internal.h |  6 +++---
 fs/namei.c    | 10 +++++-----
 2 files changed, 8 insertions(+), 8 deletions(-)

diff --git a/fs/internal.h b/fs/internal.h
index 848e165ef0f1..207a455e32d3 100644
--- a/fs/internal.h
+++ b/fs/internal.h
@@ -72,12 +72,12 @@ extern int filename_lookup(int dfd, struct filename *name, unsigned flags,
 			   struct path *path, struct path *root);
 extern int vfs_path_lookup(struct dentry *, struct vfsmount *,
 			   const char *, unsigned int, struct path *);
-long do_rmdir(int dfd, struct filename *name);
-long do_unlinkat(int dfd, struct filename *name);
+int do_rmdir(int dfd, struct filename *name);
+int do_unlinkat(int dfd, struct filename *name);
 int may_linkat(struct user_namespace *mnt_userns, 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);
+int do_mkdirat(int dfd, struct filename *name, umode_t mode);
 
 /*
  * namespace.c
diff --git a/fs/namei.c b/fs/namei.c
index c4e13bd8652f..d06aeaf5da00 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -3745,7 +3745,7 @@ static int may_mknod(umode_t mode)
 	}
 }
 
-static long do_mknodat(int dfd, struct filename *name, umode_t mode,
+static int do_mknodat(int dfd, struct filename *name, umode_t mode,
 		unsigned int dev)
 {
 	struct user_namespace *mnt_userns;
@@ -3850,7 +3850,7 @@ int vfs_mkdir(struct user_namespace *mnt_userns, struct inode *dir,
 }
 EXPORT_SYMBOL(vfs_mkdir);
 
-long do_mkdirat(int dfd, struct filename *name, umode_t mode)
+int do_mkdirat(int dfd, struct filename *name, umode_t mode)
 {
 	struct dentry *dentry;
 	struct path path;
@@ -3947,7 +3947,7 @@ int vfs_rmdir(struct user_namespace *mnt_userns, struct inode *dir,
 }
 EXPORT_SYMBOL(vfs_rmdir);
 
-long do_rmdir(int dfd, struct filename *name)
+int do_rmdir(int dfd, struct filename *name)
 {
 	struct user_namespace *mnt_userns;
 	int error;
@@ -4085,7 +4085,7 @@ EXPORT_SYMBOL(vfs_unlink);
  * writeout happening, and we don't want to prevent access to the directory
  * while waiting on the I/O.
  */
-long do_unlinkat(int dfd, struct filename *name)
+int do_unlinkat(int dfd, struct filename *name)
 {
 	int error;
 	struct dentry *dentry;
@@ -4213,7 +4213,7 @@ int vfs_symlink(struct user_namespace *mnt_userns, struct inode *dir,
 }
 EXPORT_SYMBOL(vfs_symlink);
 
-static long do_symlinkat(struct filename *from, int newdfd,
+static int do_symlinkat(struct filename *from, int newdfd,
 		  struct filename *to)
 {
 	int error;
-- 
2.30.2

[PATCH v8 10/11] io_uring: add support for IORING_OP_SYMLINKAT

[Dmitry Kadashev]

IORING_OP_SYMLINKAT behaves like symlinkat(2) and takes the same flags
and arguments.

Cc: Linus Torvalds <torvalds@linux-foundation.org>
Suggested-by: Christian Brauner <christian.brauner@ubuntu.com>
Link: https://lore.kernel.org/io-uring/20210514145259.wtl4xcsp52woi6ab@wittgenstein/
Signed-off-by: Dmitry Kadashev <dkadashev@gmail.com>
Acked-by: Christian Brauner <christian.brauner@ubuntu.com>
---
 fs/internal.h                 |  1 +
 fs/io_uring.c                 | 66 +++++++++++++++++++++++++++++++++++
 fs/namei.c                    |  3 +-
 include/uapi/linux/io_uring.h |  1 +
 4 files changed, 69 insertions(+), 2 deletions(-)

diff --git a/fs/internal.h b/fs/internal.h
index 207a455e32d3..3b3954214385 100644
--- a/fs/internal.h
+++ b/fs/internal.h
@@ -78,6 +78,7 @@ int may_linkat(struct user_namespace *mnt_userns, struct path *link);
 int do_renameat2(int olddfd, struct filename *oldname, int newdfd,
 		 struct filename *newname, unsigned int flags);
 int do_mkdirat(int dfd, struct filename *name, umode_t mode);
+int do_symlinkat(struct filename *from, int newdfd, struct filename *to);
 
 /*
  * namespace.c
diff --git a/fs/io_uring.c b/fs/io_uring.c
index 42d54f9bbbb2..a0f681ec25bb 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -681,6 +681,13 @@ struct io_mkdir {
 	struct filename			*filename;
 };
 
+struct io_symlink {
+	struct file			*file;
+	int				new_dfd;
+	struct filename			*oldpath;
+	struct filename			*newpath;
+};
+
 struct io_completion {
 	struct file			*file;
 	struct list_head		list;
@@ -839,6 +846,7 @@ struct io_kiocb {
 		struct io_rename	rename;
 		struct io_unlink	unlink;
 		struct io_mkdir		mkdir;
+		struct io_symlink	symlink;
 		/* use only after cleaning per-op data, see io_clean_op() */
 		struct io_completion	compl;
 	};
@@ -1051,6 +1059,7 @@ static const struct io_op_def io_op_defs[] = {
 	[IORING_OP_RENAMEAT] = {},
 	[IORING_OP_UNLINKAT] = {},
 	[IORING_OP_MKDIRAT] = {},
+	[IORING_OP_SYMLINKAT] = {},
 };
 
 static bool io_disarm_next(struct io_kiocb *req);
@@ -3596,6 +3605,54 @@ static int io_mkdirat(struct io_kiocb *req, int issue_flags)
 	return 0;
 }
 
+static int io_symlinkat_prep(struct io_kiocb *req,
+			    const struct io_uring_sqe *sqe)
+{
+	struct io_symlink *sl = &req->symlink;
+	const char __user *oldpath, *newpath;
+
+	if (unlikely(req->ctx->flags & IORING_SETUP_IOPOLL))
+		return -EINVAL;
+	if (sqe->ioprio || sqe->len || sqe->rw_flags || sqe->buf_index)
+		return -EINVAL;
+	if (unlikely(req->flags & REQ_F_FIXED_FILE))
+		return -EBADF;
+
+	sl->new_dfd = READ_ONCE(sqe->fd);
+	oldpath = u64_to_user_ptr(READ_ONCE(sqe->addr));
+	newpath = u64_to_user_ptr(READ_ONCE(sqe->addr2));
+
+	sl->oldpath = getname(oldpath);
+	if (IS_ERR(sl->oldpath))
+		return PTR_ERR(sl->oldpath);
+
+	sl->newpath = getname(newpath);
+	if (IS_ERR(sl->newpath)) {
+		putname(sl->oldpath);
+		return PTR_ERR(sl->newpath);
+	}
+
+	req->flags |= REQ_F_NEED_CLEANUP;
+	return 0;
+}
+
+static int io_symlinkat(struct io_kiocb *req, int issue_flags)
+{
+	struct io_symlink *sl = &req->symlink;
+	int ret;
+
+	if (issue_flags & IO_URING_F_NONBLOCK)
+		return -EAGAIN;
+
+	ret = do_symlinkat(sl->oldpath, sl->new_dfd, sl->newpath);
+
+	req->flags &= ~REQ_F_NEED_CLEANUP;
+	if (ret < 0)
+		req_set_fail(req);
+	io_req_complete(req, ret);
+	return 0;
+}
+
 static int io_shutdown_prep(struct io_kiocb *req,
 			    const struct io_uring_sqe *sqe)
 {
@@ -6006,6 +6063,8 @@ static int io_req_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe)
 		return io_unlinkat_prep(req, sqe);
 	case IORING_OP_MKDIRAT:
 		return io_mkdirat_prep(req, sqe);
+	case IORING_OP_SYMLINKAT:
+		return io_symlinkat_prep(req, sqe);
 	}
 
 	printk_once(KERN_WARNING "io_uring: unhandled opcode %d\n",
@@ -6170,6 +6229,10 @@ static void io_clean_op(struct io_kiocb *req)
 		case IORING_OP_MKDIRAT:
 			putname(req->mkdir.filename);
 			break;
+		case IORING_OP_SYMLINKAT:
+			putname(req->symlink.oldpath);
+			putname(req->symlink.newpath);
+			break;
 		}
 	}
 	if ((req->flags & REQ_F_POLLED) && req->apoll) {
@@ -6301,6 +6364,9 @@ static int io_issue_sqe(struct io_kiocb *req, unsigned int issue_flags)
 	case IORING_OP_MKDIRAT:
 		ret = io_mkdirat(req, issue_flags);
 		break;
+	case IORING_OP_SYMLINKAT:
+		ret = io_symlinkat(req, issue_flags);
+		break;
 	default:
 		ret = -EINVAL;
 		break;
diff --git a/fs/namei.c b/fs/namei.c
index d06aeaf5da00..f241348e64f4 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -4213,8 +4213,7 @@ int vfs_symlink(struct user_namespace *mnt_userns, struct inode *dir,
 }
 EXPORT_SYMBOL(vfs_symlink);
 
-static int do_symlinkat(struct filename *from, int newdfd,
-		  struct filename *to)
+int do_symlinkat(struct filename *from, int newdfd, struct filename *to)
 {
 	int error;
 	struct dentry *dentry;
diff --git a/include/uapi/linux/io_uring.h b/include/uapi/linux/io_uring.h
index a926407c230e..61fd347ab176 100644
--- a/include/uapi/linux/io_uring.h
+++ b/include/uapi/linux/io_uring.h
@@ -134,6 +134,7 @@ enum {
 	IORING_OP_RENAMEAT,
 	IORING_OP_UNLINKAT,
 	IORING_OP_MKDIRAT,
+	IORING_OP_SYMLINKAT,
 
 	/* this goes last, obviously */
 	IORING_OP_LAST,
-- 
2.30.2

[PATCH v8 11/11] io_uring: add support for IORING_OP_LINKAT

[Dmitry Kadashev]

IORING_OP_LINKAT behaves like linkat(2) and takes the same flags and
arguments.

In some internal places 'hardlink' is used instead of 'link' to avoid
confusion with the SQE links. Name 'link' conflicts with the existing
'link' member of io_kiocb.

Cc: Linus Torvalds <torvalds@linux-foundation.org>
Suggested-by: Christian Brauner <christian.brauner@ubuntu.com>
Link: https://lore.kernel.org/io-uring/20210514145259.wtl4xcsp52woi6ab@wittgenstein/
Signed-off-by: Dmitry Kadashev <dkadashev@gmail.com>
Acked-by: Christian Brauner <christian.brauner@ubuntu.com>
---
 fs/internal.h                 |  2 +
 fs/io_uring.c                 | 71 +++++++++++++++++++++++++++++++++++
 fs/namei.c                    |  2 +-
 include/uapi/linux/io_uring.h |  2 +
 4 files changed, 76 insertions(+), 1 deletion(-)

diff --git a/fs/internal.h b/fs/internal.h
index 3b3954214385..15a7d210cc67 100644
--- a/fs/internal.h
+++ b/fs/internal.h
@@ -79,6 +79,8 @@ int do_renameat2(int olddfd, struct filename *oldname, int newdfd,
 		 struct filename *newname, unsigned int flags);
 int do_mkdirat(int dfd, struct filename *name, umode_t mode);
 int do_symlinkat(struct filename *from, int newdfd, struct filename *to);
+int do_linkat(int olddfd, struct filename *old, int newdfd,
+			struct filename *new, int flags);
 
 /*
  * namespace.c
diff --git a/fs/io_uring.c b/fs/io_uring.c
index a0f681ec25bb..d18ca8afd1fb 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -688,6 +688,15 @@ struct io_symlink {
 	struct filename			*newpath;
 };
 
+struct io_hardlink {
+	struct file			*file;
+	int				old_dfd;
+	int				new_dfd;
+	struct filename			*oldpath;
+	struct filename			*newpath;
+	int				flags;
+};
+
 struct io_completion {
 	struct file			*file;
 	struct list_head		list;
@@ -847,6 +856,7 @@ struct io_kiocb {
 		struct io_unlink	unlink;
 		struct io_mkdir		mkdir;
 		struct io_symlink	symlink;
+		struct io_hardlink	hardlink;
 		/* use only after cleaning per-op data, see io_clean_op() */
 		struct io_completion	compl;
 	};
@@ -1060,6 +1070,7 @@ static const struct io_op_def io_op_defs[] = {
 	[IORING_OP_UNLINKAT] = {},
 	[IORING_OP_MKDIRAT] = {},
 	[IORING_OP_SYMLINKAT] = {},
+	[IORING_OP_LINKAT] = {},
 };
 
 static bool io_disarm_next(struct io_kiocb *req);
@@ -3653,6 +3664,57 @@ static int io_symlinkat(struct io_kiocb *req, int issue_flags)
 	return 0;
 }
 
+static int io_linkat_prep(struct io_kiocb *req,
+			    const struct io_uring_sqe *sqe)
+{
+	struct io_hardlink *lnk = &req->hardlink;
+	const char __user *oldf, *newf;
+
+	if (unlikely(req->ctx->flags & IORING_SETUP_IOPOLL))
+		return -EINVAL;
+	if (sqe->ioprio || sqe->rw_flags || sqe->buf_index)
+		return -EINVAL;
+	if (unlikely(req->flags & REQ_F_FIXED_FILE))
+		return -EBADF;
+
+	lnk->old_dfd = READ_ONCE(sqe->fd);
+	lnk->new_dfd = READ_ONCE(sqe->len);
+	oldf = u64_to_user_ptr(READ_ONCE(sqe->addr));
+	newf = u64_to_user_ptr(READ_ONCE(sqe->addr2));
+	lnk->flags = READ_ONCE(sqe->hardlink_flags);
+
+	lnk->oldpath = getname(oldf);
+	if (IS_ERR(lnk->oldpath))
+		return PTR_ERR(lnk->oldpath);
+
+	lnk->newpath = getname(newf);
+	if (IS_ERR(lnk->newpath)) {
+		putname(lnk->oldpath);
+		return PTR_ERR(lnk->newpath);
+	}
+
+	req->flags |= REQ_F_NEED_CLEANUP;
+	return 0;
+}
+
+static int io_linkat(struct io_kiocb *req, int issue_flags)
+{
+	struct io_hardlink *lnk = &req->hardlink;
+	int ret;
+
+	if (issue_flags & IO_URING_F_NONBLOCK)
+		return -EAGAIN;
+
+	ret = do_linkat(lnk->old_dfd, lnk->oldpath, lnk->new_dfd,
+				lnk->newpath, lnk->flags);
+
+	req->flags &= ~REQ_F_NEED_CLEANUP;
+	if (ret < 0)
+		req_set_fail(req);
+	io_req_complete(req, ret);
+	return 0;
+}
+
 static int io_shutdown_prep(struct io_kiocb *req,
 			    const struct io_uring_sqe *sqe)
 {
@@ -6065,6 +6127,8 @@ static int io_req_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe)
 		return io_mkdirat_prep(req, sqe);
 	case IORING_OP_SYMLINKAT:
 		return io_symlinkat_prep(req, sqe);
+	case IORING_OP_LINKAT:
+		return io_linkat_prep(req, sqe);
 	}
 
 	printk_once(KERN_WARNING "io_uring: unhandled opcode %d\n",
@@ -6233,6 +6297,10 @@ static void io_clean_op(struct io_kiocb *req)
 			putname(req->symlink.oldpath);
 			putname(req->symlink.newpath);
 			break;
+		case IORING_OP_LINKAT:
+			putname(req->hardlink.oldpath);
+			putname(req->hardlink.newpath);
+			break;
 		}
 	}
 	if ((req->flags & REQ_F_POLLED) && req->apoll) {
@@ -6367,6 +6435,9 @@ static int io_issue_sqe(struct io_kiocb *req, unsigned int issue_flags)
 	case IORING_OP_SYMLINKAT:
 		ret = io_symlinkat(req, issue_flags);
 		break;
+	case IORING_OP_LINKAT:
+		ret = io_linkat(req, issue_flags);
+		break;
 	default:
 		ret = -EINVAL;
 		break;
diff --git a/fs/namei.c b/fs/namei.c
index f241348e64f4..b5adfd4f7de6 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -4358,7 +4358,7 @@ EXPORT_SYMBOL(vfs_link);
  * with linux 2.0, and to avoid hard-linking to directories
  * and other special files.  --ADM
  */
-static int do_linkat(int olddfd, struct filename *old, int newdfd,
+int do_linkat(int olddfd, struct filename *old, int newdfd,
 	      struct filename *new, int flags)
 {
 	struct user_namespace *mnt_userns;
diff --git a/include/uapi/linux/io_uring.h b/include/uapi/linux/io_uring.h
index 61fd347ab176..10eb38d2864f 100644
--- a/include/uapi/linux/io_uring.h
+++ b/include/uapi/linux/io_uring.h
@@ -44,6 +44,7 @@ struct io_uring_sqe {
 		__u32		splice_flags;
 		__u32		rename_flags;
 		__u32		unlink_flags;
+		__u32		hardlink_flags;
 	};
 	__u64	user_data;	/* data to be passed back at completion time */
 	/* pack this to avoid bogus arm OABI complaints */
@@ -135,6 +136,7 @@ enum {
 	IORING_OP_UNLINKAT,
 	IORING_OP_MKDIRAT,
 	IORING_OP_SYMLINKAT,
+	IORING_OP_LINKAT,
 
 	/* this goes last, obviously */
 	IORING_OP_LAST,
-- 
2.30.2

Re: [PATCH v8 00/11] io_uring: add mkdir and [sym]linkat support

[Dmitry Kadashev]

On Wed, Jul 7, 2021 at 7:27 PM Dmitry Kadashev <dkadashev@gmail.com> wrote:
>
> This started out as an attempt to add mkdirat support to io_uring which
> is heavily based on renameat() / unlinkat() support.
>
> During the review process more operations were added (linkat, symlinkat,
> mknodat) mainly to keep things uniform internally (in namei.c), and
> with things changed in namei.c adding support for these operations to
> io_uring is trivial, so that was done too (except for mknodat). See
> https://lore.kernel.org/io-uring/20210514145259.wtl4xcsp52woi6ab@wittgenstein/

Christian, I've kept your Acked-by on the old commits, since changes
there are pretty minimal (conditional putname()s are unconditional now,
and with __filename_* functions not consuming their args now some goto
labels changed). Hope that was the right thing to do.

-- 
Dmitry Kadashev

Re: [PATCH v8 00/11] io_uring: add mkdir and [sym]linkat support

[Linus Torvalds]

On Wed, Jul 7, 2021 at 5:28 AM Dmitry Kadashev <dkadashev@gmail.com> wrote:
>
> This started out as an attempt to add mkdirat support to io_uring which
> is heavily based on renameat() / unlinkat() support.

Ok, sorry for having made you go through all the different versions,
but I like the new series and think it's a marked improvement.

I did send out a few comments to the individual patches that I think
it can all now be made to be even more legible by avoiding some of the
goto spaghetti, but I think that would be a series on top.

(And I'd like to note again that I based all that on just reading the
patches, so there may be something there that makes it not work well).

One final request: can you keep the fs/namei.c patches as one entirely
separate series, and then do the io_uring parts at the end, rather
than intermixing them?

But at least I am generally happy with this version.

Al - please holler now if you see any issues.

               Linus

Re: [PATCH v8 00/11] io_uring: add mkdir and [sym]linkat support

[Dmitry Kadashev]

On Thu, Jul 8, 2021 at 2:26 AM Linus Torvalds
<torvalds@linux-foundation.org> wrote:
>
> On Wed, Jul 7, 2021 at 5:28 AM Dmitry Kadashev <dkadashev@gmail.com> wrote:
> >
> > This started out as an attempt to add mkdirat support to io_uring which
> > is heavily based on renameat() / unlinkat() support.
>
> Ok, sorry for having made you go through all the different versions,
> but I like the new series and think it's a marked improvement.

No worries at all!

> I did send out a few comments to the individual patches that I think
> it can all now be made to be even more legible by avoiding some of the
> goto spaghetti, but I think that would be a series on top.
>
> (And I'd like to note again that I based all that on just reading the
> patches, so there may be something there that makes it not work well).

I'm happy to do that. I suppose it will have to go through Al's tree
rather than Jens' one?

> One final request: can you keep the fs/namei.c patches as one entirely
> separate series, and then do the io_uring parts at the end, rather
> than intermixing them?

Sure. I'm a bit confused if you mean splitting the series into two or
just moving io_uring bits to the end though. I'll send a v9 with
io_uring bits moved to the end of the series and if you prefer it to be
a completely separate series then let me know.

Thank you for the help again, Linus!

-- 
Dmitry Kadashev