When binding a non-abstract AF_UNIX socket it will gain a representation in the filesystem. Enable the socket infrastructure to handle idmapped mounts by passing down the user namespace of the mount the socket will be created from. If the initial user namespace is passed nothing changes so non-idmapped mounts will see identical behavior as before. Link: https://lore.kernel.org/r/20210112220124.837960-25-christian.brauner@xxxxxxxxxx Cc: Christoph Hellwig <hch@xxxxxx> Cc: David Howells <dhowells@xxxxxxxxxx> Cc: Al Viro <viro@xxxxxxxxxxxxxxxxxx> Cc: linux-fsdevel@xxxxxxxxxxxxxxx Reviewed-by: Christoph Hellwig <hch@xxxxxx> Signed-off-by: Christian Brauner <christian.brauner@xxxxxxxxxx> --- /* v2 */ unchanged /* v3 */ unchanged /* v4 */ unchanged /* v5 */ unchanged base-commit: 7c53f6b671f4aba70ff15e1b05148b10d58c2837 /* v6 */ unchanged base-commit: 19c329f6808995b142b3966301f217c831e7cf31 --- net/unix/af_unix.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/net/unix/af_unix.c b/net/unix/af_unix.c index 9a1f3c04402e..5a31307ceb76 100644 --- a/net/unix/af_unix.c +++ b/net/unix/af_unix.c @@ -996,8 +996,8 @@ static int unix_mknod(const char *sun_path, umode_t mode, struct path *res) */ err = security_path_mknod(&path, dentry, mode, 0); if (!err) { - err = vfs_mknod(&init_user_ns, d_inode(path.dentry), dentry, - mode, 0); + err = vfs_mknod(mnt_user_ns(path.mnt), d_inode(path.dentry), + dentry, mode, 0); if (!err) { res->mnt = mntget(path.mnt); res->dentry = dget(dentry); -- 2.30.0