Messages in this thread | | | Date | Wed, 4 Sep 2013 02:03:02 +0100 | From | Al Viro <> | Subject | Re: [PATCH 1/3] autofs4 - fix device ioctl mount lookup |
| |
On Wed, Sep 04, 2013 at 08:54:57AM +0800, Ian Kent wrote: > +static int kern_path_top(const char *pathname, > + unsigned int flags, struct path *path) > +{ > + struct dentry *dentry; > + struct qstr name; > + const char *tmp; > + unsigned int len; > + int err; > + > + len = strlen(pathname); > + if (len <= 1) > + return -EINVAL; > + > + tmp = pathname + len - 1; > + len = 0; > + if (*tmp == '/') > + tmp--; > + do { > + if (*tmp == '/') > + break; > + len++; > + } while (--tmp >= pathname); > + tmp++; > + > + err = kern_path(pathname, flags | LOOKUP_PARENT, path); > + if (err) > + return err; > + > + name.name = tmp; > + name.len = len; > + name.hash = full_name_hash(tmp, len); > + > + dentry = d_lookup(path->dentry, &name); > + if (!dentry) { > + path_put(path); > + return -ENOENT; > + } > + dput(path->dentry); > + path->dentry = dentry; > + > + while (follow_down_one(path)) > + ; > + > + return 0; > +}
Ewwww... NAK in that form. Just what will happen if the last component given to that sucker will be . or .., for starters? Or a symlink, with '/' added to it to force following the damn thing?
| |