Re: [PATCH v2 bpf-next 1/3] bpf: support O_PATH FDs in BPF_OBJ_PIN and BPF_OBJ_GET commands

[Andrii Nakryiko <andrii.nakryiko@gmail.com>]

On Thu, May 18, 2023 at 4:59 PM kernel test robot <lkp@intel.com> wrote:
>
> Hi Andrii,
>
> kernel test robot noticed the following build warnings:
>
> [auto build test WARNING on bpf-next/master]
>
> url:    https://github.com/intel-lab-lkp/linux/commits/Andrii-Nakryiko/bpf-support-O_PATH-FDs-in-BPF_OBJ_PIN-and-BPF_OBJ_GET-commands/20230519-060110
> base:   https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next.git master
> patch link:    https://lore.kernel.org/r/20230518215444.1418789-2-andrii%40kernel.org
> patch subject: [PATCH v2 bpf-next 1/3] bpf: support O_PATH FDs in BPF_OBJ_PIN and BPF_OBJ_GET commands
> config: m68k-allyesconfig
> compiler: m68k-linux-gcc (GCC) 12.1.0
> reproduce (this is a W=1 build):
>         wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
>         chmod +x ~/bin/make.cross
>         # https://github.com/intel-lab-lkp/linux/commit/74f08e59c3fccac04b3c080831615209e11be4fb
>         git remote add linux-review https://github.com/intel-lab-lkp/linux
>         git fetch --no-tags linux-review Andrii-Nakryiko/bpf-support-O_PATH-FDs-in-BPF_OBJ_PIN-and-BPF_OBJ_GET-commands/20230519-060110
>         git checkout 74f08e59c3fccac04b3c080831615209e11be4fb
>         # save the config file
>         mkdir build_dir && cp config build_dir/.config
>         COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-12.1.0 make.cross W=1 O=build_dir ARCH=m68k olddefconfig
>         COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-12.1.0 make.cross W=1 O=build_dir ARCH=m68k SHELL=/bin/bash kernel/
>
> If you fix the issue, kindly add following tag where applicable
> | Reported-by: kernel test robot <lkp@intel.com>
> | Closes: https://lore.kernel.org/oe-kbuild-all/202305190724.nnh1ZV2F-lkp@intel.com/
>
> All warnings (new ones prefixed by >>):
>
>    kernel/bpf/syscall.c: In function 'bpf_obj_get':
> >> kernel/bpf/syscall.c:2720:13: warning: variable 'path_fd' set but not used [-Wunused-but-set-variable]
>     2720 |         int path_fd;
>          |             ^~~~~~~
>
>
> vim +/path_fd +2720 kernel/bpf/syscall.c
>
>   2717
>   2718  static int bpf_obj_get(const union bpf_attr *attr)
>   2719  {
> > 2720          int path_fd;
>   2721
>   2722          if (CHECK_ATTR(BPF_OBJ) || attr->bpf_fd != 0 ||
>   2723              attr->file_flags & ~(BPF_OBJ_FLAG_MASK | BPF_F_PATH_FD))
>   2724                  return -EINVAL;
>   2725
>   2726          /* path_fd has to be accompanied by BPF_F_PATH_FD flag */
>   2727          if (!(attr->file_flags & BPF_F_PATH_FD) && attr->path_fd)
>   2728                  return -EINVAL;
>   2729
>   2730          path_fd = attr->file_flags & BPF_F_PATH_FD ? attr->path_fd : AT_FDCWD;
>   2731          return bpf_obj_get_user(attr->path_fd, u64_to_user_ptr(attr->pathname),

argh.... s/attr->path_fd/path_fd/ here, but it's curious how tests
didn't catch this because we always provide absolute path in
attr->pathname, and from openat() man page:

  If pathname is absolute, then dirfd is ignored.

So whatever garbage FD is passed in (in this case 0), the kernel won't
complain. Interesting. I'll send v3 with a fix.

>   2732                                  attr->file_flags);
>   2733  }
>   2734
>
> --
> 0-DAY CI Kernel Test Service
> https://github.com/intel/lkp-tests/wiki