This is v2 of the CephFS decoder-bound series. The first two patches are
unchanged code-wise and now carry Slava's Reviewed-by. Patches 3 and 4
address the review feedback on overflow-safe sizing and aggregate delegated
inode bounds.

The four bugs are still independent:

  1/4 rejects a final xattr value length that runs past the xattr blob.
  2/4 bounds MDSCapAuth path and fs_name copies in handle_session().
  3/4 bounds the mdsmap export_targets array for info_v 2/3.
  4/4 bounds delegated-inode parsing by session population and by one
      reply's aggregate interval length.

Changes in v2:

  - Add Reviewed-by: Viacheslav Dubeyko to patches 1 and 2.
  - Patch 3 computes the export-targets byte count with size_mul() and
    reuses the checked length for the cursor advance.
  - Patch 4 replaces the per-interval cap with a per-session population
    counter and a per-reply interval budget, so repeated replies and
    duplicate ranges are bounded too. The cap stays a fixed client-side
    constant because the kernel client never sees the userspace
    mds_client_prealloc_inos option; it is sized as a generous multiple of
    that option's documented default of 1000.

Michael Bommarito (4):
  ceph: bound xattr value length in __build_xattrs()
  ceph: bound MDSCapAuth path and fs_name decode in handle_session()
  ceph: bound num_export_targets array for mds info v2/v3
  ceph: cap delegated inode count in ceph_parse_deleg_inos()

 fs/ceph/mds_client.c | 59 ++++++++++++++++++++++++++++++++++++++------
 fs/ceph/mds_client.h |  1 +
 fs/ceph/mdsmap.c     |  7 +++++-
 fs/ceph/super.h      |  9 +++++++
 fs/ceph/xattr.c      |  1 +
 5 files changed, 68 insertions(+), 9 deletions(-)


base-commit: f72c95f3a516d87483e225ae081a402a09fd0127
-- 
2.53.0