[PATCH v2 0/3] Fix shmem_rename2 directory offset calculation

[cel@kernel.org]

From: Chuck Lever <chuck.lever@oracle.com>

The existing code in shmem_rename2() allocates a fresh directory
offset value when renaming over an existing destination entry. User
space does not expect this behavior. In particular, applications
that rename while walking a directory can loop indefinitely because
they never reach the end of the directory.

The only test that is problematic at the moment is generic/449,
which live-locks (interruptibly). I don't have a baseline yet, so
I can't say whether the fix introduces this behavior or pre-dates
the shmem conversion to simple_offset.


--
Changes since v1:
- Patches reorganized for easier review and backport
- Passes git regression and fstests (with scratch device)
- Dropped the API clean-up patch for now

Chuck Lever (3):
  libfs: Fix simple_offset_rename_exchange()
  libfs: Add simple_offset_rename() API
  shmem: Fix shmem_rename2()

 fs/libfs.c         | 55 +++++++++++++++++++++++++++++++++++++++++-----
 include/linux/fs.h |  2 ++
 mm/shmem.c         |  3 +--
 3 files changed, 52 insertions(+), 8 deletions(-)


base-commit: fec50db7033ea478773b159e0e2efb135270e3b7
-- 
2.44.0