Hi all, There is ABBA deadlock bug between the AGI and AGF when performing rename() with RENAME_WHITEOUT flag, so add test to check that whether the rename() call works well. We add the renameat2 syscall support to fsstress, and then reproduce the deadlock problem by using fsstress. Changes for v3: - Add ancestor-descendant relationship checks for two dirs in RENAME_EXCHANGE. - Rebase the patchset to the latest xfstests. Changes for v2: - Fix the xattr_count value of the original devnode in RENAME_WHITEOUT. - Fix the parent ids swap problem in RENAME_EXCHANGE. - Add the necessary comments. kaixuxia (4): fsstress: show the real file id and parid in rename_f() fsstress: add NOREPLACE and WHITEOUT renameat2 support fsstress: add EXCHANGE renameat2 support xfs: test the deadlock between the AGI and AGF with RENAME_WHITEOUT ltp/fsstress.c | 231 +++++++++++++++++++++++++++++++++++++++++--------- tests/generic/585 | 56 ++++++++++++ tests/generic/585.out | 2 + tests/generic/group | 1 + 4 files changed, 251 insertions(+), 39 deletions(-) create mode 100755 tests/generic/585 create mode 100644 tests/generic/585.out -- 1.8.3.1
