[PATCH v2 0/2] task_work_queue() && keyctl_session_to_parent()

[Oleg Nesterov <oleg@redhat.com>]

Hello.

So, task_work_queue() again, now with the first user.

Compared to v1, 1/2 adds task_work_cancel(func), it removes
the first work with the same callback. May be it should remove
all and return the list, or we can add task_work_requeue()
instead. Easy to change if needed, although I think a simple
_cancel is enough.

checkpatch.pl dislikes replace_session_keyring() in 2/2, but
I disagree. This code was copy-and-past'ed, and I really think
that the "unnecessary" spaces make it more readable.

Of course, I can "fix" the code to make checkpatch.pl happy.

I tried to test this, seems to work. But please review the
the error-handling in keyctl_session_to_parent().

I do not like when diffstat shows the code bloat, but note that
the numbers will be much better after we kill the no longer used
->replacement_session_keyring and cleanup arch/*/kernel/signal.c

Oleg.

 include/linux/key.h          |    2 +-
 include/linux/sched.h        |    2 +
 include/linux/task_work.h    |   48 +++++++++++++++++++++++
 include/linux/tracehook.h    |   10 ++++-
 kernel/Makefile              |    2 +-
 kernel/exit.c                |    5 ++-
 kernel/fork.c                |    1 +
 kernel/task_work.c           |   77 +++++++++++++++++++++++++++++++++++++
 security/keys/keyctl.c       |   87 +++++++++++++++++++++++++++++++-----------
 security/keys/process_keys.c |   49 -----------------------
 10 files changed, 207 insertions(+), 76 deletions(-)