Documentation/litmus-tests/atomic: Add a test for atomic_set()
authorBoqun Feng <boqun.feng@gmail.com>
Thu, 26 Mar 2020 02:40:21 +0000 (10:40 +0800)
committerPaul E. McKenney <paulmck@kernel.org>
Mon, 29 Jun 2020 19:05:18 +0000 (12:05 -0700)
commit4dcd4d36ddb1fa7fa7257ffe9e711608119b9785
tree75832a38c12f4abf413b30ebfdd91fc01ecd6ed1
parentefff6150209694a78c8af8c2a7557af682086220
Documentation/litmus-tests/atomic: Add a test for atomic_set()

We already use a litmus test in atomic_t.txt to describe the behavior of
an atomic_set() with the an atomic RMW, so add it into atomic-tests
directory to make it easily accessible for anyone who cares about the
semantics of our atomic APIs.

Besides currently the litmus test "atomic-set" in atomic_t.txt has a few
things to be improved:

1) The CPU/Processor numbers "P1,P2" are not only inconsistent with
the rest of the document, which uses "CPU0" and "CPU1", but also
unacceptable by the herd tool, which requires processors start
at "P0".

2) The initialization block uses a "atomic_set()", which is OK, but
it's better to use ATOMIC_INIT() to make clear this is an
initialization.

3) The return value of atomic_add_unless() is discarded
inexplicitly, which is OK for C language, but it will be helpful
to the herd tool if we use a void cast to make the discard
explicit.

4) The name and the paragraph describing the test need to be more
accurate and aligned with our wording in LKMM.

Therefore fix these in both atomic_t.txt and the new added litmus test.

Acked-by: Andrea Parri <parri.andrea@gmail.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Boqun Feng <boqun.feng@gmail.com>
Reviewed-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Documentation/atomic_t.txt
Documentation/litmus-tests/atomic/Atomic-RMW-ops-are-atomic-WRT-atomic_set.litmus [new file with mode: 0644]
Documentation/litmus-tests/atomic/README