1===================================
2refcount_t API compared to atomic_t
3===================================
4
5.. contents:: :local:
6
7Introduction
8============
9
10The goal of refcount_t API is to provide a minimal API for implementing
11an object's reference counters. While a generic architecture-independent
12implementation from lib/refcount.c uses atomic operations underneath,
13there are a number of differences between some of the ``refcount_*()`` and
14``atomic_*()`` functions with regards to the memory ordering guarantees.
15This document outlines the differences and provides respective examples
16in order to help maintainers validate their code against the change in
17these memory ordering guarantees.
18
19The terms used through this document try to follow the formal LKMM defined in
20github.com/aparri/memory-model/blob/master/Documentation/explanation.txt
21
22memory-barriers.txt and atomic_t.txt provide more background to the
23memory ordering in general and for atomic operations specifically.
24
25Relevant types of memory ordering
26=================================
27
28.. note:: The following section only covers some of the memory
29   ordering types that are relevant for the atomics and reference
30   counters and used through this document. For a much broader picture
31   please consult memory-barriers.txt document.
32
33In the absence of any memory ordering guarantees (i.e. fully unordered)
34atomics & refcounters only provide atomicity and
35program order (po) relation (on the same CPU). It guarantees that
36each ``atomic_*()`` and ``refcount_*()`` operation is atomic and instructions
37are executed in program order on a single CPU.
38This is implemented using :c:func:`READ_ONCE`/:c:func:`WRITE_ONCE` and
39compare-and-swap primitives.
40
41A strong (full) memory ordering guarantees that all prior loads and
42stores (all po-earlier instructions) on the same CPU are completed
43before any po-later instruction is executed on the same CPU.
44It also guarantees that all po-earlier stores on the same CPU
45and all propagated stores from other CPUs must propagate to all
46other CPUs before any po-later instruction is executed on the original
47CPU (A-cumulative property). This is implemented using :c:func:`smp_mb`.
48
49A RELEASE memory ordering guarantees that all prior loads and
50stores (all po-earlier instructions) on the same CPU are completed
51before the operation. It also guarantees that all po-earlier
52stores on the same CPU and all propagated stores from other CPUs
53must propagate to all other CPUs before the release operation
54(A-cumulative property). This is implemented using
55:c:func:`smp_store_release`.
56
57A control dependency (on success) for refcounters guarantees that
58if a reference for an object was successfully obtained (reference
59counter increment or addition happened, function returned true),
60then further stores are ordered against this operation.
61Control dependency on stores are not implemented using any explicit
62barriers, but rely on CPU not to speculate on stores. This is only
63a single CPU relation and provides no guarantees for other CPUs.
64
65
66Comparison of functions
67=======================
68
69case 1) - non-"Read/Modify/Write" (RMW) ops
70-------------------------------------------
71
72Function changes:
73
74 * :c:func:`atomic_set` --> :c:func:`refcount_set`
75 * :c:func:`atomic_read` --> :c:func:`refcount_read`
76
77Memory ordering guarantee changes:
78
79 * none (both fully unordered)
80
81
82case 2) - increment-based ops that return no value
83--------------------------------------------------
84
85Function changes:
86
87 * :c:func:`atomic_inc` --> :c:func:`refcount_inc`
88 * :c:func:`atomic_add` --> :c:func:`refcount_add`
89
90Memory ordering guarantee changes:
91
92 * none (both fully unordered)
93
94case 3) - decrement-based RMW ops that return no value
95------------------------------------------------------
96
97Function changes:
98
99 * :c:func:`atomic_dec` --> :c:func:`refcount_dec`
100
101Memory ordering guarantee changes:
102
103 * fully unordered --> RELEASE ordering
104
105
106case 4) - increment-based RMW ops that return a value
107-----------------------------------------------------
108
109Function changes:
110
111 * :c:func:`atomic_inc_not_zero` --> :c:func:`refcount_inc_not_zero`
112 * no atomic counterpart --> :c:func:`refcount_add_not_zero`
113
114Memory ordering guarantees changes:
115
116 * fully ordered --> control dependency on success for stores
117
118.. note:: We really assume here that necessary ordering is provided as a
119   result of obtaining pointer to the object!
120
121
122case 5) - decrement-based RMW ops that return a value
123-----------------------------------------------------
124
125Function changes:
126
127 * :c:func:`atomic_dec_and_test` --> :c:func:`refcount_dec_and_test`
128 * :c:func:`atomic_sub_and_test` --> :c:func:`refcount_sub_and_test`
129 * no atomic counterpart --> :c:func:`refcount_dec_if_one`
130 * ``atomic_add_unless(&var, -1, 1)`` --> ``refcount_dec_not_one(&var)``
131
132Memory ordering guarantees changes:
133
134 * fully ordered --> RELEASE ordering + control dependency
135
136.. note:: :c:func:`atomic_add_unless` only provides full order on success.
137
138
139case 6) - lock-based RMW
140------------------------
141
142Function changes:
143
144 * :c:func:`atomic_dec_and_lock` --> :c:func:`refcount_dec_and_lock`
145 * :c:func:`atomic_dec_and_mutex_lock` --> :c:func:`refcount_dec_and_mutex_lock`
146
147Memory ordering guarantees changes:
148
149 * fully ordered --> RELEASE ordering + control dependency + hold
150   :c:func:`spin_lock` on success
151