1 .. SPDX-License-Identifier: GPL-2.0
2 .. Copyright (C) 2019, Google LLC.
3
4 Kernel Concurrency Sanitizer (KCSAN)
5 ====================================
6
7 The Kernel Concurrency Sanitizer (KCSAN) is a dynamic race detector, which
8 relies on compile-time instrumentation, and uses a watchpoint-based sampling
9 approach to detect races. KCSAN's primary purpose is to detect `data races`_.
10
11 Usage
12 -----
13
14 KCSAN is supported by both GCC and Clang. With GCC we require version 11 or
15 later, and with Clang also require version 11 or later.
16
17 To enable KCSAN configure the kernel with::
18
19 CONFIG_KCSAN = y
20
21 KCSAN provides several other configuration options to customize behaviour (see
22 the respective help text in ``lib/Kconfig.kcsan`` for more info).
23
24 Error reports
25 ~~~~~~~~~~~~~
26
27 A typical data race report looks like this::
28
29 ==================================================================
30 BUG: KCSAN: data-race in test_kernel_read / test_kernel_write
31
32 write to 0xffffffffc009a628 of 8 bytes by task 487 on cpu 0:
33 test_kernel_write+0x1d/0x30
34 access_thread+0x89/0xd0
35 kthread+0x23e/0x260
36 ret_from_fork+0x22/0x30
37
38 read to 0xffffffffc009a628 of 8 bytes by task 488 on cpu 6:
39 test_kernel_read+0x10/0x20
40 access_thread+0x89/0xd0
41 kthread+0x23e/0x260
42 ret_from_fork+0x22/0x30
43
44 value changed: 0x00000000000009a6 -> 0x00000000000009b2
45
46 Reported by Kernel Concurrency Sanitizer on:
47 CPU: 6 PID: 488 Comm: access_thread Not tainted 5.12.0-rc2+ #1
48 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.14.0-2 04/01/2014
49 ==================================================================
50
51 The header of the report provides a short summary of the functions involved in
52 the race. It is followed by the access types and stack traces of the 2 threads
53 involved in the data race. If KCSAN also observed a value change, the observed
54 old value and new value are shown on the "value changed" line respectively.
55
56 The other less common type of data race report looks like this::
57
58 ==================================================================
59 BUG: KCSAN: data-race in test_kernel_rmw_array+0x71/0xd0
60
61 race at unknown origin, with read to 0xffffffffc009bdb0 of 8 bytes by task 515 on cpu 2:
62 test_kernel_rmw_array+0x71/0xd0
63 access_thread+0x89/0xd0
64 kthread+0x23e/0x260
65 ret_from_fork+0x22/0x30
66
67 value changed: 0x0000000000002328 -> 0x0000000000002329
68
69 Reported by Kernel Concurrency Sanitizer on:
70 CPU: 2 PID: 515 Comm: access_thread Not tainted 5.12.0-rc2+ #1
71 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.14.0-2 04/01/2014
72 ==================================================================
73
74 This report is generated where it was not possible to determine the other
75 racing thread, but a race was inferred due to the data value of the watched
76 memory location having changed. These reports always show a "value changed"
77 line. A common reason for reports of this type are missing instrumentation in
78 the racing thread, but could also occur due to e.g. DMA accesses. Such reports
79 are shown only if ``CONFIG_KCSAN_REPORT_RACE_UNKNOWN_ORIGIN=y``, which is
80 enabled by default.
81
82 Selective analysis
83 ~~~~~~~~~~~~~~~~~~
84
85 It may be desirable to disable data race detection for specific accesses,
86 functions, compilation units, or entire subsystems. For static blacklisting,
87 the below options are available:
88
89 * KCSAN understands the ``data_race(expr)`` annotation, which tells KCSAN that
90 any data races due to accesses in ``expr`` should be ignored and resulting
91 behaviour when encountering a data race is deemed safe. Please see
92 `"Marking Shared-Memory Accesses" in the LKMM`_ for more information.
93
94 * Similar to ``data_race(...)``, the type qualifier ``__data_racy`` can be used
95 to document that all data races due to accesses to a variable are intended
96 and should be ignored by KCSAN::
97
98 struct foo {
99 ...
100 int __data_racy stats_counter;
101 ...
102 };
103
104 * Disabling data race detection for entire functions can be accomplished by
105 using the function attribute ``__no_kcsan``::
106
107 __no_kcsan
108 void foo(void) {
109 ...
110
111 To dynamically limit for which functions to generate reports, see the
112 `DebugFS interface`_ blacklist/whitelist feature.
113
114 * To disable data race detection for a particular compilation unit, add to the
115 ``Makefile``::
116
117 KCSAN_SANITIZE_file.o := n
118
119 * To disable data race detection for all compilation units listed in a
120 ``Makefile``, add to the respective ``Makefile``::
121
122 KCSAN_SANITIZE := n
123
124 .. _"Marking Shared-Memory Accesses" in the LKMM: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/tools/memory-model/Documentation/access-marking.txt
125
126 Furthermore, it is possible to tell KCSAN to show or hide entire classes of
127 data races, depending on preferences. These can be changed via the following
128 Kconfig options:
129
130 * ``CONFIG_KCSAN_REPORT_VALUE_CHANGE_ONLY``: If enabled and a conflicting write
131 is observed via a watchpoint, but the data value of the memory location was
132 observed to remain unchanged, do not report the data race.
133
134 * ``CONFIG_KCSAN_ASSUME_PLAIN_WRITES_ATOMIC``: Assume that plain aligned writes
135 up to word size are atomic by default. Assumes that such writes are not
136 subject to unsafe compiler optimizations resulting in data races. The option
137 causes KCSAN to not report data races due to conflicts where the only plain
138 accesses are aligned writes up to word size.
139
140 * ``CONFIG_KCSAN_PERMISSIVE``: Enable additional permissive rules to ignore
141 certain classes of common data races. Unlike the above, the rules are more
142 complex involving value-change patterns, access type, and address. This
143 option depends on ``CONFIG_KCSAN_REPORT_VALUE_CHANGE_ONLY=y``. For details
144 please see the ``kernel/kcsan/permissive.h``. Testers and maintainers that
145 only focus on reports from specific subsystems and not the whole kernel are
146 recommended to disable this option.
147
148 To use the strictest possible rules, select ``CONFIG_KCSAN_STRICT=y``, which
149 configures KCSAN to follow the Linux-kernel memory consistency model (LKMM) as
150 closely as possible.
151
152 DebugFS interface
153 ~~~~~~~~~~~~~~~~~
154
155 The file ``/sys/kernel/debug/kcsan`` provides the following interface:
156
157 * Reading ``/sys/kernel/debug/kcsan`` returns various runtime statistics.
158
159 * Writing ``on`` or ``off`` to ``/sys/kernel/debug/kcsan`` allows turning KCSAN
160 on or off, respectively.
161
162 * Writing ``!some_func_name`` to ``/sys/kernel/debug/kcsan`` adds
163 ``some_func_name`` to the report filter list, which (by default) blacklists
164 reporting data races where either one of the top stackframes are a function
165 in the list.
166
167 * Writing either ``blacklist`` or ``whitelist`` to ``/sys/kernel/debug/kcsan``
168 changes the report filtering behaviour. For example, the blacklist feature
169 can be used to silence frequently occurring data races; the whitelist feature
170 can help with reproduction and testing of fixes.
171
172 Tuning performance
173 ~~~~~~~~~~~~~~~~~~
174
175 Core parameters that affect KCSAN's overall performance and bug detection
176 ability are exposed as kernel command-line arguments whose defaults can also be
177 changed via the corresponding Kconfig options.
178
179 * ``kcsan.skip_watch`` (``CONFIG_KCSAN_SKIP_WATCH``): Number of per-CPU memory
180 operations to skip, before another watchpoint is set up. Setting up
181 watchpoints more frequently will result in the likelihood of races to be
182 observed to increase. This parameter has the most significant impact on
183 overall system performance and race detection ability.
184
185 * ``kcsan.udelay_task`` (``CONFIG_KCSAN_UDELAY_TASK``): For tasks, the
186 microsecond delay to stall execution after a watchpoint has been set up.
187 Larger values result in the window in which we may observe a race to
188 increase.
189
190 * ``kcsan.udelay_interrupt`` (``CONFIG_KCSAN_UDELAY_INTERRUPT``): For
191 interrupts, the microsecond delay to stall execution after a watchpoint has
192 been set up. Interrupts have tighter latency requirements, and their delay
193 should generally be smaller than the one chosen for tasks.
194
195 They may be tweaked at runtime via ``/sys/module/kcsan/parameters/``.
196
197 Data Races
198 ----------
199
200 In an execution, two memory accesses form a *data race* if they *conflict*,
201 they happen concurrently in different threads, and at least one of them is a
202 *plain access*; they *conflict* if both access the same memory location, and at
203 least one is a write. For a more thorough discussion and definition, see `"Plain
204 Accesses and Data Races" in the LKMM`_.
205
206 .. _"Plain Accesses and Data Races" in the LKMM: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/tools/memory-model/Documentation/explanation.txt#n1922
207
208 Relationship with the Linux-Kernel Memory Consistency Model (LKMM)
209 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
210
211 The LKMM defines the propagation and ordering rules of various memory
212 operations, which gives developers the ability to reason about concurrent code.
213 Ultimately this allows to determine the possible executions of concurrent code,
214 and if that code is free from data races.
215
216 KCSAN is aware of *marked atomic operations* (``READ_ONCE``, ``WRITE_ONCE``,
217 ``atomic_*``, etc.), and a subset of ordering guarantees implied by memory
218 barriers. With ``CONFIG_KCSAN_WEAK_MEMORY=y``, KCSAN models load or store
219 buffering, and can detect missing ``smp_mb()``, ``smp_wmb()``, ``smp_rmb()``,
220 ``smp_store_release()``, and all ``atomic_*`` operations with equivalent
221 implied barriers.
222
223 Note, KCSAN will not report all data races due to missing memory ordering,
224 specifically where a memory barrier would be required to prohibit subsequent
225 memory operation from reordering before the barrier. Developers should
226 therefore carefully consider the required memory ordering requirements that
227 remain unchecked.
228
229 Race Detection Beyond Data Races
230 --------------------------------
231
232 For code with complex concurrency design, race-condition bugs may not always
233 manifest as data races. Race conditions occur if concurrently executing
234 operations result in unexpected system behaviour. On the other hand, data races
235 are defined at the C-language level. The following macros can be used to check
236 properties of concurrent code where bugs would not manifest as data races.
237
238 .. kernel-doc:: include/linux/kcsan-checks.h
239 :functions: ASSERT_EXCLUSIVE_WRITER ASSERT_EXCLUSIVE_WRITER_SCOPED
240 ASSERT_EXCLUSIVE_ACCESS ASSERT_EXCLUSIVE_ACCESS_SCOPED
241 ASSERT_EXCLUSIVE_BITS
242
243 Implementation Details
244 ----------------------
245
246 KCSAN relies on observing that two accesses happen concurrently. Crucially, we
247 want to (a) increase the chances of observing races (especially for races that
248 manifest rarely), and (b) be able to actually observe them. We can accomplish
249 (a) by injecting various delays, and (b) by using address watchpoints (or
250 breakpoints).
251
252 If we deliberately stall a memory access, while we have a watchpoint for its
253 address set up, and then observe the watchpoint to fire, two accesses to the
254 same address just raced. Using hardware watchpoints, this is the approach taken
255 in `DataCollider
256 <http://usenix.org/legacy/events/osdi10/tech/full_papers/Erickson.pdf>`_.
257 Unlike DataCollider, KCSAN does not use hardware watchpoints, but instead
258 relies on compiler instrumentation and "soft watchpoints".
259
260 In KCSAN, watchpoints are implemented using an efficient encoding that stores
261 access type, size, and address in a long; the benefits of using "soft
262 watchpoints" are portability and greater flexibility. KCSAN then relies on the
263 compiler instrumenting plain accesses. For each instrumented plain access:
264
265 1. Check if a matching watchpoint exists; if yes, and at least one access is a
266 write, then we encountered a racing access.
267
268 2. Periodically, if no matching watchpoint exists, set up a watchpoint and
269 stall for a small randomized delay.
270
271 3. Also check the data value before the delay, and re-check the data value
272 after delay; if the values mismatch, we infer a race of unknown origin.
273
274 To detect data races between plain and marked accesses, KCSAN also annotates
275 marked accesses, but only to check if a watchpoint exists; i.e. KCSAN never
276 sets up a watchpoint on marked accesses. By never setting up watchpoints for
277 marked operations, if all accesses to a variable that is accessed concurrently
278 are properly marked, KCSAN will never trigger a watchpoint and therefore never
279 report the accesses.
280
281 Modeling Weak Memory
282 ~~~~~~~~~~~~~~~~~~~~
283
284 KCSAN's approach to detecting data races due to missing memory barriers is
285 based on modeling access reordering (with ``CONFIG_KCSAN_WEAK_MEMORY=y``).
286 Each plain memory access for which a watchpoint is set up, is also selected for
287 simulated reordering within the scope of its function (at most 1 in-flight
288 access).
289
290 Once an access has been selected for reordering, it is checked along every
291 other access until the end of the function scope. If an appropriate memory
292 barrier is encountered, the access will no longer be considered for simulated
293 reordering.
294
295 When the result of a memory operation should be ordered by a barrier, KCSAN can
296 then detect data races where the conflict only occurs as a result of a missing
297 barrier. Consider the example::
298
299 int x, flag;
300 void T1(void)
301 {
302 x = 1; // data race!
303 WRITE_ONCE(flag, 1); // correct: smp_store_release(&flag, 1)
304 }
305 void T2(void)
306 {
307 while (!READ_ONCE(flag)); // correct: smp_load_acquire(&flag)
308 ... = x; // data race!
309 }
310
311 When weak memory modeling is enabled, KCSAN can consider ``x`` in ``T1`` for
312 simulated reordering. After the write of ``flag``, ``x`` is again checked for
313 concurrent accesses: because ``T2`` is able to proceed after the write of
314 ``flag``, a data race is detected. With the correct barriers in place, ``x``
315 would not be considered for reordering after the proper release of ``flag``,
316 and no data race would be detected.
317
318 Deliberate trade-offs in complexity but also practical limitations mean only a
319 subset of data races due to missing memory barriers can be detected. With
320 currently available compiler support, the implementation is limited to modeling
321 the effects of "buffering" (delaying accesses), since the runtime cannot
322 "prefetch" accesses. Also recall that watchpoints are only set up for plain
323 accesses, and the only access type for which KCSAN simulates reordering. This
324 means reordering of marked accesses is not modeled.
325
326 A consequence of the above is that acquire operations do not require barrier
327 instrumentation (no prefetching). Furthermore, marked accesses introducing
328 address or control dependencies do not require special handling (the marked
329 access cannot be reordered, later dependent accesses cannot be prefetched).
330
331 Key Properties
332 ~~~~~~~~~~~~~~
333
334 1. **Memory Overhead:** The overall memory overhead is only a few MiB
335 depending on configuration. The current implementation uses a small array of
336 longs to encode watchpoint information, which is negligible.
337
338 2. **Performance Overhead:** KCSAN's runtime aims to be minimal, using an
339 efficient watchpoint encoding that does not require acquiring any shared
340 locks in the fast-path. For kernel boot on a system with 8 CPUs:
341
342 - 5.0x slow-down with the default KCSAN config;
343 - 2.8x slow-down from runtime fast-path overhead only (set very large
344 ``KCSAN_SKIP_WATCH`` and unset ``KCSAN_SKIP_WATCH_RANDOMIZE``).
345
346 3. **Annotation Overheads:** Minimal annotations are required outside the KCSAN
347 runtime. As a result, maintenance overheads are minimal as the kernel
348 evolves.
349
350 4. **Detects Racy Writes from Devices:** Due to checking data values upon
351 setting up watchpoints, racy writes from devices can also be detected.
352
353 5. **Memory Ordering:** KCSAN is aware of only a subset of LKMM ordering rules;
354 this may result in missed data races (false negatives).
355
356 6. **Analysis Accuracy:** For observed executions, due to using a sampling
357 strategy, the analysis is *unsound* (false negatives possible), but aims to
358 be complete (no false positives).
359
360 Alternatives Considered
361 -----------------------
362
363 An alternative data race detection approach for the kernel can be found in the
364 `Kernel Thread Sanitizer (KTSAN) <https://github.com/google/ktsan/wiki>`_.
365 KTSAN is a happens-before data race detector, which explicitly establishes the
366 happens-before order between memory operations, which can then be used to
367 determine data races as defined in `Data Races`_.
368
369 To build a correct happens-before relation, KTSAN must be aware of all ordering
370 rules of the LKMM and synchronization primitives. Unfortunately, any omission
371 leads to large numbers of false positives, which is especially detrimental in
372 the context of the kernel which includes numerous custom synchronization
373 mechanisms. To track the happens-before relation, KTSAN's implementation
374 requires metadata for each memory location (shadow memory), which for each page
375 corresponds to 4 pages of shadow memory, and can translate into overhead of
376 tens of GiB on a large system.
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.