1 # SPDX-License-Identifier: GPL-2.0-only
6 config HAVE_KCSAN_COMPILER
7 def_bool (CC_IS_CLANG && $(cc-option,-fsanitize=thread -mllvm -tsan-distinguish-volatile=1)) || \
8 (CC_IS_GCC && $(cc-option,-fsanitize=thread --param tsan-distinguish-volatile=1))
10 For the list of compilers that support KCSAN, please see
11 <file:Documentation/dev-tools/kcsan.rst>.
13 config KCSAN_KCOV_BROKEN
14 def_bool KCOV && CC_HAS_SANCOV_TRACE_PC
15 depends on CC_IS_CLANG
16 depends on !$(cc-option,-Werror=unused-command-line-argument -fsanitize=thread -fsanitize-coverage=trace-pc)
18 Some versions of clang support either KCSAN and KCOV but not the
19 combination of the two.
20 See https://bugs.llvm.org/show_bug.cgi?id=45831 for the status
24 bool "KCSAN: dynamic data race detector"
25 depends on HAVE_ARCH_KCSAN && HAVE_KCSAN_COMPILER
26 depends on DEBUG_KERNEL && !KASAN
27 depends on !KCSAN_KCOV_BROKEN
30 The Kernel Concurrency Sanitizer (KCSAN) is a dynamic
31 data-race detector that relies on compile-time instrumentation.
32 KCSAN uses a watchpoint-based sampling approach to detect races.
34 While KCSAN's primary purpose is to detect data races, it
35 also provides assertions to check data access constraints.
36 These assertions can expose bugs that do not manifest as
39 See <file:Documentation/dev-tools/kcsan.rst> for more details.
43 # Compiler capabilities that should not fail the test if they are unavailable.
44 config CC_HAS_TSAN_COMPOUND_READ_BEFORE_WRITE
45 def_bool (CC_IS_CLANG && $(cc-option,-fsanitize=thread -mllvm -tsan-compound-read-before-write=1)) || \
46 (CC_IS_GCC && $(cc-option,-fsanitize=thread --param tsan-compound-read-before-write=1))
49 bool "Show verbose reports with more information about system state"
50 depends on PROVE_LOCKING
52 If enabled, reports show more information about the system state that
53 may help better analyze and debug races. This includes held locks and
56 While this option should generally be benign, we call into more
57 external functions on report generation; if a race report is
58 generated from any one of them, system stability may suffer due to
59 deadlocks or recursion. If in doubt, say N.
62 bool "Debugging of KCSAN internals"
65 bool "Perform short selftests on boot"
68 Run KCSAN selftests on boot. On test failure, causes the kernel to
69 panic. Recommended to be enabled, ensuring critical functionality
73 tristate "KCSAN test for integrated runtime behaviour"
74 depends on TRACEPOINTS && KUNIT
77 KCSAN test focusing on behaviour of the integrated runtime. Tests
78 various race scenarios, and verifies the reports generated to
79 console. Makes use of KUnit for test organization, and the Torture
80 framework for test thread control.
82 Each test case may run at least up to KCSAN_REPORT_ONCE_IN_MS
83 milliseconds. Test run duration may be optimized by building the
84 kernel and KCSAN test with KCSAN_REPORT_ONCE_IN_MS set to a lower
87 Say Y here if you want the test to be built into the kernel and run
88 during boot; say M if you want the test to build as a module; say N
91 config KCSAN_EARLY_ENABLE
92 bool "Early enable during boot"
95 If KCSAN should be enabled globally as soon as possible. KCSAN can
96 later be enabled/disabled via debugfs.
98 config KCSAN_NUM_WATCHPOINTS
99 int "Number of available watchpoints"
102 Total number of available watchpoints. An address range maps into a
103 specific watchpoint slot as specified in kernel/kcsan/encoding.h.
104 Although larger number of watchpoints may not be usable due to
105 limited number of CPUs, a larger value helps to improve performance
106 due to reducing cache-line contention. The chosen default is a
107 conservative value; we should almost never observe "no_capacity"
108 events (see /sys/kernel/debug/kcsan).
110 config KCSAN_UDELAY_TASK
111 int "Delay in microseconds (for tasks)"
114 For tasks, the microsecond delay after setting up a watchpoint.
116 config KCSAN_UDELAY_INTERRUPT
117 int "Delay in microseconds (for interrupts)"
120 For interrupts, the microsecond delay after setting up a watchpoint.
121 Interrupts have tighter latency requirements, and their delay should
122 be lower than for tasks.
124 config KCSAN_DELAY_RANDOMIZE
125 bool "Randomize above delays"
128 If delays should be randomized, where the maximum is KCSAN_UDELAY_*.
129 If false, the chosen delays are always the KCSAN_UDELAY_* values
132 config KCSAN_SKIP_WATCH
133 int "Skip instructions before setting up watchpoint"
136 The number of per-CPU memory operations to skip, before another
137 watchpoint is set up, i.e. one in KCSAN_WATCH_SKIP per-CPU
138 memory operations are used to set up a watchpoint. A smaller value
139 results in more aggressive race detection, whereas a larger value
140 improves system performance at the cost of missing some races.
142 config KCSAN_SKIP_WATCH_RANDOMIZE
143 bool "Randomize watchpoint instruction skip count"
146 If instruction skip count should be randomized, where the maximum is
147 KCSAN_WATCH_SKIP. If false, the chosen value is always
150 config KCSAN_INTERRUPT_WATCHER
151 bool "Interruptible watchers"
153 If enabled, a task that set up a watchpoint may be interrupted while
154 delayed. This option will allow KCSAN to detect races between
155 interrupted tasks and other threads of execution on the same CPU.
157 Currently disabled by default, because not all safe per-CPU access
158 primitives and patterns may be accounted for, and therefore could
159 result in false positives.
161 config KCSAN_REPORT_ONCE_IN_MS
162 int "Duration in milliseconds, in which any given race is only reported once"
165 Any given race is only reported once in the defined time window.
166 Different races may still generate reports within a duration that is
167 smaller than the duration defined here. This allows rate limiting
168 reporting to avoid flooding the console with reports. Setting this
169 to 0 disables rate limiting.
171 # The main purpose of the below options is to control reported data races (e.g.
172 # in fuzzer configs), and are not expected to be switched frequently by other
173 # users. We could turn some of them into boot parameters, but given they should
174 # not be switched normally, let's keep them here to simplify configuration.
176 # The defaults below are chosen to be very conservative, and may miss certain
179 config KCSAN_REPORT_RACE_UNKNOWN_ORIGIN
180 bool "Report races of unknown origin"
183 If KCSAN should report races where only one access is known, and the
184 conflicting access is of unknown origin. This type of race is
185 reported if it was only possible to infer a race due to a data value
186 change while an access is being delayed on a watchpoint.
188 config KCSAN_REPORT_VALUE_CHANGE_ONLY
189 bool "Only report races where watcher observed a data value change"
192 If enabled and a conflicting write is observed via a watchpoint, but
193 the data value of the memory location was observed to remain
194 unchanged, do not report the data race.
196 config KCSAN_ASSUME_PLAIN_WRITES_ATOMIC
197 bool "Assume that plain aligned writes up to word size are atomic"
200 Assume that plain aligned writes up to word size are atomic by
201 default, and also not subject to other unsafe compiler optimizations
202 resulting in data races. This will cause KCSAN to not report data
203 races due to conflicts where the only plain accesses are aligned
204 writes up to word size: conflicts between marked reads and plain
205 aligned writes up to word size will not be reported as data races;
206 notice that data races between two conflicting plain aligned writes
207 will also not be reported.
209 config KCSAN_IGNORE_ATOMICS
210 bool "Do not instrument marked atomic accesses"
212 Never instrument marked atomic accesses. This option can be used for
213 additional filtering. Conflicting marked atomic reads and plain
214 writes will never be reported as a data race, however, will cause
215 plain reads and marked writes to result in "unknown origin" reports.
216 If combined with CONFIG_KCSAN_REPORT_RACE_UNKNOWN_ORIGIN=n, data
217 races where at least one access is marked atomic will never be
220 Similar to KCSAN_ASSUME_PLAIN_WRITES_ATOMIC, but including unaligned
221 accesses, conflicting marked atomic reads and plain writes will not
222 be reported as data races; however, unlike that option, data races
223 due to two conflicting plain writes will be reported (aligned and
224 unaligned, if CONFIG_KCSAN_ASSUME_PLAIN_WRITES_ATOMIC=n).