1 C11Tester: A Testing tool for C11 and C++11 Atomics
2 =====================================================
4 C11Tester is a testing tool for C11/C++11 which randomly explores the
5 behaviors of code under the C/C++ memory model.
7 C11Tester is constructed as a dynamically-linked shared library which
8 implements the C and C++ atomic types and portions of the other thread-support
9 libraries of C/C++ (e.g., std::atomic, std::mutex, etc.).
11 C11Tester compiles on Linux. Instrumenting programs requires using
12 our LLVM pass. It likely can be ported to other \*NIX flavors.
18 If you haven't done so already, you may download C11Tester using git:
20 git clone git://plrg.eecs.uci.edu/c11tester.git
22 Get the benchmarks (not required; distributed separately):
24 git clone git://plrg.eecs.uci.edu/c11concurrency-benchmarks.git
26 Get the LLVM frontend using git and follow its directions to build:
28 git clone git://plrg.eecs.uci.edu/c11llvm.git
34 To see the help message on how to run C11Tester, execute:
44 > Verbose: show all executions and not just buggy ones.
48 > Specify the number number of executions to run.
52 > Value to provide to atomics loads from uninitialized memory locations. The
53 > default is 0, but this may cause some programs to throw exceptions
54 > (segfault) before the model checker prints a trace.
59 Many simple tests are located in the `test/` directory. These are
60 manually instrumented and can just be run.
62 You may also want to try the larger benchmarks (distributed
63 separately). These require LLVM to instrument.
69 You likely want to test your own code, not just our tests. You will
70 likely need to use our LLVM pass to instrument your program. You will
71 have to modify your build environment to do this.
73 Test programs should be compiled against our shared library
74 (libmodel.so). Then the shared library must be made available to the
75 dynamic linker, using the `LD_LIBRARY_PATH` environment variable, for
79 Reading an execution trace
80 --------------------------
82 When C11Tester detects a bug in your program (or when run with the `--verbose`
83 flag), it prints the output of the program run (STDOUT) along with some summary
84 trace information for the execution in question. The trace is given as a
85 sequence of lines, where each line represents an operation in the execution
86 trace. These lines are ordered by the order in which they were run by C11Tester
87 (i.e., the "execution order"), which does not necessarily align with the "order"
88 of the values observed (i.e., the modification order or the reads-from
91 The following list describes each of the columns in the execution trace output:
93 * \#: The sequence number within the execution. That is, sequence number "9"
94 means the operation was the 9th operation executed by C11Tester. Note that
95 this represents the execution order, not necessarily any other order (e.g.,
96 modification order or reads-from).
98 * t: The thread number
100 * Action type: The type of operation performed
102 * MO: The memory-order for this operation (i.e., `memory_order_XXX`, where `XXX` is
103 `relaxed`, `release`, `acquire`, `rel_acq`, or `seq_cst`)
105 * Location: The memory location on which this operation is operating. This is
106 well-defined for atomic write/read/RMW, but other operations are subject to
107 C11Tester implementation details.
109 * Value: For reads/writes/RMW, the value returned by the operation. Note that
110 for RMW, this is the value that is *read*, not the value that was *written*.
111 For other operations, 'value' may have some C11Tester-internal meaning, or
112 it may simply be a don't-care (such as `0xdeadbeef`).
114 * Rf: For reads, the sequence number of the operation from which it reads.
115 [Note: If the execution is a partial, infeasible trace (labeled INFEASIBLE),
116 as printed during `--verbose` execution, reads may not be resolved and so may
117 have Rf=? or Rf=Px, where x is a promised future value.]
119 * CV: The clock vector, encapsulating the happens-before relation (see our
120 paper, or the C/C++ memory model itself). We use a Lamport-style clock vector
121 similar to [1]. The "clock" is just the sequence number (#). The clock vector
122 can be read as follows:
124 Each entry is indexed as CV[i], where
126 i = 0, 1, 2, ..., <number of threads>
128 So for any thread i, we say CV[i] is the sequence number of the most recent
129 operation in thread i such that operation i happens-before this operation.
130 Notably, thread 0 is reserved as a dummy thread for certain C11Tester
133 See the following example trace:
135 ------------------------------------------------------------------------------------
136 # t Action type MO Location Value Rf CV
137 ------------------------------------------------------------------------------------
138 1 1 thread start seq_cst 0x7f68ff11e7c0 0xdeadbeef ( 0, 1)
139 2 1 init atomic relaxed 0x601068 0 ( 0, 2)
140 3 1 init atomic relaxed 0x60106c 0 ( 0, 3)
141 4 1 thread create seq_cst 0x7f68fe51c710 0x7f68fe51c6e0 ( 0, 4)
142 5 2 thread start seq_cst 0x7f68ff11ebc0 0xdeadbeef ( 0, 4, 5)
143 6 2 atomic read relaxed 0x60106c 0 3 ( 0, 4, 6)
144 7 1 thread create seq_cst 0x7f68fe51c720 0x7f68fe51c6e0 ( 0, 7)
145 8 3 thread start seq_cst 0x7f68ff11efc0 0xdeadbeef ( 0, 7, 0, 8)
146 9 2 atomic write relaxed 0x601068 0 ( 0, 4, 9)
147 10 3 atomic read relaxed 0x601068 0 2 ( 0, 7, 0, 10)
148 11 2 thread finish seq_cst 0x7f68ff11ebc0 0xdeadbeef ( 0, 4, 11)
149 12 3 atomic write relaxed 0x60106c 0x2a ( 0, 7, 0, 12)
150 13 1 thread join seq_cst 0x7f68ff11ebc0 0x2 ( 0, 13, 11)
151 14 3 thread finish seq_cst 0x7f68ff11efc0 0xdeadbeef ( 0, 7, 0, 14)
152 15 1 thread join seq_cst 0x7f68ff11efc0 0x3 ( 0, 15, 11, 14)
153 16 1 thread finish seq_cst 0x7f68ff11e7c0 0xdeadbeef ( 0, 16, 11, 14)
155 ------------------------------------------------------------------------------------
157 Now consider, for example, operation 10:
159 This is the 10th operation in the execution order. It is an atomic read-relaxed
160 operation performed by thread 3 at memory address `0x601068`. It reads the value
161 "0", which was written by the 2nd operation in the execution order. Its clock
162 vector consists of the following values:
164 CV[0] = 0, CV[1] = 7, CV[2] = 0, CV[3] = 10
166 End of Execution Summary
167 ------------------------
169 C11Tester prints summary statistics at the end of each execution. These
170 summaries are based off of a few different properties of an execution, which we
171 will break down here:
173 * A _buggy_ execution is an execution in which C11Tester has found a real
174 bug: a data race, a deadlock, failure of a user-provided assertion, or an
175 uninitialized load, for instance. C11Tester will only report bugs in feasible
179 Other Notes and Pitfalls
180 ------------------------
182 * Data races may be reported as multiple bugs, one for each byte-address of the
183 data race in question. See, for example, this run:
185 $ ./run.sh test/releaseseq.o
187 Bug report: 4 bugs detected
188 [BUG] Data race detected @ address 0x601078:
189 Access 1: write in thread 2 @ clock 4
190 Access 2: read in thread 3 @ clock 9
191 [BUG] Data race detected @ address 0x601079:
192 Access 1: write in thread 2 @ clock 4
193 Access 2: read in thread 3 @ clock 9
194 [BUG] Data race detected @ address 0x60107a:
195 Access 1: write in thread 2 @ clock 4
196 Access 2: read in thread 3 @ clock 9
197 [BUG] Data race detected @ address 0x60107b:
198 Access 1: write in thread 2 @ clock 4
199 Access 2: read in thread 3 @ clock 9
205 The C11Tester project page:
207 > <http://demsky.eecs.uci.edu/c11tester.html>
209 The C11Tester source and accompanying benchmarks on Gitweb:
211 > <http://plrg.eecs.uci.edu/git/?p=c11tester.git>
213 > <http://plrg.eecs.uci.edu/git/?p=c11llvm.git>
215 > <http://plrg.eecs.uci.edu/git/?p=c11concurrency-benchmarks.git>
222 Please feel free to contact us for more information. Bug reports are welcome,
223 and we are happy to hear from our users. We are also very interested to know if
224 C11Tester catches bugs in your programs.
226 Contact Weiyu Luo at <weiyul7@uci.edu> or Brian Demsky at <bdemsky@uci.edu>.
232 Copyright © 2013 and 2019 Regents of the University of California. All rights reserved.
234 C11Tester is distributed under the GPL v2. See the LICENSE file for details.
240 [1] L. Lamport. Time, clocks, and the ordering of events in a distributed
241 system. CACM, 21(7):558-565, July 1978.