1 /*===- DataFlow.cpp - a standalone DataFlow tracer                  -------===//
2 //
3 //                     The LLVM Compiler Infrastructure
4 //
5 // This file is distributed under the University of Illinois Open Source
6 // License. See LICENSE.TXT for details.
7 //
8 //===----------------------------------------------------------------------===//
9 // An experimental data-flow tracer for fuzz targets.
10 // It is based on DFSan and SanitizerCoverage.
11 // https://clang.llvm.org/docs/DataFlowSanitizer.html
12 // https://clang.llvm.org/docs/SanitizerCoverage.html#tracing-data-flow
13 //
14 // It executes the fuzz target on the given input while monitoring the
15 // data flow for every instrumented comparison instruction.
16 //
17 // The output shows which functions depend on which bytes of the input.
18 //
19 // Build:
20 //   1. Compile this file with -fsanitize=dataflow
21 //   2. Build the fuzz target with -g -fsanitize=dataflow
22 //       -fsanitize-coverage=trace-pc-guard,pc-table,func,trace-cmp
23 //   3. Link those together with -fsanitize=dataflow
24 //
25 //  -fsanitize-coverage=trace-cmp inserts callbacks around every comparison
26 //  instruction, DFSan modifies the calls to pass the data flow labels.
27 //  The callbacks update the data flow label for the current function.
28 //  See e.g. __dfsw___sanitizer_cov_trace_cmp1 below.
29 //
30 //  -fsanitize-coverage=trace-pc-guard,pc-table,func instruments function
31 //  entries so that the comparison callback knows that current function.
32 //
33 //
34 // Run:
35 //   # Collect data flow for INPUT_FILE, write to OUTPUT_FILE (default: stdout)
36 //   ./a.out INPUT_FILE [OUTPUT_FILE]
37 //
38 //   # Print all instrumented functions. llvm-symbolizer must be present in PATH
39 //   ./a.out
40 //
41 // Example output:
42 // ===============
43 //  F0 11111111111111
44 //  F1 10000000000000
45 //  ===============
46 // "FN xxxxxxxxxx": tells what bytes of the input does the function N depend on.
47 //    The byte string is LEN+1 bytes. The last byte is set if the function
48 //    depends on the input length.
49 //===----------------------------------------------------------------------===*/
50 
51 #include <assert.h>
52 #include <stdio.h>
53 #include <stdlib.h>
54 #include <stdint.h>
55 #include <string.h>
56 
57 #include <execinfo.h>  // backtrace_symbols_fd
58 
59 #include <sanitizer/dfsan_interface.h>
60 
61 extern "C" {
62 extern int LLVMFuzzerTestOneInput(const unsigned char *Data, size_t Size);
63 __attribute__((weak)) extern int LLVMFuzzerInitialize(int *argc, char ***argv);
64 } // extern "C"
65 
66 static size_t InputLen;
67 static size_t NumFuncs;
68 static const uintptr_t *FuncsBeg;
69 static __thread size_t CurrentFunc;
70 static dfsan_label *FuncLabels;  // Array of NumFuncs elements.
71 static char *PrintableStringForLabel;  // InputLen + 2 bytes.
72 
73 // Prints all instrumented functions.
74 static int PrintFunctions() {
75   // We don't have the symbolizer integrated with dfsan yet.
76   // So use backtrace_symbols_fd and pipe it through llvm-symbolizer.
77   // TODO(kcc): this is pretty ugly and may break in lots of ways.
78   //      We'll need to make a proper in-process symbolizer work with DFSan.
79   FILE *Pipe = popen("sed 's/(+/ /g; s/).*//g' "
80                      "| llvm-symbolizer "
81                      "| grep 'dfs\\$' "
82                      "| sed 's/dfs\\$//g'", "w");
83   for (size_t I = 0; I < NumFuncs; I++) {
84     uintptr_t PC = FuncsBeg[I * 2];
85     void *const Buf[1] = {(void*)PC};
86     backtrace_symbols_fd(Buf, 1, fileno(Pipe));
87   }
88   pclose(Pipe);
89   return 0;
90 }
91 
92 static void SetBytesForLabel(dfsan_label L, char *Bytes) {
93   assert(L);
94   if (L <= InputLen + 1) {
95     Bytes[L - 1] = '1';
96   } else {
97     auto *DLI = dfsan_get_label_info(L);
98     SetBytesForLabel(DLI->l1, Bytes);
99     SetBytesForLabel(DLI->l2, Bytes);
100   }
101 }
102 
103 static char *GetPrintableStringForLabel(dfsan_label L) {
104   memset(PrintableStringForLabel, '0', InputLen + 1);
105   PrintableStringForLabel[InputLen + 1] = 0;
106   SetBytesForLabel(L, PrintableStringForLabel);
107   return PrintableStringForLabel;
108 }
109 
110 static void PrintDataFlow(FILE *Out) {
111   for (size_t I = 0; I < NumFuncs; I++)
112     if (FuncLabels[I])
113       fprintf(Out, "F%zd %s\n", I, GetPrintableStringForLabel(FuncLabels[I]));
114 }
115 
116 int main(int argc, char **argv) {
117   if (LLVMFuzzerInitialize)
118     LLVMFuzzerInitialize(&argc, &argv);
119   if (argc == 1)
120     return PrintFunctions();
121   assert(argc == 2 || argc == 3);
122 
123   const char *Input = argv[1];
124   fprintf(stderr, "INFO: reading '%s'\n", Input);
125   FILE *In = fopen(Input, "r");
126   assert(In);
127   fseek(In, 0, SEEK_END);
128   InputLen = ftell(In);
129   fseek(In, 0, SEEK_SET);
130   unsigned char *Buf = (unsigned char*)malloc(InputLen);
131   size_t NumBytesRead = fread(Buf, 1, InputLen, In);
132   assert(NumBytesRead == InputLen);
133   PrintableStringForLabel = (char*)malloc(InputLen + 2);
134   fclose(In);
135 
136   fprintf(stderr, "INFO: running '%s'\n", Input);
137   for (size_t I = 1; I <= InputLen; I++) {
138     dfsan_label L = dfsan_create_label("", nullptr);
139     assert(L == I);
140     dfsan_set_label(L, Buf + I - 1, 1);
141   }
142   dfsan_label SizeL = dfsan_create_label("", nullptr);
143   assert(SizeL == InputLen + 1);
144   dfsan_set_label(SizeL, &InputLen, sizeof(InputLen));
145 
146   LLVMFuzzerTestOneInput(Buf, InputLen);
147   free(Buf);
148 
149   bool OutIsStdout = argc == 2;
150   fprintf(stderr, "INFO: writing dataflow to %s\n",
151           OutIsStdout ? "<stdout>" : argv[2]);
152   FILE *Out = OutIsStdout ? stdout : fopen(argv[2], "w");
153   PrintDataFlow(Out);
154   if (!OutIsStdout) fclose(Out);
155 }
156 
157 extern "C" {
158 
159 void __sanitizer_cov_trace_pc_guard_init(uint32_t *start,
160                                          uint32_t *stop) {
161   assert(NumFuncs == 0 && "This tool does not support DSOs");
162   assert(start < stop && "The code is not instrumented for coverage");
163   if (start == stop || *start) return;  // Initialize only once.
164   for (uint32_t *x = start; x < stop; x++)
165     *x = ++NumFuncs;  // The first index is 1.
166   FuncLabels = (dfsan_label*)calloc(NumFuncs, sizeof(dfsan_label));
167   fprintf(stderr, "INFO: %zd instrumented function(s) observed\n", NumFuncs);
168 }
169 
170 void __sanitizer_cov_pcs_init(const uintptr_t *pcs_beg,
171                               const uintptr_t *pcs_end) {
172   assert(NumFuncs == (pcs_end - pcs_beg) / 2);
173   FuncsBeg = pcs_beg;
174 }
175 
176 void __sanitizer_cov_trace_pc_indir(uint64_t x){}  // unused.
177 
178 void __sanitizer_cov_trace_pc_guard(uint32_t *guard){
179   uint32_t FuncNum = *guard - 1;  // Guards start from 1.
180   assert(FuncNum < NumFuncs);
181   CurrentFunc = FuncNum;
182 }
183 
184 void __dfsw___sanitizer_cov_trace_switch(uint64_t Val, uint64_t *Cases,
185                                          dfsan_label L1, dfsan_label UnusedL) {
186   assert(CurrentFunc < NumFuncs);
187   FuncLabels[CurrentFunc] = dfsan_union(FuncLabels[CurrentFunc], L1);
188 }
189 
190 #define HOOK(Name, Type)                                                       \
191   void Name(Type Arg1, Type Arg2, dfsan_label L1, dfsan_label L2) {            \
192     assert(CurrentFunc < NumFuncs);                                            \
193     FuncLabels[CurrentFunc] =                                                  \
194         dfsan_union(FuncLabels[CurrentFunc], dfsan_union(L1, L2));             \
195   }
196 
197 HOOK(__dfsw___sanitizer_cov_trace_const_cmp1, uint8_t)
198 HOOK(__dfsw___sanitizer_cov_trace_const_cmp2, uint16_t)
199 HOOK(__dfsw___sanitizer_cov_trace_const_cmp4, uint32_t)
200 HOOK(__dfsw___sanitizer_cov_trace_const_cmp8, uint64_t)
201 HOOK(__dfsw___sanitizer_cov_trace_cmp1, uint8_t)
202 HOOK(__dfsw___sanitizer_cov_trace_cmp2, uint16_t)
203 HOOK(__dfsw___sanitizer_cov_trace_cmp4, uint32_t)
204 HOOK(__dfsw___sanitizer_cov_trace_cmp8, uint64_t)
205 
206 } // extern "C"
207