1 /* 2 * Simple trace backend 3 * 4 * Copyright IBM, Corp. 2010 5 * 6 * This work is licensed under the terms of the GNU GPL, version 2. See 7 * the COPYING file in the top-level directory. 8 * 9 */ 10 11 #include <stdlib.h> 12 #include <stdint.h> 13 #include <stdio.h> 14 #include <time.h> 15 #ifndef _WIN32 16 #include <signal.h> 17 #include <pthread.h> 18 #endif 19 #include "qemu-timer.h" 20 #include "trace.h" 21 #include "trace/control.h" 22 23 /** Trace file header event ID */ 24 #define HEADER_EVENT_ID (~(uint64_t)0) /* avoids conflicting with TraceEventIDs */ 25 26 /** Trace file magic number */ 27 #define HEADER_MAGIC 0xf2b177cb0aa429b4ULL 28 29 /** Trace file version number, bump if format changes */ 30 #define HEADER_VERSION 0 31 32 /** Records were dropped event ID */ 33 #define DROPPED_EVENT_ID (~(uint64_t)0 - 1) 34 35 /** Trace record is valid */ 36 #define TRACE_RECORD_VALID ((uint64_t)1 << 63) 37 38 /** Trace buffer entry */ 39 typedef struct { 40 uint64_t event; 41 uint64_t timestamp_ns; 42 uint64_t x1; 43 uint64_t x2; 44 uint64_t x3; 45 uint64_t x4; 46 uint64_t x5; 47 uint64_t x6; 48 } TraceRecord; 49 50 enum { 51 TRACE_BUF_LEN = 4096, 52 TRACE_BUF_FLUSH_THRESHOLD = TRACE_BUF_LEN / 4, 53 }; 54 55 /* 56 * Trace records are written out by a dedicated thread. The thread waits for 57 * records to become available, writes them out, and then waits again. 58 */ 59 static GStaticMutex trace_lock = G_STATIC_MUTEX_INIT; 60 static GCond *trace_available_cond; 61 static GCond *trace_empty_cond; 62 static bool trace_available; 63 static bool trace_writeout_enabled; 64 65 static TraceRecord trace_buf[TRACE_BUF_LEN]; 66 static unsigned int trace_idx; 67 static FILE *trace_fp; 68 static char *trace_file_name = NULL; 69 70 /** 71 * Read a trace record from the trace buffer 72 * 73 * @idx Trace buffer index 74 * @record Trace record to fill 75 * 76 * Returns false if the record is not valid. 77 */ 78 static bool get_trace_record(unsigned int idx, TraceRecord *record) 79 { 80 if (!(trace_buf[idx].event & TRACE_RECORD_VALID)) { 81 return false; 82 } 83 84 __sync_synchronize(); /* read memory barrier before accessing record */ 85 86 *record = trace_buf[idx]; 87 record->event &= ~TRACE_RECORD_VALID; 88 return true; 89 } 90 91 /** 92 * Kick writeout thread 93 * 94 * @wait Whether to wait for writeout thread to complete 95 */ 96 static void flush_trace_file(bool wait) 97 { 98 g_static_mutex_lock(&trace_lock); 99 trace_available = true; 100 g_cond_signal(trace_available_cond); 101 102 if (wait) { 103 g_cond_wait(trace_empty_cond, g_static_mutex_get_mutex(&trace_lock)); 104 } 105 106 g_static_mutex_unlock(&trace_lock); 107 } 108 109 static void wait_for_trace_records_available(void) 110 { 111 g_static_mutex_lock(&trace_lock); 112 while (!(trace_available && trace_writeout_enabled)) { 113 g_cond_signal(trace_empty_cond); 114 g_cond_wait(trace_available_cond, 115 g_static_mutex_get_mutex(&trace_lock)); 116 } 117 trace_available = false; 118 g_static_mutex_unlock(&trace_lock); 119 } 120 121 static gpointer writeout_thread(gpointer opaque) 122 { 123 TraceRecord record; 124 unsigned int writeout_idx = 0; 125 unsigned int num_available, idx; 126 size_t unused __attribute__ ((unused)); 127 128 for (;;) { 129 wait_for_trace_records_available(); 130 131 num_available = trace_idx - writeout_idx; 132 if (num_available > TRACE_BUF_LEN) { 133 record = (TraceRecord){ 134 .event = DROPPED_EVENT_ID, 135 .x1 = num_available, 136 }; 137 unused = fwrite(&record, sizeof(record), 1, trace_fp); 138 writeout_idx += num_available; 139 } 140 141 idx = writeout_idx % TRACE_BUF_LEN; 142 while (get_trace_record(idx, &record)) { 143 trace_buf[idx].event = 0; /* clear valid bit */ 144 unused = fwrite(&record, sizeof(record), 1, trace_fp); 145 idx = ++writeout_idx % TRACE_BUF_LEN; 146 } 147 148 fflush(trace_fp); 149 } 150 return NULL; 151 } 152 153 static void trace(TraceEventID event, uint64_t x1, uint64_t x2, uint64_t x3, 154 uint64_t x4, uint64_t x5, uint64_t x6) 155 { 156 unsigned int idx; 157 uint64_t timestamp; 158 159 if (!trace_list[event].state) { 160 return; 161 } 162 163 timestamp = get_clock(); 164 165 idx = g_atomic_int_exchange_and_add((gint *)&trace_idx, 1) % TRACE_BUF_LEN; 166 trace_buf[idx] = (TraceRecord){ 167 .event = event, 168 .timestamp_ns = timestamp, 169 .x1 = x1, 170 .x2 = x2, 171 .x3 = x3, 172 .x4 = x4, 173 .x5 = x5, 174 .x6 = x6, 175 }; 176 __sync_synchronize(); /* write barrier before marking as valid */ 177 trace_buf[idx].event |= TRACE_RECORD_VALID; 178 179 if ((idx + 1) % TRACE_BUF_FLUSH_THRESHOLD == 0) { 180 flush_trace_file(false); 181 } 182 } 183 184 void trace0(TraceEventID event) 185 { 186 trace(event, 0, 0, 0, 0, 0, 0); 187 } 188 189 void trace1(TraceEventID event, uint64_t x1) 190 { 191 trace(event, x1, 0, 0, 0, 0, 0); 192 } 193 194 void trace2(TraceEventID event, uint64_t x1, uint64_t x2) 195 { 196 trace(event, x1, x2, 0, 0, 0, 0); 197 } 198 199 void trace3(TraceEventID event, uint64_t x1, uint64_t x2, uint64_t x3) 200 { 201 trace(event, x1, x2, x3, 0, 0, 0); 202 } 203 204 void trace4(TraceEventID event, uint64_t x1, uint64_t x2, uint64_t x3, uint64_t x4) 205 { 206 trace(event, x1, x2, x3, x4, 0, 0); 207 } 208 209 void trace5(TraceEventID event, uint64_t x1, uint64_t x2, uint64_t x3, uint64_t x4, uint64_t x5) 210 { 211 trace(event, x1, x2, x3, x4, x5, 0); 212 } 213 214 void trace6(TraceEventID event, uint64_t x1, uint64_t x2, uint64_t x3, uint64_t x4, uint64_t x5, uint64_t x6) 215 { 216 trace(event, x1, x2, x3, x4, x5, x6); 217 } 218 219 void st_set_trace_file_enabled(bool enable) 220 { 221 if (enable == !!trace_fp) { 222 return; /* no change */ 223 } 224 225 /* Halt trace writeout */ 226 flush_trace_file(true); 227 trace_writeout_enabled = false; 228 flush_trace_file(true); 229 230 if (enable) { 231 static const TraceRecord header = { 232 .event = HEADER_EVENT_ID, 233 .timestamp_ns = HEADER_MAGIC, 234 .x1 = HEADER_VERSION, 235 }; 236 237 trace_fp = fopen(trace_file_name, "wb"); 238 if (!trace_fp) { 239 return; 240 } 241 242 if (fwrite(&header, sizeof header, 1, trace_fp) != 1) { 243 fclose(trace_fp); 244 trace_fp = NULL; 245 return; 246 } 247 248 /* Resume trace writeout */ 249 trace_writeout_enabled = true; 250 flush_trace_file(false); 251 } else { 252 fclose(trace_fp); 253 trace_fp = NULL; 254 } 255 } 256 257 /** 258 * Set the name of a trace file 259 * 260 * @file The trace file name or NULL for the default name-<pid> set at 261 * config time 262 */ 263 bool st_set_trace_file(const char *file) 264 { 265 st_set_trace_file_enabled(false); 266 267 free(trace_file_name); 268 269 if (!file) { 270 if (asprintf(&trace_file_name, CONFIG_TRACE_FILE, getpid()) < 0) { 271 trace_file_name = NULL; 272 return false; 273 } 274 } else { 275 if (asprintf(&trace_file_name, "%s", file) < 0) { 276 trace_file_name = NULL; 277 return false; 278 } 279 } 280 281 st_set_trace_file_enabled(true); 282 return true; 283 } 284 285 void st_print_trace_file_status(FILE *stream, int (*stream_printf)(FILE *stream, const char *fmt, ...)) 286 { 287 stream_printf(stream, "Trace file \"%s\" %s.\n", 288 trace_file_name, trace_fp ? "on" : "off"); 289 } 290 291 void st_print_trace(FILE *stream, int (*stream_printf)(FILE *stream, const char *fmt, ...)) 292 { 293 unsigned int i; 294 295 for (i = 0; i < TRACE_BUF_LEN; i++) { 296 TraceRecord record; 297 298 if (!get_trace_record(i, &record)) { 299 continue; 300 } 301 stream_printf(stream, "Event %" PRIu64 " : %" PRIx64 " %" PRIx64 302 " %" PRIx64 " %" PRIx64 " %" PRIx64 " %" PRIx64 "\n", 303 record.event, record.x1, record.x2, 304 record.x3, record.x4, record.x5, 305 record.x6); 306 } 307 } 308 309 void st_flush_trace_buffer(void) 310 { 311 flush_trace_file(true); 312 } 313 314 void trace_print_events(FILE *stream, fprintf_function stream_printf) 315 { 316 unsigned int i; 317 318 for (i = 0; i < NR_TRACE_EVENTS; i++) { 319 stream_printf(stream, "%s [Event ID %u] : state %u\n", 320 trace_list[i].tp_name, i, trace_list[i].state); 321 } 322 } 323 324 bool trace_event_set_state(const char *name, bool state) 325 { 326 unsigned int i; 327 328 for (i = 0; i < NR_TRACE_EVENTS; i++) { 329 if (!strcmp(trace_list[i].tp_name, name)) { 330 trace_list[i].state = state; 331 return true; 332 } 333 } 334 return false; 335 } 336 337 /* Helper function to create a thread with signals blocked. Use glib's 338 * portable threads since QEMU abstractions cannot be used due to reentrancy in 339 * the tracer. Also note the signal masking on POSIX hosts so that the thread 340 * does not steal signals when the rest of the program wants them blocked. 341 */ 342 static GThread *trace_thread_create(GThreadFunc fn) 343 { 344 GThread *thread; 345 #ifndef _WIN32 346 sigset_t set, oldset; 347 348 sigfillset(&set); 349 pthread_sigmask(SIG_SETMASK, &set, &oldset); 350 #endif 351 thread = g_thread_create(writeout_thread, NULL, FALSE, NULL); 352 #ifndef _WIN32 353 pthread_sigmask(SIG_SETMASK, &oldset, NULL); 354 #endif 355 356 return thread; 357 } 358 359 bool trace_backend_init(const char *events, const char *file) 360 { 361 GThread *thread; 362 363 if (!g_thread_supported()) { 364 g_thread_init(NULL); 365 } 366 367 trace_available_cond = g_cond_new(); 368 trace_empty_cond = g_cond_new(); 369 370 thread = trace_thread_create(writeout_thread); 371 if (!thread) { 372 fprintf(stderr, "warning: unable to initialize simple trace backend\n"); 373 return false; 374 } 375 376 atexit(st_flush_trace_buffer); 377 trace_backend_init_events(events); 378 st_set_trace_file(file); 379 return true; 380 } 381