kernel/trace/trace_irqsoff.c

   1 /*
   2  * trace irqs off critical timings
   3  *
   4  * Copyright (C) 2007-2008 Steven Rostedt <srostedt@redhat.com>
   5  * Copyright (C) 2008 Ingo Molnar <mingo@redhat.com>
   6  *
   7  * From code in the latency_tracer, that is:
   8  *
   9  *  Copyright (C) 2004-2006 Ingo Molnar
  10  *  Copyright (C) 2004 Nadia Yvette Chambers
  11  */
  12 #include <linux/kallsyms.h>
  13 #include <linux/debugfs.h>
  14 #include <linux/uaccess.h>
  15 #include <linux/module.h>
  16 #include <linux/ftrace.h>
  17 #include <linux/fs.h>
  18
  19 #include "trace.h"
  20
  21 static struct trace_array               *irqsoff_trace __read_mostly;
  22 static int                              tracer_enabled __read_mostly;
  23
  24 static DEFINE_PER_CPU(int, tracing_cpu);
  25
  26 static DEFINE_RAW_SPINLOCK(max_trace_lock);
  27
  28 enum {
  29         TRACER_IRQS_OFF         = (1 << 1),
  30         TRACER_PREEMPT_OFF      = (1 << 2),
  31 };
  32
  33 static int trace_type __read_mostly;
  34
  35 static int save_lat_flag;
  36
  37 static void stop_irqsoff_tracer(struct trace_array *tr, int graph);
  38 static int start_irqsoff_tracer(struct trace_array *tr, int graph);
  39
  40 #ifdef CONFIG_PREEMPT_TRACER
  41 static inline int
  42 preempt_trace(void)
  43 {
  44         return ((trace_type & TRACER_PREEMPT_OFF) && preempt_count());
  45 }
  46 #else
  47 # define preempt_trace() (0)
  48 #endif
  49
  50 #ifdef CONFIG_IRQSOFF_TRACER
  51 static inline int
  52 irq_trace(void)
  53 {
  54         return ((trace_type & TRACER_IRQS_OFF) &&
  55                 irqs_disabled());
  56 }
  57 #else
  58 # define irq_trace() (0)
  59 #endif
  60
  61 #define TRACE_DISPLAY_GRAPH     1
  62
  63 static struct tracer_opt trace_opts[] = {
  64 #ifdef CONFIG_FUNCTION_GRAPH_TRACER
  65         /* display latency trace as call graph */
  66         { TRACER_OPT(display-graph, TRACE_DISPLAY_GRAPH) },
  67 #endif
  68         { } /* Empty entry */
  69 };
  70
  71 static struct tracer_flags tracer_flags = {
  72         .val  = 0,
  73         .opts = trace_opts,
  74 };
  75
  76 #define is_graph() (tracer_flags.val & TRACE_DISPLAY_GRAPH)
  77
  78 /*
  79  * Sequence count - we record it when starting a measurement and
  80  * skip the latency if the sequence has changed - some other section
  81  * did a maximum and could disturb our measurement with serial console
  82  * printouts, etc. Truly coinciding maximum latencies should be rare
  83  * and what happens together happens separately as well, so this doesn't
  84  * decrease the validity of the maximum found:
  85  */
  86 static __cacheline_aligned_in_smp       unsigned long max_sequence;
  87
  88 #ifdef CONFIG_FUNCTION_TRACER
  89 /*
  90  * Prologue for the preempt and irqs off function tracers.
  91  *
  92  * Returns 1 if it is OK to continue, and data->disabled is
  93  *            incremented.
  94  *         0 if the trace is to be ignored, and data->disabled
  95  *            is kept the same.
  96  *
  97  * Note, this function is also used outside this ifdef but
  98  *  inside the #ifdef of the function graph tracer below.
  99  *  This is OK, since the function graph tracer is
 100  *  dependent on the function tracer.
 101  */
 102 static int func_prolog_dec(struct trace_array *tr,
 103                            struct trace_array_cpu **data,
 104                            unsigned long *flags)
 105 {
 106         long disabled;
 107         int cpu;
 108
 109         /*
 110          * Does not matter if we preempt. We test the flags
 111          * afterward, to see if irqs are disabled or not.
 112          * If we preempt and get a false positive, the flags
 113          * test will fail.
 114          */
 115         cpu = raw_smp_processor_id();
 116         if (likely(!per_cpu(tracing_cpu, cpu)))
 117                 return 0;
 118
 119         local_save_flags(*flags);
 120         /* slight chance to get a false positive on tracing_cpu */
 121         if (!irqs_disabled_flags(*flags))
 122                 return 0;
 123
 124         *data = tr->data[cpu];
 125         disabled = atomic_inc_return(&(*data)->disabled);
 126
 127         if (likely(disabled == 1))
 128                 return 1;
 129
 130         atomic_dec(&(*data)->disabled);
 131
 132         return 0;
 133 }
 134
 135 /*
 136  * irqsoff uses its own tracer function to keep the overhead down:
 137  */
 138 static void
 139 irqsoff_tracer_call(unsigned long ip, unsigned long parent_ip,
 140                     struct ftrace_ops *op, struct pt_regs *pt_regs)
 141 {
 142         struct trace_array *tr = irqsoff_trace;
 143         struct trace_array_cpu *data;
 144         unsigned long flags;
 145
 146         if (!func_prolog_dec(tr, &data, &flags))
 147                 return;
 148
 149         trace_function(tr, ip, parent_ip, flags, preempt_count());
 150
 151         atomic_dec(&data->disabled);
 152 }
 153
 154 static struct ftrace_ops trace_ops __read_mostly =
 155 {
 156         .func = irqsoff_tracer_call,
 157         .flags = FTRACE_OPS_FL_GLOBAL | FTRACE_OPS_FL_RECURSION_SAFE,
 158 };
 159 #endif /* CONFIG_FUNCTION_TRACER */
 160
 161 #ifdef CONFIG_FUNCTION_GRAPH_TRACER
 162 static int irqsoff_set_flag(u32 old_flags, u32 bit, int set)
 163 {
 164         int cpu;
 165
 166         if (!(bit & TRACE_DISPLAY_GRAPH))
 167                 return -EINVAL;
 168
 169         if (!(is_graph() ^ set))
 170                 return 0;
 171
 172         stop_irqsoff_tracer(irqsoff_trace, !set);
 173
 174         for_each_possible_cpu(cpu)
 175                 per_cpu(tracing_cpu, cpu) = 0;
 176
 177         tracing_max_latency = 0;
 178         tracing_reset_online_cpus(irqsoff_trace);
 179
 180         return start_irqsoff_tracer(irqsoff_trace, set);
 181 }
 182
 183 static int irqsoff_graph_entry(struct ftrace_graph_ent *trace)
 184 {
 185         struct trace_array *tr = irqsoff_trace;
 186         struct trace_array_cpu *data;
 187         unsigned long flags;
 188         int ret;
 189         int pc;
 190
 191         if (!func_prolog_dec(tr, &data, &flags))
 192                 return 0;
 193
 194         pc = preempt_count();
 195         ret = __trace_graph_entry(tr, trace, flags, pc);
 196         atomic_dec(&data->disabled);
 197
 198         return ret;
 199 }
 200
 201 static void irqsoff_graph_return(struct ftrace_graph_ret *trace)
 202 {
 203         struct trace_array *tr = irqsoff_trace;
 204         struct trace_array_cpu *data;
 205         unsigned long flags;
 206         int pc;
 207
 208         if (!func_prolog_dec(tr, &data, &flags))
 209                 return;
 210
 211         pc = preempt_count();
 212         __trace_graph_return(tr, trace, flags, pc);
 213         atomic_dec(&data->disabled);
 214 }
 215
 216 static void irqsoff_trace_open(struct trace_iterator *iter)
 217 {
 218         if (is_graph())
 219                 graph_trace_open(iter);
 220
 221 }
 222
 223 static void irqsoff_trace_close(struct trace_iterator *iter)
 224 {
 225         if (iter->private)
 226                 graph_trace_close(iter);
 227 }
 228
 229 #define GRAPH_TRACER_FLAGS (TRACE_GRAPH_PRINT_CPU | \
 230                             TRACE_GRAPH_PRINT_PROC | \
 231                             TRACE_GRAPH_PRINT_ABS_TIME | \
 232                             TRACE_GRAPH_PRINT_DURATION)
 233
 234 static enum print_line_t irqsoff_print_line(struct trace_iterator *iter)
 235 {
 236         /*
 237          * In graph mode call the graph tracer output function,
 238          * otherwise go with the TRACE_FN event handler
 239          */
 240         if (is_graph())
 241                 return print_graph_function_flags(iter, GRAPH_TRACER_FLAGS);
 242
 243         return TRACE_TYPE_UNHANDLED;
 244 }
 245
 246 static void irqsoff_print_header(struct seq_file *s)
 247 {
 248         if (is_graph())
 249                 print_graph_headers_flags(s, GRAPH_TRACER_FLAGS);
 250         else
 251                 trace_default_header(s);
 252 }
 253
 254 static void
 255 __trace_function(struct trace_array *tr,
 256                  unsigned long ip, unsigned long parent_ip,
 257                  unsigned long flags, int pc)
 258 {
 259         if (is_graph())
 260                 trace_graph_function(tr, ip, parent_ip, flags, pc);
 261         else
 262                 trace_function(tr, ip, parent_ip, flags, pc);
 263 }
 264
 265 #else
 266 #define __trace_function trace_function
 267
 268 static int irqsoff_set_flag(u32 old_flags, u32 bit, int set)
 269 {
 270         return -EINVAL;
 271 }
 272
 273 static int irqsoff_graph_entry(struct ftrace_graph_ent *trace)
 274 {
 275         return -1;
 276 }
 277
 278 static enum print_line_t irqsoff_print_line(struct trace_iterator *iter)
 279 {
 280         return TRACE_TYPE_UNHANDLED;
 281 }
 282
 283 static void irqsoff_graph_return(struct ftrace_graph_ret *trace) { }
 284 static void irqsoff_trace_open(struct trace_iterator *iter) { }
 285 static void irqsoff_trace_close(struct trace_iterator *iter) { }
 286
 287 #ifdef CONFIG_FUNCTION_TRACER
 288 static void irqsoff_print_header(struct seq_file *s)
 289 {
 290         trace_default_header(s);
 291 }
 292 #else
 293 static void irqsoff_print_header(struct seq_file *s)
 294 {
 295         trace_latency_header(s);
 296 }
 297 #endif /* CONFIG_FUNCTION_TRACER */
 298 #endif /* CONFIG_FUNCTION_GRAPH_TRACER */
 299
 300 /*
 301  * Should this new latency be reported/recorded?
 302  */
 303 static int report_latency(cycle_t delta)
 304 {
 305         if (tracing_thresh) {
 306                 if (delta < tracing_thresh)
 307                         return 0;
 308         } else {
 309                 if (delta <= tracing_max_latency)
 310                         return 0;
 311         }
 312         return 1;
 313 }
 314
 315 static void
 316 check_critical_timing(struct trace_array *tr,
 317                       struct trace_array_cpu *data,
 318                       unsigned long parent_ip,
 319                       int cpu)
 320 {
 321         cycle_t T0, T1, delta;
 322         unsigned long flags;
 323         int pc;
 324
 325         T0 = data->preempt_timestamp;
 326         T1 = ftrace_now(cpu);
 327         delta = T1-T0;
 328
 329         local_save_flags(flags);
 330
 331         pc = preempt_count();
 332
 333         if (!report_latency(delta))
 334                 goto out;
 335
 336         raw_spin_lock_irqsave(&max_trace_lock, flags);
 337
 338         /* check if we are still the max latency */
 339         if (!report_latency(delta))
 340                 goto out_unlock;
 341
 342         __trace_function(tr, CALLER_ADDR0, parent_ip, flags, pc);
 343         /* Skip 5 functions to get to the irq/preempt enable function */
 344         __trace_stack(tr, flags, 5, pc);
 345
 346         if (data->critical_sequence != max_sequence)
 347                 goto out_unlock;
 348
 349         data->critical_end = parent_ip;
 350
 351         if (likely(!is_tracing_stopped())) {
 352                 tracing_max_latency = delta;
 353                 update_max_tr_single(tr, current, cpu);
 354         }
 355
 356         max_sequence++;
 357
 358 out_unlock:
 359         raw_spin_unlock_irqrestore(&max_trace_lock, flags);
 360
 361 out:
 362         data->critical_sequence = max_sequence;
 363         data->preempt_timestamp = ftrace_now(cpu);
 364         __trace_function(tr, CALLER_ADDR0, parent_ip, flags, pc);
 365 }
 366
 367 static inline void
 368 start_critical_timing(unsigned long ip, unsigned long parent_ip)
 369 {
 370         int cpu;
 371         struct trace_array *tr = irqsoff_trace;
 372         struct trace_array_cpu *data;
 373         unsigned long flags;
 374
 375         if (likely(!tracer_enabled))
 376                 return;
 377
 378         cpu = raw_smp_processor_id();
 379
 380         if (per_cpu(tracing_cpu, cpu))
 381                 return;
 382
 383         data = tr->data[cpu];
 384
 385         if (unlikely(!data) || atomic_read(&data->disabled))
 386                 return;
 387
 388         atomic_inc(&data->disabled);
 389
 390         data->critical_sequence = max_sequence;
 391         data->preempt_timestamp = ftrace_now(cpu);
 392         data->critical_start = parent_ip ? : ip;
 393
 394         local_save_flags(flags);
 395
 396         __trace_function(tr, ip, parent_ip, flags, preempt_count());
 397
 398         per_cpu(tracing_cpu, cpu) = 1;
 399
 400         atomic_dec(&data->disabled);
 401 }
 402
 403 static inline void
 404 stop_critical_timing(unsigned long ip, unsigned long parent_ip)
 405 {
 406         int cpu;
 407         struct trace_array *tr = irqsoff_trace;
 408         struct trace_array_cpu *data;
 409         unsigned long flags;
 410
 411         cpu = raw_smp_processor_id();
 412         /* Always clear the tracing cpu on stopping the trace */
 413         if (unlikely(per_cpu(tracing_cpu, cpu)))
 414                 per_cpu(tracing_cpu, cpu) = 0;
 415         else
 416                 return;
 417
 418         if (!tracer_enabled)
 419                 return;
 420
 421         data = tr->data[cpu];
 422
 423         if (unlikely(!data) ||
 424             !data->critical_start || atomic_read(&data->disabled))
 425                 return;
 426
 427         atomic_inc(&data->disabled);
 428
 429         local_save_flags(flags);
 430         __trace_function(tr, ip, parent_ip, flags, preempt_count());
 431         check_critical_timing(tr, data, parent_ip ? : ip, cpu);
 432         data->critical_start = 0;
 433         atomic_dec(&data->disabled);
 434 }
 435
 436 /* start and stop critical timings used to for stoppage (in idle) */
 437 void start_critical_timings(void)
 438 {
 439         if (preempt_trace() || irq_trace())
 440                 start_critical_timing(CALLER_ADDR0, CALLER_ADDR1);
 441 }
 442 EXPORT_SYMBOL_GPL(start_critical_timings);
 443
 444 void stop_critical_timings(void)
 445 {
 446         if (preempt_trace() || irq_trace())
 447                 stop_critical_timing(CALLER_ADDR0, CALLER_ADDR1);
 448 }
 449 EXPORT_SYMBOL_GPL(stop_critical_timings);
 450
 451 #ifdef CONFIG_IRQSOFF_TRACER
 452 #ifdef CONFIG_PROVE_LOCKING
 453 void time_hardirqs_on(unsigned long a0, unsigned long a1)
 454 {
 455         if (!preempt_trace() && irq_trace())
 456                 stop_critical_timing(a0, a1);
 457 }
 458
 459 void time_hardirqs_off(unsigned long a0, unsigned long a1)
 460 {
 461         if (!preempt_trace() && irq_trace())
 462                 start_critical_timing(a0, a1);
 463 }
 464
 465 #else /* !CONFIG_PROVE_LOCKING */
 466
 467 /*
 468  * Stubs:
 469  */
 470
 471 void trace_softirqs_on(unsigned long ip)
 472 {
 473 }
 474
 475 void trace_softirqs_off(unsigned long ip)
 476 {
 477 }
 478
 479 inline void print_irqtrace_events(struct task_struct *curr)
 480 {
 481 }
 482
 483 /*
 484  * We are only interested in hardirq on/off events:
 485  */
 486 void trace_hardirqs_on(void)
 487 {
 488         if (!preempt_trace() && irq_trace())
 489                 stop_critical_timing(CALLER_ADDR0, CALLER_ADDR1);
 490 }
 491 EXPORT_SYMBOL(trace_hardirqs_on);
 492
 493 void trace_hardirqs_off(void)
 494 {
 495         if (!preempt_trace() && irq_trace())
 496                 start_critical_timing(CALLER_ADDR0, CALLER_ADDR1);
 497 }
 498 EXPORT_SYMBOL(trace_hardirqs_off);
 499
 500 void trace_hardirqs_on_caller(unsigned long caller_addr)
 501 {
 502         if (!preempt_trace() && irq_trace())
 503                 stop_critical_timing(CALLER_ADDR0, caller_addr);
 504 }
 505 EXPORT_SYMBOL(trace_hardirqs_on_caller);
 506
 507 void trace_hardirqs_off_caller(unsigned long caller_addr)
 508 {
 509         if (!preempt_trace() && irq_trace())
 510                 start_critical_timing(CALLER_ADDR0, caller_addr);
 511 }
 512 EXPORT_SYMBOL(trace_hardirqs_off_caller);
 513
 514 #endif /* CONFIG_PROVE_LOCKING */
 515 #endif /*  CONFIG_IRQSOFF_TRACER */
 516
 517 #ifdef CONFIG_PREEMPT_TRACER
 518 void trace_preempt_on(unsigned long a0, unsigned long a1)
 519 {
 520         if (preempt_trace() && !irq_trace())
 521                 stop_critical_timing(a0, a1);
 522 }
 523
 524 void trace_preempt_off(unsigned long a0, unsigned long a1)
 525 {
 526         if (preempt_trace() && !irq_trace())
 527                 start_critical_timing(a0, a1);
 528 }
 529 #endif /* CONFIG_PREEMPT_TRACER */
 530
 531 static int start_irqsoff_tracer(struct trace_array *tr, int graph)
 532 {
 533         int ret = 0;
 534
 535         if (!graph)
 536                 ret = register_ftrace_function(&trace_ops);
 537         else
 538                 ret = register_ftrace_graph(&irqsoff_graph_return,
 539                                             &irqsoff_graph_entry);
 540
 541         if (!ret && tracing_is_enabled())
 542                 tracer_enabled = 1;
 543         else
 544                 tracer_enabled = 0;
 545
 546         return ret;
 547 }
 548
 549 static void stop_irqsoff_tracer(struct trace_array *tr, int graph)
 550 {
 551         tracer_enabled = 0;
 552
 553         if (!graph)
 554                 unregister_ftrace_function(&trace_ops);
 555         else
 556                 unregister_ftrace_graph();
 557 }
 558
 559 static void __irqsoff_tracer_init(struct trace_array *tr)
 560 {
 561         save_lat_flag = trace_flags & TRACE_ITER_LATENCY_FMT;
 562         trace_flags |= TRACE_ITER_LATENCY_FMT;
 563
 564         tracing_max_latency = 0;
 565         irqsoff_trace = tr;
 566         /* make sure that the tracer is visible */
 567         smp_wmb();
 568         tracing_reset_online_cpus(tr);
 569
 570         if (start_irqsoff_tracer(tr, is_graph()))
 571                 printk(KERN_ERR "failed to start irqsoff tracer\n");
 572 }
 573
 574 static void irqsoff_tracer_reset(struct trace_array *tr)
 575 {
 576         stop_irqsoff_tracer(tr, is_graph());
 577
 578         if (!save_lat_flag)
 579                 trace_flags &= ~TRACE_ITER_LATENCY_FMT;
 580 }
 581
 582 static void irqsoff_tracer_start(struct trace_array *tr)
 583 {
 584         tracer_enabled = 1;
 585 }
 586
 587 static void irqsoff_tracer_stop(struct trace_array *tr)
 588 {
 589         tracer_enabled = 0;
 590 }
 591
 592 #ifdef CONFIG_IRQSOFF_TRACER
 593 static int irqsoff_tracer_init(struct trace_array *tr)
 594 {
 595         trace_type = TRACER_IRQS_OFF;
 596
 597         __irqsoff_tracer_init(tr);
 598         return 0;
 599 }
 600 static struct tracer irqsoff_tracer __read_mostly =
 601 {
 602         .name           = "irqsoff",
 603         .init           = irqsoff_tracer_init,
 604         .reset          = irqsoff_tracer_reset,
 605         .start          = irqsoff_tracer_start,
 606         .stop           = irqsoff_tracer_stop,
 607         .print_max      = true,
 608         .print_header   = irqsoff_print_header,
 609         .print_line     = irqsoff_print_line,
 610         .flags          = &tracer_flags,
 611         .set_flag       = irqsoff_set_flag,
 612 #ifdef CONFIG_FTRACE_SELFTEST
 613         .selftest    = trace_selftest_startup_irqsoff,
 614 #endif
 615         .open           = irqsoff_trace_open,
 616         .close          = irqsoff_trace_close,
 617         .use_max_tr     = true,
 618 };
 619 # define register_irqsoff(trace) register_tracer(&trace)
 620 #else
 621 # define register_irqsoff(trace) do { } while (0)
 622 #endif
 623
 624 #ifdef CONFIG_PREEMPT_TRACER
 625 static int preemptoff_tracer_init(struct trace_array *tr)
 626 {
 627         trace_type = TRACER_PREEMPT_OFF;
 628
 629         __irqsoff_tracer_init(tr);
 630         return 0;
 631 }
 632
 633 static struct tracer preemptoff_tracer __read_mostly =
 634 {
 635         .name           = "preemptoff",
 636         .init           = preemptoff_tracer_init,
 637         .reset          = irqsoff_tracer_reset,
 638         .start          = irqsoff_tracer_start,
 639         .stop           = irqsoff_tracer_stop,
 640         .print_max      = true,
 641         .print_header   = irqsoff_print_header,
 642         .print_line     = irqsoff_print_line,
 643         .flags          = &tracer_flags,
 644         .set_flag       = irqsoff_set_flag,
 645 #ifdef CONFIG_FTRACE_SELFTEST
 646         .selftest    = trace_selftest_startup_preemptoff,
 647 #endif
 648         .open           = irqsoff_trace_open,
 649         .close          = irqsoff_trace_close,
 650         .use_max_tr     = true,
 651 };
 652 # define register_preemptoff(trace) register_tracer(&trace)
 653 #else
 654 # define register_preemptoff(trace) do { } while (0)
 655 #endif
 656
 657 #if defined(CONFIG_IRQSOFF_TRACER) && \
 658         defined(CONFIG_PREEMPT_TRACER)
 659
 660 static int preemptirqsoff_tracer_init(struct trace_array *tr)
 661 {
 662         trace_type = TRACER_IRQS_OFF | TRACER_PREEMPT_OFF;
 663
 664         __irqsoff_tracer_init(tr);
 665         return 0;
 666 }
 667
 668 static struct tracer preemptirqsoff_tracer __read_mostly =
 669 {
 670         .name           = "preemptirqsoff",
 671         .init           = preemptirqsoff_tracer_init,
 672         .reset          = irqsoff_tracer_reset,
 673         .start          = irqsoff_tracer_start,
 674         .stop           = irqsoff_tracer_stop,
 675         .print_max      = true,
 676         .print_header   = irqsoff_print_header,
 677         .print_line     = irqsoff_print_line,
 678         .flags          = &tracer_flags,
 679         .set_flag       = irqsoff_set_flag,
 680 #ifdef CONFIG_FTRACE_SELFTEST
 681         .selftest    = trace_selftest_startup_preemptirqsoff,
 682 #endif
 683         .open           = irqsoff_trace_open,
 684         .close          = irqsoff_trace_close,
 685         .use_max_tr     = true,
 686 };
 687
 688 # define register_preemptirqsoff(trace) register_tracer(&trace)
 689 #else
 690 # define register_preemptirqsoff(trace) do { } while (0)
 691 #endif
 692
 693 __init static int init_irqsoff_tracer(void)
 694 {
 695         register_irqsoff(irqsoff_tracer);
 696         register_preemptoff(preemptoff_tracer);
 697         register_preemptirqsoff(preemptirqsoff_tracer);
 698
 699         return 0;
 700 }
 701 core_initcall(init_irqsoff_tracer);