xref: /linux/tools/perf/util/thread-stack.c (revision da1d9caf95def6f0320819cf941c9fd1069ba9e1)
1 // SPDX-License-Identifier: GPL-2.0-only
2 /*
3  * thread-stack.c: Synthesize a thread's stack using call / return events
4  * Copyright (c) 2014, Intel Corporation.
5  */
6 
7 #include <linux/rbtree.h>
8 #include <linux/list.h>
9 #include <linux/log2.h>
10 #include <linux/zalloc.h>
11 #include <errno.h>
12 #include <stdlib.h>
13 #include <string.h>
14 #include "thread.h"
15 #include "event.h"
16 #include "machine.h"
17 #include "env.h"
18 #include "debug.h"
19 #include "symbol.h"
20 #include "comm.h"
21 #include "call-path.h"
22 #include "thread-stack.h"
23 
24 #define STACK_GROWTH 2048
25 
26 /*
27  * State of retpoline detection.
28  *
29  * RETPOLINE_NONE: no retpoline detection
30  * X86_RETPOLINE_POSSIBLE: x86 retpoline possible
31  * X86_RETPOLINE_DETECTED: x86 retpoline detected
32  */
33 enum retpoline_state_t {
34 	RETPOLINE_NONE,
35 	X86_RETPOLINE_POSSIBLE,
36 	X86_RETPOLINE_DETECTED,
37 };
38 
39 /**
40  * struct thread_stack_entry - thread stack entry.
41  * @ret_addr: return address
42  * @timestamp: timestamp (if known)
43  * @ref: external reference (e.g. db_id of sample)
44  * @branch_count: the branch count when the entry was created
45  * @insn_count: the instruction count when the entry was created
46  * @cyc_count the cycle count when the entry was created
47  * @db_id: id used for db-export
48  * @cp: call path
49  * @no_call: a 'call' was not seen
50  * @trace_end: a 'call' but trace ended
51  * @non_call: a branch but not a 'call' to the start of a different symbol
52  */
53 struct thread_stack_entry {
54 	u64 ret_addr;
55 	u64 timestamp;
56 	u64 ref;
57 	u64 branch_count;
58 	u64 insn_count;
59 	u64 cyc_count;
60 	u64 db_id;
61 	struct call_path *cp;
62 	bool no_call;
63 	bool trace_end;
64 	bool non_call;
65 };
66 
67 /**
68  * struct thread_stack - thread stack constructed from 'call' and 'return'
69  *                       branch samples.
70  * @stack: array that holds the stack
71  * @cnt: number of entries in the stack
72  * @sz: current maximum stack size
73  * @trace_nr: current trace number
74  * @branch_count: running branch count
75  * @insn_count: running  instruction count
76  * @cyc_count running  cycle count
77  * @kernel_start: kernel start address
78  * @last_time: last timestamp
79  * @crp: call/return processor
80  * @comm: current comm
81  * @arr_sz: size of array if this is the first element of an array
82  * @rstate: used to detect retpolines
83  * @br_stack_rb: branch stack (ring buffer)
84  * @br_stack_sz: maximum branch stack size
85  * @br_stack_pos: current position in @br_stack_rb
86  * @mispred_all: mark all branches as mispredicted
87  */
88 struct thread_stack {
89 	struct thread_stack_entry *stack;
90 	size_t cnt;
91 	size_t sz;
92 	u64 trace_nr;
93 	u64 branch_count;
94 	u64 insn_count;
95 	u64 cyc_count;
96 	u64 kernel_start;
97 	u64 last_time;
98 	struct call_return_processor *crp;
99 	struct comm *comm;
100 	unsigned int arr_sz;
101 	enum retpoline_state_t rstate;
102 	struct branch_stack *br_stack_rb;
103 	unsigned int br_stack_sz;
104 	unsigned int br_stack_pos;
105 	bool mispred_all;
106 };
107 
108 /*
109  * Assume pid == tid == 0 identifies the idle task as defined by
110  * perf_session__register_idle_thread(). The idle task is really 1 task per cpu,
111  * and therefore requires a stack for each cpu.
112  */
113 static inline bool thread_stack__per_cpu(struct thread *thread)
114 {
115 	return !(thread->tid || thread->pid_);
116 }
117 
118 static int thread_stack__grow(struct thread_stack *ts)
119 {
120 	struct thread_stack_entry *new_stack;
121 	size_t sz, new_sz;
122 
123 	new_sz = ts->sz + STACK_GROWTH;
124 	sz = new_sz * sizeof(struct thread_stack_entry);
125 
126 	new_stack = realloc(ts->stack, sz);
127 	if (!new_stack)
128 		return -ENOMEM;
129 
130 	ts->stack = new_stack;
131 	ts->sz = new_sz;
132 
133 	return 0;
134 }
135 
136 static int thread_stack__init(struct thread_stack *ts, struct thread *thread,
137 			      struct call_return_processor *crp,
138 			      bool callstack, unsigned int br_stack_sz)
139 {
140 	int err;
141 
142 	if (callstack) {
143 		err = thread_stack__grow(ts);
144 		if (err)
145 			return err;
146 	}
147 
148 	if (br_stack_sz) {
149 		size_t sz = sizeof(struct branch_stack);
150 
151 		sz += br_stack_sz * sizeof(struct branch_entry);
152 		ts->br_stack_rb = zalloc(sz);
153 		if (!ts->br_stack_rb)
154 			return -ENOMEM;
155 		ts->br_stack_sz = br_stack_sz;
156 	}
157 
158 	if (thread->maps && thread->maps->machine) {
159 		struct machine *machine = thread->maps->machine;
160 		const char *arch = perf_env__arch(machine->env);
161 
162 		ts->kernel_start = machine__kernel_start(machine);
163 		if (!strcmp(arch, "x86"))
164 			ts->rstate = X86_RETPOLINE_POSSIBLE;
165 	} else {
166 		ts->kernel_start = 1ULL << 63;
167 	}
168 	ts->crp = crp;
169 
170 	return 0;
171 }
172 
173 static struct thread_stack *thread_stack__new(struct thread *thread, int cpu,
174 					      struct call_return_processor *crp,
175 					      bool callstack,
176 					      unsigned int br_stack_sz)
177 {
178 	struct thread_stack *ts = thread->ts, *new_ts;
179 	unsigned int old_sz = ts ? ts->arr_sz : 0;
180 	unsigned int new_sz = 1;
181 
182 	if (thread_stack__per_cpu(thread) && cpu > 0)
183 		new_sz = roundup_pow_of_two(cpu + 1);
184 
185 	if (!ts || new_sz > old_sz) {
186 		new_ts = calloc(new_sz, sizeof(*ts));
187 		if (!new_ts)
188 			return NULL;
189 		if (ts)
190 			memcpy(new_ts, ts, old_sz * sizeof(*ts));
191 		new_ts->arr_sz = new_sz;
192 		zfree(&thread->ts);
193 		thread->ts = new_ts;
194 		ts = new_ts;
195 	}
196 
197 	if (thread_stack__per_cpu(thread) && cpu > 0 &&
198 	    (unsigned int)cpu < ts->arr_sz)
199 		ts += cpu;
200 
201 	if (!ts->stack &&
202 	    thread_stack__init(ts, thread, crp, callstack, br_stack_sz))
203 		return NULL;
204 
205 	return ts;
206 }
207 
208 static struct thread_stack *thread__cpu_stack(struct thread *thread, int cpu)
209 {
210 	struct thread_stack *ts = thread->ts;
211 
212 	if (cpu < 0)
213 		cpu = 0;
214 
215 	if (!ts || (unsigned int)cpu >= ts->arr_sz)
216 		return NULL;
217 
218 	ts += cpu;
219 
220 	if (!ts->stack)
221 		return NULL;
222 
223 	return ts;
224 }
225 
226 static inline struct thread_stack *thread__stack(struct thread *thread,
227 						    int cpu)
228 {
229 	if (!thread)
230 		return NULL;
231 
232 	if (thread_stack__per_cpu(thread))
233 		return thread__cpu_stack(thread, cpu);
234 
235 	return thread->ts;
236 }
237 
238 static int thread_stack__push(struct thread_stack *ts, u64 ret_addr,
239 			      bool trace_end)
240 {
241 	int err = 0;
242 
243 	if (ts->cnt == ts->sz) {
244 		err = thread_stack__grow(ts);
245 		if (err) {
246 			pr_warning("Out of memory: discarding thread stack\n");
247 			ts->cnt = 0;
248 		}
249 	}
250 
251 	ts->stack[ts->cnt].trace_end = trace_end;
252 	ts->stack[ts->cnt++].ret_addr = ret_addr;
253 
254 	return err;
255 }
256 
257 static void thread_stack__pop(struct thread_stack *ts, u64 ret_addr)
258 {
259 	size_t i;
260 
261 	/*
262 	 * In some cases there may be functions which are not seen to return.
263 	 * For example when setjmp / longjmp has been used.  Or the perf context
264 	 * switch in the kernel which doesn't stop and start tracing in exactly
265 	 * the same code path.  When that happens the return address will be
266 	 * further down the stack.  If the return address is not found at all,
267 	 * we assume the opposite (i.e. this is a return for a call that wasn't
268 	 * seen for some reason) and leave the stack alone.
269 	 */
270 	for (i = ts->cnt; i; ) {
271 		if (ts->stack[--i].ret_addr == ret_addr) {
272 			ts->cnt = i;
273 			return;
274 		}
275 	}
276 }
277 
278 static void thread_stack__pop_trace_end(struct thread_stack *ts)
279 {
280 	size_t i;
281 
282 	for (i = ts->cnt; i; ) {
283 		if (ts->stack[--i].trace_end)
284 			ts->cnt = i;
285 		else
286 			return;
287 	}
288 }
289 
290 static bool thread_stack__in_kernel(struct thread_stack *ts)
291 {
292 	if (!ts->cnt)
293 		return false;
294 
295 	return ts->stack[ts->cnt - 1].cp->in_kernel;
296 }
297 
298 static int thread_stack__call_return(struct thread *thread,
299 				     struct thread_stack *ts, size_t idx,
300 				     u64 timestamp, u64 ref, bool no_return)
301 {
302 	struct call_return_processor *crp = ts->crp;
303 	struct thread_stack_entry *tse;
304 	struct call_return cr = {
305 		.thread = thread,
306 		.comm = ts->comm,
307 		.db_id = 0,
308 	};
309 	u64 *parent_db_id;
310 
311 	tse = &ts->stack[idx];
312 	cr.cp = tse->cp;
313 	cr.call_time = tse->timestamp;
314 	cr.return_time = timestamp;
315 	cr.branch_count = ts->branch_count - tse->branch_count;
316 	cr.insn_count = ts->insn_count - tse->insn_count;
317 	cr.cyc_count = ts->cyc_count - tse->cyc_count;
318 	cr.db_id = tse->db_id;
319 	cr.call_ref = tse->ref;
320 	cr.return_ref = ref;
321 	if (tse->no_call)
322 		cr.flags |= CALL_RETURN_NO_CALL;
323 	if (no_return)
324 		cr.flags |= CALL_RETURN_NO_RETURN;
325 	if (tse->non_call)
326 		cr.flags |= CALL_RETURN_NON_CALL;
327 
328 	/*
329 	 * The parent db_id must be assigned before exporting the child. Note
330 	 * it is not possible to export the parent first because its information
331 	 * is not yet complete because its 'return' has not yet been processed.
332 	 */
333 	parent_db_id = idx ? &(tse - 1)->db_id : NULL;
334 
335 	return crp->process(&cr, parent_db_id, crp->data);
336 }
337 
338 static int __thread_stack__flush(struct thread *thread, struct thread_stack *ts)
339 {
340 	struct call_return_processor *crp = ts->crp;
341 	int err;
342 
343 	if (!crp) {
344 		ts->cnt = 0;
345 		ts->br_stack_pos = 0;
346 		if (ts->br_stack_rb)
347 			ts->br_stack_rb->nr = 0;
348 		return 0;
349 	}
350 
351 	while (ts->cnt) {
352 		err = thread_stack__call_return(thread, ts, --ts->cnt,
353 						ts->last_time, 0, true);
354 		if (err) {
355 			pr_err("Error flushing thread stack!\n");
356 			ts->cnt = 0;
357 			return err;
358 		}
359 	}
360 
361 	return 0;
362 }
363 
364 int thread_stack__flush(struct thread *thread)
365 {
366 	struct thread_stack *ts = thread->ts;
367 	unsigned int pos;
368 	int err = 0;
369 
370 	if (ts) {
371 		for (pos = 0; pos < ts->arr_sz; pos++) {
372 			int ret = __thread_stack__flush(thread, ts + pos);
373 
374 			if (ret)
375 				err = ret;
376 		}
377 	}
378 
379 	return err;
380 }
381 
382 static void thread_stack__update_br_stack(struct thread_stack *ts, u32 flags,
383 					  u64 from_ip, u64 to_ip)
384 {
385 	struct branch_stack *bs = ts->br_stack_rb;
386 	struct branch_entry *be;
387 
388 	if (!ts->br_stack_pos)
389 		ts->br_stack_pos = ts->br_stack_sz;
390 
391 	ts->br_stack_pos -= 1;
392 
393 	be              = &bs->entries[ts->br_stack_pos];
394 	be->from        = from_ip;
395 	be->to          = to_ip;
396 	be->flags.value = 0;
397 	be->flags.abort = !!(flags & PERF_IP_FLAG_TX_ABORT);
398 	be->flags.in_tx = !!(flags & PERF_IP_FLAG_IN_TX);
399 	/* No support for mispredict */
400 	be->flags.mispred = ts->mispred_all;
401 
402 	if (bs->nr < ts->br_stack_sz)
403 		bs->nr += 1;
404 }
405 
406 int thread_stack__event(struct thread *thread, int cpu, u32 flags, u64 from_ip,
407 			u64 to_ip, u16 insn_len, u64 trace_nr, bool callstack,
408 			unsigned int br_stack_sz, bool mispred_all)
409 {
410 	struct thread_stack *ts = thread__stack(thread, cpu);
411 
412 	if (!thread)
413 		return -EINVAL;
414 
415 	if (!ts) {
416 		ts = thread_stack__new(thread, cpu, NULL, callstack, br_stack_sz);
417 		if (!ts) {
418 			pr_warning("Out of memory: no thread stack\n");
419 			return -ENOMEM;
420 		}
421 		ts->trace_nr = trace_nr;
422 		ts->mispred_all = mispred_all;
423 	}
424 
425 	/*
426 	 * When the trace is discontinuous, the trace_nr changes.  In that case
427 	 * the stack might be completely invalid.  Better to report nothing than
428 	 * to report something misleading, so flush the stack.
429 	 */
430 	if (trace_nr != ts->trace_nr) {
431 		if (ts->trace_nr)
432 			__thread_stack__flush(thread, ts);
433 		ts->trace_nr = trace_nr;
434 	}
435 
436 	if (br_stack_sz)
437 		thread_stack__update_br_stack(ts, flags, from_ip, to_ip);
438 
439 	/*
440 	 * Stop here if thread_stack__process() is in use, or not recording call
441 	 * stack.
442 	 */
443 	if (ts->crp || !callstack)
444 		return 0;
445 
446 	if (flags & PERF_IP_FLAG_CALL) {
447 		u64 ret_addr;
448 
449 		if (!to_ip)
450 			return 0;
451 		ret_addr = from_ip + insn_len;
452 		if (ret_addr == to_ip)
453 			return 0; /* Zero-length calls are excluded */
454 		return thread_stack__push(ts, ret_addr,
455 					  flags & PERF_IP_FLAG_TRACE_END);
456 	} else if (flags & PERF_IP_FLAG_TRACE_BEGIN) {
457 		/*
458 		 * If the caller did not change the trace number (which would
459 		 * have flushed the stack) then try to make sense of the stack.
460 		 * Possibly, tracing began after returning to the current
461 		 * address, so try to pop that. Also, do not expect a call made
462 		 * when the trace ended, to return, so pop that.
463 		 */
464 		thread_stack__pop(ts, to_ip);
465 		thread_stack__pop_trace_end(ts);
466 	} else if ((flags & PERF_IP_FLAG_RETURN) && from_ip) {
467 		thread_stack__pop(ts, to_ip);
468 	}
469 
470 	return 0;
471 }
472 
473 void thread_stack__set_trace_nr(struct thread *thread, int cpu, u64 trace_nr)
474 {
475 	struct thread_stack *ts = thread__stack(thread, cpu);
476 
477 	if (!ts)
478 		return;
479 
480 	if (trace_nr != ts->trace_nr) {
481 		if (ts->trace_nr)
482 			__thread_stack__flush(thread, ts);
483 		ts->trace_nr = trace_nr;
484 	}
485 }
486 
487 static void __thread_stack__free(struct thread *thread, struct thread_stack *ts)
488 {
489 	__thread_stack__flush(thread, ts);
490 	zfree(&ts->stack);
491 	zfree(&ts->br_stack_rb);
492 }
493 
494 static void thread_stack__reset(struct thread *thread, struct thread_stack *ts)
495 {
496 	unsigned int arr_sz = ts->arr_sz;
497 
498 	__thread_stack__free(thread, ts);
499 	memset(ts, 0, sizeof(*ts));
500 	ts->arr_sz = arr_sz;
501 }
502 
503 void thread_stack__free(struct thread *thread)
504 {
505 	struct thread_stack *ts = thread->ts;
506 	unsigned int pos;
507 
508 	if (ts) {
509 		for (pos = 0; pos < ts->arr_sz; pos++)
510 			__thread_stack__free(thread, ts + pos);
511 		zfree(&thread->ts);
512 	}
513 }
514 
515 static inline u64 callchain_context(u64 ip, u64 kernel_start)
516 {
517 	return ip < kernel_start ? PERF_CONTEXT_USER : PERF_CONTEXT_KERNEL;
518 }
519 
520 void thread_stack__sample(struct thread *thread, int cpu,
521 			  struct ip_callchain *chain,
522 			  size_t sz, u64 ip, u64 kernel_start)
523 {
524 	struct thread_stack *ts = thread__stack(thread, cpu);
525 	u64 context = callchain_context(ip, kernel_start);
526 	u64 last_context;
527 	size_t i, j;
528 
529 	if (sz < 2) {
530 		chain->nr = 0;
531 		return;
532 	}
533 
534 	chain->ips[0] = context;
535 	chain->ips[1] = ip;
536 
537 	if (!ts) {
538 		chain->nr = 2;
539 		return;
540 	}
541 
542 	last_context = context;
543 
544 	for (i = 2, j = 1; i < sz && j <= ts->cnt; i++, j++) {
545 		ip = ts->stack[ts->cnt - j].ret_addr;
546 		context = callchain_context(ip, kernel_start);
547 		if (context != last_context) {
548 			if (i >= sz - 1)
549 				break;
550 			chain->ips[i++] = context;
551 			last_context = context;
552 		}
553 		chain->ips[i] = ip;
554 	}
555 
556 	chain->nr = i;
557 }
558 
559 /*
560  * Hardware sample records, created some time after the event occurred, need to
561  * have subsequent addresses removed from the call chain.
562  */
563 void thread_stack__sample_late(struct thread *thread, int cpu,
564 			       struct ip_callchain *chain, size_t sz,
565 			       u64 sample_ip, u64 kernel_start)
566 {
567 	struct thread_stack *ts = thread__stack(thread, cpu);
568 	u64 sample_context = callchain_context(sample_ip, kernel_start);
569 	u64 last_context, context, ip;
570 	size_t nr = 0, j;
571 
572 	if (sz < 2) {
573 		chain->nr = 0;
574 		return;
575 	}
576 
577 	if (!ts)
578 		goto out;
579 
580 	/*
581 	 * When tracing kernel space, kernel addresses occur at the top of the
582 	 * call chain after the event occurred but before tracing stopped.
583 	 * Skip them.
584 	 */
585 	for (j = 1; j <= ts->cnt; j++) {
586 		ip = ts->stack[ts->cnt - j].ret_addr;
587 		context = callchain_context(ip, kernel_start);
588 		if (context == PERF_CONTEXT_USER ||
589 		    (context == sample_context && ip == sample_ip))
590 			break;
591 	}
592 
593 	last_context = sample_ip; /* Use sample_ip as an invalid context */
594 
595 	for (; nr < sz && j <= ts->cnt; nr++, j++) {
596 		ip = ts->stack[ts->cnt - j].ret_addr;
597 		context = callchain_context(ip, kernel_start);
598 		if (context != last_context) {
599 			if (nr >= sz - 1)
600 				break;
601 			chain->ips[nr++] = context;
602 			last_context = context;
603 		}
604 		chain->ips[nr] = ip;
605 	}
606 out:
607 	if (nr) {
608 		chain->nr = nr;
609 	} else {
610 		chain->ips[0] = sample_context;
611 		chain->ips[1] = sample_ip;
612 		chain->nr = 2;
613 	}
614 }
615 
616 void thread_stack__br_sample(struct thread *thread, int cpu,
617 			     struct branch_stack *dst, unsigned int sz)
618 {
619 	struct thread_stack *ts = thread__stack(thread, cpu);
620 	const size_t bsz = sizeof(struct branch_entry);
621 	struct branch_stack *src;
622 	struct branch_entry *be;
623 	unsigned int nr;
624 
625 	dst->nr = 0;
626 
627 	if (!ts)
628 		return;
629 
630 	src = ts->br_stack_rb;
631 	if (!src->nr)
632 		return;
633 
634 	dst->nr = min((unsigned int)src->nr, sz);
635 
636 	be = &dst->entries[0];
637 	nr = min(ts->br_stack_sz - ts->br_stack_pos, (unsigned int)dst->nr);
638 	memcpy(be, &src->entries[ts->br_stack_pos], bsz * nr);
639 
640 	if (src->nr >= ts->br_stack_sz) {
641 		sz -= nr;
642 		be = &dst->entries[nr];
643 		nr = min(ts->br_stack_pos, sz);
644 		memcpy(be, &src->entries[0], bsz * ts->br_stack_pos);
645 	}
646 }
647 
648 /* Start of user space branch entries */
649 static bool us_start(struct branch_entry *be, u64 kernel_start, bool *start)
650 {
651 	if (!*start)
652 		*start = be->to && be->to < kernel_start;
653 
654 	return *start;
655 }
656 
657 /*
658  * Start of branch entries after the ip fell in between 2 branches, or user
659  * space branch entries.
660  */
661 static bool ks_start(struct branch_entry *be, u64 sample_ip, u64 kernel_start,
662 		     bool *start, struct branch_entry *nb)
663 {
664 	if (!*start) {
665 		*start = (nb && sample_ip >= be->to && sample_ip <= nb->from) ||
666 			 be->from < kernel_start ||
667 			 (be->to && be->to < kernel_start);
668 	}
669 
670 	return *start;
671 }
672 
673 /*
674  * Hardware sample records, created some time after the event occurred, need to
675  * have subsequent addresses removed from the branch stack.
676  */
677 void thread_stack__br_sample_late(struct thread *thread, int cpu,
678 				  struct branch_stack *dst, unsigned int sz,
679 				  u64 ip, u64 kernel_start)
680 {
681 	struct thread_stack *ts = thread__stack(thread, cpu);
682 	struct branch_entry *d, *s, *spos, *ssz;
683 	struct branch_stack *src;
684 	unsigned int nr = 0;
685 	bool start = false;
686 
687 	dst->nr = 0;
688 
689 	if (!ts)
690 		return;
691 
692 	src = ts->br_stack_rb;
693 	if (!src->nr)
694 		return;
695 
696 	spos = &src->entries[ts->br_stack_pos];
697 	ssz  = &src->entries[ts->br_stack_sz];
698 
699 	d = &dst->entries[0];
700 	s = spos;
701 
702 	if (ip < kernel_start) {
703 		/*
704 		 * User space sample: start copying branch entries when the
705 		 * branch is in user space.
706 		 */
707 		for (s = spos; s < ssz && nr < sz; s++) {
708 			if (us_start(s, kernel_start, &start)) {
709 				*d++ = *s;
710 				nr += 1;
711 			}
712 		}
713 
714 		if (src->nr >= ts->br_stack_sz) {
715 			for (s = &src->entries[0]; s < spos && nr < sz; s++) {
716 				if (us_start(s, kernel_start, &start)) {
717 					*d++ = *s;
718 					nr += 1;
719 				}
720 			}
721 		}
722 	} else {
723 		struct branch_entry *nb = NULL;
724 
725 		/*
726 		 * Kernel space sample: start copying branch entries when the ip
727 		 * falls in between 2 branches (or the branch is in user space
728 		 * because then the start must have been missed).
729 		 */
730 		for (s = spos; s < ssz && nr < sz; s++) {
731 			if (ks_start(s, ip, kernel_start, &start, nb)) {
732 				*d++ = *s;
733 				nr += 1;
734 			}
735 			nb = s;
736 		}
737 
738 		if (src->nr >= ts->br_stack_sz) {
739 			for (s = &src->entries[0]; s < spos && nr < sz; s++) {
740 				if (ks_start(s, ip, kernel_start, &start, nb)) {
741 					*d++ = *s;
742 					nr += 1;
743 				}
744 				nb = s;
745 			}
746 		}
747 	}
748 
749 	dst->nr = nr;
750 }
751 
752 struct call_return_processor *
753 call_return_processor__new(int (*process)(struct call_return *cr, u64 *parent_db_id, void *data),
754 			   void *data)
755 {
756 	struct call_return_processor *crp;
757 
758 	crp = zalloc(sizeof(struct call_return_processor));
759 	if (!crp)
760 		return NULL;
761 	crp->cpr = call_path_root__new();
762 	if (!crp->cpr)
763 		goto out_free;
764 	crp->process = process;
765 	crp->data = data;
766 	return crp;
767 
768 out_free:
769 	free(crp);
770 	return NULL;
771 }
772 
773 void call_return_processor__free(struct call_return_processor *crp)
774 {
775 	if (crp) {
776 		call_path_root__free(crp->cpr);
777 		free(crp);
778 	}
779 }
780 
781 static int thread_stack__push_cp(struct thread_stack *ts, u64 ret_addr,
782 				 u64 timestamp, u64 ref, struct call_path *cp,
783 				 bool no_call, bool trace_end)
784 {
785 	struct thread_stack_entry *tse;
786 	int err;
787 
788 	if (!cp)
789 		return -ENOMEM;
790 
791 	if (ts->cnt == ts->sz) {
792 		err = thread_stack__grow(ts);
793 		if (err)
794 			return err;
795 	}
796 
797 	tse = &ts->stack[ts->cnt++];
798 	tse->ret_addr = ret_addr;
799 	tse->timestamp = timestamp;
800 	tse->ref = ref;
801 	tse->branch_count = ts->branch_count;
802 	tse->insn_count = ts->insn_count;
803 	tse->cyc_count = ts->cyc_count;
804 	tse->cp = cp;
805 	tse->no_call = no_call;
806 	tse->trace_end = trace_end;
807 	tse->non_call = false;
808 	tse->db_id = 0;
809 
810 	return 0;
811 }
812 
813 static int thread_stack__pop_cp(struct thread *thread, struct thread_stack *ts,
814 				u64 ret_addr, u64 timestamp, u64 ref,
815 				struct symbol *sym)
816 {
817 	int err;
818 
819 	if (!ts->cnt)
820 		return 1;
821 
822 	if (ts->cnt == 1) {
823 		struct thread_stack_entry *tse = &ts->stack[0];
824 
825 		if (tse->cp->sym == sym)
826 			return thread_stack__call_return(thread, ts, --ts->cnt,
827 							 timestamp, ref, false);
828 	}
829 
830 	if (ts->stack[ts->cnt - 1].ret_addr == ret_addr &&
831 	    !ts->stack[ts->cnt - 1].non_call) {
832 		return thread_stack__call_return(thread, ts, --ts->cnt,
833 						 timestamp, ref, false);
834 	} else {
835 		size_t i = ts->cnt - 1;
836 
837 		while (i--) {
838 			if (ts->stack[i].ret_addr != ret_addr ||
839 			    ts->stack[i].non_call)
840 				continue;
841 			i += 1;
842 			while (ts->cnt > i) {
843 				err = thread_stack__call_return(thread, ts,
844 								--ts->cnt,
845 								timestamp, ref,
846 								true);
847 				if (err)
848 					return err;
849 			}
850 			return thread_stack__call_return(thread, ts, --ts->cnt,
851 							 timestamp, ref, false);
852 		}
853 	}
854 
855 	return 1;
856 }
857 
858 static int thread_stack__bottom(struct thread_stack *ts,
859 				struct perf_sample *sample,
860 				struct addr_location *from_al,
861 				struct addr_location *to_al, u64 ref)
862 {
863 	struct call_path_root *cpr = ts->crp->cpr;
864 	struct call_path *cp;
865 	struct symbol *sym;
866 	u64 ip;
867 
868 	if (sample->ip) {
869 		ip = sample->ip;
870 		sym = from_al->sym;
871 	} else if (sample->addr) {
872 		ip = sample->addr;
873 		sym = to_al->sym;
874 	} else {
875 		return 0;
876 	}
877 
878 	cp = call_path__findnew(cpr, &cpr->call_path, sym, ip,
879 				ts->kernel_start);
880 
881 	return thread_stack__push_cp(ts, ip, sample->time, ref, cp,
882 				     true, false);
883 }
884 
885 static int thread_stack__pop_ks(struct thread *thread, struct thread_stack *ts,
886 				struct perf_sample *sample, u64 ref)
887 {
888 	u64 tm = sample->time;
889 	int err;
890 
891 	/* Return to userspace, so pop all kernel addresses */
892 	while (thread_stack__in_kernel(ts)) {
893 		err = thread_stack__call_return(thread, ts, --ts->cnt,
894 						tm, ref, true);
895 		if (err)
896 			return err;
897 	}
898 
899 	return 0;
900 }
901 
902 static int thread_stack__no_call_return(struct thread *thread,
903 					struct thread_stack *ts,
904 					struct perf_sample *sample,
905 					struct addr_location *from_al,
906 					struct addr_location *to_al, u64 ref)
907 {
908 	struct call_path_root *cpr = ts->crp->cpr;
909 	struct call_path *root = &cpr->call_path;
910 	struct symbol *fsym = from_al->sym;
911 	struct symbol *tsym = to_al->sym;
912 	struct call_path *cp, *parent;
913 	u64 ks = ts->kernel_start;
914 	u64 addr = sample->addr;
915 	u64 tm = sample->time;
916 	u64 ip = sample->ip;
917 	int err;
918 
919 	if (ip >= ks && addr < ks) {
920 		/* Return to userspace, so pop all kernel addresses */
921 		err = thread_stack__pop_ks(thread, ts, sample, ref);
922 		if (err)
923 			return err;
924 
925 		/* If the stack is empty, push the userspace address */
926 		if (!ts->cnt) {
927 			cp = call_path__findnew(cpr, root, tsym, addr, ks);
928 			return thread_stack__push_cp(ts, 0, tm, ref, cp, true,
929 						     false);
930 		}
931 	} else if (thread_stack__in_kernel(ts) && ip < ks) {
932 		/* Return to userspace, so pop all kernel addresses */
933 		err = thread_stack__pop_ks(thread, ts, sample, ref);
934 		if (err)
935 			return err;
936 	}
937 
938 	if (ts->cnt)
939 		parent = ts->stack[ts->cnt - 1].cp;
940 	else
941 		parent = root;
942 
943 	if (parent->sym == from_al->sym) {
944 		/*
945 		 * At the bottom of the stack, assume the missing 'call' was
946 		 * before the trace started. So, pop the current symbol and push
947 		 * the 'to' symbol.
948 		 */
949 		if (ts->cnt == 1) {
950 			err = thread_stack__call_return(thread, ts, --ts->cnt,
951 							tm, ref, false);
952 			if (err)
953 				return err;
954 		}
955 
956 		if (!ts->cnt) {
957 			cp = call_path__findnew(cpr, root, tsym, addr, ks);
958 
959 			return thread_stack__push_cp(ts, addr, tm, ref, cp,
960 						     true, false);
961 		}
962 
963 		/*
964 		 * Otherwise assume the 'return' is being used as a jump (e.g.
965 		 * retpoline) and just push the 'to' symbol.
966 		 */
967 		cp = call_path__findnew(cpr, parent, tsym, addr, ks);
968 
969 		err = thread_stack__push_cp(ts, 0, tm, ref, cp, true, false);
970 		if (!err)
971 			ts->stack[ts->cnt - 1].non_call = true;
972 
973 		return err;
974 	}
975 
976 	/*
977 	 * Assume 'parent' has not yet returned, so push 'to', and then push and
978 	 * pop 'from'.
979 	 */
980 
981 	cp = call_path__findnew(cpr, parent, tsym, addr, ks);
982 
983 	err = thread_stack__push_cp(ts, addr, tm, ref, cp, true, false);
984 	if (err)
985 		return err;
986 
987 	cp = call_path__findnew(cpr, cp, fsym, ip, ks);
988 
989 	err = thread_stack__push_cp(ts, ip, tm, ref, cp, true, false);
990 	if (err)
991 		return err;
992 
993 	return thread_stack__call_return(thread, ts, --ts->cnt, tm, ref, false);
994 }
995 
996 static int thread_stack__trace_begin(struct thread *thread,
997 				     struct thread_stack *ts, u64 timestamp,
998 				     u64 ref)
999 {
1000 	struct thread_stack_entry *tse;
1001 	int err;
1002 
1003 	if (!ts->cnt)
1004 		return 0;
1005 
1006 	/* Pop trace end */
1007 	tse = &ts->stack[ts->cnt - 1];
1008 	if (tse->trace_end) {
1009 		err = thread_stack__call_return(thread, ts, --ts->cnt,
1010 						timestamp, ref, false);
1011 		if (err)
1012 			return err;
1013 	}
1014 
1015 	return 0;
1016 }
1017 
1018 static int thread_stack__trace_end(struct thread_stack *ts,
1019 				   struct perf_sample *sample, u64 ref)
1020 {
1021 	struct call_path_root *cpr = ts->crp->cpr;
1022 	struct call_path *cp;
1023 	u64 ret_addr;
1024 
1025 	/* No point having 'trace end' on the bottom of the stack */
1026 	if (!ts->cnt || (ts->cnt == 1 && ts->stack[0].ref == ref))
1027 		return 0;
1028 
1029 	cp = call_path__findnew(cpr, ts->stack[ts->cnt - 1].cp, NULL, 0,
1030 				ts->kernel_start);
1031 
1032 	ret_addr = sample->ip + sample->insn_len;
1033 
1034 	return thread_stack__push_cp(ts, ret_addr, sample->time, ref, cp,
1035 				     false, true);
1036 }
1037 
1038 static bool is_x86_retpoline(const char *name)
1039 {
1040 	const char *p = strstr(name, "__x86_indirect_thunk_");
1041 
1042 	return p == name || !strcmp(name, "__indirect_thunk_start");
1043 }
1044 
1045 /*
1046  * x86 retpoline functions pollute the call graph. This function removes them.
1047  * This does not handle function return thunks, nor is there any improvement
1048  * for the handling of inline thunks or extern thunks.
1049  */
1050 static int thread_stack__x86_retpoline(struct thread_stack *ts,
1051 				       struct perf_sample *sample,
1052 				       struct addr_location *to_al)
1053 {
1054 	struct thread_stack_entry *tse = &ts->stack[ts->cnt - 1];
1055 	struct call_path_root *cpr = ts->crp->cpr;
1056 	struct symbol *sym = tse->cp->sym;
1057 	struct symbol *tsym = to_al->sym;
1058 	struct call_path *cp;
1059 
1060 	if (sym && is_x86_retpoline(sym->name)) {
1061 		/*
1062 		 * This is a x86 retpoline fn. It pollutes the call graph by
1063 		 * showing up everywhere there is an indirect branch, but does
1064 		 * not itself mean anything. Here the top-of-stack is removed,
1065 		 * by decrementing the stack count, and then further down, the
1066 		 * resulting top-of-stack is replaced with the actual target.
1067 		 * The result is that the retpoline functions will no longer
1068 		 * appear in the call graph. Note this only affects the call
1069 		 * graph, since all the original branches are left unchanged.
1070 		 */
1071 		ts->cnt -= 1;
1072 		sym = ts->stack[ts->cnt - 2].cp->sym;
1073 		if (sym && sym == tsym && to_al->addr != tsym->start) {
1074 			/*
1075 			 * Target is back to the middle of the symbol we came
1076 			 * from so assume it is an indirect jmp and forget it
1077 			 * altogether.
1078 			 */
1079 			ts->cnt -= 1;
1080 			return 0;
1081 		}
1082 	} else if (sym && sym == tsym) {
1083 		/*
1084 		 * Target is back to the symbol we came from so assume it is an
1085 		 * indirect jmp and forget it altogether.
1086 		 */
1087 		ts->cnt -= 1;
1088 		return 0;
1089 	}
1090 
1091 	cp = call_path__findnew(cpr, ts->stack[ts->cnt - 2].cp, tsym,
1092 				sample->addr, ts->kernel_start);
1093 	if (!cp)
1094 		return -ENOMEM;
1095 
1096 	/* Replace the top-of-stack with the actual target */
1097 	ts->stack[ts->cnt - 1].cp = cp;
1098 
1099 	return 0;
1100 }
1101 
1102 int thread_stack__process(struct thread *thread, struct comm *comm,
1103 			  struct perf_sample *sample,
1104 			  struct addr_location *from_al,
1105 			  struct addr_location *to_al, u64 ref,
1106 			  struct call_return_processor *crp)
1107 {
1108 	struct thread_stack *ts = thread__stack(thread, sample->cpu);
1109 	enum retpoline_state_t rstate;
1110 	int err = 0;
1111 
1112 	if (ts && !ts->crp) {
1113 		/* Supersede thread_stack__event() */
1114 		thread_stack__reset(thread, ts);
1115 		ts = NULL;
1116 	}
1117 
1118 	if (!ts) {
1119 		ts = thread_stack__new(thread, sample->cpu, crp, true, 0);
1120 		if (!ts)
1121 			return -ENOMEM;
1122 		ts->comm = comm;
1123 	}
1124 
1125 	rstate = ts->rstate;
1126 	if (rstate == X86_RETPOLINE_DETECTED)
1127 		ts->rstate = X86_RETPOLINE_POSSIBLE;
1128 
1129 	/* Flush stack on exec */
1130 	if (ts->comm != comm && thread->pid_ == thread->tid) {
1131 		err = __thread_stack__flush(thread, ts);
1132 		if (err)
1133 			return err;
1134 		ts->comm = comm;
1135 	}
1136 
1137 	/* If the stack is empty, put the current symbol on the stack */
1138 	if (!ts->cnt) {
1139 		err = thread_stack__bottom(ts, sample, from_al, to_al, ref);
1140 		if (err)
1141 			return err;
1142 	}
1143 
1144 	ts->branch_count += 1;
1145 	ts->insn_count += sample->insn_cnt;
1146 	ts->cyc_count += sample->cyc_cnt;
1147 	ts->last_time = sample->time;
1148 
1149 	if (sample->flags & PERF_IP_FLAG_CALL) {
1150 		bool trace_end = sample->flags & PERF_IP_FLAG_TRACE_END;
1151 		struct call_path_root *cpr = ts->crp->cpr;
1152 		struct call_path *cp;
1153 		u64 ret_addr;
1154 
1155 		if (!sample->ip || !sample->addr)
1156 			return 0;
1157 
1158 		ret_addr = sample->ip + sample->insn_len;
1159 		if (ret_addr == sample->addr)
1160 			return 0; /* Zero-length calls are excluded */
1161 
1162 		cp = call_path__findnew(cpr, ts->stack[ts->cnt - 1].cp,
1163 					to_al->sym, sample->addr,
1164 					ts->kernel_start);
1165 		err = thread_stack__push_cp(ts, ret_addr, sample->time, ref,
1166 					    cp, false, trace_end);
1167 
1168 		/*
1169 		 * A call to the same symbol but not the start of the symbol,
1170 		 * may be the start of a x86 retpoline.
1171 		 */
1172 		if (!err && rstate == X86_RETPOLINE_POSSIBLE && to_al->sym &&
1173 		    from_al->sym == to_al->sym &&
1174 		    to_al->addr != to_al->sym->start)
1175 			ts->rstate = X86_RETPOLINE_DETECTED;
1176 
1177 	} else if (sample->flags & PERF_IP_FLAG_RETURN) {
1178 		if (!sample->addr) {
1179 			u32 return_from_kernel = PERF_IP_FLAG_SYSCALLRET |
1180 						 PERF_IP_FLAG_INTERRUPT;
1181 
1182 			if (!(sample->flags & return_from_kernel))
1183 				return 0;
1184 
1185 			/* Pop kernel stack */
1186 			return thread_stack__pop_ks(thread, ts, sample, ref);
1187 		}
1188 
1189 		if (!sample->ip)
1190 			return 0;
1191 
1192 		/* x86 retpoline 'return' doesn't match the stack */
1193 		if (rstate == X86_RETPOLINE_DETECTED && ts->cnt > 2 &&
1194 		    ts->stack[ts->cnt - 1].ret_addr != sample->addr)
1195 			return thread_stack__x86_retpoline(ts, sample, to_al);
1196 
1197 		err = thread_stack__pop_cp(thread, ts, sample->addr,
1198 					   sample->time, ref, from_al->sym);
1199 		if (err) {
1200 			if (err < 0)
1201 				return err;
1202 			err = thread_stack__no_call_return(thread, ts, sample,
1203 							   from_al, to_al, ref);
1204 		}
1205 	} else if (sample->flags & PERF_IP_FLAG_TRACE_BEGIN) {
1206 		err = thread_stack__trace_begin(thread, ts, sample->time, ref);
1207 	} else if (sample->flags & PERF_IP_FLAG_TRACE_END) {
1208 		err = thread_stack__trace_end(ts, sample, ref);
1209 	} else if (sample->flags & PERF_IP_FLAG_BRANCH &&
1210 		   from_al->sym != to_al->sym && to_al->sym &&
1211 		   to_al->addr == to_al->sym->start) {
1212 		struct call_path_root *cpr = ts->crp->cpr;
1213 		struct call_path *cp;
1214 
1215 		/*
1216 		 * The compiler might optimize a call/ret combination by making
1217 		 * it a jmp. Make that visible by recording on the stack a
1218 		 * branch to the start of a different symbol. Note, that means
1219 		 * when a ret pops the stack, all jmps must be popped off first.
1220 		 */
1221 		cp = call_path__findnew(cpr, ts->stack[ts->cnt - 1].cp,
1222 					to_al->sym, sample->addr,
1223 					ts->kernel_start);
1224 		err = thread_stack__push_cp(ts, 0, sample->time, ref, cp, false,
1225 					    false);
1226 		if (!err)
1227 			ts->stack[ts->cnt - 1].non_call = true;
1228 	}
1229 
1230 	return err;
1231 }
1232 
1233 size_t thread_stack__depth(struct thread *thread, int cpu)
1234 {
1235 	struct thread_stack *ts = thread__stack(thread, cpu);
1236 
1237 	if (!ts)
1238 		return 0;
1239 	return ts->cnt;
1240 }
1241