perf thread-stack: Represent jmps to the start of a different symbol

The compiler might optimize a call/ret combination by making it a jmp.
However the thread-stack does not presently cater for that, so that such
control flow is not visible in the call graph. Make it visible by
recording on the stack a branch to the start of a different symbol.
Note, that means when a ret pops the stack, all jmps must be popped off
first.

Example:

  $ cat jmp-to-fn.c
  __attribute__((noinline)) int bar(void)
  {
          return -1;
  }

  __attribute__((noinline)) int foo(void)
  {
          return bar() + 1;
  }

  int main()
  {
          return foo();
  }
  $ gcc -ggdb3 -Wall -Wextra -O2 -o jmp-to-fn jmp-to-fn.c
  $ objdump -d jmp-to-fn
  <SNIP>
  0000000000001040 <main>:
      1040:       31 c0                   xor    %eax,%eax
      1042:       e9 09 01 00 00          jmpq   1150 <foo>
  <SNIP>
  0000000000001140 <bar>:
      1140:       b8 ff ff ff ff          mov    $0xffffffff,%eax
      1145:       c3                      retq
  <SNIP>
  0000000000001150 <foo>:
      1150:       31 c0                   xor    %eax,%eax
      1152:       e8 e9 ff ff ff          callq  1140 <bar>
      1157:       83 c0 01                add    $0x1,%eax
      115a:       c3                      retq
  <SNIP>
  $ perf record -o jmp-to-fn.perf.data -e intel_pt/cyc/u ./jmp-to-fn
  [ perf record: Woken up 1 times to write data ]
  [ perf record: Captured and wrote 0,017 MB jmp-to-fn.perf.data ]
  $ perf script -i jmp-to-fn.perf.data --itrace=be -s ~/libexec/perf-core/scripts/python/export-to-sqlite.py jmp-to-fn.db branches calls
  2019-01-08 13:24:58.783069 Creating database...
  2019-01-08 13:24:58.794650 Writing records...
  2019-01-08 13:24:59.008050 Adding indexes
  2019-01-08 13:24:59.015802 Done
  $  ~/libexec/perf-core/scripts/python/exported-sql-viewer.py jmp-to-fn.db

Before:

    main
        -> bar

After:

    main
        -> foo
            -> bar

Committer testing:

Install the python2-pyside package, then select these menu options
on the GUI:

   "Reports"
      "Context sensitive callgraphs"

Then go on expanding the symbols, to get, full picture when doing this
on a fedora:29 with gcc version 8.2.1 20181215 (Red Hat 8.2.1-6) (GCC):

jmp-to-fn
  PID:TID
    _start                (ld-2.28.so)
      __libc_start_main
        main
          foo
            bar

To verify that indeed, this fixes the problem.

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Link: http://lkml.kernel.org/r/20190109091835.5570-5-adrian.hunter@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

1 files changed, 28 insertions, 2 deletions

diff --git a/tools/perf/util/thread-stack.c b/tools/perf/util/thread-stack.c
index 7f8eff018c16..f52c0f90915d 100644
--- a/tools/perf/util/thread-stack.c
+++ b/tools/perf/util/thread-stack.c
@@ -38,6 +38,7 @@
  * @cp: call path
  * @no_call: a 'call' was not seen
  * @trace_end: a 'call' but trace ended
+ * @non_call: a branch but not a 'call' to the start of a different symbol
  */
 struct thread_stack_entry {
 	u64 ret_addr;
@@ -47,6 +48,7 @@ struct thread_stack_entry {
 	struct call_path *cp;
 	bool no_call;
 	bool trace_end;
+	bool non_call;
 };
 
 /**
@@ -268,6 +270,8 @@ static int thread_stack__call_return(struct thread *thread,
 		cr.flags |= CALL_RETURN_NO_CALL;
 	if (no_return)
 		cr.flags |= CALL_RETURN_NO_RETURN;
+	if (tse->non_call)
+		cr.flags |= CALL_RETURN_NON_CALL;
 
 	return crp->process(&cr, crp->data);
 }
@@ -510,6 +514,7 @@ static int thread_stack__push_cp(struct thread_stack *ts, u64 ret_addr,
 	tse->cp = cp;
 	tse->no_call = no_call;
 	tse->trace_end = trace_end;
+	tse->non_call = false;
 
 	return 0;
 }
@@ -531,14 +536,16 @@ static int thread_stack__pop_cp(struct thread *thread, struct thread_stack *ts,
 							 timestamp, ref, false);
 	}
 
-	if (ts->stack[ts->cnt - 1].ret_addr == ret_addr) {
+	if (ts->stack[ts->cnt - 1].ret_addr == ret_addr &&
+	    !ts->stack[ts->cnt - 1].non_call) {
 		return thread_stack__call_return(thread, ts, --ts->cnt,
 						 timestamp, ref, false);
 	} else {
 		size_t i = ts->cnt - 1;
 
 		while (i--) {
-			if (ts->stack[i].ret_addr != ret_addr)
+			if (ts->stack[i].ret_addr != ret_addr ||
+			    ts->stack[i].non_call)
 				continue;
 			i += 1;
 			while (ts->cnt > i) {
@@ -757,6 +764,25 @@ int thread_stack__process(struct thread *thread, struct comm *comm,
 		err = thread_stack__trace_begin(thread, ts, sample->time, ref);
 	} else if (sample->flags & PERF_IP_FLAG_TRACE_END) {
 		err = thread_stack__trace_end(ts, sample, ref);
+	} else if (sample->flags & PERF_IP_FLAG_BRANCH &&
+		   from_al->sym != to_al->sym && to_al->sym &&
+		   to_al->addr == to_al->sym->start) {
+		struct call_path_root *cpr = ts->crp->cpr;
+		struct call_path *cp;
+
+		/*
+		 * The compiler might optimize a call/ret combination by making
+		 * it a jmp. Make that visible by recording on the stack a
+		 * branch to the start of a different symbol. Note, that means
+		 * when a ret pops the stack, all jmps must be popped off first.
+		 */
+		cp = call_path__findnew(cpr, ts->stack[ts->cnt - 1].cp,
+					to_al->sym, sample->addr,
+					ts->kernel_start);
+		err = thread_stack__push_cp(ts, 0, sample->time, ref, cp, false,
+					    false);
+		if (!err)
+			ts->stack[ts->cnt - 1].non_call = true;
 	}
 
 	return err;