Skip to contentSkip to conversation

perf script: Fix obtaining next event

PATCH In review Chandan Rajendra 3 participants 3 messages Oct 30, 2019 perf events

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 perf script: Fix obtaining next event

The current code segfaults when perf.data file contains two or more events. This happens due to incorrect pointer arithmetic being performed in trace_find_next_event().

tep_handle->events is an array of pointers to 'struct tep_event'. The pointer arithmetic interprets tep_handle->events as an array of 'struct tep_event' elements.

This commit replaces the usage of pointer arithmetic with calls to tep_get_event().

Fixes bb3dd7e ("tools lib traceevent, perf tools: Move struct tep_handler definition in a local header file")
---
 tools/perf/util/trace-event-parse.c | 24 +++++++-----------------
 1 file changed, 7 insertions(+), 17 deletions(-)
tools/perf/util/trace-event-parse.c +7 -17
index 5d6bfc70b210..7bf423a3631e 100644
@@ -176,31 +176,21 @@ int parse_event_file(struct tep_handle *pevent,
 struct tep_event *trace_find_next_event(struct tep_handle *pevent,
 					struct tep_event *event)
 {
-	static int idx;
+	int idx;
 	int events_count;
-	struct tep_event *all_events;
 
-	all_events = tep_get_first_event(pevent);
 	events_count = tep_get_events_count(pevent);
-	if (!pevent || !all_events || events_count < 1)
+	if (!pevent || events_count < 1)
 		return NULL;
 
-	if (!event) {
-		idx = 0;
-		return all_events;
-	}
+	if (!event)
+		return tep_get_event(pevent, 0);
 
-	if (idx < events_count && event == (all_events + idx)) {
-		idx++;
-		if (idx == events_count)
-			return NULL;
-		return (all_events + idx);
+	for (idx = 0; idx < events_count - 1; idx++) {
+		if (event == tep_get_event(pevent, idx))
+			return tep_get_event(pevent, idx + 1);
 	}
 
-	for (idx = 1; idx < events_count; idx++) {
-		if (event == (all_events + (idx - 1)))
-			return (all_events + idx);
-	}
 	return NULL;
 }
 
Signature · 1 line
2.19.1
#1 lore raw
Quoted 8 lines · Chandan Rajendra

The current code segfaults when perf.data file contains two or more events. This happens due to incorrect pointer arithmetic being performed in trace_find_next_event().

tep_handle->events is an array of pointers to 'struct tep_event'. The pointer arithmetic interprets tep_handle->events as an array of 'struct tep_event' elements.

Quoted 5 lines

This commit replaces the usage of pointer arithmetic with calls to tep_get_event().

Fixes bb3dd7e ("tools lib traceevent, perf tools: Move struct tep_handler definition in a local header file")
  $ sudo ./perf record -e sched:sched_switch -e syscalls:sys_enter_openat -- make

Without patch:

  $ sudo ./perf script -g python
  Segmentation fault

With patch:

  $ sudo ./perf script -g python
  generated Python script: perf-script.py
Reviewed-and-tested-by Ravi Bangoria <[email protected]>
#2 lore raw

Em Wed, Oct 30, 2019 at 03:16:10PM +0530, Ravi Bangoria escreveu:

Quoted 20 lines · Chandan Rajendra
Quoted 8 lines

The current code segfaults when perf.data file contains two or more events. This happens due to incorrect pointer arithmetic being performed in trace_find_next_event().

tep_handle->events is an array of pointers to 'struct tep_event'. The pointer arithmetic interprets tep_handle->events as an array of 'struct tep_event' elements.

Quoted 5 lines

This commit replaces the usage of pointer arithmetic with calls to tep_get_event().

Fixes bb3dd7e ("tools lib traceevent, perf tools: Move struct tep_handler definition in a local header file")
  $ sudo ./perf record -e sched:sched_switch -e syscalls:sys_enter_openat -- make

Without patch:

Quoted 8 lines
  $ sudo ./perf script -g python
  Segmentation fault

With patch:

  $ sudo ./perf script -g python
  generated Python script: perf-script.py
Reviewed-and-tested-by Ravi Bangoria <[email protected]>

This was fixed already in perf/core, by Steven:

commit 9bdff5b6436655d42dd30253c521e86ce07b9961

Author: Steven Rostedt (VMware) <[email protected]>
Date:   Thu Oct 17 17:05:23 2019 -0400

perf tools: Remove unused trace_find_next_event()

trace_find_next_event() was buggy and pretty much a useless helper. As there are no more users, just remove it.

commit a5e05abc6b8d81148b35cd8632a4a6252383d968

Author: Steven Rostedt (VMware) <[email protected]>
Date:   Thu Oct 17 17:05:22 2019 -0400

perf scripting engines: Iterate on tep event arrays directly

Instead of calling a useless (and broken) helper function to get the next event of a tep event array, just get the array directly and iterate over it.

Note, the broken part was from trace_find_next_event() which after this will no longer be used, and can be removed.

Committer notes:

This fixes a segfault when generating python scripts from perf.data files with multiple tracepoint events, i.e. the following use case is fixed by this patch:

# perf record -e sched:* sleep 1
[ perf record: Woken up 31 times to write data ]
[ perf record: Captured and wrote 0.031 MB perf.data (9 samples) ] # perf script -g python
Segmentation fault (core dumped)
#