Skip to contentSkip to conversation

perf tools: Fix segfault accessing sample_id xyarray

PATCH Applied Adrian Hunter 3 participants 3 messages Apr 13, 2022 – Apr 14, 2022 perf events

lore
Share & tools
lore

Adrian Hunter

author reviewer
#0 lore raw

PATCH 1/1 perf tools: Fix segfault accessing sample_id xyarray

perf_evsel sample_id is an xyarray which can cause a segfault when accessed beyond its size. e.g.

 # perf record -e intel_pt// -C 1 sleep 1
 Segmentation fault (core dumped)

That is happening because a dummy event is opened to capture text poke events accoss all CPUs, however the mmap logic is allocating according to the number of user_requested_cpus.

In general, perf sometimes uses the evsel cpus to open events, and sometimes the evlist user_requested_cpus. However, it is not necessary to determine which case is which because the opened event file descriptors are also in an xyarray, the size of whch can be used to correctly allocate the size of the sample_id xyarray, because there is one ID per file descriptor. Note, in the affected code path, perf_evsel fd array is subsequently used to get the file descriptor for the mmap, so it makes sense for the xyarrays to be the same size there.

Fixes 246eba8e9041c4 ("perf tools: Add support for PERF_RECORD_TEXT_POKE")
Fixes d1a177595b3a82 ("libperf: Adopt perf_evlist__mmap()/munmap() from tools/perf")
---
 tools/lib/perf/evlist.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
tools/lib/perf/evlist.c +1 -2
index 1b15ba13c477..a09315538a30 100644
@@ -577,7 +577,6 @@ int perf_evlist__mmap_ops(struct perf_evlist *evlist,
 {
 	struct perf_evsel *evsel;
 	const struct perf_cpu_map *cpus = evlist->user_requested_cpus;
-	const struct perf_thread_map *threads = evlist->threads;
 
 	if (!ops || !ops->get || !ops->mmap)
 		return -EINVAL;
@@ -589,7 +588,7 @@ int perf_evlist__mmap_ops(struct perf_evlist *evlist,
 	perf_evlist__for_each_entry(evlist, evsel) {
 		if ((evsel->attr.read_format & PERF_FORMAT_ID) &&
 		    evsel->sample_id == NULL &&
-		    perf_evsel__alloc_id(evsel, perf_cpu_map__nr(cpus), threads->nr) < 0)
+		    perf_evsel__alloc_id(evsel, evsel->fd->max_x, evsel->fd->max_y) < 0)
 			return -ENOMEM;
 	}
 
Signature · 1 line
2.25.1
reviewer Acked-by on v1 1/1
#1 lore raw
Quoted 10 lines · Adrian Hunter

perf_evsel sample_id is an xyarray which can cause a segfault when accessed beyond its size. e.g.

 # perf record -e intel_pt// -C 1 sleep 1
 Segmentation fault (core dumped)

That is happening because a dummy event is opened to capture text poke events accoss all CPUs, however the mmap logic is allocating according to the number of user_requested_cpus.

Nit: typo on 'accoss'

Quoted 9 lines

In general, perf sometimes uses the evsel cpus to open events, and sometimes the evlist user_requested_cpus. However, it is not necessary to determine which case is which because the opened event file descriptors are also in an xyarray, the size of whch can be used to correctly allocate the size of the sample_id xyarray, because there is one ID per file descriptor. Note, in the affected code path, perf_evsel fd array is subsequently used to get the file descriptor for the mmap, so it makes sense for the xyarrays to be the same size there.

Quoted 4 lines
Fixes 246eba8e9041c4 ("perf tools: Add support for PERF_RECORD_TEXT_POKE")
Fixes d1a177595b3a82 ("libperf: Adopt perf_evlist__mmap()/munmap() from tools/perf")

Thanks,
Ian

Quoted 28 lines
---
 tools/lib/perf/evlist.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
tools/lib/perf/evlist.c +1 -3
index 1b15ba13c477..a09315538a30 100644
@@ -577,7 +577,6 @@ int perf_evlist__mmap_ops(struct perf_evlist *evlist,
 {
        struct perf_evsel *evsel;
        const struct perf_cpu_map *cpus = evlist->user_requested_cpus;
-       const struct perf_thread_map *threads = evlist->threads;
 
        if (!ops || !ops->get || !ops->mmap)
                return -EINVAL;
@@ -589,7 +588,7 @@ int perf_evlist__mmap_ops(struct perf_evlist *evlist,
        perf_evlist__for_each_entry(evlist, evsel) {
                if ((evsel->attr.read_format & PERF_FORMAT_ID) &&
                    evsel->sample_id == NULL &&
-                   perf_evsel__alloc_id(evsel, perf_cpu_map__nr(cpus), threads->nr) < 0)
+                   perf_evsel__alloc_id(evsel, evsel->fd->max_x, evsel->fd->max_y) < 0)
                        return -ENOMEM;
        }
 
--

2.25.1

maintainer Applied
#2 lore raw

Em Wed, Apr 13, 2022 at 02:11:41PM -0700, Ian Rogers escreveu:

Quoted 5 lines · Adrian Hunter
Quoted 4 lines

perf_evsel sample_id is an xyarray which can cause a segfault when accessed beyond its size. e.g.

Quoted 8 lines
Quoted 6 lines
 # perf record -e intel_pt// -C 1 sleep 1
 Segmentation fault (core dumped)

That is happening because a dummy event is opened to capture text poke events accoss all CPUs, however the mmap logic is allocating according to the number of user_requested_cpus.

Nit: typo on 'accoss'

Fixed

Quoted 9 lines
Quoted 9 lines

In general, perf sometimes uses the evsel cpus to open events, and sometimes the evlist user_requested_cpus. However, it is not necessary to determine which case is which because the opened event file descriptors are also in an xyarray, the size of whch can be used to correctly allocate the size of the sample_id xyarray, because there is one ID per file descriptor. Note, in the affected code path, perf_evsel fd array is subsequently used to get the file descriptor for the mmap, so it makes sense for the xyarrays to be the same size there.

Quoted 6 lines
Quoted 4 lines
Fixes 246eba8e9041c4 ("perf tools: Add support for PERF_RECORD_TEXT_POKE")
Fixes d1a177595b3a82 ("libperf: Adopt perf_evlist__mmap()/munmap() from tools/perf")

Thanks, applied.

- Arnaldo

Quoted 31 lines

Thanks,
Ian

Quoted 28 lines
---
 tools/lib/perf/evlist.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
tools/lib/perf/evlist.c +1 -3
index 1b15ba13c477..a09315538a30 100644
@@ -577,7 +577,6 @@ int perf_evlist__mmap_ops(struct perf_evlist *evlist,
 {
        struct perf_evsel *evsel;
        const struct perf_cpu_map *cpus = evlist->user_requested_cpus;
-       const struct perf_thread_map *threads = evlist->threads;
 
        if (!ops || !ops->get || !ops->mmap)
                return -EINVAL;
@@ -589,7 +588,7 @@ int perf_evlist__mmap_ops(struct perf_evlist *evlist,
        perf_evlist__for_each_entry(evlist, evsel) {
                if ((evsel->attr.read_format & PERF_FORMAT_ID) &&
                    evsel->sample_id == NULL &&
-                   perf_evsel__alloc_id(evsel, perf_cpu_map__nr(cpus), threads->nr) < 0)
+                   perf_evsel__alloc_id(evsel, evsel->fd->max_x, evsel->fd->max_y) < 0)
                        return -ENOMEM;
        }
 
--

2.25.1

Signature · 1 line
- Arnaldo