perf tools: Assorted fixes
Arnaldo Carvalho de Melo
maintainerv2 cover perf tools: Assorted fixes
Hi,
Sixth batch of pre-existing bug fixes found by sashiko-bot AI review during the perf-data-validation hardening series. All bugs are latent in existing code — none were introduced by the hardening patches.
Three broad categories:
1. snprintf() accumulation overflows (patches 2, 9, 10, 11):
Several functions accumulate formatted output via ret += snprintf(). snprintf() returns the would-have-been-written count, so on truncation ret overshoots the buffer size and the next 'size - ret' underflows to a huge unsigned value, disabling bounds checking. Switched to scnprintf() which returns actual bytes written.
Affected: cpu_map__snprint(), snprintf_hex(), synthesize_bpf_prog_name(), hists__scnprintf_title(), build_id__snprintf(), hwmon_pmu__describe_items().
2. Missing safety checks on untrusted data (patches 1, 3, 6, 7):
- get_max_num(): size_t underflow on empty sysfs file causes heap over-read.
- machine__resolve(): unguarded env->cpu[] access with untrusted CPU index — switched to perf_env__get_cpu_topology() accessor.
- timehist: test_bit(prio, ...) without bounds check on untrusted tracepoint priority.
- idle-hist: rb_first_cached() on a tree populated with plain rb_insert_color() — rb_leftmost never set, callchains silently dropped.
3. Resource hygiene (patches 4, 5, 8):
- mbind() bitmap allocation one bit short of what kernel reads.
- bitmap_free() without NULLing the pointer (3 call sites).
- O_CLOEXEC missing from open() calls in DSO and ELF code (12 call sites across 2 files).
Also expanded the libperf ABI TODO (tools/lib/perf/TODO) to emphasize the code simplification argument for widening struct perf_cpu.cpu from int16_t to int — the narrow type forces defensive truncation checks at every boundary where wider CPU indices are narrowed.
tools/lib/perf/TODO | 7 +++++++ tools/perf/builtin-record.c | 1 + tools/perf/builtin-sched.c | 7 +++++-- tools/perf/util/bpf-event.c | 11 ++++++----- tools/perf/util/build-id.c | 2 +- tools/perf/util/cpumap.c | 24 +++++++++++++++--------- tools/perf/util/dso.c | 4 ++-- tools/perf/util/event.c | 9 +++++++-- tools/perf/util/header.c | 4 +++- tools/perf/util/hist.c | 7 ++++--- tools/perf/util/hwmon_pmu.c | 12 ++++++------ tools/perf/util/mmap.c | 4 +++- tools/perf/util/symbol-elf.c | 20 ++++++++++---------- 13 files changed, 70 insertions(+), 42 deletions(-)
Changes since v1:
- Patch 6: fix prio bounds-check logic — the v1 condition
(prio < 0 || prio >= MAX_PRIO || !test_bit(...)) incorrectly
skipped events with unknown priority (prio == -1). Changed to
(prio >= 0 && (prio >= MAX_PRIO || !test_bit(...))) to preserve
the original pass-through for events without priority info.
(Found by sashiko-bot lore review)Thanks,
- Arnaldo
Arnaldo Carvalho de Melo
maintainerPATCH v2 1/11 perf tools: Fix get_max_num() size_t underflow on empty sysfs file
From: Arnaldo Carvalho de Melo <[email protected]>
get_max_num() reads a sysfs file (cpu/possible, cpu/present, or node/possible) and scans backward from the end to find the last number. If the file is empty, filename__read_str() returns num == 0. The loop while (--num) decrements the size_t from 0 to SIZE_MAX, reading backward across the heap until a comma or hyphen is found or unmapped memory is hit.
Add an early return for empty files before the backward scan.
--- tools/perf/util/cpumap.c | 6 ++++++ 1 file changed, 6 insertions(+)
tools/perf/util/cpumap.c +6 -0
@@ -448,6 +448,12 @@ static int get_max_num(char *path, int *max) buf[num] ='\0' ; +/* empty file — nothing to parse */ +if (num ==0 ) { + err =-1 ; + goto out; + } +/* start on the right, to find highest node num */ while (--num) {if ((buf[num] ==',' ) || (buf[num] =='-' )) {
Signature · 1 line
2.54.0
[email protected]
AI reviewRe: [PATCH 01/11] perf tools: Fix get_max_num() size_t underflow on empty sysfs file
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
Pre-existing issues:
- [High] Heap out-of-bounds write (poison null byte) due to size_t underflow when reading empty files --
commit 1a2f44329fa687f2b52540b18c56a66a0d4e9628
Author: Arnaldo Carvalho de Melo <[email protected]>
perf tools: Fix get_max_num() size_t underflow on empty sysfs file
This commit fixes a size_t underflow in get_max_num() by adding an early return when a sysfs file is empty. Previously, reading an empty file would result in num == 0, causing a backward scan to underflow and read out-of-bounds memory.
tools/perf/util/cpumap.c static int get_max_num(char *path, int *max) · line 456
Quoted 13 lines
tools/perf/util/cpumap.c +6
-0@@ -448,6 +448,12 @@ static int get_max_num(char *path, int *max) buf[num] ='\0' ; +/* empty file — nothing to parse */ +if (num ==0 ) { + err =-1 ; + goto out; + } +
[Severity: High]
This isn't a bug introduced by this patch, but while you are fixing empty file reads here, does this same size_t underflow pattern occur in other places that read strings?
For instance, in tools/perf/util/pmu.c:pmu_id():
if (filename__read_str(path, &str, &len) < 0) return NULL; str[len - 1] = 0; /* remove line feed */
If len is 0 because the identifier file is completely empty, len - 1 underflows to SIZE_MAX. Since the string is heap-allocated and returned with a length of 0, writing to index SIZE_MAX effectively writes one byte before the string, which corrupts the malloc chunk size metadata.
A similar pattern seems to exist in tools/perf/util/thread.c:thread__set_comm_from_proc():
if (procfs__read_str(path, &comm, &sz) == 0) {
comm[sz - 1] = '\0';
err = thread__set_comm(thread, comm, 0);
}Could these empty file underflows trigger similar heap corruption if the identifier or comm files happen to be empty?
Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=1
Arnaldo Carvalho de Melo
maintainerPATCH v2 2/11 perf tools: Use scnprintf() in cpu_map__snprint() to prevent overflow
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
cpu_map__snprint() accumulates snprintf() return values in ret. snprintf() returns the number of characters that *would have been written* on truncation, not the actual count. When a fragmented CPU list exceeds the buffer, ret grows past size, causing size - ret to underflow (both are size_t), and subsequent snprintf() calls write past the end of the caller's stack buffer.
Switch to scnprintf() which returns the actual number of characters written, making ret accumulation safe by construction.
--- tools/perf/util/cpumap.c | 18 +++++++++--------- 1 file changed, 9 insertions(+), 9 deletions(-)
tools/perf/util/cpumap.c +9 -9
@@ -692,21 +692,21 @@ size_t cpu_map__snprint(struct perf_cpu_map *map, char *buf, size_t size)if (start ==-1 ) { start = i;if (last) { - ret +=snprintf (buf + ret, size - ret, -"%s%d" ,COMMA , -perf_cpu_map__cpu (map, i).cpu); + ret +=scnprintf (buf + ret, size - ret, +"%s%d" ,COMMA , +perf_cpu_map__cpu (map, i).cpu); } }else if (((i - start) != (cpu.cpu -perf_cpu_map__cpu (map, start).cpu)) || last) {int end = i -1 ;if (start == end) { - ret +=snprintf (buf + ret, size - ret, -"%s%d" ,COMMA , -perf_cpu_map__cpu (map, start).cpu); + ret +=scnprintf (buf + ret, size - ret, +"%s%d" ,COMMA , +perf_cpu_map__cpu (map, start).cpu); }else { - ret +=snprintf (buf + ret, size - ret, -"%s%d-%d" ,COMMA , -perf_cpu_map__cpu (map, start).cpu,perf_cpu_map__cpu (map, end).cpu); + ret +=scnprintf (buf + ret, size - ret, +"%s%d-%d" ,COMMA , +perf_cpu_map__cpu (map, start).cpu,perf_cpu_map__cpu (map, end).cpu); } first = false; start = i;
Signature · 1 line
2.54.0
Arnaldo Carvalho de Melo
maintainerPATCH v2 3/11 perf tools: Use perf_env__get_cpu_topology() in machine__resolve()
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
machine__resolve() accesses env->cpu[al->cpu].socket_id after checking al->cpu >= 0 and env->cpu != NULL, but without validating al->cpu against env->nr_cpus_avail. Since al->cpu comes from the untrusted perf.data sample, a crafted file with a large CPU index causes an out-of-bounds heap read.
Use perf_env__get_cpu_topology() which validates both NULL and bounds.
--- tools/perf/util/event.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-)
tools/perf/util/event.c +7 -2
@@ -14,6 +14,7 @@#include <linux/perf_event.h> #include "cpumap.h" #include "dso.h" +#include "env.h" #include "event.h" #include "debug.h" #include "hist.h"
@@ -835,9 +836,13 @@ int machine__resolve(struct machine *machine, struct addr_location *al,if (al->cpu >=0 ) {struct perf_env *env = machine->env; +struct cpu_topology_map *topo; -if (env && env->cpu) - al->socket = env->cpu[al->cpu].socket_id; +if (env) { + topo =perf_env__get_cpu_topology (env, (struct perf_cpu ){ al->cpu }); +if (topo) + al->socket = topo->socket_id; + } }/* Account for possible out-of-order switch events. */
Signature · 1 line
2.54.0
[email protected]
AI reviewRe: [PATCH 03/11] perf tools: Use perf_env__get_cpu_topology() in machine__resolve()
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider: - [Medium] Silent truncation of a 32-bit CPU index (`al->cpu`) down to a 16-bit `struct perf_cpu` bypasses…
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] Silent truncation of a 32-bit CPU index (al->cpu) down to a 16-bit struct perf_cpu bypasses intended bounds validation for malformed inputs. --
commit 93fba7be75f567784ab8e6d8ce5f966760c09265
Author: Arnaldo Carvalho de Melo <[email protected]>
perf tools: Use perf_env__get_cpu_topology() in machine__resolve()
This commit updates machine__resolve() to use perf_env__get_cpu_topology(), replacing direct array access to prevent out-of-bounds heap reads from untrusted perf.data samples.
Quoted 3 lines
tools/perf/util/event.c +0
-0
[ ... ]
tools/perf/util/event.c int machine__resolve(struct machine *machine, struct addr_location *al, · line 842
Quoted 10 lines
tools/perf/util/event.c +3
-2@@ -835,9 +836,13 @@ int machine__resolve(struct machine *machine, struct addr_location *al,if (al->cpu >=0 ) {struct perf_env *env = machine->env; +struct cpu_topology_map *topo; -if (env && env->cpu) - al->socket = env->cpu[al->cpu].socket_id; +if (env) { + topo =perf_env__get_cpu_topology (env, (struct perf_cpu ){ al->cpu });
[Severity: Medium]
Does this cast silently truncate the 32-bit al->cpu?
Since al->cpu is an s32 from the untrusted sample, and struct perf_cpu uses a 16-bit integer for the cpu index, an attacker providing an index like 65536 would truncate it to 0. Wouldn't this bypass the downstream bounds validation and falsely associate the sample with CPU 0's topology?
Also, for systems with more than 32,767 CPUs, wouldn't this wrap to a negative number and incorrectly fail the bounds check?
Quoted 6 lines
tools/perf/util/event.c +3
-0+if (topo) + al->socket = topo->socket_id; + } }/* Account for possible out-of-order switch events. */
Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=3
Arnaldo Carvalho de Melo
maintainerPATCH v2 4/11 perf mmap: Fix mbind() maxnode vs bitmap allocation mismatch in aio_bind
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
perf_mmap__aio_bind() allocates a node mask bitmap with bitmap_zalloc(node_index + 1) bits, but passes node_index + 2 as the maxnode argument to mbind(). The mbind syscall interprets maxnode as the number of bits to read from the mask.
When node_index + 2 crosses a BITS_PER_LONG boundary (e.g. node_index = 63 on 64-bit), the bitmap occupies 8 bytes but mbind reads 16 — an out-of-bounds read of user heap memory into kernel space.
Allocate node_index + 2 bits to match what mbind will actually read.
--- tools/perf/util/mmap.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
tools/perf/util/mmap.c +2 -1
@@ -113,7 +113,8 @@ static int perf_mmap__aio_bind(struct mmap *map, int idx, struct perf_cpu cpu, iif (node <0 )return 0 ; node_index = node; - node_mask =bitmap_zalloc (node_index +1 ); +/* mbind's maxnode is node_index + 2 — allocate to match */ + node_mask =bitmap_zalloc (node_index +2 );if (!node_mask) {pr_err ("Failed to allocate node mask for mbind: error %m\n" );return -1 ;
Signature · 1 line
2.54.0
Arnaldo Carvalho de Melo
maintainerPATCH v2 5/11 perf tools: NULL bitmap pointers after bitmap_free()
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
Three call sites free bitmaps without NULLing the pointer, risking double-free if the structure is reused or cleanup is called twice:
- mmap__munmap(): map->affinity_mask.bits - record__mmap_cpu_mask_free(): mask->bits - memory_node__delete_nodes(): nodesp[i].set
Set each pointer to NULL after bitmap_free().
--- tools/perf/builtin-record.c | 1 + tools/perf/util/header.c | 4 +++- tools/perf/util/mmap.c | 1 + 3 files changed, 5 insertions(+), 1 deletion(-)
tools/perf/builtin-record.c +1 -0
@@ -3084,6 +3084,7 @@ static int record__mmap_cpu_mask_alloc(struct mmap_cpu_mask *mask, int nr_bits)static void record__mmap_cpu_mask_free (struct mmap_cpu_mask *mask) {bitmap_free (mask->bits); + mask->bits =NULL ; mask->nbits =0 ; }
tools/perf/util/header.c +3 -1
@@ -1481,8 +1481,10 @@ static int memory_node__read(struct memory_node *n, unsigned long idx)static void memory_node__delete_nodes (struct memory_node *nodesp,u64 cnt) { -for (u64 i =0 ; i < cnt; i++) +for (u64 i =0 ; i < cnt; i++) {bitmap_free (nodesp[i].set); + nodesp[i].set =NULL ; + }free (nodesp); }
tools/perf/util/mmap.c +1 -0
@@ -239,6 +239,7 @@ static void perf_mmap__aio_munmap(struct mmap *map __maybe_unused)void mmap__munmap (struct mmap *map) {bitmap_free (map->affinity_mask.bits); + map->affinity_mask.bits =NULL ;zstd_fini (&map->zstd_data);
Signature · 1 line
2.54.0
[email protected]
AI reviewRe: [PATCH 05/11] perf tools: NULL bitmap pointers after bitmap_free()
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider: - [Low] Setting `nodesp[i].set = NULL` in `memory_node__delete_nodes` is a dead store and fails to prevent…
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Low] Setting nodesp[i].set = NULL in memory_node__delete_nodes is a dead store and fails to prevent double-frees because the parent nodesp array is immediately freed. --
commit 96984a1b2dbcef0f5989e5cb40505523a7726dd3
Author: Arnaldo Carvalho de Melo <[email protected]>
perf tools: NULL bitmap pointers after bitmap_free()
This commit modifies three call sites to set bitmap pointers to NULL after they are freed using bitmap_free(). The purpose is to avoid double-free scenarios if the structure is later reused or cleaned up multiple times.
[ ... ]
tools/perf/util/header.c static int memory_node__read(struct memory_node *n, unsigned long idx) · line 1486
Quoted 12 lines
tools/perf/util/header.c +2
-1@@ -1481,8 +1481,10 @@ static int memory_node__read(struct memory_node *n, unsigned long idx)static void memory_node__delete_nodes (struct memory_node *nodesp,u64 cnt) { -for (u64 i =0 ; i < cnt; i++) +for (u64 i =0 ; i < cnt; i++) {bitmap_free (nodesp[i].set); + nodesp[i].set =NULL ;
[Severity: Low]
Since the entire nodesp array is freed immediately after this loop, does nullifying this pointer provide any protection against double-frees?
If a cleanup function were called twice on this structure, wouldn't it result in a use-after-free on the nodesp array itself before the set pointer could even be evaluated?
Quoted 4 lines
tools/perf/util/header.c +1
-0+ }free (nodesp); }
Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=5
Arnaldo Carvalho de Melo
maintainerPATCH v2 6/11 perf sched: Bounds-check prio before test_bit() in timehist
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
timehist_skip_sample() reads prio from untrusted tracepoint data via perf_sample__intval(sample, "prev_prio") without bounds validation. A crafted perf.data with prev_prio >= MAX_PRIO (140) causes test_bit() to read past the end of the prio_bitmap, which is only MAX_PRIO bits.
Add a prio >= 0 && prio < MAX_PRIO check before the test_bit() call. This also makes the != -1 sentinel check explicit as >= 0.
--- tools/lib/perf/TODO | 7 +++++++ tools/perf/builtin-sched.c | 4 +++- 2 files changed, 10 insertions(+), 1 deletion(-)
tools/lib/perf/TODO +7 -0
@@ -11,6 +11,13 @@ together. (x86_64 max is 8192, arm64 is 4096), but NR_CPUS limits keep growing. perf clamps to INT16_MAX in set_max_cpu_num() as a safety net. + - Code simplification: the int16_t forces defensive truncation + checks at every boundary where a wider CPU index (int from + sample->cpu, al->cpu, etc.) is narrowed into struct perf_cpu. + Without these checks, values > 32767 silently wrap to small + positive numbers, bypassing bounds validation. Widening to int + eliminates this entire class of silent truncation bugs and + removes the need for the INT16_MAX clamp in set_max_cpu_num(). - Scope: struct perf_cpu is embedded everywhere — perf_cpu_map__cpu(), perf_cpu_map__min(), perf_cpu_map__max(), perf_cpu_map__has(), the for_each_cpu macros, and all internal callers. The perf_cpu_map
tools/perf/builtin-sched.c +3 -1
@@ -2645,7 +2645,9 @@ static bool timehist_skip_sample(struct perf_sched *sched,else if (evsel__name_is (sample->evsel,"sched:sched_switch" )) prio =perf_sample__intval (sample,"prev_prio" ); -if (prio !=-1 && !test_bit (prio, sched->prio_bitmap)) { +/* prio comes from untrusted tracepoint data — bounds-check before test_bit */ +if (prio >=0 && + (prio >=MAX_PRIO || !test_bit (prio, sched->prio_bitmap))) { rc = true; sched->skipped_samples++; }
Signature · 1 line
2.54.0
[email protected]
AI reviewRe: [PATCH 06/11] perf sched: Bounds-check prio before test_bit() in timehist
Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
Pre-existing issues:
- [Medium] Integer truncation of sample->cpu into int16_t bypasses bounds checks and silently misattributes data
- [Medium] perf sched subcommands fail to register required API callbacks (.attr and .feature) for pipe mode --
commit ccee4e7598c7486d2e3310b0d4a29fa61c041674
Author: Arnaldo Carvalho de Melo <[email protected]>
perf sched: Bounds-check prio before test_bit() in timehist
timehist_skip_sample() reads prio from untrusted tracepoint data via perf_sample__intval(sample, "prev_prio") without bounds validation. A crafted perf.data with prev_prio >= MAX_PRIO (140) causes test_bit() to read past the end of the prio_bitmap, which is only MAX_PRIO bits. Add a prio >= 0 && prio < MAX_PRIO check before the test_bit() call.
Quoted 5 lines
tools/perf/builtin-sched.c +0
-0
tools/perf/builtin-sched.c static bool timehist_skip_sample(struct perf_sched *sched, · line 2653
Quoted 10 lines
tools/perf/builtin-sched.c +3
-1@@ -2645,7 +2645,9 @@ static bool timehist_skip_sample(struct perf_sched *sched,else if (evsel__name_is (sample->evsel,"sched:sched_switch" )) prio =perf_sample__intval (sample,"prev_prio" ); -if (prio !=-1 && !test_bit (prio, sched->prio_bitmap)) { +/* prio comes from untrusted tracepoint data — bounds-check before test_bit */ +if (prio >=0 && + (prio >=MAX_PRIO || !test_bit (prio, sched->prio_bitmap))) { rc = true; sched->skipped_samples++; }
[Severity: Medium]
This is a pre-existing issue, but I noticed that the sample->cpu value is narrowed to a 16-bit integer during struct initialization in map_switch_event() and perf_timehist__process_sample() in tools/perf/builtin-sched.c.
Does this narrowing regression allow wrapped out-of-bounds cpu values to bypass the subsequent validation check?
For example in map_switch_event():
struct perf_cpu this_cpu = {
.cpu = sample->cpu,
};
/* perf.data is untrusted input — CPU may be absent or corrupted */
if (this_cpu.cpu >= MAX_CPUS || this_cpu.cpu < 0) {If sample->cpu is larger than 32767, it will silently wrap, passing the bounds check but potentially misattributing data. A similar pattern exists in perf_timehist__process_sample().
[Severity: Medium]
This isn't a bug introduced by this patch, but looking at cmd_sched() and perf_sched__timehist() in tools/perf/builtin-sched.c, it appears the perf sched subcommands fail to register required API callbacks like .attr and .feature for pipe mode.
In cmd_sched():
perf_tool__init(&sched.tool, /*ordered_events=*/true); sched.tool.sample = perf_sched__process_tracepoint_sample; sched.tool.comm = perf_sched__process_comm; sched.tool.namespaces = perf_event__process_namespaces; sched.tool.lost = perf_event__process_lost; sched.tool.fork = perf_sched__process_fork_event;
And in perf_sched__timehist():
sched->tool.lost = process_lost; sched->tool.attr = perf_event__process_attr; sched->tool.tracing_data = perf_event__process_tracing_data; sched->tool.build_id = perf_event__process_build_id;
Without the .attr callback, no evsel structures are created, and without the .feature callback, perf_env remains unpopulated in pipe mode.
Does this omission lead to a regression where dropped events or uninitialized system metadata occur when processing pipelined inputs?
Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=6
Arnaldo Carvalho de Melo
maintainerPATCH v2 7/11 perf sched: Fix idle-hist callchain display using wrong rb_first variant
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
timehist_print_idlehist_callchain() calls rb_first_cached() on sorted_root, but the sort function (callchain_param.sort) populates it via rb_insert_color() on the plain rb_root member — not the cached variant. This means rb_leftmost is never set, so rb_first_cached() always returns NULL and the entire callchain summary is silently dropped from --idle-hist output.
Use rb_first(&root->rb_root) to match how the tree was populated.
--- tools/perf/builtin-sched.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
tools/perf/builtin-sched.c +2 -1
@@ -3130,7 +3130,8 @@ static size_t timehist_print_idlehist_callchain(struct rb_root_cached *root)size_t ret =0 ;FILE *fp = stdout;struct callchain_node *chain; -struct rb_node *rb_node =rb_first_cached (root); +/* sort() uses rb_insert_color() on rb_root, not rb_root_cached */ +struct rb_node *rb_node =rb_first (&root->rb_root);printf (" %16s %8s %s\n" ,"Idle time (msec)" ,"Count" ,"Callchains" ); printf(" %.16s %.8s %.50s\n", graph_dotted_line, graph_dotted_line,
Signature · 1 line
2.54.0
Arnaldo Carvalho de Melo
maintainerPATCH v2 8/11 perf tools: Add O_CLOEXEC to open() calls in DSO and ELF code
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
open() calls in dso.c and symbol-elf.c omit O_CLOEXEC, which leaks file descriptors to child processes spawned during symbol resolution (e.g., addr2line, objdump). This can exhaust the fd limit during long profiling sessions or when processing many DSOs.
Add O_CLOEXEC to all open() calls in both files (12 call sites).
--- tools/perf/util/dso.c | 4 ++-- tools/perf/util/symbol-elf.c | 20 ++++++++++---------- 2 files changed, 12 insertions(+), 12 deletions(-)
tools/perf/util/dso.c +2 -2
@@ -344,7 +344,7 @@ int filename__decompress(const char *name, char *pathname,* descriptor to the uncompressed file. */ if (!compressions[comp].is_compressed (name)) -return open (name,O_RDONLY ); +return open (name,O_RDONLY |O_CLOEXEC ); fd =mkstemp (tmpbuf);if (fd <0 ) {
@@ -1911,7 +1911,7 @@ static const u8 *__dso__read_symbol(struct dso *dso, const char *symfs_filename,int saved_errno;nsinfo__mountns_enter (dso__nsinfo (dso), &nsc); - fd =open (symfs_filename,O_RDONLY ); + fd =open (symfs_filename,O_RDONLY |O_CLOEXEC ); saved_errno = errno;nsinfo__mountns_exit (&nsc);if (fd <0 ) {
tools/perf/util/symbol-elf.c +10 -10
@@ -217,7 +217,7 @@ bool filename__has_section(const char *filename, const char *sec)GElf_Shdr shdr;bool found = false; - fd =open (filename,O_RDONLY ); + fd =open (filename,O_RDONLY |O_CLOEXEC );if (fd <0 )return false;
@@ -872,7 +872,7 @@ static int read_build_id(const char *filename, struct build_id *bid)if (size <BUILD_ID_SIZE ) goto out; - fd =open (filename,O_RDONLY ); + fd =open (filename,O_RDONLY |O_CLOEXEC );if (fd <0 ) goto out;
@@ -935,7 +935,7 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)size_t size =sizeof (bid->data);int fd, err =-1 ; - fd =open (filename,O_RDONLY ); + fd =open (filename,O_RDONLY |O_CLOEXEC );if (fd <0 ) goto out;
@@ -995,7 +995,7 @@ int filename__read_debuglink(const char *filename, char *debuglink,if (err >=0 ) goto out; - fd =open (filename,O_RDONLY ); + fd =open (filename,O_RDONLY |O_CLOEXEC );if (fd <0 ) goto out;
@@ -1153,7 +1153,7 @@ int symsrc__init(struct symsrc *ss, struct dso *dso, const char *name, type =dso__symtab_type (dso); }else { - fd =open (name,O_RDONLY ); + fd =open (name,O_RDONLY |O_CLOEXEC );if (fd <0 ) { *dso__load_errno (dso) = errno;return -1 ;
@@ -1952,7 +1952,7 @@ static int kcore__open(struct kcore *kcore, const char *filename) {GElf_Ehdr *ehdr; - kcore->fd =open (filename,O_RDONLY ); + kcore->fd =open (filename,O_RDONLY |O_CLOEXEC );if (kcore->fd ==-1 )return -1 ;
@@ -1985,7 +1985,7 @@ static int kcore__init(struct kcore *kcore, char *filename, int elfclass,if (temp) kcore->fd =mkstemp (filename);else - kcore->fd =open (filename,O_WRONLY |O_CREAT |O_EXCL ,0400 ); + kcore->fd =open (filename,O_WRONLY |O_CREAT |O_EXCL |O_CLOEXEC ,0400 );if (kcore->fd ==-1 )return -1 ;
@@ -2461,11 +2461,11 @@ static int kcore_copy__compare_files(const char *from_filename, {int from, to, err =-1 ; - from =open (from_filename,O_RDONLY ); + from =open (from_filename,O_RDONLY |O_CLOEXEC );if (from <0 )return -1 ; - to =open (to_filename,O_RDONLY ); + to =open (to_filename,O_RDONLY |O_CLOEXEC );if (to <0 ) goto out_close_from;
@@ -2883,7 +2883,7 @@ int get_sdt_note_list(struct list_head *head, const char *target)Elf *elf;int fd, ret; - fd =open (target,O_RDONLY ); + fd =open (target,O_RDONLY |O_CLOEXEC );if (fd <0 )return -EBADF ;
Signature · 1 line
2.54.0
[email protected]
AI reviewRe: [PATCH 08/11] perf tools: Add O_CLOEXEC to open() calls in DSO and ELF code
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
Pre-existing issues:
- [High] The patch incompletely fixes the file descriptor leak because it misses multiple mkstemp() calls that do not set O_CLOEXEC. --
commit 76a844002251898e2934fafe58f422e4df6f5517
Author: Arnaldo Carvalho de Melo <[email protected]>
perf tools: Add O_CLOEXEC to open() calls in DSO and ELF code
This patch adds O_CLOEXEC to open() calls in dso.c and symbol-elf.c to prevent file descriptors from leaking to child processes spawned during symbol resolution. This aims to prevent fd exhaustion during long profiling sessions or when processing many DSOs.
tools/perf/util/dso.c int filename__decompress(const char *name, char *pathname, · line 349
Quoted 12 lines
tools/perf/util/dso.c +1
-1@@ -344,7 +344,7 @@ int filename__decompress(const char *name, char *pathname,* descriptor to the uncompressed file. */ if (!compressions[comp].is_compressed (name)) -return open (name,O_RDONLY ); +return open (name,O_RDONLY |O_CLOEXEC ); fd =mkstemp (tmpbuf);
[Severity: High]
This is a pre-existing issue, but does mkstemp() also leak file descriptors to child processes?
Since mkstemp() does not automatically set the O_CLOEXEC flag, the file descriptor created here for the decompressed kernel module might still remain open and leak to child processes (like addr2line) spawned during symbol resolution.
Should a subsequent fcntl(fd, F_SETFD, FD_CLOEXEC) be added to fully resolve the file descriptor exhaustion issue this patch targets?
Quoted 1 line
tools/perf/util/dso.c +0
-0if (fd <0 ) {
[ ... ]
Quoted 4 lines
tools/perf/util/symbol-elf.c +0
-0
[ ... ]
tools/perf/util/symbol-elf.c static int kcore__init(struct kcore *kcore, char *filename, int elfclass, · line 1986
Quoted 3 lines
tools/perf/util/symbol-elf.c +0
-0@@ -1985,7 +1985,7 @@ static int kcore__init(struct kcore *kcore, char *filename, int elfclass,if (temp) kcore->fd =mkstemp (filename);
[Severity: High]
This is a pre-existing issue, but this mkstemp() call also creates a temporary file for kcore without setting O_CLOEXEC.
Additionally, looking at read_gnu_debugdata() in the same file, there is another call to mkstemp() used for extracted .gnu_debugdata:
tools/perf/util/symbol-elf.c:read_gnu_debugdata() {
...
temp_fd = mkstemp(temp_filename);
...
}Do these mkstemp() descriptors also need to have the FD_CLOEXEC flag set to prevent them from leaking to child processes?
Quoted 5 lines
tools/perf/util/symbol-elf.c +1
-1else - kcore->fd =open (filename,O_WRONLY |O_CREAT |O_EXCL ,0400 ); + kcore->fd =open (filename,O_WRONLY |O_CREAT |O_EXCL |O_CLOEXEC ,0400 );if (kcore->fd ==-1 )return -1 ;
Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=8
Arnaldo Carvalho de Melo
maintainerPATCH v2 9/11 perf bpf: Use scnprintf() in snprintf_hex() and synthesize_bpf_prog_name()
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
Both functions accumulate formatted output via ret += snprintf(buf + ret, size - ret, ...). If the buffer is too small and snprintf() returns more than the remaining space, ret exceeds size and the next 'size - ret' underflows, causing snprintf() to write past the buffer end.
Switch to scnprintf() which returns the actual number of bytes written, making the accumulation safe.
--- tools/perf/util/bpf-event.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-)
tools/perf/util/bpf-event.c +6 -5
@@ -36,7 +36,7 @@ static int snprintf_hex(char *buf, size_t size, unsigned char *data, size_t len)size_t i;for (i =0 ; i < len; i++) - ret +=snprintf (buf + ret, size - ret,"%02x" , data[i]); + ret +=scnprintf (buf + ret, size - ret,"%02x" , data[i]);return ret; }
@@ -140,7 +140,7 @@ static int synthesize_bpf_prog_name(char *buf, int size,const struct btf_type *t;int name_len; - name_len =snprintf (buf, size,"bpf_prog_" ); + name_len =scnprintf (buf, size,"bpf_prog_" ); name_len +=snprintf_hex (buf + name_len, size - name_len, prog_tags[sub_id],BPF_TAG_SIZE );if (btf) {
@@ -153,9 +153,10 @@ static int synthesize_bpf_prog_name(char *buf, int size, short_name = info->name; }else short_name ="F" ; -if (short_name) - name_len +=snprintf (buf + name_len, size - name_len, -"_%s" , short_name); +if (short_name) { + name_len +=scnprintf (buf + name_len, size - name_len, +"_%s" , short_name); + }return name_len; }
Signature · 1 line
2.54.0
Arnaldo Carvalho de Melo
maintainerPATCH v2 10/11 perf hists: Fix snprintf() in hists__scnprintf_title() UID filter path
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
hists__scnprintf_title() accumulates formatted output into a buffer using scnprintf() for all filter clauses except the UID filter, which uses snprintf(). If the buffer fills up and snprintf() returns more than the remaining space, printed exceeds size and the next 'size - printed' underflows, causing later scnprintf() calls to write past the buffer.
Switch the UID filter clause to scnprintf() to match the rest of the function.
--- tools/perf/util/hist.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-)
tools/perf/util/hist.c +4 -3
@@ -2963,9 +2963,10 @@ int __hists__scnprintf_title(struct hists *hists, char *bf, size_t size, bool sh ev_name, sample_freq_str, enable_ref ? ref : " ", nr_events); -if (hists->uid_filter_str) - printed +=snprintf (bf + printed, size - printed, -", UID: %s" , hists->uid_filter_str); +if (hists->uid_filter_str) { + printed +=scnprintf (bf + printed, size - printed, +", UID: %s" , hists->uid_filter_str); + }if (thread) {if (hists__has (hists, thread)) { printed += scnprintf(bf + printed, size - printed,
Signature · 1 line
2.54.0
[email protected]
AI reviewRe: [PATCH 10/11] perf hists: Fix snprintf() in hists__scnprintf_title() UID filter path
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
Pre-existing issues:
- [High] Stack buffer overflow in perf_pmu__for_each_event due to unchecked snprintf accumulation. --
commit 6971cf80165017520e4fc2dbee2d01c9c2ff4efb
Author: Arnaldo Carvalho de Melo <[email protected]>
perf hists: Fix snprintf() in hists__scnprintf_title() UID filter path
This patch fixes a potential buffer overflow by switching an unchecked snprintf call to scnprintf. The original code accumulated the return value of snprintf, which could underflow the remaining size on subsequent calls.
Quoted 5 lines
tools/perf/util/hist.c +0
-0
tools/perf/util/hist.c int __hists__scnprintf_title(struct hists *hists, char *bf, size_t size, bool sh · line 2969
Quoted 10 lines
tools/perf/util/hist.c +4
-3@@ -2963,9 +2963,10 @@ int __hists__scnprintf_title(struct hists *hists, char *bf, size_t size, bool sh ev_name, sample_freq_str, enable_ref ? ref : " ", nr_events); -if (hists->uid_filter_str) - printed +=snprintf (bf + printed, size - printed, -", UID: %s" , hists->uid_filter_str); +if (hists->uid_filter_str) { + printed +=scnprintf (bf + printed, size - printed, +", UID: %s" , hists->uid_filter_str); + }
[Severity: High]
This is a pre-existing issue, but while you are fixing this unchecked snprintf accumulation here, there appears to be another instance of the exact same pattern in tools/perf/util/pmu.c:perf_pmu__for_each_event().
When accumulating buf_used across consecutive snprintf calls, if the first formatted string is large enough:
tools/perf/util/pmu.c:perf_pmu__for_each_event() {
...
char buf[1024];
...
buf_used += snprintf(buf + buf_used, sizeof(buf) - buf_used,
"%G%s", event->scale, event->unit) + 1;
...
buf_used += snprintf(buf + buf_used, sizeof(buf) - buf_used,
"%.*s/%s/", (int)pmu_name_len, info.pmu_name,
event->terms) + 1;
...
}If the first snprintf returns a value causing buf_used to exceed sizeof(buf), the second snprintf will execute with an underflowed limit and an out-of-bounds destination pointer.
Could this write past the stack buffer if a crafted PMU sysfs entry is encountered?
Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=10
Arnaldo Carvalho de Melo
maintainerPATCH v2 11/11 perf tools: Use scnprintf() in build_id__snprintf() and hwmon read_events()
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
build_id__snprintf() and hwmon_pmu__read_events() accumulate formatted output via snprintf(), which returns the would-have-been-written count on truncation. In build_id__snprintf(), this inflates the return value beyond the buffer size. In hwmon_pmu__read_events(), len overshoots out_buf_len and the next 'out_buf_len - len' underflows.
Switch both to scnprintf() which returns actual bytes written.
--- tools/perf/util/build-id.c | 2 +- tools/perf/util/hwmon_pmu.c | 12 ++++++------ 2 files changed, 7 insertions(+), 7 deletions(-)
tools/perf/util/build-id.c +1 -1
@@ -94,7 +94,7 @@ int build_id__snprintf(const struct build_id *build_id, char *bf, size_t bf_size }for (size_t i =0 ; i < build_id->size && offs < bf_size; ++i) - offs +=snprintf (bf + offs, bf_size - offs,"%02x" , build_id->data[i]); + offs +=scnprintf (bf + offs, bf_size - offs,"%02x" , build_id->data[i]);return offs; }
tools/perf/util/hwmon_pmu.c +6 -6
@@ -442,12 +442,12 @@ static size_t hwmon_pmu__describe_items(struct hwmon_pmu *hwm, char *out_buf, si buf[read_len] ='\0' ; val =strtoll (buf,/*endptr=*/ NULL ,10 ); - len +=snprintf (out_buf + len, out_buf_len - len,"%s%s%s=%g%s" , - len ==0 ?" " :", " , - hwmon_item_strs[bit], - is_alarm ?"_alarm" :"" , - (double )val /1000.0 , - hwmon_units[key.type]); + len +=scnprintf (out_buf + len, out_buf_len - len,"%s%s%s=%g%s" , + len ==0 ?" " :", " , + hwmon_item_strs[bit], + is_alarm ?"_alarm" :"" , + (double )val /1000.0 , + hwmon_units[key.type]); }close (fd); }
Signature · 1 line
2.54.0
[email protected]
AI reviewRe: [PATCH 11/11] perf tools: Use scnprintf() in build_id__snprintf() and hwmon read_events()
Thank you for your contribution! Sashiko AI review found 3 potential issue(s) to consider:
Thank you for your contribution! Sashiko AI review found 3 potential issue(s) to consider:
Pre-existing issues:
- [High] Missing size adjustment in read() and pread() size arguments leads to out-of-bounds null termination on stack buffers.
- [High] If the read() system call returns <= 0, uninitialized stack memory is evaluated and potentially duplicated, leading to an information leak or crash.
- [Medium] A flawed check if (fd > 0) fails to process or close file descriptor 0, causing a resource leak if openat returns 0.
--
commit 63dc76c4229ffb5cbb25da7f0bdc34b17ef66fae
Author: Arnaldo Carvalho de Melo <[email protected]>
perf tools: Use scnprintf() in build_id__snprintf() and hwmon read_events()
build_id__snprintf() and hwmon_pmu__read_events() accumulate formatted output via snprintf(), which returns the would-have-been-written count on truncation. In hwmon_pmu__read_events(), len overshoots out_buf_len and the next 'out_buf_len - len' underflows. Switch both to scnprintf() which returns actual bytes written.
[ ... ]
tools/perf/util/hwmon_pmu.c static size_t hwmon_pmu__describe_items(struct hwmon_pmu *hwm, char *out_buf, si · line 443
Quoted 7 lines
tools/perf/util/hwmon_pmu.c +0
-0@@ -442,12 +442,12 @@ static size_t hwmon_pmu__describe_items(struct hwmon_pmu *hwm, char *out_buf, si buf[read_len] ='\0' ;
^^^^^^^^^^^^^^^^^^^^
[Severity: High]
This isn't a bug introduced by this patch, but doesn't this result in an out-of-bounds write?
Since read() can return up to sizeof(buf), if the sysfs file contents completely fill the buffer (i.e. read_len == sizeof(buf)), writing the null terminator here will overwrite one byte past the end of the buf array on the stack.
Could we pass sizeof(buf) - 1 to read() to avoid this?
This same pre-existing issue seems to be present in hwmon_pmu__read_events() and evsel__hwmon_pmu_read() as well, where read() and pread() are called with the full buffer size and similarly null-terminated.
[Severity: Medium]
Also a pre-existing issue in hwmon_pmu__describe_items() just above this hunk, can this leak a file descriptor?
hwmon_pmu__describe_items() opens a sysfs file using openat(), and then checks:
fd = openat(dir, buf, O_RDONLY);
if (fd > 0) {If the tool runs in an environment where standard input has been closed, openat() can successfully return 0. The current logic treats 0 as an error, skips file processing, and fails to call close(fd), causing a resource leak.
Should this check be if (fd >= 0) instead?
[Severity: High]
Another pre-existing issue in hwmon_pmu__read_events(), can this leak uninitialized stack memory?
If read() returns 0 or less (for example, on an error or empty file), read_len will be 0 or less. The code skips null-terminating the buffer and proceeds to check:
if (buf[0] == '\0') {
...
value->label = strdup(buf);Since char buf[128] is uninitialized, wouldn't evaluating buf[0] and passing buf to strdup() result in reading and copying garbage stack memory until it hits a random null byte?
Quoted 16 lines
tools/perf/util/hwmon_pmu.c +6
-6val =strtoll (buf,/*endptr=*/ NULL ,10 ); - len +=snprintf (out_buf + len, out_buf_len - len,"%s%s%s=%g%s" , - len ==0 ?" " :", " , - hwmon_item_strs[bit], - is_alarm ?"_alarm" :"" , - (double )val /1000.0 , - hwmon_units[key.type]); + len +=scnprintf (out_buf + len, out_buf_len - len,"%s%s%s=%g%s" , + len ==0 ?" " :", " , + hwmon_item_strs[bit], + is_alarm ?"_alarm" :"" , + (double )val /1000.0 , + hwmon_units[key.type]); }close (fd); }
Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=11