[GIT PULL] perf fixes

From: Ingo Molnar
Date: Sat Oct 14 2017 - 12:04:51 EST

Next message: Ingo Molnar: "[GIT PULL] RAS fixes"
Previous message: tedheadster: "[REGRESSION]: x86/idt: Move early IDT setup out of 32-bit asm"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Linus,

Please pull the latest perf-urgent-for-linus git tree from:

git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git perf-urgent-for-linus

# HEAD: 47a74bdcbfeff543f706dc0e385eebbb5d655ed2 Merge tag 'perf-urgent-for-mingo-4.14-20171010' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux into perf/urgent

Some tooling fixes plus three kernel fixes: a memory leak fix, a statistics fix
and a crash fix.

Thanks,

Ingo

------------------>
Arnaldo Carvalho de Melo (1):
tools include uapi bpf.h: Sync kernel ABI header with tooling header

Colin Ian King (1):
perf/x86/intel/uncore: Fix memory leaks on allocation failures

Mark Rutland (1):
perf pmu: Unbreak perf record for arm/arm64 with events with explicit PMU

Mark Santaniello (1):
perf script: Add missing separator for "-F ip,brstack" (and brstackoff)

Ravi Bangoria (1):
perf callchain: Compare dsos (as well) for CCKEY_FUNCTION

Will Deacon (1):
perf/core: Avoid freeing static PMU contexts when PMU is unregistered

leilei.lin (1):
perf/core: Fix cgroup time when scheduling descendants

arch/x86/events/intel/uncore.c | 12 +++++++--
kernel/events/core.c | 10 +++++++-
tools/include/uapi/linux/bpf.h | 2 +-
tools/perf/builtin-script.c | 4 +--
tools/perf/util/callchain.c | 6 ++++-
tools/perf/util/parse-events.c | 9 ++++---
tools/perf/util/pmu.c | 56 +++++++++++++++++++++++++++++++-----------
tools/perf/util/pmu.h | 1 +
8 files changed, 74 insertions(+), 26 deletions(-)

diff --git a/arch/x86/events/intel/uncore.c b/arch/x86/events/intel/uncore.c
index 1c5390f1cf09..d45e06346f14 100644
--- a/arch/x86/events/intel/uncore.c
+++ b/arch/x86/events/intel/uncore.c
@@ -822,7 +822,7 @@ static int __init uncore_type_init(struct intel_uncore_type *type, bool setid)
pmus[i].type = type;
pmus[i].boxes = kzalloc(size, GFP_KERNEL);
if (!pmus[i].boxes)
- return -ENOMEM;
+ goto err;
}

type->pmus = pmus;
@@ -836,7 +836,7 @@ static int __init uncore_type_init(struct intel_uncore_type *type, bool setid)
attr_group = kzalloc(sizeof(struct attribute *) * (i + 1) +
sizeof(*attr_group), GFP_KERNEL);
if (!attr_group)
- return -ENOMEM;
+ goto err;

attrs = (struct attribute **)(attr_group + 1);
attr_group->name = "events";
@@ -849,7 +849,15 @@ static int __init uncore_type_init(struct intel_uncore_type *type, bool setid)
}

type->pmu_group = &uncore_pmu_attr_group;
+
return 0;
+
+err:
+ for (i = 0; i < type->num_boxes; i++)
+ kfree(pmus[i].boxes);
+ kfree(pmus);
+
+ return -ENOMEM;
}

static int __init
diff --git a/kernel/events/core.c b/kernel/events/core.c
index 6bc21e202ae4..9d93db81fa36 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -662,7 +662,7 @@ static inline void update_cgrp_time_from_event(struct perf_event *event)
/*
* Do not update time when cgroup is not active
*/
- if (cgrp == event->cgrp)
+ if (cgroup_is_descendant(cgrp->css.cgroup, event->cgrp->css.cgroup))
__update_cgrp_time(event->cgrp);
}

@@ -8955,6 +8955,14 @@ static struct perf_cpu_context __percpu *find_pmu_context(int ctxn)

static void free_pmu_context(struct pmu *pmu)
{
+ /*
+ * Static contexts such as perf_sw_context have a global lifetime
+ * and may be shared between different PMUs. Avoid freeing them
+ * when a single PMU is going away.
+ */
+ if (pmu->task_ctx_nr > perf_invalid_context)
+ return;
+
mutex_lock(&pmus_lock);
free_percpu(pmu->pmu_cpu_context);
mutex_unlock(&pmus_lock);
diff --git a/tools/include/uapi/linux/bpf.h b/tools/include/uapi/linux/bpf.h
index 43ab5c402f98..f90860d1f897 100644
--- a/tools/include/uapi/linux/bpf.h
+++ b/tools/include/uapi/linux/bpf.h
@@ -312,7 +312,7 @@ union bpf_attr {
* jump into another BPF program
* @ctx: context pointer passed to next program
* @prog_array_map: pointer to map which type is BPF_MAP_TYPE_PROG_ARRAY
- * @index: index inside array that selects specific program to run
+ * @index: 32-bit index inside array that selects specific program to run
* Return: 0 on success or negative error
*
* int bpf_clone_redirect(skb, ifindex, flags)
diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c
index 3d4c3b5e1868..0c977b6e0f8b 100644
--- a/tools/perf/builtin-script.c
+++ b/tools/perf/builtin-script.c
@@ -586,7 +586,7 @@ static void print_sample_brstack(struct perf_sample *sample,
thread__find_addr_map(thread, sample->cpumode, MAP__FUNCTION, to, &alt);
}

- printf("0x%"PRIx64, from);
+ printf(" 0x%"PRIx64, from);
if (PRINT_FIELD(DSO)) {
printf("(");
map__fprintf_dsoname(alf.map, stdout);
@@ -681,7 +681,7 @@ static void print_sample_brstackoff(struct perf_sample *sample,
if (alt.map && !alt.map->dso->adjust_symbols)
to = map__map_ip(alt.map, to);

- printf("0x%"PRIx64, from);
+ printf(" 0x%"PRIx64, from);
if (PRINT_FIELD(DSO)) {
printf("(");
map__fprintf_dsoname(alf.map, stdout);
diff --git a/tools/perf/util/callchain.c b/tools/perf/util/callchain.c
index be09d77cade0..a971caf3759d 100644
--- a/tools/perf/util/callchain.c
+++ b/tools/perf/util/callchain.c
@@ -685,6 +685,8 @@ static enum match_result match_chain(struct callchain_cursor_node *node,
{
struct symbol *sym = node->sym;
u64 left, right;
+ struct dso *left_dso = NULL;
+ struct dso *right_dso = NULL;

if (callchain_param.key == CCKEY_SRCLINE) {
enum match_result match = match_chain_srcline(node, cnode);
@@ -696,12 +698,14 @@ static enum match_result match_chain(struct callchain_cursor_node *node,
if (cnode->ms.sym && sym && callchain_param.key == CCKEY_FUNCTION) {
left = cnode->ms.sym->start;
right = sym->start;
+ left_dso = cnode->ms.map->dso;
+ right_dso = node->map->dso;
} else {
left = cnode->ip;
right = node->ip;
}

- if (left == right) {
+ if (left == right && left_dso == right_dso) {
if (node->branch) {
cnode->branch_count++;

diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c
index f6257fb4f08c..39b15968eab1 100644
--- a/tools/perf/util/parse-events.c
+++ b/tools/perf/util/parse-events.c
@@ -309,10 +309,11 @@ static char *get_config_name(struct list_head *head_terms)
static struct perf_evsel *
__add_event(struct list_head *list, int *idx,
struct perf_event_attr *attr,
- char *name, struct cpu_map *cpus,
+ char *name, struct perf_pmu *pmu,
struct list_head *config_terms, bool auto_merge_stats)
{
struct perf_evsel *evsel;
+ struct cpu_map *cpus = pmu ? pmu->cpus : NULL;

event_attr_init(attr);

@@ -323,7 +324,7 @@ __add_event(struct list_head *list, int *idx,
(*idx)++;
evsel->cpus = cpu_map__get(cpus);
evsel->own_cpus = cpu_map__get(cpus);
- evsel->system_wide = !!cpus;
+ evsel->system_wide = pmu ? pmu->is_uncore : false;
evsel->auto_merge_stats = auto_merge_stats;

if (name)
@@ -1233,7 +1234,7 @@ static int __parse_events_add_pmu(struct parse_events_state *parse_state,

if (!head_config) {
attr.type = pmu->type;
- evsel = __add_event(list, &parse_state->idx, &attr, NULL, pmu->cpus, NULL, auto_merge_stats);
+ evsel = __add_event(list, &parse_state->idx, &attr, NULL, pmu, NULL, auto_merge_stats);
return evsel ? 0 : -ENOMEM;
}

@@ -1254,7 +1255,7 @@ static int __parse_events_add_pmu(struct parse_events_state *parse_state,
return -EINVAL;

evsel = __add_event(list, &parse_state->idx, &attr,
- get_config_name(head_config), pmu->cpus,
+ get_config_name(head_config), pmu,
&config_terms, auto_merge_stats);
if (evsel) {
evsel->unit = info.unit;
diff --git a/tools/perf/util/pmu.c b/tools/perf/util/pmu.c
index ac16a9db1fb5..1c4d7b4e4fb5 100644
--- a/tools/perf/util/pmu.c
+++ b/tools/perf/util/pmu.c
@@ -470,17 +470,36 @@ static void pmu_read_sysfs(void)
closedir(dir);
}

+static struct cpu_map *__pmu_cpumask(const char *path)
+{
+ FILE *file;
+ struct cpu_map *cpus;
+
+ file = fopen(path, "r");
+ if (!file)
+ return NULL;
+
+ cpus = cpu_map__read(file);
+ fclose(file);
+ return cpus;
+}
+
+/*
+ * Uncore PMUs have a "cpumask" file under sysfs. CPU PMUs (e.g. on arm/arm64)
+ * may have a "cpus" file.
+ */
+#define CPUS_TEMPLATE_UNCORE "%s/bus/event_source/devices/%s/cpumask"
+#define CPUS_TEMPLATE_CPU "%s/bus/event_source/devices/%s/cpus"
+
static struct cpu_map *pmu_cpumask(const char *name)
{
- struct stat st;
char path[PATH_MAX];
- FILE *file;
struct cpu_map *cpus;
const char *sysfs = sysfs__mountpoint();
const char *templates[] = {
- "%s/bus/event_source/devices/%s/cpumask",
- "%s/bus/event_source/devices/%s/cpus",
- NULL
+ CPUS_TEMPLATE_UNCORE,
+ CPUS_TEMPLATE_CPU,
+ NULL
};
const char **template;

@@ -489,20 +508,25 @@ static struct cpu_map *pmu_cpumask(const char *name)

for (template = templates; *template; template++) {
snprintf(path, PATH_MAX, *template, sysfs, name);
- if (stat(path, &st) == 0)
- break;
+ cpus = __pmu_cpumask(path);
+ if (cpus)
+ return cpus;
}

- if (!*template)
- return NULL;
+ return NULL;
+}

- file = fopen(path, "r");
- if (!file)
- return NULL;
+static bool pmu_is_uncore(const char *name)
+{
+ char path[PATH_MAX];
+ struct cpu_map *cpus;
+ const char *sysfs = sysfs__mountpoint();

- cpus = cpu_map__read(file);
- fclose(file);
- return cpus;
+ snprintf(path, PATH_MAX, CPUS_TEMPLATE_UNCORE, sysfs, name);
+ cpus = __pmu_cpumask(path);
+ cpu_map__put(cpus);
+
+ return !!cpus;
}

/*
@@ -617,6 +641,8 @@ static struct perf_pmu *pmu_lookup(const char *name)

pmu->cpus = pmu_cpumask(name);

+ pmu->is_uncore = pmu_is_uncore(name);
+
INIT_LIST_HEAD(&pmu->format);
INIT_LIST_HEAD(&pmu->aliases);
list_splice(&format, &pmu->format);
diff --git a/tools/perf/util/pmu.h b/tools/perf/util/pmu.h
index 389e9729331f..fe0de0502ce2 100644
--- a/tools/perf/util/pmu.h
+++ b/tools/perf/util/pmu.h
@@ -22,6 +22,7 @@ struct perf_pmu {
char *name;
__u32 type;
bool selectable;
+ bool is_uncore;
struct perf_event_attr *default_config;
struct cpu_map *cpus;
struct list_head format; /* HEAD struct perf_pmu_format -> list */

Next message: Ingo Molnar: "[GIT PULL] RAS fixes"
Previous message: tedheadster: "[REGRESSION]: x86/idt: Move early IDT setup out of 32-bit asm"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]