[for-next][PATCH] trace_uprobe: support reference counter in fd-based uprobe

From: Steven Rostedt
Date: Thu Oct 11 2018 - 22:37:45 EST



Added for the 4.20/5.0 queue.

-- Steve


git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace.git
for-next

Head SHA1: a6ca88b241d5e929e6e60b12ad8cd288f0ffa256


Song Liu (1):
trace_uprobe: support reference counter in fd-based uprobe

----
include/linux/trace_events.h | 3 ++-
kernel/events/core.c | 49 ++++++++++++++++++++++++++++++++---------
kernel/trace/trace_event_perf.c | 7 +++---
kernel/trace/trace_probe.h | 3 ++-
kernel/trace/trace_uprobe.c | 4 +++-
5 files changed, 50 insertions(+), 16 deletions(-)
---------------------------
commit a6ca88b241d5e929e6e60b12ad8cd288f0ffa256
Author: Song Liu <songliubraving@xxxxxx>
Date: Mon Oct 1 22:36:36 2018 -0700

trace_uprobe: support reference counter in fd-based uprobe

This patch enables uprobes with reference counter in fd-based uprobe.
Highest 32 bits of perf_event_attr.config is used to stored offset
of the reference count (semaphore).

Format information in /sys/bus/event_source/devices/uprobe/format/ is
updated to reflect this new feature.

Link: http://lkml.kernel.org/r/20181002053636.1896903-1-songliubraving@xxxxxx

Cc: Oleg Nesterov <oleg@xxxxxxxxxx>
Acked-by: Peter Zijlstra (Intel) <peterz@xxxxxxxxxxxxx>
Reviewed-and-tested-by: Ravi Bangoria <ravi.bangoria@xxxxxxxxxxxxx>
Signed-off-by: Song Liu <songliubraving@xxxxxx>
Signed-off-by: Steven Rostedt (VMware) <rostedt@xxxxxxxxxxx>

diff --git a/include/linux/trace_events.h b/include/linux/trace_events.h
index 78a010e19ed4..4130a5497d40 100644
--- a/include/linux/trace_events.h
+++ b/include/linux/trace_events.h
@@ -575,7 +575,8 @@ extern int bpf_get_kprobe_info(const struct perf_event *event,
bool perf_type_tracepoint);
#endif
#ifdef CONFIG_UPROBE_EVENTS
-extern int perf_uprobe_init(struct perf_event *event, bool is_retprobe);
+extern int perf_uprobe_init(struct perf_event *event,
+ unsigned long ref_ctr_offset, bool is_retprobe);
extern void perf_uprobe_destroy(struct perf_event *event);
extern int bpf_get_uprobe_info(const struct perf_event *event,
u32 *fd_type, const char **filename,
diff --git a/kernel/events/core.c b/kernel/events/core.c
index c80549bf82c6..65b30773af3e 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -8368,30 +8368,39 @@ static struct pmu perf_tracepoint = {
*
* PERF_PROBE_CONFIG_IS_RETPROBE if set, create kretprobe/uretprobe
* if not set, create kprobe/uprobe
+ *
+ * The following values specify a reference counter (or semaphore in the
+ * terminology of tools like dtrace, systemtap, etc.) Userspace Statically
+ * Defined Tracepoints (USDT). Currently, we use 40 bit for the offset.
+ *
+ * PERF_UPROBE_REF_CTR_OFFSET_BITS # of bits in config as th offset
+ * PERF_UPROBE_REF_CTR_OFFSET_SHIFT # of bits to shift left
*/
enum perf_probe_config {
PERF_PROBE_CONFIG_IS_RETPROBE = 1U << 0, /* [k,u]retprobe */
+ PERF_UPROBE_REF_CTR_OFFSET_BITS = 32,
+ PERF_UPROBE_REF_CTR_OFFSET_SHIFT = 64 - PERF_UPROBE_REF_CTR_OFFSET_BITS,
};

PMU_FORMAT_ATTR(retprobe, "config:0");
+#endif

-static struct attribute *probe_attrs[] = {
+#ifdef CONFIG_KPROBE_EVENTS
+static struct attribute *kprobe_attrs[] = {
&format_attr_retprobe.attr,
NULL,
};

-static struct attribute_group probe_format_group = {
+static struct attribute_group kprobe_format_group = {
.name = "format",
- .attrs = probe_attrs,
+ .attrs = kprobe_attrs,
};

-static const struct attribute_group *probe_attr_groups[] = {
- &probe_format_group,
+static const struct attribute_group *kprobe_attr_groups[] = {
+ &kprobe_format_group,
NULL,
};
-#endif

-#ifdef CONFIG_KPROBE_EVENTS
static int perf_kprobe_event_init(struct perf_event *event);
static struct pmu perf_kprobe = {
.task_ctx_nr = perf_sw_context,
@@ -8401,7 +8410,7 @@ static struct pmu perf_kprobe = {
.start = perf_swevent_start,
.stop = perf_swevent_stop,
.read = perf_swevent_read,
- .attr_groups = probe_attr_groups,
+ .attr_groups = kprobe_attr_groups,
};

static int perf_kprobe_event_init(struct perf_event *event)
@@ -8433,6 +8442,24 @@ static int perf_kprobe_event_init(struct perf_event *event)
#endif /* CONFIG_KPROBE_EVENTS */

#ifdef CONFIG_UPROBE_EVENTS
+PMU_FORMAT_ATTR(ref_ctr_offset, "config:32-63");
+
+static struct attribute *uprobe_attrs[] = {
+ &format_attr_retprobe.attr,
+ &format_attr_ref_ctr_offset.attr,
+ NULL,
+};
+
+static struct attribute_group uprobe_format_group = {
+ .name = "format",
+ .attrs = uprobe_attrs,
+};
+
+static const struct attribute_group *uprobe_attr_groups[] = {
+ &uprobe_format_group,
+ NULL,
+};
+
static int perf_uprobe_event_init(struct perf_event *event);
static struct pmu perf_uprobe = {
.task_ctx_nr = perf_sw_context,
@@ -8442,12 +8469,13 @@ static struct pmu perf_uprobe = {
.start = perf_swevent_start,
.stop = perf_swevent_stop,
.read = perf_swevent_read,
- .attr_groups = probe_attr_groups,
+ .attr_groups = uprobe_attr_groups,
};

static int perf_uprobe_event_init(struct perf_event *event)
{
int err;
+ unsigned long ref_ctr_offset;
bool is_retprobe;

if (event->attr.type != perf_uprobe.type)
@@ -8463,7 +8491,8 @@ static int perf_uprobe_event_init(struct perf_event *event)
return -EOPNOTSUPP;

is_retprobe = event->attr.config & PERF_PROBE_CONFIG_IS_RETPROBE;
- err = perf_uprobe_init(event, is_retprobe);
+ ref_ctr_offset = event->attr.config >> PERF_UPROBE_REF_CTR_OFFSET_SHIFT;
+ err = perf_uprobe_init(event, ref_ctr_offset, is_retprobe);
if (err)
return err;

diff --git a/kernel/trace/trace_event_perf.c b/kernel/trace/trace_event_perf.c
index 69a3fe926e8c..76217bbef815 100644
--- a/kernel/trace/trace_event_perf.c
+++ b/kernel/trace/trace_event_perf.c
@@ -290,7 +290,8 @@ void perf_kprobe_destroy(struct perf_event *p_event)
#endif /* CONFIG_KPROBE_EVENTS */

#ifdef CONFIG_UPROBE_EVENTS
-int perf_uprobe_init(struct perf_event *p_event, bool is_retprobe)
+int perf_uprobe_init(struct perf_event *p_event,
+ unsigned long ref_ctr_offset, bool is_retprobe)
{
int ret;
char *path = NULL;
@@ -312,8 +313,8 @@ int perf_uprobe_init(struct perf_event *p_event, bool is_retprobe)
goto out;
}

- tp_event = create_local_trace_uprobe(
- path, p_event->attr.probe_offset, is_retprobe);
+ tp_event = create_local_trace_uprobe(path, p_event->attr.probe_offset,
+ ref_ctr_offset, is_retprobe);
if (IS_ERR(tp_event)) {
ret = PTR_ERR(tp_event);
goto out;
diff --git a/kernel/trace/trace_probe.h b/kernel/trace/trace_probe.h
index 5f52668e165d..03b10f3201a5 100644
--- a/kernel/trace/trace_probe.h
+++ b/kernel/trace/trace_probe.h
@@ -412,6 +412,7 @@ create_local_trace_kprobe(char *func, void *addr, unsigned long offs,
extern void destroy_local_trace_kprobe(struct trace_event_call *event_call);

extern struct trace_event_call *
-create_local_trace_uprobe(char *name, unsigned long offs, bool is_return);
+create_local_trace_uprobe(char *name, unsigned long offs,
+ unsigned long ref_ctr_offset, bool is_return);
extern void destroy_local_trace_uprobe(struct trace_event_call *event_call);
#endif
diff --git a/kernel/trace/trace_uprobe.c b/kernel/trace/trace_uprobe.c
index 3a7c73c40007..d09638706fe0 100644
--- a/kernel/trace/trace_uprobe.c
+++ b/kernel/trace/trace_uprobe.c
@@ -1405,7 +1405,8 @@ static int unregister_uprobe_event(struct trace_uprobe *tu)

#ifdef CONFIG_PERF_EVENTS
struct trace_event_call *
-create_local_trace_uprobe(char *name, unsigned long offs, bool is_return)
+create_local_trace_uprobe(char *name, unsigned long offs,
+ unsigned long ref_ctr_offset, bool is_return)
{
struct trace_uprobe *tu;
struct path path;
@@ -1437,6 +1438,7 @@ create_local_trace_uprobe(char *name, unsigned long offs, bool is_return)

tu->offset = offs;
tu->path = path;
+ tu->ref_ctr_offset = ref_ctr_offset;
tu->filename = kstrdup(name, GFP_KERNEL);
init_trace_event_call(tu, &tu->tp.call);