perf_counter: Use PID namespaces properly
Stop using task_struct::pid and start using PID namespaces.
PIDs will be reported in the PID namespace of the monitoring
task at the moment of counter creation.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Eric W. Biederman <ebiederm@xmission.com>
Cc: Oleg Nesterov <oleg@tv-sign.ru>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: John Kacur <jkacur@redhat.com>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
diff --git a/include/linux/perf_counter.h b/include/linux/perf_counter.h
index d970fbc..9ec20fc 100644
--- a/include/linux/perf_counter.h
+++ b/include/linux/perf_counter.h
@@ -317,6 +317,7 @@
#include <linux/spinlock.h>
#include <linux/hrtimer.h>
#include <linux/fs.h>
+#include <linux/pid_namespace.h>
#include <asm/atomic.h>
struct task_struct;
@@ -500,6 +501,8 @@
void (*destroy)(struct perf_counter *);
struct rcu_head rcu_head;
+
+ struct pid_namespace *ns;
#endif
};
diff --git a/kernel/perf_counter.c b/kernel/perf_counter.c
index fbed4d2..caa012c 100644
--- a/kernel/perf_counter.c
+++ b/kernel/perf_counter.c
@@ -1432,6 +1432,8 @@
struct perf_counter *counter;
counter = container_of(head, struct perf_counter, rcu_head);
+ if (counter->ns)
+ put_pid_ns(counter->ns);
kfree(counter);
}
@@ -2267,6 +2269,28 @@
rcu_read_unlock();
}
+static u32 perf_counter_pid(struct perf_counter *counter, struct task_struct *p)
+{
+ /*
+ * only top level counters have the pid namespace they were created in
+ */
+ if (counter->parent)
+ counter = counter->parent;
+
+ return task_tgid_nr_ns(p, counter->ns);
+}
+
+static u32 perf_counter_tid(struct perf_counter *counter, struct task_struct *p)
+{
+ /*
+ * only top level counters have the pid namespace they were created in
+ */
+ if (counter->parent)
+ counter = counter->parent;
+
+ return task_pid_nr_ns(p, counter->ns);
+}
+
static void perf_counter_output(struct perf_counter *counter,
int nmi, struct pt_regs *regs, u64 addr)
{
@@ -2303,8 +2327,8 @@
if (record_type & PERF_RECORD_TID) {
/* namespace issues */
- tid_entry.pid = current->group_leader->pid;
- tid_entry.tid = current->pid;
+ tid_entry.pid = perf_counter_pid(counter, current);
+ tid_entry.tid = perf_counter_tid(counter, current);
header.type |= PERF_RECORD_TID;
header.size += sizeof(tid_entry);
@@ -2432,6 +2456,9 @@
if (ret)
return;
+ comm_event->event.pid = perf_counter_pid(counter, comm_event->task);
+ comm_event->event.tid = perf_counter_tid(counter, comm_event->task);
+
perf_output_put(&handle, comm_event->event);
perf_output_copy(&handle, comm_event->comm,
comm_event->comm_size);
@@ -2504,8 +2531,6 @@
.task = task,
.event = {
.header = { .type = PERF_EVENT_COMM, },
- .pid = task->group_leader->pid,
- .tid = task->pid,
},
};
@@ -2542,6 +2567,9 @@
if (ret)
return;
+ mmap_event->event.pid = perf_counter_pid(counter, current);
+ mmap_event->event.tid = perf_counter_tid(counter, current);
+
perf_output_put(&handle, mmap_event->event);
perf_output_copy(&handle, mmap_event->file_name,
mmap_event->file_size);
@@ -2641,8 +2669,6 @@
.file = file,
.event = {
.header = { .type = PERF_EVENT_MMAP, },
- .pid = current->group_leader->pid,
- .tid = current->pid,
.start = addr,
.len = len,
.pgoff = pgoff,
@@ -2664,8 +2690,6 @@
.file = file,
.event = {
.header = { .type = PERF_EVENT_MUNMAP, },
- .pid = current->group_leader->pid,
- .tid = current->pid,
.start = addr,
.len = len,
.pgoff = pgoff,
@@ -3445,6 +3469,8 @@
list_add_tail(&counter->owner_entry, ¤t->perf_counter_list);
mutex_unlock(¤t->perf_counter_mutex);
+ counter->ns = get_pid_ns(current->nsproxy->pid_ns);
+
fput_light(counter_file, fput_needed2);
out_fput: