[PATCH] document while_each_thread(), change first_tid() to use for_each_thread()

From: Oleg Nesterov
Date: Wed Aug 23 2023 - 13:09:48 EST


Add the comment to explain that while_each_thread(g,t) is not rcu-safe
unless g is stable (e.g. current). Even if g is a group leader and thus
can't exit before t, t or another sub-thread can exec and remove g from
the thread_group list.

The only lockless user of while_each_thread() is first_tid() and it is
fine in that it can't loop forever, yet for_each_thread() looks better
and I am going to change while_each_thread/next_thread.

Signed-off-by: Oleg Nesterov <oleg@xxxxxxxxxx>
---
fs/proc/base.c | 5 ++---
include/linux/sched/signal.h | 4 ++++
2 files changed, 6 insertions(+), 3 deletions(-)

diff --git a/fs/proc/base.c b/fs/proc/base.c
index 9df3f4839662..69dbb03ad55b 100644
--- a/fs/proc/base.c
+++ b/fs/proc/base.c
@@ -3813,11 +3813,10 @@ static struct task_struct *first_tid(struct pid *pid, int tid, loff_t f_pos,
/* If we haven't found our starting place yet start
* with the leader and walk nr threads forward.
*/
- pos = task = task->group_leader;
- do {
+ for_each_thread(task, pos) {
if (!nr--)
goto found;
- } while_each_thread(task, pos);
+ };
fail:
pos = NULL;
goto out;
diff --git a/include/linux/sched/signal.h b/include/linux/sched/signal.h
index 0deebe2ab07d..0014d3adaf84 100644
--- a/include/linux/sched/signal.h
+++ b/include/linux/sched/signal.h
@@ -648,6 +648,10 @@ extern void flush_itimer_signals(void);

extern bool current_is_single_threaded(void);

+/*
+ * Without tasklist/siglock it is only rcu-safe if g can't exit/exec,
+ * otherwise next_thread(t) will never reach g after list_del_rcu(g).
+ */
#define while_each_thread(g, t) \
while ((t = next_thread(t)) != g)

--
2.25.1.362.g51ebf55