« Return to Thread: [PATCH] rework/fix is_single_threaded()

[PATCH] rework/fix is_single_threaded()

by Oleg Nesterov-3 :: Rate this Message:

Reply to Author | View in Thread

- Fix the comment, is_single_threaded(p) actually means that nobody shares
  ->mm with p.

  I think this helper should be renamed, and it should not have arguments.
  With or without this patch it must not be used unless p == current,
  otherwise we can't safely use p->signal or p->mm.

- "if (atomic_read(&p->signal->count) != 1)" is not right when we have a
  zombie group leader, use signal->live instead.

- Add PF_KTHREAD check to skip kernel threads which may borrow p->mm,
  otherwise we can return the wrong "false".

- Use for_each_process() instead of do_each_thread(), all threads must use
  the same ->mm.

- Use down_write(mm->mmap_sem) + rcu_read_lock() instead of tasklist_lock
  to iterate over the process list. If there is another CLONE_VM process
  it can't pass exit_mm() which takes the same mm->mmap_sem. We can miss
  a freshly forked CLONE_VM task, but this doesn't matter because we must
  see its parent and return false.

Signed-off-by: Oleg Nesterov <oleg@...>

--- 6.30/lib/is_single_threaded.c~ISS 2009-04-06 00:03:42.000000000 +0200
+++ 6.30/lib/is_single_threaded.c 2009-04-13 23:39:56.000000000 +0200
@@ -12,34 +12,44 @@
 
 #include <linux/sched.h>
 
-/**
- * is_single_threaded - Determine if a thread group is single-threaded or not
- * @p: A task in the thread group in question
- *
- * This returns true if the thread group to which a task belongs is single
- * threaded, false if it is not.
+/*
+ * Returns true if the task does not share ->mm with another thread/process.
  */
-bool is_single_threaded(struct task_struct *p)
+bool is_single_threaded(struct task_struct *task)
 {
- struct task_struct *g, *t;
- struct mm_struct *mm = p->mm;
+ struct mm_struct *mm = task->mm;
+ struct task_struct *p, *t;
+ bool ret;
 
- if (atomic_read(&p->signal->count) != 1)
- goto no;
+ might_sleep();
 
- if (atomic_read(&p->mm->mm_users) != 1) {
- read_lock(&tasklist_lock);
- do_each_thread(g, t) {
- if (t->mm == mm && t != p)
- goto no_unlock;
- } while_each_thread(g, t);
- read_unlock(&tasklist_lock);
- }
+ if (atomic_read(&task->signal->live) != 1)
+ return false;
 
- return true;
+ if (atomic_read(&mm->mm_users) == 1)
+ return true;
 
-no_unlock:
- read_unlock(&tasklist_lock);
-no:
- return false;
+ ret = false;
+ down_write(&mm->mmap_sem);
+ rcu_read_lock();
+ for_each_process(p) {
+ if (unlikely(p->flags & PF_KTHREAD))
+ continue;
+ if (unlikely(p == task->group_leader))
+ continue;
+
+ t = p;
+ do {
+ if (unlikely(t->mm == mm))
+ goto found;
+ if (likely(t->mm))
+ break;
+ } while_each_thread(p, t);
+ }
+ ret = true;
+found:
+ rcu_read_unlock();
+ up_write(&mm->mmap_sem);
+
+ return ret;
 }

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

 « Return to Thread: [PATCH] rework/fix is_single_threaded()