Re: [PATCH 7/7] proc: Ensure we see the exit of each process tid exactly once
flat view
From: Oleg Nesterov <oleg@redhat.com>
Date: 2020-02-21 16:50:53
Also in:
linux-fsdevel, linux-security-module, lkml
Possibly related (same subject, not in this thread)
- 2020-02-22 · Re: [PATCH 7/7] proc: Ensure we see the exit of each process tid exactly once · Eric W. Biederman <hidden>
On 02/20, Eric W. Biederman wrote:
quoted hunk ↗ jump to hunk
+void exchange_tids(struct task_struct *ntask, struct task_struct *otask) +{ + /* pid_links[PIDTYPE_PID].next is always NULL */ + struct pid *npid = READ_ONCE(ntask->thread_pid); + struct pid *opid = READ_ONCE(otask->thread_pid); + + rcu_assign_pointer(opid->tasks[PIDTYPE_PID].first, &ntask->pid_links[PIDTYPE_PID]); + rcu_assign_pointer(npid->tasks[PIDTYPE_PID].first, &otask->pid_links[PIDTYPE_PID]); + rcu_assign_pointer(ntask->thread_pid, opid); + rcu_assign_pointer(otask->thread_pid, npid);
this breaks has_group_leader_pid()... proc_pid_readdir() can miss a process doing mt-exec but this looks fixable, just we need to update ntask->thread_pid before updating ->first. The more problematic case is __exit_signal() which does if (unlikely(has_group_leader_pid(tsk))) posix_cpu_timers_exit_group(tsk); Oleg.