linux-kernel - Re: [PATCH 2/5] sched,ptrace: Fix ptrace_check_attach() vs PREEMPT

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20220413185909.GB30360@redhat.com>
Date:   Wed, 13 Apr 2022 20:59:10 +0200
From:   Oleg Nesterov <oleg@...hat.com>
To:     Peter Zijlstra <peterz@...radead.org>
Cc:     rjw@...ysocki.net, mingo@...nel.org, vincent.guittot@...aro.org,
        dietmar.eggemann@....com, rostedt@...dmis.org, mgorman@...e.de,
        ebiederm@...ssion.com, bigeasy@...utronix.de,
        Will Deacon <will@...nel.org>, linux-kernel@...r.kernel.org,
        tj@...nel.org, linux-pm@...r.kernel.org
Subject: Re: [PATCH 2/5] sched,ptrace: Fix ptrace_check_attach() vs PREEMPT_RT

On 04/13, Oleg Nesterov wrote:
>
> On 04/13, Oleg Nesterov wrote:
> >
> > I like 1-2 but I need to read them (and other patches) again, a
> > couple of nits right now.
>
> Sorry, didn't have time to do this today, and now I am already sleeping.
>
> But... on a second thought, it seems there is a better solution. If nothing
> else it is simpler and doesn't duplicate the wait_task_inactive() logic.
>
> How about the patch below instead? On top of 1/5.
>
> Yes,yes, incomplete. in particular see the "!!!!!!!!!" comments. Just to
> explain the idea.

Cough. forget to attach the patch, sorry for noise.

Oleg.
---

diff --git a/include/linux/sched/jobctl.h b/include/linux/sched/jobctl.h
index ec8b312f7506..1b5a57048e13 100644
--- a/include/linux/sched/jobctl.h
+++ b/include/linux/sched/jobctl.h
@@ -22,7 +22,8 @@ struct task_struct;
 
 #define JOBCTL_STOPPED_BIT	24
 #define JOBCTL_TRACED_BIT	25
-#define JOBCTL_TRACED_FROZEN_BIT 26
+#define JOBCTL_TRACED_XXX_BIT	25
+#define JOBCTL_TRACED_FROZEN_BIT 27
 
 #define JOBCTL_STOP_DEQUEUED	(1UL << JOBCTL_STOP_DEQUEUED_BIT)
 #define JOBCTL_STOP_PENDING	(1UL << JOBCTL_STOP_PENDING_BIT)
@@ -35,6 +36,7 @@ struct task_struct;
 
 #define JOBCTL_STOPPED		(1UL << JOBCTL_STOPPED_BIT)
 #define JOBCTL_TRACED		(1UL << JOBCTL_TRACED_BIT)
+#define JOBCTL_TRACED_XXX	(1UL << JOBCTL_TRACED_XXX_BIT)
 #define JOBCTL_TRACED_FROZEN	(1UL << JOBCTL_TRACED_FROZEN_BIT)
 
 #define JOBCTL_TRAP_MASK	(JOBCTL_TRAP_STOP | JOBCTL_TRAP_NOTIFY)
diff --git a/kernel/ptrace.c b/kernel/ptrace.c
index 626f96d275c7..86b5226e6ba2 100644
--- a/kernel/ptrace.c
+++ b/kernel/ptrace.c
@@ -255,6 +255,19 @@ static int ptrace_check_attach(struct task_struct *child, bool ignore_state)
 {
 	int ret = -ESRCH;
 
+	if (!(child->ptrace && child->parent == current))
+		return ret;
+
+	if (ignore_state)
+		return 0;
+
+	if (wait_on_bit(&task->jobctl, JOBCTL_TRACED_XXX_BIT, TASK_KILLABLE))
+		return -EINTR;
+	// now that the tracee cleared JOBCTL_TRACED_XXX_BIT
+	// wait_task_inactive() should succeed or fail "really soon".
+	if (!wait_task_inactive(child, __TASK_TRACED))
+		return ret;
+
 	/*
 	 * We take the read lock around doing both checks to close a
 	 * possible race where someone else was tracing our child and
@@ -269,23 +282,11 @@ static int ptrace_check_attach(struct task_struct *child, bool ignore_state)
 		 * child->sighand can't be NULL, release_task()
 		 * does ptrace_unlink() before __exit_signal().
 		 */
-		if (ignore_state || ptrace_freeze_traced(child))
+		if (ptrace_freeze_traced(child))
 			ret = 0;
 	}
 	read_unlock(&tasklist_lock);
 
-	if (!ret && !ignore_state) {
-		if (!wait_task_inactive(child, __TASK_TRACED)) {
-			/*
-			 * This can only happen if may_ptrace_stop() fails and
-			 * ptrace_stop() changes ->state back to TASK_RUNNING,
-			 * so we should not worry about leaking __TASK_TRACED.
-			 */
-			WARN_ON(READ_ONCE(child->__state) == __TASK_TRACED);
-			ret = -ESRCH;
-		}
-	}
-
 	return ret;
 }
 
diff --git a/kernel/signal.c b/kernel/signal.c
index 0aea3f0a8002..5ca6235e5231 100644
--- a/kernel/signal.c
+++ b/kernel/signal.c
@@ -2220,7 +2220,7 @@ static int ptrace_stop(int exit_code, int why, int clear_code,
 	 * schedule() will not sleep if there is a pending signal that
 	 * can awaken the task.
 	 */
-	current->jobctl |= JOBCTL_TRACED;
+	current->jobctl |= (JOBCTL_TRACED | JOBCTL_TRACED_XXX);
 	set_special_state(TASK_TRACED);
 
 	/*
@@ -2291,6 +2291,10 @@ static int ptrace_stop(int exit_code, int why, int clear_code,
 		preempt_disable();
 		read_unlock(&tasklist_lock);
 		cgroup_enter_frozen();
+		// !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
+		// wrong, needs siglock
+		current->jobctl &= ~JOBCTL_TRACED_XXX;
+		wake_up_bit(&current->jobctl, ~JOBCTL_TRACED_XXX_BIT);
 		preempt_enable_no_resched();
 		freezable_schedule();
 		cgroup_leave_frozen(true);
@@ -2308,6 +2312,8 @@ static int ptrace_stop(int exit_code, int why, int clear_code,
 		if (gstop_done)
 			do_notify_parent_cldstop(current, false, why);
 
+		// !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
+		// need to clear ~JOBCTL_TRACED_XXX
 		/* tasklist protects us from ptrace_freeze_traced() */
 		__set_current_state(TASK_RUNNING);
 		read_code = false;