mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Oleg Nesterov <oleg@redhat.com>
To: Ingo Molnar <mingo@elte.hu>
Cc: arjan@infradead.org, jeremy@goop.org, mschmidt@redhat.com,
	mingo@redhat.com, hpa@zytor.com, linux-kernel@vger.kernel.org,
	tj@kernel.org, tglx@linutronix.de,
	Linus Torvalds <torvalds@linux-foundation.org>,
	Andrew Morton <akpm@linux-foundation.org>,
	linux-tip-commits@vger.kernel.org
Subject: Re: [PATCH] kthreads: Fix startup synchronization boot crash
Date: Tue, 1 Sep 2009 15:37:09 +0200	[thread overview]
Message-ID: <20090901133709.GA24041@redhat.com> (raw)
In-Reply-To: <20090901131440.GA29783@elte.hu>

On 09/01, Ingo Molnar wrote:
>
> * Oleg Nesterov <oleg@redhat.com> wrote:
>
> > Yes, this should work. But I _think_ we can make the better fix...
> >
> > I'll try to make the patch soon. Afaics we don't need
> > kthreadd_task_init_done.
>
> ok.

Just in case, the patch is ready. I need to re-check my thinking
and test it somehow...

- remove kthreadd_task initialization from rest_init()

- change kthreadd() to initialize kthreadd_task = current

- change the main loop in kthreadd() to take kthread_create_lock
  before the first schedule() (just shift schedule() down)

This way, if kthreadd_task needs the wakeup, kthread_create()
must see kthreadd_task != NULL after unlock(kthread_create_lock).

If kthread_create() sees kthreadd_task == NULL we can just sleep
on create.done, kthreadd() must notice the new request before
it calls schedule().

Note that with this change it is possible to use kthread_create()
at any time, but the caller will sleep until rest_init() creates
kthreadd.

Oleg.

 init/main.c      |    5 +----
 kernel/kthread.c |   30 ++++++++++++++++++------------
 2 files changed, 19 insertions(+), 16 deletions(-)


--- a/init/main.c
+++ b/init/main.c
@@ -449,12 +449,9 @@ static void __init setup_command_line(ch
 static noinline void __init_refok rest_init(void)
 	__releases(kernel_lock)
 {
-	int pid;
-
 	kernel_thread(kernel_init, NULL, CLONE_FS | CLONE_SIGHAND);
 	numa_default_policy();
-	pid = kernel_thread(kthreadd, NULL, CLONE_FS | CLONE_FILES);
-	kthreadd_task = find_task_by_pid_ns(pid, &init_pid_ns);
+	kernel_thread(kthreadd, NULL, CLONE_FS | CLONE_FILES);
 	unlock_kernel();
 
 	/*
--- a/kernel/kthread.c
+++ b/kernel/kthread.c
@@ -128,8 +128,14 @@ struct task_struct *kthread_create(int (
 	spin_lock(&kthread_create_lock);
 	list_add_tail(&create.list, &kthread_create_list);
 	spin_unlock(&kthread_create_lock);
-
-	wake_up_process(kthreadd_task);
+	/*
+	 * If kthreadd was not created yet, kthreadd() must see the result
+	 * of list_add_tail() later, it takes kthread_create_lock before the
+	 * first schedule(). If kthreadd() locked kthread_create_lock at
+	 * least once, we must see kthreadd_task != NULL.
+	 */
+	if (likely(kthreadd_task))
+		wake_up_process(kthreadd_task);
 	wait_for_completion(&create.done);
 
 	if (!IS_ERR(create.result)) {
@@ -216,23 +222,18 @@ EXPORT_SYMBOL(kthread_stop);
 
 int kthreadd(void *unused)
 {
-	struct task_struct *tsk = current;
+	kthreadd_task = current;
 
 	/* Setup a clean context for our children to inherit. */
-	set_task_comm(tsk, "kthreadd");
-	ignore_signals(tsk);
-	set_user_nice(tsk, KTHREAD_NICE_LEVEL);
-	set_cpus_allowed_ptr(tsk, cpu_all_mask);
+	set_task_comm(kthreadd_task, "kthreadd");
+	ignore_signals(kthreadd_task);
+	set_user_nice(kthreadd_task, KTHREAD_NICE_LEVEL);
+	set_cpus_allowed_ptr(kthreadd_task, cpu_all_mask);
 	set_mems_allowed(node_possible_map);
 
 	current->flags |= PF_NOFREEZE | PF_FREEZER_NOSIG;
 
 	for (;;) {
-		set_current_state(TASK_INTERRUPTIBLE);
-		if (list_empty(&kthread_create_list))
-			schedule();
-		__set_current_state(TASK_RUNNING);
-
 		spin_lock(&kthread_create_lock);
 		while (!list_empty(&kthread_create_list)) {
 			struct kthread_create_info *create;
@@ -247,6 +248,11 @@ int kthreadd(void *unused)
 			spin_lock(&kthread_create_lock);
 		}
 		spin_unlock(&kthread_create_lock);
+
+		set_current_state(TASK_INTERRUPTIBLE);
+		if (list_empty(&kthread_create_list))
+			schedule();
+		__set_current_state(TASK_RUNNING);
 	}
 
 	return 0;


  reply	other threads:[~2009-09-01 13:42 UTC|newest]

Thread overview: 30+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2009-08-29 16:27 [PATCH] x86: detect stack protector for i386 builds on x86_64 Michal Schmidt
2009-08-30 13:39 ` Américo Wang
2009-08-30 18:43 ` [tip:x86/asm] x86: Detect " tip-bot for Michal Schmidt
2009-09-01 10:03   ` Ingo Molnar
2009-09-01 11:39     ` [PATCH] kthreads: Fix startup synchronization boot crash Ingo Molnar
2009-09-01 13:04       ` Oleg Nesterov
2009-09-01 13:14         ` Ingo Molnar
2009-09-01 13:37           ` Oleg Nesterov [this message]
2009-09-01 13:59             ` Ingo Molnar
2009-09-01 14:55               ` Oleg Nesterov
2009-09-01 15:54                 ` Ingo Molnar
2009-09-01 16:00                   ` Oleg Nesterov
2009-09-02 13:06                     ` Ingo Molnar
2009-09-01 16:52                 ` [PATCH 0/1] kthreads: simplify !kthreadd_task logic, kill kthreadd_task_init_done Oleg Nesterov
2009-09-01 16:53                   ` [PATCH 1/1] " Oleg Nesterov
2009-09-01 23:22                   ` [PATCH 0/1] " Eric W. Biederman
2009-09-02  9:13                     ` Oleg Nesterov
2009-09-04  7:37                       ` Ingo Molnar
2009-09-18 16:32                         ` Wu Fei
2009-09-18 18:54                           ` Oleg Nesterov
2009-09-18 19:17                             ` Linus Torvalds
2009-09-18 21:12                               ` Oleg Nesterov
2009-09-18 21:15                                 ` Oleg Nesterov
2009-09-18 22:06                                 ` Linus Torvalds
2009-09-18 23:11                                   ` Eric W. Biederman
2009-09-18 23:22                                   ` Oleg Nesterov
2009-09-18 23:38                                     ` Linus Torvalds
2009-09-01 15:08             ` [PATCH] kthreads: Fix startup synchronization boot crash Américo Wang
2009-09-01 15:19               ` Oleg Nesterov
2009-08-31  6:21 ` [PATCH] x86: detect stack protector for i386 builds on x86_64 Tejun Heo

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20090901133709.GA24041@redhat.com \
    --to=oleg@redhat.com \
    --cc=akpm@linux-foundation.org \
    --cc=arjan@infradead.org \
    --cc=hpa@zytor.com \
    --cc=jeremy@goop.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-tip-commits@vger.kernel.org \
    --cc=mingo@elte.hu \
    --cc=mingo@redhat.com \
    --cc=mschmidt@redhat.com \
    --cc=tglx@linutronix.de \
    --cc=tj@kernel.org \
    --cc=torvalds@linux-foundation.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

Powered by JetHome