linux-kernel - Re: Questions about wake_up[_interruptible][

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: <13851363-0dc9-465c-9ced-3ede4904eef0@samba.org>
Date: Thu, 14 Aug 2025 09:57:14 +0200
From: Stefan Metzmacher <metze@...ba.org>
To: Steven Rostedt <rostedt@...dmis.org>
Cc: Ingo Molnar <mingo@...hat.com>, Peter Zijlstra <peterz@...radead.org>,
 Juri Lelli <juri.lelli@...hat.com>,
 Dietmar Eggemann <dietmar.eggemann@....com>, Ben Segall
 <bsegall@...gle.com>, Mel Gorman <mgorman@...e.de>,
 Valentin Schneider <vschneid@...hat.com>, Steve French <smfrench@...il.com>,
 Namjae Jeon <linkinjeon@...nel.org>,
 "linux-cifs@...r.kernel.org" <linux-cifs@...r.kernel.org>,
 "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: Questions about wake_up[_interruptible][_all]

Am 13.08.25 um 23:37 schrieb Steven Rostedt:
> On Wed, 13 Aug 2025 22:28:08 +0200
> Stefan Metzmacher <metze@...ba.org> wrote:
> 
>> I guess I understand the difference between
>> wait_event() and wait_event_interruptible(),
>> the first ignores any signal even kill and the
>> 2nd returns -ERESTARTSYS on any signal.
> 
> The main difference is what the code does after the wait_event*().
> 
> If you use wait_event_interruptible() the first thing the code should do is
> to check if a signal is pending or not. Or at least check some status to
> know that what it is waiting for did not happen and handle it properly.
> 
> But there's places in the kernel where the task is waiting for something
> and it expects that whatever it is waiting for *must* happen eventually and
> it should not continue until it does.
> 
> Looking at one example: fs/jbd2/journal.c: jbd2_journal_start_thread()
> 
> It creates a thread, tests that it is created, and then waits for that
> thread to acknowledge that it is running, and the function should not
> return until it does.
> 
> If someone were to send a signal to that waiter and wake it up prematurely,
> the following code may become buggy as it expects the thread to be
> initialized and active when it is not.

Thanks!

Via a private channel I also got this answer:

wake_up_interruptible() only wakes tasks that are in the
TASK_INTERRUPTIBLE state.

wake_up() wakes tasks that are in either the TASK_INTERRUPTIBLE or
TASK_UNINTERRUPTIBLE state, as per the TASK_NORMAL macro used in the
definition of wake_up().

Call chain:

wake_up_interruptible
   __wake_up(mode = TASK_INTERRUPTIBLE)
     __wake_up_common_lock(mode = TASK_INTERRUPTIBLE)
       __wake_up_common(mode = TASK_INTERRUPTIBLE)
         curr->func(mode = TASK_INTERRUPTIBLE)
           // curr->func is usually default_wake_function
           default_wake_function(mode = TASK_INTERRUPTIBLE)
             try_to_wake_up(state = TASK_INTERRUPTIBLE)
               ttwu_state_match(state = TASK_INTERRUPTIBLE)
                 __task_state_match(state = TASK_INTERRUPTIBLE):
                 if (READ_ONCE(p->__state) & state) ...

metze