public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Thomas Gleixner <tglx@linutronix.de>
To: Frederic Weisbecker <frederic@kernel.org>
Cc: LKML <linux-kernel@vger.kernel.org>,
	Anna-Maria Behnsen <anna-maria@linutronix.de>,
	John Stultz <jstultz@google.com>,
	Peter Zijlstra <peterz@infradead.org>,
	Ingo Molnar <mingo@kernel.org>, Stephen Boyd <sboyd@kernel.org>,
	Eric Biederman <ebiederm@xmission.com>,
	Oleg Nesterov <oleg@redhat.com>
Subject: Re: [patch v6 17/20] signal: Queue ignored posixtimers on ignore list
Date: Sat, 02 Nov 2024 15:49:50 +0100	[thread overview]
Message-ID: <87bjyxr8q9.ffs@tglx> (raw)
In-Reply-To: <871pzuvfzg.ffs@tglx>

On Fri, Nov 01 2024 at 21:47, Thomas Gleixner wrote:
> On Fri, Nov 01 2024 at 15:21, Frederic Weisbecker wrote:
>> Le Thu, Oct 31, 2024 at 04:46:43PM +0100, Thomas Gleixner a écrit :
>>> +static void sigqueue_free_ignored(struct task_struct *tsk, struct sigqueue *q)
>>> +{
>>> +	if (likely(!(q->flags & SIGQUEUE_PREALLOC) || q->info.si_code != SI_TIMER))
>>> +		__sigqueue_free(q);
>>> +	else
>>> +		posixtimer_sig_ignore(tsk, q);
>>
>> So this happens when the signal is ignored and delays it to when it will be
>> unignored. But the comment on do_sigaction() says:
>>
>> 		/*
>> 		 * POSIX 3.3.1.3:
>> 		 *  "Setting a signal action to SIG_IGN for a signal that is
>> 		 *   pending shall cause the pending signal to be discarded,
>> 		 *   whether or not it is blocked."
>> 		 *
>> 		 */
>>
>> Are posix timers an exception to that rule?
>>
>> Also I see flush_sigqueue_mask() called on other occasions, for example
>> when a STOP signal happens to remove pending CONT, not sure if posix
>> timers can set SIGCONT...
>
> No. The problem with posix timers is that they are magically different
> from regular signals in the case of periodic timers.
>
> When the signal is ignored at expiry, then the signal is not delivered
> and is 'dropped'. But when SIG_IGN is removed then the following period
> expiry has to deliver the signal.
>
> Right now the kernel ensures that by keeping the timer self rearming and
> rate limiting it for obvious reasons. That's a completely pointless
> exercise up to the point where SIG_IGN is removed.
>
> The only way to avoid the self rearming is to actually stop the timer
> when the signal is ignored and rearm it when SIG_IGN for the specific
> signal is removed.
>
> That's what this magic does...

But it does not exclude oneshot timers from this. Their signal has to be
dropped for real.

Delta patch below.

Thanks,

        tglx
---
--- a/kernel/signal.c
+++ b/kernel/signal.c
@@ -731,7 +731,7 @@ void signal_wake_up_state(struct task_st
 		kick_process(t);
 }
 
-static inline void posixtimer_sig_ignore(struct task_struct *tsk, struct sigqueue *q);
+static inline bool posixtimer_sig_ignore(struct task_struct *tsk, struct sigqueue *q);
 
 static void sigqueue_free_ignored(struct task_struct *tsk, struct sigqueue *q)
 {
@@ -1999,8 +1999,8 @@ int posixtimer_send_sigqueue(struct k_it
 			goto out;
 
 		if (hlist_unhashed(&tmr->ignored_list)) {
-			hlist_add_head(&tmr->ignored_list, &t->signal->ignored_posix_timers);
-			posixtimer_sigqueue_getref(q);
+			if (posixtimer_sig_ignore(t, tmr))
+				posixtimer_sigqueue_getref(q);
 		}
 		goto out;
 	}
@@ -2024,11 +2024,15 @@ int posixtimer_send_sigqueue(struct k_it
 	return 0;
 }
 
-static inline void posixtimer_sig_ignore(struct task_struct *tsk, struct sigqueue *q)
+static inline bool posixtimer_sig_ignore(struct task_struct *tsk, struct sigqueue *q)
 {
 	struct k_itimer *tmr = container_of(q, struct k_itimer, sigq);
 
+	/* Only queue periodic timer signals */
+	if (tmr->status != POSIX_TIMER_REQUEUE_PENDING)
+		return false;
 	hlist_add_head(&tmr->ignored_list, &tsk->signal->ignored_posix_timers);
+	return true;
 }
 
 static void posixtimer_sig_unignore(struct task_struct *tsk, int sig)

  reply	other threads:[~2024-11-02 14:49 UTC|newest]

Thread overview: 43+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-10-31 15:46 [patch v6 00/20] posix-timers: Cure the SIG_IGN mess Thomas Gleixner
2024-10-31 15:46 ` [patch v6 01/20] posix-timers: Make signal delivery consistent Thomas Gleixner
2024-11-01 12:26   ` Frederic Weisbecker
2024-10-31 15:46 ` [patch v6 02/20] posix-timers: Make signal overrun accounting sensible Thomas Gleixner
2024-11-01 12:51   ` Frederic Weisbecker
2024-11-01 20:36     ` Thomas Gleixner
2024-11-02 19:41       ` Thomas Gleixner
2024-11-02 22:57         ` Frederic Weisbecker
2024-10-31 15:46 ` [patch v6 03/20] posix-cpu-timers: Cleanup the firing logic Thomas Gleixner
2024-11-01 13:14   ` Frederic Weisbecker
2024-10-31 15:46 ` [patch v6 04/20] posix-cpu-timers: Use dedicated flag for CPU timer nanosleep Thomas Gleixner
2024-10-31 15:46 ` [patch v6 05/20] posix-timers: Add a refcount to struct k_itimer Thomas Gleixner
2024-10-31 15:46 ` [patch v6 06/20] signal: Split up __sigqueue_alloc() Thomas Gleixner
2024-10-31 15:46 ` [patch v6 07/20] signal: Provide posixtimer_sigqueue_init() Thomas Gleixner
2024-10-31 15:46 ` [patch v6 08/20] posix-timers: Store PID type in the timer Thomas Gleixner
2024-10-31 15:46 ` [patch v6 09/20] signal: Refactor send_sigqueue() Thomas Gleixner
2024-10-31 15:46 ` [patch v6 10/20] signal: Replace resched_timer logic Thomas Gleixner
2024-11-01 13:25   ` Frederic Weisbecker
2024-10-31 15:46 ` [patch v6 11/20] posix-timers: Embed sigqueue in struct k_itimer Thomas Gleixner
2024-10-31 15:46 ` [patch v6 12/20] signal: Cleanup unused posix-timer leftovers Thomas Gleixner
2024-10-31 15:46 ` [patch v6 13/20] posix-timers: Move sequence logic into struct k_itimer Thomas Gleixner
2024-10-31 15:46 ` [patch v6 14/20] signal: Provide ignored_posix_timers list Thomas Gleixner
2024-10-31 15:46 ` [patch v6 15/20] posix-timers: Handle ignored list on delete and exit Thomas Gleixner
2024-11-01 13:47   ` Frederic Weisbecker
2024-11-01 20:38     ` Thomas Gleixner
2024-10-31 15:46 ` [patch v6 16/20] signal: Handle ignored signals in do_sigaction(action != SIG_IGN) Thomas Gleixner
2024-11-01 14:04   ` Frederic Weisbecker
2024-10-31 15:46 ` [patch v6 17/20] signal: Queue ignored posixtimers on ignore list Thomas Gleixner
2024-11-01 14:21   ` Frederic Weisbecker
2024-11-01 20:47     ` Thomas Gleixner
2024-11-02 14:49       ` Thomas Gleixner [this message]
2024-11-02 20:57         ` Thomas Gleixner
2024-11-02 23:46       ` Frederic Weisbecker
2024-11-03  9:44         ` Thomas Gleixner
2024-11-03 19:55           ` Frederic Weisbecker
2024-11-02 21:05   ` [patch v6.1 " Thomas Gleixner
2024-11-04 11:42     ` Frederic Weisbecker
2024-11-04 15:21       ` Thomas Gleixner
2024-11-04 21:31         ` Thomas Gleixner
2024-11-04 23:02           ` Frederic Weisbecker
2024-10-31 15:46 ` [patch v6 18/20] posix-timers: Cleanup SIG_IGN workaround leftovers Thomas Gleixner
2024-10-31 15:46 ` [patch v6 19/20] alarmtimers: Remove the throttle mechanism from alarm_forward_now() Thomas Gleixner
2024-10-31 15:46 ` [patch v6 20/20] alarmtimers: Remove return value from alarm functions Thomas Gleixner

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=87bjyxr8q9.ffs@tglx \
    --to=tglx@linutronix.de \
    --cc=anna-maria@linutronix.de \
    --cc=ebiederm@xmission.com \
    --cc=frederic@kernel.org \
    --cc=jstultz@google.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@kernel.org \
    --cc=oleg@redhat.com \
    --cc=peterz@infradead.org \
    --cc=sboyd@kernel.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