linux-kernel - Re: [PATCH v6 15/21] timer: Add get next timer interrupt functionality for remote CPUs

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: <c92c4491-d183-4b15-5c94-504c9c8c681@linutronix.de>
Date:   Thu, 11 May 2023 15:06:43 +0200 (CEST)
From:   Anna-Maria Behnsen <anna-maria@...utronix.de>
To:     Frederic Weisbecker <frederic@...nel.org>
cc:     linux-kernel@...r.kernel.org,
        Peter Zijlstra <peterz@...radead.org>,
        John Stultz <jstultz@...gle.com>,
        Thomas Gleixner <tglx@...utronix.de>,
        Eric Dumazet <edumazet@...gle.com>,
        "Rafael J . Wysocki" <rafael.j.wysocki@...el.com>,
        Arjan van de Ven <arjan@...radead.org>,
        "Paul E . McKenney" <paulmck@...nel.org>,
        Frederic Weisbecker <fweisbec@...il.com>,
        Rik van Riel <riel@...riel.com>,
        Steven Rostedt <rostedt@...dmis.org>,
        Sebastian Siewior <bigeasy@...utronix.de>,
        Giovanni Gherdovich <ggherdovich@...e.cz>,
        Lukasz Luba <lukasz.luba@....com>,
        "Gautham R . Shenoy" <gautham.shenoy@....com>
Subject: Re: [PATCH v6 15/21] timer: Add get next timer interrupt functionality
 for remote CPUs

On Wed, 10 May 2023, Frederic Weisbecker wrote:

> Le Wed, May 10, 2023 at 09:28:11AM +0200, Anna-Maria Behnsen a écrit :
> > +/**
> > + * fetch_next_timer_interrupt_remote
> > + * @basej:	base time jiffies
> > + * @basem:	base time clock monotonic
> > + * @tevt:	Pointer to the storage for the expiry values
> > + * @cpu:	Remote CPU
> > + *
> > + * Stores the next pending local and global timer expiry values in the
> > + * struct pointed to by @tevt. If a queue is empty the corresponding
> > + * field is set to KTIME_MAX. If local event expires before global
> > + * event, global event is set to KTIME_MAX as well.
> > + *
> > + * Caller needs to make sure timer base locks are held (use
> > + * timer_lock_remote_bases() for this purpose). Caller must make sure
> > + * interrupts are reopened, if required.
> > + */
> > +void fetch_next_timer_interrupt_remote(unsigned long basej, u64 basem,
> > +				       struct timer_events *tevt,
> > +				       unsigned int cpu)
> > +{
> > +	struct timer_base *base_local, *base_global;
> > +
> > +	/* Preset local / global events */
> > +	tevt->local = tevt->global = KTIME_MAX;
> > +
> > +	base_local = per_cpu_ptr(&timer_bases[BASE_LOCAL], cpu);
> > +	base_global = per_cpu_ptr(&timer_bases[BASE_GLOBAL], cpu);
> > +
> > +	lockdep_assert_held(&base_local->lock);
> > +	lockdep_assert_held(&base_global->lock);
> > +
> > +	fetch_next_timer_interrupt(base_local, base_global, basej, basem, tevt);
> > +
> > +	raw_spin_unlock(&base_global->lock);
> > +	raw_spin_unlock(&base_local->lock);
> 
> Oh so that makes:
> 
> LOCK(baseL)
> LOCK(baseG)
> LOCK(tmc)
> UNLOCK(baseG)
> UNLOCK(baseL)
> UNLOCK(tmc)
> 
> I guess you can keep the bases locks locked until the end of
> tmigr_handle_remote_cpu(). After all that's what get_next_timer_interrupt()
> does. I'm not sure the above early release of bases locks will bring much
> in case of contention...
> 
> Then a timer_unlock_remote_bases() would restore symmetry.

When the walk of the hierarchy for updating the new timer is done,
additional locks has to be taken. So then 5 locks are held during the
update in lvl1 and higher: baseL, baseG, tmc, child, group

I don't see a problem releasing the baseL and baseG lock earlier. But I
will add an extra function for releasing the base locks to make it more
clear. The whole section is protected by a local_irq_disable() and
irq_enable is done when unlocking the tmc lock.

No?

> > +/**
> > + * timer_lock_remote_bases - lock timer bases of cpu
> > + * @cpu:	Remote CPU
> > + *
> > + * Locks the remote timer bases.
> > + *
> > + * Returns false if cpu is offline, otherwise true is returned.
> > + */
> > +bool timer_lock_remote_bases(unsigned int cpu)
> > +{
> > +	struct timer_base *base_local, *base_global;
> > +
> > +	/*
> > +	 * Pretend that there is no timer pending if the cpu is offline.
> > +	 * Possible pending timers will be migrated later to an active cpu.
> > +	 */
> > +	if (cpu_is_offline(cpu))
> > +		return false;
> 
> This value is never checked and the caller assumes the bases are
> always locked upon calling this (more on this on the big patch).
> 
> Thanks.
> 
> > +
> > +	base_local = per_cpu_ptr(&timer_bases[BASE_LOCAL], cpu);
> > +	base_global = per_cpu_ptr(&timer_bases[BASE_GLOBAL], cpu);
> > +
> > +	raw_spin_lock_irq(&base_local->lock);
> > +	raw_spin_lock_nested(&base_global->lock, SINGLE_DEPTH_NESTING);
> > +
> > +	return true;
> > +}
> 
>