From: "Paul E. McKenney" <paulmck@kernel.org>
To: Frederic Weisbecker <frederic@kernel.org>
Cc: rcu@vger.kernel.org, linux-kernel@vger.kernel.org,
kernel-team@fb.com, rostedt@goodmis.org,
Randy Dunlap <rdunlap@infradead.org>,
Thomas Gleixner <tglx@linutronix.de>,
John Ogness <john.ogness@linutronix.de>,
Petr Mladek <pmladek@suse.com>
Subject: Re: [PATCH RFC v2 rcu 2/8] srcu: Create an srcu_read_lock_nmisafe() and srcu_read_unlock_nmisafe()
Date: Mon, 3 Oct 2022 04:52:31 -0700 [thread overview]
Message-ID: <20221003115231.GX4196@paulmck-ThinkPad-P17-Gen-1> (raw)
In-Reply-To: <20221003095535.GA298829@lothringen>
On Mon, Oct 03, 2022 at 11:55:35AM +0200, Frederic Weisbecker wrote:
> On Sun, Oct 02, 2022 at 04:46:55PM -0700, Paul E. McKenney wrote:
> > On Sun, Oct 02, 2022 at 11:47:10PM +0200, Frederic Weisbecker wrote:
> > > On Sun, Oct 02, 2022 at 09:09:57AM -0700, Paul E. McKenney wrote:
> > > > On Sun, Oct 02, 2022 at 05:55:16PM +0200, Frederic Weisbecker wrote:
> > > > > On Thu, Sep 29, 2022 at 11:07:25AM -0700, Paul E. McKenney wrote:
> > > > > > @@ -1090,7 +1121,7 @@ static unsigned long srcu_gp_start_if_needed(struct srcu_struct *ssp,
> > > > > > int ss_state;
> > > > > >
> > > > > > check_init_srcu_struct(ssp);
> > > > > > - idx = srcu_read_lock(ssp);
> > > > > > + idx = __srcu_read_lock_nmisafe(ssp);
> > > > >
> > > > > Why do we need to force the atomic based version here (even if CONFIG_NEED_SRCU_NMI_SAFE=y)?
> > > >
> > > > In kernels built with CONFIG_NEED_SRCU_NMI_SAFE=n, we of course need it.
> > > > As you say, in kernels built with CONFIG_NEED_SRCU_NMI_SAFE=y, we don't.
> > > > But it doesn't hurt to always use __srcu_read_lock_nmisafe() here, and
> > > > this is nowhere near a fastpath, so there is little benefit to using
> > > > __srcu_read_lock() when it is safe to do so.
> > > >
> > > > In addition, note that it is possible that a given srcu_struct structure's
> > > > first grace period is executed before its first reader. In that
> > > > case, we have no way of knowing which of __srcu_read_lock_nmisafe()
> > > > or __srcu_read_lock() to choose.
> > > >
> > > > So this code always does it the slow(ish) safe way.
> > >
> > > But then srcu_read_lock_nmisafe() would work as well, right?
> >
> > Almost.
> >
> > The problem is that without the leading "__", this would convince SRCU
> > that this is an NMI-safe srcu_struct. Which it might not be. Worse yet,
> > if this srcu_struct had already done an srcu_read_lock(), it would splat.
>
> Ah ok, now that makes sense.
>
> >
> > > > > > ss_state = smp_load_acquire(&ssp->srcu_size_state);
> > > > > > if (ss_state < SRCU_SIZE_WAIT_CALL)
> > > > > > sdp = per_cpu_ptr(ssp->sda, 0);
> > > > > > @@ -1123,7 +1154,7 @@ static unsigned long srcu_gp_start_if_needed(struct srcu_struct *ssp,
> > > > > > srcu_funnel_gp_start(ssp, sdp, s, do_norm);
> > > > > > else if (needexp)
> > > > > > srcu_funnel_exp_start(ssp, sdp_mynode, s);
> > > > > > - srcu_read_unlock(ssp, idx);
> > > > > > + __srcu_read_unlock_nmisafe(ssp, idx);
> > > > > > return s;
> > > > > > }
> > > > > >
> > > > > > @@ -1427,13 +1458,13 @@ void srcu_barrier(struct srcu_struct *ssp)
> > > > > > /* Initial count prevents reaching zero until all CBs are posted. */
> > > > > > atomic_set(&ssp->srcu_barrier_cpu_cnt, 1);
> > > > > >
> > > > > > - idx = srcu_read_lock(ssp);
> > > > > > + idx = __srcu_read_lock_nmisafe(ssp);
> > > > >
> > > > > And same here?
> > > >
> > > > Yes, same here. ;-)
> > >
> > > Now bonus question: why do SRCU grace period starting/tracking
> > > need to be in an SRCU read side critical section? :o)
> >
> > Because I am lazy and like to keep things simple? ;-)
> >
> > More seriously, take a look at srcu_gp_start_if_needed() and the functions
> > it calls and ask yourself what bad things could happen if they were
> > preempted for an arbitrarily long period of time.
>
> I can see a risk for ssp->srcu_gp_seq to overflow. Can't say that was obvious
> though, at least for me. Am I missing something else?
That is what I recall. There might also be something else, of course. ;-)
Thanx, Paul
next prev parent reply other threads:[~2022-10-03 11:52 UTC|newest]
Thread overview: 60+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-09-21 14:46 [PATCH rcu 0/4] NMI-safe SRCU reader API Paul E. McKenney
2022-09-21 14:46 ` [PATCH RFC rcu 1/4] srcu: Convert ->srcu_lock_count and ->srcu_unlock_count to atomic Paul E. McKenney
2022-09-21 14:46 ` [PATCH RFC rcu 2/4] srcu: Create and srcu_read_lock_nmisafe() and srcu_read_unlock_nmisafe() Paul E. McKenney
2022-09-21 14:46 ` [PATCH RFC rcu 3/4] srcu: Check for consistent per-CPU per-srcu_struct NMI safety Paul E. McKenney
2022-09-21 14:46 ` [PATCH RFC rcu 4/4] srcu: Check for consistent global " Paul E. McKenney
2022-09-29 18:07 ` [PATCH v2 rcu 0/8] NMI-safe SRCU reader API Paul E. McKenney
2022-09-29 18:07 ` [PATCH RFC v2 rcu 1/8] srcu: Convert ->srcu_lock_count and ->srcu_unlock_count to atomic Paul E. McKenney
2022-09-30 15:02 ` John Ogness
2022-09-30 15:35 ` Paul E. McKenney
2022-09-30 20:37 ` John Ogness
2022-10-01 16:51 ` Paul E. McKenney
2022-09-29 18:07 ` [PATCH RFC v2 rcu 2/8] srcu: Create an srcu_read_lock_nmisafe() and srcu_read_unlock_nmisafe() Paul E. McKenney
2022-10-02 15:55 ` Frederic Weisbecker
2022-10-02 15:57 ` Frederic Weisbecker
2022-10-02 16:10 ` Paul E. McKenney
2022-10-02 16:09 ` Paul E. McKenney
2022-10-02 21:47 ` Frederic Weisbecker
2022-10-02 23:46 ` Paul E. McKenney
2022-10-03 9:55 ` Frederic Weisbecker
2022-10-03 11:52 ` Paul E. McKenney [this message]
2022-10-18 14:31 ` John Ogness
2022-10-18 15:18 ` Paul E. McKenney
2022-09-29 18:07 ` [PATCH RFC v2 rcu 3/8] srcu: Check for consistent per-CPU per-srcu_struct NMI safety Paul E. McKenney
2022-10-02 22:06 ` Frederic Weisbecker
2022-10-02 23:51 ` Paul E. McKenney
2022-10-03 10:13 ` Frederic Weisbecker
2022-10-03 11:57 ` Paul E. McKenney
2022-10-03 12:37 ` Frederic Weisbecker
2022-10-03 13:32 ` Paul E. McKenney
2022-10-03 13:36 ` Frederic Weisbecker
2022-09-29 18:07 ` [PATCH RFC v2 rcu 4/8] srcu: Check for consistent global " Paul E. McKenney
2022-09-29 18:07 ` [PATCH RFC v2 rcu 5/8] arch/x86: Add ARCH_HAS_NMI_SAFE_THIS_CPU_OPS Kconfig option Paul E. McKenney
2022-09-29 18:07 ` [PATCH RFC v2 rcu 6/8] arch/arm64: " Paul E. McKenney
2022-10-05 11:12 ` Mark Rutland
2022-09-29 18:07 ` [PATCH RFC v2 rcu 7/8] arch/loongarch: " Paul E. McKenney
2022-09-29 18:07 ` [PATCH RFC v2 rcu 8/8] arch/s390: " Paul E. McKenney
2022-10-03 14:11 ` [PATCH v2 rcu 0/8] NMI-safe SRCU reader API Frederic Weisbecker
2022-10-03 16:38 ` Paul E. McKenney
2022-10-14 22:47 ` Joel Fernandes
2022-10-14 22:52 ` Joel Fernandes
2022-10-18 10:33 ` John Ogness
2022-10-18 15:24 ` Paul E. McKenney
2022-10-18 18:44 ` John Ogness
2022-10-18 18:59 ` Paul E. McKenney
2022-10-18 21:57 ` Paul E. McKenney
2022-10-19 11:13 ` John Ogness
2022-10-19 19:14 ` Paul E. McKenney
2022-10-19 21:38 ` John Ogness
2022-10-19 22:05 ` Frederic Weisbecker
2022-10-20 22:27 ` Paul E. McKenney
2022-10-20 22:41 ` Paul E. McKenney
2022-10-21 12:27 ` John Ogness
2022-10-21 13:59 ` Paul E. McKenney
2022-10-21 18:41 ` Paul E. McKenney
2022-10-24 6:15 ` John Ogness
2022-10-24 13:47 ` Paul E. McKenney
2022-10-27 9:31 ` John Ogness
2022-10-27 14:10 ` Paul E. McKenney
2022-10-27 14:39 ` John Ogness
2022-10-27 16:01 ` Paul E. McKenney
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=20221003115231.GX4196@paulmck-ThinkPad-P17-Gen-1 \
--to=paulmck@kernel.org \
--cc=frederic@kernel.org \
--cc=john.ogness@linutronix.de \
--cc=kernel-team@fb.com \
--cc=linux-kernel@vger.kernel.org \
--cc=pmladek@suse.com \
--cc=rcu@vger.kernel.org \
--cc=rdunlap@infradead.org \
--cc=rostedt@goodmis.org \
--cc=tglx@linutronix.de \
/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;
as well as URLs for NNTP newsgroup(s).