From: "Michael S. Tsirkin" <mst@redhat.com>
To: "Paul E. McKenney" <paulmck@linux.ibm.com>
Cc: aarcange@redhat.com, akpm@linux-foundation.org,
christian@brauner.io, davem@davemloft.net, ebiederm@xmission.com,
elena.reshetova@intel.com, guro@fb.com, hch@infradead.org,
james.bottomley@hansenpartnership.com, jasowang@redhat.com,
jglisse@redhat.com, keescook@chromium.org, ldv@altlinux.org,
linux-arm-kernel@lists.infradead.org,
linux-kernel@vger.kernel.org, linux-mm@kvack.org,
linux-parisc@vger.kernel.org, luto@amacapital.net,
mhocko@suse.com, mingo@kernel.org, namit@vmware.com,
peterz@infradead.org, syzkaller-bugs@googlegroups.com,
viro@zeniv.linux.org.uk, wad@chromium.org
Subject: Re: RFC: call_rcu_outstanding (was Re: WARNING in __mmdrop)
Date: Mon, 22 Jul 2019 03:56:22 -0400 [thread overview]
Message-ID: <20190722035236-mutt-send-email-mst@kernel.org> (raw)
In-Reply-To: <20190721192841.GT14271@linux.ibm.com>
On Sun, Jul 21, 2019 at 12:28:41PM -0700, Paul E. McKenney wrote:
> On Sun, Jul 21, 2019 at 01:53:23PM -0400, Michael S. Tsirkin wrote:
> > On Sun, Jul 21, 2019 at 06:17:25AM -0700, Paul E. McKenney wrote:
> > > On Sun, Jul 21, 2019 at 08:28:05AM -0400, Michael S. Tsirkin wrote:
> > > > Hi Paul, others,
> > > >
> > > > So it seems that vhost needs to call kfree_rcu from an ioctl. My worry
> > > > is what happens if userspace starts cycling through lots of these
> > > > ioctls. Given we actually use rcu as an optimization, we could just
> > > > disable the optimization temporarily - but the question would be how to
> > > > detect an excessive rate without working too hard :) .
> > > >
> > > > I guess we could define as excessive any rate where callback is
> > > > outstanding at the time when new structure is allocated. I have very
> > > > little understanding of rcu internals - so I wanted to check that the
> > > > following more or less implements this heuristic before I spend time
> > > > actually testing it.
> > > >
> > > > Could others pls take a look and let me know?
> > >
> > > These look good as a way of seeing if there are any outstanding callbacks,
> > > but in the case of Tree RCU, call_rcu_outstanding() would almost never
> > > return false on a busy system.
> >
> > Hmm, ok. Maybe I could rename this to e.g. call_rcu_busy
> > and change the tree one to do rcu_segcblist_n_lazy_cbs > 1000?
>
> Or the function could simply return the number of callbacks queued
> on the current CPU, and let the caller decide how many is too many.
>
> > > Here are some alternatives:
> > >
> > > o RCU uses some pieces of Rao Shoaib kfree_rcu() patches.
> > > The idea is to make kfree_rcu() locally buffer requests into
> > > batches of (say) 1,000, but processing smaller batches when RCU
> > > is idle, or when some smallish amout of time has passed with
> > > no more kfree_rcu() request from that CPU. RCU than takes in
> > > the batch using not call_rcu(), but rather queue_rcu_work().
> > > The resulting batch of kfree() calls would therefore execute in
> > > workqueue context rather than in softirq context, which should
> > > be much easier on the system.
> > >
> > > In theory, this would allow people to use kfree_rcu() without
> > > worrying quite so much about overload. It would also not be
> > > that hard to implement.
> > >
> > > o Subsystems vulnerable to user-induced kfree_rcu() flooding use
> > > call_rcu() instead of kfree_rcu(). Keep a count of the number
> > > of things waiting for a grace period, and when this gets too
> > > large, disable the optimization. It will then drain down, at
> > > which point the optimization can be re-enabled.
> > >
> > > But please note that callbacks are -not- guaranteed to run on
> > > the CPU that queued them. So yes, you would need a per-CPU
> > > counter, but you would need to periodically sum it up to check
> > > against the global state. Or keep track of the CPU that
> > > did the call_rcu() so that you can atomically decrement in
> > > the callback the same counter that was atomically incremented
> > > just before the call_rcu(). Or any number of other approaches.
> >
> > I'm really looking for something we can do this merge window
> > and without adding too much code, and kfree_rcu is intended to
> > fix a bug.
> > Adding call_rcu and careful accounting is something that I'm not
> > happy adding with merge window already open.
>
> OK, then I suggest having the interface return you the number of
> callbacks. That allows you to experiment with the cutoff.
>
> Give or take the ioctl overhead...
OK - and for tiny just assume 1 is too much?
> > > Also, the overhead is important. For example, as far as I know,
> > > current RCU gracefully handles close(open(...)) in a tight userspace
> > > loop. But there might be trouble due to tight userspace loops around
> > > lighter-weight operations.
> > >
> > > So an important question is "Just how fast is your ioctl?" If it takes
> > > (say) 100 microseconds to execute, there should be absolutely no problem.
> > > On the other hand, if it can execute in 50 nanoseconds, this very likely
> > > does need serious attention.
> > >
> > > Other thoughts?
> > >
> > > Thanx, Paul
> >
> > Hmm the answer to this would be I'm not sure.
> > It's setup time stuff we never tested it.
>
> Is it possible to measure it easily?
>
> Thanx, Paul
>
> > > > Thanks!
> > > >
> > > > Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> > > >
> > > >
> > > > diff --git a/kernel/rcu/tiny.c b/kernel/rcu/tiny.c
> > > > index 477b4eb44af5..067909521d72 100644
> > > > --- a/kernel/rcu/tiny.c
> > > > +++ b/kernel/rcu/tiny.c
> > > > @@ -125,6 +125,25 @@ void synchronize_rcu(void)
> > > > }
> > > > EXPORT_SYMBOL_GPL(synchronize_rcu);
> > > >
> > > > +/*
> > > > + * Helpful for rate-limiting kfree_rcu/call_rcu callbacks.
> > > > + */
> > > > +bool call_rcu_outstanding(void)
> > > > +{
> > > > + unsigned long flags;
> > > > + struct rcu_data *rdp;
> > > > + bool outstanding;
> > > > +
> > > > + local_irq_save(flags);
> > > > + rdp = this_cpu_ptr(&rcu_data);
> > > > + outstanding = rcu_segcblist_empty(&rdp->cblist);
> > > > + outstanding = rcu_ctrlblk.donetail != rcu_ctrlblk.curtail;
> > > > + local_irq_restore(flags);
> > > > +
> > > > + return outstanding;
> > > > +}
> > > > +EXPORT_SYMBOL_GPL(call_rcu_outstanding);
> > > > +
> > > > /*
> > > > * Post an RCU callback to be invoked after the end of an RCU grace
> > > > * period. But since we have but one CPU, that would be after any
> > > > diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c
> > > > index a14e5fbbea46..d4b9d61e637d 100644
> > > > --- a/kernel/rcu/tree.c
> > > > +++ b/kernel/rcu/tree.c
> > > > @@ -2482,6 +2482,24 @@ static void rcu_leak_callback(struct rcu_head *rhp)
> > > > {
> > > > }
> > > >
> > > > +/*
> > > > + * Helpful for rate-limiting kfree_rcu/call_rcu callbacks.
> > > > + */
> > > > +bool call_rcu_outstanding(void)
> > > > +{
> > > > + unsigned long flags;
> > > > + struct rcu_data *rdp;
> > > > + bool outstanding;
> > > > +
> > > > + local_irq_save(flags);
> > > > + rdp = this_cpu_ptr(&rcu_data);
> > > > + outstanding = rcu_segcblist_empty(&rdp->cblist);
> > > > + local_irq_restore(flags);
> > > > +
> > > > + return outstanding;
> > > > +}
> > > > +EXPORT_SYMBOL_GPL(call_rcu_outstanding);
> > > > +
> > > > /*
> > > > * Helper function for call_rcu() and friends. The cpu argument will
> > > > * normally be -1, indicating "currently running CPU". It may specify
> >
next prev parent reply other threads:[~2019-07-22 7:56 UTC|newest]
Thread overview: 87+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <0000000000008dd6bb058e006938@google.com>
2019-07-20 10:08 ` WARNING in __mmdrop syzbot
2019-07-21 10:02 ` Michael S. Tsirkin
2019-07-21 12:18 ` Michael S. Tsirkin
2019-07-22 5:24 ` Jason Wang
2019-07-22 8:08 ` Michael S. Tsirkin
2019-07-23 4:01 ` Jason Wang
2019-07-23 5:01 ` Michael S. Tsirkin
2019-07-23 5:47 ` Jason Wang
2019-07-23 7:23 ` Michael S. Tsirkin
2019-07-23 7:53 ` Jason Wang
2019-07-23 8:10 ` Michael S. Tsirkin
2019-07-23 8:49 ` Jason Wang
2019-07-23 9:26 ` Michael S. Tsirkin
2019-07-23 13:31 ` Jason Wang
2019-07-25 5:52 ` Michael S. Tsirkin
2019-07-25 7:43 ` Jason Wang
2019-07-25 8:28 ` Michael S. Tsirkin
2019-07-25 13:21 ` Jason Wang
2019-07-25 13:26 ` Michael S. Tsirkin
2019-07-25 14:25 ` Jason Wang
2019-07-26 11:49 ` Michael S. Tsirkin
2019-07-26 12:00 ` Jason Wang
2019-07-26 12:38 ` Michael S. Tsirkin
2019-07-26 12:53 ` Jason Wang
2019-07-26 13:36 ` Jason Wang
2019-07-26 13:49 ` Michael S. Tsirkin
2019-07-29 5:54 ` Jason Wang
2019-07-29 8:59 ` Michael S. Tsirkin
2019-07-29 14:24 ` Jason Wang
2019-07-29 14:44 ` Michael S. Tsirkin
2019-07-30 7:44 ` Jason Wang
2019-07-30 8:03 ` Jason Wang
2019-07-30 15:08 ` Michael S. Tsirkin
2019-07-31 8:49 ` Jason Wang
2019-07-31 23:00 ` Jason Gunthorpe
2019-07-26 13:47 ` Michael S. Tsirkin
2019-07-26 14:00 ` Jason Wang
2019-07-26 14:10 ` Michael S. Tsirkin
2019-07-26 15:03 ` Jason Gunthorpe
2019-07-29 5:56 ` Jason Wang
2019-07-21 12:28 ` RFC: call_rcu_outstanding (was Re: WARNING in __mmdrop) Michael S. Tsirkin
2019-07-21 13:17 ` Paul E. McKenney
2019-07-21 17:53 ` Michael S. Tsirkin
2019-07-21 19:28 ` Paul E. McKenney
2019-07-22 7:56 ` Michael S. Tsirkin [this message]
2019-07-22 11:57 ` Paul E. McKenney
2019-07-21 21:08 ` Matthew Wilcox
2019-07-21 23:31 ` Paul E. McKenney
2019-07-22 7:52 ` Michael S. Tsirkin
2019-07-22 11:51 ` Paul E. McKenney
2019-07-22 13:41 ` Jason Gunthorpe
2019-07-22 15:52 ` Paul E. McKenney
2019-07-22 16:04 ` Jason Gunthorpe
2019-07-22 16:15 ` Michael S. Tsirkin
2019-07-22 16:15 ` Paul E. McKenney
2019-07-22 15:14 ` Joel Fernandes
2019-07-22 15:47 ` Michael S. Tsirkin
2019-07-22 15:55 ` Paul E. McKenney
2019-07-22 16:13 ` Michael S. Tsirkin
2019-07-22 16:25 ` Paul E. McKenney
2019-07-22 16:32 ` Michael S. Tsirkin
2019-07-22 18:58 ` Paul E. McKenney
2019-07-22 5:21 ` WARNING in __mmdrop Jason Wang
2019-07-22 8:02 ` Michael S. Tsirkin
2019-07-23 3:55 ` Jason Wang
2019-07-23 5:02 ` Michael S. Tsirkin
2019-07-23 5:48 ` Jason Wang
2019-07-23 7:25 ` Michael S. Tsirkin
2019-07-23 7:55 ` Jason Wang
2019-07-23 7:56 ` Michael S. Tsirkin
2019-07-23 8:42 ` Jason Wang
2019-07-23 10:27 ` Michael S. Tsirkin
2019-07-23 13:34 ` Jason Wang
2019-07-23 15:02 ` Michael S. Tsirkin
2019-07-24 2:17 ` Jason Wang
2019-07-24 8:05 ` Michael S. Tsirkin
2019-07-24 10:08 ` Jason Wang
2019-07-24 18:25 ` Michael S. Tsirkin
2019-07-25 3:44 ` Jason Wang
2019-07-25 5:09 ` Michael S. Tsirkin
2019-07-24 16:53 ` Jason Gunthorpe
2019-07-24 18:25 ` Michael S. Tsirkin
2019-07-23 10:42 ` Michael S. Tsirkin
2019-07-23 13:37 ` Jason Wang
2019-07-22 14:11 ` Jason Gunthorpe
2019-07-25 6:02 ` Michael S. Tsirkin
2019-07-25 7:44 ` Jason Wang
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=20190722035236-mutt-send-email-mst@kernel.org \
--to=mst@redhat.com \
--cc=aarcange@redhat.com \
--cc=akpm@linux-foundation.org \
--cc=christian@brauner.io \
--cc=davem@davemloft.net \
--cc=ebiederm@xmission.com \
--cc=elena.reshetova@intel.com \
--cc=guro@fb.com \
--cc=hch@infradead.org \
--cc=james.bottomley@hansenpartnership.com \
--cc=jasowang@redhat.com \
--cc=jglisse@redhat.com \
--cc=keescook@chromium.org \
--cc=ldv@altlinux.org \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linux-parisc@vger.kernel.org \
--cc=luto@amacapital.net \
--cc=mhocko@suse.com \
--cc=mingo@kernel.org \
--cc=namit@vmware.com \
--cc=paulmck@linux.ibm.com \
--cc=peterz@infradead.org \
--cc=syzkaller-bugs@googlegroups.com \
--cc=viro@zeniv.linux.org.uk \
--cc=wad@chromium.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;
as well as URLs for NNTP newsgroup(s).