From: "Paul E. McKenney" <paulmck@kernel.org>
To: Joel Fernandes <joel@joelfernandes.org>
Cc: "Uladzislau Rezki (Sony)" <urezki@gmail.com>,
LKML <linux-kernel@vger.kernel.org>,
linux-mm@kvack.org, Andrew Morton <akpm@linux-foundation.org>,
"Theodore Y . Ts'o" <tytso@mit.edu>,
Matthew Wilcox <willy@infradead.org>, RCU <rcu@vger.kernel.org>,
Oleksiy Avramchenko <oleksiy.avramchenko@sonymobile.com>
Subject: Re: [PATCH 19/24] rcu/tree: Support reclaim for head-less object
Date: Sun, 3 May 2020 17:28:55 -0700 [thread overview]
Message-ID: <20200504002855.GF2869@paulmck-ThinkPad-P72> (raw)
In-Reply-To: <20200504001258.GD197097@google.com>
On Sun, May 03, 2020 at 08:12:58PM -0400, Joel Fernandes wrote:
> On Fri, May 01, 2020 at 03:39:09PM -0700, Paul E. McKenney wrote:
> > On Tue, Apr 28, 2020 at 10:58:58PM +0200, Uladzislau Rezki (Sony) wrote:
> > > Update the kvfree_call_rcu() with head-less support, it
> > > means an object without any rcu_head structure can be
> > > reclaimed after GP.
> > >
> > > To store pointers there are two chain-arrays maintained
> > > one for SLAB and another one is for vmalloc. Both types
> > > of objects(head-less variant and regular one) are placed
> > > there based on the type.
> > >
> > > It can be that maintaining of arrays becomes impossible
> > > due to high memory pressure. For such reason there is an
> > > emergency path. In that case objects with rcu_head inside
> > > are just queued building one way list. Later on that list
> > > is drained.
> > >
> > > As for head-less variant. Such objects do not have any
> > > rcu_head helper inside. Thus it is dynamically attached.
> > > As a result an object consists of back-pointer and regular
> > > rcu_head. It implies that emergency path can detect such
> > > object type, therefore they are tagged. So a back-pointer
> > > could be freed as well as dynamically attached wrapper.
> > >
> > > Even though such approach requires dynamic memory it needs
> > > only sizeof(unsigned long *) + sizeof(struct rcu_head) bytes,
> > > thus SLAB is used to obtain it. Finally if attaching of the
> > > rcu_head and queuing get failed, the current context has
> > > to follow might_sleep() annotation, thus below steps could
> > > be applied:
> > > a) wait until a grace period has elapsed;
> > > b) direct inlining of the kvfree() call.
> > >
> > > Reviewed-by: Joel Fernandes (Google) <joel@joelfernandes.org>
> > > Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
> > > Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
> > > Co-developed-by: Joel Fernandes (Google) <joel@joelfernandes.org>
> > > ---
> > > kernel/rcu/tree.c | 102 ++++++++++++++++++++++++++++++++++++++++++++--
> > > 1 file changed, 98 insertions(+), 4 deletions(-)
> > >
> > > diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c
> > > index 51726e4c3b4d..501cac02146d 100644
> > > --- a/kernel/rcu/tree.c
> > > +++ b/kernel/rcu/tree.c
> > > @@ -3072,15 +3072,31 @@ static void kfree_rcu_work(struct work_struct *work)
> > > */
> > > for (; head; head = next) {
> > > unsigned long offset = (unsigned long)head->func;
> > > - void *ptr = (void *)head - offset;
> > > + bool headless;
> > > + void *ptr;
> > >
> > > next = head->next;
> > > +
> > > + /* We tag the headless object, if so adjust offset. */
> > > + headless = (((unsigned long) head - offset) & BIT(0));
> > > + if (headless)
> > > + offset -= 1;
> > > +
> > > + ptr = (void *) head - offset;
> > > +
> > > debug_rcu_head_unqueue((struct rcu_head *)ptr);
> > > rcu_lock_acquire(&rcu_callback_map);
> > > trace_rcu_invoke_kvfree_callback(rcu_state.name, head, offset);
> > >
> > > - if (!WARN_ON_ONCE(!__is_kvfree_rcu_offset(offset)))
> > > + if (!WARN_ON_ONCE(!__is_kvfree_rcu_offset(offset))) {
> > > + /*
> > > + * If headless free the back-pointer first.
> > > + */
> > > + if (headless)
> > > + kvfree((void *) *((unsigned long *) ptr));
> > > +
> > > kvfree(ptr);
> > > + }
> > >
> > > rcu_lock_release(&rcu_callback_map);
> > > cond_resched_tasks_rcu_qs();
> > > @@ -3221,6 +3237,13 @@ kvfree_call_rcu_add_ptr_to_bulk(struct kfree_rcu_cpu *krcp, void *ptr)
> > > if (IS_ENABLED(CONFIG_PREEMPT_RT))
> > > return false;
> > >
> > > + /*
> > > + * TODO: For one argument of kvfree_rcu() we can
> > > + * drop the lock and get the page in sleepable
> > > + * context. That would allow to maintain an array
> > > + * for the CONFIG_PREEMPT_RT as well. Thus we could
> > > + * get rid of dynamic rcu_head attaching code.
> > > + */
> > > bnode = (struct kvfree_rcu_bulk_data *)
> > > __get_free_page(GFP_NOWAIT | __GFP_NOWARN);
> > > }
> > > @@ -3244,6 +3267,23 @@ kvfree_call_rcu_add_ptr_to_bulk(struct kfree_rcu_cpu *krcp, void *ptr)
> > > return true;
> > > }
> > >
> > > +static inline struct rcu_head *
> > > +attach_rcu_head_to_object(void *obj)
> > > +{
> > > + unsigned long *ptr;
> > > +
> > > + ptr = kmalloc(sizeof(unsigned long *) +
> > > + sizeof(struct rcu_head), GFP_NOWAIT |
> > > + __GFP_RECLAIM | /* can do direct reclaim. */
> > > + __GFP_NORETRY | /* only lightweight one. */
> > > + __GFP_NOWARN); /* no failure reports. */
> >
> > Again, let's please not do this single-pointer-sized allocation. If
> > a full page is not available and this is a single-argument kfree_rcu(),
> > just call synchronize_rcu() and then free the object directly.
>
> With the additional caching, lack of full page should not be very likely. I
> agree we can avoid doing any allocation and just straight to
> synchroize_rcu().
That sounds good to me!
> > It should not be -that- hard to adjust locking for CONFIG_PREEMPT_RT!
> > For example, have some kind of reservation protocol so that a task
> > that drops the lock can retry the page allocation and be sure of having
> > a place to put it. This might entail making CONFIG_PREEMPT_RT reserve
> > more pages per CPU. Or maybe that would not be necessary.
>
> If we are not doing single-pointer allocation, then that would also eliminate
> entering the low-level page allocator for single-pointer allocations.
>
> Or did you mean entry into the allocator for the full-page allocations
> related to the pointer array for PREEMPT_RT? Even if we skip entry into the
> allocator for those, we will still have additional caching which further
> reduces chances of getting a full page. In the event of such failure, we can
> simply queue the rcu_head.
>
> Thoughts?
I was just trying to guess why you kept the single-pointer allocation.
It looks like I guessed wrong. ;-)
If, as you say above, you make it go straight to synchronize_rcu()
upon full-page allocation failure, that would be good!
Thanx, Paul
> thanks,
>
> - Joel
>
> >
> > Thanx, Paul
> >
> > > + if (!ptr)
> > > + return NULL;
> > > +
> > > + ptr[0] = (unsigned long) obj;
> > > + return ((struct rcu_head *) ++ptr);
> > > +}
> > > +
> > > /*
> > > * Queue a request for lazy invocation of appropriate free routine after a
> > > * grace period. Please note there are three paths are maintained, two are the
> > > @@ -3260,16 +3300,34 @@ void kvfree_call_rcu(struct rcu_head *head, rcu_callback_t func)
> > > {
> > > unsigned long flags;
> > > struct kfree_rcu_cpu *krcp;
> > > + bool success;
> > > void *ptr;
> > >
> > > + if (head) {
> > > + ptr = (void *) head - (unsigned long) func;
> > > + } else {
> > > + /*
> > > + * Please note there is a limitation for the head-less
> > > + * variant, that is why there is a clear rule for such
> > > + * objects:
> > > + *
> > > + * it can be used from might_sleep() context only. For
> > > + * other places please embed an rcu_head to your data.
> > > + */
> > > + might_sleep();
> > > + ptr = (unsigned long *) func;
> > > + }
> > > +
> > > krcp = krc_this_cpu_lock(&flags);
> > > - ptr = (void *)head - (unsigned long)func;
> > >
> > > /* Queue the object but don't yet schedule the batch. */
> > > if (debug_rcu_head_queue(ptr)) {
> > > /* Probable double kfree_rcu(), just leak. */
> > > WARN_ONCE(1, "%s(): Double-freed call. rcu_head %p\n",
> > > __func__, head);
> > > +
> > > + /* Mark as success and leave. */
> > > + success = true;
> > > goto unlock_return;
> > > }
> > >
> > > @@ -3277,10 +3335,34 @@ void kvfree_call_rcu(struct rcu_head *head, rcu_callback_t func)
> > > * Under high memory pressure GFP_NOWAIT can fail,
> > > * in that case the emergency path is maintained.
> > > */
> > > - if (unlikely(!kvfree_call_rcu_add_ptr_to_bulk(krcp, ptr))) {
> > > + success = kvfree_call_rcu_add_ptr_to_bulk(krcp, ptr);
> > > + if (!success) {
> > > + if (head == NULL) {
> > > + /*
> > > + * Headless(one argument kvfree_rcu()) can sleep.
> > > + * Drop the lock and tack it back. So it can do
> > > + * direct lightweight reclaim.
> > > + */
> > > + krc_this_cpu_unlock(krcp, flags);
> > > + head = attach_rcu_head_to_object(ptr);
> > > + krcp = krc_this_cpu_lock(&flags);
> > > +
> > > + if (head == NULL)
> > > + goto unlock_return;
> > > +
> > > + /*
> > > + * Tag the headless object. Such objects have a
> > > + * back-pointer to the original allocated memory,
> > > + * that has to be freed as well as dynamically
> > > + * attached wrapper/head.
> > > + */
> > > + func = (rcu_callback_t) (sizeof(unsigned long *) + 1);
> > > + }
> > > +
> > > head->func = func;
> > > head->next = krcp->head;
> > > krcp->head = head;
> > > + success = true;
> > > }
> > >
> > > WRITE_ONCE(krcp->count, krcp->count + 1);
> > > @@ -3294,6 +3376,18 @@ void kvfree_call_rcu(struct rcu_head *head, rcu_callback_t func)
> > >
> > > unlock_return:
> > > krc_this_cpu_unlock(krcp, flags);
> > > +
> > > + /*
> > > + * High memory pressure, so inline kvfree() after
> > > + * synchronize_rcu(). We can do it from might_sleep()
> > > + * context only, so the current CPU can pass the QS
> > > + * state.
> > > + */
> > > + if (!success) {
> > > + debug_rcu_head_unqueue(ptr);
> > > + synchronize_rcu();
> > > + kvfree(ptr);
> > > + }
> > > }
> > > EXPORT_SYMBOL_GPL(kvfree_call_rcu);
> > >
> > > --
> > > 2.20.1
> > >
next prev parent reply other threads:[~2020-05-04 0:28 UTC|newest]
Thread overview: 78+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-04-28 20:58 [PATCH 00/24] Introduce kvfree_rcu(1 or 2 arguments) Uladzislau Rezki (Sony)
2020-04-28 20:58 ` [PATCH 01/24] rcu/tree: Keep kfree_rcu() awake during lock contention Uladzislau Rezki (Sony)
2020-04-28 20:58 ` [PATCH 02/24] rcu/tree: Skip entry into the page allocator for PREEMPT_RT Uladzislau Rezki (Sony)
2020-04-28 20:58 ` [PATCH 03/24] rcu/tree: Use consistent style for comments Uladzislau Rezki (Sony)
2020-05-01 19:05 ` Paul E. McKenney
2020-05-01 20:52 ` Joe Perches
2020-05-03 23:44 ` Joel Fernandes
2020-05-04 0:23 ` Paul E. McKenney
2020-05-04 0:34 ` Joe Perches
2020-05-04 0:41 ` Joel Fernandes
2020-05-03 23:52 ` Joel Fernandes
2020-05-04 0:26 ` Paul E. McKenney
2020-05-04 0:39 ` Joel Fernandes
2020-04-28 20:58 ` [PATCH 04/24] rcu/tree: Repeat the monitor if any free channel is busy Uladzislau Rezki (Sony)
2020-04-28 20:58 ` [PATCH 05/24] rcu/tree: Simplify debug_objects handling Uladzislau Rezki (Sony)
2020-04-28 20:58 ` [PATCH 06/24] rcu/tree: Simplify KFREE_BULK_MAX_ENTR macro Uladzislau Rezki (Sony)
2020-04-28 20:58 ` [PATCH 07/24] rcu/tree: move locking/unlocking to separate functions Uladzislau Rezki (Sony)
2020-04-28 20:58 ` [PATCH 08/24] rcu/tree: Use static initializer for krc.lock Uladzislau Rezki (Sony)
2020-05-01 21:17 ` Paul E. McKenney
2020-05-04 12:10 ` Uladzislau Rezki
2020-04-28 20:58 ` [PATCH 09/24] rcu/tree: cache specified number of objects Uladzislau Rezki (Sony)
2020-05-01 21:27 ` Paul E. McKenney
2020-05-04 12:43 ` Uladzislau Rezki
2020-05-04 15:24 ` Paul E. McKenney
2020-05-04 17:48 ` Uladzislau Rezki
2020-05-04 18:07 ` Paul E. McKenney
2020-05-04 18:08 ` Joel Fernandes
2020-05-04 19:01 ` Paul E. McKenney
2020-05-04 19:37 ` Joel Fernandes
2020-05-04 19:51 ` Uladzislau Rezki
2020-05-04 20:15 ` joel
2020-05-04 20:16 ` Paul E. McKenney
2020-05-05 11:03 ` Uladzislau Rezki
2020-04-28 20:58 ` [PATCH 10/24] rcu/tree: add rcutree.rcu_min_cached_objs description Uladzislau Rezki (Sony)
2020-05-01 22:25 ` Paul E. McKenney
2020-05-04 12:44 ` Uladzislau Rezki
2020-04-28 20:58 ` [PATCH 11/24] rcu/tree: Maintain separate array for vmalloc ptrs Uladzislau Rezki (Sony)
2020-05-01 21:37 ` Paul E. McKenney
2020-05-03 23:42 ` Joel Fernandes
2020-05-04 0:20 ` Paul E. McKenney
2020-05-04 0:58 ` Joel Fernandes
2020-05-04 2:20 ` Paul E. McKenney
2020-05-04 14:25 ` Uladzislau Rezki
2020-04-28 20:58 ` [PATCH 12/24] rcu/tiny: support vmalloc in tiny-RCU Uladzislau Rezki (Sony)
2020-04-28 20:58 ` [PATCH 13/24] rcu: Rename rcu_invoke_kfree_callback/rcu_kfree_callback Uladzislau Rezki (Sony)
2020-04-28 20:58 ` [PATCH 14/24] rcu: Rename __is_kfree_rcu_offset() macro Uladzislau Rezki (Sony)
2020-04-28 20:58 ` [PATCH 15/24] rcu: Rename kfree_call_rcu() to the kvfree_call_rcu() Uladzislau Rezki (Sony)
2020-04-28 20:58 ` [PATCH 16/24] mm/list_lru.c: Rename kvfree_rcu() to local variant Uladzislau Rezki (Sony)
2020-04-28 20:58 ` [PATCH 17/24] rcu: Introduce 2 arg kvfree_rcu() interface Uladzislau Rezki (Sony)
2020-04-28 20:58 ` [PATCH 18/24] mm/list_lru.c: Remove kvfree_rcu_local() function Uladzislau Rezki (Sony)
2020-04-28 20:58 ` [PATCH 19/24] rcu/tree: Support reclaim for head-less object Uladzislau Rezki (Sony)
2020-05-01 22:39 ` Paul E. McKenney
2020-05-04 0:12 ` Joel Fernandes
2020-05-04 0:28 ` Paul E. McKenney [this message]
2020-05-04 0:32 ` Joel Fernandes
2020-05-04 14:21 ` Uladzislau Rezki
2020-05-04 15:31 ` Paul E. McKenney
2020-05-04 16:56 ` Uladzislau Rezki
2020-05-04 17:08 ` Paul E. McKenney
2020-05-04 12:57 ` Uladzislau Rezki
2020-04-28 20:58 ` [PATCH 20/24] rcu/tree: Make kvfree_rcu() tolerate any alignment Uladzislau Rezki (Sony)
2020-05-01 23:00 ` Paul E. McKenney
2020-05-04 0:24 ` Joel Fernandes
2020-05-04 0:29 ` Paul E. McKenney
2020-05-04 0:31 ` Joel Fernandes
2020-05-04 12:56 ` Uladzislau Rezki
2020-04-28 20:59 ` [PATCH 21/24] rcu/tiny: move kvfree_call_rcu() out of header Uladzislau Rezki (Sony)
2020-05-01 23:03 ` Paul E. McKenney
2020-05-04 12:45 ` Uladzislau Rezki
2020-05-06 18:29 ` Uladzislau Rezki
2020-05-06 18:45 ` Paul E. McKenney
2020-05-07 17:34 ` Uladzislau Rezki
2020-04-28 20:59 ` [PATCH 22/24] rcu/tiny: support reclaim for head-less object Uladzislau Rezki (Sony)
2020-05-01 23:06 ` Paul E. McKenney
2020-05-04 0:27 ` Joel Fernandes
2020-05-04 12:45 ` Uladzislau Rezki
2020-04-28 20:59 ` [PATCH 23/24] rcu: Introduce 1 arg kvfree_rcu() interface Uladzislau Rezki (Sony)
2020-04-28 20:59 ` [PATCH 24/24] lib/test_vmalloc.c: Add test cases for kvfree_rcu() Uladzislau Rezki (Sony)
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=20200504002855.GF2869@paulmck-ThinkPad-P72 \
--to=paulmck@kernel.org \
--cc=akpm@linux-foundation.org \
--cc=joel@joelfernandes.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=oleksiy.avramchenko@sonymobile.com \
--cc=rcu@vger.kernel.org \
--cc=tytso@mit.edu \
--cc=urezki@gmail.com \
--cc=willy@infradead.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).