From: joel@joelfernandes.org
To: Michal Hocko <mhocko@kernel.org>
Cc: linux-kernel@vger.kernel.org, linux-mm@kvack.org,
rcu@vger.kernel.org, willy@infradead.org, peterz@infradead.org,
neilb@suse.com, vbabka@suse.cz, mgorman@suse.de,
Andrew Morton <akpm@linux-foundation.org>,
Josh Triplett <josh@joshtriplett.org>,
Lai Jiangshan <jiangshanlai@gmail.com>,
Mathieu Desnoyers <mathieu.desnoyers@efficios.com>,
"Paul E. McKenney" <paulmck@kernel.org>,
Steven Rostedt <rostedt@goodmis.org>
Subject: Re: [PATCH RFC] rcu/tree: Use GFP_MEMALLOC for alloc memory to free memory pattern
Date: Wed, 01 Apr 2020 07:14:16 -0400 [thread overview]
Message-ID: <30295C90-34DB-469C-9DCD-55DB91938BA9@joelfernandes.org> (raw)
In-Reply-To: <20200401072359.GC22681@dhcp22.suse.cz>
On April 1, 2020 3:23:59 AM EDT, Michal Hocko <mhocko@kernel.org> wrote:
>On Tue 31-03-20 12:01:17, Joel Fernandes wrote:
>> On Tue, Mar 31, 2020 at 05:34:50PM +0200, Michal Hocko wrote:
>> > On Tue 31-03-20 10:58:06, Joel Fernandes wrote:
>> > [...]
>> > > > diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c
>> > > > index 4be763355c9fb..965deefffdd58 100644
>> > > > --- a/kernel/rcu/tree.c
>> > > > +++ b/kernel/rcu/tree.c
>> > > > @@ -3149,7 +3149,7 @@ static inline struct rcu_head
>*attach_rcu_head_to_object(void *obj)
>> > > >
>> > > > if (!ptr)
>> > > > ptr = kmalloc(sizeof(unsigned long *) +
>> > > > - sizeof(struct rcu_head), GFP_ATOMIC | __GFP_NOWARN);
>> > > > + sizeof(struct rcu_head), GFP_MEMALLOC);
>> > >
>> > > Just to add, the main requirements here are:
>> > > 1. Allocation should be bounded in time.
>> > > 2. Allocation should try hard (possibly tapping into reserves)
>> > > 3. Sleeping is Ok but should not affect the time bound.
>> >
>> >
>> > __GFP_ATOMIC | __GFP_HIGH is the way to get an additional access to
>> > memory reserves regarless of the sleeping status.
>> >
>> > Using __GFP_MEMALLOC is quite dangerous because it can deplete
>_all_ the
>> > memory. What does prevent the above code path to do that?
>
>Neil has provided a nice explanation down the email thread. But let me
>clarify few things here.
>
>> Can you suggest what prevents other users of GFP_MEMALLOC from doing
>that
>> also?
>
>There is no explicit mechanism which is indeed unfortunate. The only
>user real user of the flag is Swap over NFS AFAIK. I have never dared
>to
>look into details on how the complete reserves depletion is prevented.
>Mel would be much better fit here.
>
>> That's the whole point of having a reserve, in normal usage no one
>will
>> use it, but some times you need to use it. Keep in mind this is not a
>common
>> case in this code here, this is triggered only if earlier allocation
>attempts
>> failed. Only *then* we try with GFP_MEMALLOC with promises to free
>additional
>> memory soon.
>
>You are right that this is the usecase for the flag. But this should be
>done with an extreme care because the core MM relies on those reserves
>so any other users should better make sure they do not consume a lot
>from reserves as well.
>
Understood and agreed.
>> > If a partial access to reserves is sufficient then why the existing
>> > modifiers (mentioned above are not sufficient?
>>
>> The point with using GFP_MEMALLOC is it is useful for situations
>where you
>> are about to free memory and needed some memory temporarily, to free
>that. It
>> depletes it a bit temporarily to free even more. Is that not the
>point of
>> PF_MEMALLOC?
>> * %__GFP_MEMALLOC allows access to all memory. This should only be
>used when
>> * the caller guarantees the allocation will allow more memory to be
>freed
>> * very shortly e.g. process exiting or swapping. Users either should
>> * be the MM or co-ordinating closely with the VM (e.g. swap over
>NFS).
>>
>> I was just recommending usage of this flag here because it fits the
>> requirement of allocating some memory to free some memory. I am also
>Ok with
>> GFP_ATOMIC with the GFP_NOWARN removed, if you are Ok with that.
>
>Maybe we need to refine this documentation to be more explicit about an
>extreme care to be taken when using the flag.
>
>diff --git a/include/linux/gfp.h b/include/linux/gfp.h
>index e5b817cb86e7..e436a7e28392 100644
>--- a/include/linux/gfp.h
>+++ b/include/linux/gfp.h
>@@ -110,6 +110,9 @@ struct vm_area_struct;
>* the caller guarantees the allocation will allow more memory to be
>freed
> * very shortly e.g. process exiting or swapping. Users either should
> * be the MM or co-ordinating closely with the VM (e.g. swap over NFS).
>+ * Users of this flag have to be extremely careful to not deplete the
>reserve
>+ * completely and implement a throttling mechanism which controls the
>consumption
>+ * based on the amount of freed memory.
> *
>* %__GFP_NOMEMALLOC is used to explicitly forbid access to emergency
>reserves.
> * This takes precedence over the %__GFP_MEMALLOC flag if both are set.
I am in support of this documentation patch. I would say "consumption of the reserve".
Thanks,
- Joel
--
Sent from my Android device with K-9 Mail. Please excuse my brevity.
next prev parent reply other threads:[~2020-04-01 11:14 UTC|newest]
Thread overview: 38+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-03-31 13:16 [PATCH RFC] rcu/tree: Use GFP_MEMALLOC for alloc memory to free memory pattern Joel Fernandes (Google)
2020-03-31 14:04 ` Uladzislau Rezki
2020-03-31 15:09 ` Joel Fernandes
2020-03-31 16:01 ` Uladzislau Rezki
2020-03-31 17:02 ` Uladzislau Rezki
2020-03-31 17:49 ` Paul E. McKenney
2020-03-31 18:30 ` Joel Fernandes
2020-04-01 12:25 ` Uladzislau Rezki
2020-04-01 13:47 ` Paul E. McKenney
2020-04-01 18:16 ` Uladzislau Rezki
2020-04-01 18:26 ` Paul E. McKenney
2020-04-01 18:37 ` Uladzislau Rezki
2020-04-01 18:54 ` Paul E. McKenney
2020-04-01 19:05 ` Uladzislau Rezki
2020-04-01 19:34 ` Paul E. McKenney
2020-04-01 19:35 ` Uladzislau Rezki
2020-03-31 14:58 ` Joel Fernandes
2020-03-31 15:34 ` Michal Hocko
2020-03-31 16:01 ` Joel Fernandes
2020-03-31 22:19 ` NeilBrown
2020-04-01 3:25 ` Joel Fernandes
2020-04-01 4:52 ` NeilBrown
2020-04-01 7:23 ` Michal Hocko
2020-04-01 11:14 ` joel [this message]
2020-04-01 12:05 ` Michal Hocko
2020-04-01 13:14 ` Mel Gorman
2020-04-01 14:45 ` Joel Fernandes
2020-03-31 16:12 ` Uladzislau Rezki
2020-04-01 7:09 ` Michal Hocko
2020-04-01 12:32 ` Uladzislau Rezki
2020-04-01 12:55 ` Michal Hocko
2020-04-01 13:08 ` Uladzislau Rezki
2020-04-01 13:15 ` Michal Hocko
2020-04-01 13:22 ` Uladzislau Rezki
2020-04-01 15:28 ` Michal Hocko
2020-04-01 15:46 ` Uladzislau Rezki
2020-04-01 15:57 ` Paul E. McKenney
2020-04-01 16:10 ` Michal Hocko
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=30295C90-34DB-469C-9DCD-55DB91938BA9@joelfernandes.org \
--to=joel@joelfernandes.org \
--cc=akpm@linux-foundation.org \
--cc=jiangshanlai@gmail.com \
--cc=josh@joshtriplett.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mathieu.desnoyers@efficios.com \
--cc=mgorman@suse.de \
--cc=mhocko@kernel.org \
--cc=neilb@suse.com \
--cc=paulmck@kernel.org \
--cc=peterz@infradead.org \
--cc=rcu@vger.kernel.org \
--cc=rostedt@goodmis.org \
--cc=vbabka@suse.cz \
--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).