Re: [PATCH RFC] rcu/tree: Use GFP_MEMALLOC for alloc memory to free memory pattern

From: Michal Hocko <mhocko@kernel.org>
To: Uladzislau Rezki <urezki@gmail.com>
Cc: Joel Fernandes <joel@joelfernandes.org>,
	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, 1 Apr 2020 15:15:28 +0200	[thread overview]
Message-ID: <20200401131528.GK22681@dhcp22.suse.cz> (raw)
In-Reply-To: <20200401130816.GA1320@pc636>

On Wed 01-04-20 15:08:16, Uladzislau Rezki wrote:
> On Wed, Apr 01, 2020 at 02:55:03PM +0200, Michal Hocko wrote:
> > On Wed 01-04-20 14:32:30, Uladzislau Rezki wrote:
> > > On Wed, Apr 01, 2020 at 09:09:58AM +0200, Michal Hocko wrote:
> > > > On Tue 31-03-20 18:12:15, Uladzislau Rezki wrote:
> > > > > > 
> > > > > > __GFP_ATOMIC | __GFP_HIGH is the way to get an additional access to
> > > > > > memory reserves regarless of the sleeping status.
> > > > > > 
> > > > > Michal, just one question here regarding proposed flags. Can we also
> > > > > tight it with __GFP_RETRY_MAYFAIL flag? Means it also can repeat a few
> > > > > times in order to increase the chance of being success.
> > > > 
> > > > yes, __GFP_RETRY_MAYFAIL is perfectly valid with __GFP_ATOMIC. Please
> > > > note that __GFP_ATOMIC, despite its name, doesn't imply an atomic
> > > > allocation which cannot sleep. Quite confusing, I know. A much better
> > > > name would be __GFP_RESERVES or something like that.
> > > > 
> > > OK. Then we can use GFP_ATOMIC | __GFP_RETRY_MAYFAIL to try in more harder
> > > way.
> > 
> > Please note the difference between __GFP_ATOMIC and GFP_ATOMIC. The
> > later is a highlevel flag to use for atomic contexts. The former is an
> > explicit way to give an access to memory reserves. I am not familiar
> > with your code but if you have an existing gfp context coming from the
> > caller then just do (gfp | __GFP_ATOMIC | __GFP_HIGH | __GFP_RETRY_MAYFAIL).
> > If you do not have any gfp then decide based on whether the current
> > context is allowed to sleep
> > 	gfp = GFP_KERNEL | __GFP_ATOMIC | __GFP_HIGH | __GFP_RETRY_MAYFAIL;
> > 	if (!sleepable)
> > 		gfp &= ~__GFP_DIRECT_RECLAIM;
> 
> We call it from atomic context, so we can not sleep, also we do not have
> any existing context coming from the caller. I see that GFP_ATOMIC is high-level
> flag and is differ from __GFP_ATOMIC. It is defined as:
> 
> #define GFP_ATOMIC (__GFP_HIGH|__GFP_ATOMIC|__GFP_KSWAPD_RECLAIM)
> 
> so basically we would like to have __GFP_KSWAPD_RECLAIM that is included in it,
> because it will also help in case of high memory pressure and wake-up kswapd to
> reclaim memory.
> 
> We also can extract:
> 
> __GFP_ATOMIC | __GFP_HIGH | __GFP_RETRY_MAYFAIL | __GFP_KSWAPD_RECLAIM
> 
> but that is longer then
> 
> GFP_ATMOC |  __GFP_RETRY_MAYFAIL

OK, if you are always in the atomic context then GFP_ATOMIC is
sufficient. __GFP_RETRY_MAYFAIL will make no difference for allocations
which do not reclaim (and thus not retry). Sorry this was not clear to
me from the previous description.

-- 
Michal Hocko
SUSE Labs