Date: Wed, 21 Dec 2016 14:39:56 +0000
From: Wei Yang <richard.weiyang@gmail.com>
To: Michal Hocko <mhocko@kernel.org>
Cc: Wei Yang <richard.weiyang@gmail.com>, trivial@kernel.org,
        akpm@linux-foundation.org, linux-mm@kvack.org,
        linux-kernel@vger.kernel.org
Subject: Re: [PATCH V2 2/2] mm/memblock.c: check return value of
 memblock_reserve() in memblock_virt_alloc_internal()
Message-ID: <20161221143956.GA23331@vultr.guest>
Reply-To: Wei Yang <richard.weiyang@gmail.com>
References: <1482072470-26151-1-git-send-email-richard.weiyang@gmail.com>
 <1482072470-26151-3-git-send-email-richard.weiyang@gmail.com>
 <20161219152156.GC5175@dhcp22.suse.cz>
 <20161220164823.GB13224@vultr.guest>
 <20161221075115.GE16502@dhcp22.suse.cz>
 <20161221131332.GB23096@vultr.guest>
 <20161221132200.GK31118@dhcp22.suse.cz>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20161221132200.GK31118@dhcp22.suse.cz>
User-Agent: Mutt/1.5.23 (2014-03-12)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1756
Lines: 49

On Wed, Dec 21, 2016 at 02:22:01PM +0100, Michal Hocko wrote:
>On Wed 21-12-16 13:13:32, Wei Yang wrote:
>> On Wed, Dec 21, 2016 at 08:51:16AM +0100, Michal Hocko wrote:
>> >On Tue 20-12-16 16:48:23, Wei Yang wrote:
>> >> On Mon, Dec 19, 2016 at 04:21:57PM +0100, Michal Hocko wrote:
>> >> >On Sun 18-12-16 14:47:50, Wei Yang wrote:
>> >> >> memblock_reserve() may fail in case there is not enough regions.
>> >> >
>> >> >Have you seen this happenning in the real setups or this is a by-review
>> >> >driven change?
>> >> 
>> >> This is a by-review driven change.
>> >> 
>> >> >[...]
>> >> >>  again:
>> >> >>  	alloc = memblock_find_in_range_node(size, align, min_addr, max_addr,
>> >> >>  					    nid, flags);
>> >> >> -	if (alloc)
>> >> >> +	if (alloc && !memblock_reserve(alloc, size))
>> >> >>  		goto done;
>> >
>> >So how exactly does the reserve fail when memblock_find_in_range_node
>> >found a suitable range for the given size?
>> >
>> 
>> Even memblock_find_in_range_node() gets a suitable range, memblock_reserve()
>> still could fail. And the case just happens when memblock can't resize.
>> memblock_reserve() reserve a range by adding a range to memblock.reserved. In
>> case the memblock.reserved is full and can't resize, this fails.
>
>Sorry for being dense but what does it mean that the reserved will get
>full? Also how probable is such a situation? Is it even real? In other
>words does this fix a real or only a theoretical problem?
>

This is a theoretical problem. While if happens, it is hard to detect. Future
allocator will think this range is still available.

>Anyway this all should be part of the changelog.

Ok, let me add this in changelog in next version.

>-- 
>Michal Hocko
>SUSE Labs

-- 
Wei Yang
Help you, Help me