Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Mon, 7 Jan 2019 09:52:30 +0000
From:   Catalin Marinas <catalin.marinas@arm.com>
To:     Nathan Royce <nroycea+kernel@gmail.com>
Cc:     linux-kernel@vger.kernel.org, linux-mm@kvack.org
Subject: Re: kmemleak: Cannot allocate a kmemleak_object structure - Kernel
 4.19.13
Message-ID: <20190107095229.uvfuxpglreibxlo4@mbp>
References: <CALaQ_hpCKoLxp-0cgxw9TqPGBSzY7RhrnFZ0jGAQ11HbOZkZ3w@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <CALaQ_hpCKoLxp-0cgxw9TqPGBSzY7RhrnFZ0jGAQ11HbOZkZ3w@mail.gmail.com>
User-Agent: NeoMutt/20170113 (1.7.2)
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Hi Nathan,

On Tue, Jan 01, 2019 at 01:17:06PM -0600, Nathan Royce wrote:
> I had a leak somewhere and I was directed to look into SUnreclaim
> which was 5.5 GB after an uptime of a little over 1 month on an 8 GB
> system. kmalloc-2048 was a problem.
> I just had enough and needed to find out the cause for my lagging system.
> 
> I finally upgraded from 4.18.16 to 4.19.13 and enabled kmemleak to
> hunt for the culprit. I don't think a day had elapsed before kmemleak
> crashed and disabled itself.

Under memory pressure, kmemleak may fail to allocate memory. See this
patch for an attempt to slightly improve things but it's not a proper
solution:

http://lkml.kernel.org/r/20190102180619.12392-1-cai@lca.pw

-- 
Catalin