Re: WARNING: at mm/mremap.c:211 move_page_tables in i386

From: Naresh Kamboju <naresh.kamboju@linaro.org>
To: Arnd Bergmann <arnd@arndb.de>
Cc: "linux- stable" <stable@vger.kernel.org>,
	"open list" <linux-kernel@vger.kernel.org>,
	linux-mm <linux-mm@kvack.org>,
	"Andrew Morton" <akpm@linux-foundation.org>,
	"Roman Gushchin" <guro@fb.com>,
	"Michal Hocko" <mhocko@kernel.org>,
	lkft-triage@lists.linaro.org, "Chris Down" <chris@chrisdown.name>,
	"Linus Torvalds" <torvalds@linux-foundation.org>,
	"Michel Lespinasse" <walken@google.com>,
	Fan_Yang@sjtu.edu.cn, bgeffon@google.com,
	"Anshuman Khandual" <anshuman.khandual@arm.com>,
	"Will Deacon" <will@kernel.org>,
	"Catalin Marinas" <catalin.marinas@arm.com>,
	pugaowei@gmail.com, "Jérôme Glisse" <jglisse@redhat.com>,
	"Joel Fernandes" <joel@joelfernandes.org>,
	"Greg Kroah-Hartman" <gregkh@linuxfoundation.org>,
	"Mel Gorman" <mgorman@techsingularity.net>,
	"Hugh Dickins" <hughd@google.com>,
	"Al Viro" <viro@zeniv.linux.org.uk>, "Tejun Heo" <tj@kernel.org>,
	"Sasha Levin" <sashal@kernel.org>
Subject: Re: WARNING: at mm/mremap.c:211 move_page_tables in i386
Date: Fri, 10 Jul 2020 09:47:35 +0530	[thread overview]
Message-ID: <CA+G9fYsoEiQ-8ECKxMqQqPZKUbTQStp6wZi7ZiJDyi0YahFAvg@mail.gmail.com> (raw)
In-Reply-To: <CAK8P3a0ii1Z-UG8NpwTvkmOEcOvvTo4+m9xjW0JqR6LPvUZ4=g@mail.gmail.com>

On Thu, 9 Jul 2020 at 13:55, Arnd Bergmann <arnd@arndb.de> wrote:
>
> On Thu, Jul 9, 2020 at 7:28 AM Naresh Kamboju <naresh.kamboju@linaro.org> wrote:
> >
> > While running LTP mm test suite on i386 or qemu_i386 this kernel warning
> > has been noticed from stable 5.4 to stable 5.7 branches and mainline 5.8.0-rc4
> > and linux next.
>
> Are you able to correlate this with any particular test case in LTP, or does
> it happen for random processes?
>
> In the log you linked to, it happens once for ksm05.c and multiple times for
> thp01.c, sources here:
>
> https://github.com/linux-test-project/ltp/blob/master/testcases/kernel/mem/ksm/ksm05.c
> https://github.com/linux-test-project/ltp/blob/master/testcases/kernel/mem/thp/thp01.c
>
> Is it always these tests that trigger the warning, or sometimes others?

These two test cases are causing this warning multiple times on i386.

>
> When you say it happens with linux-5.4 stable, does that mean you don't see
> it with older versions? What is the last known working version?

I do not notice on stable-4.19 and below versions.
Sorry i did not get the known working commit id or version.

It started happening from stable-rc 5.0 first release.
I have evidence [1] showing it on 5.0.1

>
> I also see that you give the virtual machine 16GB of RAM, but as you are
> running a 32-bit kernel without PAE, only 2.3GB end up being available,
> and some other LTP tests in the log run out of memory.
>
> You could check if the behavior changes if you give the kernel less memory,
> e.g. 768MB (lowmem only), or enable CONFIG_X86_PAE to let it use the
> entire 16GB.

Warning is still happening after enabling PAE config.
But the oom-killer messages are gone. Thank you.

CONFIG_HIGHMEM=y
CONFIG_X86_PAE=y

full test log oom-killer messages are gone and kernel warning is still there,
https://lkft.validation.linaro.org/scheduler/job/1552606#L10357

build location:
https://builds.tuxbuild.com/puilcMcGVwzFMN5fDUhY4g/

[1] https://qa-reports.linaro.org/lkft/linux-stable-rc-5.0-oe/build/v5.0.1/testrun/1324990/suite/ltp-mm-tests/test/ksm02/log

---
[  775.646689] WARNING: CPU: 3 PID: 10858 at
/srv/oe/build/tmp-lkft-glibc/work-shared/intel-core2-32/kernel-source/mm/mremap.c:211
move_page_tables+0x553/0x570
[  775.647006] Modules linked in: fuse
[  775.647006] CPU: 3 PID: 10858 Comm: true Not tainted 5.0.1 #1
[  775.647006] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996),
BIOS 1.10.2-1 04/01/2014
[  775.647006] EIP: move_page_tables+0x553/0x570

- Naresh