coredumps in syscall 279

I'm noticing numerous coredumps in dmesg relating to syscall 279:

$this->bbcode_second_pass_quote('', '
')$ uname -a
Linux odroid-001e06336dd6 3.14.79-11-ARCH #1 SMP PREEMPT Thu Nov 3 19:23:19 MDT 2016 aarch64 GNU/Linux
$ dmesg |grep 'syscall 279'[ 125.002267] pactl[498]: syscall 279
[ 136.069361] chromium[666]: syscall 279
[ 393.013636] chromium[666]: syscall 279
[ 689.150825] pcmanfm[482]: syscall 279
[ 692.600774] lxterminal[1108]: syscall 279
[ 5335.157281] chromium[3008]: syscall 279

[WarheadsSE]

Compared to mainline, http://lxr.free-electrons.com/source/in ... std.h#L709

Compared to mainline 3.14, http://lxr.free-electrons.com/source/in ... =3.14#L701

Compared to the headers for the ODROID-C2, https://github.com/hardkernel/linux/blo ... std.h#L704

Happened to notice the message as well with thunar, and while I was at it in a terminal, I got the extract
below via strace:
$this->bbcode_second_pass_quote('', '
')memfd_create("pulseaudio", MFD_ALLOW_SEALING) = -1 ENOSYS (Function not implemented)
write(2, "shared memfd open() failed: Fonc"..., 52) = 52

[kroylar]

These messages also appear with lxqt-panel.

I modified the pulseaudio PKGBUILD and added --disable-memfd to the configure arguments and I no longer receive these messages.

What's the proper way to make this work within archlinuxarm besides waiting for amlogic to update their kernel or for odroids to get mainlined? A new pulseaudio-odroid package?

Ideally, glibc would be built on an odroid-c2 running the 3.14 kernel so that syscall would not exist in bits/syscall.h. And when pulseaudio is built (again, on the odroid-c2), its configure script would detect that memfd isn't available and automatically exclude it.

Of course, that can't be expected. But what is the right way to do this kind of thing?

[WarheadsSE]

It may have been built on something more mainline, and capable than an ODROID-C2 (think server-grade AArch64 hardware)