linux-kernel - [Fwd: 11 minute NTP hw clock update racy?]

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [day] [month] [year] [list]

Message-ID: <6e9971f10c75537582abd5c77d798944080b716a.camel@infinera.com>
Date:   Tue, 28 Aug 2018 07:43:34 +0000
From:   Joakim Tjernlund <Joakim.Tjernlund@...inera.com>
To:     "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: [Fwd: 11 minute NTP hw clock update racy?]

No luck on linuxppc-dev, trying LKML ...

-------- Forwarded Message --------
From: Joakim Tjernlund <joakim.tjernlund@...inera.com>
To: linuxppc-dev linuxppc-dev <linuxppc-dev@...ts.ozlabs.org>
Subject: 11 minute NTP hw clock update racy?
Date: Mon, 27 Aug 2018 10:01:12 +0200

We see corrupt HW clock time every now and then(really hard to reproduce)
Our RTC is a DS1388 on an I2C bus.

Looking at ntp_notify_cmos_timer() and it's delayed work queue impl. I wonder
if there could be a race here w.r.t reboot ?

Could the 11 minute update kick in just as the system is about to reset
the CPU?

I am on 4.14.51, ppc32 and using the ppc_md.restart() hook which will
reset the CPU immediately.

Question, is safe to call ntp_notify_cmos_timer() when the work queue is already
armed(like do_adjtimex() does) ? 

 Jocke