linux-kernel - Re: [PATCH] clocksource: Skip watchdog check for large watchdog intervals

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

Message-ID: <20240104164828.GD3303@incl>
Date: Thu, 4 Jan 2024 17:48:29 +0100
From: Jiri Wiesner <jwiesner@...e.de>
To: Feng Tang <feng.tang@...el.com>
Cc: linux-kernel@...r.kernel.org, John Stultz <jstultz@...gle.com>,
	Thomas Gleixner <tglx@...utronix.de>,
	Stephen Boyd <sboyd@...nel.org>,
	"Paul E. McKenney" <paulmck@...nel.org>
Subject: Re: [PATCH] clocksource: Skip watchdog check for large watchdog
 intervals

On Thu, Jan 04, 2024 at 01:55:21PM +0800, Feng Tang wrote:
> On Wed, Jan 03, 2024 at 12:21:13PM +0100, Jiri Wiesner wrote:
> > Fixes: 2e27e793e280 ("clocksource: Reduce clocksource-skew threshold")
> 
> IIUC, the issue could happen even before this commit? Though I do think 
> this is good stuff for stable kernel.

Yes, it could happen before 2e27e793e280 but the threshold was so loose 
that it probably was an extremely rare event. Commit 2e27e793e280 lowered 
the threshold and this significantly increased the likelihood of skew 
detection failing for large watchdog intervals. We got a report of a 
kernel with 2e27e793e280 switching to the HPET but there were no failures 
with a kernel without 2e27e793e280 on the same fleet of 8 NUMA node machines.

> 0Day robot reported some compiling issue.

I am going to submit a v2 of the patch fixing the error after we clarify 
the jiffies concern.
-- 
Jiri Wiesner
SUSE Labs