clocksource: Check per-CPU clock synchronization when marked unstable
authorPaul E. McKenney <paulmck@kernel.org>
Thu, 27 May 2021 19:01:20 +0000 (12:01 -0700)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Wed, 14 Jul 2021 14:56:01 +0000 (16:56 +0200)
commitd9b40ebd448e437ffbc65f013836f98252279a82
treeabf4258039d336339ff1df3ceef1847eb230014a
parent03a65c14ab47b287bede2f414ca212db1ebdc8c0
clocksource: Check per-CPU clock synchronization when marked unstable

[ Upstream commit 7560c02bdffb7c52d1457fa551b9e745d4b9e754 ]

Some sorts of per-CPU clock sources have a history of going out of
synchronization with each other.  However, this problem has purportedy been
solved in the past ten years.  Except that it is all too possible that the
problem has instead simply been made less likely, which might mean that
some of the occasional "Marking clocksource 'tsc' as unstable" messages
might be due to desynchronization.  How would anyone know?

Therefore apply CPU-to-CPU synchronization checking to newly unstable
clocksource that are marked with the new CLOCK_SOURCE_VERIFY_PERCPU flag.
Lists of desynchronized CPUs are printed, with the caveat that if it
is the reporting CPU that is itself desynchronized, it will appear that
all the other clocks are wrong.  Just like in real life.

Reported-by: Chris Mason <clm@fb.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Feng Tang <feng.tang@intel.com>
Link: https://lore.kernel.org/r/20210527190124.440372-2-paulmck@kernel.org
Signed-off-by: Sasha Levin <sashal@kernel.org>
arch/x86/kernel/tsc.c
include/linux/clocksource.h
kernel/time/clocksource.c