triggers/cache-error-trigger

#!/bin/sh
# cache error trigger. This shell script is executed by mcelog in daemon mode
# when a CPU reports excessive corrected cache errors. This could be a indication
# for future uncorrected errors.
# 
# environment:
# MESSAGE	  Human readable error message
# CPU		  Linux CPU number that triggered the error
# LEVEL		  Cache level affected by error
# TYPE		  Cache type affected by error (Data,Instruction,Generic)
# AFFECTED_CPUS   List of CPUs sharing the affected cache
# SOCKETID	  Socket ID of affected CPU
#
# note: will run as mcelog configured user
# this can be changed in mcelog.conf

#
# Change "DO_OFFLINE" to "1" if you want to
# offline the CPUs sharing the affected cache
#
DO_OFFLINE=0

EXIT=0

logger -s -p daemon.crit -t mcelog "Level $LEVEL cache error threshold affects CPUs: $AFFECTED_CPUS"

if [ $DO_OFFLINE = 0 -o $AFFECTED_CPUS = unknown ]
then
	exit $EXIT
fi

if [ $(echo $AFFECTED_CPUS | wc -w) -ge $(grep -c ^processor /proc/cpuinfo) ]
then
	logger -s -p daemon.crit -t mcelog "Too many CPUs affected by cache error threshold to offline"
	exit $EXIT
fi

for i in $AFFECTED_CPUS ; do
	logger -s -p daemon.crit -t mcelog "Offlining CPU $i due to cache error threshold"
	F=$(printf "/sys/devices/system/cpu/cpu%d/online" $i)
	echo 0 > $F
	if [ "$(cat $F)" != "0" ] ; then
		logger -s -p daemon.warn -t mcelog "Offlining CPU $i failed"
		EXIT=1
	fi
done

[ -x ./cache-error-trigger.local ] && . ./cache-error-trigger.local

exit $EXIT