obmc-initfs shutdown: ping watchdog watchdog during flash update
Add code to start a watchdog daemon before calling the update script
as erasing the flash can take over 10 minutes. This allows us to
reduce the shutdown timeout to a much more useful value.
The code kills the watchdog after the update script returns and
also resets the timeout to a different value by invoking and killing
the watchdog command a second time.
Tested with the following code added before the reboot -f code and
changing the timoeuts:
while :
do
let count=$count+1
echo -ne "\rloop $count"
usleep 100
done
Change-Id: Ibe2337339af67421733230b6a5b6df962618a00a
Signed-off-by: Milton Miller <miltonm@us.ibm.com>
diff --git a/meta-phosphor/common/recipes-phosphor/initfs/files/obmc-shutdown.sh b/meta-phosphor/common/recipes-phosphor/initfs/files/obmc-shutdown.sh
index 90e5dfb..c91f5ef 100644
--- a/meta-phosphor/common/recipes-phosphor/initfs/files/obmc-shutdown.sh
+++ b/meta-phosphor/common/recipes-phosphor/initfs/files/obmc-shutdown.sh
@@ -33,10 +33,21 @@
update=/run/initramfs/update
image=/run/initramfs/image-
+wdt="-t 1 -T 5"
+wdrst="-T 15"
+
if ls $image* > /dev/null 2>&1
then
if test -x $update
then
+ if test -c /dev/watchdog
+ then
+ echo Pinging watchdog ${wdt+with args $wdt}
+ watchdog $wdt -F /dev/watchdog &
+ wd=$!
+ else
+ wd=
+ fi
$update --clean-saved-files
remaining=$(ls $image*)
if test -n "$remaining"
@@ -46,6 +57,20 @@
else
echo "Flash update completed."
fi
+
+ if test -n "$wd"
+ then
+ kill -9 $wd
+ if test -n "$wdrst"
+ then
+ echo Reseting watchdog timeouts to $wdrst
+ watchdog $wdrst -F /dev/watchdog &
+ sleep 1
+ # Kill the watchdog daemon, setting a timeout
+ # for the remaining shutdown work
+ kill -9 $!
+ fi
+ fi
else
echo 1>&2 "Flash update requested but $update program missing!"
fi