add support for control groups version 2

This is for #94.

4 files changed, 182 insertions, 29 deletions

diff --git a/etc/rc.conf b/etc/rc.conf
index 689e6be2..d9ad911d 100644
--- a/etc/rc.conf
+++ b/etc/rc.conf
@@ -191,13 +191,43 @@ rc_tty_number=12
 ##############################################################################
 # LINUX CGROUPS RESOURCE MANAGEMENT
 
-# If you have cgroups turned on in your kernel, this switch controls
-# whether or not a group for each controller is mounted under
-# /sys/fs/cgroup.
-# None of the other options in this section work if this is set to "NO".
+# This sets the mode used to mount cgroups.
+# "hybrid" mounts cgroups version 2 on /sys/fs/cgroup/unified and
+# cgroups version 1 on /sys/fs/cgroup.
+# "legacy" mounts cgroups version 1 on /sys/fs/cgroup
+# "unified" mounts cgroups version 2 on /sys/fs/cgroup
+#rc_cgroup_mode="hybrid"
+
+# This is a list of controllers which should be enabled for cgroups version 2.
+# If hybrid mode is being used, controllers listed here will not be
+# available for cgroups version 1.
+# This is a global setting.
+#rc_cgroup_controllers=""
+
+# This variable contains the cgroups version 2 settings for your services.
+# If this is set in this file, the settings will apply to all services.
+# If you want different settings for each service, place the settings in
+# /etc/conf.d/foo for service foo.
+# The format is to specify the setting and value followed by a newline.
+# Multiple settings and values can be specified.
+# For example, you would use this to set the maximum memory and maximum
+# number of pids for a service.
+#rc_cgroup_settings="
+#memory.max 10485760
+#pids.max max
+#"
+#
+# For more information about the adjustments that can be made with
+# cgroups version 2, see Documentation/cgroups-v2.txt in the linux kernel
+# source tree.
+#rc_cgroup_settings=""
+
+# This switch controls whether or not cgroups version 1 controllers are
+# individually mounted under
+# /sys/fs/cgroup in hybrid or legacy mode.
 #rc_controller_cgroups="YES"
 
-# The following settings allow you to set up values for the cgroup
+# The following settings allow you to set up values for the cgroups version 1
 # controllers for your services.
 # They can be set in this file;, however, if you do this, the settings
 # will apply to all of your services.
@@ -211,8 +241,9 @@ rc_tty_number=12
 # cpu.shares 512
 # "
 #
-#For more information about the adjustments that can be made with
-#cgroups, see Documentation/cgroups/* in the linux kernel source tree.
+# For more information about the adjustments that can be made with
+# cgroups version 1, see Documentation/cgroups-v1/* in the linux kernel
+# source tree.
 
 # Set the blkio controller settings for this service.
 #rc_cgroup_blkio=""
diff --git a/init.d/sysfs.in b/init.d/sysfs.in
index a2538114..9f39fb57 100644
--- a/init.d/sysfs.in
+++ b/init.d/sysfs.in
@@ -107,20 +107,16 @@ mount_misc()
 	fi
 }
 
-mount_cgroups()
+cgroup1_base()
 {
-	# set up kernel support for cgroups
-	if [ -d /sys/fs/cgroup ] && ! mountinfo -q /sys/fs/cgroup; then
-		if grep -qs cgroup /proc/filesystems; then
-			ebegin "Mounting cgroup filesystem"
-			local opts="${sysfs_opts},mode=755,size=${rc_cgroupsize:-10m}"
-			mount -n -t tmpfs -o ${opts} cgroup_root /sys/fs/cgroup
-			eend $?
-		fi
+	grep -qw cgroup /proc/filesystems || return 0
+	if ! mountinfo -q /sys/fs/cgroup; then
+		ebegin "Mounting cgroup filesystem"
+		local opts="${sysfs_opts},mode=755,size=${rc_cgroupsize:-10m}"
+		mount -n -t tmpfs -o "${opts}" cgroup_root /sys/fs/cgroup
+		eend $?
 	fi
 
-	mountinfo -q /sys/fs/cgroup || return 0
-
 	if ! mountinfo -q /sys/fs/cgroup/openrc; then
 		local agent="@LIBEXECDIR@/sh/cgroup-release-agent.sh"
 		mkdir /sys/fs/cgroup/openrc
@@ -129,17 +125,87 @@ mount_cgroups()
 			openrc /sys/fs/cgroup/openrc
 		printf 1 > /sys/fs/cgroup/openrc/notify_on_release
 	fi
+	return 0
+}
 
-	yesno ${rc_controller_cgroups:-YES} && [ -e /proc/cgroups ] || return 0
-	while read name hier groups enabled rest; do
+cgroup1_controllers()
+{
+	yesno "${rc_controller_cgroups:-YES}" && [ -e /proc/cgroups ] || return 0
+	while read -r name _ _ enabled rest; do
 		case "${enabled}" in
-			1)	mountinfo -q /sys/fs/cgroup/${name} && continue
-				mkdir /sys/fs/cgroup/${name}
-				mount -n -t cgroup -o ${sysfs_opts},${name} \
-					${name} /sys/fs/cgroup/${name}
+			1)	mountinfo -q "/sys/fs/cgroup/${name}" && continue
+				local x
+				for x in $rc_cgroup_controllers; do
+				[ "${name}" = "blkio" ] && [ "${x}" = "io" ] &&
+					continue 2
+				[ "${name}" = "${x}" ] &&
+				continue 2
+				done
+				mkdir "/sys/fs/cgroup/${name}"
+				mount -n -t cgroup -o "${sysfs_opts},${name}" \
+					"${name}" "/sys/fs/cgroup/${name}"
 				;;
 		esac
 	done < /proc/cgroups
+	return 0
+}
+
+cgroup2_controllers()
+{
+	local active cgroup_path x y
+	cgroup_path="$(cgroup2_find_path)"
+	[ -z "${cgroup_path}" ] && return 0
+	[ -e "${cgroup_path}/cgroup.controllers" ] &&
+	read -r active < "${cgroup_path}/cgroup.controllers"
+	for x in ${rc_cgroup_controllers}; do
+		for y in ${active}; do
+		[ "$x" = "$y" ] &&
+			[ -e "${cgroup_path}/cgroup.subtree_control" ]&&
+			echo "+${x}"  > "${cgroup_path}/cgroup.subtree_control"
+		done
+	done
+	return 0
+}
+
+cgroups_hybrid()
+{
+	grep -qw cgroup /proc/filesystems &&
+		grep -qw cgroup2 /proc/filesystems ||
+		return 0
+	cgroup1_base
+	mkdir /sys/fs/cgroup/unified
+	mount -t cgroup2 none -o "${sysfs_opts},nsdelegate" /sys/fs/cgroup/unified
+	cgroup2_controllers
+	cgroup1_controllers
+	return 0
+}
+
+cgroups_legacy()
+{
+	grep -qw cgroup /proc/filesystems || return 0
+	cgroup1_base
+	cgroup1_controllers
+	return 0
+}
+
+cgroups_unified()
+{
+	grep -qw cgroup2 /proc/filesystems || return 0
+	mount -t cgroup2 none -o "${sysfs_opts},nsdelegate" /sys/fs/cgroup
+	return 0
+}
+
+mount_cgroups()
+{
+	# set up kernel support for cgroups
+	if [ -d /sys/fs/cgroup ]; then
+		case "${rc_cgroup_mode:-hybrid}" in
+		hybrid) cgroups_hybrid ;;
+		legacy) cgroups_legacy ;;
+		unified) cgroups_unified ;;
+		esac
+	fi
+	return 0
 }
 
 restorecon_sys()
diff --git a/sh/openrc-run.sh.in b/sh/openrc-run.sh.in
index a38d46d6..e778bd09 100644
--- a/sh/openrc-run.sh.in
+++ b/sh/openrc-run.sh.in
@@ -258,8 +258,7 @@ for _cmd; do
 		[ -n "${rc_ulimit:-$RC_ULIMIT}" ] && \
 			ulimit ${rc_ulimit:-$RC_ULIMIT}
 		# Apply cgroups settings if defined
-		if [ "$(command -v cgroup_add_service)" = \
-		    "cgroup_add_service" ]
+		if [ "$(command -v cgroup_add_service)" = "cgroup_add_service" ]
 		then
 			if [ -d /sys/fs/cgroup -a ! -w /sys/fs/cgroup ]; then
 				eerror "No permission to apply cgroup settings"
@@ -268,9 +267,11 @@ for _cmd; do
 			cgroup_add_service /sys/fs/cgroup/openrc
 			cgroup_add_service /sys/fs/cgroup/systemd/system
 		fi
-		[ "$(command -v cgroup_set_limits)" = \
-		    "cgroup_set_limits" ] && \
-		    cgroup_set_limits
+		[ "$(command -v cgroup_set_limits)" = "cgroup_set_limits" ] &&
+			cgroup_set_limits
+		[ "$(command -v cgroup2_set_limits)" = "cgroup2_set_limits" ] &&
+			[ "$_cmd" = start ] &&
+			cgroup2_set_limits
 		break
 	fi
 done
@@ -368,6 +369,10 @@ while [ -n "$1" ]; do
 				"$1" = "stop" ] && \
 				yesno "${rc_cgroup_cleanup}" && \
 					cgroup_cleanup
+				if [ "$(command -v cgroup2_remove)" = "cgroup2_remove" ]; then
+					[ "$1" = stop ] || [ -z "${command}" ] &&
+					cgroup2_remove
+				fi
 				shift
 				continue 2
 			else
diff --git a/sh/rc-cgroup.sh.in b/sh/rc-cgroup.sh.in
index 5987f966..40501f22 100644
--- a/sh/rc-cgroup.sh.in
+++ b/sh/rc-cgroup.sh.in
@@ -152,3 +152,54 @@ cgroup_cleanup()
 	kill -9 $pids
 	eend $(cgroup_running && echo 1 || echo 0) "fail to stop all processes"
 }
+
+cgroup2_find_path()
+{
+	case "${rc_cgroup_mode:-hybrid}" in
+		hybrid) printf "/sys/fs/cgroup/unified" ;;
+		unified) printf "/sys/fs/cgroup" ;;
+		esac
+		return 0
+}
+
+cgroup2_remove()
+{
+	local cgroup_path="$(cgroup2_find_path)" rc_cgroup_path
+	[ -z "${cgroup_path}" ] && return 0
+	rc_cgroup_path="${cgroup_path}/${RC_SVCNAME}"
+	[ ! -d "${rc_cgroup_path}" ] ||
+		[ ! -e "${rc_cgroup_path}"/cgroup.events ] &&
+		return 0
+	grep -qx "$$" "${rc_cgroup_path}/cgroup.procs" &&
+		echo 0 > "${cgroup_path}/cgroup.procs"
+	local key populated vvalue
+	while read key value; do
+		case "${key}" in
+			populated) populated=${value} ;;
+			*) ;;
+		esac
+	done < "${rc_cgroup_path}/cgroup.events"
+	[ "${populated}" = 1 ] && return 0
+	rmdir "${rc_cgroup_path}"
+	return 0
+}
+
+cgroup2_set_limits()
+{
+	local cgroup_path="$(cgroup2_find_path)"
+	[ -z "${cgroup_path}" ] && return 0
+	rc_cgroup_path="${cgroup_path}/${RC_SVCNAME}"
+	local OIFS="$IFS"
+	IFS="
+"
+	[ ! -d "${rc_cgroup_path}" ] && mkdir "${rc_cgroup_path}"
+	echo 0 > "${rc_cgroup_path}/cgroup.procs"
+	echo "${rc_cgroup_settings}" | while IFS="$OIFS" read key value; do
+		[ -z "${key}" ] || [ -z "${value}" ] && continue
+		[ ! -e "${rc_cgroup_path}/${key}" ] && continue
+		veinfo "${RC_SVCNAME}: cgroups: ${key} ${value}"
+		echo "${value}" > "${rc_cgroup_path}/${key}"
+	done
+	IFS="$OIFS"
+	return 0
+}