chreipl-fcp-mpath: if event subject is not re-IPL target, test if WWID matches

When the first test, whether the current event subject represents the
re-IPL target failed, we attempt a second test, using the recorded WWID
of a previously successful attempt.

The new helper `chreipl-fcp-mpath-is-ipl-vol` utilizes the information
previously stored in the ID-file (read under lock via `flock`).

The recorded WWID is compared with the one of the event subject; if it
matches we know that it addresses the same volume, and might be used as
alternative re-IPL target.

Additionally we also compare the current re-IPL triplet with the one
recorded in the ID-file. If it doesn't match anymore, we assume the
recorded WWID is stale - e.g., because the operator changed the re-IPL
target manually - and the previous check invalid.

If both the WWID match, and the re-IPL triplet is still the same as
previously recorded, the toolset may continue with the subject, and use
it as replacement for the current re-IPL target.

Reviewed-by: Steffen Maier <maier@linux.ibm.com>
Signed-off-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>

chreipl-fcp-mpath/.gitignore

@@ -1,5 +1,6 @@
 # build artifacts
 /chreipl-fcp-mpath-common.sh
 /chreipl-fcp-mpath-is-ipl-tgt
+/chreipl-fcp-mpath-is-ipl-vol
 /chreipl-fcp-mpath-is-reipl-zfcp
 /chreipl-fcp-mpath-record-volume-identifier

chreipl-fcp-mpath/Makefile

@@ -45,6 +45,7 @@ CHREIPL_FCP_MPATH_COMMON	:=					\
 	chreipl-fcp-mpath-common.sh
 CHREIPL_FCP_MPATH_UDEV_HELPER	:=					\
 	chreipl-fcp-mpath-is-ipl-tgt					\
+	chreipl-fcp-mpath-is-ipl-vol					\
 	chreipl-fcp-mpath-is-reipl-zfcp					\
 	chreipl-fcp-mpath-record-volume-identifier
 

chreipl-fcp-mpath/chreipl-fcp-mpath-common.sh.in

@@ -112,6 +112,37 @@ function id_file_lock_exclusive_create() {
 	return 0
 }
 
+# Output variables:
+#  id_file_unlock_shared_no_create() - call to unlock when finished with
+#                                      critical section
+#
+# XXX: `id_file_lock_*` can't be taken recursively
+function id_file_lock_shared_no_create() {
+	declare -g ID_FILE_LOCK=""
+
+	# Open the file defined in ${ID_FILE} for reading, and store the
+	# corresponding file descriptor in ${ID_FILE_LOCK}.
+	{ exec {ID_FILE_LOCK}<"${ID_FILE}"; } 2>/dev/null	|| return 1
+
+	declare -gf id_file_unlock_shared_no_create 1>/dev/null
+	function id_file_unlock_shared_no_create() {
+		if [ -v ID_FILE_LOCK ]; then
+			# release file and implicitly the lock, if taken
+			exec {ID_FILE_LOCK}<&-
+			unset ID_FILE_LOCK
+		fi
+
+		unset "TRAP_EXIT_FN[id_file_unlock_shared_no_create]"
+	}
+	TRAP_EXIT_FN+=(
+		[id_file_unlock_shared_no_create]=id_file_unlock_shared_no_create
+	)
+
+	flock --shared --timeout 5 "${ID_FILE_LOCK}"		|| return 2
+
+	return 0
+}
+
 # Output variables:
 #  firmware_unlock_shared() - call to unlock when finished with critical section
 #

chreipl-fcp-mpath/chreipl-fcp-mpath-is-ipl-vol.in

@@ -0,0 +1,86 @@
+#!/bin/bash
+# SPDX-License-Identifier: MIT
+#
+# chreipl-fcp-mpath: use multipath information to change FCP IPL target
+# (C) Copyright IBM Corp. 2021
+#
+# Uses the following system-utilities (and shell-builtins):
+#   Those necessary for sourced library:
+#     - chreipl-fcp-mpath-common.sh
+
+# Find out whether the device in environment variable ${DEVPATH} represents the
+# _volume_ that we IPL'ed from. We do this by comparing its WWID to the one
+# recorded in `@chreiplzfcpmp-id-file@`.
+#
+# Makes use of udev event environment variables:
+#	DM_UUID
+#	SUBSYSTEM
+#	DEVPATH
+
+# shellcheck disable=SC2034
+declare -gr debug_trace_tag=11iilv
+# shellcheck disable=SC1091
+source '@chreiplzfcpmp-lib@' || exit 127
+
+function id_file_read_ipl_information() {
+	local -a records
+
+	declare -g REC_WWID="" REC_BUSID="" REC_WWPN="" REC_LUN=""
+
+	# lock file before reading ID, so we don't see any intermediate state
+	id_file_lock_shared_no_create				|| return 1
+
+	{ readarray -d "" -t -u "${ID_FILE_LOCK}" records; } 2>/dev/null \
+								|| return 2
+	if '@DEBUG@'; then declare -p records 1>&2; fi
+
+	id_file_unlock_shared_no_create
+
+	[ "${#records[@]}" = "4" ]				|| return 3
+	# check that none of the array fields contains whitespace only
+	[ "${records[0]/#*([[:space:]])}" != "" ]		|| return 4
+	[ "${records[1]/#*([[:space:]])}" != "" ]		|| return 5
+	[ "${records[2]/#*([[:space:]])}" != "" ]		|| return 6
+	[ "${records[3]/#*([[:space:]])}" != "" ]		|| return 7
+
+	REC_WWID="${records[0]}"
+	REC_BUSID="${records[1]}"
+	REC_WWPN="${records[2]}"
+	REC_LUN="${records[3]}"
+	return 0
+}
+
+id_file_read_ipl_information					|| exit 1
+
+if [[ "${DM_UUID}" == mpath-* ]]; then
+	# Assume Multipath Device Mapper Device;
+	# e.g.: DEVPATH = /devices/virtual/block/dm-0
+	declare sdev found=false
+
+	for sdev in /sys/"${DEVPATH}"/slaves/sd*/device; do
+		if sdev_get_wwid "${sdev}"; then
+			found=true
+			break
+		fi
+	done
+	unset sdev
+
+	"${found}"						|| exit 2
+
+elif [ "${SUBSYSTEM}" = block ]; then
+	# Assume SCSI Disk;
+	# e.g.: DEVPATH = /devices/css0/0.0.0014/0.0.1700/host0/rport-0:0-0/target0:0:0/0:0:0:1074806808/block/sds
+
+	sdev_get_wwid /sys/"${DEVPATH}"/device			|| exit 3
+fi
+
+# set by `sdev_get_wwid` and `id_file_read_ipl_information`
+[ "${SDEV_WWID}" = "${REC_WWID}" ]				|| exit 4
+
+firmware_get_ipl_information					|| exit 5
+# set by `firmware_get_ipl_information` and `id_file_read_ipl_information`
+[ "${IPL_BUSID}" = "${REC_BUSID}" ]				|| exit 6
+[ "${IPL_WWPN}" = "${REC_WWPN}" ]				|| exit 7
+[ "${IPL_LUN}" = "${REC_LUN}" ]					|| exit 8
+
+exit 0

chreipl-fcp-mpath/udev/rules.d/70-chreipl-fcp-mpath.rules

@@ -27,7 +27,9 @@ PROGRAM!="chreipl-fcp-mpath-is-reipl-zfcp", GOTO="chreipl_fcp_mpath_end"
 # Either:
 #
 # (A) We recognized a new SCSI Disk. This might represent:
-#     (a) the path we want to ReIPL from.
+#     (a) the path we want to ReIPL from;
+#     (b) an alternative path to the volume we want to ReIPL from;
+#     (c) a path to some unrelated volume.
 #
 # Or:
 #
@@ -37,7 +39,8 @@ PROGRAM!="chreipl-fcp-mpath-is-reipl-zfcp", GOTO="chreipl_fcp_mpath_end"
 #         (b) came back online;
 #     an alternative path to the volume we want to ReIPL from:
 #         (c) went away;
-#         (d) came back online.
+#         (d) came back online;
+#     (e) some unrelated multipath device saw an event.
 
 # Test whether the affected device is, or contains, the current IPL target.
 #
@@ -68,6 +71,27 @@ GOTO="chreipl_fcp_mpath_try_change_ipl_path"
 # ReIPL target, and not a mpath device that contains the current ReIPL target)
 LABEL="chreipl_fcp_mpath_not_direct_match"
 
+# While this sdev/mpath device doesn't directly correspond to the path
+# currently set as ReIPL target, it might still point to the same volume.
+#
+# For mpath devices this can happen if the original ReIPL target is completely
+# gone from the machine, and so there is no way we can successfully, directly
+# compare the ReIPL parameters to the sdevs of the mpath device.
+#
+# For cases like these we recorded the volume identifier, which we now can
+# compare, and so still decide whether we are addressing the correct volume.
+#
+# This covers scenarios:
+#     (A) (b)/(c),
+#     (B) (c)/(d)/(e)
+#
+# XXX: we recorded WWID, Device-Bus-ID, Remote WWPN, LUN of the ReIPL target at
+#      the time; if the latter three don't match the current ReIPL setting
+#      anymore, we have to assume that someone changed the ReIPL target
+#      manually, and we cannot use the WWID anymore since we can't possibly
+#      know whether that stayed the same when the change was done.
+PROGRAM!="chreipl-fcp-mpath-is-ipl-vol", GOTO="chreipl_fcp_mpath_end"
+
 LABEL="chreipl_fcp_mpath_try_change_ipl_path"
 
 LABEL="chreipl_fcp_mpath_end"