fix(nvmf): move connect logic to initqueue script

Rather than trying to connect immediately from parse-nvmf-boot-connections.sh,
move the connect logic to an initqueue script which is called at various
stages (settle, online, and timeout). In the timeout case, just try every
possible connection. Otherwise, use the existing priority logic.

This retry logic is the same that the iscsi module uses.

The nvme connect-nbft command will do "the right thing" when the connections
specified in the NBFT are already established: Already existing connections
will be skipped. Currently nvme-cli will print an error message and exit
with error status even if all targets are already connected. This should be
handled more gracefully on the nvme-cli side.

modules.d/95nvmf/nvmf-autoconnect.sh

@@ -1,5 +1,33 @@
-#!/bin/bash
+#!/bin/sh
+# Argument $1 is "settled", "online", or "timeout", indicating
+# the queue from which the script is called.
+# In the "timeout" case, try everything.
+# Otherwise, try options according to the priorities below.
 
-[ -f /sys/class/fc/fc_udev_device/nvme_discovery ] || exit 1
-echo add > /sys/class/fc/fc_udev_device/nvme_discovery
+[ "$RD_DEBUG" != yes ] || set -x
+
+NVMF_HOSTNQN_OK=
+[ ! -f "/etc/nvme/hostnqn" ] || [ ! -f "/etc/nvme/hostid" ] || NVMF_HOSTNQN_OK=1
+
+if [ -e /tmp/nvmf-fc-auto ] && [ "$NVMF_HOSTNQN_OK" ] \
+    && [ -f /sys/class/fc/fc_udev_device/nvme_discovery ]; then
+    # prio 1: cmdline override "rd.nvmf.discovery=fc,auto"
+    echo add > /sys/class/fc/fc_udev_device/nvme_discovery
+    [ "$1" = timeout ] || exit 0
+fi
+if [ -e /tmp/valid_nbft_entry_found ]; then
+    # prio 2: NBFT
+    /usr/sbin/nvme connect-nbft
+    [ "$1" = timeout ] || exit 0
+fi
+if [ -f /etc/nvme/discovery.conf ] && [ $NVMF_HOSTNQN_OK ]; then
+    # prio 3: discovery.conf from initrd
+    /usr/sbin/nvme connect-all
+    [ "$1" = timeout ] || exit 0
+fi
+if [ "$NVMF_HOSTNQN_OK" ] \
+    && [ -f /sys/class/fc/fc_udev_device/nvme_discovery ]; then
+    # prio 4: no discovery entries, try NVMeoFC autoconnect
+    echo add > /sys/class/fc/fc_udev_device/nvme_discovery
+fi
 exit 0

modules.d/95nvmf/parse-nvmf-boot-connections.sh

@@ -301,10 +301,10 @@ if [ -n "$nvmf_hostid" ]; then
     echo "$nvmf_hostid" > /etc/nvme/hostid
 fi
 
-NVMF_FC_AUTO=
+rm -f /tmp/nvmf-fc-auto
 for d in $(getargs rd.nvmf.discover -d nvmf.discover=); do
     parse_nvmf_discover "$d" || {
-        NVMF_FC_AUTO=1
+        : > /tmp/nvmf-fc-auto
         break
     }
 done
@@ -314,19 +314,6 @@ if [ -e /tmp/nvmf_needs_network ] || [ -e /tmp/valid_nbft_entry_found ]; then
     rm -f /tmp/nvmf_needs_network
 fi
 
-NVMF_HOSTNQN_OK=
-[ ! -f "/etc/nvme/hostnqn" ] || [ ! -f "/etc/nvme/hostid" ] || NVMF_HOSTNQN_OK=1
-
-if [ $NVMF_FC_AUTO ] && [ $NVMF_HOSTNQN_OK ]; then
-    # prio 1: cmdline override "rd.nvmf.discovery=fc,auto"
-    /sbin/initqueue --settled --onetime --unique --name nvme-fc-autoconnect /sbin/nvmf-autoconnect.sh
-elif [ -e /tmp/valid_nbft_entry_found ]; then
-    # prio 2: NBFT
-    /sbin/initqueue --settled --onetime --unique --name nvme-connect-nbft /usr/sbin/nvme connect-nbft
-elif [ -f /etc/nvme/discovery.conf ] && [ $NVMF_HOSTNQN_OK ]; then
-    # prio 3: discovery.conf from initrd
-    /sbin/initqueue --settled --onetime --unique --name nvme-discover /usr/sbin/nvme connect-all
-elif [ $NVMF_HOSTNQN_OK ]; then
-    # prio 4: no discovery entries, try NVMeoFC autoconnect
-    /sbin/initqueue --settled --onetime --unique --name nvme-fc-autoconnect /sbin/nvmf-autoconnect.sh
-fi
+/sbin/initqueue --online --name nvmf-connect-online /sbin/nvmf-autoconnect.sh online
+/sbin/initqueue --settled --onetime --name nvmf-connect-settled /sbin/nvmf-autoconnect.sh settled
+/sbin/initqueue --timeout --onetime --name nvmf-connect-timeout /sbin/nvmf-autoconnect.sh timeout