SUSE patches which should show up mainline

SUSE patches which should show up mainline

[Thomas Renninger]

Hi,

in the past weeks/months SUSE somewhat diverged from mainline development
due to SLE12 work.

This was/is not intended and here are general fixes/enhancements which should
show up mainline.

I merged quite some (on top fixes/enhancements) together to avoid to submit
half baken or broken patches.
I also put iscsi, multipath (forgot fcoe ones, but that should not matter
right now) patches together for easier review.

S390 specific patches were left out. There have been some specialities and
dracut would be the wrong place to start consolidating things.

Be carefule with patches:
[PATCH 08/33] 90multipath: Install libgcc_s library
[PATCH 12/33] 95iscsi: Install libgcc_s library
these might already be fixed, compare with mail:
Multithreaded programs crashing due to libgcc_s missing
from 06/06/14 and 07/22/14

It may make sense to first do a version jump to 041 before merging these?

Would be nice to see as much of these pushed already as possible.
I am willing to re-submit remaining patches which are questionable or need
re-work after discussing.

Thanks a lot in advance,

       Thomas

[PATCH 01/33] mkinitrd-suse: Update to sles12 SUSE state

[Thomas Renninger]

Incorporates these patches:
- 0013 mkinitrd-suse: do not update bootloader if no kernel was found
  bnc#858268
- 0018 mkinitrd-suse: Fix whitespaces when adding drivers
- 0032 mkinitrd-suse.sh: Use '--hostonly' and '--hostonly-cmdline' correctly
  bnc#874000,bnc#874905,bnc#874363
- 0039 Remove --force parameter from mkinitrd
- 0041 mkinitd-suse: remove --hostonly and --hostonly-cmdline
  Should be set via configuration files, not in the script itself.
- 0051 mkinitrd-suse: add 'update bootloader' message
- 0120 mkinitrd-suse.sh: Bail out with exit 1 if initrd cannot be generated
  bnc#886630

Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 mkinitrd-suse.8.asc |  7 +++---
 mkinitrd-suse.sh    | 66 +++++++++++++++++++++++++----------------------------
 2 files changed, 34 insertions(+), 39 deletions(-)

diff --git a/mkinitrd-suse.8.asc b/mkinitrd-suse.8.asc
index 9eb8fc9..cdb8e3a 100644
--- a/mkinitrd-suse.8.asc
+++ b/mkinitrd-suse.8.asc
@@ -76,7 +76,9 @@ OPTIONS
 
 **-A**::
     Create a so called "monster initrd" which includes all available
-    features and modules.
+    features and modules. This calls dracut with --no-hostonly and
+    --no-hostonly-cmdline parameters internally, instead of the default
+    --hostonly and --hostonly-cmdline.
 
 **-v, --verbose**::
     increase verbosity level
@@ -85,9 +87,6 @@ OPTIONS
     Disable logging to _/var/log/YaST2/mkinitrd.log_. This is useful for
     testing if you don’t want to clutter the system log.
 
-**--force**::
-    overwrite existing initramfs file.
-
 **--help**::
     print a help message and exit.
 
diff --git a/mkinitrd-suse.sh b/mkinitrd-suse.sh
index 23dc262..af42cbe 100755
--- a/mkinitrd-suse.sh
+++ b/mkinitrd-suse.sh
@@ -20,8 +20,6 @@
 
 boot_dir="/boot"
 quiet=0
-host_only=1
-force=0
 logfile=/var/log/YaST2/mkinitrd.log
 dracut_cmd=dracut
 
@@ -68,10 +66,7 @@ usage () {
     $cmd "	-j device		Journal device"
     $cmd "	-D interface		Run dhcp on the specified interface."
     $cmd "	-I interface		Configure the specified interface statically."
-    $cmd "	-a acpi_dsdt		Attach compiled ACPI DSDT (Differentiated"
-    $cmd "				System Description Table) to initrd. This"
-    $cmd "				replaces the DSDT of the BIOS. Defaults to"
-    $cmd "				the ACPI_DSDT variable in /etc/sysconfig/kernel."
+    $cmd "	-a acpi_dsdt		Obsolete, do not use."
     $cmd "	-s size			Add splash animation and bootscreen to initrd."
 
     [[ $1 = '-n' ]] && exit 0
@@ -204,8 +199,6 @@ default_kernel_images() {
     for initrd_image in $initrd_images;do
 	targets="$targets $initrd_image"
     done
-    host_only=1
-    force=1
 }
 
 while (($# > 0)); do
@@ -218,8 +211,6 @@ while (($# > 0)); do
 	    for kernel_image in $kernel_images;do
 		kernels="$kernels ${kernel_image#*-}"
 	    done
-	    host_only=1
-	    force=1
 	    ;;
 	-i) read_arg initrd_images "$@" || shift $?
 	    for initrd_image in $initrd_images;do
@@ -261,7 +252,7 @@ while (($# > 0)); do
 	    dracut_cmdline="${dracut_cmdline} ip=$(ipconfig $static_if)":
 	    ;;
 	-a) read_arg acpi_dsdt "$@" || shift $?
-	    echo "mkinitrd: custom DSDT not yet supported"
+	    echo "Obsolete -a param, use acpi_table_dir= and acpi_override= variables in /etc/dracut.conf.d/"
 	    exit 1
 	    ;;
 	-s) read_arg boot_splash "$@" || shift $?
@@ -275,7 +266,6 @@ while (($# > 0)); do
         --version|-R)
             echo "mkinitrd: dracut compatibility wrapper"
             exit 0;;
-        --force) force=1;;
 	--quiet|-q) quiet=1;;
         *)  if [[ ! $targets ]]; then
             targets=$1
@@ -289,15 +279,17 @@ while (($# > 0)); do
 done
 
 [[ $targets && $kernels ]] || default_kernel_images
-[[ $targets && $kernels ]] || (error "No kernel found in $boot_dir" && usage)
+if [[ ! $targets || ! $kernels ]];then
+    error "No kernel found in $boot_dir"
+fi
 
 # We can have several targets/kernels, transform the list to an array
 targets=( $targets )
 [[ $kernels ]] && kernels=( $kernels )
 
 [[ $logfile ]]        && dracut_args="${dracut_args} --logfile $logfile"
-[[ $host_only == 1 ]] && dracut_args="${dracut_args} --hostonly"
-[[ $force == 1 ]]     && dracut_args="${dracut_args} --force"
+dracut_args="${dracut_args} --force"
+
 [[ $dracut_cmdline ]] && dracut_args="${dracut_args} --kernel-cmdline ${dracut_cmdline}"
 [ -z "$(type -p update-bootloader)" ] && skip_update_bootloader=1
 
@@ -306,11 +298,9 @@ if [ -f /etc/sysconfig/kernel ] ; then
     . /etc/sysconfig/kernel
 fi
 [[ $module_list ]] || module_list="${INITRD_MODULES}"
-basicmodules="$basicmodules ${module_list}"
 [[ $domu_module_list ]] || domu_module_list="${DOMU_INITRD_MODULES}"
-[[ $acpi_dsdt ]] || acpi_dsdt="${ACPI_DSDT}"
+shopt -s extglob
 
-echo "Creating: target|kernel|dracut args|basicmodules "
 for ((i=0 ; $i<${#targets[@]} ; i++)); do
 
     if [[ $img_vers ]];then
@@ -320,28 +310,34 @@ for ((i=0 ; $i<${#targets[@]} ; i++)); do
     fi
     kernel="${kernels[$i]}"
 
+    if is_xen_kernel $kernel $rootfs ; then
+	modules_all="${module_list} ${domu_module_list}"
+    else
+        modules_all="${module_list}"
+    fi
+
+    # Remove leading and trailing spaces needs (set above): shopt -s extglob
+    modules_all=${modules_all%%+([[:space:]])}
+    modules_all=${modules_all##+([[:space:]])}
+
+    echo "Creating initrd: $target"
+
     # Duplicate code: No way found how to redirect output based on $quiet
     if [[ $quiet == 1 ]];then
-	echo "$target|$kernel|$dracut_args|$basicmodules"
-	if is_xen_kernel $kernel $rootfs ; then
-	    basicmodules="$basicmodules ${domu_module_list}"
-	fi
-	if [[ $basicmodules ]]; then
-            $dracut_cmd $dracut_args --add-drivers "$basicmodules" "$target" \
-		"$kernel" &>/dev/null
-	else
+        # Duplicate code: --force-drivers must not be called with empty string
+        # -> dracut bug workarounded ugly, because of complex whitespace
+        # expansion magics
+        if [ -n "${modules_all}" ];then
+            $dracut_cmd $dracut_args --force-drivers "${modules_all}" "$target" "$kernel" &>/dev/null
+        else
             $dracut_cmd $dracut_args "$target" "$kernel" &>/dev/null
-	fi
+        fi
     else
-	if is_xen_kernel $kernel $rootfs ; then
-	    basicmodules="$basicmodules ${domu_module_list}"
-	fi
-	if [[ $basicmodules ]]; then
-            $dracut_cmd $dracut_args --add-drivers "$basicmodules" "$target" \
-		"$kernel"
-	else
+        if [ -n "${modules_all}" ];then
+            $dracut_cmd $dracut_args --force-drivers "${modules_all}" "$target" "$kernel"
+        else
             $dracut_cmd $dracut_args "$target" "$kernel"
-	fi
+        fi
     fi
 done
 
-- 
1.8.5.2

[PATCH 02/33] Update dracut manpage for GRUB Legacy and GRUB2

[Thomas Renninger]

From: Shawn Dunn <sfalken-stAJ6ESoqRxg9hUCZPvPmw@public.gmane.org>

Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 dracut.usage.asc | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/dracut.usage.asc b/dracut.usage.asc
index 9e788c1..1a79638 100644
--- a/dracut.usage.asc
+++ b/dracut.usage.asc
@@ -99,9 +99,9 @@ raid with encryption and LVM on top), as long as you specify the correct
 filesystem LABEL or UUID on the kernel command line for your root device, dracut
 will find it and boot from it.
 
-The kernel command line usually can be configured in _/boot/grub/grub.conf_ or
-_/boot/grub2/grub.cfg_, if grub is your bootloader and it also can be edited in
-the real boot process in the grub menu.
+The kernel command line usually can be configured in _/boot/grub/grub.conf_,
+for GRUB legacy, or _/boot/grub2/grub.cfg_, for GRUB2 if grub is your
+bootloader and it also can be edited in the real boot process in the grub menu.
 
 The kernel command line can also be provided by the dhcp server with the
 root-path option. See <<NetworkBoot>>.
-- 
1.8.5.2

[PATCH 03/33] 00warpclock: Set correct timezone

[Thomas Renninger]

From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>

Add module for setting correct timezone.
References: bnc#830060

Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/00warpclock/module-setup.sh | 25 +++++++++++++++++++++++++
 modules.d/00warpclock/warpclock.sh    |  9 +++++++++
 2 files changed, 34 insertions(+)
 create mode 100755 modules.d/00warpclock/module-setup.sh
 create mode 100755 modules.d/00warpclock/warpclock.sh

diff --git a/modules.d/00warpclock/module-setup.sh b/modules.d/00warpclock/module-setup.sh
new file mode 100755
index 0000000..6bd7abd
--- /dev/null
+++ b/modules.d/00warpclock/module-setup.sh
@@ -0,0 +1,25 @@
+#!/bin/bash
+
+# called by dracut
+check() {
+    # hwclock does not exist on S390(x), bail out silently then
+    local _arch=$(uname -m)
+    [ "$_arch" = "s390" -o "$_arch" = "s390x" ] && return 1
+
+    [ -e /etc/localtime -a -e /etc/adjtime ] || return 1
+    require_binaries /sbin/hwclock || return 1
+}
+
+# called by dracut
+depends() {
+    return 0
+}
+
+# called by dracut
+install() {
+    inst /usr/share/zoneinfo/UTC
+    inst /etc/localtime
+    inst /etc/adjtime
+    inst_hook pre-trigger 00 "$moddir/warpclock.sh"
+    inst /sbin/hwclock
+}
diff --git a/modules.d/00warpclock/warpclock.sh b/modules.d/00warpclock/warpclock.sh
new file mode 100755
index 0000000..f64818c
--- /dev/null
+++ b/modules.d/00warpclock/warpclock.sh
@@ -0,0 +1,9 @@
+#!/bin/sh
+
+if test -e /etc/adjtime ; then
+    while read line ; do
+	if test "$line" = LOCAL ; then
+	    hwclock --systz
+	fi
+    done < /etc/adjtime
+fi
-- 
1.8.5.2

[PATCH 04/33] systemd always tries to load autofs4

[Thomas Renninger]

During boot systemd tries to load autofs4, but dracut will only
add it if it's loaded by the time dracut is run.
Modify dracut to always load autofs4.

References: bnc#869411

Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/98systemd/module-setup.sh | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/modules.d/98systemd/module-setup.sh b/modules.d/98systemd/module-setup.sh
index bd4433e..045bf8b 100755
--- a/modules.d/98systemd/module-setup.sh
+++ b/modules.d/98systemd/module-setup.sh
@@ -18,7 +18,7 @@ depends() {
 }
 
 installkernel() {
-    instmods autofs4 ipv6
+    hostonly='' instmods autofs4 ipv6
     instmods -s efivarfs
 }
 
-- 
1.8.5.2

[PATCH 05/33] Fixup mdraid setup

[Thomas Renninger]

From: NeilBrown <neilb-l3A5Bk7waGM@public.gmane.org>

This sed script to edit 64-md-raid-assemble.rules needs to
be adjusted for latest mdadm

References: bnc#866660

Signed-off-by: NeilBrown <neilb-l3A5Bk7waGM@public.gmane.org>
Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/90mdraid/module-setup.sh | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/modules.d/90mdraid/module-setup.sh b/modules.d/90mdraid/module-setup.sh
index bc4c514..ce3394a 100755
--- a/modules.d/90mdraid/module-setup.sh
+++ b/modules.d/90mdraid/module-setup.sh
@@ -86,7 +86,7 @@ install() {
     for rule in 64-md-raid.rules 64-md-raid-assembly.rules; do
         rule_path="${initdir}${udevdir}/rules.d/${rule}"
         [ -f "${rule_path}" ] && sed -i -r \
-            -e '/RUN\+?="[[:alpha:]/]*mdadm[[:blank:]]+(--incremental|-I)[[:blank:]]+(\$env\{DEVNAME\}|\$tempnode|\$devnode)/d' \
+            -e '/(RUN|IMPORT\{program\})\+?="[[:alpha:]/]*mdadm[[:blank:]]+(--incremental|-I)[[:blank:]]+(--export )?(\$env\{DEVNAME\}|\$tempnode|\$devnode)/d' \
             "${rule_path}"
     done
 
-- 
1.8.5.2

[PATCH 06/33] 95udev-rules: Include correct sg3_utils rules

[Thomas Renninger]

From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>

sg3_utils now provides two rules 55-scsi-sg3_id.rules and
58-scsi-sg3_symlink.rules, which need to be included instead
of the older 59-scsi-sg3_utils.rules.

References: bnc#873151

Signed-off-by: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/95udev-rules/module-setup.sh | 1 +
 1 file changed, 1 insertion(+)

diff --git a/modules.d/95udev-rules/module-setup.sh b/modules.d/95udev-rules/module-setup.sh
index 9d6f960..e50709d 100755
--- a/modules.d/95udev-rules/module-setup.sh
+++ b/modules.d/95udev-rules/module-setup.sh
@@ -31,6 +31,7 @@ install() {
         50-udev.rules 95-late.rules \
         50-firmware.rules \
         59-scsi-sg3_utils.rules \
+        55-scsi-sg3_id.rules 58-scsi-sg3_symlink.rules \
         70-uaccess.rules 71-seat.rules 73-seat-late.rules \
         75-net-description.rules \
         80-net-name-slot.rules 80-net-setup-link.rules \
-- 
1.8.5.2

[PATCH 07/33] 90multipath: install correct multipath rules

[Thomas Renninger]

From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>

The multipath and kpartx rules have different numbers in SUSE.

The 11-dm-mpath.rules file had been missing, causing
blacklisting to not work properly.

References: bnc#873151, bnc#872662, bnc#883149

Signed-off-by: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/90multipath/module-setup.sh | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/modules.d/90multipath/module-setup.sh b/modules.d/90multipath/module-setup.sh
index c86c5d9..3c2e3a9 100755
--- a/modules.d/90multipath/module-setup.sh
+++ b/modules.d/90multipath/module-setup.sh
@@ -100,6 +100,9 @@ install() {
 
     inst_hook cleanup   80 "$moddir/multipathd-needshutdown.sh"
 
-    inst_rules 40-multipath.rules 62-multipath.rules 65-multipath.rules 66-kpartx.rules
+    inst_rules 40-multipath.rules 56-multipath.rules \
+	62-multipath.rules 65-multipath.rules \
+	66-kpartx.rules 67-kpartx-compat.rules \
+	11-dm-mpath.rules
 }
 
-- 
1.8.5.2

[PATCH 08/33] 90multipath: Install libgcc_s library

[Thomas Renninger]

From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>

multipathd is using pthreads, which require libgcc_s for
pthread_cancel to work. Without it multipathd might crash
with SIGABRT.

Signed-off-by: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/90multipath/module-setup.sh | 1 +
 1 file changed, 1 insertion(+)

diff --git a/modules.d/90multipath/module-setup.sh b/modules.d/90multipath/module-setup.sh
index 3c2e3a9..7a60b2e 100755
--- a/modules.d/90multipath/module-setup.sh
+++ b/modules.d/90multipath/module-setup.sh
@@ -87,6 +87,7 @@ install() {
     inst $(command -v partx) /sbin/partx
 
     inst_libdir_file "libmultipath*" "multipath/*"
+    inst_libdir_file 'libgcc_s.so*'
 
     if dracut_module_included "systemd"; then
         inst_multiple \
-- 
1.8.5.2

[PATCH 09/33] 90multipath: Load device_handler modules early during boot

[Thomas Renninger]

From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>

The device handler modules need to be loaded early during boot
to avoid I/O errors being printed to the system log.

References: bnc#871617

Signed-off-by: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/90multipath/module-setup.sh | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/modules.d/90multipath/module-setup.sh b/modules.d/90multipath/module-setup.sh
index 7a60b2e..bbee5d9 100755
--- a/modules.d/90multipath/module-setup.sh
+++ b/modules.d/90multipath/module-setup.sh
@@ -28,6 +28,15 @@ depends() {
 }
 
 # called by dracut
+cmdline() {
+    for m in scsi_dh_alua scsi_dh_emc scsi_dh_rdac ; do
+        if module_is_host_only $m ; then
+            printf 'rd.driver.pre=%s ' "$m"
+        fi
+    done
+}
+
+# called by dracut
 installkernel() {
     local _ret
     local _arch=$(uname -m)
@@ -89,6 +98,11 @@ install() {
     inst_libdir_file "libmultipath*" "multipath/*"
     inst_libdir_file 'libgcc_s.so*'
 
+    if [[ $hostonly_cmdline ]] ; then
+        local _conf=$(cmdline)
+        [[ $_conf ]] && echo "$_conf" >> "${initdir}/etc/cmdline.d/90multipath.conf"
+    fi
+
     if dracut_module_included "systemd"; then
         inst_multiple \
             $systemdsystemunitdir/multipathd.service
-- 
1.8.5.2

[PATCH 10/33] 90multipath: install dracut-specific service file

[Thomas Renninger]

From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>

The multipathd package might install a service and a socket
file. Using the original service file from the installed
system without the socket file triggers a bug in systemd,
causing systemd to crash.

As we don't actually need to socket file in the initrd we
should be installing our own service file which does not
reference the socket file at all.

References: bnc#871610

Signed-off-by: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/90multipath/module-setup.sh    |  7 +++----
 modules.d/90multipath/multipathd.service | 15 +++++++++++++++
 2 files changed, 18 insertions(+), 4 deletions(-)
 create mode 100644 modules.d/90multipath/multipathd.service

diff --git a/modules.d/90multipath/module-setup.sh b/modules.d/90multipath/module-setup.sh
index bbee5d9..0f88f82 100755
--- a/modules.d/90multipath/module-setup.sh
+++ b/modules.d/90multipath/module-setup.sh
@@ -104,10 +104,9 @@ install() {
     fi
 
     if dracut_module_included "systemd"; then
-        inst_multiple \
-            $systemdsystemunitdir/multipathd.service
-        mkdir -p "${initdir}${systemdsystemconfdir}/sysinit.target.wants"
-        ln -rfs "${initdir}${systemdsystemunitdir}/multipathd.service" "${initdir}${systemdsystemconfdir}/sysinit.target.wants/multipathd.service"
+        inst_simple "${moddir}/multipathd.service" "${systemdsystemunitdir}/multipathd.service"
+        mkdir -p "${initdir}${systemdsystemunitdir}/sysinit.target.wants"
+        ln -rfs "${initdir}${systemdsystemunitdir}/multipathd.service" "${initdir}${systemdsystemunitdir}/sysinit.target.wants/multipathd.service"
     else
         inst_hook pre-trigger 02 "$moddir/multipathd.sh"
         inst_hook cleanup   02 "$moddir/multipathd-stop.sh"
diff --git a/modules.d/90multipath/multipathd.service b/modules.d/90multipath/multipathd.service
new file mode 100644
index 0000000..44a7b80
--- /dev/null
+++ b/modules.d/90multipath/multipathd.service
@@ -0,0 +1,15 @@
+[Unit]
+Description=Device-Mapper Multipath Device Controller
+Before=iscsi.service iscsid.service lvm2-activation-early.service
+DefaultDependencies=no
+Conflicts=shutdown.target
+
+[Service]
+Type=simple
+ExecStartPre=/sbin/modprobe dm-multipath
+ExecStart=/sbin/multipathd -s -d
+ExecReload=/sbin/multipathd reconfigure
+ExecStop=/sbin/multipathd shutdown
+
+[Install]
+WantedBy=sysinit.target
-- 
1.8.5.2

[PATCH 11/33] 95iscsi: Autodetect iSCSI firmware

[Thomas Renninger]

From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>

Some iSCSI initiator present firmware information in
/sys/firmware/ibft or /sys/firmware/iscsi_bootX

Whenever we detect one of those directories we should assume
that the iSCSI devices should be activated.

Also incorporates SUSE patches:
0049:
95iscsi: Use 'ip=ibft' for ibft autoconfiguration

For iBFT autoconfiguration we should be setting 'ip=ibft'
instead of rd.neednet. This should instruct dracut to only
enable the iBFT interfaces and leave the rest alone.

References: bnc#879038

0054:
95iscsi: update commandline printing

dracut has a separate callout 'cmdline' which should be used
for printing out the generated commandline.


Signed-off-by: Pavel Wieczorkiewicz <pwieczorkiewicz-IBi9RG/b67k@public.gmane.org>
Signed-off-by: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/95iscsi/module-setup.sh | 26 ++++++++++++++++++++++++++
 1 file changed, 26 insertions(+)

diff --git a/modules.d/95iscsi/module-setup.sh b/modules.d/95iscsi/module-setup.sh
index 8b42d3e..5d2a39b 100755
--- a/modules.d/95iscsi/module-setup.sh
+++ b/modules.d/95iscsi/module-setup.sh
@@ -28,6 +28,21 @@ check() {
     return 0
 }
 
+install_ibft() {
+    # When iBFT / iscsi_boot is detected:
+    # - Use 'ip=ibft' to set up iBFT network interface
+    # - specify firmware booting cmdline parameter
+
+    for d in /sys/firmware/* ; do
+        if [ -d ${d}/initiator ] ; then
+            if [ ${d##*/} = "ibft" ] ; then
+                echo -n "ip=ibft "
+            fi
+            echo -n "rd.iscsi.firmware=1"
+        fi
+    done
+}
+
 # called by dracut
 depends() {
     echo network rootfs-block
@@ -37,6 +52,12 @@ depends() {
 installkernel() {
     local _arch=$(uname -m)
 
+    # Detect iBFT and perform mandatory steps
+    if [[ $hostonly_cmdline == "yes" ]] ; then
+        install_ibft > "${initdir}/etc/cmdline.d/95iscsi.conf"
+        echo >> "${initdir}/etc/cmdline.d/95iscsi.conf"
+    fi
+
     instmods bnx2i qla4xxx cxgb3i cxgb4i be2iscsi
     hostonly="" instmods iscsi_tcp iscsi_ibft crc32c iscsi_boot_sysfs
     iscsi_module_filter() {
@@ -77,6 +98,11 @@ installkernel() {
 }
 
 # called by dracut
+cmdline() {
+    install_ibft
+}
+
+# called by dracut
 install() {
     inst_multiple umount iscsistart hostname iscsi-iname
     inst_multiple -o iscsiuio
-- 
1.8.5.2

[PATCH 12/33] 95iscsi: Install libgcc_s library

[Thomas Renninger]

From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>

iscsiuio is using pthread, which requires libgcc_s for
pthread_cancel to work.
Without this library iscsiuio will crash with SIGABRT.

References: bnc#881692

Signed-off-by: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/95iscsi/module-setup.sh | 1 +
 1 file changed, 1 insertion(+)

diff --git a/modules.d/95iscsi/module-setup.sh b/modules.d/95iscsi/module-setup.sh
index 5d2a39b..3641fb8 100755
--- a/modules.d/95iscsi/module-setup.sh
+++ b/modules.d/95iscsi/module-setup.sh
@@ -106,6 +106,7 @@ cmdline() {
 install() {
     inst_multiple umount iscsistart hostname iscsi-iname
     inst_multiple -o iscsiuio
+    inst_libdir_file 'libgcc_s.so*'
     inst_hook cmdline 90 "$moddir/parse-iscsiroot.sh"
     inst_hook cleanup 90 "$moddir/cleanup-iscsi.sh"
     inst "$moddir/iscsiroot.sh" "/sbin/iscsiroot"
-- 
1.8.5.2

[PATCH 13/33] 95iscsi: Set correct iscsi_started value for iSCSI firmware

[Thomas Renninger]

From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>

When iSCSI firmware booting is selected we should not rely on
'netroot' or 'iscsiroot' variables to be set.

References: bnc#873448

Signed-off-by: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/95iscsi/parse-iscsiroot.sh | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/modules.d/95iscsi/parse-iscsiroot.sh b/modules.d/95iscsi/parse-iscsiroot.sh
index ca57668..85c3027 100755
--- a/modules.d/95iscsi/parse-iscsiroot.sh
+++ b/modules.d/95iscsi/parse-iscsiroot.sh
@@ -90,8 +90,12 @@ if [ -n "$netroot" ] && [ "$root" != "/dev/root" ] && [ "$root" != "dhcp" ]; the
     fi
 fi
 
-netroot_enc=$(str_replace "$netroot" '/' '\2f')
-echo "[ -f '/tmp/iscsistarted-$netroot_enc' ]" > $hookdir/initqueue/finished/iscsi_started.sh
+if [ -n "$iscsi_firmware" ] ; then
+    echo "[ -f '/tmp/iscsistarted-firmware' ]" > $hookdir/initqueue/finished/iscsi_started.sh
+else
+    netroot_enc=$(str_replace "$netroot" '/' '\2f')
+    echo "[ -f '/tmp/iscsistarted-$netroot_enc' ]" > $hookdir/initqueue/finished/iscsi_started.sh
+fi
 
 # Done, all good!
 rootok=1
-- 
1.8.5.2

[PATCH 14/33] 95iscsi: strip one set of quotes when calling initqueue

[Thomas Renninger]

From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>

The bad initqueue parameter passing is already addressed by git commit:
commit bb8c16d12178373f17553294603ff5187e318b51
and
commit 486a8f33e2edaf249e3a09c98a910094e201cf9e

However this appraoch is slightly different:
"$netroot" instead of "'$netroot'", etc.

initqueue will process the arguments, assuming the first non-option
argument to be the program to be executed. Putting the entire
commandline into this argument confuses both, initqueue and the
program in question.

References: bnc#879038

Signed-off-by: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/95iscsi/parse-iscsiroot.sh | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/modules.d/95iscsi/parse-iscsiroot.sh b/modules.d/95iscsi/parse-iscsiroot.sh
index 85c3027..37e46ad 100755
--- a/modules.d/95iscsi/parse-iscsiroot.sh
+++ b/modules.d/95iscsi/parse-iscsiroot.sh
@@ -59,10 +59,10 @@ fi
 
 # iscsi_firmware does not need argument checking
 if [ -n "$iscsi_firmware" ] ; then
-    netroot=${netroot:-iscsi:}
+    [ -z "$netroot" ] && netroot=iscsi:
     modprobe -q iscsi_boot_sysfs 2>/dev/null
     modprobe -q iscsi_ibft
-    initqueue --onetime --timeout /sbin/iscsiroot dummy "'$netroot'" "'$NEWROOT'"
+    initqueue --onetime --timeout /sbin/iscsiroot dummy "$netroot" "$NEWROOT"
 fi
 
 # If it's not iscsi we don't continue
@@ -86,7 +86,7 @@ fi
 
 if [ -n "$netroot" ] && [ "$root" != "/dev/root" ] && [ "$root" != "dhcp" ]; then
     if ! getargbool 1 rd.neednet >/dev/null || ! getarg "ip="; then
-        initqueue --onetime --settled /sbin/iscsiroot dummy "'$netroot'" "'$NEWROOT'"
+        initqueue --onetime --settled /sbin/iscsiroot dummy "$netroot" "$NEWROOT"
     fi
 fi
 
-- 
1.8.5.2

[PATCH 15/33] 95iscsi: Fixup bnx2i offload booting

[Thomas Renninger]

From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>

bnx2i is using a separate iSCSI offload engine with a separate
MAC address. As a result, the iBFT information is displaying
a MAC address which does not relate to any MAC address from
the network interfaces.
In addition, the iSCSI offload engine works independently on
the NIC, so we do not need to enable the NIC for iSCSI offload
to work.
This patch modifies the automatic iBFT detection to not set
the 'ip=ibft' flag when bnx2i offload is detected.

References: bnc#855747

Signed-off-by: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/95iscsi/module-setup.sh | 26 +++++++++++++++++++++++++-
 1 file changed, 25 insertions(+), 1 deletion(-)

diff --git a/modules.d/95iscsi/module-setup.sh b/modules.d/95iscsi/module-setup.sh
index 3641fb8..d30e64f 100755
--- a/modules.d/95iscsi/module-setup.sh
+++ b/modules.d/95iscsi/module-setup.sh
@@ -28,14 +28,38 @@ check() {
     return 0
 }
 
+get_ibft_mod() {
+    local ibft_mac=$1
+    # Return the iSCSI offload module for a given MAC address
+    iscsiadm -m iface | while read iface_name iface_desc ; do
+        IFS=$','
+        set -- $iface_desc
+        if [ "$ibft_mac" = "$2" ] ; then
+            echo $1
+            return 0
+        fi
+        unset IFS
+    done
+}
+
 install_ibft() {
     # When iBFT / iscsi_boot is detected:
     # - Use 'ip=ibft' to set up iBFT network interface
+    #   Note: bnx2i is using a different MAC address of iSCSI offloading
+    #         so the 'ip=ibft' parameter must not be set
     # - specify firmware booting cmdline parameter
 
     for d in /sys/firmware/* ; do
+        if [ -d ${d}/ethernet0 ] ; then
+            read ibft_mac < ${d}/ethernet0/mac
+            ibft_mod=$(get_ibft_mod $ibft_mac)
+        fi
+        if [ -z "$ibft_mod" ] && [ -d ${d}/ethernet1 ] ; then
+            read ibft_mac < ${d}/ethernet1/mac
+            ibft_mod=$(get_ibft_mod $ibft_mac)
+        fi
         if [ -d ${d}/initiator ] ; then
-            if [ ${d##*/} = "ibft" ] ; then
+            if [ ${d##*/} = "ibft" ] && [ "$ibft_mod" != "bnx2i" ] ; then
                 echo -n "ip=ibft "
             fi
             echo -n "rd.iscsi.firmware=1"
-- 
1.8.5.2

[PATCH 16/33] 95iscsi: More empty cmdline fixes

[Thomas Renninger]

From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>

This fixes up some 95iscsi/module-setup.sh which might print out empty
commandline files.

Signed-off-by: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/95iscsi/module-setup.sh | 13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)

diff --git a/modules.d/95iscsi/module-setup.sh b/modules.d/95iscsi/module-setup.sh
index d30e64f..29bba82 100755
--- a/modules.d/95iscsi/module-setup.sh
+++ b/modules.d/95iscsi/module-setup.sh
@@ -76,12 +76,6 @@ depends() {
 installkernel() {
     local _arch=$(uname -m)
 
-    # Detect iBFT and perform mandatory steps
-    if [[ $hostonly_cmdline == "yes" ]] ; then
-        install_ibft > "${initdir}/etc/cmdline.d/95iscsi.conf"
-        echo >> "${initdir}/etc/cmdline.d/95iscsi.conf"
-    fi
-
     instmods bnx2i qla4xxx cxgb3i cxgb4i be2iscsi
     hostonly="" instmods iscsi_tcp iscsi_ibft crc32c iscsi_boot_sysfs
     iscsi_module_filter() {
@@ -131,6 +125,13 @@ install() {
     inst_multiple umount iscsistart hostname iscsi-iname
     inst_multiple -o iscsiuio
     inst_libdir_file 'libgcc_s.so*'
+
+    # Detect iBFT and perform mandatory steps
+    if [[ $hostonly_cmdline == "yes" ]] ; then
+        local _ibftconf=$(install_ibft)
+        [[ $_ibftconf ]] && printf "%s\n" "$_ibftconf" >> "${initdir}/etc/cmdline.d/95iscsi.conf"
+    fi
+
     inst_hook cmdline 90 "$moddir/parse-iscsiroot.sh"
     inst_hook cleanup 90 "$moddir/cleanup-iscsi.sh"
     inst "$moddir/iscsiroot.sh" "/sbin/iscsiroot"
-- 
1.8.5.2

[PATCH 17/33] 95iscsi: generate commandline for software iscsi

[Thomas Renninger]

From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>

When installing on an software iscsi root we need to generate
the correct commandline, otherwise the system cannot boot.

References: bnc#880108
Patch: 0093

When using software iscsi we should be using the existing
configuration from the system for the time being.

References: bnc#884768
Patch: 0095

Avoid bad ip route call on empty address
Patch: 0143

Signed-off-by: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/95iscsi/module-setup.sh | 91 +++++++++++++++++++++++++++++++++++++--
 1 file changed, 88 insertions(+), 3 deletions(-)

diff --git a/modules.d/95iscsi/module-setup.sh b/modules.d/95iscsi/module-setup.sh
index 29bba82..cd6cb4b 100755
--- a/modules.d/95iscsi/module-setup.sh
+++ b/modules.d/95iscsi/module-setup.sh
@@ -67,6 +67,84 @@ install_ibft() {
     done
 }
 
+install_iscsiroot() {
+    local devpath=$1
+    local scsi_path iscsi_lun session c d conn
+    local iscsi_session iscsi_address iscsi_port iscsi_targetname iscsi_tpgt
+
+    scsi_path=${devpath%%/block*}
+    [ "$scsi_path" = "$devpath" ] && return 1
+    iscsi_lun=${scsi_path##*:}
+    [ "$iscsi_lun" = "$scsi_path" ] && return 1
+    session=${devpath%%/target*}
+    [ "$session" = "$devpath" ] && return 1
+    iscsi_session=${session##*/}
+    [ "$iscsi_session" = "$session" ] && return 1
+
+    for d in ${session}/* ; do
+        case $d in
+	    *connection*)
+	        c=${d##*/}
+	        conn=${d}/iscsi_connection/${c}
+	        if [ -d ${conn} ] ; then
+		    iscsi_address=$(cat ${conn}/persistent_address)
+		    iscsi_port=$(cat ${conn}/persistent_port)
+	        fi
+	        ;;
+	    *session)
+	        if [ -d ${d}/${iscsi_session} ] ; then
+                    iscsi_initiator=$(cat ${d}/${iscsi_session}/initiatorname)
+		    iscsi_targetname=$(cat ${d}/${iscsi_session}/targetname)
+	        fi
+	        ;;
+        esac
+    done
+
+    [ -z "$iscsi_address" ] && return
+    local_address=$(ip -o route get to $iscsi_address | sed -n 's/.*src \([0-9a-f.:]*\).*/\1/p')
+    ifname=$(ip -o route get to $iscsi_address | sed -n 's/.*dev \([^ ]*\).*/\1/p')
+    printf 'ip=%s:static ' ${ifname}
+
+    if [ -e /sys/class/net/$ifname/address ] ; then
+        ifmac=$(cat /sys/class/net/$ifname/address)
+        printf 'ifname=%s:%s ' ${ifname} ${ifmac}
+    fi
+
+    if [ -n "$iscsi_address" -a -n "$iscsi_targetname" ] ; then
+        if [ -n "$iscsi_port" -a "$iscsi_port" -eq 3260 ] ; then
+            iscsi_port=
+        fi
+        if [ -n "$iscsi_lun" -a "$iscsi_lun" -eq 0 ] ; then
+            iscsi_lun=
+        fi
+        # In IPv6 case rd.iscsi.initatior= must pass address in [] brackets
+        case "$iscsi_address" in
+            *:*)
+                iscsi_address="[$iscsi_address]"
+                ;;
+        esac
+        echo "rd.iscsi.initiator=${iscsi_initiator} netroot=iscsi:${iscsi_address}::${iscsi_port}:${iscsi_lun}:${iscsi_targetname}"
+    fi
+    return 0
+}
+
+
+install_softiscsi() {
+    [ -d /sys/firmware/ibft ] && return 0
+
+    is_softiscsi() {
+        local _dev=$1
+        local iscsi_dev
+
+        [[ -L "/sys/dev/block/$_dev" ]] || return
+        iscsi_dev=$(cd -P /sys/dev/block/$_dev; echo $PWD)
+        install_iscsiroot $iscsi_dev
+    }
+
+    for_each_host_dev_and_slaves_all is_softiscsi || return 255
+    return 0
+}
+
 # called by dracut
 depends() {
     echo network rootfs-block
@@ -117,7 +195,14 @@ installkernel() {
 
 # called by dracut
 cmdline() {
-    install_ibft
+    local _iscsiconf=$(install_ibft)
+    {
+        if [ "$_iscsiconf" ] ; then
+            echo ${_iscsiconf}
+        else
+            install_softiscsi
+        fi
+    } | sort | uniq
 }
 
 # called by dracut
@@ -128,8 +213,8 @@ install() {
 
     # Detect iBFT and perform mandatory steps
     if [[ $hostonly_cmdline == "yes" ]] ; then
-        local _ibftconf=$(install_ibft)
-        [[ $_ibftconf ]] && printf "%s\n" "$_ibftconf" >> "${initdir}/etc/cmdline.d/95iscsi.conf"
+        local _iscsiconf=$(cmdline)
+        [[ $_iscsiconf ]] && printf "%s\n" "$_iscsiconf" >> "${initdir}/etc/cmdline.d/95iscsi.conf"
     fi
 
     inst_hook cmdline 90 "$moddir/parse-iscsiroot.sh"
-- 
1.8.5.2

[PATCH 18/33] iscsi: iscsi.initiator and others can and must only show up once

[Thomas Renninger]

Make sure duplicates of iscsi.initiator vanish.
Only get one rd.iscsi.* paramter value. If getargs is used and several
parameters are parsed, one gets two values separated by whitespace in a
variable which breaks later code and is not suppported.

Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/95iscsi/iscsiroot.sh    | 18 +++++++++---------
 modules.d/95iscsi/module-setup.sh |  5 ++++-
 2 files changed, 13 insertions(+), 10 deletions(-)

diff --git a/modules.d/95iscsi/iscsiroot.sh b/modules.d/95iscsi/iscsiroot.sh
index 7351427..5bfb017 100755
--- a/modules.d/95iscsi/iscsiroot.sh
+++ b/modules.d/95iscsi/iscsiroot.sh
@@ -81,23 +81,23 @@ handle_netroot()
     local p
 
     # override conf settings by command line options
-    arg=$(getargs rd.iscsi.initiator -d iscsi_initiator=)
+    arg=$(getarg rd.iscsi.initiator -d iscsi_initiator=)
     [ -n "$arg" ] && iscsi_initiator=$arg
-    arg=$(getargs rd.iscsi.target.name -d iscsi_target_name=)
+    arg=$(getarg rd.iscsi.target.name -d iscsi_target_name=)
     [ -n "$arg" ] && iscsi_target_name=$arg
-    arg=$(getargs rd.iscsi.target.ip -d iscsi_target_ip)
+    arg=$(getarg rd.iscsi.target.ip -d iscsi_target_ip)
     [ -n "$arg" ] && iscsi_target_ip=$arg
-    arg=$(getargs rd.iscsi.target.port -d iscsi_target_port=)
+    arg=$(getarg rd.iscsi.target.port -d iscsi_target_port=)
     [ -n "$arg" ] && iscsi_target_port=$arg
-    arg=$(getargs rd.iscsi.target.group -d iscsi_target_group=)
+    arg=$(getarg rd.iscsi.target.group -d iscsi_target_group=)
     [ -n "$arg" ] && iscsi_target_group=$arg
-    arg=$(getargs rd.iscsi.username -d iscsi_username=)
+    arg=$(getarg rd.iscsi.username -d iscsi_username=)
     [ -n "$arg" ] && iscsi_username=$arg
-    arg=$(getargs rd.iscsi.password -d iscsi_password)
+    arg=$(getarg rd.iscsi.password -d iscsi_password)
     [ -n "$arg" ] && iscsi_password=$arg
-    arg=$(getargs rd.iscsi.in.username -d iscsi_in_username=)
+    arg=$(getarg rd.iscsi.in.username -d iscsi_in_username=)
     [ -n "$arg" ] && iscsi_in_username=$arg
-    arg=$(getargs rd.iscsi.in.password -d iscsi_in_password=)
+    arg=$(getarg rd.iscsi.in.password -d iscsi_in_password=)
     [ -n "$arg" ] && iscsi_in_password=$arg
     for p in $(getargs rd.iscsi.param -d iscsi_param); do
 	iscsi_param="$iscsi_param --param $p"
diff --git a/modules.d/95iscsi/module-setup.sh b/modules.d/95iscsi/module-setup.sh
index cd6cb4b..9975e59 100755
--- a/modules.d/95iscsi/module-setup.sh
+++ b/modules.d/95iscsi/module-setup.sh
@@ -123,7 +123,10 @@ install_iscsiroot() {
                 iscsi_address="[$iscsi_address]"
                 ;;
         esac
-        echo "rd.iscsi.initiator=${iscsi_initiator} netroot=iscsi:${iscsi_address}::${iscsi_port}:${iscsi_lun}:${iscsi_targetname}"
+        # Must be two separate lines, so that "sort | uniq" commands later
+        # can sort out rd.iscsi.initiator= duplicates
+        echo "rd.iscsi.initiator=${iscsi_initiator}"
+        echo "netroot=iscsi:${iscsi_address}::${iscsi_port}:${iscsi_lun}:${iscsi_targetname}"
     fi
     return 0
 }
-- 
1.8.5.2

[PATCH 19/33] Restore original IFS value

[Thomas Renninger]

From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>

Whenever the IFS setting is modified it needs to be restored to
its original value. Otherwise it'll lead to unexpected results.

References: bnc#883770

Signed-off-by: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/90kernel-modules/insmodpost.sh   | 2 ++
 modules.d/90kernel-modules/parse-kernel.sh | 4 ++++
 modules.d/95iscsi/module-setup.sh          | 6 ++++--
 modules.d/98systemd/dracut-pre-udev.sh     | 4 ++++
 4 files changed, 14 insertions(+), 2 deletions(-)

diff --git a/modules.d/90kernel-modules/insmodpost.sh b/modules.d/90kernel-modules/insmodpost.sh
index 4362f0a..b4bcdd0 100755
--- a/modules.d/90kernel-modules/insmodpost.sh
+++ b/modules.d/90kernel-modules/insmodpost.sh
@@ -4,9 +4,11 @@
 
 for modlist in $(getargs rd.driver.post -d rdinsmodpost=); do
     (
+        OLDIFS="$IFS"
         IFS=,
         for m in $modlist; do
             modprobe $m
         done
+        IFS="$OLDIFS"
     )
 done
diff --git a/modules.d/90kernel-modules/parse-kernel.sh b/modules.d/90kernel-modules/parse-kernel.sh
index cbe7c85..81f7131 100755
--- a/modules.d/90kernel-modules/parse-kernel.sh
+++ b/modules.d/90kernel-modules/parse-kernel.sh
@@ -11,10 +11,12 @@ fi
 
 for i in $(getargs rd.driver.pre -d rdloaddriver=); do
     (
+        OLDIFS="$IFS"
         IFS=,
         for p in $i; do
             modprobe $p 2>&1 | vinfo
         done
+        IFS="$OLDIFS"
     )
 done
 
@@ -23,10 +25,12 @@ done
 
 for i in $(getargs rd.driver.blacklist -d rdblacklist=); do
     (
+        OLDIFS="$IFS"
         IFS=,
         for p in $i; do
             echo "blacklist $p" >>  $_modprobe_d/initramfsblacklist.conf
         done
+        IFS="$OLDIFS"
     )
 done
 
diff --git a/modules.d/95iscsi/module-setup.sh b/modules.d/95iscsi/module-setup.sh
index 9975e59..f551c49 100755
--- a/modules.d/95iscsi/module-setup.sh
+++ b/modules.d/95iscsi/module-setup.sh
@@ -29,16 +29,18 @@ check() {
 }
 
 get_ibft_mod() {
-    local ibft_mac=$1
+    local ibft_mac=$1 IFS
     # Return the iSCSI offload module for a given MAC address
     iscsiadm -m iface | while read iface_name iface_desc ; do
+        OLDIFS="$IFS"
         IFS=$','
         set -- $iface_desc
+        IFS="$OLDIFS"
         if [ "$ibft_mac" = "$2" ] ; then
             echo $1
             return 0
         fi
-        unset IFS
+        
     done
 }
 
diff --git a/modules.d/98systemd/dracut-pre-udev.sh b/modules.d/98systemd/dracut-pre-udev.sh
index ca13048..b2be0e0 100755
--- a/modules.d/98systemd/dracut-pre-udev.sh
+++ b/modules.d/98systemd/dracut-pre-udev.sh
@@ -24,10 +24,12 @@ fi
 
 for i in $(getargs rd.driver.pre -d rdloaddriver=); do
     (
+        OLDIFS="$IFS"
         IFS=,
         for p in $i; do
             modprobe $p 2>&1 | vinfo
         done
+        IFS="$OLDIFS"
     )
 done
 
@@ -36,10 +38,12 @@ done
 
 for i in $(getargs rd.driver.blacklist -d rdblacklist=); do
     (
+        OLDIFS="$IFS"
         IFS=,
         for p in $i; do
             echo "blacklist $p" >>  $_modprobe_d/initramfsblacklist.conf
         done
+        IFS="$OLDIFS"
     )
 done
 
-- 
1.8.5.2

[PATCH 20/33] Implement 'rd.timeout' to modify the device timeout

[Thomas Renninger]

From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>

When generating units for devices the administrator might
want to use a different timeout than the default.
So implement a new parameter 'rd.timeout' for this.

References: bnc#878770

Signed-off-by: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 dracut.cmdline.7.asc                    | 7 ++++++-
 modules.d/98systemd/rootfs-generator.sh | 7 +++++--
 modules.d/99base/dracut-lib.sh          | 6 +++++-
 3 files changed, 16 insertions(+), 4 deletions(-)

diff --git a/dracut.cmdline.7.asc b/dracut.cmdline.7.asc
index 48356c7..9741cf7 100644
--- a/dracut.cmdline.7.asc
+++ b/dracut.cmdline.7.asc
@@ -142,11 +142,16 @@ Misc
     have been loaded. This parameter can be specified multiple times.
 
 **rd.retry=**__<seconds>__::
-    specify how long dracut should wait for devices to appear.
+    specify how long dracut should retry the initqueue to configure devices.
     The default is 30 seconds. After 2/3 of the time, degraded raids are force
     started. If you have hardware, which takes a very long time to announce its
     drives, you might want to extend this value.
 
+**rd.timeout=**__<seconds>__::
+    specify how long dracut should wait for devices to appear. The
+    default is '0', which means 'forever'. Note that this timeout
+    should be longer than rd.retry to allow for proper configuration.
+
 **rd.noverifyssl**::
     accept self-signed certificates for ssl downloads.
 
diff --git a/modules.d/98systemd/rootfs-generator.sh b/modules.d/98systemd/rootfs-generator.sh
index 9523729..5b06137 100755
--- a/modules.d/98systemd/rootfs-generator.sh
+++ b/modules.d/98systemd/rootfs-generator.sh
@@ -5,8 +5,11 @@ type getarg >/dev/null 2>&1 || . /lib/dracut-lib.sh
 generator_wait_for_dev()
 {
     local _name
+    local _timeout
 
     _name="$(str_replace "$1" '/' '\x2f')"
+    _timeout=$(getarg rd.timeout)
+    _timeout=${_timeout:-0}
 
     [ -e "$hookdir/initqueue/finished/devexists-${_name}.sh" ] && return 0
 
@@ -27,7 +30,7 @@ generator_wait_for_dev()
         mkdir -p /run/systemd/generator/${_name}.device.d
         {
             echo "[Unit]"
-            echo "JobTimeoutSec=0"
+            echo "JobTimeoutSec=$_timeout"
         } > /run/systemd/generator/${_name}.device.d/timeout.conf
     fi
 }
@@ -58,6 +61,6 @@ case "$root" in
         rootok=1 ;;
 esac
 
-[ "${root%%:*}" = "block" ] && generator_wait_for_dev "${root#block:}"
+[ "${root%%:*}" = "block" ] && generator_wait_for_dev "${root#block:}" "$RDRETRY"
 
 exit 0
diff --git a/modules.d/99base/dracut-lib.sh b/modules.d/99base/dracut-lib.sh
index 8bc6b26..7080f1a 100755
--- a/modules.d/99base/dracut-lib.sh
+++ b/modules.d/99base/dracut-lib.sh
@@ -886,12 +886,16 @@ wait_for_dev()
     local _name
     local _needreload
     local _noreload
+    local _timeout
 
     if [ "$1" = "-n" ]; then
         _noreload=1
         shift
     fi
 
+    _timeout=$(getarg rd.timeout)
+    _timeout=${_timeout:-0}
+
     _name="$(str_replace "$1" '/' '\x2f')"
 
     type mark_hostonly >/dev/null 2>&1 && mark_hostonly "$hookdir/initqueue/finished/devexists-${_name}.sh"
@@ -918,7 +922,7 @@ wait_for_dev()
             mkdir -p ${PREFIX}/etc/systemd/system/${_name}.device.d
             {
                 echo "[Unit]"
-                echo "JobTimeoutSec=0"
+                echo "JobTimeoutSec=$_timeout"
             } > ${PREFIX}/etc/systemd/system/${_name}.device.d/timeout.conf
             type mark_hostonly >/dev/null 2>&1 && mark_hostonly /etc/systemd/system/${_name}.device.d/timeout.conf
             _needreload=1
-- 
1.8.5.2

[PATCH 21/33] Do not call 'lvm' for non-LVM device-mapper devices

[Thomas Renninger]

From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>

If a device-mapper device is not created by LVM it's pointless
to call any 'lvm' programs got extract details; they'll be
failing anyway. So check the UUID before calling 'lvm'.
This speeds up initrd creation and avoids I/O errors on
multipath devices.

Signed-off-by: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 dracut-functions.sh | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/dracut-functions.sh b/dracut-functions.sh
index b43b766..02aba04 100755
--- a/dracut-functions.sh
+++ b/dracut-functions.sh
@@ -698,10 +698,12 @@ for_each_host_dev_and_slaves()
 # but you cannot create the logical volume without the volume group.
 # And the volume group might be bigger than the devices the LV needs.
 check_vol_slaves() {
-    local _lv _vg _pv
+    local _lv _vg _pv _dm
     for i in /dev/mapper/*; do
         [[ $i == /dev/mapper/control ]] && continue
         _lv=$(get_maj_min $i)
+        _dm=/sys/dev/block/$_lv/dm
+        [[ -f $_dm/uuid  && $(<$_dm/uuid) =~ LVM-* ]] || continue
         if [[ $_lv = $2 ]]; then
             _vg=$(lvm lvs --noheadings -o vg_name $i 2>/dev/null)
             # strip space
-- 
1.8.5.2

[PATCH 22/33] Print stored dracut commandline during initramfs build

[Thomas Renninger]

From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>

When generating the initramfs we should be printing out the
generated dracut commandline used for booting.
This will simplify debugging.

Signed-off-by: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 dracut.sh | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/dracut.sh b/dracut.sh
index 4590f79..9a19b43 100755
--- a/dracut.sh
+++ b/dracut.sh
@@ -1576,6 +1576,16 @@ if ! ( echo $PARMS_TO_STORE > $initdir/lib/dracut/build-parameter.txt ); then
     exit 1
 fi
 
+if [[ $hostonly_cmdline ]] ; then
+    if [ -d $initdir/etc/cmdline.d ];then
+        dinfo "Stored kernel commandline:"
+        for conf in $initdir/etc/cmdline.d/*.conf ; do
+            dinfo "$(< $conf)"
+        done
+    else
+        dinfo "No dracut internal kernel commandline stored in initrd"
+    fi
+fi
 rm -f -- "$outfile"
 dinfo "*** Creating image file ***"
 
-- 
1.8.5.2

[PATCH 23/33] Align dev_unit_name() with systemd's function.

[Thomas Renninger]

From: Thorsten Behrens <tbehrens-IBi9RG/b67k@public.gmane.org>

Add more corner cases from systemd's
unit_name_from_path_instance() C function.

Signed-off-by: Thorsten Behrens <tbehrens-IBi9RG/b67k@public.gmane.org>
Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/99base/dracut-lib.sh | 24 ++++++++++++++++++------
 1 file changed, 18 insertions(+), 6 deletions(-)

diff --git a/modules.d/99base/dracut-lib.sh b/modules.d/99base/dracut-lib.sh
index 7080f1a..405e3c7 100755
--- a/modules.d/99base/dracut-lib.sh
+++ b/modules.d/99base/dracut-lib.sh
@@ -862,18 +862,30 @@ wait_for_mount()
     } >> "$hookdir/emergency/90-${_name}.sh"
 }
 
+# get a systemd-compatible unit name from a path
+# (mimicks unit_name_from_path_instance())
 dev_unit_name()
 {
+    local dev="$1"
+
     if command -v systemd-escape >/dev/null; then
-        systemd-escape -p  "$1"
+        systemd-escape -p  "$dev"
         return
     fi
 
-    _name="${1%%/}"
-    _name="${_name##/}"
-    _name="$(str_replace "$_name" '-' '\x2d')"
-    _name="$(str_replace "$_name" '/' '-')"
-    echo "$_name"
+    if [ "$dev" = "/" -o -z "$dev" ]; then
+        printf -- "-"
+        exit 0
+    fi
+
+    dev="${1%%/}"
+    dev="${dev##/}"
+    dev="$(str_replace "$dev" '\' '\x5c')"
+    dev="$(str_replace "$dev" '-' '\x2d')"
+    dev=${dev/#\./\\x2e}
+    dev="$(str_replace "$dev" '/' '-')"
+
+    printf -- "%s" "$dev"
 }
 
 # wait_for_dev <dev>
-- 
1.8.5.2

[PATCH 24/33] Generate fallback mount unit for root filesystem

[Thomas Renninger]

From: Thorsten Behrens <tbehrens-IBi9RG/b67k@public.gmane.org>

If kernel cmdline has no root= provided, *and* the initrd carries
a default value, create a systemd mount unit to have it available
in /sysroot

References: bnc#855258

Signed-off-by: Thorsten Behrens <tbehrens-IBi9RG/b67k@public.gmane.org>
Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/98systemd/rootfs-generator.sh | 34 ++++++++++++++++++++++++++++++++-
 1 file changed, 33 insertions(+), 1 deletion(-)

diff --git a/modules.d/98systemd/rootfs-generator.sh b/modules.d/98systemd/rootfs-generator.sh
index 5b06137..f3c7d1f 100755
--- a/modules.d/98systemd/rootfs-generator.sh
+++ b/modules.d/98systemd/rootfs-generator.sh
@@ -35,6 +35,35 @@ generator_wait_for_dev()
     fi
 }
 
+generator_mount_rootfs()
+{
+    local _type=$2
+    local _flags=$3
+    local _name
+
+    [ -z "$1" ] && return 0
+
+    _name=$(dev_unit_name "$1")
+    [ -d /run/systemd/generator ] || mkdir -p /run/systemd/generator
+    if ! [ -f /run/systemd/generator/sysroot.mount ]; then
+        {
+            echo "[Unit]"
+            echo "Before=initrd-root-fs.target"
+            echo "RequiresOverridable=systemd-fsck@${_name}.service"
+            echo "After=systemd-fsck@${_name}.service"
+            echo "[Mount]"
+            echo "Where=/sysroot"
+            echo "What=$1"
+            echo "Options=${_flags}"
+            echo "Type=${_type}"
+        } > /run/systemd/generator/sysroot.mount
+    fi
+    if ! [ -L /run/systemd/generator/initrd-root-fs.target.requires/sysroot.mount ]; then
+        [ -d /run/systemd/generator/initrd-root-fs.target.requires ] || mkdir -p /run/systemd/generator/initrd-root-fs.target.requires
+        ln -s ../sysroot.mount /run/systemd/generator/initrd-root-fs.target.requires/sysroot.mount
+    fi
+}
+
 root=$(getarg root=)
 case "$root" in
     block:LABEL=*|LABEL=*)
@@ -61,6 +90,9 @@ case "$root" in
         rootok=1 ;;
 esac
 
-[ "${root%%:*}" = "block" ] && generator_wait_for_dev "${root#block:}" "$RDRETRY"
+if [ "${root%%:*}" = "block" ]; then
+   generator_wait_for_dev "${root#block:}" "$RDRETRY"
+   grep -q 'root=' /proc/cmdline || generator_mount_rootfs "${root#block:}" "$(getarg rootfstype=)" "$(getarg rootflags=)"
+fi
 
 exit 0
-- 
1.8.5.2

[PATCH 25/33] 99base: warn on invalid command for initqueue

[Thomas Renninger]

From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>

When 'initqueue' is called with an invalid command it'll generate
invalid job scripts. This will lead to confusing error messages
later on.
So abort in these cases and print out a warning.

Signed-off-by: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/99base/initqueue.sh | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/modules.d/99base/initqueue.sh b/modules.d/99base/initqueue.sh
index 4a07da3..193cc5c 100755
--- a/modules.d/99base/initqueue.sh
+++ b/modules.d/99base/initqueue.sh
@@ -42,6 +42,10 @@ exe=$1
 shift
 
 [ -x "$exe" ] || exe=$(command -v $exe)
+if [ -z "$exe" ] ; then
+    echo "Invalid command"
+    exit 1
+fi
 
 {
     [ -n "$onetime" ] && echo '[ -e "$job" ] && rm -f -- "$job"'
-- 
1.8.5.2

[PATCH 26/33] Handle module alias properly

[Thomas Renninger]

From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>

Some modules (like ext4) provide aliases by which the modules
can be accessed, too. But when using aliases directly dracut
fails to include the correct module. So translate the alias
into the correct module name before checking the module.

References: bnc#886839

Signed-off-by: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 dracut-functions.sh | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/dracut-functions.sh b/dracut-functions.sh
index 02aba04..d85c639 100755
--- a/dracut-functions.sh
+++ b/dracut-functions.sh
@@ -1689,6 +1689,13 @@ instmods() {
             --*) _mpargs+=" $_mod" ;;
             *)
                 _mod=${_mod##*/}
+                # Check for aliased modules
+                _modalias=$(modinfo -k $kernel -F filename $_mod 2> /dev/null)
+                _modalias=${_modalias%.ko}
+                if [ "${_modalias##*/}" != "$_mod" ] ; then
+                    _mod=${_modalias##*/}
+                fi
+
                 # if we are already installed, skip this module and go on
                 # to the next one.
                 if [[ $DRACUT_KERNEL_LAZY_HASHDIR ]] && \
-- 
1.8.5.2

[PATCH 27/33] 99base: Increase initqueue timeout in non systemd case

[Thomas Renninger]

In case of systemd is used the timeout already is set to 180s, compare
with file: modules.d/98systemd/dracut-initqueue.sh

Do the same if systemd is not used, e.g. in kdump case.

Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/99base/init.sh | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/modules.d/99base/init.sh b/modules.d/99base/init.sh
index 3afe6d5..76b1247 100755
--- a/modules.d/99base/init.sh
+++ b/modules.d/99base/init.sh
@@ -165,7 +165,7 @@ make_trace_mem "hook initqueue" '1:shortmem' '2+:mem' '3+:slab'
 getarg 'rd.break=initqueue' -d 'rdbreak=initqueue' && emergency_shell -n initqueue "Break before initqueue"
 
 RDRETRY=$(getarg rd.retry -d 'rd_retry=')
-RDRETRY=${RDRETRY:-30}
+RDRETRY=${RDRETRY:-180}
 RDRETRY=$(($RDRETRY*2))
 export RDRETRY
 main_loop=0
-- 
1.8.5.2

[PATCH 28/33] 90lvm: Install dm-snapshot module

[Thomas Renninger]

bnc#888530

Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/90lvm/module-setup.sh | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/modules.d/90lvm/module-setup.sh b/modules.d/90lvm/module-setup.sh
index 9ddc56e..5515ac0 100755
--- a/modules.d/90lvm/module-setup.sh
+++ b/modules.d/90lvm/module-setup.sh
@@ -42,6 +42,10 @@ cmdline() {
     done
 }
 
+installkernel() {
+    instmods dm-snapshot
+}
+
 # called by dracut
 install() {
     local _i
-- 
1.8.5.2

[PATCH 29/33] Check for logfile (--logfile option) and create it if necessary

[Thomas Renninger]

From: jloeser <jloeser-l3A5Bk7waGM@public.gmane.org>

If a logfile is passed to dracut via --logfile option and doesn't
exist, dracut doesn't create it and logs nothing. Instead, dracut
should try to touch the file and print a warning if creating fails.

References: bnc#892191
Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 dracut.sh | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/dracut.sh b/dracut.sh
index 9a19b43..c657649 100755
--- a/dracut.sh
+++ b/dracut.sh
@@ -816,6 +816,16 @@ stdloglvl=$((stdloglvl + verbosity_mod_l))
 # eliminate IFS hackery when messing with fw_dir
 fw_dir=${fw_dir//:/ }
 
+# check for logfile and try to create one if it doesn't exist
+if [[ -n "$logfile" ]];then
+    if [[ ! -f "$logfile" ]];then
+        touch "$logfile"
+        if [ ! $? -eq 0 ] ;then
+            printf "%s\n" "dracut: touch $logfile failed." >&2
+        fi
+    fi
+fi
+
 # handle compression options.
 [[ $compress ]] || compress="gzip"
 case $compress in
-- 
1.8.5.2

[PATCH 30/33] dracut.sh: Avoid duplicate devices in host_devs

[Thomas Renninger]

Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 dracut.sh | 26 ++++++++++++++++++--------
 1 file changed, 18 insertions(+), 8 deletions(-)

diff --git a/dracut.sh b/dracut.sh
index c657649..f384f73 100755
--- a/dracut.sh
+++ b/dracut.sh
@@ -219,6 +219,16 @@ push() {
     done
 }
 
+# Fills up host_devs stack variable and makes sure there are no duplicates
+push_host_devs() {
+    local _dev
+    for _dev in ${host_devs[@]}; do
+        [ "$_dev" = "$1" ] && return
+    done
+    push host_devs "$1"
+}
+
+
 # function pop()
 # pops the last value from a stack
 # assigns value to second argument variable
@@ -1062,25 +1072,25 @@ declare -A host_fs_types
 for line in "${fstab_lines[@]}"; do
     set -- $line
     #dev mp fs fsopts
-    push host_devs "$1"
+    push_host_devs "$1"
     host_fs_types["$1"]="$3"
 done
 
 for f in $add_fstab; do
     [[ -e $f ]] || continue
     while read dev rest; do
-        push host_devs "$dev"
+        push_host_devs "$dev"
     done < "$f"
 done
 
 for dev in $add_device; do
-    push host_devs "$dev"
+    push_host_devs "$dev"
 done
 
 if (( ${#add_device_l[@]} )); then
     while pop add_device_l val; do
         add_device+=" $val "
-        push host_devs "$val"
+        push_host_devs "$val"
     done
 fi
 
@@ -1109,9 +1119,9 @@ if [[ $hostonly ]]; then
         _dev=$(find_block_device "$mp")
         _bdev=$(readlink -f "/dev/block/$_dev")
         [[ -b $_bdev ]] && _dev=$_bdev
-        push host_devs $_dev
+        push_host_devs $_dev
         [[ "$mp" == "/" ]] && root_dev="$_dev"
-        push host_devs "$_dev"
+        push_host_devs "$_dev"
     done
 
     if [[ -f /proc/swaps ]] && [[ -f /etc/fstab ]]; then
@@ -1140,7 +1150,7 @@ if [[ $hostonly ]]; then
                     done < /etc/crypttab
                 fi
 
-                push host_devs "$(readlink -f "$dev")"
+                push_host_devs "$(readlink -f "$dev")"
                 break
             done < /etc/fstab
         done < /proc/swaps
@@ -1205,7 +1215,7 @@ for dev in "${!host_fs_types[@]}"; do
     fi
     if [[ $journaldev ]]; then
         dev="$(readlink -f "$dev")"
-        push host_devs "$dev"
+        push_host_devs "$dev"
         _get_fs_type "$dev"
         check_block_and_slaves_all _get_fs_type "$(get_maj_min "$dev")"
     fi
-- 
1.8.5.2

[PATCH 31/33] dracut.sh: Fix UUID= fstab parsing in case --mount option is passed

[Thomas Renninger]

Dracut parses /etc/fstab when --mount is option is passed (e.g. kdump).

In host_devs variable the real block device must be stored, not UUID=
There are other /etc/fstab syntax possibilities we now warn that they
are not correctly parsed. This will be fixed by another patch
when there is time to test this properly.

Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 dracut.sh | 20 ++++++++++++++++++--
 1 file changed, 18 insertions(+), 2 deletions(-)

diff --git a/dracut.sh b/dracut.sh
index f384f73..3c3efc1 100755
--- a/dracut.sh
+++ b/dracut.sh
@@ -1071,9 +1071,25 @@ declare -A host_fs_types
 
 for line in "${fstab_lines[@]}"; do
     set -- $line
+    dev="$1"
     #dev mp fs fsopts
-    push_host_devs "$1"
-    host_fs_types["$1"]="$3"
+    case "$dev" in
+        UUID=*)
+            dev=$(blkid -l -t UUID=${dev#UUID=} -o device)
+            ;;
+        LABEL=*)
+            dev=$(blkid -l -t LABEL=${dev#LABEL=} -o device)
+            ;;
+        PARTUUID=*)
+            dev=$(blkid -l -t PARTUUID=${dev#PARTUUID=} -o device)
+            ;;
+        PARTLABEL=*)
+            dev=$(blkid -l -t PARTLABEL=${dev#PARTLABEL=} -o device)
+            ;;
+    esac
+    [ -z "$dev" ] && dwarn "Bad fstab entry $@" && continue
+    push_host_devs "$dev"
+    host_fs_types["$dev"]="$3"
 done
 
 for f in $add_fstab; do
-- 
1.8.5.2

[PATCH 32/33] dracut.usage.asc: Remove distro specific help from manpage

[Thomas Renninger]

Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 dracut.usage.asc | 19 +------------------
 1 file changed, 1 insertion(+), 18 deletions(-)

diff --git a/dracut.usage.asc b/dracut.usage.asc
index 1a79638..e2182d9 100644
--- a/dracut.usage.asc
+++ b/dracut.usage.asc
@@ -246,23 +246,6 @@ specified multiple times.
 If your root partition is on a network drive, you have to have the network
 dracut modules installed to create a network aware initramfs image.
 
-On a Red Hat Enterprise Linux or Fedora system, this means, you have to install
-the _dracut-network_ rpm package:
-
-
-----
-# yum install dracut-network
-----
-
-The resulting initramfs image can be served by a boot manager residing on your
-local hard drive or it can be served by a PXE/TFTP server.
-
-How to setup your PXE/TFTP server can be found in the
-http://docs.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/6/html/Storage_Administration_Guide/[Red
-Hat Enterprise Linux 6 Storage Administration Guide]
-or http://docs.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/6/html/Storage_Administration_Guide/[Red
-Hat Enterprise Linux 7 Storage Administration Guide].
-
 If you specify ip=dhcp on the kernel command line, then dracut asks a dhcp
 server about the ip adress for the machine. The dhcp server can also serve an
 additional root-path, which will set the root device for dracut. With this
@@ -302,7 +285,7 @@ This will reduce the size of the initramfs image significantly.
 If the boot process does not succeed, you have several options to debug the
 situation. Some of the basic operations are covered here. For more information
 you should also visit:
-http://fedoraproject.org/wiki/How_to_debug_Dracut_problems
+https://www.kernel.org/pub/linux/utils/boot/dracut/dracut.html
 
 
 [[identifying-your-problem-area]]
-- 
1.8.5.2

[PATCH 33/33] nfs: Add ip=... and root=nfs... parameters to internal dracut

[Thomas Renninger]

 cmdline

If the rootfs is an nfs mount, also know as nfsroot, add the correct
parameter to the dracut cmdline.

Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
---
 modules.d/95nfs/module-setup.sh | 46 ++++++++++++++++++++++++++++++++++++++---
 1 file changed, 43 insertions(+), 3 deletions(-)

diff --git a/modules.d/95nfs/module-setup.sh b/modules.d/95nfs/module-setup.sh
index e1340b3..41c6ef7 100755
--- a/modules.d/95nfs/module-setup.sh
+++ b/modules.d/95nfs/module-setup.sh
@@ -14,7 +14,6 @@ check() {
         done
         return 255
     }
-
     return 0
 }
 
@@ -29,14 +28,56 @@ installkernel() {
     instmods nfs sunrpc ipv6 nfsv2 nfsv3 nfsv4 nfs_acl nfs_layout_nfsv41_files
 }
 
+cmdline() {
+    local nfs_device
+    local nfs_options
+    local nfs_root
+    local nfs_address
+    local lookup
+    local ifname
+
+    ### nfsroot= ###
+    nfs_device=$(findmnt -t nfs4 -n -o SOURCE /)
+    if [ -n "$nfs_device" ];then
+        nfs_root="root=nfs4:$nfs_device"
+    else
+        nfs_device=$(findmnt -t nfs -n -o SOURCE /)
+        [ -z "$nfs_device" ] && return
+        nfs_root="root=nfs:$nfs_device"
+    fi
+    nfs_options=$(findmnt -t nfs4,nfs -n -o OPTIONS /)
+    [ -n "$nfs_options" ] && nfs_root="$nfs_root:$nfs_options"
+    echo "$nfs_root"
+
+    ### ip= ###
+    if [[ $nfs_device = [0-9]*\.[0-9]*\.[0-9]*.[0-9]* ]] || [[ $nfs_device = \[.*\] ]]; then
+        nfs_address="$nfs_device"
+    else
+        lookup=$(host $(echo ${nfs_device%%:*})| head -n1)
+        nfs_address=${lookup##* }
+    fi
+    ifname=$(ip -o route get to $nfs_address | sed -n 's/.*dev \([^ ]*\).*/\1/p')
+    if [ -e /sys/class/net/$ifname/address ] ; then
+        ifmac=$(cat /sys/class/net/$ifname/address)
+        printf 'ifname=%s:%s ' ${ifname} ${ifmac}
+    fi
+
+    printf 'ip=%s:static\n' ${ifname}
+}
+
 # called by dracut
 install() {
     local _i
     local _nsslibs
     inst_multiple -o portmap rpcbind rpc.statd mount.nfs \
-        mount.nfs4 umount rpc.idmapd sed /etc/netconfig
+        mount.nfs4 umount rpc.idmapd sed /etc/netconfig host sed
     inst_multiple /etc/services /etc/nsswitch.conf /etc/rpc /etc/protocols /etc/idmapd.conf
 
+    if [[ $hostonly_cmdline == "yes" ]]; then
+        local _netconf="$(cmdline)"
+        [[ $_netconf ]] && printf "%s\n" "$_netconf" >> "${initdir}/etc/cmdline.d/95nfs.conf"
+    fi
+
     if [ -f /lib/modprobe.d/nfs.conf ]; then
         inst_multiple /lib/modprobe.d/nfs.conf
     else
@@ -75,4 +116,3 @@ install() {
         && chown rpc.rpc "$initdir/var/lib/rpcbind"
     dracut_need_initqueue
 }
-
-- 
1.8.5.2

Re: [PATCH 11/33] 95iscsi: Autodetect iSCSI firmware

[WANG Chao]

Hi,

On 11/20/14 at 03:56pm, Thomas Renninger wrote:
> From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
> 
> Some iSCSI initiator present firmware information in
> /sys/firmware/ibft or /sys/firmware/iscsi_bootX
> 
> Whenever we detect one of those directories we should assume
> that the iSCSI devices should be activated.
> 
> Also incorporates SUSE patches:
> 0049:
> 95iscsi: Use 'ip=ibft' for ibft autoconfiguration
> 
> For iBFT autoconfiguration we should be setting 'ip=ibft'
> instead of rd.neednet. This should instruct dracut to only
> enable the iBFT interfaces and leave the rest alone.

I think,

/sys/firmware/ibft is created for iBFT if detected.
/sys/firmware/iscsi_bootX is created for hardware iSCSI if detected.

ip=ibft denotes we use iBFT, which is software based iscsi.
rd.iscsi.firmware denotes we use hardware iscsi.

Though "iscsistart -b" will handle both the cases, I don't think it's
good idea to mix these concepts.

One more thing, it seems ip=ibft is deprecated now and "rd.iscsi.ibft"
will take over.

Thanks
WANG Chao

> 
> References: bnc#879038
> 
> 0054:
> 95iscsi: update commandline printing
> 
> dracut has a separate callout 'cmdline' which should be used
> for printing out the generated commandline.
> 
> 
> Signed-off-by: Pavel Wieczorkiewicz <pwieczorkiewicz-IBi9RG/b67k@public.gmane.org>
> Signed-off-by: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
> Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
> ---
>  modules.d/95iscsi/module-setup.sh | 26 ++++++++++++++++++++++++++
>  1 file changed, 26 insertions(+)
> 
> diff --git a/modules.d/95iscsi/module-setup.sh b/modules.d/95iscsi/module-setup.sh
> index 8b42d3e..5d2a39b 100755
> --- a/modules.d/95iscsi/module-setup.sh
> +++ b/modules.d/95iscsi/module-setup.sh
> @@ -28,6 +28,21 @@ check() {
>      return 0
>  }
>  
> +install_ibft() {
> +    # When iBFT / iscsi_boot is detected:
> +    # - Use 'ip=ibft' to set up iBFT network interface
> +    # - specify firmware booting cmdline parameter
> +
> +    for d in /sys/firmware/* ; do
> +        if [ -d ${d}/initiator ] ; then
> +            if [ ${d##*/} = "ibft" ] ; then
> +                echo -n "ip=ibft "
> +            fi
> +            echo -n "rd.iscsi.firmware=1"
> +        fi
> +    done
> +}
> +
>  # called by dracut
>  depends() {
>      echo network rootfs-block
> @@ -37,6 +52,12 @@ depends() {
>  installkernel() {
>      local _arch=$(uname -m)
>  
> +    # Detect iBFT and perform mandatory steps
> +    if [[ $hostonly_cmdline == "yes" ]] ; then
> +        install_ibft > "${initdir}/etc/cmdline.d/95iscsi.conf"
> +        echo >> "${initdir}/etc/cmdline.d/95iscsi.conf"
> +    fi
> +
>      instmods bnx2i qla4xxx cxgb3i cxgb4i be2iscsi
>      hostonly="" instmods iscsi_tcp iscsi_ibft crc32c iscsi_boot_sysfs
>      iscsi_module_filter() {
> @@ -77,6 +98,11 @@ installkernel() {
>  }
>  
>  # called by dracut
> +cmdline() {
> +    install_ibft
> +}
> +
> +# called by dracut
>  install() {
>      inst_multiple umount iscsistart hostname iscsi-iname
>      inst_multiple -o iscsiuio
> -- 
> 1.8.5.2
> 
> --
> To unsubscribe from this list: send the line "unsubscribe initramfs" in
> the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 11/33] 95iscsi: Autodetect iSCSI firmware

[Hannes Reinecke]

On 11/21/2014 06:43 AM, WANG Chao wrote:
> Hi,
> 
> On 11/20/14 at 03:56pm, Thomas Renninger wrote:
>> From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
>>
>> Some iSCSI initiator present firmware information in
>> /sys/firmware/ibft or /sys/firmware/iscsi_bootX
>>
>> Whenever we detect one of those directories we should assume
>> that the iSCSI devices should be activated.
>>
>> Also incorporates SUSE patches:
>> 0049:
>> 95iscsi: Use 'ip=ibft' for ibft autoconfiguration
>>
>> For iBFT autoconfiguration we should be setting 'ip=ibft'
>> instead of rd.neednet. This should instruct dracut to only
>> enable the iBFT interfaces and leave the rest alone.
> 
> I think,
> 
> /sys/firmware/ibft is created for iBFT if detected.
> /sys/firmware/iscsi_bootX is created for hardware iSCSI if detected.
> 
Nope.

ibft is created whenever an iBFT section (either in system ROM
or via EFI) is detected.
iscsi_boot is created for be2iscsi devices, which use an iBFT-like
interface.

Both provide the same attributes.

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		      zSeries & Storage
hare-l3A5Bk7waGM@public.gmane.org			      +49 911 74053 688
SUSE LINUX GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: J. Hawn, J. Guild, F. Imendörffer, HRB 21284 (AG Nürnberg)

Re: [PATCH 11/33] 95iscsi: Autodetect iSCSI firmware

[WANG Chao]

On 11/21/14 at 07:51am, Hannes Reinecke wrote:
> On 11/21/2014 06:43 AM, WANG Chao wrote:
> > Hi,
> > 
> > On 11/20/14 at 03:56pm, Thomas Renninger wrote:
> >> From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
> >>
> >> Some iSCSI initiator present firmware information in
> >> /sys/firmware/ibft or /sys/firmware/iscsi_bootX
> >>
> >> Whenever we detect one of those directories we should assume
> >> that the iSCSI devices should be activated.
> >>
> >> Also incorporates SUSE patches:
> >> 0049:
> >> 95iscsi: Use 'ip=ibft' for ibft autoconfiguration
> >>
> >> For iBFT autoconfiguration we should be setting 'ip=ibft'
> >> instead of rd.neednet. This should instruct dracut to only
> >> enable the iBFT interfaces and leave the rest alone.
> > 
> > I think,
> > 
> > /sys/firmware/ibft is created for iBFT if detected.
> > /sys/firmware/iscsi_bootX is created for hardware iSCSI if detected.
> > 
> Nope.
> 
> ibft is created whenever an iBFT section (either in system ROM
> or via EFI) is detected.
> iscsi_boot is created for be2iscsi devices, which use an iBFT-like
> interface.
> 
> Both provide the same attributes.

Yes, they have same attributes.

But IIRC iBFT is provided by ACPI or EFI and the other is provided by
HBA's firmware. Right?

Thanks
WANG Chao

Re: [PATCH 19/33] Restore original IFS value

[Harald Hoyer]

On 20.11.2014 15:56, Thomas Renninger wrote:
> From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
> 
> Whenever the IFS setting is modified it needs to be restored to
> its original value. Otherwise it'll lead to unexpected results.
> 
> References: bnc#883770
> 
> Signed-off-by: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
> Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
> ---
>  modules.d/90kernel-modules/insmodpost.sh   | 2 ++
>  modules.d/90kernel-modules/parse-kernel.sh | 4 ++++
>  modules.d/95iscsi/module-setup.sh          | 6 ++++--
>  modules.d/98systemd/dracut-pre-udev.sh     | 4 ++++
>  4 files changed, 14 insertions(+), 2 deletions(-)
> 
> diff --git a/modules.d/90kernel-modules/insmodpost.sh b/modules.d/90kernel-modules/insmodpost.sh
> index 4362f0a..b4bcdd0 100755
> --- a/modules.d/90kernel-modules/insmodpost.sh
> +++ b/modules.d/90kernel-modules/insmodpost.sh
> @@ -4,9 +4,11 @@
>  
>  for modlist in $(getargs rd.driver.post -d rdinsmodpost=); do
>      (
> +        OLDIFS="$IFS"
>          IFS=,
>          for m in $modlist; do
>              modprobe $m
>          done
> +        IFS="$OLDIFS"
>      )

Well, "()" ensures, that IFS is local only.
$ a=foo; echo $a; ( a=bar; echo $a); echo $a
foo
bar
foo


> diff --git a/modules.d/95iscsi/module-setup.sh b/modules.d/95iscsi/module-setup.sh
> index 9975e59..f551c49 100755
> --- a/modules.d/95iscsi/module-setup.sh
> +++ b/modules.d/95iscsi/module-setup.sh
> @@ -29,16 +29,18 @@ check() {
>  }
>  
>  get_ibft_mod() {
> -    local ibft_mac=$1
> +    local ibft_mac=$1 IFS
>      # Return the iSCSI offload module for a given MAC address
>      iscsiadm -m iface | while read iface_name iface_desc ; do
> +        OLDIFS="$IFS"
>          IFS=$','
>          set -- $iface_desc
> +        IFS="$OLDIFS"
>          if [ "$ibft_mac" = "$2" ] ; then
>              echo $1
>              return 0
>          fi
> -        unset IFS
> +        
>      done
>  }

That's a valid point.

Re: [PATCH 33/33] nfs: Add ip=... and root=nfs... parameters to internal dracut

[Harald Hoyer]

On 20.11.2014 15:57, Thomas Renninger wrote:
>  cmdline
> 
> If the rootfs is an nfs mount, also know as nfsroot, add the correct
> parameter to the dracut cmdline.
> 
> Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
> ---
>  modules.d/95nfs/module-setup.sh | 46 ++++++++++++++++++++++++++++++++++++++---
>  1 file changed, 43 insertions(+), 3 deletions(-)
> 
> diff --git a/modules.d/95nfs/module-setup.sh b/modules.d/95nfs/module-setup.sh
> index e1340b3..41c6ef7 100755
> --- a/modules.d/95nfs/module-setup.sh
> +++ b/modules.d/95nfs/module-setup.sh
> @@ -14,7 +14,6 @@ check() {
>          done
>          return 255
>      }
> -
>      return 0
>  }
>  
> @@ -29,14 +28,56 @@ installkernel() {
>      instmods nfs sunrpc ipv6 nfsv2 nfsv3 nfsv4 nfs_acl nfs_layout_nfsv41_files
>  }
>  
> +cmdline() {
> +    local nfs_device
> +    local nfs_options
> +    local nfs_root
> +    local nfs_address
> +    local lookup
> +    local ifname
> +
> +    ### nfsroot= ###
> +    nfs_device=$(findmnt -t nfs4 -n -o SOURCE /)
> +    if [ -n "$nfs_device" ];then
> +        nfs_root="root=nfs4:$nfs_device"
> +    else
> +        nfs_device=$(findmnt -t nfs -n -o SOURCE /)
> +        [ -z "$nfs_device" ] && return
> +        nfs_root="root=nfs:$nfs_device"
> +    fi
> +    nfs_options=$(findmnt -t nfs4,nfs -n -o OPTIONS /)
> +    [ -n "$nfs_options" ] && nfs_root="$nfs_root:$nfs_options"
> +    echo "$nfs_root"
> +
> +    ### ip= ###
> +    if [[ $nfs_device = [0-9]*\.[0-9]*\.[0-9]*.[0-9]* ]] || [[ $nfs_device = \[.*\] ]]; then
> +        nfs_address="$nfs_device"
> +    else
> +        lookup=$(host $(echo ${nfs_device%%:*})| head -n1)
> +        nfs_address=${lookup##* }
> +    fi
> +    ifname=$(ip -o route get to $nfs_address | sed -n 's/.*dev \([^ ]*\).*/\1/p')
> +    if [ -e /sys/class/net/$ifname/address ] ; then
> +        ifmac=$(cat /sys/class/net/$ifname/address)
> +        printf 'ifname=%s:%s ' ${ifname} ${ifmac}
> +    fi
> +
> +    printf 'ip=%s:static\n' ${ifname}
> +}
> +
>  # called by dracut
>  install() {
>      local _i
>      local _nsslibs
>      inst_multiple -o portmap rpcbind rpc.statd mount.nfs \
> -        mount.nfs4 umount rpc.idmapd sed /etc/netconfig
> +        mount.nfs4 umount rpc.idmapd sed /etc/netconfig host sed

Why "host" and "sed" ?? Just because it's used in cmdline() does not mean it
has to be installed in the initramfs?

Re: [PATCH 33/33] nfs: Add ip=... and root=nfs... parameters to internal dracut

[Thomas Renninger]

On Friday, November 21, 2014 10:26:30 AM Harald Hoyer wrote:
> On 20.11.2014 15:57, Thomas Renninger wrote:
...
> > -        mount.nfs4 umount rpc.idmapd sed /etc/netconfig
> > +        mount.nfs4 umount rpc.idmapd sed /etc/netconfig host sed
> 
> Why "host" and "sed" ?? Just because it's used in cmdline() does not mean it
> has to be installed in the initramfs?
Right and sed is already added anyway.
Corrected version will show up on re-submit.

Thanks,

      Thomas

Re: [PATCH 19/33] Restore original IFS value

[Thomas Renninger]

On Friday, November 21, 2014 10:10:54 AM Harald Hoyer wrote:
> On 20.11.2014 15:56, Thomas Renninger wrote:
> > From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
..
> > diff --git a/modules.d/95iscsi/module-setup.sh
> > b/modules.d/95iscsi/module-setup.sh index 9975e59..f551c49 100755
> > --- a/modules.d/95iscsi/module-setup.sh
> > +++ b/modules.d/95iscsi/module-setup.sh
> > @@ -29,16 +29,18 @@ check() {
> > 
> >  }
> >  
> >  get_ibft_mod() {
> > 
> > -    local ibft_mac=$1
> > +    local ibft_mac=$1 IFS
> > 
...
> >      done
> >  
> >  }
> 
> That's a valid point.

Wait, this function is touched in a later patch, nuking
IFS modifcations totally.
I removed the whole patch and adjusted the later patch now.
Which, eh..., is missing.
I had not the whole series applied when sending.
Patches 34..42 are missing.
I'll resend. No need to go through all. It's just the
ones in the end which are new.
New series will have your 2 mentioned fixes added already.

I also add the fcoe patches, will take the one or other hour...

Thanks,

      Thomas

Re: [PATCH 19/33] Restore original IFS value

[Harald Hoyer]

On 21.11.2014 11:53, Thomas Renninger wrote:> On Friday, November 21, 2014
10:10:54 AM Harald Hoyer wrote:
>> On 20.11.2014 15:56, Thomas Renninger wrote:
>>> From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
> ..
>>> diff --git a/modules.d/95iscsi/module-setup.sh
>>> b/modules.d/95iscsi/module-setup.sh index 9975e59..f551c49 100755
>>> --- a/modules.d/95iscsi/module-setup.sh
>>> +++ b/modules.d/95iscsi/module-setup.sh
>>> @@ -29,16 +29,18 @@ check() {
>>>
>>>  }
>>>
>>>  get_ibft_mod() {
>>>
>>> -    local ibft_mac=$1
>>> +    local ibft_mac=$1 IFS
>>>
> ...
>>>      done
>>>
>>>  }
>>
>> That's a valid point.
>
> Wait, this function is touched in a later patch, nuking
> IFS modifcations totally.
> I removed the whole patch and adjusted the later patch now.
> Which, eh..., is missing.

Then squash them please.

> I had not the whole series applied when sending.
> Patches 34..42 are missing.
> I'll resend. No need to go through all. It's just the
> ones in the end which are new.
> New series will have your 2 mentioned fixes added already.
>
> I also add the fcoe patches, will take the one or other hour...
>
> Thanks,
>
>       Thomas
>

Re: [PATCH 09/33] 90multipath: Load device_handler modules early during boot

[Harald Hoyer]

On 20.11.2014 15:56, Thomas Renninger wrote:
> From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
> 
> The device handler modules need to be loaded early during boot
> to avoid I/O errors being printed to the system log.
> 
> References: bnc#871617
> 
> Signed-off-by: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
> Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
> ---
>  modules.d/90multipath/module-setup.sh | 14 ++++++++++++++
>  1 file changed, 14 insertions(+)
> 
> diff --git a/modules.d/90multipath/module-setup.sh b/modules.d/90multipath/module-setup.sh
> index 7a60b2e..bbee5d9 100755
> --- a/modules.d/90multipath/module-setup.sh
> +++ b/modules.d/90multipath/module-setup.sh
> @@ -28,6 +28,15 @@ depends() {
>  }
>  
>  # called by dracut
> +cmdline() {
> +    for m in scsi_dh_alua scsi_dh_emc scsi_dh_rdac ; do
> +        if module_is_host_only $m ; then
> +            printf 'rd.driver.pre=%s ' "$m"
> +        fi
> +    done
> +}
> +
> +# called by dracut
>  installkernel() {
>      local _ret
>      local _arch=$(uname -m)
> @@ -89,6 +98,11 @@ install() {
>      inst_libdir_file "libmultipath*" "multipath/*"
>      inst_libdir_file 'libgcc_s.so*'
>  
> +    if [[ $hostonly_cmdline ]] ; then
> +        local _conf=$(cmdline)
> +        [[ $_conf ]] && echo "$_conf" >> "${initdir}/etc/cmdline.d/90multipath.conf"
> +    fi
> +
>      if dracut_module_included "systemd"; then
>          inst_multiple \
>              $systemdsystemunitdir/multipathd.service
> 

applied, thanks!

Re: [PATCH 08/33] 90multipath: Install libgcc_s library

[Harald Hoyer]

On 20.11.2014 15:56, Thomas Renninger wrote:
> From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
> 
> multipathd is using pthreads, which require libgcc_s for
> pthread_cancel to work. Without it multipathd might crash
> with SIGABRT.
> 
> Signed-off-by: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
> Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
> ---
>  modules.d/90multipath/module-setup.sh | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/modules.d/90multipath/module-setup.sh b/modules.d/90multipath/module-setup.sh
> index 3c2e3a9..7a60b2e 100755
> --- a/modules.d/90multipath/module-setup.sh
> +++ b/modules.d/90multipath/module-setup.sh
> @@ -87,6 +87,7 @@ install() {
>      inst $(command -v partx) /sbin/partx
>  
>      inst_libdir_file "libmultipath*" "multipath/*"
> +    inst_libdir_file 'libgcc_s.so*'
>  
>      if dracut_module_included "systemd"; then
>          inst_multiple \
> 

applied, thanks!

Re: [PATCH 10/33] 90multipath: install dracut-specific service file

[Harald Hoyer]

On 20.11.2014 15:56, Thomas Renninger wrote:
> From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
> 
> The multipathd package might install a service and a socket
> file. Using the original service file from the installed
> system without the socket file triggers a bug in systemd,
> causing systemd to crash.
> 
> As we don't actually need to socket file in the initrd we
> should be installing our own service file which does not
> reference the socket file at all.
> 
> References: bnc#871610
> 
> Signed-off-by: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
> Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
> ---
>  modules.d/90multipath/module-setup.sh    |  7 +++----
>  modules.d/90multipath/multipathd.service | 15 +++++++++++++++
>  2 files changed, 18 insertions(+), 4 deletions(-)
>  create mode 100644 modules.d/90multipath/multipathd.service
> 
> diff --git a/modules.d/90multipath/module-setup.sh b/modules.d/90multipath/module-setup.sh
> index bbee5d9..0f88f82 100755
> --- a/modules.d/90multipath/module-setup.sh
> +++ b/modules.d/90multipath/module-setup.sh
> @@ -104,10 +104,9 @@ install() {
>      fi
>  
>      if dracut_module_included "systemd"; then
> -        inst_multiple \
> -            $systemdsystemunitdir/multipathd.service
> -        mkdir -p "${initdir}${systemdsystemconfdir}/sysinit.target.wants"
> -        ln -rfs "${initdir}${systemdsystemunitdir}/multipathd.service" "${initdir}${systemdsystemconfdir}/sysinit.target.wants/multipathd.service"
> +        inst_simple "${moddir}/multipathd.service" "${systemdsystemunitdir}/multipathd.service"
> +        mkdir -p "${initdir}${systemdsystemunitdir}/sysinit.target.wants"
> +        ln -rfs "${initdir}${systemdsystemunitdir}/multipathd.service" "${initdir}${systemdsystemunitdir}/sysinit.target.wants/multipathd.service"
>      else
>          inst_hook pre-trigger 02 "$moddir/multipathd.sh"
>          inst_hook cleanup   02 "$moddir/multipathd-stop.sh"
> diff --git a/modules.d/90multipath/multipathd.service b/modules.d/90multipath/multipathd.service
> new file mode 100644
> index 0000000..44a7b80
> --- /dev/null
> +++ b/modules.d/90multipath/multipathd.service
> @@ -0,0 +1,15 @@
> +[Unit]
> +Description=Device-Mapper Multipath Device Controller
> +Before=iscsi.service iscsid.service lvm2-activation-early.service
> +DefaultDependencies=no
> +Conflicts=shutdown.target
> +
> +[Service]
> +Type=simple
> +ExecStartPre=/sbin/modprobe dm-multipath
> +ExecStart=/sbin/multipathd -s -d
> +ExecReload=/sbin/multipathd reconfigure
> +ExecStop=/sbin/multipathd shutdown
> +
> +[Install]
> +WantedBy=sysinit.target
> 

applied, thanks!

Also added a patch to turn off the service via kernel command line options.

Re: [PATCH 02/33] Update dracut manpage for GRUB Legacy and GRUB2

[Dave Young]

On 11/20/14 at 03:56pm, Thomas Renninger wrote:
> From: Shawn Dunn <sfalken-stAJ6ESoqRxg9hUCZPvPmw@public.gmane.org>
> 
> Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
> ---
>  dracut.usage.asc | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/dracut.usage.asc b/dracut.usage.asc
> index 9e788c1..1a79638 100644
> --- a/dracut.usage.asc
> +++ b/dracut.usage.asc
> @@ -99,9 +99,9 @@ raid with encryption and LVM on top), as long as you specify the correct
>  filesystem LABEL or UUID on the kernel command line for your root device, dracut
>  will find it and boot from it.
>  
> -The kernel command line usually can be configured in _/boot/grub/grub.conf_ or
> -_/boot/grub2/grub.cfg_, if grub is your bootloader and it also can be edited in
> -the real boot process in the grub menu.
> +The kernel command line usually can be configured in _/boot/grub/grub.conf_,
> +for GRUB legacy, or _/boot/grub2/grub.cfg_, for GRUB2 if grub is your
> +bootloader and it also can be edited in the real boot process in the grub menu.

In case you are enhance the documentation, how about add case for uefi boot

>  
>  The kernel command line can also be provided by the dhcp server with the
>  root-path option. See <<NetworkBoot>>.
> -- 
> 1.8.5.2
> 
> --
> To unsubscribe from this list: send the line "unsubscribe initramfs" in
> the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 03/33] 00warpclock: Set correct timezone

[Dave Young]

Ccing Chao who was handling the Fedora kdump timezone issues..

On 11/20/14 at 03:56pm, Thomas Renninger wrote:
> From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
> 
> Add module for setting correct timezone.
> References: bnc#830060
> 
> Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
> ---
>  modules.d/00warpclock/module-setup.sh | 25 +++++++++++++++++++++++++
>  modules.d/00warpclock/warpclock.sh    |  9 +++++++++
>  2 files changed, 34 insertions(+)
>  create mode 100755 modules.d/00warpclock/module-setup.sh
>  create mode 100755 modules.d/00warpclock/warpclock.sh
> 
> diff --git a/modules.d/00warpclock/module-setup.sh b/modules.d/00warpclock/module-setup.sh
> new file mode 100755
> index 0000000..6bd7abd
> --- /dev/null
> +++ b/modules.d/00warpclock/module-setup.sh
> @@ -0,0 +1,25 @@
> +#!/bin/bash
> +
> +# called by dracut
> +check() {
> +    # hwclock does not exist on S390(x), bail out silently then
> +    local _arch=$(uname -m)
> +    [ "$_arch" = "s390" -o "$_arch" = "s390x" ] && return 1
> +
> +    [ -e /etc/localtime -a -e /etc/adjtime ] || return 1
> +    require_binaries /sbin/hwclock || return 1
> +}
> +
> +# called by dracut
> +depends() {
> +    return 0
> +}
> +
> +# called by dracut
> +install() {
> +    inst /usr/share/zoneinfo/UTC
> +    inst /etc/localtime
> +    inst /etc/adjtime
> +    inst_hook pre-trigger 00 "$moddir/warpclock.sh"
> +    inst /sbin/hwclock
> +}
> diff --git a/modules.d/00warpclock/warpclock.sh b/modules.d/00warpclock/warpclock.sh
> new file mode 100755
> index 0000000..f64818c
> --- /dev/null
> +++ b/modules.d/00warpclock/warpclock.sh
> @@ -0,0 +1,9 @@
> +#!/bin/sh
> +
> +if test -e /etc/adjtime ; then
> +    while read line ; do
> +	if test "$line" = LOCAL ; then
> +	    hwclock --systz
> +	fi
> +    done < /etc/adjtime
> +fi
> -- 
> 1.8.5.2
> 
> --
> To unsubscribe from this list: send the line "unsubscribe initramfs" in
> the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 03/33] 00warpclock: Set correct timezone

[WANG Chao]

Hi,

On 11/20/14 at 03:56pm, Thomas Renninger wrote:
> From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
> 
> Add module for setting correct timezone.
> References: bnc#830060

I used to implement such functionality but it's reverted later:

commit d27cd4d
Author: Harald Hoyer <harald-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
Date:   Wed Jul 31 15:55:13 2013 +0200

    Revert "base: setup correct system time and time zone in initrd"

    This reverts commit
    77364fd6bb01ea184db73b1a5ad05da984752293.

    Too many problems with adjtime in the initramfs

    https://bugzilla.redhat.com/show_bug.cgi?id=981617

After that what we do for kdump is to copy /etc/localtime and
/etc/adjtime to initramfs ourselves. In 2nd kernel, systemd will take
care of time zone setup according to these two files.

Thanks
WANG Chao

> 
> Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
> ---
>  modules.d/00warpclock/module-setup.sh | 25 +++++++++++++++++++++++++
>  modules.d/00warpclock/warpclock.sh    |  9 +++++++++
>  2 files changed, 34 insertions(+)
>  create mode 100755 modules.d/00warpclock/module-setup.sh
>  create mode 100755 modules.d/00warpclock/warpclock.sh
> 
> diff --git a/modules.d/00warpclock/module-setup.sh b/modules.d/00warpclock/module-setup.sh
> new file mode 100755
> index 0000000..6bd7abd
> --- /dev/null
> +++ b/modules.d/00warpclock/module-setup.sh
> @@ -0,0 +1,25 @@
> +#!/bin/bash
> +
> +# called by dracut
> +check() {
> +    # hwclock does not exist on S390(x), bail out silently then
> +    local _arch=$(uname -m)
> +    [ "$_arch" = "s390" -o "$_arch" = "s390x" ] && return 1
> +
> +    [ -e /etc/localtime -a -e /etc/adjtime ] || return 1
> +    require_binaries /sbin/hwclock || return 1
> +}
> +
> +# called by dracut
> +depends() {
> +    return 0
> +}
> +
> +# called by dracut
> +install() {
> +    inst /usr/share/zoneinfo/UTC
> +    inst /etc/localtime
> +    inst /etc/adjtime
> +    inst_hook pre-trigger 00 "$moddir/warpclock.sh"
> +    inst /sbin/hwclock
> +}
> diff --git a/modules.d/00warpclock/warpclock.sh b/modules.d/00warpclock/warpclock.sh
> new file mode 100755
> index 0000000..f64818c
> --- /dev/null
> +++ b/modules.d/00warpclock/warpclock.sh
> @@ -0,0 +1,9 @@
> +#!/bin/sh
> +
> +if test -e /etc/adjtime ; then
> +    while read line ; do
> +	if test "$line" = LOCAL ; then
> +	    hwclock --systz
> +	fi
> +    done < /etc/adjtime
> +fi
> -- 
> 1.8.5.2
> 
> --
> To unsubscribe from this list: send the line "unsubscribe initramfs" in
> the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 02/33] Update dracut manpage for GRUB Legacy and GRUB2

[Thomas Renninger]

On Wednesday, November 26, 2014 01:32:16 PM Dave Young wrote:
> On 11/20/14 at 03:56pm, Thomas Renninger wrote:
> > From: Shawn Dunn <sfalken-stAJ6ESoqRxg9hUCZPvPmw@public.gmane.org>
> > 
> > Signed-off-by: Thomas Renninger <trenn-l3A5Bk7waGM@public.gmane.org>
> > ---
> > 
> >  dracut.usage.asc | 6 +++---
> >  1 file changed, 3 insertions(+), 3 deletions(-)
> > 
> > diff --git a/dracut.usage.asc b/dracut.usage.asc
> > index 9e788c1..1a79638 100644
> > --- a/dracut.usage.asc
> > +++ b/dracut.usage.asc
> > @@ -99,9 +99,9 @@ raid with encryption and LVM on top), as long as you
> > specify the correct> 
> >  filesystem LABEL or UUID on the kernel command line for your root device,
> >  dracut will find it and boot from it.
> > 
> > -The kernel command line usually can be configured in
> > _/boot/grub/grub.conf_ or -_/boot/grub2/grub.cfg_, if grub is your
> > bootloader and it also can be edited in -the real boot process in the
> > grub menu.
> > +The kernel command line usually can be configured in
> > _/boot/grub/grub.conf_, +for GRUB legacy, or _/boot/grub2/grub.cfg_, for
> > GRUB2 if grub is your +bootloader and it also can be edited in the real
> > boot process in the grub menu.
> In case you are enhance the documentation, how about add case for uefi boot

I am not aware of any UEFI specialities.
This also sounds as if a seperate submit makes sense.
It's also a non trivial task to get all these patches in line and submitted.
In this case adjusting is unneeded overhead, please submit the changes
you have in mind on top.
If you think the patch is not worth it, leave it out.

Thanks,

   Thomas

Re: [PATCH 11/33] 95iscsi: Autodetect iSCSI firmware

[Thomas Renninger]

On Friday, November 21, 2014 03:32:55 PM WANG Chao wrote:
> On 11/21/14 at 07:51am, Hannes Reinecke wrote:
> > On 11/21/2014 06:43 AM, WANG Chao wrote:
> > > Hi,
> > > 
> > > On 11/20/14 at 03:56pm, Thomas Renninger wrote:
> > >> From: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>
> > >> 
> > >> Some iSCSI initiator present firmware information in
> > >> /sys/firmware/ibft or /sys/firmware/iscsi_bootX
> > >> 
> > >> Whenever we detect one of those directories we should assume
> > >> that the iSCSI devices should be activated.
> > >> 
> > >> Also incorporates SUSE patches:
> > >> 0049:
> > >> 95iscsi: Use 'ip=ibft' for ibft autoconfiguration
> > >> 
> > >> For iBFT autoconfiguration we should be setting 'ip=ibft'
> > >> instead of rd.neednet. This should instruct dracut to only
> > >> enable the iBFT interfaces and leave the rest alone.
> > > 
> > > I think,
> > > 
> > > /sys/firmware/ibft is created for iBFT if detected.
> > > /sys/firmware/iscsi_bootX is created for hardware iSCSI if detected.
> > 
> > Nope.
> > 
> > ibft is created whenever an iBFT section (either in system ROM
> > or via EFI) is detected.
> > iscsi_boot is created for be2iscsi devices, which use an iBFT-like
> > interface.
> > 
> > Both provide the same attributes.
> 
> Yes, they have same attributes.

The same sysfs attributes?

> But IIRC iBFT is provided by ACPI or EFI and the other is provided by
> HBA's firmware. Right?

But this should not matter.
If the kernel differs both APIs and provides a unified sysfs interface,
not differing in userspace again would be the right thing to do.

     Thomas

PS: I continue to collect missing patches tomorrow and do a big resend...