- Applied backport patches from s390-tools 2.37 to 2.36 ( jsc#PED-11870 )

( jsc#IBM-1447, jsc#IBM-1062 )
  * s390-tools-General-update-01.patch
  * s390-tools-General-update-02.patch
  * s390-tools-General-update-03.patch
  * s390-tools-General-update-04.patch
  * s390-tools-General-update-05.patch
  * s390-tools-General-update-06.patch
  * s390-tools-General-update-07.patch
  * s390-tools-General-update-08.patch
  * s390-tools-General-update-09.patch
  * s390-tools-General-update-10.patch
  * s390-tools-General-update-11.patch
  * s390-tools-General-update-12.patch
  * s390-tools-Additional-update-01.patch
  * s390-tools-Additional-update-02.patch
    ( jsc#IBM-1570, jsc#IBM-1571 )
  * s390-tools-Support-unencrypted-SE-images-01.patch
    ( jsc#IBM-1572, jsc#IBM-1573 )
  * s390-tools-pvimg-info-command-01.patch
  * s390-tools-pvimg-info-command-02.patch
  * s390-tools-pvimg-info-command-03.patch
  * s390-tools-pvimg-info-command-04.patch
    ( jsc#IBM-1576, jsc#IBM-1577 )
  * s390-tools-pvimg-additional-01.patch
- Renamed patches from - to
  * s390-tools-01-opticsmon-Fix-runaway-loop-in-on_link_change.patch 
      to
      s390-tools-Additional-update-01.patch
  * s390-tools-02-libzpci-opticsmon-Refactor-on_link_change-using-new.patch
      to
      s390-tools-Additional-update-02.patch
  * s390-tools-03-rust-pvimg-Add-enable-disable-image-encryption-flags-to-pvimg-create.patch
      to
      s390-tools-Support-unencrypted-SE-images-01.patch
- Revendored vendor.tar.gz

OBS-URL: https://build.opensuse.org/package/show/Base:System/s390-tools?expand=0&rev=243
This commit is contained in:
Nikolay Gueorguiev 2025-01-09 11:12:54 +00:00 committed by Git OBS Bridge
commit a45e3d9c5b
123 changed files with 23458 additions and 0 deletions

23
.gitattributes vendored Normal file
View File

@ -0,0 +1,23 @@
## Default LFS
*.7z filter=lfs diff=lfs merge=lfs -text
*.bsp filter=lfs diff=lfs merge=lfs -text
*.bz2 filter=lfs diff=lfs merge=lfs -text
*.gem filter=lfs diff=lfs merge=lfs -text
*.gz filter=lfs diff=lfs merge=lfs -text
*.jar filter=lfs diff=lfs merge=lfs -text
*.lz filter=lfs diff=lfs merge=lfs -text
*.lzma filter=lfs diff=lfs merge=lfs -text
*.obscpio filter=lfs diff=lfs merge=lfs -text
*.oxt filter=lfs diff=lfs merge=lfs -text
*.pdf filter=lfs diff=lfs merge=lfs -text
*.png filter=lfs diff=lfs merge=lfs -text
*.rpm filter=lfs diff=lfs merge=lfs -text
*.tbz filter=lfs diff=lfs merge=lfs -text
*.tbz2 filter=lfs diff=lfs merge=lfs -text
*.tgz filter=lfs diff=lfs merge=lfs -text
*.ttf filter=lfs diff=lfs merge=lfs -text
*.txz filter=lfs diff=lfs merge=lfs -text
*.whl filter=lfs diff=lfs merge=lfs -text
*.xz filter=lfs diff=lfs merge=lfs -text
*.zip filter=lfs diff=lfs merge=lfs -text
*.zst filter=lfs diff=lfs merge=lfs -text

1
.gitignore vendored Normal file
View File

@ -0,0 +1 @@
.osc

13
59-graf.rules.opensuse Normal file
View File

@ -0,0 +1,13 @@
#
# Rules for unique 3270 device nodes created in /dev/3270/
# This file should be installed in /usr/lib/udev/rules.d
#
SUBSYSTEM!="ccw", GOTO="graf_end"
DRIVER!="3270", GOTO="graf_end"
# Configure 3270 device
ACTION=="add", SUBSYSTEM=="ccw", PROGRAM="/usr/sbin/chccwdev -e $kernel"
ACTION=="remove", SUBSYSTEM=="ccw", PROGRAM="/usr/sbin/chccwdev -d $kernel"
LABEL="graf_end"

13
59-graf.rules.suse Normal file
View File

@ -0,0 +1,13 @@
#
# Rules for unique 3270 device nodes created in /dev/3270/
# This file should be installed in /usr/lib/udev/rules.d
#
SUBSYSTEM!="ccw", GOTO="graf_end"
DRIVER!="3270", GOTO="graf_end"
# Configure 3270 device
ACTION=="add", SUBSYSTEM=="ccw", PROGRAM="/sbin/chccwdev -e $kernel"
ACTION=="remove", SUBSYSTEM=="ccw", PROGRAM="/sbin/chccwdev -d $kernel"
LABEL="graf_end"

5
59-prng.rules Normal file
View File

@ -0,0 +1,5 @@
#
# Rule for prandom character device node permissions
# This file should be installed in /usr/lib/udev/rules.d
#
ACTION=="add", SUBSYSTEM=="misc", KERNEL=="prandom", MODE="0444"

23
59-zfcp-compat.rules Normal file
View File

@ -0,0 +1,23 @@
# Rules for creating the ID_PATH for SCSI devices based on the CCW bus
# using the form: ccw-<BUS_ID>-zfcp-<WWPN>:<LUN>
#
ACTION=="remove", GOTO="zfcp_scsi_device_end"
#
# Set environment variable "ID_ZFCP_BUS" to "1" if the devices
# (both disk and partition) are SCSI devices based on FCP devices
#
KERNEL=="sd*", SUBSYSTEMS=="ccw", DRIVERS=="zfcp", ENV{.ID_ZFCP_BUS}="1"
# For SCSI disks
KERNEL=="sd*[!0-9]", SUBSYSTEMS=="scsi", \
ENV{.ID_ZFCP_BUS}=="1", ENV{DEVTYPE}=="disk", \
SYMLINK+="disk/by-path/ccw-$attr{hba_id}-zfcp-$attr{wwpn}:$attr{fcp_lun}"
# For partitions on a SCSI disk
KERNEL=="sd*[0-9]", SUBSYSTEMS=="scsi", \
ENV{.ID_ZFCP_BUS}=="1", ENV{DEVTYPE}=="partition", \
SYMLINK+="disk/by-path/ccw-$attr{hba_id}-zfcp-$attr{wwpn}:$attr{fcp_lun}-part%n"
LABEL="zfcp_scsi_device_end"

25
90-s390-tools.conf Normal file
View File

@ -0,0 +1,25 @@
#
# Copyright (c) 2016 SUSE LINUX GmbH, Nuernberg, Germany.
# All rights reserved.
#
# Please don't edit this file. Place your settings into
# /etc/modprobe.d/99-local.conf instead.
#
# The dasd_diag_mod kernel module will not function properly
# unless the dasd_fba_mod module is also loaded. However,
# there are no cross-module symbol dependencies that would
# cause and entry to be placed in
# /lib/modules/$(uname -r)/modules.dep
# So, we're adding this "soft" dependency here to make sure that
# any time dasd_diag_mod gets loaded, so will dasd_fba_mod.
#
# Additionally, DASD devices that are supposed to be used in
# DIAG250 mode will have problems because as far as the kernel
# is concerned, and hence udev, the driver is dasd_fba_mod. So,
# we need to also have the reverse dependency so that when
# dasd_fba_mod gets loaded, so will dasd_diag_mod. This will
# prevent problems that would show up in the system log as:
# Setting the DASD online failed because of missing DIAG discipline
#
softdep dasd_diag_mod pre: dasd_fba_mod
softdep dasd_fba_mod pre: dasd_diag_mod

57
README.SUSE.opensuse Normal file
View File

@ -0,0 +1,57 @@
ls - Addons by SUSE
The following utility and its man page have been added to make it
easier to determine the machine type on which Linux is running.
* cputype
Usage: cputype
The following utilities and their man pages have been added by SUSE to
ease the activation and deactivation of devices. These scripts are also
used by YaST. Functionality not provided by these scripts cannot be
provided by YaST.
These scripts also create/delete the needed udev rules.
Detailed information on some parameters are in the
"Device Drivers, Features and Commands" for this release.
General parameters
channel numbers are with lower letters
parameters switching things on or off are
1 for on and 0 for off
* ctc_configure
Usage: /usr/sbin/ctc_configure <read channel> <write channel> <online> [<protocol>]
To configure CTC connections
Valid Parameters for the protocal are 0, 1 and 3
For a detailed explanation please look in the Device Driver book
* dasd_configure
Usage: dasd_configure <ccwid> <online> <use_diag>
To set DASDs online/offline
The use_diag makes only sense under z/VM. In an
LPAR just set it to 0
* iucv_configure
Usage: /usr/sbin/iucv_configure <router> <online>
To set an IUCV IP-network online/offline
* qeth_configure
Usage: /usr/sbin/qeth_configure [options] <read chan> <write chan> <control chan> <online>
Set qeth, hipersocket adapter online/offline.
options could be one of the following:
-i Configure IP takeover
-l Configure Layer2 support
-p NAME QETH Portname to use
-n 1/0 QETH port number to use
* zfcp_disk_configure
Usage: /usr/sbin/zfcp_disk_configure <ccwid> <wwpn> <lun> <online>
set a disk online/offline. This require that the repective
Adapter is online. See command below.
* zfcp_host_configure
Usage: /usr/sbin/zfcp_host_configure <ccwid> <online>
Set a zfcp Adapter online/offline

57
README.SUSE.suse Normal file
View File

@ -0,0 +1,57 @@
ls - Addons by SUSE
The following utility and its man page have been added to make it
easier to determine the machine type on which Linux is running.
* cputype
Usage: cputype
The following utilities and their man pages have been added by SUSE to
ease the activation and deactivation of devices. These scripts are also
used by YaST. Functionality not provided by these scripts cannot be
provided by YaST.
These scripts also create/delete the needed udev rules.
Detailed information on some parameters are in the
"Device Drivers, Features and Commands" for this release.
General parameters
channel numbers are with lower letters
parameters switching things on or off are
1 for on and 0 for off
* ctc_configure
Usage: /sbin/ctc_configure <read channel> <write channel> <online> [<protocol>]
To configure CTC connections
Valid Parameters for the protocal are 0, 1 and 3
For a detailed explanation please look in the Device Driver book
* dasd_configure
Usage: dasd_configure <ccwid> <online> <use_diag>
To set DASDs online/offline
The use_diag makes only sense under z/VM. In an
LPAR just set it to 0
* iucv_configure
Usage: /sbin/iucv_configure <router> <online>
To set an IUCV IP-network online/offline
* qeth_configure
Usage: /sbin/qeth_configure [options] <read chan> <write chan> <control chan> <online>
Set qeth, hipersocket adapter online/offline.
options could be one of the following:
-i Configure IP takeover
-l Configure Layer2 support
-p NAME QETH Portname to use
-n 1/0 QETH port number to use
* zfcp_disk_configure
Usage: /sbin/zfcp_disk_configure <ccwid> <wwpn> <lun> <online>
set a disk online/offline. This require that the repective
Adapter is online. See command below.
* zfcp_host_configure
Usage: /sbin/zfcp_host_configure <ccwid> <online>
Set a zfcp Adapter online/offline

8
_service Normal file
View File

@ -0,0 +1,8 @@
<services>
<service name="cargo_vendor" mode="manual">
<param name="srctar">s390-tools-2.29.0.tar.gz</param>
<param name="compression">zst</param>
<param name="update">true</param>
</service>
<service name="cargo_audit" mode="manual" />
</services>

126
appldata Normal file
View File

@ -0,0 +1,126 @@
#!/bin/sh
# Copyright (c) 2003 SUSE LINUX AG Nuernberg, Germany.
#
# Submit feedback to http://www.suse.de/feedback/
# Local settings
LOCKFILE=/var/lock/appldata
CONFIGFILE=/etc/sysconfig/appldata
# Source config file
if [ -f $CONFIGFILE ]; then
. $CONFIGFILE
else
echo "No config file found (should be $CONFIGFILE)."
exit 1
fi
RETVAL=0
start() {
echo "Starting \"Linux - z/VM Monitor Stream\" ..."
echo -n "(interval $APPLDATA_INTERVAL milliseconds) "
echo $APPLDATA_INTERVAL > /proc/sys/appldata/interval
if [ "$APPLDATA_MEM" = "yes" ]; then
if [ ! -e /proc/sys/appldata/mem ]; then
echo -n "(mem) "
modprobe appldata_mem 2>&1
if [ "$?" -ne 0 ] ; then
exit 1
else
echo 1 > /proc/sys/appldata/mem
fi
fi
fi
if [ "$APPLDATA_OS" = "yes" ]; then
if [ ! -e /proc/sys/appldata/os ]; then
echo -n "(os) "
modprobe appldata_os 2>&1
if [ "$?" -ne 0 ]; then
exit 1
else
echo 1 > /proc/sys/appldata/os
fi
fi
fi
if [ "$APPLDATA_NET_SUM" = "yes" ]; then
if [ ! -e /proc/sys/appldata/net_sum ]; then
echo -n "(net_sum) "
modprobe appldata_net_sum 2>&1
if [ "$?" -ne 0 ]; then
exit 1
else
echo 1 > /proc/sys/appldata/net_sum
fi
fi
fi
echo -n "(timer)"
echo 1 > /proc/sys/appldata/timer
touch $LOCKFILE
}
stop() {
echo -n "Stopping \"Linux - z/VM Monitor Stream\" "
echo -n "(timer"
echo 0 > /proc/sys/appldata/timer
if [ -e /proc/sys/appldata/mem ]; then
echo -n ",mem"
echo 0 > /proc/sys/appldata/mem
rmmod appldata_mem
fi
if [ -e /proc/sys/appldata/os ]; then
echo -n ",os"
echo 0 > /proc/sys/appldata/os
rmmod appldata_os
fi
if [ -e /proc/sys/appldata/net_sum ]; then
echo -n ",net_sum"
echo 0 > /proc/sys/appldata/net_sum
rmmod appldata_net_sum
fi
echo -n ")"
rm -f $LOCKFILE
}
status() {
echo "\"Linux - z/VM Monitor Stream\" status..."
echo -n "interval "
cat /proc/sys/appldata/interval
echo -n "timer "
cat /proc/sys/appldata/timer
echo -n "mem "
if [ -e /proc/sys/appldata/mem ]; then
cat /proc/sys/appldata/mem
else
echo 0
fi
echo -n "os "
if [ -e /proc/sys/appldata/os ]; then
cat /proc/sys/appldata/os
else
echo 0
fi
echo -n "net_sum "
if [ -e /proc/sys/appldata/net_sum ]; then
cat /proc/sys/appldata/net_sum
else
echo 0
fi
}
# How are we called?
case "$1" in
start)
start
;;
stop)
stop
;;
status)
status
;;
*)
RETVAL=1
esac
exit $RETVAL

17
appldata.service Normal file
View File

@ -0,0 +1,17 @@
[Unit]
Description=Linux - z/VM Monitor Stream
After=network-online.target remote-fs.target
Wants=network-online.target remote-fs.target
ConditionPathExists=/proc/sys/appldata/interval
ConditionPathExists=!/var/lock/appldata
[Service]
Type=oneshot
RemainAfterExit=yes
ExecStart=/usr/lib/systemd/scripts/appldata start
ExecStartPost=/usr/lib/systemd/scripts/appldata status
ExecStop=/usr/lib/systemd/scripts/appldata stop
[Install]
WantedBy=default.target

5
cargo_config Normal file
View File

@ -0,0 +1,5 @@
[source.crates-io]
replace-with = "vendored-sources"
[source.vendored-sources]
directory = "vendor/"

13
cio_ignore.service Normal file
View File

@ -0,0 +1,13 @@
[Unit]
Description=Setup devices for cio_ignore
DefaultDependencies=no
Before=local-fs.target
ConditionKernelCommandLine=cio_ignore
[Service]
Type=oneshot
RemainAfterExit=yes
ExecStart=/usr/lib/systemd/scripts/setup_cio_ignore.sh
[Install]
WantedBy=sysinit.target

73
cputype Normal file
View File

@ -0,0 +1,73 @@
#!/bin/sh
#
# cputype
#
# Copyright (c) 2014-2017, 2019, 2023 SUSE LINUX GmbH, Nuernberg, Germany.
#
# Based on the IBM machine model, returns a (hopefully) human understandable
# string that identifies the processor.
#
# Usage:
# cputype
#
# Return values:
# 1 The script was executed on a system that is a non-IBM mainframe
# architecture
# 2 The search for the machine type in /proc/cpuinfo returned a null string
# 3 The parsing of the machine type returned a null string
# 4 The machine type found is (probably) a new one, and the script needs to
# be updated to handle it.
#
architecture=$(/bin/uname -m)
if [ "${architecture}" != "s390x" -a "${architecture}" != "s390" ]; then
echo "This command is only useful on IBM mainframes." >&2
exit 1
fi
args=$(/usr/bin/grep machine /proc/cpuinfo | awk '{print $11}' )
if [ -z "${args}" ]; then
echo "I couldn't find the machine type. Please report a bug with this output:" >&2
/bin/cat /proc/cpuinfo >&2
echo "******************" >&2
/usr/bin/grep machine /proc/cpuinfo >&2
exit 2
fi
machine=${args:0:4}
if [ -z "${machine}" ] ; then
echo "The machine type came out null. Please report a bug with this output:" >&2
/bin/cat /proc/cpuinfo >&2
exit 3
fi
case "${machine}" in
2064) echo "${machine} = z900 IBM eServer zSeries 900" ;;
2066) echo "${machine} = z800 IBM eServer zSeries 800" ;;
2084) echo "${machine} = z990 IBM eServer zSeries 990" ;;
2086) echo "${machine} = z890 IBM eServer zSeries 890" ;;
2094) echo "${machine} = z9-EC IBM System z9 Enterprise Class" ;;
2096) echo "${machine} = z9-BC IBM System z9 Business Class" ;;
2097) echo "${machine} = z10-EC IBM System z10 Enterprise Class" ;;
2098) echo "${machine} = z10-BC IBM System z10 Business Class" ;;
2817) echo "${machine} = z196 IBM zEnterprise 196" ;;
2818) echo "${machine} = z114 IBM zEnterprise 114" ;;
2827) echo "${machine} = z12-EC IBM zEnterprise EC12" ;;
2828) echo "${machine} = z12-BC IBM zEnterprise BC12" ;;
2964) echo "${machine} = z13 IBM z13" ;;
2965) echo "${machine} = z13s IBM z13s (single frame)" ;;
3906) echo "${machine} = z14 IBM z14" ;;
3907) echo "${machine} = z14 ZR1 IBM z14 ZR1" ;;
8561) echo "${machine} = z15 T01 IBM z15 T01" ;;
8562) echo "${machine} = z15 T02 IBM z15 T02" ;;
3931) echo "${machine} = z16 A01 IBM z16 A01" ;;
3932) echo "${machine} = z16 A02 IBM z16 A02" ;;
*) echo "An unknown machine type was reported: ${machine}" >&2
echo "Please file a bug report with this output:" >&2
/bin/cat /proc/cpuinfo >&2
exit 4
;;
esac

50
cputype.1 Normal file
View File

@ -0,0 +1,50 @@
.TH cputype 1 "April 2014" "s390-tools"
.SH NAME
cputype \- Based on the IBM machine model, returns a (hopefully) human understandable string that identifies the processor.
.SH SYNOPSIS
.B cputype
.SH DESCRIPTION
.B cputype
is intended to make it easy to find out the type of the mainframe system in use, by examining /proc/cpuinfo and converting that to the name typically known by people familiar with the IBM mainframe.
.SH PARAMETERS
.IP None
.SH FILES
.I /proc/cpuinfo
.RS
Read to determine the IBM machine model for the running system.
.RE
.SH DIAGNOSTICS
The following messages may be issued on stderr:
.IP
.B This command is only useful on IBM mainframes.
.RS
The command was executed on a system that is running on a non-IBM mainframe architecture.
Return code 1 is set.
.RE
.IP
.B I couldn't find the machine type. Please report a bug with this output:
.RS
The contents of /proc/cpuinfo are printed as well as the output from the grep command used.
Return code 2 is set.
.RE
.IP
.B The machine type came out null. Please report a bug with this output:
.RS
The contents of /proc/cpuinfo are printed. Return code 3 is set.
.RE
.IP
.B An unknown machine type was reported: mmmm
.RS
.B Please file a bug report with this output:
.RE
.RS
This is most likely seen because the command was run on a newer generation processor
and the script has not been updated with the new model number.
The contents of /proc/cpuinfo are printed. Return code 4 is set.
.RE
.SH Author
Mark Post (mpost@suse.com)
.SH Copyright
Copyright (c) 2014 SUSE LINUX Products GmbH, Nuernberg, Germany.
.SH BUGS
Gotta be some, I'm sure. If you find one, please open a bug report.

128
ctc_configure Normal file
View File

@ -0,0 +1,128 @@
#! /bin/sh
#
# ctc_configure
#
# Copyright (c) 2017 SUSE LINUX GmbH, Nuernberg, Germany.
#
# Configures a CTC device by calling the IBM-provided chzdev command.
# Whereas this script used to validate the parameters provided to it,
# we now rely on chzdev to do that instead. The script is intended only
# as a "translation layer" to provide backward compatability for older
# scripts and tools that invoke it.
#
# Usage:
# ctc_configure <read channel> <write channel> <online> [<protocol>]
#
# read/write channel = x.y.ssss where
# x is always 0 until IBM creates something that
# uses that number
# y is the logical channel subsystem (lcss) number.
# Most often this is 0, but it could be non-zero
# ssss is the four digit subchannel address of the
# device, in hexidecimal, with leading zeros.
# online = 0 to take the device offline
# 1 to bring the device online
# protocol = 0 Compatibility with peers other than OS/390®, or z/OS, for
# example, a z/VM TCP service machine. This is the default.
# 1 Enhanced package checking for Linux peers.
# 3 For compatibility with OS/390 or z/OS peers.
# 4 For MPC connections to VTAM on traditional mainframe
# operating systems.
#
# Return values:
# Return codes are determined by the chzdev command.
#
mesg () {
echo "$@"
}
debug_mesg () {
case "${DEBUG}" in
yes) mesg "$@" ;;
*) ;;
esac
}
add_cio_channel() {
echo "$* # ${DATE}" >> /boot/zipl/active_devices.txt
}
remove_cio_channel() {
[ -w /boot/zipl/active_devices.txt ] && sed -i -e "/^${1}/d" /boot/zipl/active_devices.txt
}
usage(){
echo "Usage: ${0} <read channel> <write channel> <online> [<protocol>]"
echo " read/write channel = x.y.ssss where"
echo " x is always 0 until IBM creates something that"
echo " uses that number"
echo " y is the logical channel subsystem (lcss) number."
echo " Most often this is 0, but it could be non-zero"
echo " ssss is the four digit subchannel address of the"
echo " device, in hexidecimal, with leading zeros."
echo " online = 0 to take the device offline"
echo " 1 to bring the device online"
echo " protocol = 0 Compatibility with peers other than OS/390®, or z/OS, for"
echo " example, a z/VM TCP service machine. This is the default."
echo " 1 Enhanced package checking for Linux peers."
echo " 3 For compatibility with OS/390 or z/OS peers."
echo " 4 For MPC connections to VTAM on traditional mainframe"
echo " operating systems."
}
if [ "${DEBUG}" != "yes" ]; then
DEBUG="no"
fi
DATE=$(date)
CTC_READ_CHAN=${1}
CTC_WRITE_CHAN=${2}
ON_OFF=${3}
CTC_MODE=${4}
if [ -z "${CTC_READ_CHAN}" ] || [ -z "${CTC_WRITE_CHAN}" ] || [ -z "${ON_OFF}" ]; then
mesg "You didn't specify all the needed parameters."
usage
exit 1
fi
if [ -f /sys/bus/ccw/devices/${CTC_READ_CHAN}/cutype ]; then
read CU_TYPE < /sys/bus/ccw/devices/${CTC_READ_CHAN}/cutype
else mesg "Psuedo file/sys/bus/ccw/devices/${CTC_READ_CHAN}/cutype doesn't exist."
mesg "Check to see if sysfs is mounted."
exit 1
fi
PARM_LIST=""
if [ "${CU_TYPE}" == "3088/01" ] || [ "${CU_TYPE}" == "3088/60" ]; then
DEV_TYPE="lcs"
else DEV_TYPE="ctc"
if [ -z "${CTC_MODE}" ]; then
PARM_LIST="${PARM_LIST} protocol=0"
else PARM_LIST="${PARM_LIST} protocol=${CTC_MODE}"
fi
fi
if [ "${ON_OFF}" == 0 ]; then
debug_mesg "chzdev -d ${DEV_TYPE} --no-root-update ${CTC_READ_CHAN}"
chzdev -d ${DEV_TYPE} --no-root-update ${CTC_READ_CHAN}
elif [ "${ON_OFF}" == 1 ]; then
debug_mesg "chzdev -e ${DEV_TYPE} --no-root-update ${CTC_READ_CHAN} ${PARM_LIST}"
chzdev -e ${DEV_TYPE} --no-root-update ${CTC_READ_CHAN} ${PARM_LIST}
else mesg "You must specify a 0 or a 1 for the online/offline attribute."
usage
exit 1
fi
RC=${?}
if [ ${RC} -ne 0 ]; then
exit ${RC}
fi
if [ ${ON_OFF} == 1 ]; then
add_cio_channel "${CTC_READ_CHAN},${CTC_WRITE_CHAN}"
else remove_cio_channel "${CTC_READ_CHAN}"
remove_cio_channel "${CTC_WRITE_CHAN}"
fi

155
ctc_configure.8 Normal file
View File

@ -0,0 +1,155 @@
.TH ctc_configure "8" "July 2013" "s390-tools"
.SH NAME
ctc_configure \- Configures or deconfigures a Channel-to-Channel adapter (CTC) or LAN Channel Station adapter (LCS)
.SH SYNOPSIS
.B ctc_configure read_channel write_channel online [protocol]
.SH DESCRIPTION
.B ctc_configure
is intended to make it easy to persistently add and remove IBM CTC and LCS adapters. In addition to bringing the adapter online or offline, it will also create or delete the necessary udev rules for the adapter.
.SH PARAMETERS
.IP read_channel
The device number of the read channel of the adapter. Takes the form x.y.ssss.
.IP write_channel
The device number of the write channel of the adapter.Takes the form x.y.ssss.
.RS
where
.RS
.B x
is always 0 until IBM creates something that uses that number.
.RE
.RS
.B y
is the logical channel subsystem (lcss) number. Most often this is 0, but it could be non-zero.
.RE
.RS
.B ssss
is the four digit subchannel address of the device, in hexidecimal, with leading zeros. If entered in upper/mixed case, this is automatically converted to lower case.
.RE
.RE
.RS
Keep in mind that for a CTC, the read channel needs to be coupled to the write channel of the peer, and vice versa. This does not apply to LCS adapters.
.RE
.RE
.IP online
Either a literal 1 to bring the adapter online or a literal 0 to take it offline
.IP protocol
.RS
0 Compatibility with peers other than OS/390®, or z/OS, for example, a z/VM TCP service machine. This is the default.
.RE
.RS
1 Enhanced package checking for Linux peers.
.RE
.RS
3 For compatibility with OS/390 or z/OS peers.
.RE
.RS
4 For MPC connections to VTAM on traditional mainframe operating systems.
.RE
.RS
Not needed for LCS adapters.
.SH FILES
.I /etc/udev/rules.d/51-ctcm-<ccwid>.rules
.RE
.I /etc/udev/rules.d/51-lcs-<ccwid>.rules
.RS
These files provide the udev rules necessary to activate a specific CTC or LCS.
.RE
.SH ENVIRONMENT
.IP DEBUG
If set to "yes" some minimal debugging information is output during execution.
.SH DIAGNOSTICS
The following messages may be issued on stdout:
.IP
.B /sysfs not present
.RS
The sysfs file system could not be found in /proc/mounts, so there's nothing the script can
do. Return code 1 is set.
.RE
.IP
.B Invalid device status ${ONLINE}
.RS
A value other than 0 or 1 was specified for the third parameter, online. Return code 2 is set.
.RE
.IP
.B Device ${CTC_READ_CHAN} does not exist
.RS
A non-existent <ccwid> was specified for the first parameter. Remember the x.y.ssss format is necessary. Return code 3 is set.
.RE
.IP
.B Device ${CTC_READ_CHAN} does not exist
.RS
A non-existent <ccwid> was specified for the second parameter. Remember the x.y.ssss format is necessary. Return code 4 is set.
.RE
.IP
.B Not a valid CTC device (cu ${_cutype}, dev ${_devtype})
.RS
The device number specified does not correspond to a valid CTC or LCS device type. Return code 5 is st.
.RE
.IP
.B CTC type mismatch (read: ${tmp_chan}, write: ${CCW_CHAN_GROUP})
.RS
The device number specified for the read channel has a different device type than the device number specified for the write channel. Return code 6 is set.
.RE
.IP
.B Could not load module ${CCW_CHAN_GROUP}
.RS
The kernel module for the device type failed to load. Try "dmesg" to see if there is any indication why. Return code 7 is set.
.RE
.IP
.B CCW devices grouped to different devices
.RS
The read and write channels are already grouped, but not within the same interface. Try again with different devices. Return code 8 is set.
.RE
.IP
.B Could not group ${CCW_CHAN_GROUP} devices ${CTC_READ_CHAN}/${CTC_WRITE_CHAN}
.RS
The attempt to group the read and write channels into an interface failed. Try "dmesg" to see if there is any indication why. Return code 9 is set.
.RE
.IP
.B Could not set device ${CCW_CHAN_ID} online
.RS
The attempt to bring the grouped devices online failed. Try "dmesg" to see if there is any indication why. Return code 10 is set.
.RE
.IP
.B Could not set device ${CCW_CHAN_ID} offline
.RS
The attempt to take the grouped devices offline failed. Try "dmesg" to see if there is any indication why. Return code 11 is set.
.RE
If environment variable DEBUG is set to "yes," the following messages may be issued on stdout:
.IP
.B
Configuring CTC/LCS device ${CTC_READ_CHAN}/${CTC_WRITE_CHAN}
.RS
Just a little bit of verbosity, since it just indicates that we got past certain error checks and will now try to do something useful.
.RE
.IP
.B Group is ${_ccw_groupdir}/drivers/${CCW_CHAN_GROUP}/group
.RS
Just a little bit of verbosity.
.RE
.IP
.B Setting device online
.RS
Just a little bit of verbosity.
.RE
.IP
.B Device ${CCW_CHAN_ID} is already online
.RS
An attempt was made to bring the adapter online when it was already online.
.RE
.IP
.B Setting device offline
.RS
Just a little bit of verbosity.
.RE
.IP
.B Device ${CCW_CHAN_ID} is already offline
.RS
An attempt was made to take the adapter offline when it was already offline.
.RE
.SH BUGS
Gotta be some, I'm sure. If you find one, please open a bug report.

60
dasd_configure.8 Normal file
View File

@ -0,0 +1,60 @@
.TH dasd_configure "8" "February 2013" "s390-tools"
.SH NAME
dasd_configure \- Configures or deconfigures a Direct Access Storage Device (DASD) volume.
.SH SYNOPSIS
.B dasd_configure [-f -t dasd_type ] ccwid online [use_diag]
.SH DESCRIPTION
.B dasd_configure
is intended to make it easy to persistently add and remove DASD volumes. In addition to bringing the volume online or offline, it will also create or delete the necessary udev rules for the volume.
.SH PARAMETERS
.IP -f
Force creation of udev rules, do not check values in /sys.
.IP -t
Must be either dasd-eckd or dasd-fba. Must be provided if -f is used.
.IP ccwid
The device number of the DASD volume. Takes the form x.y.ssss where
.RS
.B x
is always 0 until IBM creates something that uses that number.
.RE
.RS
.B y
is the subchannel set ID (SSID). Most often this is 0, but it could be non-zero.
.RE
.RS
.B ssss
is the four digit device address of the subchannel, in hexidecimal, with leading zeros. If entered in upper/mixed case, this is automatically converted to lower case.
.RE
.IP online
Either a literal 1 to bring the volume online or a literal 0 to take it offline
.RE
.IP use_diag
Either a literal 1 to use the DIAG driver for this device, or a literal 0 to use the "normal" driver.
.RE
.SH FILES
Please see the documentation of
.B chzdev.
.SH ENVIRONMENT
.IP DEBUG
If set to "yes" some minimal debugging information is output during execution.
.SH DIAGNOSTICS
Messages and return codes are determined by the
.B chzdev
command.
Except for:
.IP
.B Device ${CCW_CHAN_ID} is unformatted
.RS
The DASD volume was brought online, but it has not been formatted with dasdfmt. This condition is really only important for YaST to determine if it should prompt the user to decide if they want to format it or not at that point. Return code 8 is set.
.RE
If environment variable DEBUG is set to "yes," it shows the command line of the invoked
.B chzdev.
Additionally, the following messages may be issued on stdout:
.IP
.B DASD ${CCW_CHAN_ID} did not come online.
.RS
The DASD volume did not come online within the waiting time. Could not check if the DASD is formatted (see above). Return code 17 is set.
.RE
.SH BUGS
Gotta be some, I'm sure. If you find one, please open a bug report.

173
dasd_configure.opensuse Normal file
View File

@ -0,0 +1,173 @@
#! /bin/sh
#
# dasd_configure
#
# Copyright (c) 2017 SUSE LINUX GmbH, Nuernberg, Germany.
#
# Configures a DASD device by calling the IBM-provided chzdev command.
# Whereas this script used to validate the parameters provided to it,
# we now rely on chzdev to do that instead. The script is intended only
# as a "translation layer" to provide backward compatability for older
# scripts and tools that invoke it.
#
# Usage:
# dasd_configure [-f -t <dasd_type> ] <ccwid> <online> [use_diag]
#
# -f Override safety checks
# -t DASD type. Must be provided if -f is used. Only dasd-eckd and
# dasd-fba are supported - Deprecated
# ccwid = x.y.ssss where
# x is always 0 until IBM creates something that uses that number
# y is the subchannel set ID (SSID). Most often
# this is 0, but it could be non-zero
# ssss is the four digit device address of the subchannel, in
# hexidecimal, with leading zeros.
# online = 0 to take the device offline
# 1 to bring the device online
# use_diag = 0 to _not_ use z/VM DIAG250 I/O, which is the default
# 1 to use z/VM DIAG250 I/O
#
# Return values:
# Return codes are determined by the chzdev command, with one exception: If a
# DASD volume is not formatted, we will issue a return code of 8.
#
mesg () {
echo "$@"
}
debug_mesg () {
case "${DEBUG}" in
yes) mesg "$@" ;;
*) ;;
esac
}
add_cio_channel() {
echo "$* # ${DATE}" >> /boot/zipl/active_devices.txt
}
remove_cio_channel() {
[ -w /boot/zipl/active_devices.txt ] && sed -i -e "/^${1}/d" /boot/zipl/active_devices.txt
}
usage(){
echo "Usage: ${0} [-f -t <dasd_type> ] <ccwid> <online> [use_diag]"
echo
echo " -f Override safety checks"
echo " -t DASD type. Must be provided if -f is used. Only dasd-eckd and"
echo " dasd-fba are supported - Deprecated"
echo " ccwid = x.y.ssss where"
echo " x is always 0 until IBM creates something that uses that number"
echo " y is the subchannel set ID (SSID). Most often"
echo " this is 0, but it could be non-zero"
echo " ssss is the four digit device address of the subchannel, in"
echo " hexidecimal, with leading zeros."
echo " online = 0 to take the device offline"
echo " 1 to bring the device online"
echo " use_diag = 0 to _not_ use z/VM DIAG250 I/O, which is the default"
echo " 1 to use z/VM DIAG250 I/O"
}
if [ "${DEBUG}" != "yes" ]; then
DEBUG="no"
fi
DATE=$(date)
DASD_FORCE=0
############################################################
# Parse the parameters from the command line
#
ARGS=$(getopt --options ft: -n "dasd_configure" -- "$@")
if [ $? != 0 ] ; then echo "Terminating..." >&2 ; exit 1 ; fi
eval set -- "${ARGS}"
debug_mesg "All the parms passed were ${ARGS}"
while true; do
case "${1}" in
-f) debug_mesg "This used to mean udev rules will always be generated."
debug_mesg "For chzdev, it means safety checks will be overridden."
debug_mesg "Kinda sorta the same thing, really."
PARM_LIST="${PARM_LIST} -f"
DASD_FORCE=1
shift 1
;;
-t) debug_mesg "This used to set the card type to ${2}"
debug_mesg "Now it gets ignored."
shift 2
;;
--) debug_mesg "Found the end of parms indicator: --"
shift 1
break
;;
*) debug_mesg "At the catch-all select entry"
debug_mesg "What was selected was ${1}"
shift 1
;;
esac
done
CCW_CHAN_ID=${1}
ON_OFF=${2}
USE_DIAG=${3}
if [ -z "${CCW_CHAN_ID}" ] || [ -z "${ON_OFF}" ]; then
mesg "You didn't specify all the needed parameters."
usage
exit 1
fi
if [ -n "${USE_DIAG}" ]; then
PARM_LIST="${PARM_LIST} use_diag=${USE_DIAG}"
else PARM_LIST="${PARM_LIST} use_diag=0"
fi
if [ "${ON_OFF}" == 0 ]; then
debug_mesg "chzdev -d dasd --no-root-update ${CCW_CHAN_ID}"
chzdev -d dasd --no-root-update ${CCW_CHAN_ID}
elif [ "${ON_OFF}" == 1 ]; then
debug_mesg "chzdev -e dasd --no-root-update ${CCW_CHAN_ID} ${PARM_LIST}"
chzdev -e dasd --no-root-update ${CCW_CHAN_ID} ${PARM_LIST}
else mesg "You must specify a 0 or a 1 for the online/offline attribute."
usage
exit 1
fi
RC=${?}
if [ ${RC} -ne 0 ]; then
exit ${RC}
elif [ ${ON_OFF} == 1 ]; then
exitcode=0
# Extract the full busid so that we can reference the proper entries in /sys
BUSID=$(/usr/sbin/lszdev dasd ${CCW_CHAN_ID} | /usr/bin/sed -e 1d | /usr/bin/tr -s " " | /usr/bin/cut -f2 -d" " )
# Make sure the DASD volume came online
for ((counter=0; counter<30; counter++)); do
sleep 0.1
read online < /sys/bus/ccw/devices/${BUSID}/online
if [ ${online} -eq 1 ] ; then
break
fi
done
if [ ${online} -ne 1 ]; then
debug_mesg "DASD ${CCW_CHAN_ID} did not come online."
exit 17
fi
# Check to see if the DASD volume is unformatted. If so, let YaST know.
read status < /sys/bus/ccw/devices/${BUSID}/status
if [ "${status}" == "unformatted" ]; then
mesg "DASD ${CCW_CHAN_ID} is unformatted."
exitcode=8
fi
fi
if [ ${ON_OFF} == 1 ]; then
add_cio_channel "${CCW_CHAN_ID}"
else remove_cio_channel "${CCW_CHAN_ID}"
fi
exit ${exitcode}

173
dasd_configure.suse Normal file
View File

@ -0,0 +1,173 @@
#! /bin/sh
#
# dasd_configure
#
# Copyright (c) 2017 SUSE LINUX GmbH, Nuernberg, Germany.
#
# Configures a DASD device by calling the IBM-provided chzdev command.
# Whereas this script used to validate the parameters provided to it,
# we now rely on chzdev to do that instead. The script is intended only
# as a "translation layer" to provide backward compatability for older
# scripts and tools that invoke it.
#
# Usage:
# dasd_configure [-f -t <dasd_type> ] <ccwid> <online> [use_diag]
#
# -f Override safety checks
# -t DASD type. Must be provided if -f is used. Only dasd-eckd and
# dasd-fba are supported - Deprecated
# ccwid = x.y.ssss where
# x is always 0 until IBM creates something that uses that number
# y is the subchannel set ID (SSID). Most often
# this is 0, but it could be non-zero
# ssss is the four digit device address of the subchannel, in
# hexidecimal, with leading zeros.
# online = 0 to take the device offline
# 1 to bring the device online
# use_diag = 0 to _not_ use z/VM DIAG250 I/O, which is the default
# 1 to use z/VM DIAG250 I/O
#
# Return values:
# Return codes are determined by the chzdev command, with one exception: If a
# DASD volume is not formatted, we will issue a return code of 8.
#
mesg () {
echo "$@"
}
debug_mesg () {
case "${DEBUG}" in
yes) mesg "$@" ;;
*) ;;
esac
}
add_cio_channel() {
echo "$* # ${DATE}" >> /boot/zipl/active_devices.txt
}
remove_cio_channel() {
[ -w /boot/zipl/active_devices.txt ] && sed -i -e "/^${1}/d" /boot/zipl/active_devices.txt
}
usage(){
echo "Usage: ${0} [-f -t <dasd_type> ] <ccwid> <online> [use_diag]"
echo
echo " -f Override safety checks"
echo " -t DASD type. Must be provided if -f is used. Only dasd-eckd and"
echo " dasd-fba are supported - Deprecated"
echo " ccwid = x.y.ssss where"
echo " x is always 0 until IBM creates something that uses that number"
echo " y is the subchannel set ID (SSID). Most often"
echo " this is 0, but it could be non-zero"
echo " ssss is the four digit device address of the subchannel, in"
echo " hexidecimal, with leading zeros."
echo " online = 0 to take the device offline"
echo " 1 to bring the device online"
echo " use_diag = 0 to _not_ use z/VM DIAG250 I/O, which is the default"
echo " 1 to use z/VM DIAG250 I/O"
}
if [ "${DEBUG}" != "yes" ]; then
DEBUG="no"
fi
DATE=$(date)
DASD_FORCE=0
############################################################
# Parse the parameters from the command line
#
ARGS=$(getopt --options ft: -n "dasd_configure" -- "$@")
if [ $? != 0 ] ; then echo "Terminating..." >&2 ; exit 1 ; fi
eval set -- "${ARGS}"
debug_mesg "All the parms passed were ${ARGS}"
while true; do
case "${1}" in
-f) debug_mesg "This used to mean udev rules will always be generated."
debug_mesg "For chzdev, it means safety checks will be overridden."
debug_mesg "Kinda sorta the same thing, really."
PARM_LIST="${PARM_LIST} -f"
DASD_FORCE=1
shift 1
;;
-t) debug_mesg "This used to set the card type to ${2}"
debug_mesg "Now it gets ignored."
shift 2
;;
--) debug_mesg "Found the end of parms indicator: --"
shift 1
break
;;
*) debug_mesg "At the catch-all select entry"
debug_mesg "What was selected was ${1}"
shift 1
;;
esac
done
CCW_CHAN_ID=${1}
ON_OFF=${2}
USE_DIAG=${3}
if [ -z "${CCW_CHAN_ID}" ] || [ -z "${ON_OFF}" ]; then
mesg "You didn't specify all the needed parameters."
usage
exit 1
fi
if [ -n "${USE_DIAG}" ]; then
PARM_LIST="${PARM_LIST} use_diag=${USE_DIAG}"
else PARM_LIST="${PARM_LIST} use_diag=0"
fi
if [ "${ON_OFF}" == 0 ]; then
debug_mesg "chzdev -d dasd --no-root-update ${CCW_CHAN_ID}"
chzdev -d dasd --no-root-update ${CCW_CHAN_ID}
elif [ "${ON_OFF}" == 1 ]; then
debug_mesg "chzdev -e dasd --no-root-update ${CCW_CHAN_ID} ${PARM_LIST}"
chzdev -e dasd --no-root-update ${CCW_CHAN_ID} ${PARM_LIST}
else mesg "You must specify a 0 or a 1 for the online/offline attribute."
usage
exit 1
fi
RC=${?}
if [ ${RC} -ne 0 ]; then
exit ${RC}
elif [ ${ON_OFF} == 1 ]; then
exitcode=0
# Extract the full busid so that we can reference the proper entries in /sys
BUSID=$(/sbin/lszdev dasd ${CCW_CHAN_ID} | /usr/bin/sed -e 1d | /usr/bin/tr -s " " | /usr/bin/cut -f2 -d" " )
# Make sure the DASD volume came online
for ((counter=0; counter<30; counter++)); do
sleep 0.1
read online < /sys/bus/ccw/devices/${BUSID}/online
if [ ${online} -eq 1 ] ; then
break
fi
done
if [ ${online} -ne 1 ]; then
debug_mesg "DASD ${CCW_CHAN_ID} did not come online."
exit 17
fi
# Check to see if the DASD volume is unformatted. If so, let YaST know.
read status < /sys/bus/ccw/devices/${BUSID}/status
if [ "${status}" == "unformatted" ]; then
mesg "DASD ${CCW_CHAN_ID} is unformatted."
exitcode=8
fi
fi
if [ ${ON_OFF} == 1 ]; then
add_cio_channel "${CCW_CHAN_ID}"
else remove_cio_channel "${CCW_CHAN_ID}"
fi
exit ${exitcode}

156
dasd_reload.opensuse Normal file
View File

@ -0,0 +1,156 @@
#!/bin/sh
#
# dasd_reload
# $Id: dasd_reload,v 1.2 2004/05/26 15:17:09 hare Exp $
#
# Deconfigures all active DASDs, unloads the modules
# and activates the configured DASDs again.
# Needed to establish an identical device mapping
# in the installation system and in the running system.
# All DASD access need to be cancelled prior to running
# this script.
#
# Usage:
# dasd_reload
#
# Return values:
# 1 Cannot read /proc/modules
# 2 Missing module programs
# 3 /sys not mounted
# 4 Failure on deactivate DASDs
#
if [ ! -r /proc/modules ]; then
echo "Cannot read /proc/modules"
exit 1
fi
if [ ! -x /usr/sbin/rmmod -o ! -x /usr/sbin/modprobe ]; then
echo "Missing module programs"
exit 2
fi
if [ ! -d /sys/bus ]; then
echo "sysfs not mounted"
exit 3
fi
let anymd=0
if [ -f /proc/mdstat ]; then
for mddevice in $(grep active /proc/mdstat | cut -f1 -d:); do
mdadm -S /dev/${mddevice}
let anymd=1
done
udevadm settle
fi
#
# Setting HyperPAV alias devices offline
#
dasd_alias=
let EXITRC=0
for dev in /sys/bus/ccw/devices/*; do
if [ -f ${dev}/use_diag ]; then
read _online < ${dev}/online
read _alias < ${dev}/alias
if [ "$_online" -eq 1 -a "$_alias" -eq 1 ]; then
echo "setting DASD HyperPAV alias $(basename ${dev}) offline"
echo "0" > ${dev}/online
read _online < ${dev}/online
dasd_alias="${dasd_alias} $(basename ${dev})"
if [ "$_online" -eq 1 ]; then
echo "failure on setting DASD HyperPAV alias $(basename ${dev}) offline !"
let EXITRC=4
fi
fi
fi
done
#
# Setting "normal" DASD and HyperPAV base devices offline
#
dasd_base=
for dev in /sys/bus/ccw/devices/*; do
if [ -f ${dev}/use_diag ]; then
read _online < ${dev}/online
read _alias < ${dev}/alias
if [ "$_online" -eq 1 -a "$_alias" -eq 0 ]; then
echo "setting DASD $(basename ${dev}) offline"
echo "0" > ${dev}/online
read _online < ${dev}/online
dasd_base="${dasd_base} $(basename ${dev})"
if [ "$_online" -eq 1 ]; then
echo "failure on setting DASD $(basename ${dev}) offline !"
let EXITRC=4
fi
fi
fi
done
udevadm settle
module_list=
module_test_list="dasd_diag_mod dasd_eckd_mod dasd_fba_mod dasd_mod"
for module in ${module_test_list}; do
if grep -q "${module}" /proc/modules; then
module_list="${module} ${module_list}"
: Unloading ${module}
/usr/sbin/rmmod ${module}
fi
done
udevadm settle
sleep 2
if [ -d /etc/udev/rules.d ]; then
cd /etc/udev/rules.d
#
# Re-activating "normal" DASD and HyperPAV base devices
#
# We need to move all the DASD udev rules out from /etc/udev/rules.d
# because if we don't, then when the first DASD volume gets brought
# back online, they are all brought back online, in a non-deterministic
# order, not the numeric order we expect.
#
mv -i 41-dasd-*.rules 51-dasd-*.rules /tmp
cd /tmp
for dasd in ${dasd_base}; do
for file in 41-dasd-*-${dasd}.rules 51-dasd-${dasd}.rules; do
[ -f "${file}" ] || continue
#
# Special handling is needed for old udev rules that start with 51-
# since the chzdev command won't look for that name
#
prefix="$(echo ${file} | cut -f1 -d-)"
if [ "${prefix}" == "51" ]; then
if [ -h /sys/bus/ccw/drivers/dasd-eckd/${dasd} ]; then
mv -i ${file} 41-dasd-eckd-${dasd}.rules
elif [ -h /sys/bus/ccw/drivers/dasd-fba/${dasd} ]; then
mv -i ${file} 41-dasd-fba-${dasd}.rules
else echo "DASD volume ${dasd} is neither an ECKD or FBA device."
let EXITRC=4
fi
fi
echo Activating ${dasd}
mv -i "${file}" /etc/udev/rules.d/
/usr/sbin/chzdev dasd --apply --configured -q --no-root-update ${dasd}
lsdasd
break
done
done
#
# Re-activating HyperPAV alias devices
#
for dasd in ${dasd_alias}; do
for file in 41-dasd-*-${dasd}.rules 51-dasd-${dasd}.rules; do
[ -f "${file}" ] || continue
echo Activating ${dasd}
mv -i "${file}" /etc/udev/rules.d/
/usr/sbin/chzdev dasd --apply --configured -q --no-root-update ${dasd}
break
done
done
fi
exit ${EXITRC}

156
dasd_reload.suse Normal file
View File

@ -0,0 +1,156 @@
#!/bin/sh
#
# dasd_reload
# $Id: dasd_reload,v 1.2 2004/05/26 15:17:09 hare Exp $
#
# Deconfigures all active DASDs, unloads the modules
# and activates the configured DASDs again.
# Needed to establish an identical device mapping
# in the installation system and in the running system.
# All DASD access need to be cancelled prior to running
# this script.
#
# Usage:
# dasd_reload
#
# Return values:
# 1 Cannot read /proc/modules
# 2 Missing module programs
# 3 /sys not mounted
# 4 Failure on deactivate DASDs
#
if [ ! -r /proc/modules ]; then
echo "Cannot read /proc/modules"
exit 1
fi
if [ ! -x /sbin/rmmod -o ! -x /sbin/modprobe ]; then
echo "Missing module programs"
exit 2
fi
if [ ! -d /sys/bus ]; then
echo "sysfs not mounted"
exit 3
fi
let anymd=0
if [ -f /proc/mdstat ]; then
for mddevice in $(grep active /proc/mdstat | cut -f1 -d:); do
mdadm -S /dev/${mddevice}
let anymd=1
done
udevadm settle
fi
#
# Setting HyperPAV alias devices offline
#
dasd_alias=
let EXITRC=0
for dev in /sys/bus/ccw/devices/*; do
if [ -f ${dev}/use_diag ]; then
read _online < ${dev}/online
read _alias < ${dev}/alias
if [ "$_online" -eq 1 -a "$_alias" -eq 1 ]; then
echo "setting DASD HyperPAV alias $(basename ${dev}) offline"
echo "0" > ${dev}/online
read _online < ${dev}/online
dasd_alias="${dasd_alias} $(basename ${dev})"
if [ "$_online" -eq 1 ]; then
echo "failure on setting DASD HyperPAV alias $(basename ${dev}) offline !"
let EXITRC=4
fi
fi
fi
done
#
# Setting "normal" DASD and HyperPAV base devices offline
#
dasd_base=
for dev in /sys/bus/ccw/devices/*; do
if [ -f ${dev}/use_diag ]; then
read _online < ${dev}/online
read _alias < ${dev}/alias
if [ "$_online" -eq 1 -a "$_alias" -eq 0 ]; then
echo "setting DASD $(basename ${dev}) offline"
echo "0" > ${dev}/online
read _online < ${dev}/online
dasd_base="${dasd_base} $(basename ${dev})"
if [ "$_online" -eq 1 ]; then
echo "failure on setting DASD $(basename ${dev}) offline !"
let EXITRC=4
fi
fi
fi
done
udevadm settle
module_list=
module_test_list="dasd_diag_mod dasd_eckd_mod dasd_fba_mod dasd_mod"
for module in ${module_test_list}; do
if grep -q "${module}" /proc/modules; then
module_list="${module} ${module_list}"
: Unloading ${module}
/sbin/rmmod ${module}
fi
done
udevadm settle
sleep 2
if [ -d /etc/udev/rules.d ]; then
cd /etc/udev/rules.d
#
# Re-activating "normal" DASD and HyperPAV base devices
#
# We need to move all the DASD udev rules out from /etc/udev/rules.d
# because if we don't, then when the first DASD volume gets brought
# back online, they are all brought back online, in a non-deterministic
# order, not the numeric order we expect.
#
mv -i 41-dasd-*.rules 51-dasd-*.rules /tmp
cd /tmp
for dasd in ${dasd_base}; do
for file in 41-dasd-*-${dasd}.rules 51-dasd-${dasd}.rules; do
[ -f "${file}" ] || continue
#
# Special handling is needed for old udev rules that start with 51-
# since the chzdev command won't look for that name
#
prefix="$(echo ${file} | cut -f1 -d-)"
if [ "${prefix}" == "51" ]; then
if [ -h /sys/bus/ccw/drivers/dasd-eckd/${dasd} ]; then
mv -i ${file} 41-dasd-eckd-${dasd}.rules
elif [ -h /sys/bus/ccw/drivers/dasd-fba/${dasd} ]; then
mv -i ${file} 41-dasd-fba-${dasd}.rules
else echo "DASD volume ${dasd} is neither an ECKD or FBA device."
let EXITRC=4
fi
fi
echo Activating ${dasd}
mv -i "${file}" /etc/udev/rules.d/
/sbin/chzdev dasd --apply --configured -q --no-root-update ${dasd}
lsdasd
break
done
done
#
# Re-activating HyperPAV alias devices
#
for dasd in ${dasd_alias}; do
for file in 41-dasd-*-${dasd}.rules 51-dasd-${dasd}.rules; do
[ -f "${file}" ] || continue
echo Activating ${dasd}
mv -i "${file}" /etc/udev/rules.d/
/sbin/chzdev dasd --apply --configured -q --no-root-update ${dasd}
break
done
done
fi
exit ${EXITRC}

20
dasdro Normal file
View File

@ -0,0 +1,20 @@
#!/bin/bash
# checks DASD accessibility in VM and sets Linux-side readonly attributes
# accordingly
modprobe -q vmcp
vmcp q v dasd 2>/dev/null >/dev/null || exit 0 # not running in VM
vmcp q v dasd | while read x dev rest
do
dev=`echo $dev|tr A-F a-f`
roattr=/sys/bus/ccw/devices/?.?.$dev/readonly
test -e $roattr || continue
if echo "$rest"|grep -q R/O
then
echo 1 >$roattr
else
echo 0 >$roattr
fi
done

157
detach_disks.sh.opensuse Normal file
View File

@ -0,0 +1,157 @@
#!/bin/sh
DASDFILE=/tmp/dasd.list.$(mcookie)
DETFILE=/tmp/detach.disks.$(mcookie)
KEEPFILE=/tmp/keep.disks.$(mcookie)
NICFILE=/tmp/nic.list.$(mcookie)
FAILFILE=/tmp/error.$(mcookie)
function expand_RANGE(){
local RANGE=${1}
local RANGE_SAVE=${RANGE}
local DEVNO
local BEGIN=0
local END=0
RANGE=$(IFS=":-"; echo ${RANGE} | cut -f1-2 -d" " )
set -- ${RANGE}
let BEGIN=0x$1 2>/dev/null
let END=0x$2 2>/dev/null
if [ ${BEGIN} -eq 0 ] || [ ${END} -eq 0 ]; then
${msg} "An invalid device number range was specified: ${RANGE_SAVE}" >&2
touch ${FAILFILE}
return
fi
for DEVNO in $(eval echo {${BEGIN}..${END}})
do printf "%d\n" ${DEVNO}
done
}
function usage(){
echo "Usage: ${0} [ -F ] [ -q ] [ -h ]"
echo " -F Exit with failure if any invalid parms are detected."
echo " -q Don't generate any output."
echo " -h Display this help message."
}
msg="echo"
let FORCE_FAIL=0
############################################################
# Parse the parameters from the command line
#
ARGS=$(getopt -a --options Fhq -n "detach_devices" -- "$@")
if [ $? -ne 0 ]; then
usage
exit 3
fi
eval set -- "${ARGS}"
for ARG; do
case "${ARG}" in
-F) let FORCE_FAIL=1
shift 1
;;
-h) usage;
exit 0
;;
-q) msg="/bin/true"
shift 1
;;
--) shift 1
;;
*) ${msg} "Extraneous input detected: ${1}"
shift 1
;;
esac
done
if [ -r /etc/sysconfig/virtsetup ]; then
. /etc/sysconfig/virtsetup
else ${msg} "No /etc/sysconfig/virtsetup file was found."
exit 1
fi
# First, get a list of all the DASD devices we have for this guest, in decimal.
# (Trying to handle things in hex gets complicated.)
/usr/sbin/vmcp -b1048576 q v dasd | cut -f2 -d" " |\
while read HEXNO
do let DECNO=0x${HEXNO}
echo ${DECNO}
done > ${DASDFILE} 2>/dev/null
# If the system administrator specified certain devices to be detached
# let's put those device numbers in a file, one per line.
touch ${DETFILE}
for ADDR in $(IFS=", " ; echo ${ZVM_DISKS_TO_DETACH})
do if $(echo ${ADDR} | grep -iqE ":|-" 2>/dev/null)
then expand_RANGE ${ADDR} >> ${DETFILE}
else let DEVNO=0
let DEVNO=0x${ADDR} 2>/dev/null
if [ ${DEVNO} -eq 0 ]; then
${msg} "An invalid device number was specified: ${ADDR}" >&2
touch ${FAILFILE}
else printf "%d\n" ${DEVNO}
fi
fi
done > ${DETFILE}
# If the system administrator specified certain devices that should _not_
# be detached, let's put those in another file, one per line.
touch ${KEEPFILE}
for ADDR in $(IFS=", " ; echo ${ZVM_DISKS_TO_NOT_DETACH})
do if $(echo ${ADDR} | grep -iqE ":|-" 2>/dev/null)
then expand_RANGE ${ADDR} >> ${KEEPFILE}
else let DEVNO=0
let DEVNO=0x${ADDR} 2>/dev/null
if [ ${DEVNO} -eq 0 ]; then
${msg} "An invalid device number was specified: ${ADDR}" >&2
touch ${FAILFILE}
else printf "%d\n" ${DEVNO}
fi
fi
done > ${KEEPFILE}
if [ ${FORCE_FAIL} -eq 1 ] && [ -e ${FAILFILE} ]; then
let RETURN_CODE=1
${msg} "Terminating detach_disk because of input errors."
else
# If the system administrator specified that all "unused" disks should be
# detached, compare the disks lsdasd show as activated to the complete
# list of disks we have currently, and add the inactive ones to the
# file containing devices to be detached
if [ "${ZVM_DETACH_ALL_UNUSED}" == "yes" ]; then
lsdasd -s | sed -e 1,2d | cut -f1 -d" " | \
while read ADDR
do let DEVNO=0x${ADDR}
sed -i -e "/^${DEVNO}$/d" ${DASDFILE}
done
cat ${DASDFILE} >> ${DETFILE}
fi
# Now remove any "to be kept" disks from the detach file
while read DEVNO
do sed -i -e "/^${DEVNO}/d" ${DETFILE}
done < ${KEEPFILE}
# Get a list of all the virtual NICs since they require an
# extra keyword to detach. Contrary to what we've done before
# these will be hex values
/usr/sbin/vmcp -b1048576 q nic | grep Adapter | cut -f2 -d" " | cut -f1 -d. > ${NICFILE}
# Now we sort the device numbers and detach them.
sort -un ${DETFILE} | \
while read DEVNO
do HEXNO=$(printf %04X ${DEVNO})
if grep -q ^${HEXNO}$ ${NICFILE} 2>/dev/null ; then
vmcp detach nic ${HEXNO} 2>/dev/null
else vmcp detach ${HEXNO} 2>/dev/null
fi
done
let RETURN_CODE=0
fi
rm -f ${DASDFILE} ${DETFILE} ${KEEPFILE} ${NICFILE} ${FAILFILE}
exit ${RETURN_CODE}

157
detach_disks.sh.suse Normal file
View File

@ -0,0 +1,157 @@
#!/bin/sh
DASDFILE=/tmp/dasd.list.$(mcookie)
DETFILE=/tmp/detach.disks.$(mcookie)
KEEPFILE=/tmp/keep.disks.$(mcookie)
NICFILE=/tmp/nic.list.$(mcookie)
FAILFILE=/tmp/error.$(mcookie)
function expand_RANGE(){
local RANGE=${1}
local RANGE_SAVE=${RANGE}
local DEVNO
local BEGIN=0
local END=0
RANGE=$(IFS=":-"; echo ${RANGE} | cut -f1-2 -d" " )
set -- ${RANGE}
let BEGIN=0x$1 2>/dev/null
let END=0x$2 2>/dev/null
if [ ${BEGIN} -eq 0 ] || [ ${END} -eq 0 ]; then
${msg} "An invalid device number range was specified: ${RANGE_SAVE}" >&2
touch ${FAILFILE}
return
fi
for DEVNO in $(eval echo {${BEGIN}..${END}})
do printf "%d\n" ${DEVNO}
done
}
function usage(){
echo "Usage: ${0} [ -F ] [ -q ] [ -h ]"
echo " -F Exit with failure if any invalid parms are detected."
echo " -q Don't generate any output."
echo " -h Display this help message."
}
msg="echo"
let FORCE_FAIL=0
############################################################
# Parse the parameters from the command line
#
ARGS=$(getopt -a --options Fhq -n "detach_devices" -- "$@")
if [ $? -ne 0 ]; then
usage
exit 3
fi
eval set -- "${ARGS}"
for ARG; do
case "${ARG}" in
-F) let FORCE_FAIL=1
shift 1
;;
-h) usage;
exit 0
;;
-q) msg="/bin/true"
shift 1
;;
--) shift 1
;;
*) ${msg} "Extraneous input detected: ${1}"
shift 1
;;
esac
done
if [ -r /etc/sysconfig/virtsetup ]; then
. /etc/sysconfig/virtsetup
else ${msg} "No /etc/sysconfig/virtsetup file was found."
exit 1
fi
# First, get a list of all the DASD devices we have for this guest, in decimal.
# (Trying to handle things in hex gets complicated.)
/sbin/vmcp -b1048576 q v dasd | cut -f2 -d" " |\
while read HEXNO
do let DECNO=0x${HEXNO}
echo ${DECNO}
done > ${DASDFILE} 2>/dev/null
# If the system administrator specified certain devices to be detached
# let's put those device numbers in a file, one per line.
touch ${DETFILE}
for ADDR in $(IFS=", " ; echo ${ZVM_DISKS_TO_DETACH})
do if $(echo ${ADDR} | grep -iqE ":|-" 2>/dev/null)
then expand_RANGE ${ADDR} >> ${DETFILE}
else let DEVNO=0
let DEVNO=0x${ADDR} 2>/dev/null
if [ ${DEVNO} -eq 0 ]; then
${msg} "An invalid device number was specified: ${ADDR}" >&2
touch ${FAILFILE}
else printf "%d\n" ${DEVNO}
fi
fi
done > ${DETFILE}
# If the system administrator specified certain devices that should _not_
# be detached, let's put those in another file, one per line.
touch ${KEEPFILE}
for ADDR in $(IFS=", " ; echo ${ZVM_DISKS_TO_NOT_DETACH})
do if $(echo ${ADDR} | grep -iqE ":|-" 2>/dev/null)
then expand_RANGE ${ADDR} >> ${KEEPFILE}
else let DEVNO=0
let DEVNO=0x${ADDR} 2>/dev/null
if [ ${DEVNO} -eq 0 ]; then
${msg} "An invalid device number was specified: ${ADDR}" >&2
touch ${FAILFILE}
else printf "%d\n" ${DEVNO}
fi
fi
done > ${KEEPFILE}
if [ ${FORCE_FAIL} -eq 1 ] && [ -e ${FAILFILE} ]; then
let RETURN_CODE=1
${msg} "Terminating detach_disk because of input errors."
else
# If the system administrator specified that all "unused" disks should be
# detached, compare the disks lsdasd show as activated to the complete
# list of disks we have currently, and add the inactive ones to the
# file containing devices to be detached
if [ "${ZVM_DETACH_ALL_UNUSED}" == "yes" ]; then
lsdasd -s | sed -e 1,2d | cut -f1 -d" " | \
while read ADDR
do let DEVNO=0x${ADDR}
sed -i -e "/^${DEVNO}$/d" ${DASDFILE}
done
cat ${DASDFILE} >> ${DETFILE}
fi
# Now remove any "to be kept" disks from the detach file
while read DEVNO
do sed -i -e "/^${DEVNO}/d" ${DETFILE}
done < ${KEEPFILE}
# Get a list of all the virtual NICs since they require an
# extra keyword to detach. Contrary to what we've done before
# these will be hex values
/sbin/vmcp -b1048576 q nic | grep Adapter | cut -f2 -d" " | cut -f1 -d. > ${NICFILE}
# Now we sort the device numbers and detach them.
sort -un ${DETFILE} | \
while read DEVNO
do HEXNO=$(printf %04X ${DEVNO})
if grep -q ^${HEXNO}$ ${NICFILE} 2>/dev/null ; then
vmcp detach nic ${HEXNO} 2>/dev/null
else vmcp detach ${HEXNO} 2>/dev/null
fi
done
let RETURN_CODE=0
fi
rm -f ${DASDFILE} ${DETFILE} ${KEEPFILE} ${NICFILE} ${FAILFILE}
exit ${RETURN_CODE}

181
hsnc Normal file
View File

@ -0,0 +1,181 @@
#! /bin/sh
# Copyright (c) 2003 SUSE LINUX AG Nuernberg, Germany.
#
# Please send feedback to http://www.suse.de/feedback/
#
# /etc/init.d/hsnc
#
# and symbolic its link
#
# /usr/sbin/ip_watcher.pl
# /usr/sbin/xcec-bridge
# /usr/sbin/start_hsnc.sh
# /use/sbin/rchsnc
#
# System startup script for the HiperSockets Network Concentrator
#
# /etc/hsnc.conf should contain the following lines:
#
# operating_mode=[unicast|full|no]
# unicast means, only unicast forwarded between the hsint's and osaint's.
# this is the default mode
# full means, unicast, multicast and broadcast are forwarded, if supported
# by the hardware
#
# hsi_int="<interface> [<interface> [...]]"
# described all the HiperSockets interfaces involved in the HSN
#
# osa_int="<interface>"
# describes the OSA interface connecting to other LANs
#
START_HSNC_BIN=/usr/sbin/start_hsnc.sh
IP_WATCHER_BIN=/usr/sbin/ip_watcher.pl
XCEC_BRIDGE_BIN=/usr/sbin/xcec-bridge
HSNC_CONFIG_FILE=/etc/sysconfig/hsnc
HSNC_CLEANUP_FILE=/var/run/hsnc.cleanup
test -x $START_HSNC_BIN || exit 5
test -x $IP_WATCHER_BIN || exit 5
test -x $XCEC_BRIDGE_BIN || exit 5
# Return values acc. to LSB for all commands but status:
# 0 - success
# 1 - generic or unspecified error
# 2 - invalid or excess argument(s)
# 3 - unimplemented feature (e.g. "reload")
# 4 - insufficient privilege
# 5 - program is not installed
# 6 - program is not configured
# 7 - program is not running
#
# Note that starting an already running service, stopping
# or restarting a not-running service as well as the restart
# with force-reload (in case signalling is not supported) are
# considered a success.
#call with cleanup or not
read_config_file() {
if [ "$1" == "cleanup" ]; then
file=$HSNC_CLEANUP_FILE
else
file=$HSNC_CONFIG_FILE
fi
if [ -s $file ]; then
source $file
else
echo -ne "\nCannot read $file: empty or nonexistant! "
# Means not configured:
exit 3
fi
}
#call with cleanup or not
set_osa_mode() {
# for full mode, we set up the osa as multicast router. otherwise, no
# special setup is required for the osa.
if [ "$operating_mode" == "full" ]; then
if [ "$1" == "cleanup" ]; then
echo no_router > /sys/class/net/$osa_int/device/route4
else
echo multicast_router > /sys/class/net/$osa_int/device/route4
fi
fi
}
#call with cleanup or not
set_hsi_mode() {
# set all the involved HiperSockets interfaces as primary_connector. For
# special HA setups, some more tweaking is needed, but then a handcarved
# solution should be used anyway.
for i in $hsi_int ; do
if [ "$1" == "cleanup" ]; then
echo no_router > /sys/class/net/$i/device/route4
else
echo primary_connector > /sys/class/net/$i/device/route4
fi
done
}
do_start_hsnc() {
set_osa_mode
set_hsi_mode
if [ "$operating_mode" == "full" ]; then
$IP_WATCHER_BIN --check
else
$IP_WATCHER_BIN --check $osa_int
fi
CODE=$?
if [ $CODE != 0 ]; then
return $CODE
else
cp $HSNC_CONFIG_FILE $HSNC_CLEANUP_FILE
#
# To match the LSB spec, startproc returns 0,
# even if the program it already running.
#
if [ "$operating_mode" == "full" ]; then
startproc $START_HSNC_BIN
else
startproc $START_HSNC_BIN $osa_int
fi
return $?
fi
}
service="HiperSockets Network concentrator"
case "$1" in
start)
if checkproc $START_HSNC_BIN; then
# Starting an already running service is success:
echo -n "(already running)"
else
if read_config_file; then
do_start_hsnc
RETVAL=$?
exit $RETVAL
fi
fi
;;
stop)
echo -n "Shutting down $service "
# kill ip_watcher, start_hsnc, which started it needs cleans up
# then:
killproc -TERM $IP_WATCHER_BIN
if [ -f $HSNC_CLEANUP_FILE ]; then
read_config_file cleanup
# remove all connector settings(not yet implemented):
set_osa_mode cleanup
set_hsi_mode cleanup
# remove the file in /var/run
rm -f $HSNC_CLEANUP_FILE
else
echo -n "- no cleanup file found "
fi
;;
status)
echo -n "Checking $service "
## Check status with checkproc(8), if process is running
## checkproc will return with exit status 0.
# Status has a slightly different for the status command:
# 0 - service running
# 1 - service dead, but /var/run/ pid file exists
# 2 - service dead, but /var/lock/ lock file exists
# 3 - service not running
# NOTE: checkproc returns LSB compliant status values.
checkproc $START_HSNC_BIN
;;
*)
exit 1
;;
esac

16
hsnc.service Normal file
View File

@ -0,0 +1,16 @@
[Unit]
Description=Start the qeth HiperSockets Network Concentrator
After=network-online.target remote-fs.target
Wants=network-online.target remote-fs.target
ConditionPathExists=/sys/devices/qeth
[Service]
Type=oneshot
RemainAfterExit=yes
ExecStart=/usr/lib/systemd/scripts/hsnc start
ExecStartPost=/usr/lib/systemd/scripts/hsnc status
ExecStop=/usr/lib/systemd/scripts/hsnc stop
[Install]
WantedBy=default.target

64
iucv_configure.8 Normal file
View File

@ -0,0 +1,64 @@
.TH iucv_configure "8" "July 2013" "s390-tools"
.SH NAME
iucv_configure \- Configures or deconfigures a z/VM Inter-User Communications Vehicle (IUCV) network adapter
.SH SYNOPSIS
.B iucv_configure peer_userid online
.SH DESCRIPTION
.B iucv_configure
is intended to make it easy to persistently add and remove z/VM point-to-point IUCV network adapters. In addition to bringing the adapter online or offline, it will also create or delete the necessary udev rules for the adapter.
.SH PARAMETERS
.IP peer_userid
The z/VM userid of the virtual machine on the other end of the point-to-point connection.
.IP online
Either a literal 1 to bring the adapter online or a literal 0 to take it offline
.SH FILES
.I /etc/udev/rules.d/51-iucv-<peer_userid>.rules
.RS
This file provide the udev rules necessary to activate a specific IUCV adapter.
.RE
.SH ENVIRONMENT
.IP DEBUG
If set to "yes" some minimal debugging information is output during execution.
.SH DIAGNOSTICS
The following messages may be issued on stdout:
.IP
.B /sysfs not present
.RS
The sysfs file system could not be found in /proc/mounts, so there's nothing the script can
do. Return code 1 is set.
.RE
.IP
.B No IUCV user name given
.RS
You didn't specify the z/VM userid of the virtual machine to which you want to connect. Return code 2 is set.
.RE
.IP
.B Unable to connect to $PEER_USERID
.RS
The attempt to connect to the IUCV peer failed. Try "dmesg" to see if there is any indication why. Return code 3 is set.
.RE
.IP
.B Unable to remove device $netdev
.RS
The attempt to remove the IUCV adapter failed. Try "dmesg" to see if there is any indication why. Return code 4 is set.
.RE
If environment variable DEBUG is set to "yes," the following messages may be issued on stdout:
.IP
.B
Configuring IUCV device ${PEER_USERID}
.RS
Just a little bit of verbosity, since it just indicates that we got past certain error checks and will now try to do something useful.
.RE
.IP
.B Configured device $iucvdev
.RS
The attempt to create the IUCV adapter was successful.
.RE
.IP
.B Removed device $iucvdev
.RS
The attempt to remove the IUCV adapter was successful.
.RE
.SH BUGS
Gotta be some, I'm sure. If you find one, please open a bug report.

133
iucv_configure.opensuse Normal file
View File

@ -0,0 +1,133 @@
#! /bin/sh
#
# iucv_configure
#
# Configures a z/VM IUCV network adapter
#
# Usage:
# iucv_configure <peer_userid> <online>
#
# peer_userid = z/VM userid of the IUCV peer
# online = 0 to take the device offline
# 1 to bring the device online
#
# Return values:
# 1 sysfs not mounted
# 2 Invalid status for <online>
# 3 Could not create iucv device
# 4 Could not remove iucv device
#
if [ "${DEBUG}" != "yes" ]; then
DEBUG="no"
fi
mesg () {
echo "$@"
}
debug_mesg () {
case "$DEBUG" in
yes) mesg "$@" ;;
*) ;;
esac
}
if [ $# -ne 2 ] ; then
echo "Usage: $0 <peer_userid> <online>"
echo " peer_userid = z/VM userid of the IUCV peer"
echo " online = 0 to take the device offline"
echo " 1 to bring the device online"
exit 1
fi
# Get the mount point for sysfs
while read MNTPT MNTDIR MNTSYS MNTTYPE; do
if test "$MNTSYS" = "sysfs"; then
SYSFS="$MNTDIR"
break;
fi
done </proc/mounts
if [ -z "$SYSFS" ]; then
mesg "/sysfs not present"
exit 1
fi
PEER_USERID_LOWER=$1
PEER_USERID=$(echo $1 | tr "a-z" "A-Z")
ONLINE=$2
if [ -z "$PEER_USERID" ] ; then
mesg "No IUCV user name given"
exit 2
fi
if [ -z "$ONLINE" ] ; then
ONLINE=1
fi
_iucv_dir=${SYSFS}/bus/iucv/devices
_iucv_drv=${SYSFS}/bus/iucv/drivers/netiucv
if [ ! -d "$_iucv_drv" ] ; then
modprobe -q netiucv
fi
debug_mesg "Configuring IUCV device ${PEER_USERID}"
for _iucv_dev in $_iucv_dir/netiucv?* ; do
[ -d $_iucv_dev ] || continue
read user < $_iucv_dev/user
if [ "$user" = "$PEER_USERID" ] ; then
# Already configured, ok
iucvdev=${_iucv_dev##*/}
break;
fi
done
if [ -z "$iucvdev" -a $ONLINE -eq 1 ] ; then
echo $PEER_USERID > $_iucv_drv/connection
if [ $? -ne 0 ] ; then
mesg "Unable to connect to $PEER_USERID"
exit 3
fi
for _iucv_dev in $_iucv_dir/netiucv?* ; do
[ -d $_iucv_dev ] || continue
read user < $_iucv_dev/user
if [ "$user" = "$PEER_USERID" ] ; then
iucvdev=${_iucv_dev##*/}
break;
fi
done
if [ "$iucvdev" ] ; then
debug_mesg "Configured device $iucvdev"
fi
elif [ "$iucvdev" -a $ONLINE -eq 0 ] ; then
for _net_dev in $_iucv_dir/$iucvdev/net/* ; do
[ -d $_net_dev ] || continue
netdev=${_net_dev##*/}
break;
done
if [ "$netdev" ] ; then
echo $netdev > $_iucv_drv/remove
if [ $? -ne 0 ] ; then
mesg "Unable to remove device $netdev"
exit 4
else
debug_mesg "Removed device $iucvdev"
rm -f /etc/udev/rules.d/51-iucv-$PEER_USERID.rules /etc/udev/rules.d/51-iucv-$PEER_USERID_LOWER.rules
iucvdev=
fi
fi
fi
if [ "$iucvdev" ] ; then
cat > /etc/udev/rules.d/51-iucv-$PEER_USERID.rules <<EOF
ACTION=="add", SUBSYSTEM=="subsystem", KERNEL=="iucv", RUN+="/usr/sbin/modprobe netiucv"
ACTION=="add", SUBSYSTEM=="drivers", KERNEL=="netiucv", ATTR{connection}="$PEER_USERID"
EOF
fi
exit

133
iucv_configure.suse Normal file
View File

@ -0,0 +1,133 @@
#! /bin/sh
#
# iucv_configure
#
# Configures a z/VM IUCV network adapter
#
# Usage:
# iucv_configure <peer_userid> <online>
#
# peer_userid = z/VM userid of the IUCV peer
# online = 0 to take the device offline
# 1 to bring the device online
#
# Return values:
# 1 sysfs not mounted
# 2 Invalid status for <online>
# 3 Could not create iucv device
# 4 Could not remove iucv device
#
if [ "${DEBUG}" != "yes" ]; then
DEBUG="no"
fi
mesg () {
echo "$@"
}
debug_mesg () {
case "$DEBUG" in
yes) mesg "$@" ;;
*) ;;
esac
}
if [ $# -ne 2 ] ; then
echo "Usage: $0 <peer_userid> <online>"
echo " peer_userid = z/VM userid of the IUCV peer"
echo " online = 0 to take the device offline"
echo " 1 to bring the device online"
exit 1
fi
# Get the mount point for sysfs
while read MNTPT MNTDIR MNTSYS MNTTYPE; do
if test "$MNTSYS" = "sysfs"; then
SYSFS="$MNTDIR"
break;
fi
done </proc/mounts
if [ -z "$SYSFS" ]; then
mesg "/sysfs not present"
exit 1
fi
PEER_USERID_LOWER=$1
PEER_USERID=$(echo $1 | tr "a-z" "A-Z")
ONLINE=$2
if [ -z "$PEER_USERID" ] ; then
mesg "No IUCV user name given"
exit 2
fi
if [ -z "$ONLINE" ] ; then
ONLINE=1
fi
_iucv_dir=${SYSFS}/bus/iucv/devices
_iucv_drv=${SYSFS}/bus/iucv/drivers/netiucv
if [ ! -d "$_iucv_drv" ] ; then
modprobe -q netiucv
fi
debug_mesg "Configuring IUCV device ${PEER_USERID}"
for _iucv_dev in $_iucv_dir/netiucv?* ; do
[ -d $_iucv_dev ] || continue
read user < $_iucv_dev/user
if [ "$user" = "$PEER_USERID" ] ; then
# Already configured, ok
iucvdev=${_iucv_dev##*/}
break;
fi
done
if [ -z "$iucvdev" -a $ONLINE -eq 1 ] ; then
echo $PEER_USERID > $_iucv_drv/connection
if [ $? -ne 0 ] ; then
mesg "Unable to connect to $PEER_USERID"
exit 3
fi
for _iucv_dev in $_iucv_dir/netiucv?* ; do
[ -d $_iucv_dev ] || continue
read user < $_iucv_dev/user
if [ "$user" = "$PEER_USERID" ] ; then
iucvdev=${_iucv_dev##*/}
break;
fi
done
if [ "$iucvdev" ] ; then
debug_mesg "Configured device $iucvdev"
fi
elif [ "$iucvdev" -a $ONLINE -eq 0 ] ; then
for _net_dev in $_iucv_dir/$iucvdev/net/* ; do
[ -d $_net_dev ] || continue
netdev=${_net_dev##*/}
break;
done
if [ "$netdev" ] ; then
echo $netdev > $_iucv_drv/remove
if [ $? -ne 0 ] ; then
mesg "Unable to remove device $netdev"
exit 4
else
debug_mesg "Removed device $iucvdev"
rm -f /etc/udev/rules.d/51-iucv-$PEER_USERID.rules /etc/udev/rules.d/51-iucv-$PEER_USERID_LOWER.rules
iucvdev=
fi
fi
fi
if [ "$iucvdev" ] ; then
cat > /etc/udev/rules.d/51-iucv-$PEER_USERID.rules <<EOF
ACTION=="add", SUBSYSTEM=="subsystem", KERNEL=="iucv", RUN+="/sbin/modprobe netiucv"
ACTION=="add", SUBSYSTEM=="drivers", KERNEL=="netiucv", ATTR{connection}="$PEER_USERID"
EOF
fi
exit

217
killcdl.opensuse Normal file
View File

@ -0,0 +1,217 @@
#!/bin/sh
#
# Copyright (c) 2016 SUSE LINUX GmbH, Nuernberg, Germany.
# Released under the GNU General Public License version 2.
#
let FORCE=0
DEVPARM=""
usage(){
echo "Usage: ${0} [ -f ] devno|busid"
echo " -f Force unformatting for DASD volumes in the CMS device range of 19x."
echo " devno The \"plain\" device number of the volume, e.g. 3184."
echo " busid The full specification of the volume, e.g., 0.0.3184."
}
ARCH="$(/usr/bin/uname -m)"
if [ "${ARCH}" != "s390x" ] && [ "${ARCH}" != "s390" ]; then
echo "This script is only useful on IBM mainframes."
exit 1
fi
############################################################
# Parse the parameters from the command line
#
ARGS=$(getopt -a --options f -n "killcdl" -- "$@")
if [ $? -ne 0 ]; then
usage
exit 3
fi
eval set -- "${ARGS}"
for ARG; do
case "${ARG}" in
-f) FORCE=1
shift 1
;;
--) shift 1
;;
[0-9a-fA-F]*) if [ ! -z "${DEVPARM}" ]; then
echo "More than one parameter specified."
usage
exit 4
fi
DEVPARM=${1}
shift 1
;;
*) echo "That looks invalid"
usage
exit 5
;;
esac
done
if [ -z "${DEVPARM}" ]; then
echo "You must specify the device number of the DASD volume to be unformatted."
usage
exit 6
fi
DEVNO=$(echo "${DEVPARM}" | tr A-Z a-z)
# Validate the device number or busid provided
set -- $(IFS='.'; echo ${DEVNO})
let NUMPARMS=${#}
if [ ${NUMPARMS} -ne 1 ] && [ ${NUMPARMS} -ne 3 ]; then
echo "You have not specified the device number in a recognizable format."
echo "It must either be the plain device number, e.g., 0123, or in"
echo "so-called busid format, e.g., 0.0.0123"
exit 7
fi
# Just a device number, SIMPLE=1. A busid, SIMPLE=0
SIMPLE=0
if [ ${NUMPARMS} -eq 1 ]; then
let SIMPLE=1
let FIRST=0
let FIRSTLEN=1
let SECOND=0
let SECONDLEN=1
DEVNO="${1}"
let DEVNOLEN=${#1}
else FIRST="${1}"
let FIRSTLEN=${#FIRST}
SECOND="${2}"
let SECONDLEN=${#SECOND}
DEVNO="${3}"
let DEVNOLEN=${#3}
fi
if [ ${FIRSTLEN} -ne 1 ] || [ ${SECONDLEN} -ne 1 ]; then
echo "The first and second fields of the busid may only be one digit long."
exit 8
fi
if [ ${DEVNOLEN} -gt 4 ]; then
echo "The device number may only be 4 digits long."
exit 9
fi
if [ ${DEVNOLEN} -lt 4 ]; then
DEVNO=$(echo "0000${DEVNO}" | rev | cut -c1-4 | rev)
fi
BUSID="${FIRST}.${SECOND}.${DEVNO}"
if [ ! -h /sys/bus/ccw/devices/${BUSID} ]; then
echo "Busid ${BUSID} was not found."
/usr/sbin/cio_ignore -i ${BUSID} > /dev/null
if [ $? -eq 0 ]; then
echo "That device is in the cio_ignore list."
echo "Please remove it with \"cio_ignore -r ${BUSID}\" before trying again."
fi
exit 10
fi
case ${DEVNO:0:3} in
019) if grep -q "version = FF" /proc/cpuinfo 2>/dev/null; then
echo "That looks like a CMS disk."
if [ ${FORCE} -eq 0 ]; then
echo "Specify the -f option to force the operation."
exit 11
fi
echo "But you specified -f so we'll kill it anyway."
fi
;;
esac
read ORIG_ONLINE_STATUS < /sys/bus/ccw/devices/${BUSID}/online
DISCIPLINE="none"
if [ -r /sys/bus/ccw/devices/${BUSID}/discipline ]; then
# We have to bring the device online before the kernel will fill in
# the value for discipline.
if [ ${ORIG_ONLINE_STATUS} -eq 0 ]; then
/usr/sbin/chccwdev -e ${BUSID}
/usr/sbin/udevadm settle
fi
read STATUS < /sys/bus/ccw/devices/${BUSID}/status
if [ "${STATUS}" == "unformatted" ]; then
echo "DASD device ${BUSID} is already in an unformatted state."
if [ ${ORIG_ONLINE_STATUS} -eq 0 ]; then
/usr/sbin/chccwdev -d -s ${BUSID}
/usr/sbin/udevadm settle
fi
exit 0
fi
read DISCIPLINE < /sys/bus/ccw/devices/${BUSID}/discipline
else read CU_TYPE < /sys/bus/ccw/devices/${BUSID}/cutype
read DEV_TYPE < /sys/bus/ccw/devices/${BUSID}/devtype
case "${CU_TYPE}" in
3990/*|2105/*|2107/*|1750/*|9343/*)
DISCIPLINE=ECKD
;;
3880/*)
case "${DEV_TYPE}" in
3390/*)
DISCIPLINE=ECKD
;;
esac
;;
esac
fi
if [ "${DISCIPLINE}" != "ECKD" ]; then
echo "This script only works on ECKD DASD."
if [ ${ORIG_ONLINE_STATUS} -eq 0 ]; then
/usr/sbin/chccwdev -d -s ${BUSID}
fi
exit 12
fi
read STATUS < /sys/bus/ccw/devices/${BUSID}/online
if [ ${STATUS} -eq 1 ]; then
if [ ! -h /dev/disk/by-path/ccw-${BUSID} ]; then
echo "The udev-generated symbolic link in /dev/disk/by-path was not found."
exit 13
fi
/usr/sbin/chccwdev -d -s ${BUSID}
/usr/sbin/udevadm settle
read STATUS < /sys/bus/ccw/devices/${BUSID}/online
if [ ${STATUS} -ne 0 ]; then
echo "Device number ${DEVNO} didn't go offline. Unable to continue."
exit 14
fi
fi
/usr/sbin/chccwdev -a raw_track_access=1 -e ${BUSID}
/usr/sbin/udevadm settle
read STATUS < /sys/bus/ccw/devices/${BUSID}/online
if [ ${STATUS} -ne 1 ]; then
echo "Unable to bring ${DEVNO} online. Unable to continue."
exit 15
fi
# After this point, we will kill the formatting on the device
perl -e 'for ($h=0;$h<2;$h++){printf "\0\0\0%c\0\0\0\x8%s",$h,(("\0"x8).("\xff"x8).("\0"x65512))}' | dd bs=65536 count=2 oflag=direct of=/dev/disk/by-path/ccw-${BUSID} >/dev/null 2>&1
if [ "$?" -ne 0 ]; then
echo "The writing of the null record 0 failed."
exit 16
fi
echo "Setting ${BUSID} back offline with raw track access disabled."
/usr/sbin/chccwdev -d -s -a raw_track_access=0 ${BUSID}
/usr/sbin/udevadm settle
if [ ${ORIG_ONLINE_STATUS} -eq 1 ]; then
/usr/sbin/chccwdev -e ${BUSID}
/usr/sbin/udevadm settle
fi

217
killcdl.suse Normal file
View File

@ -0,0 +1,217 @@
#!/bin/sh
#
# Copyright (c) 2016 SUSE LINUX GmbH, Nuernberg, Germany.
# Released under the GNU General Public License version 2.
#
let FORCE=0
DEVPARM=""
usage(){
echo "Usage: ${0} [ -f ] devno|busid"
echo " -f Force unformatting for DASD volumes in the CMS device range of 19x."
echo " devno The \"plain\" device number of the volume, e.g. 3184."
echo " busid The full specification of the volume, e.g., 0.0.3184."
}
ARCH="$(/bin/uname -m)"
if [ "${ARCH}" != "s390x" ] && [ "${ARCH}" != "s390" ]; then
echo "This script is only useful on IBM mainframes."
exit 1
fi
############################################################
# Parse the parameters from the command line
#
ARGS=$(getopt -a --options f -n "killcdl" -- "$@")
if [ $? -ne 0 ]; then
usage
exit 3
fi
eval set -- "${ARGS}"
for ARG; do
case "${ARG}" in
-f) FORCE=1
shift 1
;;
--) shift 1
;;
[0-9a-fA-F]*) if [ ! -z "${DEVPARM}" ]; then
echo "More than one parameter specified."
usage
exit 4
fi
DEVPARM=${1}
shift 1
;;
*) echo "That looks invalid"
usage
exit 5
;;
esac
done
if [ -z "${DEVPARM}" ]; then
echo "You must specify the device number of the DASD volume to be unformatted."
usage
exit 6
fi
DEVNO=$(echo "${DEVPARM}" | tr A-Z a-z)
# Validate the device number or busid provided
set -- $(IFS='.'; echo ${DEVNO})
let NUMPARMS=${#}
if [ ${NUMPARMS} -ne 1 ] && [ ${NUMPARMS} -ne 3 ]; then
echo "You have not specified the device number in a recognizable format."
echo "It must either be the plain device number, e.g., 0123, or in"
echo "so-called busid format, e.g., 0.0.0123"
exit 7
fi
# Just a device number, SIMPLE=1. A busid, SIMPLE=0
SIMPLE=0
if [ ${NUMPARMS} -eq 1 ]; then
let SIMPLE=1
let FIRST=0
let FIRSTLEN=1
let SECOND=0
let SECONDLEN=1
DEVNO="${1}"
let DEVNOLEN=${#1}
else FIRST="${1}"
let FIRSTLEN=${#FIRST}
SECOND="${2}"
let SECONDLEN=${#SECOND}
DEVNO="${3}"
let DEVNOLEN=${#3}
fi
if [ ${FIRSTLEN} -ne 1 ] || [ ${SECONDLEN} -ne 1 ]; then
echo "The first and second fields of the busid may only be one digit long."
exit 8
fi
if [ ${DEVNOLEN} -gt 4 ]; then
echo "The device number may only be 4 digits long."
exit 9
fi
if [ ${DEVNOLEN} -lt 4 ]; then
DEVNO=$(echo "0000${DEVNO}" | rev | cut -c1-4 | rev)
fi
BUSID="${FIRST}.${SECOND}.${DEVNO}"
if [ ! -h /sys/bus/ccw/devices/${BUSID} ]; then
echo "Busid ${BUSID} was not found."
/sbin/cio_ignore -i ${BUSID} > /dev/null
if [ $? -eq 0 ]; then
echo "That device is in the cio_ignore list."
echo "Please remove it with \"cio_ignore -r ${BUSID}\" before trying again."
fi
exit 10
fi
case ${DEVNO:0:3} in
019) if grep -q "version = FF" /proc/cpuinfo 2>/dev/null; then
echo "That looks like a CMS disk."
if [ ${FORCE} -eq 0 ]; then
echo "Specify the -f option to force the operation."
exit 11
fi
echo "But you specified -f so we'll kill it anyway."
fi
;;
esac
read ORIG_ONLINE_STATUS < /sys/bus/ccw/devices/${BUSID}/online
DISCIPLINE="none"
if [ -r /sys/bus/ccw/devices/${BUSID}/discipline ]; then
# We have to bring the device online before the kernel will fill in
# the value for discipline.
if [ ${ORIG_ONLINE_STATUS} -eq 0 ]; then
/sbin/chccwdev -e ${BUSID}
/sbin/udevadm settle
fi
read STATUS < /sys/bus/ccw/devices/${BUSID}/status
if [ "${STATUS}" == "unformatted" ]; then
echo "DASD device ${BUSID} is already in an unformatted state."
if [ ${ORIG_ONLINE_STATUS} -eq 0 ]; then
/sbin/chccwdev -d -s ${BUSID}
/sbin/udevadm settle
fi
exit 0
fi
read DISCIPLINE < /sys/bus/ccw/devices/${BUSID}/discipline
else read CU_TYPE < /sys/bus/ccw/devices/${BUSID}/cutype
read DEV_TYPE < /sys/bus/ccw/devices/${BUSID}/devtype
case "${CU_TYPE}" in
3990/*|2105/*|2107/*|1750/*|9343/*)
DISCIPLINE=ECKD
;;
3880/*)
case "${DEV_TYPE}" in
3390/*)
DISCIPLINE=ECKD
;;
esac
;;
esac
fi
if [ "${DISCIPLINE}" != "ECKD" ]; then
echo "This script only works on ECKD DASD."
if [ ${ORIG_ONLINE_STATUS} -eq 0 ]; then
/sbin/chccwdev -d -s ${BUSID}
fi
exit 12
fi
read STATUS < /sys/bus/ccw/devices/${BUSID}/online
if [ ${STATUS} -eq 1 ]; then
if [ ! -h /dev/disk/by-path/ccw-${BUSID} ]; then
echo "The udev-generated symbolic link in /dev/disk/by-path was not found."
exit 13
fi
/sbin/chccwdev -d -s ${BUSID}
/sbin/udevadm settle
read STATUS < /sys/bus/ccw/devices/${BUSID}/online
if [ ${STATUS} -ne 0 ]; then
echo "Device number ${DEVNO} didn't go offline. Unable to continue."
exit 14
fi
fi
/sbin/chccwdev -a raw_track_access=1 -e ${BUSID}
/sbin/udevadm settle
read STATUS < /sys/bus/ccw/devices/${BUSID}/online
if [ ${STATUS} -ne 1 ]; then
echo "Unable to bring ${DEVNO} online. Unable to continue."
exit 15
fi
# After this point, we will kill the formatting on the device
perl -e 'for ($h=0;$h<2;$h++){printf "\0\0\0%c\0\0\0\x8%s",$h,(("\0"x8).("\xff"x8).("\0"x65512))}' | dd bs=65536 count=2 oflag=direct of=/dev/disk/by-path/ccw-${BUSID} >/dev/null 2>&1
if [ "$?" -ne 0 ]; then
echo "The writing of the null record 0 failed."
exit 16
fi
echo "Setting ${BUSID} back offline with raw track access disabled."
/sbin/chccwdev -d -s -a raw_track_access=0 ${BUSID}
/sbin/udevadm settle
if [ ${ORIG_ONLINE_STATUS} -eq 1 ]; then
/sbin/chccwdev -e ${BUSID}
/sbin/udevadm settle
fi

335
lgr_check Normal file
View File

@ -0,0 +1,335 @@
#!/bin/sh
function check_sysoper(){
local SYSOPER=$(vmcp q sysoper | cut -f4 -d" ")
local USERID=$(vmcp q userid | cut -f1 -d" ")"."
if [ "${SYSOPER}" == "${USERID}" ]; then
return 0
else return 1
fi
}
function check_disc(){
local USERID=$(vmcp q userid | cut -f1 -d" ")
local DISCONNECTED=$(vmcp q ${USERID} | cut -f2 -d-)
if [ "${DISCONNECTED}" == " DSC" ]; then
return 1
else return 0
fi
}
function check_3270(){
local CONMODE=$(vmcp q term | sed -n -e '/CONMODE/ {s/CONMODE \([0-9]*\), .*$/\1/;p}')
if [ "${CONMODE}" == "3270" ]; then
return 0
else return 1
fi
}
function check_graf(){
local GRAF
GRAF=$(vmcp -b 1048576 q v graf 2>/dev/null | grep -v "^CONS" | grep "ON LDEV" )
if [ ${?} -eq 0 ] && [ -n "${GRAF}" ]; then
return 0
else return 1
fi
}
function check_ascii_console(){
local SYSASCII=$(vmcp q v sysascii 2>/dev/null | grep "not attached to you")
if [ -z "${SYSASCII}" ]; then
return 0
else return 1
fi
}
function check_tdisk(){
local TDISK=$(vmcp -b 1048576 q v dasd 2>/dev/null | grep TEMP)
if [ -n "${TDISK}" ]; then
return 0
else return 1
fi
}
function check_ctc(){
local CTC
CTC=$(vmcp -b 1048576 q v ctc 2>/dev/null)
if [ ${?} -eq 0 ]; then
return 0
else return 1
fi
}
function check_dynamic_switch(){
local SWCH
SWCH=$(vmcp -b 1048576 q v switches 2>/dev/null)
if [ ${?} -eq 0 ]; then
return 0
else return 1
fi
}
function check_maint_mdisks(){
local MDISKS
MDISKS=$(vmcp -b 1048576 q v dasd | grep -E "0190|0191|0193|019D|019E")
if [ -n "${MDISKS}" ]; then
return 0
else return 1
fi
}
function check_wrkalleg(){
local WRKALLEG
WRKALLEG=$(vmcp -b 1048576 q wrkalleg | grep "is not simulated")
if [ -z "${WRKALLEG}" ]; then
return 0
else return 1
fi
}
function check_isolated_vswitch(){
local ISOLATED=0
local VSWITCH
# Find out if we have any NICs coupled to any VSWITCH or not. If not, we're OK.
VSWITCH=$(vmcp -b 1048576 q nic | sed -e '1~2 {N;s/\n//;}' | grep VSWITCH)
if [ -z "${VSWITCH}" ]; then
return 1
fi
ISOLATED=$(vmcp -b 1048576 q nic | sed -e '1~2 {N;s/\n//;}' | \
grep VSWITCH | \
sed -e 's/^.* VSWITCH: //' | \
while read owner name
do VSWITCH=$(vmcp -b 1048576 q vswitch $name | grep RDEV)
if [ -z "${VSWITCH}" ]; then
echo 1
fi
done)
if [ "${ISOLATED}" == "1" ]; then
return 0
else return 1
fi
}
function check_chpidvirt(){
local CHPIDV
CHPIDV=$(vmcp q chpidv 2>/dev/null | grep "CHPID Virtualization is on")
if [ -z "${CHPIDV}" ]; then
return 0
else return 1
fi
}
function check_pci_functions(){
local PCIF
local RETCODE
PCIF=$(vmcp -b 1048576 q v pcif 2>/dev/null | grep "A PCI function was not found.")
RETCODE=${?}
if [ ${RETCODE} -eq 0 ] && [ -z "${PCIF}" ]; then
return 0
else return 1
fi
}
function check_tape_assign(){
local TAPES
local RETCODE=1
TAPES=$(vmcp -b 1048576 q v tapes 2>/dev/null | grep "Device TAPE does not exist")
if [ -n "${TAPES}" ]; then
return 1
fi
TAPES=$(vmcp -b 1048576 q v tapes 2>/dev/null | grep "NOASSIGN")
if [ -n "${TAPES}" ]; then
return 0
else return 1
fi
}
function check_open_spool(){
local QSPOOL
local OPENSPOOL=0
QSPOOL=$(vmcp -b 1048576 q pun \* all 2>/dev/null | grep "OPEN")
if [ -n "${QSPOOL}" ]; then
let OPENSPOOL=1
fi
QSPOOL=$(vmcp -b 1048576 q prt \* all 2>/dev/null | grep "OPEN" | grep -v " CON ")
if [ -n "${QSPOOL}" ]; then
let OPENSPOOL=1
fi
if [ ${OPENSPOOL} -eq 1 ]; then
return 0
else return 1
fi
}
function check_xstore(){
local XSTOR
XSTOR=$(vmcp -b 1048576 q v xstor 2>/dev/null | grep "XSTORE = none")
if [ -z "${XSTOR}" ]; then
return 0
else return 1
fi
}
function check_iucv(){
local QIUCV
QIUCV=$(vmcp -b 1048576 q iucv 2>/dev/null | grep -vE "^No IUCV paths exist|^Source| *MSG| *MSGALL")
if [ -n "${QIUCV}" ]; then
return 0
else return 1
fi
}
function usage(){
echo "Usage: ${0} [ -f ] [ -h ] devno|busid"
echo " -q Don't generate any output, just set a return code."
echo " -m Suppress the check for local minidisks."
echo " Only use this if you know for certain all minidisks for this"
echo " guest are NOT local to this instance of z/VM."
echo " -h Display this help message."
}
ARCH="$(/bin/uname -m)"
if [ "${ARCH}" != "s390x" ] && [ "${ARCH}" != "s390" ]; then
echo "This script is only useful on IBM mainframes."
exit 1
fi
HYPERVISOR=$(systemd-detect-virt)
if [ "${HYPERVISOR}" != "zvm" ]; then
echo "This script is only useful for guests of the z/VM hypervisor."
exit 1
fi
MDISK_SUPPRESS=0
msg="echo"
############################################################
# Parse the parameters from the command line
#
ARGS=$(getopt -a --options qhm -n "lgr_check" -- "$@")
if [ $? -ne 0 ]; then
usage
exit 3
fi
eval set -- "${ARGS}"
for ARG; do
case "${ARG}" in
-h) usage;
exit 0
;;
-m) let MDISK_SUPPRESS=1;
shift 1
;;
-q) msg="/bin/true"
shift 1
;;
--) shift 1
;;
*) ${msg} "Extraneous input detected: ${1}"
shift 1
;;
esac
done
let FAIL=0
##let COLS=$(stty -a | sed -n -e '/columns/{s/^.*columns \([0-9]*\);.*$/\1/;p}')
if [ ! -c /dev/vmcp ]; then
${msg} "Cannot find /dev/vmcp to issue z/VM CP commands."
exit 1
fi
${msg} "Checking for conditions that might prevent Live Guest Relocation"
if check_chpidvirt; then
${msg} "This guest does not have CHPID Virtualization set for it in the CP directory."
${msg} "Live Guest Relocation is absolutely not possible for this guest."
# exit 99
let FAIL=1
fi
if check_sysoper ; then
${msg} "This guest is currently the z/VM system operator."
let FAIL=1
fi
let GUEST_CONN=0
if check_disc; then
${msg} "The guest is not running disconnected."
let GUEST_CONN=1
let FAIL=1
fi
if check_3270; then
${msg} -n "The guest has a 3270 console, "
if [ ${GUEST_CONN} -eq 0 ]; then
${msg} "but it is not currently in use."
else ${msg} "and it is currently in use."
let FAIL=1
fi
fi
if check_graf; then
${msg} "The guest has a DIALED 3270 device in current use."
let FAIL=1
fi
if check_ascii_console; then
${msg} "The guest has the ASCII console attached to it."
let FAIL=1
fi
if check_tdisk; then
${msg} "The guest has a temporary disk (T-disk) attached to it."
let FAIL=1
fi
if check_ctc; then
${msg} "The guest has a Channel-to-Channel device (CTC) attached to it."
let FAIL=1
fi
if check_dynamic_switch; then
${msg} "The guest has a dynamic switching device attached to it."
let FAIL=1
fi
if check_wrkalleg; then
${msg} "The guest is currently using virtual working allegiance."
let FAIL=1
fi
if [ ${MDISK_SUPPRESS} -eq 0 ] && check_maint_mdisks; then
${msg} "The guest currently has one or more Minidisks that might be local to this instance of z/VM."
let FAIL=1
fi
if check_isolated_vswitch; then
${msg} "The guest is currently coupled to an isolated VSWITCH."
let FAIL=1
fi
if check_pci_functions; then
${msg} "The guest has PCI functions available to it."
let FAIL=1
fi
if check_tape_assign; then
${msg} "The guest has potential problems with a tape."
let FAIL=1
fi
if check_open_spool; then
${msg} "The guest has an open SPOOL file that is not from the virtual console."
let FAIL=1
fi
if check_xstore; then
${msg} "The guest has Expanded Storage attached to it."
let FAIL=1
fi
if check_iucv; then
${msg} "The guest has an IUCV connection to a z/VM system service or another z/VM user."
let FAIL=1
fi
if [ ${FAIL} == 1 ]; then
${msg} "The guest is currently not eligible for Live Guest Relocation."
exit 1
else ${msg} "As far as can be determined from within the guest, it is currently eligible for Live Guest Relocation."
${msg} "This is not a guarantee. Other factors that cannot be checked from within the guest may prevent it from being eligible for LGR."
fi

72
mkdump.8 Normal file
View File

@ -0,0 +1,72 @@
.\" DO NOT MODIFY THIS FILE! It was generated by help2man 1.36.
.TH MKDUMP "8" "August 2011" "mkdump 2.0" "System Administration Utilities"
.SH NAME
mkdump \- Preparing disks for use as S/390 dump device.
.SH SYNOPSIS
.B mkdump
[\fIOPTIONS\fR] [\fIDEVICE\fR]...
.SH DESCRIPTION
mkdump 2.0.3
.PP
Prepare one or more volumes for use as S/390 dump device. Supported devices
are ECKD DASD and SCSI over zFCP disks, while multi\-volumes are limited to DASD.
.PP
Only whole disks can be used, no partitions! If the device is incompatible
formatted/partioned, the script will refuse to install the dump record
unless the \fB\-\-force\fR switch is given.
.PP
Disks which are in use or have mounted partitions will not be listed and can't be used.
The mentioning of "dumpdevice" after a disk indicates that it is an already usable dump device. Additionally multi\-volume dump devices are indicated by the list of including DASD ids.
.SH OPTIONS
.TP
\fB\-h\fR, \fB\-\-help\fR
display this help and exit
.TP
\fB\-V\fR, \fB\-\-version\fR
display version information and exit
.TP
\fB\-d\fR, \fB\-\-debug\fR
debug mode, do not run programs which commit changes
.TP
\fB\-v\fR, \fB\-\-verbose\fR
be verbose and show command outputs
.TP
\fB\-f\fR, \fB\-\-force\fR
force overwrite of the disk
.TP
\fB\-l\fR, \fB\-\-list\-dump\fR
display dump disks
.TP
\fB\-D\fR, \fB\-\-list\-dasd\fR
display usable DASD disks (Device, Size, ID, Dump)
.TP
\fB\-Z\fR, \fB\-\-list\-zfcp\fR
display usable SCSI over zFCP disks (Device, Size, ID, WWPN, LUN, Dump)
.SH DIAGNOSTICS
mkdump returns the following exit codes:
.RS
.IP 0
Normal (no errors or warnings detected)
.IP 11
Invalid or unusable disk (fatal)
.IP 12
Incompatible formatting/partitioning, can be corrected with --force
.IP 13
Missing support programs
.IP 14
Bad command line parameters
.IP 15
Access problem
.IP other
Support program failed
.SH AUTHOR
Written by Tim Hardeck <thardeck@suse.de>.
.SH "REPORTING BUGS"
Report bugs on https://bugzilla.novell.com/
.SH COPYRIGHT
Copyright \(co 2011 SUSE LINUX Products GmbH
License GPLv2 or (at your option) any later version.
<http://www.gnu.org/licenses/gpl-2.0.html>
.br
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.

666
mkdump.pl.opensuse Normal file
View File

@ -0,0 +1,666 @@
#!/usr/bin/perl
########################################################################
#
# mkdump.pl - Preparing disks for use as S/390 dump device
#
# Copyright (c) 2011 Tim Hardeck, SUSE LINUX Products GmbH
# Copyright (c) 2017 SUSE LINUX GmbH, Nuernberg, Germany.
# Based on mkdump.sh (c) 2004 Hannes Reinecke, SuSE AG
#
# License:
#
# This program is free software; you can redistribute it and/or
# modify it under the terms of the GNU General Public License
# as published by the Free Software Foundation; either version 2
# of the License, or (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software
# Foundation, Inc., 51 Franklin Street, Fifth Floor,
# Boston, MA 02110-1301 USA
#
########################################################################
use strict;
use warnings;
use Fcntl;
use Getopt::Long;
my $VERSION = "2.0.4";
my $BLKID = "/usr/sbin/blkid";
my $PARTED = "/usr/sbin/parted";
my $FDASD = "/usr/sbin/fdasd";
my $DASDVIEW = "/usr/sbin/dasdview";
my $DASDFMT = "/usr/sbin/dasdfmt";
my $ZIPL = "/usr/sbin/zipl";
my $UDEVADM = "/usr/sbin/udevadm";
my $ZGETDUMP = "/usr/sbin/zgetdump";
# temporary DASD device configuration file for Zipl
my $MDPATH = "/tmp/mvdump.conf.".`mcookie`;
chomp($MDPATH);
my $OPT_DEBUG = 0;
my $OPT_FORCE = 0;
my $OPT_VERBOSE = 0;
sub cleanup
{
# DASD
if (-e $MDPATH) {
system("rm -f $MDPATH");
}
}
sub exit_with
{
my $message = shift();
my $exitcode = shift();
print STDERR "$message Exiting...\n";
cleanup();
# fdasd isn't able to create volume label interactively
# could be fixed with a reformat
if ($exitcode == 65280) {
$exitcode = 12;
}
# bigger exit codes are not supported
if ($exitcode > 255) {
$exitcode = 255;
}
exit($exitcode);
}
sub run_cmd
{
my $cmd = shift();
my $output = "";
if (! $OPT_DEBUG) {
my ($app) = $cmd =~ /\/(\w+) /;
# run command
$output = `$cmd`;
my $exit_code = $?;
# wait for udev to finish processing
system("$UDEVADM settle");
# only print output in case of an error or in verbose mode
if ($output and ($exit_code != 0 or $OPT_VERBOSE)) {
print("$output\n");
}
if ($exit_code != 0) {
exit_with("$app failed with exit code $exit_code", $exit_code);
}
} else {
# only print the command in debug mode
print("\`$cmd\`\n");
}
return($output);
}
sub check_paths
{
for my $path ($BLKID, $PARTED, $FDASD, $DASDVIEW, $DASDFMT, $ZIPL, $UDEVADM, $ZGETDUMP) {
unless ( -x $path) {
exit_with("Command $path is not available.", 13);
}
}
}
sub read_file
{
my $path = shift();
open(my $file, "<", "$path") or exit_with("Unable to access $path: $!.", 15);
my @content = <$file>;
close($file);
# no need for arrays in case of single lines
if (@content > 1) {
return @content;
} else {
chomp($content[0]);
return($content[0]);
}
}
sub is_dasd
{
# remove leading /dev/
my $device = substr(shift(), 5);
if (-r "/sys/block/$device/device/discipline") {
return(1);
} else {
return(0);
}
}
sub has_free_single_kpartx
{
my $device = substr(shift(), 5);
return(0) unless ($device =~ /^dm-[0-9]+$/);
my $blockpath = "/sys/block/$device";
my @holders = glob("$blockpath/holders/*");
return(0) unless (@holders == 1);
my $dmuuid = read_file("$holders[0]/dm/uuid");
return(0) unless ($dmuuid =~ /^part1-mpath-/);
my @holderparts = split(/\//, $holders[0]);
my $holder = "/dev/" . $holderparts[-1];
if(-b $holder and sysopen(my $blockdev, $holder, O_RDWR|O_EXCL)) {
close($blockdev);
return(1);
}
return(0);
}
sub is_zfcp
{
# remove leading /dev/
my $device = substr(shift(), 5);
my $blockpath = "/sys/block/$device";
my $dmname = undef;
# if user passed a device name on cmdline that we listed before
# convert to a dm-[0-9]+ kernel device name
if ($device =~ /^mapper\//) {
$device = substr(readlink("/dev/" . $device), 3);
$blockpath = "/sys/block/" . $device;
}
# check if dm-multipath and get one path member
if ($device =~ /^dm-[0-9]+$/) {
my $dmuuid = read_file("$blockpath/dm/uuid");
return(undef) unless $dmuuid =~ /^mpath-/;
$dmname = read_file("$blockpath/dm/name");
opendir(DIR, "$blockpath/slaves/") or return(undef);
while (defined(my $pathmember = readdir(DIR))) {
# skip ".", "..", or other non scsi disk entries
next unless $pathmember =~ /^sd[a-z]+$/;
$device = $pathmember;
last;
}
closedir(DIR);
}
my $devpath = "/sys/block/$device/device";
unless (-r "$devpath/hba_id" or -r "$devpath/type") {
return(undef);
}
my $devtype = read_file("$devpath/type");
# SCSI type '0' means disk
if ($devtype == 0) {
if (defined $dmname) {
return("/dev/mapper/$dmname");
} else {
return("/dev/$device");
}
} else {
return(undef);
}
}
sub get_partition_num
{
# remove leading /dev/
my $device = substr(shift, 5);
my $part_num = grep(/\s+$device\d+/, read_file("/proc/partitions"));
return($part_num);
}
sub print_device
{
my $device = shift();
my $only_dump_disks = shift();
my $devpath;
if ($device =~ /^\/dev\/mapper\//) {
$devpath = "/sys/block/" . substr(readlink($device), 3);
} else {
$devpath = "/sys/block/" . substr($device, 5);
}
my $output = $device;
my $dump_device = 0;
my $size = int(read_file("$devpath/size") / 2048); # 512 Byte blocks
# size can't be read this way in case of unformatted devices
if ($size != 0) {
$output .= "\t${size}MB";
} else {
$output .= "\tunknown";
}
if (is_dasd($device)) {
my ($busid) = readlink("$devpath/device") =~ /(\w\.\w\.\w{4})/;
$output .= "\t$busid";
# check for dump record and list multi volumes
my $zgetdump_output = `$ZGETDUMP -d $device 2>&1`;
my @dump_devs = $zgetdump_output =~ /(\w\.\w\.\w{4})/g;
if (@dump_devs) {
$dump_device = 1;
$output .= "\tdumpdevice";
# no need to output the dump ids for a single device
if (@dump_devs > 1) {
for my $id (@dump_devs) {
$output .= "|$id";
}
}
} else {
# check for single volume dump devices
if ($zgetdump_output =~ /Single-volume DASD dump tool/) {
$dump_device = 1;
$output .= "\tdumpdevice";
}
}
} else {
# get one path member to fill path info for "yast onpanic"
if ($device =~ /^\/dev\/mapper\//) {
my $blockdev = substr(readlink($device), 3);
my $blockpath = "/sys/block/" . $blockdev;
opendir(DIR, "$blockpath/slaves/") or return(undef);
while (defined(my $pathmember = readdir(DIR))) {
# skip ".", "..", or other non scsi disk entries
next unless $pathmember =~ /^sd[a-z]+$/;
$devpath = "/sys/block/" . $pathmember;
last;
}
closedir(DIR);
}
my $adapter = read_file("$devpath/device/hba_id");
my $wwpn = read_file("$devpath/device/wwpn");
my $lun = read_file("$devpath/device/fcp_lun");
$output .= "\t$adapter\t$wwpn\t$lun";
# check for dump record
my $zgetdump = `$ZGETDUMP -d $device 2>&1`;
if ($? == 0) {
my ($dsize) = ($zgetdump =~ /Maximum dump size\.:\s+([0-9]+) MB/m);
$dsize = $size unless (defined($dsize));
$output = "$device\t${dsize}MB\t$adapter\t$wwpn\t$lun\tdumpdevice";
$dump_device = 1;
}
}
if ($only_dump_disks) {
if ($dump_device) {
print("$output\n");
}
} else {
print("$output\n");
}
}
sub list_free_disks
{
my $devices_ref = shift();
my $type = shift();
if (@$devices_ref) {
for my $device (@$devices_ref) {
print_device($device);
}
} else {
print STDERR "No free $type devices available!\n";
}
}
sub list_dump_disks
{
my @devices = @_;
if (@devices) {
for my $device (@devices) {
print_device($device, 1);
}
} else {
print STDERR "No dump devices available!\n";
}
}
sub determine_free_disks
{
my @dasd;
my @zfcp;
my @devices;
# gather block devices
my $path="/sys/block/";
opendir(DIR, $path) or exit_with("Unable to find $path: $!", 15);
while (defined(my $file = readdir(DIR))) {
# no need to add other devices then dasd* or sd*
# or dm-multipath
if ($file =~ /^dasd[a-z]+$/ or $file =~ /^sd[a-z]+$/ or
$file =~ /^dm-[0-9]+$/) {
push(@devices, $file);
}
}
closedir(DIR);
for my $entry (@devices) {
# only allow disks, no partitions
my ($device) = $entry =~ /^([a-z]+)$/;
# dm devices other than dm-multipath are filtered by is_zfcp()
($device) = $entry =~ /^(dm-[0-9]+)$/ unless ($device);
next unless ($device);
$device = "/dev/$device";
# determine if the block device could be accessed exclusively
if(-b $device and sysopen(my $blockdev, $device, O_RDWR|O_EXCL)) {
close($blockdev);
if (is_dasd($device)) {
push(@dasd, $device);
}
my $zfcp = is_zfcp($device);
if (defined $zfcp) {
push(@zfcp, $zfcp);
}
} else {
# A dm-multipath device with a single holder
# being a kpartx partition number 1 could still
# be free or contain a zfcpdump boot record.
# Due to the kpartx linear dm mapping, such
# dm-multipath device cannot open exclusively.
if (has_free_single_kpartx($device)) {
my $zfcp = is_zfcp($device);
if (defined $zfcp) {
push(@zfcp, $zfcp);
}
}
}
# wait for udev to process all events triggered by sysopen(,O_EXCL)
system("$UDEVADM settle");
}
return(\@dasd, \@zfcp);
}
sub prepare_dasd
{
my @devices = @_;
my $format_disks = "";
# check formatting
for my $device (@devices) {
# determine disk layout
my ($fmtstr) = `$DASDVIEW -x $device` =~ /(\w\w\w) formatted/;
SWITCH:
for($fmtstr) {
if (/NOT/) {
print("Unformatted disk, formatting $device.\n");
$format_disks .= " $device";
last SWITCH;
}
if (/LDL/) {
if ($OPT_FORCE) {
print("Linux disk layout, reformatting $device.\n");
$format_disks .= " $device";
} else {
print("$device was formatted with the Linux disk layout.\n");
print("Unable to use it without reformatting.\n");
exit_with("Re-issue the mkdump command with the --force option.", 12);
}
last SWITCH;
}
if (/CDL/) {
# allow reformatting with force, since fdasd isn't able to create volume label interactively
if ($OPT_FORCE) {
print("Compatible disk layout, force reformatting $device.\n");
$format_disks .= " $device";
} else {
print("$device: Compatible disk layout, Ok to use.\n");
}
last SWITCH;
}
exit_with("Unknown layout ($fmtstr), cannot use disk.", 11);
}
}
# format devices
if ($format_disks) {
#up to eight devices in parallel
run_cmd("$DASDFMT -P 8 -b 4096 -y -f $format_disks");
}
# check partitioning and partition
for my $device (@devices) {
my $part_num = get_partition_num($device);
if ($part_num == 0 or $OPT_FORCE) {
print("Re-partitioning disk $device.\n");
run_cmd("$FDASD -a $device");
} else {
# allow disk with one partition if it don't consist a file system
if ($part_num == 1) {
my ($fstype) = `$BLKID ${device}1` =~ /TYPE=\"(\w+)\"/;
if ($fstype) {
exit_with("Device ${device}1 already contains a filesystem of type $fstype.", 12);
}
} else {
exit_with("$part_num partitions detected, cannot use disk $device.", 12);
}
}
}
}
sub setup_dasddump
{
my @devices = @_;
prepare_dasd(@devices);
# create zipl device configuration file
# don't create files in debug mode
unless ($OPT_DEBUG) {
open(my $file, ">", $MDPATH) or exit_with("Unable to access $MDPATH: $!.", 15);
for my $device (@devices) {
print{$file}("${device}1\n");
}
close($file);
}
print("Creating dump record.\n");
run_cmd("${ZIPL} -V -n -M $MDPATH");
cleanup();
}
sub setup_zfcpdump
{
my $device = shift();
# check partitioning
my $part_num = get_partition_num($device);
if ($part_num == 0 or $OPT_FORCE) {
print("Re-partitioning disk $device.\n");
run_cmd("$PARTED -s -- $device mklabel gpt mkpart primary 0 -1");
} else {
if ($part_num > 1) {
exit_with("$part_num partitions detected, cannot use disk $device.", 12);
}
}
# install bootloader
print("Creating dump record.\n");
my $partdev;
if ($device =~ /^\/dev\/mapper\//) {
$partdev = $device . "-part1"; # kpartx partition on multipath
} else {
$partdev = $device . "1"; # real partition, single path SCSI
}
run_cmd("${ZIPL} -V -d ${partdev}");
cleanup();
}
sub print_version
{
print << "EOF";
mkdump $VERSION
Copyright (c) 2011 SUSE LINUX Products GmbH
License GPLv2 or (at your option) any later version.
<http://www.gnu.org/licenses/gpl-2.0.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.
Written by Tim Hardeck <thardeck\@suse.de>.
EOF
exit(0);
}
sub print_usage
{
my $exitcode = shift();
print << "EOF";
Usage: mkdump [OPTIONS] [DEVICE]...
mkdump $VERSION
Prepare one or more volumes for use as S/390 dump device. Supported devices
are ECKD DASD and SCSI over zFCP disks, while multi-volumes are limited to DASD.
Only whole disks can be used, no partitions! If the device is incompatible
formatted/partioned, the script will refuse to install the dump record
unless the --force switch is given.
Disks which are in use or have mounted partitions will not be listed and can't be used.
The mentioning of "dumpdevice" after a disk indicates that it is an already usable dump device. Additionally multi-volume dump devices are indicated by the list of including DASD ids.
Options:
-h, --help display this help and exit
-V, --version display version information and exit
-d, --debug debug mode, do not run programs which commit changes
-v, --verbose be verbose and show command outputs
-f, --force force overwrite of the disk
-l, --list-dump display dump disks
-D, --list-dasd display usable DASD disks (Device, Size, ID, Dump)
-Z, --list-zfcp display usable SCSI over zFCP disks (Device, Size, ID, WWPN, LUN, Dump)
Report bugs on https://bugzilla.novell.com/
EOF
exit($exitcode);
}
sub analyze_cmd_parameters
{
#verbose, debug and force are global
my $opt_help = 0;
my $opt_version = 0;
my $opt_dump = 0;
my $opt_dasd = 0;
my $opt_zfcp = 0;
if (@ARGV == 0) {
print_usage(14);
}
Getopt::Long::Configure('bundling');
GetOptions(
'h|help' => \$opt_help,
'V|version' => \$opt_version,
'd|debug' => \$OPT_DEBUG,
'v|verbose' => \$OPT_VERBOSE,
'f|force' => \$OPT_FORCE,
'l|list-dump' => \$opt_dump,
'D|list-dasd' => \$opt_dasd,
'Z|list-zfcp' => \$opt_zfcp,
) or print_usage(14);
if ($opt_help) {
print_usage(0);
}
if ($opt_version) {
print_version();
}
# determine free dasd and zfcp devices
my ($dasd_ref, $zfcp_ref) = determine_free_disks();
if ($opt_dump) {
list_dump_disks(@$dasd_ref, @$zfcp_ref);
exit 0;
}
if ($opt_dasd) {
list_free_disks(\@$dasd_ref, "dasd");
}
if ($opt_zfcp) {
list_free_disks(\@$zfcp_ref, "zfcp");
}
# allow listing of both device types at the same time
if ($opt_dasd or $opt_zfcp) {
exit 0;
}
# check provided devices and be strict
my @devices;
for my $device (@ARGV) {
if (grep(/$device/, @devices)) {
exit_with("$device is mentioned more than once.", 14);
}
# dm devices other than dm-multipath are filtered by is_zfcp()
if ( $device =~ /^\/dev\/[a-z]+$/ == 0 and
$device !~ /^\/dev\/mapper\// ) {
exit_with("The device parameter $device is inaccurate. Only whole disks are allowed.", 14);
}
if (grep(/$device/, (@$dasd_ref, @$zfcp_ref))) {
my $zfcp = is_zfcp($device);
if (defined $zfcp and @ARGV > 1) {
exit_with("Multi-volume dumps aren't supported with zFCP.", 14);
}
push(@devices, (defined $zfcp) ? $zfcp : $device);
} else {
if (-b $device) {
exit_with("$device is in use or not a DASD/zFCP disk!", 14);
} else {
exit_with("$device does not exist!", 14);
}
}
}
if (@devices == 0) {
exit_with("No usable devices where provided.", 14);
}
return(@devices);
}
sub main
{
check_paths();
my @devices = analyze_cmd_parameters();
# only one dump device is possible with zFCP which is enforced in analyze_cmd_parameters
my $zfcp = is_zfcp($devices[0]);
if (defined $zfcp) {
setup_zfcpdump($zfcp);
} else {
setup_dasddump(@devices);
}
print("Creating the dump device was successful.\n");
}
main();

666
mkdump.pl.suse Normal file
View File

@ -0,0 +1,666 @@
#!/usr/bin/perl
########################################################################
#
# mkdump.pl - Preparing disks for use as S/390 dump device
#
# Copyright (c) 2011 Tim Hardeck, SUSE LINUX Products GmbH
# Copyright (c) 2017 SUSE LINUX GmbH, Nuernberg, Germany.
# Based on mkdump.sh (c) 2004 Hannes Reinecke, SuSE AG
#
# License:
#
# This program is free software; you can redistribute it and/or
# modify it under the terms of the GNU General Public License
# as published by the Free Software Foundation; either version 2
# of the License, or (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software
# Foundation, Inc., 51 Franklin Street, Fifth Floor,
# Boston, MA 02110-1301 USA
#
########################################################################
use strict;
use warnings;
use Fcntl;
use Getopt::Long;
my $VERSION = "2.0.4";
my $BLKID = "/sbin/blkid";
my $PARTED = "/usr/sbin/parted";
my $FDASD = "/sbin/fdasd";
my $DASDVIEW = "/sbin/dasdview";
my $DASDFMT = "/sbin/dasdfmt";
my $ZIPL = "/sbin/zipl";
my $UDEVADM = "/sbin/udevadm";
my $ZGETDUMP = "/sbin/zgetdump";
# temporary DASD device configuration file for Zipl
my $MDPATH = "/tmp/mvdump.conf.".`mcookie`;
chomp($MDPATH);
my $OPT_DEBUG = 0;
my $OPT_FORCE = 0;
my $OPT_VERBOSE = 0;
sub cleanup
{
# DASD
if (-e $MDPATH) {
system("rm -f $MDPATH");
}
}
sub exit_with
{
my $message = shift();
my $exitcode = shift();
print STDERR "$message Exiting...\n";
cleanup();
# fdasd isn't able to create volume label interactively
# could be fixed with a reformat
if ($exitcode == 65280) {
$exitcode = 12;
}
# bigger exit codes are not supported
if ($exitcode > 255) {
$exitcode = 255;
}
exit($exitcode);
}
sub run_cmd
{
my $cmd = shift();
my $output = "";
if (! $OPT_DEBUG) {
my ($app) = $cmd =~ /\/(\w+) /;
# run command
$output = `$cmd`;
my $exit_code = $?;
# wait for udev to finish processing
system("$UDEVADM settle");
# only print output in case of an error or in verbose mode
if ($output and ($exit_code != 0 or $OPT_VERBOSE)) {
print("$output\n");
}
if ($exit_code != 0) {
exit_with("$app failed with exit code $exit_code", $exit_code);
}
} else {
# only print the command in debug mode
print("\`$cmd\`\n");
}
return($output);
}
sub check_paths
{
for my $path ($BLKID, $PARTED, $FDASD, $DASDVIEW, $DASDFMT, $ZIPL, $UDEVADM, $ZGETDUMP) {
unless ( -x $path) {
exit_with("Command $path is not available.", 13);
}
}
}
sub read_file
{
my $path = shift();
open(my $file, "<", "$path") or exit_with("Unable to access $path: $!.", 15);
my @content = <$file>;
close($file);
# no need for arrays in case of single lines
if (@content > 1) {
return @content;
} else {
chomp($content[0]);
return($content[0]);
}
}
sub is_dasd
{
# remove leading /dev/
my $device = substr(shift(), 5);
if (-r "/sys/block/$device/device/discipline") {
return(1);
} else {
return(0);
}
}
sub has_free_single_kpartx
{
my $device = substr(shift(), 5);
return(0) unless ($device =~ /^dm-[0-9]+$/);
my $blockpath = "/sys/block/$device";
my @holders = glob("$blockpath/holders/*");
return(0) unless (@holders == 1);
my $dmuuid = read_file("$holders[0]/dm/uuid");
return(0) unless ($dmuuid =~ /^part1-mpath-/);
my @holderparts = split(/\//, $holders[0]);
my $holder = "/dev/" . $holderparts[-1];
if(-b $holder and sysopen(my $blockdev, $holder, O_RDWR|O_EXCL)) {
close($blockdev);
return(1);
}
return(0);
}
sub is_zfcp
{
# remove leading /dev/
my $device = substr(shift(), 5);
my $blockpath = "/sys/block/$device";
my $dmname = undef;
# if user passed a device name on cmdline that we listed before
# convert to a dm-[0-9]+ kernel device name
if ($device =~ /^mapper\//) {
$device = substr(readlink("/dev/" . $device), 3);
$blockpath = "/sys/block/" . $device;
}
# check if dm-multipath and get one path member
if ($device =~ /^dm-[0-9]+$/) {
my $dmuuid = read_file("$blockpath/dm/uuid");
return(undef) unless $dmuuid =~ /^mpath-/;
$dmname = read_file("$blockpath/dm/name");
opendir(DIR, "$blockpath/slaves/") or return(undef);
while (defined(my $pathmember = readdir(DIR))) {
# skip ".", "..", or other non scsi disk entries
next unless $pathmember =~ /^sd[a-z]+$/;
$device = $pathmember;
last;
}
closedir(DIR);
}
my $devpath = "/sys/block/$device/device";
unless (-r "$devpath/hba_id" or -r "$devpath/type") {
return(undef);
}
my $devtype = read_file("$devpath/type");
# SCSI type '0' means disk
if ($devtype == 0) {
if (defined $dmname) {
return("/dev/mapper/$dmname");
} else {
return("/dev/$device");
}
} else {
return(undef);
}
}
sub get_partition_num
{
# remove leading /dev/
my $device = substr(shift, 5);
my $part_num = grep(/\s+$device\d+/, read_file("/proc/partitions"));
return($part_num);
}
sub print_device
{
my $device = shift();
my $only_dump_disks = shift();
my $devpath;
if ($device =~ /^\/dev\/mapper\//) {
$devpath = "/sys/block/" . substr(readlink($device), 3);
} else {
$devpath = "/sys/block/" . substr($device, 5);
}
my $output = $device;
my $dump_device = 0;
my $size = int(read_file("$devpath/size") / 2048); # 512 Byte blocks
# size can't be read this way in case of unformatted devices
if ($size != 0) {
$output .= "\t${size}MB";
} else {
$output .= "\tunknown";
}
if (is_dasd($device)) {
my ($busid) = readlink("$devpath/device") =~ /(\w\.\w\.\w{4})/;
$output .= "\t$busid";
# check for dump record and list multi volumes
my $zgetdump_output = `$ZGETDUMP -d $device 2>&1`;
my @dump_devs = $zgetdump_output =~ /(\w\.\w\.\w{4})/g;
if (@dump_devs) {
$dump_device = 1;
$output .= "\tdumpdevice";
# no need to output the dump ids for a single device
if (@dump_devs > 1) {
for my $id (@dump_devs) {
$output .= "|$id";
}
}
} else {
# check for single volume dump devices
if ($zgetdump_output =~ /Single-volume DASD dump tool/) {
$dump_device = 1;
$output .= "\tdumpdevice";
}
}
} else {
# get one path member to fill path info for "yast onpanic"
if ($device =~ /^\/dev\/mapper\//) {
my $blockdev = substr(readlink($device), 3);
my $blockpath = "/sys/block/" . $blockdev;
opendir(DIR, "$blockpath/slaves/") or return(undef);
while (defined(my $pathmember = readdir(DIR))) {
# skip ".", "..", or other non scsi disk entries
next unless $pathmember =~ /^sd[a-z]+$/;
$devpath = "/sys/block/" . $pathmember;
last;
}
closedir(DIR);
}
my $adapter = read_file("$devpath/device/hba_id");
my $wwpn = read_file("$devpath/device/wwpn");
my $lun = read_file("$devpath/device/fcp_lun");
$output .= "\t$adapter\t$wwpn\t$lun";
# check for dump record
my $zgetdump = `$ZGETDUMP -d $device 2>&1`;
if ($? == 0) {
my ($dsize) = ($zgetdump =~ /Maximum dump size\.:\s+([0-9]+) MB/m);
$dsize = $size unless (defined($dsize));
$output = "$device\t${dsize}MB\t$adapter\t$wwpn\t$lun\tdumpdevice";
$dump_device = 1;
}
}
if ($only_dump_disks) {
if ($dump_device) {
print("$output\n");
}
} else {
print("$output\n");
}
}
sub list_free_disks
{
my $devices_ref = shift();
my $type = shift();
if (@$devices_ref) {
for my $device (@$devices_ref) {
print_device($device);
}
} else {
print STDERR "No free $type devices available!\n";
}
}
sub list_dump_disks
{
my @devices = @_;
if (@devices) {
for my $device (@devices) {
print_device($device, 1);
}
} else {
print STDERR "No dump devices available!\n";
}
}
sub determine_free_disks
{
my @dasd;
my @zfcp;
my @devices;
# gather block devices
my $path="/sys/block/";
opendir(DIR, $path) or exit_with("Unable to find $path: $!", 15);
while (defined(my $file = readdir(DIR))) {
# no need to add other devices then dasd* or sd*
# or dm-multipath
if ($file =~ /^dasd[a-z]+$/ or $file =~ /^sd[a-z]+$/ or
$file =~ /^dm-[0-9]+$/) {
push(@devices, $file);
}
}
closedir(DIR);
for my $entry (@devices) {
# only allow disks, no partitions
my ($device) = $entry =~ /^([a-z]+)$/;
# dm devices other than dm-multipath are filtered by is_zfcp()
($device) = $entry =~ /^(dm-[0-9]+)$/ unless ($device);
next unless ($device);
$device = "/dev/$device";
# determine if the block device could be accessed exclusively
if(-b $device and sysopen(my $blockdev, $device, O_RDWR|O_EXCL)) {
close($blockdev);
if (is_dasd($device)) {
push(@dasd, $device);
}
my $zfcp = is_zfcp($device);
if (defined $zfcp) {
push(@zfcp, $zfcp);
}
} else {
# A dm-multipath device with a single holder
# being a kpartx partition number 1 could still
# be free or contain a zfcpdump boot record.
# Due to the kpartx linear dm mapping, such
# dm-multipath device cannot open exclusively.
if (has_free_single_kpartx($device)) {
my $zfcp = is_zfcp($device);
if (defined $zfcp) {
push(@zfcp, $zfcp);
}
}
}
# wait for udev to process all events triggered by sysopen(,O_EXCL)
system("$UDEVADM settle");
}
return(\@dasd, \@zfcp);
}
sub prepare_dasd
{
my @devices = @_;
my $format_disks = "";
# check formatting
for my $device (@devices) {
# determine disk layout
my ($fmtstr) = `$DASDVIEW -x $device` =~ /(\w\w\w) formatted/;
SWITCH:
for($fmtstr) {
if (/NOT/) {
print("Unformatted disk, formatting $device.\n");
$format_disks .= " $device";
last SWITCH;
}
if (/LDL/) {
if ($OPT_FORCE) {
print("Linux disk layout, reformatting $device.\n");
$format_disks .= " $device";
} else {
print("$device was formatted with the Linux disk layout.\n");
print("Unable to use it without reformatting.\n");
exit_with("Re-issue the mkdump command with the --force option.", 12);
}
last SWITCH;
}
if (/CDL/) {
# allow reformatting with force, since fdasd isn't able to create volume label interactively
if ($OPT_FORCE) {
print("Compatible disk layout, force reformatting $device.\n");
$format_disks .= " $device";
} else {
print("$device: Compatible disk layout, Ok to use.\n");
}
last SWITCH;
}
exit_with("Unknown layout ($fmtstr), cannot use disk.", 11);
}
}
# format devices
if ($format_disks) {
#up to eight devices in parallel
run_cmd("$DASDFMT -P 8 -b 4096 -y -f $format_disks");
}
# check partitioning and partition
for my $device (@devices) {
my $part_num = get_partition_num($device);
if ($part_num == 0 or $OPT_FORCE) {
print("Re-partitioning disk $device.\n");
run_cmd("$FDASD -a $device");
} else {
# allow disk with one partition if it don't consist a file system
if ($part_num == 1) {
my ($fstype) = `$BLKID ${device}1` =~ /TYPE=\"(\w+)\"/;
if ($fstype) {
exit_with("Device ${device}1 already contains a filesystem of type $fstype.", 12);
}
} else {
exit_with("$part_num partitions detected, cannot use disk $device.", 12);
}
}
}
}
sub setup_dasddump
{
my @devices = @_;
prepare_dasd(@devices);
# create zipl device configuration file
# don't create files in debug mode
unless ($OPT_DEBUG) {
open(my $file, ">", $MDPATH) or exit_with("Unable to access $MDPATH: $!.", 15);
for my $device (@devices) {
print{$file}("${device}1\n");
}
close($file);
}
print("Creating dump record.\n");
run_cmd("${ZIPL} -V -n -M $MDPATH");
cleanup();
}
sub setup_zfcpdump
{
my $device = shift();
# check partitioning
my $part_num = get_partition_num($device);
if ($part_num == 0 or $OPT_FORCE) {
print("Re-partitioning disk $device.\n");
run_cmd("$PARTED -s -- $device mklabel gpt mkpart primary 0 -1");
} else {
if ($part_num > 1) {
exit_with("$part_num partitions detected, cannot use disk $device.", 12);
}
}
# install bootloader
print("Creating dump record.\n");
my $partdev;
if ($device =~ /^\/dev\/mapper\//) {
$partdev = $device . "-part1"; # kpartx partition on multipath
} else {
$partdev = $device . "1"; # real partition, single path SCSI
}
run_cmd("${ZIPL} -V -d ${partdev}");
cleanup();
}
sub print_version
{
print << "EOF";
mkdump $VERSION
Copyright (c) 2011 SUSE LINUX Products GmbH
License GPLv2 or (at your option) any later version.
<http://www.gnu.org/licenses/gpl-2.0.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.
Written by Tim Hardeck <thardeck\@suse.de>.
EOF
exit(0);
}
sub print_usage
{
my $exitcode = shift();
print << "EOF";
Usage: mkdump [OPTIONS] [DEVICE]...
mkdump $VERSION
Prepare one or more volumes for use as S/390 dump device. Supported devices
are ECKD DASD and SCSI over zFCP disks, while multi-volumes are limited to DASD.
Only whole disks can be used, no partitions! If the device is incompatible
formatted/partioned, the script will refuse to install the dump record
unless the --force switch is given.
Disks which are in use or have mounted partitions will not be listed and can't be used.
The mentioning of "dumpdevice" after a disk indicates that it is an already usable dump device. Additionally multi-volume dump devices are indicated by the list of including DASD ids.
Options:
-h, --help display this help and exit
-V, --version display version information and exit
-d, --debug debug mode, do not run programs which commit changes
-v, --verbose be verbose and show command outputs
-f, --force force overwrite of the disk
-l, --list-dump display dump disks
-D, --list-dasd display usable DASD disks (Device, Size, ID, Dump)
-Z, --list-zfcp display usable SCSI over zFCP disks (Device, Size, ID, WWPN, LUN, Dump)
Report bugs on https://bugzilla.novell.com/
EOF
exit($exitcode);
}
sub analyze_cmd_parameters
{
#verbose, debug and force are global
my $opt_help = 0;
my $opt_version = 0;
my $opt_dump = 0;
my $opt_dasd = 0;
my $opt_zfcp = 0;
if (@ARGV == 0) {
print_usage(14);
}
Getopt::Long::Configure('bundling');
GetOptions(
'h|help' => \$opt_help,
'V|version' => \$opt_version,
'd|debug' => \$OPT_DEBUG,
'v|verbose' => \$OPT_VERBOSE,
'f|force' => \$OPT_FORCE,
'l|list-dump' => \$opt_dump,
'D|list-dasd' => \$opt_dasd,
'Z|list-zfcp' => \$opt_zfcp,
) or print_usage(14);
if ($opt_help) {
print_usage(0);
}
if ($opt_version) {
print_version();
}
# determine free dasd and zfcp devices
my ($dasd_ref, $zfcp_ref) = determine_free_disks();
if ($opt_dump) {
list_dump_disks(@$dasd_ref, @$zfcp_ref);
exit 0;
}
if ($opt_dasd) {
list_free_disks(\@$dasd_ref, "dasd");
}
if ($opt_zfcp) {
list_free_disks(\@$zfcp_ref, "zfcp");
}
# allow listing of both device types at the same time
if ($opt_dasd or $opt_zfcp) {
exit 0;
}
# check provided devices and be strict
my @devices;
for my $device (@ARGV) {
if (grep(/$device/, @devices)) {
exit_with("$device is mentioned more than once.", 14);
}
# dm devices other than dm-multipath are filtered by is_zfcp()
if ( $device =~ /^\/dev\/[a-z]+$/ == 0 and
$device !~ /^\/dev\/mapper\// ) {
exit_with("The device parameter $device is inaccurate. Only whole disks are allowed.", 14);
}
if (grep(/$device/, (@$dasd_ref, @$zfcp_ref))) {
my $zfcp = is_zfcp($device);
if (defined $zfcp and @ARGV > 1) {
exit_with("Multi-volume dumps aren't supported with zFCP.", 14);
}
push(@devices, (defined $zfcp) ? $zfcp : $device);
} else {
if (-b $device) {
exit_with("$device is in use or not a DASD/zFCP disk!", 14);
} else {
exit_with("$device does not exist!", 14);
}
}
}
if (@devices == 0) {
exit_with("No usable devices where provided.", 14);
}
return(@devices);
}
sub main
{
check_paths();
my @devices = analyze_cmd_parameters();
# only one dump device is possible with zFCP which is enforced in analyze_cmd_parameters
my $zfcp = is_zfcp($devices[0]);
if (defined $zfcp) {
setup_zfcpdump($zfcp);
} else {
setup_dasddump(@devices);
}
print("Creating the dump device was successful.\n");
}
main();

10
pkey.conf Normal file
View File

@ -0,0 +1,10 @@
#
# Copyright (c) 2018-2024 SUSE LINUX GmbH, Nuernberg, Germany.
# All rights reserved.
#
# load pkey module at boot time
pkey
pkey_cca
pkey_ep11
pkey_pckmo

174
qeth_configure Normal file
View File

@ -0,0 +1,174 @@
#! /bin/sh
#
# qeth_configure
#
# Copyright (c) 2017 SUSE LINUX GmbH, Nuernberg, Germany.
#
# Configures a qeth device by calling the IBM-provided chzdev command.
# Whereas this script used to validate the parameters provided to it,
# we now rely on chzdev to do that instead. The script is intended only
# as a "translation layer" to provide backward compatability for older
# scripts and tools that invoke it.
#
# Usage:
# qeth_configure [-i] [-l] [-f -t <CARDTYPE> ] [-o "Values"] -n <portno> -p <portname> <read chan> <write chan> <data chan> <online>
#
# -i Configure IP takeover
# -l Configure Layer2 support
# -f Override safety checks
# -t Valid cardtypes are: qeth, hsi - Deprecated
# -o General QETH options, separated by spaces
# -n QETH port number to use, 0 or 1. Only needed for real, not virtual
# devices.
# -p QETH Portname to use - Deprecated. OSAs no longer need a port name.
# read/write/data chan = x.y.ssss where
# x is always 0 until IBM creates something that
# uses that number
# y is the logical channel subsystem (lcss)
# number. Most often this is 0, but it could
# be non-zero
# ssss is the four digit subchannel address of
# the device, in hexidecimal, with leading
# zeros.
# online = 0 to take the device offline
# 1 to bring the device online
#
# Return values:
# Return codes are determined by the chzdev command.
#
mesg () {
echo "$@"
}
debug_mesg () {
case "${DEBUG}" in
yes) mesg "$@" ;;
*) ;;
esac
}
add_cio_channel() {
echo "$* # ${DATE}" >> /boot/zipl/active_devices.txt
}
remove_cio_channel() {
[ -w /boot/zipl/active_devices.txt ] && sed -i -e "/^${1}/d" /boot/zipl/active_devices.txt
}
usage(){
echo "Usage: ${0} [options] <read chan> <write chan> <data chan> <online>"
echo " -i Configure IP takeover"
echo " -l Configure Layer2 support"
echo " -f Override safety checks"
echo " -t Valid cardtypes are: qeth, hsi - Deprecated."
echo " -o General QETH options, separated by spaces"
echo " -n QETH port number to use, 0 or 1. Only needed for real, not virtual"
echo " devices."
echo " -p QETH Portname to use - Deprecated. OSAs no longer need a port name."
echo " read/write/data chan = x.y.ssss where"
echo " x is always 0 until IBM creates something that"
echo " uses that number"
echo " y is the logical channel subsystem (lcss)"
echo " number. Most often this is 0, but it could"
echo " be non-zero"
echo " ssss is the four digit subchannel address of"
echo " the device, in hexidecimal, with leading"
echo " zeros."
echo " online = 0 to take the device offline"
echo " 1 to bring the device online"
}
if [ "${DEBUG}" != "yes" ]; then
DEBUG="no"
fi
DATE=$(date)
############################################################
# Parse the parameters from the command line
#
ARGS=$(getopt --options ifln:o:p:t: -n "qeth_configure" -- "$@")
if [ $? != 0 ] ; then echo "Terminating..." >&2 ; exit 1 ; fi
eval set -- "${ARGS}"
debug_mesg "All the parms passed were ${ARGS}"
# Set some defaults
LAYER_MODE="layer2=0"
while true; do
case "${1}" in
-i) debug_mesg "Configure IP takeover"
PARM_LIST="${PARM_LIST} ipa_takeover/enable=1"
shift 1
;;
-f) debug_mesg "This used to mean udev rules will always be generated."
debug_mesg "For chzdev, it means safety checks will be overridden."
debug_mesg "Kinda sorta the same thing, really."
PARM_LIST="${PARM_LIST} -f"
shift 1
;;
-l) debug_mesg "Configure Layer 2 support"
LAYER_MODE="layer2=1"
shift 1
;;
-n) debug_mesg "Set QETH port number to ${2}"
PARM_LIST="${PARM_LIST} portno=${2}"
shift 2
;;
-o) debug_mesg "Add the following arbitrary parms: ${2}"
PARM_LIST="${PARM_LIST} ${2}"
shift 2
;;
-p) debug_mesg "QETH Port name is no longer used, don't specify it: ${2}"
shift 2
;;
-t) debug_mesg "This used to set the card type to ${2}"
debug_mesg "Now it gets ignored."
shift 2
;;
--) debug_mesg "Found the end of parms indicator: --"
shift 1
break
;;
*) debug_mesg "At the catch-all select entry"
debug_mesg "What was selected was ${1}"
shift 1
;;
esac
done
QETH_READ_CHAN=${1}
QETH_WRITE_CHAN=${2}
QETH_DATA_CHAN=${3}
ON_OFF=${4}
if [ -z "${QETH_READ_CHAN}" ] || [ -z "${QETH_WRITE_CHAN}" ] || [ -z "${QETH_DATA_CHAN}" ] || [ -z "${ON_OFF}" ]; then
mesg "You didn't specify all the needed parameters."
usage
exit 1
fi
if [ "${ON_OFF}" == 0 ]; then
debug_mesg "chzdev -d qeth --no-root-update ${QETH_READ_CHAN}"
chzdev -d qeth --no-root-update ${QETH_READ_CHAN}
elif [ "${ON_OFF}" == 1 ]; then
debug_mesg "chzdev -e qeth --no-root-update ${LAYER_MODE} ${PARM_LIST} ${QETH_READ_CHAN}"
chzdev -e qeth ${LAYER_MODE} --no-root-update ${PARM_LIST} ${QETH_READ_CHAN}
else mesg "You must specify a 0 or a 1 for the online/offline attribute."
usage
exit 1
fi
RC=${?}
if [ ${RC} -ne 0 ]; then
exit ${RC}
fi
if [ ${ON_OFF} == 1 ]; then
add_cio_channel "${QETH_READ_CHAN},${QETH_WRITE_CHAN},${QETH_DATA_CHAN}"
else remove_cio_channel "${QETH_READ_CHAN}"
remove_cio_channel "${QETH_WRITE_CHAN}"
remove_cio_channel "${QETH_DATA_CHAN}"
fi

66
qeth_configure.8 Normal file
View File

@ -0,0 +1,66 @@
.TH qeth_configure "8" "July 2013" "s390-tools"
.SH NAME
qeth_configure \- Configures or deconfigures a HiperSocket adapter or an IBM Open Systems Adapter (OSA) in QDIO mode
.SH SYNOPSIS
.B qeth_configure [options] read_channel write_channel data_channel online
.SH DESCRIPTION
.B qeth_configure
is intended to make it easy to persistently add and remove HiperSocket Adapters and Open System Adapters (OSAs) that are in QDIO mode. In addition to bringing the adapter online or offline, it will also create or delete the necessary udev rules for the adapter.
.SH PARAMETERS
.IP read_channel
The device number of the read channel of the adapter. Takes the form x.y.ssss.
.IP write_channel
The device number of the write channel of the adapter.Takes the form x.y.ssss.
.IP data_channel
The device number of the data channel of the adapter.Takes the form x.y.ssss.
.RS
where
.RS
.B x
is always 0 until IBM creates something that uses that number.
.RE
.RS
.B y
is the logical channel subsystem (lcss) number. Most often this is 0, but it could be non-zero.
.RE
.RS
.B ssss
is the four digit subchannel address of the device, in hexidecimal, with leading zeros. If entered in upper/mixed case, this is automatically converted to lower case.
.RE
.RE
.RS
.RE
.RE
.IP online
Either a literal 1 to bring the adapter online or a literal 0 to take it offline
.SH OPTIONAL PARAMETERS
.IP -i
Configure IP takeover
.IP -l
Configure Layer 2 support
.IP -f
Force creation of udev rules, do not check values in /sys. Requires -t to be specfied.
.IP "-t CARDTYPE"
The type of card being configured. Valid values are: qeth, hsi, or osn.
.IP "-o ""Values"""
General/arbitrary QETH options, separated by spaces
.IP "-n portnumber"
QETH port number to use, 0 or 1. Only needed for real, not virtual devices.
.IP "-p portname"
QETH Portname to use. Only needed if sharing a real OSA with z/OS.
.SH FILES
Please see the documentation of
.B chzdev.
.SH ENVIRONMENT
.IP DEBUG
If set to "yes" some minimal debugging information is output during execution.
.SH DIAGNOSTICS
Messages and return codes are determined by the
.B chzdev
command.
If environment variable DEBUG is set to "yes," it shows the command line of the invoked
.B chzdev,
and a message for each command line option is issued on stdout.
.SH BUGS
Gotta be some, I'm sure. If you find one, please open a bug report.

50
read_values.8 Normal file
View File

@ -0,0 +1,50 @@
.TH read_values "8" "March 2015" "s390-tools"
.SH NAME
read_values \- Read information from the /sys and /proc filesystems for SUSE Customer Center (SCC) and the like.
.SH SYNOPSIS
.B read_values [-s] [-u] [-c] [-a Attribute] [-L Keyword] [-d debug] [-h]
.SH DESCRIPTION
.B read_values
is intended to make it easy to read values from the /sys and /proc filesystems. Those values may be at different places depending on the machine type and the kernel version.
.SH PARAMETERS
.IP -s
Outputs the values needed by the SCC (SUSE Customer Center)
.IP -u
Creates a uuid for this system
.IP -c
Prints the CPU type of the current system
.IP -a Attribute
Prints the value of the
.B Attribute
.IP -L Keyword
The
.B Keyword
may be
.B Attribute
or
.B Recognised.
With this option you get a list of all
.B Attributes
the programm can accept (
.B Attribute
) or a list of attributes which in turn can be used for the option -L (
.B Recognised
).
.SH FILES
.I /sys and /proc
.SH DIAGNOSTICS
The following messages may be issued on stderr:
.IP
.B Unable to open /proc/sysinfo
or
.B Unable to open sysinfo.zvm
.RS
The named file cannot be opened. This means the tool can't do anything useful. Return code 99 is set.
.RE
.IP
.B Only one of the options a, c, L, s or u can be specified.
.RS
Only one of the options a, c, L, s or u can be specified at a time. Return code 1 is set.
.RE
.SH BUGS
Gotta be some, I'm sure. If you find one, please open a bug report.

628
read_values.c Normal file
View File

@ -0,0 +1,628 @@
/********************************************************************************/
/* */
/* Copyright (C) 2014-2015, 2019-2023 SUSE LLC */
/* */
/* All rights reserved.
Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:
1. Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
2. Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#define _GNU_SOURCE
#include <sys/utsname.h>
#include <ctype.h>
#include <stddef.h>
#include <stdio.h>
#include <stdlib.h>
#include <errno.h>
#include <string.h>
#include <unistd.h>
#include <query_capacity.h>
/*
* Data types
*/
enum datatypes {
integer,
string,
floatingpoint
};
#define WITHOUT_KEY 0
#define WITH_KEY 1
static char *versionstring = "Version 1.0.5 2024-06-20 14:30";
static char *version = "1.0.5";
void *configuration_handle = NULL;
int layers = -1;
/*
* List of machine types
*/
struct machinetype {
enum qc_model_families model_families;
char *typenumber;
char *fullname;
} machinetypes[] = {
{ QC_TYPE_FAMILY_IBMZ, "2064", "2064 = z900 IBM eServer zSeries 900" },
{ QC_TYPE_FAMILY_IBMZ, "2066", "2066 = z800 IBM eServer zSeries 800" },
{ QC_TYPE_FAMILY_IBMZ, "2084", "2084 = z990 IBM eServer zSeries 990" },
{ QC_TYPE_FAMILY_IBMZ, "2086", "2086 = z890 IBM eServer zSeries 890" },
{ QC_TYPE_FAMILY_IBMZ, "2094", "2094 = z9-EC IBM System z9 Enterprise Class" },
{ QC_TYPE_FAMILY_IBMZ, "2096", "2096 = z9-BC IBM System z9 Business Class" },
{ QC_TYPE_FAMILY_IBMZ, "2097", "2097 = z10-EC IBM System z10 Enterprise Class" },
{ QC_TYPE_FAMILY_IBMZ, "2098", "2098 = z10-BC IBM System z10 Business Class" },
{ QC_TYPE_FAMILY_IBMZ, "2817", "2817 = z196 IBM zEnterprise 196" },
{ QC_TYPE_FAMILY_IBMZ, "2818", "2818 = z114 IBM zEnterprise 114" },
{ QC_TYPE_FAMILY_IBMZ, "2827", "2827 = z12-EC IBM zEnterprise EC12" },
{ QC_TYPE_FAMILY_IBMZ, "2828", "2828 = z12-BC IBM zEnterprise BC12" },
{ QC_TYPE_FAMILY_IBMZ, "2964", "2964 = z13 IBM z13" },
{ QC_TYPE_FAMILY_LINUXONE, "2964", "2964 = IBM LinuxONE Emperor" },
{ QC_TYPE_FAMILY_IBMZ, "2965", "2965 = z13s IBM z13s (single frame)" },
{ QC_TYPE_FAMILY_LINUXONE, "2965", "2965 = IBM LinuxONE Rockhopper" },
{ QC_TYPE_FAMILY_IBMZ, "3906", "3906 = z14 IBM z14" },
{ QC_TYPE_FAMILY_LINUXONE, "3906", "3906 = IBM LinuxONE Emperor II" },
{ QC_TYPE_FAMILY_IBMZ, "3907", "3907 = z14 ZR1 IBM z14 ZR1" },
{ QC_TYPE_FAMILY_LINUXONE, "3907", "3907 = IBM LinuxONE Rockhopper II" },
{ QC_TYPE_FAMILY_IBMZ, "8561", "8561 = z15 T01 IBM z15 T01" },
{ QC_TYPE_FAMILY_LINUXONE, "8561", "8561 = IBM LinuxONE III LT1" },
{ QC_TYPE_FAMILY_IBMZ, "8562", "8562 = z15 T02 IBM z15 T02" },
{ QC_TYPE_FAMILY_LINUXONE, "8562", "8562 = IBM LinuxONE III LT2" },
{ QC_TYPE_FAMILY_IBMZ, "3931", "3931 = z16 A01 IBM z16 A01" },
{ QC_TYPE_FAMILY_LINUXONE, "3931", "3931 = IBM LinuxONE Emperor 4" },
{ QC_TYPE_FAMILY_IBMZ, "3932", "3932 = z16 A02 IBM z16 A02" },
{ QC_TYPE_FAMILY_LINUXONE, "3932", "3932 = IBM LinuxONE Rockhopper 4" },
};
int debug = 0;
/******************************************************************************/
/* */
/* Print the program version */
/* */
/******************************************************************************/
void print_version()
{
printf("Version: %s\n", version);
}
/******************************************************************************/
/* */
/* Look for one attribute and print it */
/* */
/******************************************************************************/
void print_attribute(char *user_string, int level, enum qc_attr_id attribute, enum datatypes type, int print_key)
{
int erg = 0;
const char *result_string = NULL;
int result_int = 0;
float result_float = 0.0;
switch (type)
{
case integer:
erg = qc_get_attribute_int(configuration_handle, attribute, level, &result_int);
break;
case string:
erg = qc_get_attribute_string(configuration_handle, attribute, level, &result_string);
break;
case floatingpoint:
erg = qc_get_attribute_float(configuration_handle, attribute, level, &result_float);
break;
default:
break;
}
if (erg == 1) {
if (print_key == WITH_KEY) {
printf("%s : ",(user_string == NULL? "NULL": user_string));
} /* endif */
switch (type)
{
case integer:
printf("%d\n",result_int);
break;
case string:
printf("%s\n", result_string);
break;
case floatingpoint:
printf("%f\n",result_float);
break;
default:
break;
}
} /* endif */
else {
if ( erg == 0 ) {
/* printf("%s : Attribute exists, but is not set. \n", (user_string == NULL? "NULL": user_string)); */
} /* endif */
else if ( erg < 0) {
printf("%s: An error occurred retrieving the attribute. Error: erg = %d, result_string = %s \n", user_string, erg, (result_string == NULL? "NULL": result_string));
} /* end else if */
/* */
/* TODO qc_get_attribute_string returned error */
/* */
}
} /* print_attribute */
/********************************************************************************/
/* */
/* Open the lib and get the handle */
/* */
/********************************************************************************/
int read_sysinfo()
{
int return_code;
configuration_handle = qc_open(&return_code);
if (return_code < 0) {
printf("Error: Unable to open configuration, return_code =%d\n", return_code);
return -1;
} /* endif */
if (return_code > 0) {
printf("Warning: Unable to read configuration completely, return_code =%d\n", return_code);
return -2;
} /* endif */
if (configuration_handle == NULL) {
printf("Error: Unable to open configuration, return_code =%d\n", return_code);
return -3;
} /* endif */
layers = qc_get_num_layers(configuration_handle, &return_code);
if (layers < 0) {
printf("Error: Unable to retrieve number of layers, return_code =%d\n", return_code);
return -4;
} /* endif */
return 0;
} /* read_sysinfo */
/********************************************************************************/
/* */
/* Look at the type of machine we're running on and print out a user */
/* friendly string */
/* */
/********************************************************************************/
void print_cputype()
{
int i, search;
int erg;
const char *cpu_type = NULL;
int family_type = -1;
/*
* First find out whether we run on an IBM Z, or a LinuxONE system
*/
erg = qc_get_attribute_int(configuration_handle, qc_type_family, 0, &family_type);
if (erg <= 0 || family_type == -1) {
printf("Error reading family type\n");
return;
} /* endif */
/*
* Now get the machine ID
*/
erg = qc_get_attribute_string(configuration_handle, qc_type, 0, &cpu_type);
if (erg == 1 && cpu_type != NULL) {
for (i = 0, search = 1; (i < sizeof(machinetypes) / sizeof(struct machinetype)) && search ; i++)
{
if ((family_type == machinetypes[i].model_families) && (strcmp(cpu_type, machinetypes[i].typenumber) == 0)) {
printf("%s\n", machinetypes[i].fullname);
search = 0;
} /* endif */
} /* endfor */
if (search != 0) {
printf("An unknown machine type was reported: %s\n\
Please file a bug report with this output:\n" , cpu_type);
/* TODO output of /proc/sysinfo */
} /* endif */
} /* endif */
return;
} /* print_cputype */
/********************************************************************************/
/* */
/* Print out the values for SCC */
/* */
/* To uniquely identify a machine the following information is used: */
/* */
/* Type */
/* Sequence code */
/* CPUs total */
/* CPUs IFL */
/* LPAR Number */
/* LPAR Characteristics: */
/* LPAR CPUs */
/* LPAR IFLs */
/* */
/* Optional: */
/* */
/* VM00 Name */
/* VM00 Control Programm */
/* VM00 CPUs */
/* */
/********************************************************************************/
void print_scc()
{
print_version();
print_attribute("Type", 0, qc_type, string, WITH_KEY);
print_attribute("Type Name", 0, qc_type_name, string, WITH_KEY);
print_attribute("Sequence Code", 0, qc_sequence_code, string, WITH_KEY);
print_attribute("CPUs Total", 0, qc_num_ifl_total, integer, WITH_KEY);
print_attribute("CPUs IFL", 0, qc_num_ifl_total, integer, WITH_KEY);
print_attribute("LPAR Number", 1, qc_partition_number, integer, WITH_KEY);
print_attribute("LPAR Name", 1, qc_layer_name, string, WITH_KEY);
print_attribute("LPAR Characteristics", 1, qc_partition_char, string, WITH_KEY);
print_attribute("LPAR CPUs Total", 1, qc_num_ifl_total, integer, WITH_KEY);
print_attribute("LPAR CPUs IFL", 1, qc_num_ifl_total, integer, WITH_KEY);
if (layers > 2) {
/*
* This means, that eather zKVM or z/Vm is running
*/
print_attribute("VM00 Name", 3, qc_layer_name, string, WITH_KEY);
print_attribute("VM00 Control Program", 2, qc_control_program_id, string, WITH_KEY);
print_attribute("VM00 CPUs Total", 3, qc_num_cpu_total, integer, WITH_KEY);
print_attribute("VM00 IFLs", 3, qc_num_cpu_total, integer, WITH_KEY);
} /* endif */
return;
} /* print_scc */
/******************************************************************************/
/* */
/* Secure boot support models ( check_model () ) */
/* Only the following machines support secure boot: */
/* z14, z14 ZR1, z15, z16 */
/* */
/******************************************************************************/
int check_model (const char *cpu) {
#define IBM_Models 6 /* Number of IBM models listed below */
char *types[IBM_Models] = {
"3906",
"3907",
"8561",
"8562",
"3931",
"3932",
};
int i;
int models = sizeof(types) / sizeof(types[0]);
for ( i = 0; i < models; i++) {
if ( !strcmp(cpu,types[i]) ) {
return 1;
};
}
return 0;
} /* check_model */
/******************************************************************************/
/* */
/* print out whether secure boot is enabled */
/* */
/******************************************************************************/
void print_secure_mode()
{
int erg;
int release_major;
int release_sub;
int release_minor;
const char *cpu_type = NULL;
int cpu_okay = 0;
int Layer = 0;
int i = 0;
/*
* First we have to check whether we have the appropriate kernel Level (>= 5.3)
*/
struct utsname uts;
erg = uname(&uts);
if (erg != 0) {
perror ("Error executing uname(): ");
return;
} /* endif */
#if 0
printf("sysname: %s\n", uts.sysname);
printf("nodename: %s\n", uts.nodename);
printf("release: %s\n", uts.release);
#endif
/*
* A release number looks like: m.s.mi
* where m, s, mi are numbers with one ore more digits
* Minimum kernel version is 5.3
*/
erg = sscanf(uts.release,"%d.%d.%d-%*s", &release_major, &release_sub, &release_minor);
if ( release_major < 5 ) {
goto return_does_not_exist;
}
if ( release_sub < 3 ) {
goto return_does_not_exist;
}
#if 0
printf("Translated successfully: %d\n", erg);
printf("release_major: %d\n", release_major);
printf("release_sub: %d\n", release_sub);
printf("release_minor: %d\n", release_minor);
printf("version: %s\n", uts.version);
printf("machine: %s\n", uts.machine);
printf("Print_secure called\n");
#endif
/*
* Only the following machines support secure boot:
* z14, z15, z16
* 3906, 3907, 8561, 8562, 3931, 3932
*/
erg = qc_get_attribute_string(configuration_handle, qc_type, 0, &cpu_type);
if (erg == 1 && cpu_type != NULL) {
cpu_okay = check_model(cpu_type);
if ( cpu_okay == 0 ) {
goto return_does_not_exist;
} /* endif */
} /* endif */
for ( i = 0; i < layers; i++) {
erg = qc_get_attribute_int(configuration_handle, qc_layer_type_num, i, &Layer);
if (erg == 1) {
print_attribute("Secure mode on ", i, qc_has_secure, integer, WITH_KEY);
print_attribute("Secure mode used", i, qc_secure, integer, WITH_KEY);
} /* endif */
} /* endfor */
return;
return_does_not_exist:
/*
* Software or hardware does not support secure boot.
*/
puts("Secure mode on : 0\nSecure mode used : 0");
return;
} /* print_secure_mode */
/******************************************************************************/
/* */
/* print out the uuid for this machine */
/* */
/******************************************************************************/
int print_uuid()
{
const char *result_string = NULL;
int erg;
erg = qc_get_attribute_string(configuration_handle, qc_sequence_code, 0, &result_string);
if (erg != 1)
{
puts("Error reading the Serial Number");
return 1;
}
printf("%s", result_string);
result_string = NULL;
erg = qc_get_attribute_string(configuration_handle, qc_layer_name, 1, &result_string);
if (erg != 1)
{
puts("Error reading the LPAR Name");
return 1;
}
printf("-%s", result_string);
result_string = NULL;
if (layers > 2) {
erg = qc_get_attribute_string(configuration_handle, qc_layer_name, 3, &result_string);
if (erg != 1)
{
puts("Error Reading the VM Name");
return 1;
}
printf("-%s", result_string);
}
putchar('\n');
return 0;
} /* print_uuid */
/******************************************************************************/
/* */
/* print out the list of valid / found symbols */
/* */
/******************************************************************************/
void list(char * list_attribute_param)
{
return;
} /* list */
/******************************************************************************/
/* */
/* print out the requested attribute */
/* */
/******************************************************************************/
void print_user_attribute(char *key, char *attribute_param, int layer)
{
return;
} /* print_user_attribute */
/******************************************************************************/
/* */
/* Help Function */
/* */
/******************************************************************************/
void help()
{
puts("help:\n\
\n\
-a <attribute> List the value of the named attribute\n\
-c Print the cputype of this machine\n\
-d <number> Debug Level\n\
-h this help\n\
-L <keyword> List the requested list (Attribute, Recognised)\n\
-s create Info for SCC\n\
-S report whether secure boot is switched on\n\
-u create uuid\n\
-V print version string\n\
");
#if 0
if (debug != 0) {
puts("\n\
Valid values for debug:\n\
4 - read sysinfo.zvm from current directory instead of /proc/sysinfo\n\
8 - printout lines read in from source (see debug == 4)\n\
16 - printf found keys in store_value\n\
32 - Search expression in show attribute\n\
");
} /* endif */
#endif
} /* help */
/******************************************************************************/
/* */
/* Main */
/* */
/******************************************************************************/
int main(int argc, char **argv, char **envp)
{
int opt;
int read_sysinfo_opt;
int print_attr;
int print_cpu;
int print_secure;
int print_help;
int list_attr;
int create_scc;
int create_uuid;
int erg;
int return_code;
char *print_attribute_param = NULL;
char *list_attribute_param = NULL;
void *configuration_handle_tmp = NULL;
read_sysinfo_opt =
print_attr =
print_cpu =
print_secure =
print_help =
list_attr =
create_scc =
create_uuid =
return_code =
erg = 0;
if (strcmp(argv[0],"cputype") == 0) {
read_sysinfo_opt++;
print_cpu++;
} /* endif */
else {
while ((opt = getopt(argc, argv, "a:cd:hL:sSuV")) != -1) {
switch (opt)
{
case 'a':
read_sysinfo_opt++;
print_attr++;
print_attribute_param = strdup(optarg);
break;
case 'c':
read_sysinfo_opt++;
print_cpu++;
break;
case 'd':
debug = atoi(optarg);
if ((debug & 1) == 1) {
setenv("QC_DEBUG", "1", 1);
} /* endif */
if ((debug & 2) == 2) {
setenv("QC_AUTODUMP", "1", 1);
} /* endif */
debug = debug >> 2;
break;
case 'L':
read_sysinfo_opt++;
list_attr++;
list_attribute_param = strdup(optarg);
break;
case 's': /* create unique string for scc */
read_sysinfo_opt++;
create_scc++;
break;
case 'S': /* print out whether secure boot is enabled */
read_sysinfo_opt++;
print_secure++;
break;
case 'u': /* create UUID */
read_sysinfo_opt++;
create_uuid++;
break;
case 'V':
printf("%s\n",versionstring);
return 0;
break;
case 'h':
default:
print_help++;
break;
} /* endswitch */
} /* while */
} /* endlse */
if (print_help != 0) {
help();
return 0;
} /* endif */
if (read_sysinfo_opt != 0) {
if ((erg = read_sysinfo()) != 0) {
return -erg;
} /* endif */
} /* endif */
if ((print_attr + print_cpu + print_secure + list_attr + create_scc + create_uuid) > 1) {
fputs("Only one of the options a, c, L, s, S or u can be specified.\n",stderr);
return 1;
} /* endif */
/* still not implemented thatfore set to zero */
list_attr = print_attr = 0;
if (print_attr != 0) {
print_user_attribute(NULL, print_attribute_param, layers);
goto main_exit;
} /* endif */
if (print_cpu != 0) {
print_cputype();
goto main_exit;
} /* endif */
if (print_secure != 0) {
print_secure_mode();
goto main_exit;
} /* endif */
if (list_attr != 0) {
list(list_attribute_param);
goto main_exit;
} /* endif */
if (create_scc != 0) {
print_scc();
goto main_exit;
} /* endif */
if (create_uuid != 0) {
if(print_uuid() == 1){
goto main_exit_error;
}
goto main_exit;
} /* endif */
help();
main_exit:
if (configuration_handle != NULL) {
configuration_handle_tmp = qc_open(&return_code);
qc_close(configuration_handle);
setenv("QC_DEBUG", "0", 1);
setenv("QC_AUTODUMP", "0", 1);
qc_close(configuration_handle_tmp);
} /* endif */
return 0;
main_exit_error:
return 1;
} /* end main */

2
rules.hw_random Normal file
View File

@ -0,0 +1,2 @@
# Rules to add hw_random node to maintain SLES11-SP1 backward compatibility
KERNEL=="hwrng", SYMLINK+="hw_random"

2
rules.xpram Normal file
View File

@ -0,0 +1,2 @@
# Rules to add xpram* nodes to maintain SLES11-SP1 backward compatibility
KERNEL=="sl*[0-9]", SYMLINK+="xpram%n"

View File

@ -0,0 +1,64 @@
From dff965465ca9d9c4edaf0f90eadd9a6de335b354 Mon Sep 17 00:00:00 2001
From: Niklas Schnelle <schnelle@linux.ibm.com>
Date: Fri, 6 Dec 2024 15:28:08 +0100
Subject: [PATCH] opticsmon: Fix runaway loop in on_link_change()
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
When on_link_change() gets called with a netdev that would be monitored
but hasn't entered zpci_list yet, reloads is 1 after the loops and
a reload occurs. Then the netdev is found in the list and reloads
becomes -1 which incorrectly triggers more reloads until underflow.
Fix this by returning once the device is found. Also just check for
reloads being larger than zero.
Fixes: c34adb9cabee ("opticsmon: Introduce opticsmon tool")
Reviewed-by: Halil Pasic <pasic@linux.ibm.com>
Signed-off-by: Niklas Schnelle <schnelle@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
---
opticsmon/opticsmon.c | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/opticsmon/opticsmon.c b/opticsmon/opticsmon.c
index c2f355e2..50dd8d7f 100644
--- a/opticsmon/opticsmon.c
+++ b/opticsmon/opticsmon.c
@@ -280,16 +280,15 @@ void on_link_change(struct zpci_netdev *netdev, void *arg)
if (!ctx->zpci_list || util_list_is_empty(ctx->zpci_list))
zpci_list_reload(&ctx->zpci_list);
-reload:
+find:
util_list_iterate(ctx->zpci_list, zdev) {
for (i = 0; i < zdev->num_netdevs; i++) {
if (!strcmp(zdev->netdevs[i].name, netdev->name)) {
- reloads--;
/* Skip data collection if operational state is
* unchanged
*/
if (zdev->netdevs[i].operstate == netdev->operstate)
- continue;
+ return;
/* Update operation state for VFs even though
* they are skipped just for a consistent view
*/
@@ -297,14 +296,15 @@ void on_link_change(struct zpci_netdev *netdev, void *arg)
/* Only collect optics data for PFs */
if (!zpci_is_vf(zdev))
dump_adapter_data(ctx, zdev);
+ return;
}
}
}
/* Might be a new device, reload list of devices and retry */
- if (reloads) {
+ if (reloads > 0) {
zpci_list_reload(&ctx->zpci_list);
reloads--;
- goto reload;
+ goto find;
}
}

View File

@ -0,0 +1,969 @@
From d6b702d5791b47f735960ad1f6986e0a32768df6 Mon Sep 17 00:00:00 2001
From: Eduard Shishkin <edward6@linux.ibm.com>
Date: Thu, 11 Jul 2024 10:43:37 +0200
Subject: [PATCH] zipl/src: add basic support for multiple target base disks
. Modify disk_get_info() to process multiple sets of target parameters
provided by the helper script and store it in the array of "targets"
of the structure job_target_data;
. Besides the logical device, maintain an array of physical base disks
in the disk_info structure;
. Use the logical target device only to create bootmap (it is
automatically mirrored by the respective linux driver (dm, or md)
managing the mirrored target). In contrast, install bootstrap blocks
to each physical base disk individually, bypassing that driver;
. Report in verbose mode on which base disks the bootstrap
installation was performed;
. Use the following logic of setting @info->device (which is printed
as "Device...:" in verbose mode):
. source_auto - the target base disk is set;
. source_script - the target (logical) device is set;
. source_user - the device specified by user (via --targetbase
option), or config file is set.
Signed-off-by: Eduard Shishkin <edward6@linux.ibm.com>
Reviewed-by: Stefan Haberland <sth@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
---
zipl/include/disk.h | 14 +-
zipl/include/install.h | 2
zipl/include/job.h | 122 ++++++++++++++++++--
zipl/include/zipl.h | 1
zipl/src/bootmap.c | 23 ++-
zipl/src/disk.c | 295 +++++++++++++++++++++++++++++++++++++------------
zipl/src/install.c | 89 +++++++++-----
zipl/src/job.c | 82 +++++++------
8 files changed, 469 insertions(+), 159 deletions(-)
--- a/zipl/include/disk.h
+++ b/zipl/include/disk.h
@@ -56,13 +56,14 @@
/* targetbase definition */
typedef enum {
defined_as_device,
- defined_as_name
+ defined_as_name,
+ undefined
} definition_t;
/* Disk information type */
struct disk_info {
disk_type_t type;
- dev_t device;
+ dev_t device; /* logical device for bootmap creation */
dev_t partition;
int devno;
int partnum;
@@ -72,8 +73,11 @@
struct hd_geometry geo;
char* name;
char* drv_name;
- definition_t targetbase;
+ definition_t targetbase_def;
int is_nvme;
+ dev_t basedisks[MAX_TARGETS]; /* array of physical disks for
+ * bootstrap blocks recording
+ */
};
struct file_range {
@@ -113,6 +117,9 @@
struct disk_info *info, int align,
off_t *offset);
void disk_print_devt(dev_t d);
+void disk_print_devname(dev_t d);
+void prepare_footnote_ptr(int source, char *ptr);
+void print_footnote_ref(int source, const char *prefix);
void disk_print_info(struct disk_info *info, int source);
int disk_is_zero_block(disk_blockptr_t* block, struct disk_info* info);
blocknum_t disk_compact_blocklist(disk_blockptr_t* list, blocknum_t count,
@@ -122,7 +129,6 @@
disk_blockptr_t** blocklist,
struct disk_info* pinfo);
int disk_check_subchannel_set(int devno, dev_t device, char* dev_name);
-void disk_print_geo(struct disk_info *data);
int fs_map(int fd, uint64_t offset, blocknum_t *mapped, int fs_block_size);
#endif /* not DISK_H */
--- a/zipl/include/install.h
+++ b/zipl/include/install.h
@@ -71,7 +71,7 @@
struct program_component *components[NR_PROGRAM_COMPONENTS];
int nr_menu_entries;
int fd;
- char *device;
+ char *basetmp[MAX_TARGETS];
char *filename;
unsigned int tmp_filename_created:1;
unsigned int skip_prepare:1;
--- a/zipl/include/job.h
+++ b/zipl/include/job.h
@@ -18,7 +18,6 @@
#include "disk.h"
#include "zipl.h"
-
enum job_id {
job_print_usage = 1,
job_print_version = 2,
@@ -30,6 +29,21 @@
job_mvdump = 8,
};
+/*
+ * Set of parameters per physical disk, which are provided
+ * either by user, or by helper script
+ */
+struct target {
+ char *targetbase;
+ disk_type_t targettype;
+ int targetcylinders;
+ int targetheads;
+ int targetsectors;
+ int targetblocksize;
+ blocknum_t targetoffset;
+ int check_params;
+};
+
/* target information source */
typedef enum {
source_unknown = 0,
@@ -39,17 +53,21 @@
} source_t;
struct job_target_data {
- char* bootmap_dir;
- char* targetbase;
- disk_type_t targettype;
- int targetcylinders;
- int targetheads;
- int targetsectors;
- int targetblocksize;
- blocknum_t targetoffset;
+ char *bootmap_dir;
+ int nr_targets;
+ struct target targets[MAX_TARGETS];
source_t source;
};
+enum target_params {
+ TARGET_BASE,
+ TARGET_TYPE,
+ TARGET_GEOMETRY,
+ TARGET_BLOCKSIZE,
+ TARGET_OFFSET,
+ LAST_TARGET_PARAM
+};
+
struct job_common_ipl_data {
char* image;
char* parmline;
@@ -142,12 +160,94 @@
int is_ldipl_dump;
};
+static inline struct target *target_at(struct job_target_data *data,
+ int index)
+{
+ return index >= MAX_TARGETS ? NULL : &data->targets[index];
+}
+
+static inline char *get_targetbase(struct job_target_data *data, int index)
+{
+ return target_at(data, index)->targetbase;
+}
+
+static inline void set_targetbase(struct job_target_data *data, int index,
+ char *value)
+{
+ target_at(data, index)->targetbase = value;
+}
+
+static inline disk_type_t get_targettype(struct job_target_data *data,
+ int index)
+{
+ return target_at(data, index)->targettype;
+}
+
+int set_targettype(struct job_target_data *data, int index, char *value);
+
+static inline char *job_get_targetbase(struct job_data *job)
+{
+ return get_targetbase(&job->target, 0);
+}
+
+static inline void job_set_targetbase(struct job_data *job, char *value)
+{
+ set_targetbase(&job->target, 0, value);
+}
+
+static inline int job_get_nr_targets(struct job_data *job)
+{
+ return job->target.nr_targets;
+}
+
+static inline void job_set_nr_targets(struct job_data *job, int value)
+{
+ job->target.nr_targets = value;
+}
+
+static inline disk_type_t job_get_targettype(struct job_data *job)
+{
+ return get_targettype(&job->target, 0);
+}
+
+int job_set_targettype(struct job_data *job, char *value);
+
+#define define_target_param_ops(_TYPE_, _PARAM_) \
+static inline _TYPE_ get_target##_PARAM_(struct job_target_data *data, \
+ int index) \
+{ \
+ return target_at(data, index)->target##_PARAM_; \
+} \
+ \
+static inline void set_target##_PARAM_(struct job_target_data *data, \
+ int index, _TYPE_ value) \
+{ \
+ target_at(data, index)->target##_PARAM_ = value; \
+} \
+ \
+static inline _TYPE_ job_get_target##_PARAM_(struct job_data *job) \
+{ \
+ return get_target##_PARAM_(&job->target, 0); \
+} \
+ \
+static inline void job_set_target##_PARAM_(struct job_data *job, \
+ _TYPE_ value) \
+{ \
+ set_target##_PARAM_(&job->target, 0, value); \
+}
+
+define_target_param_ops(int, cylinders)
+define_target_param_ops(int, heads)
+define_target_param_ops(int, sectors)
+define_target_param_ops(int, blocksize)
+define_target_param_ops(blocknum_t, offset)
+
/**
- * Return true, if target parameters for the base disk are set
+ * Return true, if target parameters are set at least for one target base disk
*/
static inline int target_parameters_are_set(struct job_target_data *td)
{
- return td->targetbase != NULL;
+ return get_targetbase(td, 0) != NULL;
}
int job_get(int argc, char* argv[], struct job_data** data);
--- a/zipl/include/zipl.h
+++ b/zipl/include/zipl.h
@@ -41,6 +41,7 @@
#define MENU_DEFAULT_TIMEOUT 0
#define MAX_DUMP_VOLUMES 32
+#define MAX_TARGETS 32
#define SECURE_BOOT_UNDEFINED -1
#define SECURE_BOOT_DISABLED 0
--- a/zipl/src/bootmap.c
+++ b/zipl/src/bootmap.c
@@ -1477,9 +1477,9 @@
printf("Target device information\n");
disk_print_info(bis->info, job->target.source);
}
- if (misc_temp_dev(bis->info->device, 1, &bis->device))
+ if (misc_temp_dev(bis->info->device, 1, &bis->basetmp[0]))
return -1;
- if (check_dump_device(job, bis->info, bis->device))
+ if (check_dump_device(job, bis->info, bis->basetmp[0]))
return -1;
printf("Building bootmap directly on partition '%s'%s\n",
bis->filename,
@@ -1543,6 +1543,8 @@
static int prepare_build_program_table_file(struct job_data *job,
struct install_set *bis)
{
+ int i;
+
if (bis->skip_prepare)
/* skip the preparation work */
return 0;
@@ -1576,8 +1578,12 @@
printf("Target device information\n");
disk_print_info(bis->info, job->target.source);
}
- if (misc_temp_dev(bis->info->device, 1, &bis->device))
- return -1;
+ for (i = 0; i < job_get_nr_targets(job); i++) {
+ if (misc_temp_dev(bis->info->basedisks[i],
+ 1,
+ &bis->basetmp[i]))
+ return -1;
+ }
/* Check configuration number limits */
if (job->id == job_menu) {
if (check_menu_positions(&job->data.menu, job->name,
@@ -1692,9 +1698,9 @@
/* Retrieve target device information */
if (disk_get_info(job->data.dump.device, &job->target, &info))
return -1;
- if (misc_temp_dev(info->device, 1, &bis->device))
+ if (misc_temp_dev(info->device, 1, &bis->basetmp[0]))
return -1;
- if (check_dump_device(job, info, bis->device))
+ if (check_dump_device(job, info, bis->basetmp[0]))
return -1;
assert(!job->target.bootmap_dir);
@@ -1844,6 +1850,9 @@
if (bis->tmp_filename_created)
misc_free_temp_file(bis->filename);
free(bis->filename);
- misc_free_temp_dev(bis->device);
+ for (i = 0; i < MAX_TARGETS; i++) {
+ if (bis->basetmp[i])
+ misc_free_temp_dev(bis->basetmp[i]);
+ }
disk_free_info(bis->info);
}
--- a/zipl/src/disk.c
+++ b/zipl/src/disk.c
@@ -187,43 +187,134 @@
return rc;
}
+/**
+ * Process a script output represented by FH and consisting
+ * of pairs 'key=value' (each such pair is on a separate line).
+ * Check its consistency and set the extracted target parameters
+ * to the array of "targets" at TD.
+ *
+ * NOTE: this function defines specifications on valid output of
+ * zipl helper scripts. See zipl-support-for-mirrored-devices.txt
+ * for details. Before modifying this function, make sure that it
+ * won't lead to format change.
+ */
static int set_target_parameters(FILE *fh, struct job_target_data *td)
{
- int checkparm = 0;
+ int idx[LAST_TARGET_PARAM] = {0};
+ struct target *t;
char buffer[80];
char value[40];
+ char *error;
+ int i;
+ /**
+ * Process a stream of 'key=value' pairs and distribute
+ * them into groups.
+ * The i-th occurrence of some "key" in the stream means
+ * that the respective pair belongs to the group #i
+ */
+ error = "Exceeded the maximum number of base disks";
while (fgets(buffer, 80, fh)) {
if (sscanf(buffer, "targetbase=%s", value) == 1) {
- td->targetbase = misc_strdup(value);
- checkparm++;
+ t = target_at(td, idx[TARGET_BASE]++);
+ if (!t)
+ goto error;
+ t->targetbase = misc_strdup(value);
+ goto found;
}
if (sscanf(buffer, "targettype=%s", value) == 1) {
- type_from_target(value, &td->targettype);
- checkparm++;
+ t = target_at(td, idx[TARGET_TYPE]++);
+ if (!t)
+ goto error;
+ type_from_target(value, &t->targettype);
+ goto found;
}
if (sscanf(buffer, "targetgeometry=%s", value) == 1) {
- td->targetcylinders =
- atoi(strtok(value, ","));
- td->targetheads = atoi(strtok(NULL, ","));
- td->targetsectors = atoi(strtok(NULL, ","));
- checkparm++;
+ t = target_at(td, idx[TARGET_GEOMETRY]++);
+ if (!t)
+ goto error;
+ t->targetcylinders = atoi(strtok(value, ","));
+ t->targetheads = atoi(strtok(NULL, ","));
+ t->targetsectors = atoi(strtok(NULL, ","));
+ goto found;
}
if (sscanf(buffer, "targetblocksize=%s", value) == 1) {
- td->targetblocksize = atoi(value);
- checkparm++;
+ t = target_at(td, idx[TARGET_BLOCKSIZE]++);
+ if (!t)
+ goto error;
+ t->targetblocksize = atoi(value);
+ goto found;
}
if (sscanf(buffer, "targetoffset=%s", value) == 1) {
- td->targetoffset = atol(value);
- checkparm++;
+ t = target_at(td, idx[TARGET_OFFSET]++);
+ if (!t)
+ goto error;
+ t->targetoffset = atol(value);
+ goto found;
}
+ continue;
+found:
+ t->check_params++;
}
- if ((!disk_is_eckd(td->targettype) && checkparm < 4) ||
- (disk_is_eckd(td->targettype) && checkparm != 5)) {
- error_reason("Target parameters missing from script");
- return -1;
+ /* Check for consistency */
+ error = "Inconsistent script output";
+ /*
+ * First, calculate total number of groups
+ */
+ td->nr_targets = 0;
+ for (i = 0; i < MAX_TARGETS; i++) {
+ t = target_at(td, i);
+ if (t->check_params == 0)
+ break;
+ td->nr_targets++;
+ }
+ if (!td->nr_targets)
+ /* No keywords found in the stream */
+ goto error;
+ /*
+ * Each group has to include targetbase, targettype,
+ * targetblocksize and targetoffset.
+ */
+ if (td->nr_targets != idx[TARGET_BASE] ||
+ td->nr_targets != idx[TARGET_TYPE] ||
+ td->nr_targets != idx[TARGET_BLOCKSIZE] ||
+ td->nr_targets != idx[TARGET_OFFSET])
+ goto error;
+ /*
+ * In addition, any group of "ECKD" type has to include
+ * targetgeometry
+ */
+ for (i = 0; i < td->nr_targets; i++) {
+ t = target_at(td, i);
+ assert(t->check_params >= 4);
+ if (disk_is_eckd(t->targettype) && t->check_params != 5)
+ goto error;
}
return 0;
+error:
+ error_reason("%s", error);
+ return -1;
+}
+
+static void print_base_disk_params(struct job_target_data *td, int index)
+{
+ disk_type_t type = get_targettype(td, index);
+
+ if (!verbose)
+ return;
+ {
+ fprintf(stderr, "Base disk '%s':\n", get_targetbase(td, index));
+ fprintf(stderr, " layout........: %s\n", disk_get_type_name(type));
+ }
+ if (disk_is_eckd(type)) {
+ fprintf(stderr, " heads.........: %u\n", get_targetheads(td, index));
+ fprintf(stderr, " sectors.......: %u\n", get_targetsectors(td, index));
+ fprintf(stderr, " cylinders.....: %u\n", get_targetcylinders(td, index));
+ }
+ {
+ fprintf(stderr, " start.........: %lu\n", get_targetoffset(td, index));
+ fprintf(stderr, " blksize.......: %u\n", get_targetblocksize(td, index));
+ }
}
/**
@@ -235,31 +326,57 @@
{
int majnum, minnum;
struct stat stats;
-
+ int i;
+ /*
+ * Currently multiple base disks with different parameters
+ * are not supported
+ */
data->devno = -1;
- data->phy_block_size = td->targetblocksize;
- data->type = td->targettype;
- data->partnum = 0;
+ data->phy_block_size = get_targetblocksize(td, 0);
+ data->type = get_targettype(td, 0);
- if (sscanf(td->targetbase, "%d:%d", &majnum, &minnum) == 2) {
- data->device = makedev(majnum, minnum);
- data->targetbase = defined_as_device;
- data->partnum = minor(stats.st_rdev) - minnum;
- } else {
- if (stat(td->targetbase, &stats)) {
- error_reason(strerror(errno));
- error_text("Could not get information for "
- "file '%s'", td->targetbase);
+ assert(td->nr_targets != 0);
+ for (i = 1; i < td->nr_targets; i++) {
+ if (data->type != get_targettype(td, i) ||
+ data->phy_block_size != get_targetblocksize(td, i)) {
+ print_base_disk_params(td, 0);
+ print_base_disk_params(td, i);
+ error_reason("Inconsistent base disk geometry in target device");
return -1;
}
- if (!S_ISBLK(stats.st_mode)) {
- error_reason("Target base device '%s' is not "
- "a block device",
- td->targetbase);
+ }
+ data->partnum = 0;
+ data->targetbase_def = undefined;
+
+ for (i = 0; i < td->nr_targets; i++) {
+ definition_t defined_as;
+
+ if (sscanf(get_targetbase(td, i),
+ "%d:%d", &majnum, &minnum) == 2) {
+ data->basedisks[i] = makedev(majnum, minnum);
+ defined_as = defined_as_device;
+ } else {
+ if (stat(get_targetbase(td, i), &stats)) {
+ error_reason(strerror(errno));
+ error_text("Could not get information for "
+ "file '%s'", get_targetbase(td, i));
+ return -1;
+ }
+ if (!S_ISBLK(stats.st_mode)) {
+ error_reason("Target base device '%s' is not "
+ "a block device",
+ get_targetbase(td, i));
+ return -1;
+ }
+ data->basedisks[i] = stats.st_rdev;
+ defined_as = defined_as_name;
+ }
+ if (data->targetbase_def != undefined &&
+ data->targetbase_def != defined_as) {
+ error_reason("Target base disks are defined by different ways");
return -1;
}
- data->device = stats.st_rdev;
- data->targetbase = defined_as_name;
+ data->targetbase_def = defined_as;
}
if (data->type == disk_type_scsi && ioctl(fd, NVME_IOCTL_ID) >= 0)
data->is_nvme = 1;
@@ -446,11 +563,28 @@
static int disk_set_geometry_by_hint(struct job_target_data *td,
struct disk_info *data)
{
- data->geo.heads = td->targetheads;
- data->geo.sectors = td->targetsectors;
- data->geo.cylinders = td->targetcylinders;
- data->geo.start = td->targetoffset;
-
+ int i;
+ /*
+ * Currently multiple base disks with different parameters
+ * are not supported
+ */
+ data->geo.heads = get_targetheads(td, 0);
+ data->geo.sectors = get_targetsectors(td, 0);
+ data->geo.cylinders = get_targetcylinders(td, 0);
+ data->geo.start = get_targetoffset(td, 0);
+
+ assert(td->nr_targets != 0);
+ for (i = 1; i < td->nr_targets; i++) {
+ if (data->geo.heads != get_targetheads(td, i) ||
+ data->geo.sectors != get_targetsectors(td, i) ||
+ data->geo.cylinders != get_targetcylinders(td, i) ||
+ data->geo.start != get_targetoffset(td, i)) {
+ print_base_disk_params(td, 0);
+ print_base_disk_params(td, i);
+ error_reason("Inconsistent base disk geometry in target device");
+ return -1;
+ }
+ }
return 0;
}
@@ -515,14 +649,16 @@
}
/**
- * Prepare INFO required to perform IPL installation on the physical
- * disk where the logical DEVICE is located.
+ * Prepare INFO required to perform IPL installation on physical disks
+ * participating in the logical DEVICE.
* Preparation is performed in 2 steps:
*
- * 1. Find out a physical "base" disk where the logical DEVICE is
- * located. Calculate "target" parameters (type, geometry, physical
- * block size, data offset, etc);
- * 2. Complete INFO by the found base disk and target parameters.
+ * 1. Find out a set of physical "base" disks participating in the
+ * logical DEVICE. For each found disk calculate "target" parameters
+ * (type, geometry, physical block size, data offset, etc) and store
+ * it in the array of "targets" of TD;
+ * 2. Complete INFO using the found base disks and calculated target
+ * parameters.
*
* TD: optionally contains target parameters specified by user via
* config file, or special "target options" of zipl tool.
@@ -566,6 +702,7 @@
goto error;
if (disk_set_info_by_hint(td, data, fd))
goto error;
+ data->device = stats.st_rdev;
break;
case source_user:
/*
@@ -578,6 +715,12 @@
goto error;
if (disk_set_info_by_hint(td, data, fd))
goto error;
+ /*
+ * multiple base disks are not supported
+ * with this source type
+ */
+ assert(td->nr_targets == 1);
+ data->device = data->basedisks[0];
break;
case source_auto:
/* no ready target parameters are available */
@@ -585,6 +728,12 @@
goto error;
if (disk_set_info_auto(data, &stats, fd))
goto error;
+ /*
+ * multiple base disks are not supported
+ * with this source type
+ */
+ data->basedisks[0] = data->device;
+ td->nr_targets = 1;
break;
default:
assert(0);
@@ -940,6 +1089,33 @@
printf("%02x:%02x", major(d), minor(d));
}
+void disk_print_devname(dev_t dev)
+{
+ struct util_proc_part_entry part_entry;
+
+ if (!util_proc_part_get_entry(dev, &part_entry)) {
+ printf("%s", part_entry.name);
+ util_proc_part_free_entry(&part_entry);
+ } else {
+ disk_print_devt(dev);
+ }
+}
+
+void prepare_footnote_ptr(int source, char *ptr)
+{
+ if (source == source_user || source == source_script)
+ strcpy(ptr, " *)");
+ else
+ strcpy(ptr, "");
+}
+
+void print_footnote_ref(int source, const char *prefix)
+{
+ if (source == source_user)
+ printf("%s*) Data provided by user.\n", prefix);
+ else if (source == source_script)
+ printf("%s*) Data provided by script.\n", prefix);
+}
/* Return a name for a given disk TYPE. */
char *
@@ -991,12 +1167,11 @@
void disk_print_info(struct disk_info *info, int source)
{
char footnote[4] = "";
- if (source == source_user || source == source_script)
- strcpy(footnote, " *)");
+ prepare_footnote_ptr(source, footnote);
printf(" Device..........................: ");
disk_print_devt(info->device);
- if (info->targetbase == defined_as_device)
+ if (info->targetbase_def == defined_as_device)
printf("%s", footnote);
printf("\n");
if (info->partnum != 0) {
@@ -1007,7 +1182,7 @@
if (info->name) {
printf(" Device name.....................: %s",
info->name);
- if (info->targetbase == defined_as_name)
+ if (info->targetbase_def == defined_as_name)
printf("%s", footnote);
printf("\n");
}
@@ -1050,21 +1225,7 @@
info->phy_block_size, footnote);
printf(" Device size in physical blocks..: %ld\n",
(long) info->phy_blocks);
- if (source == source_user)
- printf(" *) Data provided by user.\n");
- if (source == source_script)
- printf(" *) Data provided by script.\n");
-}
-
-/* Print textual representation of geo structure. */
-void
-disk_print_geo(struct disk_info *data)
-{
- printf(" geo.heads.........:%u\n", data->geo.heads);
- printf(" geo.sectors.......:%u\n", data->geo.sectors);
- printf(" geo.cylinders.....:%u\n", data->geo.cylinders);
- printf(" geo.start.........:%lu\n", data->geo.start);
- printf(" blksize...........:%u\n", data->phy_block_size);
+ print_footnote_ref(source, " ");
}
/* Check whether a block is a zero block which identifies a hole in a file.
--- a/zipl/src/install.c
+++ b/zipl/src/install.c
@@ -434,11 +434,14 @@
{
disk_blockptr_t *scsi_dump_sb_blockptr = &bis->scsi_dump_sb_blockptr;
struct disk_info *info = bis->info;
- char *device = bis->device;
- int fd, rc;
+ char footnote[4];
+ int rc;
+ int i;
if (!info)
return 0;
+
+ prepare_footnote_ptr(job->target.source, footnote);
/* Inform user about what we're up to */
printf("Preparing boot device for %s%s: ",
disk_get_ipl_type(info->type,
@@ -455,40 +458,58 @@
disk_print_devt(info->device);
printf(".\n");
}
- /* Open device file */
- fd = open(device, O_RDWR);
- if (fd == -1) {
- error_reason(strerror(errno));
- error_text("Could not open temporary device file '%s'",
- device);
- return -1;
- }
- /* Ensure that potential cache inconsistencies between disk and
- * partition are resolved by flushing the corresponding buffers. */
- if (!dry_run) {
- if (ioctl(fd, BLKFLSBUF)) {
- fprintf(stderr, "Warning: Could not flush disk "
- "caches.\n");
+ /* Install independently on each physical target base */
+
+ for (i = 0; i < job_get_nr_targets(job); i++) {
+ int fd;
+
+ if (verbose) {
+ printf("Installing on base disk: ");
+ disk_print_devname(info->basedisks[i]);
+ printf("%s.\n", footnote);
}
+ /* Open device file */
+ fd = open(bis->basetmp[i], O_RDWR);
+ if (fd == -1) {
+ error_reason(strerror(errno));
+ error_text("Could not open temporary device file '%s'",
+ bis->basetmp[i]);
+ return -1;
+ }
+ /* Ensure that potential cache inconsistencies between disk and
+ * partition are resolved by flushing the corresponding buffers.
+ */
+ if (!dry_run) {
+ if (ioctl(fd, BLKFLSBUF)) {
+ fprintf(stderr, "Warning: Could not flush disk "
+ "caches.\n");
+ }
+ }
+ /*
+ * Depending on disk type, install one or two program tables
+ * for CCW-type IPL and (or) for List-Directed IPL (see the
+ * picture in comments above)
+ */
+ if (job->id == job_dump_partition) {
+ rc = install_bootloader_dump(bis->tables, info,
+ scsi_dump_sb_blockptr,
+ is_ngdump_enabled(job),
+ fd);
+ } else {
+ rc = install_bootloader_ipl(bis->tables, info,
+ fd);
+ }
+ if (fsync(fd))
+ error_text("Could not sync device file '%s'",
+ bis->basetmp[i]);
+ if (close(fd))
+ error_text("Could not close device file '%s'",
+ bis->basetmp[i]);
+ if (rc)
+ break;
}
- /*
- * Depending on disk type, install one or two program tables
- * for CCW-type IPL and (or) for List-Directed IPL (see the
- * picture in comments above)
- */
- if (job->id == job_dump_partition) {
- rc = install_bootloader_dump(bis->tables, info,
- scsi_dump_sb_blockptr,
- is_ngdump_enabled(job),
- fd);
- } else {
- rc = install_bootloader_ipl(bis->tables, info, fd);
- }
-
- if (fsync(fd))
- error_text("Could not sync device file '%s'", device);
- if (close(fd))
- error_text("Could not close device file '%s'", device);
+ if (verbose)
+ print_footnote_ref(job->target.source, "");
if (!dry_run && rc == 0) {
if (info->devno >= 0)
--- a/zipl/src/job.c
+++ b/zipl/src/job.c
@@ -1346,6 +1346,27 @@
}
}
+int set_targettype(struct job_target_data *data, int index, char *value)
+{
+ return type_from_target(value,
+ &target_at(data, index)->targettype);
+}
+
+int job_set_targettype(struct job_data *job, char *value)
+{
+ return set_targettype(&job->target, 0, value);
+}
+
+static int job_set_target(struct job_data *job, char *value)
+{
+ job_set_targetbase(job, value);
+ if (!job_get_targetbase(job))
+ return -1;
+ job_set_nr_targets(job, 1);
+ job->target.source = source_user;
+ return 0;
+}
+
static int
get_job_from_section_data(char* data[], struct job_data* job, char* section)
{
@@ -1362,32 +1383,28 @@
return -1;
/* Fill in target */
if (data[(int) scan_keyword_targetbase] != NULL) {
- job->target.targetbase =
- misc_strdup(data[(int)
- scan_keyword_targetbase]);
- if (job->target.targetbase == NULL)
+ if (job_set_target(job, misc_strdup(data[(int)
+ scan_keyword_targetbase])))
return -1;
- job->target.source = source_user;
}
if (data[(int) scan_keyword_targettype] != NULL) {
- if (type_from_target(
- data[(int) scan_keyword_targettype],
- &job->target.targettype))
+ if (job_set_targettype(job,
+ data[(int) scan_keyword_targettype]))
return -1;
}
if (data[(int) scan_keyword_targetgeometry] != NULL) {
- job->target.targetcylinders =
+ job_set_targetcylinders(job,
atoi(strtok(data[(int)
- scan_keyword_targetgeometry], ","));
- job->target.targetheads = atoi(strtok(NULL, ","));
- job->target.targetsectors = atoi(strtok(NULL, ","));
+ scan_keyword_targetgeometry], ",")));
+ job_set_targetheads(job, atoi(strtok(NULL, ",")));
+ job_set_targetsectors(job, atoi(strtok(NULL, ",")));
}
if (data[(int) scan_keyword_targetblocksize] != NULL)
- job->target.targetblocksize =
- atoi(data[(int) scan_keyword_targetblocksize]);
+ job_set_targetblocksize(job,
+ atoi(data[(int) scan_keyword_targetblocksize]));
if (data[(int) scan_keyword_targetoffset] != NULL)
- job->target.targetoffset =
- atol(data[(int) scan_keyword_targetoffset]);
+ job_set_targetoffset(job,
+ atol(data[(int) scan_keyword_targetoffset]));
/* Fill in name and address of image file */
job->data.ipl.common.image = misc_strdup(
@@ -1615,37 +1632,32 @@
return -1;
break;
case scan_keyword_targetbase:
- job->target.targetbase = misc_strdup(
- scan[i].content.keyword.value);
- if (job->target.targetbase == NULL)
+ if (job_set_target(job, misc_strdup(
+ scan[i].content.keyword.value)))
return -1;
- job->target.source = source_user;
break;
case scan_keyword_targettype:
- if (type_from_target(
- scan[i].content.keyword.value,
- &job->target.targettype))
+ if (job_set_targettype(job,
+ scan[i].content.keyword.value))
return -1;
break;
case scan_keyword_targetgeometry:
- job->target.targetcylinders =
+ job_set_targetcylinders(job,
atoi(strtok(
scan[i].content.keyword.value,
- ","));
- job->target.targetheads =
- atoi(strtok(NULL, ","));
- job->target.targetsectors =
- atoi(strtok(NULL, ","));
+ ",")));
+ job_set_targetheads(job,
+ atoi(strtok(NULL, ",")));
+ job_set_targetsectors(job,
+ atoi(strtok(NULL, ",")));
break;
case scan_keyword_targetblocksize:
- job->target.targetblocksize =
- atoi(
- scan[i].content.keyword.value);
+ job_set_targetblocksize(job, atoi(
+ scan[i].content.keyword.value));
break;
case scan_keyword_targetoffset:
- job->target.targetoffset =
- atol(
- scan[i].content.keyword.value);
+ job_set_targetoffset(job, atol(
+ scan[i].content.keyword.value));
break;
default:
/* Should not happen */

View File

@ -0,0 +1,67 @@
From 2d26a63806d2847f549c06276070a636a61bcb80 Mon Sep 17 00:00:00 2001
From: Eduard Shishkin <edward6@linux.ibm.com>
Date: Wed, 4 Dec 2024 13:37:46 +0100
Subject: [PATCH s390-tools] zipl_helper.device-mapper: add missed step in
logical device resolution
This fixes 670bf3e
Preparing a loop device for IPL by zipl tool, using its partition as
zipl target, leads to inconsistent installation setup. The problem is in
a missed step in the procedure of logical device resolution performed
by the script zipl_helper.device-mapper:
\# lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
loop0 7:0 0 5G 0 loop
|-loop0p1 253:15 0 128M 0 part
`-loop0p2 253:16 0 4.9G 0 part /mnt
\# ./zipl_helper.device-mapper 253:16
Expected result:
targetbase=7:0
targettype=SCSI
targetblocksize=4096
targetoffset=32784
Actual result:
targetbase=253:16
targettype=SCSI
targetblocksize=4096
targetoffset=32784
The fixup adds a missed resolution step.
Reference-ID: LTC210771
Signed-off-by: Eduard Shishkin <edward6@linux.ibm.com>
---
zipl/src/zipl_helper.device-mapper.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/zipl/src/zipl_helper.device-mapper.c b/zipl/src/zipl_helper.device-mapper.c
index aca52be1..918c5aba 100644
--- a/zipl/src/zipl_helper.device-mapper.c
+++ b/zipl/src/zipl_helper.device-mapper.c
@@ -1306,13 +1306,13 @@ static int complete_physical_device(struct physical_device *pd, dev_t *base_dev)
*base_dev = base_entry->dev.dev;
} else {
/*
- * In this case base device is the uppermost logical
+ * In this case base device is the uppermost
* device which provides access to boot sectors
*/
base_entry = find_base_entry(pd->dmpath, dc->bootsectors);
if (!base_entry)
return -1;
- *base_dev = base_entry->dev.dev;
+ *base_dev = first_device_by_target_data(base_entry->target);
}
/* Check for valid offset of filesystem */
if ((pd->offset % (dc->blocksize / SECTOR_SIZE)) != 0) {
--
2.39.0

View File

@ -0,0 +1,129 @@
From cf5560a100b5552e2eeeaac9c60a88ae77233530 Mon Sep 17 00:00:00 2001
From: Niklas Schnelle <schnelle@linux.ibm.com>
Date: Mon, 9 Dec 2024 15:08:03 +0100
Subject: [PATCH] libzpci: opticsmon: Refactor on_link_change() using new
zpci_find_by_netdev()
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Finding a PCI device given the name of a netdev seems generally useful
so pull this out into a new zpci_find_by_netdev() function in libzpci
and use this to simplify on_link_change() removing the need for
backwards goto.
Reviewed-by: Halil Pasic <pasic@linux.ibm.com>
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Niklas Schnelle <schnelle@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
---
include/lib/pci_list.h | 3 +++
libzpci/pci_list.c | 31 +++++++++++++++++++++++++++++++
opticsmon/opticsmon.c | 27 +++++++++++----------------
3 files changed, 45 insertions(+), 16 deletions(-)
diff --git a/include/lib/pci_list.h b/include/lib/pci_list.h
index 829ec244..5b2918bc 100644
--- a/include/lib/pci_list.h
+++ b/include/lib/pci_list.h
@@ -93,4 +93,7 @@ const char *zpci_pft_str(struct zpci_dev *zdev);
const char *zpci_operstate_str(operstate_t state);
operstate_t zpci_operstate_from_str(const char *oper_str);
+struct zpci_dev *zpci_find_by_netdev(struct util_list *zpci_list, char *netdev_name,
+ struct zpci_netdev **netdev);
+
#endif /* LIB_ZPCI_PCI_LIST_H */
diff --git a/libzpci/pci_list.c b/libzpci/pci_list.c
index 10f64e89..e0d56e44 100644
--- a/libzpci/pci_list.c
+++ b/libzpci/pci_list.c
@@ -356,3 +356,34 @@ void zpci_free_dev_list(struct util_list *zpci_list)
}
util_list_free(zpci_list);
}
+
+/**
+ * Find a PCI device given the name of a netdev
+ *
+ * This function allows finding a PCI device when only the name of one
+ * of its netdevs is known.
+ *
+ * @param[in] zpci_list The device list to search
+ * @param[in] netdev_name The name of the netdev
+ * @param[out] netdev Pointer to store the netdev or NULL if
+ * only the PCI device is needed
+ *
+ * @return The PCI device if one is found NULL otherwise
+ */
+struct zpci_dev *zpci_find_by_netdev(struct util_list *zpci_list, char *netdev_name,
+ struct zpci_netdev **netdev)
+{
+ struct zpci_dev *zdev = NULL;
+ int i;
+
+ util_list_iterate(zpci_list, zdev) {
+ for (i = 0; i < zdev->num_netdevs; i++) {
+ if (!strcmp(zdev->netdevs[i].name, netdev_name)) {
+ if (netdev)
+ *netdev = &zdev->netdevs[i];
+ return zdev;
+ }
+ }
+ }
+ return NULL;
+}
diff --git a/opticsmon/opticsmon.c b/opticsmon/opticsmon.c
index 50dd8d7f..7ecaa125 100644
--- a/opticsmon/opticsmon.c
+++ b/opticsmon/opticsmon.c
@@ -274,38 +274,33 @@ static int oneshot_mode(struct opticsmon_ctx *ctx)
void on_link_change(struct zpci_netdev *netdev, void *arg)
{
struct opticsmon_ctx *ctx = arg;
- struct zpci_dev *zdev;
- int i, reloads = 1;
-
- if (!ctx->zpci_list || util_list_is_empty(ctx->zpci_list))
- zpci_list_reload(&ctx->zpci_list);
+ struct zpci_netdev *found_netdev;
+ struct zpci_dev *zdev = NULL;
+ int reloads = 1;
-find:
- util_list_iterate(ctx->zpci_list, zdev) {
- for (i = 0; i < zdev->num_netdevs; i++) {
- if (!strcmp(zdev->netdevs[i].name, netdev->name)) {
+ do {
+ if (ctx->zpci_list) {
+ zdev = zpci_find_by_netdev(ctx->zpci_list, netdev->name, &found_netdev);
+ if (zdev) {
/* Skip data collection if operational state is
* unchanged
*/
- if (zdev->netdevs[i].operstate == netdev->operstate)
+ if (found_netdev->operstate == netdev->operstate)
return;
/* Update operation state for VFs even though
* they are skipped just for a consistent view
*/
- zdev->netdevs[i].operstate = netdev->operstate;
+ found_netdev->operstate = netdev->operstate;
/* Only collect optics data for PFs */
if (!zpci_is_vf(zdev))
dump_adapter_data(ctx, zdev);
return;
}
}
- }
- /* Might be a new device, reload list of devices and retry */
- if (reloads > 0) {
+ /* Could be uninitalized list or a new device, retry after reload */
zpci_list_reload(&ctx->zpci_list);
reloads--;
- goto find;
- }
+ } while (reloads > 0);
}
#define MAX_EVENTS 8

View File

@ -0,0 +1,15 @@
--- a/zipl/src/job.c 2024-09-16 14:20:09.321762661 +0200
+++ b/zipl/src/job.c 2024-09-16 14:29:28.601846724 +0200
@@ -373,8 +373,11 @@
static void
free_target_data(struct job_target_data* data)
{
+ int i;
+
free(data->bootmap_dir);
- free(data->targetbase);
+ for (i = 0; i < data->nr_targets; i++)
+ free(get_targetbase(data, i));
}
static void

View File

@ -0,0 +1,63 @@
From 592a016a1095fa9813f0bae8256433ba5af4ab9b Mon Sep 17 00:00:00 2001
From: Eduard Shishkin <edward6@linux.ibm.com>
Date: Sat, 7 Dec 2024 12:48:12 +0100
Subject: [PATCH s390-tools 2/2] zipl/src: fix imprecise check that file is on
specified device
This fixes c0f02d2
The check that file is on specified disk is imprecise: In case when
target parameters are specified by user, the check compares a logical
device with a base disk, which is incorrect.
The fixup makes the check compare base disks (a specified one with
the base disk determined by disk_get_info() procedure).
Signed-off-by: Eduard Shishkin <edward6@linux.ibm.com>
---
zipl/src/bootmap.c | 9 +++++----
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/zipl/src/bootmap.c b/zipl/src/bootmap.c
index 7d340156..880b93ce 100644
--- a/zipl/src/bootmap.c
+++ b/zipl/src/bootmap.c
@@ -299,14 +299,15 @@ create_component_header(void* buffer, component_header_type type)
}
/*
- * Not precise check that the file FILENAME locates on specified physical DISK.
+ * Not precise check that the file FILENAME locates on the physical
+ * disk specified by WHERE.
*
* Try to auto-detect parameters of the disk which the file locates on
* and compare found device-ID with DISK.
* Return 0, if auto-detection succeeded, and it is proven that the
* file does NOT locate on DISK. Otherwise, return 1.
*/
-static int file_is_on_disk(const char *filename, dev_t disk)
+static int file_is_on_disk(const char *filename, struct disk_info *where)
{
/*
* Retrieve info of the underlying disk without any user hints
@@ -331,7 +332,7 @@ static int file_is_on_disk(const char *filename, dev_t disk)
"Warning: Preparing a logical device for boot might fail\n");
return 1;
}
- if (info->device != disk) {
+ if (info->basedisks[0] != where->basedisks[0]) {
disk_free_info(info);
return 0;
}
@@ -378,7 +379,7 @@ static int add_component_file_range(struct install_set *bis,
return -1;
}
} else {
- if (!file_is_on_disk(filename, bis->info->device)) {
+ if (!file_is_on_disk(filename, bis->info)) {
error_reason("File is not on target device");
return -1;
}
--
2.39.0

View File

@ -0,0 +1,334 @@
From cf51ac786095f2a1a17d04fea9ee73271438d247 Mon Sep 17 00:00:00 2001
From: Marc Hartmayer <mhartmay@linux.ibm.com>
Date: Wed, 11 Dec 2024 19:25:59 +0100
Subject: [PATCH] rust/pvimg: Add '--(enable|disable)-image-encryption' flags
to 'pvimg create'
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
With runtime attestation it might be useful to have non-encrypted Secure
Execution images. This patch adds the support for this to the 'pvimg
create' and 'genprotimg' commands.
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Acked-by: Hendrik Brueckner <brueckner@linux.ibm.com>
Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
---
rust/pvimg/man/genprotimg.1 | 26 +++++++++++++++++++++-----
rust/pvimg/man/pvimg-create.1 | 26 +++++++++++++++++++++-----
rust/pvimg/man/pvimg-info.1 | 10 +++++-----
rust/pvimg/man/pvimg-test.1 | 10 +++++-----
rust/pvimg/man/pvimg.1 | 10 +++++-----
rust/pvimg/src/cli.rs | 18 ++++++++++++++++++
rust/pvimg/src/cmd/create.rs | 10 ++++++++++
7 files changed, 85 insertions(+), 25 deletions(-)
diff --git a/rust/pvimg/man/genprotimg.1 b/rust/pvimg/man/genprotimg.1
index 46a91aa4..3f4949e9 100644
--- a/rust/pvimg/man/genprotimg.1
+++ b/rust/pvimg/man/genprotimg.1
@@ -3,11 +3,11 @@
.\" it under the terms of the MIT license. See LICENSE for details.
.\"
-.TH genprotimg 1 "2024-12-05" "s390-tools" "Genprotimg Manual"
+.TH genprotimg 1 "2024-12-11" "s390-tools" "Genprotimg Manual"
.nh
.ad l
.SH NAME
-\fBgenprotimg\fP - Create an IBM Secure Execution image
+\fBgenprotimg\fP \- Create an IBM Secure Execution image
\fB
.SH SYNOPSIS
.nf
@@ -196,6 +196,22 @@ Disable the support for backup target keys (default).
.RE
.RE
.PP
+\-\-enable\-image\-encryption
+.RS 4
+Enable encryption of the image components (default). The image components are:
+the kernel, ramdisk, and kernel command line.
+.RE
+.RE
+.PP
+\-\-disable\-image\-encryption
+.RS 4
+Disable encryption of the image components. The image components are: the
+kernel, ramdisk, and kernel command line. Use only if the components used do not
+contain any confidential content (for example, secrets like non\-public
+cryptographic keys).
+.RE
+.RE
+.PP
\-v, \-\-verbose
.RS 4
Provide more detailed output.
@@ -222,16 +238,16 @@ Print help (see a summary with \fB\-h\fR).
.SH EXIT STATUS
.TP 8
-.B 0 - Program finished successfully
+.B 0 \- Program finished successfully
The command was executed successfully.
.RE
.TP 8
-.B 1 - Generic error
+.B 1 \- Generic error
Something went wrong during the operation. Refer to the error
message.
.RE
.TP 8
-.B 2 - Usage error
+.B 2 \- Usage error
The command was used incorrectly, for example: unsupported command
line flag, or wrong number of arguments.
.RE
diff --git a/rust/pvimg/man/pvimg-create.1 b/rust/pvimg/man/pvimg-create.1
index aba197fa..dae1cf18 100644
--- a/rust/pvimg/man/pvimg-create.1
+++ b/rust/pvimg/man/pvimg-create.1
@@ -3,11 +3,11 @@
.\" it under the terms of the MIT license. See LICENSE for details.
.\"
-.TH pvimg-create 1 "2024-12-05" "s390-tools" "Pvimg Manual"
+.TH pvimg-create 1 "2024-12-11" "s390-tools" "Pvimg Manual"
.nh
.ad l
.SH NAME
-\fBpvimg create\fP - Create an IBM Secure Execution image
+\fBpvimg create\fP \- Create an IBM Secure Execution image
\fB
.SH SYNOPSIS
.nf
@@ -195,6 +195,22 @@ Disable the support for backup target keys (default).
.RE
.RE
.PP
+\-\-enable\-image\-encryption
+.RS 4
+Enable encryption of the image components (default). The image components are:
+the kernel, ramdisk, and kernel command line.
+.RE
+.RE
+.PP
+\-\-disable\-image\-encryption
+.RS 4
+Disable encryption of the image components. The image components are: the
+kernel, ramdisk, and kernel command line. Use only if the components used do not
+contain any confidential content (for example, secrets like non\-public
+cryptographic keys).
+.RE
+.RE
+.PP
\-h, \-\-help
.RS 4
Print help (see a summary with \fB\-h\fR).
@@ -203,16 +219,16 @@ Print help (see a summary with \fB\-h\fR).
.SH EXIT STATUS
.TP 8
-.B 0 - Program finished successfully
+.B 0 \- Program finished successfully
The command was executed successfully.
.RE
.TP 8
-.B 1 - Generic error
+.B 1 \- Generic error
Something went wrong during the operation. Refer to the error
message.
.RE
.TP 8
-.B 2 - Usage error
+.B 2 \- Usage error
The command was used incorrectly, for example: unsupported command
line flag, or wrong number of arguments.
.RE
diff --git a/rust/pvimg/man/pvimg-info.1 b/rust/pvimg/man/pvimg-info.1
index e88cbe49..d2726c35 100644
--- a/rust/pvimg/man/pvimg-info.1
+++ b/rust/pvimg/man/pvimg-info.1
@@ -3,11 +3,11 @@
.\" it under the terms of the MIT license. See LICENSE for details.
.\"
-.TH pvimg-info 1 "2024-12-05" "s390-tools" "Pvimg Manual"
+.TH pvimg-info 1 "2024-12-11" "s390-tools" "Pvimg Manual"
.nh
.ad l
.SH NAME
-\fBpvimg info\fP - Print information about the IBM Secure Execution image
+\fBpvimg info\fP \- Print information about the IBM Secure Execution image
\fB
.SH SYNOPSIS
.nf
@@ -51,16 +51,16 @@ Print help (see a summary with \fB\-h\fR).
.SH EXIT STATUS
.TP 8
-.B 0 - Program finished successfully
+.B 0 \- Program finished successfully
The command was executed successfully.
.RE
.TP 8
-.B 1 - Generic error
+.B 1 \- Generic error
Something went wrong during the operation. Refer to the error
message.
.RE
.TP 8
-.B 2 - Usage error
+.B 2 \- Usage error
The command was used incorrectly, for example: unsupported command
line flag, or wrong number of arguments.
.RE
diff --git a/rust/pvimg/man/pvimg-test.1 b/rust/pvimg/man/pvimg-test.1
index 901c7edb..4fb7d73f 100644
--- a/rust/pvimg/man/pvimg-test.1
+++ b/rust/pvimg/man/pvimg-test.1
@@ -3,11 +3,11 @@
.\" it under the terms of the MIT license. See LICENSE for details.
.\"
-.TH pvimg-test 1 "2024-12-05" "s390-tools" "Pvimg Manual"
+.TH pvimg-test 1 "2024-12-11" "s390-tools" "Pvimg Manual"
.nh
.ad l
.SH NAME
-\fBpvimg test\fP - Test different aspects of an existing IBM Secure Execution image
+\fBpvimg test\fP \- Test different aspects of an existing IBM Secure Execution image
\fB
.SH SYNOPSIS
.nf
@@ -54,16 +54,16 @@ Print help (see a summary with \fB\-h\fR).
.SH EXIT STATUS
.TP 8
-.B 0 - Program finished successfully
+.B 0 \- Program finished successfully
The command was executed successfully.
.RE
.TP 8
-.B 1 - Generic error
+.B 1 \- Generic error
Something went wrong during the operation. Refer to the error
message.
.RE
.TP 8
-.B 2 - Usage error
+.B 2 \- Usage error
The command was used incorrectly, for example: unsupported command
line flag, or wrong number of arguments.
.RE
diff --git a/rust/pvimg/man/pvimg.1 b/rust/pvimg/man/pvimg.1
index 37c8e978..5676b61d 100644
--- a/rust/pvimg/man/pvimg.1
+++ b/rust/pvimg/man/pvimg.1
@@ -3,11 +3,11 @@
.\" it under the terms of the MIT license. See LICENSE for details.
.\"
-.TH pvimg 1 "2024-12-05" "s390-tools" "Pvimg Manual"
+.TH pvimg 1 "2024-12-11" "s390-tools" "Pvimg Manual"
.nh
.ad l
.SH NAME
-\fBpvimg\fP - Create and inspect IBM Secure Execution images
+\fBpvimg\fP \- Create and inspect IBM Secure Execution images
\fB
.SH SYNOPSIS
.nf
@@ -69,16 +69,16 @@ Print help (see a summary with \fB\-h\fR).
.SH EXIT STATUS
.TP 8
-.B 0 - Program finished successfully
+.B 0 \- Program finished successfully
The command was executed successfully.
.RE
.TP 8
-.B 1 - Generic error
+.B 1 \- Generic error
Something went wrong during the operation. Refer to the error
message.
.RE
.TP 8
-.B 2 - Usage error
+.B 2 \- Usage error
The command was used incorrectly, for example: unsupported command
line flag, or wrong number of arguments.
.RE
diff --git a/rust/pvimg/src/cli.rs b/rust/pvimg/src/cli.rs
index 2ca4e901..12f0b764 100644
--- a/rust/pvimg/src/cli.rs
+++ b/rust/pvimg/src/cli.rs
@@ -140,6 +140,20 @@ pub struct CreateBootImageLegacyFlags {
/// Disable the support for backup target keys (default).
#[arg(long, action = clap::ArgAction::SetTrue, conflicts_with="enable_backup_keys", group="header-flags")]
pub disable_backup_keys: Option<bool>,
+
+ /// Enable encryption of the image components (default).
+ ///
+ /// The image components are: the kernel, ramdisk, and kernel command line.
+ #[arg(long, action = clap::ArgAction::SetTrue, group="header-flags")]
+ pub enable_image_encryption: Option<bool>,
+
+ /// Disable encryption of the image components.
+ ///
+ /// The image components are: the kernel, ramdisk, and kernel command line.
+ /// Use only if the components used do not contain any confidential content
+ /// (for example, secrets like non-public cryptographic keys).
+ #[arg(long, action = clap::ArgAction::SetTrue, conflicts_with="enable_image_encryption", group="header-flags")]
+ pub disable_image_encryption: Option<bool>,
}
#[non_exhaustive]
@@ -476,6 +490,8 @@ mod test {
flat_map_collect(insert(mvca.clone(), vec![CliOption::new("enable-pckmo", ["--enable-pckmo"])])),
flat_map_collect(insert(mvca.clone(), vec![CliOption::new("enable-pckmo-hmac", ["--enable-pckmo-hmac"])])),
flat_map_collect(insert(mvca.clone(), vec![CliOption::new("enable-backup-keys", ["--enable-backup-keys"])])),
+ flat_map_collect(insert(mvca.clone(), vec![CliOption::new("disable-image-encryption", ["--disable-image-encryption"])])),
+ flat_map_collect(insert(mvca.clone(), vec![CliOption::new("enable-image-encryption", ["--enable-image-encryption"])])),
];
let invalid_create_args = [
flat_map_collect(remove(mvcanv.clone(), "no-verify")),
@@ -501,6 +517,8 @@ mod test {
CliOption::new("x-pcf2", ["--x-pcf", "0x0"])])),
flat_map_collect(insert(mvca.clone(), vec![CliOption::new("enable-pckmo", ["--enable-pckmo"]),
CliOption::new("disable-pckmo", ["--disable-pckmo"])])),
+ flat_map_collect(insert(mvca.clone(), vec![CliOption::new("enable-image-encryption", ["--enable-image-encryption"]),
+ CliOption::new("disable-image-encryption", ["--disable-image-encryption"])])),
];
let mut genprotimg_valid_args = vec![
diff --git a/rust/pvimg/src/cmd/create.rs b/rust/pvimg/src/cmd/create.rs
index b696d790..475d3523 100644
--- a/rust/pvimg/src/cmd/create.rs
+++ b/rust/pvimg/src/cmd/create.rs
@@ -80,6 +80,12 @@ fn parse_flags(
lf.enable_backup_keys
.filter(|x| *x)
.and(Some(PcfV1::all_enabled([PcfV1::BackupTargetKeys]))),
+ lf.disable_image_encryption
+ .filter(|x| *x)
+ .and(Some(PcfV1::all_enabled([PcfV1::NoComponentEncryption]))),
+ lf.enable_image_encryption
+ .filter(|x| *x)
+ .and(Some(PcfV1::all_disabled([PcfV1::NoComponentEncryption]))),
]
.into_iter()
.flatten()
@@ -135,6 +141,10 @@ pub fn create(opt: &CreateBootImageArgs) -> Result<OwnExitCode> {
read_user_provided_keys(opt.comm_key.as_deref(), &opt.experimental_args)?;
let (plaintext_flags, secret_flags) = parse_flags(opt)?;
+ if plaintext_flags.is_set(PcfV1::NoComponentEncryption) {
+ warn!("The components encryption is disabled, make sure that the components do not contain any confidential content.");
+ }
+
let mut components = components(&opt.component_paths)?;
if opt.no_component_check {
warn!("The component check is turned off!");

BIN
s390-tools-2.31.0.tar.gz (Stored with Git LFS) Normal file

Binary file not shown.

View File

@ -0,0 +1,51 @@
From 6a55d0c2e57952600164822dd100e8247b4b010f Mon Sep 17 00:00:00 2001
From: Steffen Eiden <seiden@linux.ibm.com>
Date: Fri, 23 Aug 2024 09:16:26 +0200
Subject: [PATCH] rust/pv: Lower most lints to warn
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Lower the lint level to warn for the styling lints.
This avoids compile issues during packaging for newer tooling with
potential more lint findings.
Still deny compiling if a public symbol has no documentation.
Fixes: https://github.com/ibm-s390-linux/s390-tools/issues/173
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
---
rust/pv/src/lib.rs | 4 ++--
rust/pv_core/src/lib.rs | 4 ++--
2 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/rust/pv/src/lib.rs b/rust/pv/src/lib.rs
index 9a647617..1084f8e8 100644
--- a/rust/pv/src/lib.rs
+++ b/rust/pv/src/lib.rs
@@ -2,8 +2,8 @@
//
// Copyright IBM Corp. 2023, 2024
-#![deny(
- missing_docs,
+#![deny(missing_docs)]
+#![warn(
missing_debug_implementations,
trivial_numeric_casts,
unstable_features,
diff --git a/rust/pv_core/src/lib.rs b/rust/pv_core/src/lib.rs
index 1356c1b7..b617b8f9 100644
--- a/rust/pv_core/src/lib.rs
+++ b/rust/pv_core/src/lib.rs
@@ -1,8 +1,8 @@
// SPDX-License-Identifier: MIT
//
// Copyright IBM Corp. 2023, 2024
-#![deny(
- missing_docs,
+#![deny(missing_docs)]
+#![warn(
missing_debug_implementations,
trivial_numeric_casts,
unstable_features,

3
s390-tools-2.34.0.tar.gz Normal file
View File

@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:ea4758c4e460d7f7e040e6aedf68b1be32d63fecb733358b08182f6b9b7440a2
size 2114507

BIN
s390-tools-2.35.0.tar.gz (Stored with Git LFS) Normal file

Binary file not shown.

BIN
s390-tools-2.36.0.tar.gz (Stored with Git LFS) Normal file

Binary file not shown.

View File

@ -0,0 +1,133 @@
---
zdev/dracut/96zdev-live/module-setup.sh | 32 +++++++++++++++++++++++++
zdev/dracut/96zdev-live/parse-zdev-live.sh | 36 +++++++++++++++++++++++++++++
zdev/dracut/96zdev-live/write-udev-live.sh | 11 ++++++++
zdev/dracut/Makefile | 15 ++++++++++--
4 files changed, 92 insertions(+), 2 deletions(-)
--- /dev/null
+++ b/zdev/dracut/96zdev-live/module-setup.sh
@@ -0,0 +1,32 @@
+#!/bin/bash
+
+# called by dracut
+check() {
+ arch=${DRACUT_ARCH:-$(uname -m)}
+ [ "$arch" = "s390" -o "$arch" = "s390x" ] || return 1
+
+ require_binaries chzdev || return 1
+
+ [[ $hostonly ]] || return 0
+
+ # or on request
+ return 255
+}
+
+# called by dracut
+depends() {
+ echo bash
+ return 0
+}
+
+# called by dracut
+installkernel() {
+ instmods ctcm lcs qeth qeth_l2 qeth_l3 dasd_diag_mod dasd_eckd_mod dasd_fba_mod
+}
+
+# called by dracut
+install() {
+ inst_hook cmdline 41 "$moddir/parse-zdev-live.sh"
+ inst_hook cleanup 41 "$moddir/write-udev-live.sh"
+ inst_multiple chzdev
+}
--- /dev/null
+++ b/zdev/dracut/96zdev-live/parse-zdev-live.sh
@@ -0,0 +1,36 @@
+#!/bin/bash
+#
+# 96zdev-live/parse-zdev-live.sh
+# Parse the kernel command line for rd.zdev kernel parameters. These
+# parameters are evaluated and used to configure z Systems specific devices
+# with chzdev(8), especially for use on live/installation type media.
+# Note: this is only active on no-hostonly initrds (by default).
+#
+# Format:
+# rd.zdev=TYPE,DEVICE[,SETTINGS]
+#
+# where
+#
+# TYPE: all device types supported by chzdev(8), like qeth and dasd
+# DEVICE: device specification as supported by chzdev(8) '--enable',
+# with the exception of specifying multiple devices, which
+# need to be separated by commas. Channel group members
+# (or zFCP parameters) in turn are separated by colons.
+# SETTINGS: Settings are positional arguments of chzdev in the form
+# KEY=VALUE separated by commas.
+
+zdev_enable="--persistent --enable"
+zdev_base_args="--yes --no-root-update --no-settle"
+
+for zdevs in $(getargs rd.zdev) ; do
+ IFS=',' read -r -a zdev <<< "$zdevs"
+ if [ -n "$zdev" ] && [ "$zdev" = "no-auto" -o "$zdev" = "auto" ] ; then
+ : # ignore, as it's handled by 95zdev
+ elif [ -z "$zdev" ] || [ -z "${zdev[1]}" ] ; then
+ warn "Unsupported usage of rd.zdev=$zdevs"
+ else
+ info "+ chzdev $zdev_enable [...] ${zdev[@]}"
+ chzdev $zdev_enable $zdev_base_args "${zdev[@]}"
+ fi
+done
+
--- /dev/null
+++ b/zdev/dracut/96zdev-live/write-udev-live.sh
@@ -0,0 +1,11 @@
+#!/bin/sh
+#
+# 96zdev-live/write-udev-live.sh
+# Copy udeve rules generated by chzdev for device activation starting with 41
+# to a *writable* /sysroot -- this is primarily useful for live/installation-
+# type media (and by default only active on no-hostonly initrds)
+#
+
+if [ -w /sysroot/etc/udev/rules.d ]; then
+ cp -p /etc/udev/rules.d/41-* /sysroot/etc/udev/rules.d
+fi
--- a/zdev/dracut/Makefile
+++ b/zdev/dracut/Makefile
@@ -3,17 +3,23 @@
ZDEVDIR := 95zdev
ZDEVKDUMPDIR := 95zdev-kdump
+ZDEVLIVEDIR := 96zdev-live
# HAVE_DRACUT
#
-# This install time parameter determines whether the zdev dracut module is
-# installed (HAVE_DRACUT=1) or not (default). When installed, the module
+# This install time parameter determines whether the zdev dracut modules are
+# installed (HAVE_DRACUT=1) or not (default). When installed, the 95zdev module
# performs the following functions when dracut is run:
#
# - copy the persistent root device configuration to the initial ram disk
# - install a boot-time hook to apply firmware-provided configuration data
# to the system
#
+# The 96zdev-live module performs the following functions when dracut is run:
+#
+# - install a boot-time hook to apply command-line-provided configuration data
+# to a no-hostonly built initial ram disk for use in live/installation media
+#
ifeq ($(HAVE_DRACUT),1)
install:
$(INSTALL) -m 755 -d $(DESTDIR)$(DRACUTMODDIR)/
@@ -29,4 +35,9 @@
$(INSTALL) -m 755 -d $(DESTDIR)$(DRACUTMODDIR)/$(ZDEVKDUMPDIR)
$(INSTALL) -m 755 $(ZDEVKDUMPDIR)/module-setup.sh \
$(DESTDIR)$(DRACUTMODDIR)/$(ZDEVKDUMPDIR)/
+ $(INSTALL) -m 755 -d $(DESTDIR)$(DRACUTMODDIR)/$(ZDEVLIVEDIR)
+ $(INSTALL) -m 755 $(ZDEVLIVEDIR)/module-setup.sh \
+ $(ZDEVLIVEDIR)/parse-zdev-live.sh \
+ $(ZDEVLIVEDIR)/write-udev-live.sh \
+ $(DESTDIR)$(DRACUTMODDIR)/$(ZDEVLIVEDIR)/
endif

View File

@ -0,0 +1,64 @@
From dff965465ca9d9c4edaf0f90eadd9a6de335b354 Mon Sep 17 00:00:00 2001
From: Niklas Schnelle <schnelle@linux.ibm.com>
Date: Fri, 6 Dec 2024 15:28:08 +0100
Subject: [PATCH] opticsmon: Fix runaway loop in on_link_change()
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
When on_link_change() gets called with a netdev that would be monitored
but hasn't entered zpci_list yet, reloads is 1 after the loops and
a reload occurs. Then the netdev is found in the list and reloads
becomes -1 which incorrectly triggers more reloads until underflow.
Fix this by returning once the device is found. Also just check for
reloads being larger than zero.
Fixes: c34adb9cabee ("opticsmon: Introduce opticsmon tool")
Reviewed-by: Halil Pasic <pasic@linux.ibm.com>
Signed-off-by: Niklas Schnelle <schnelle@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
---
opticsmon/opticsmon.c | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/opticsmon/opticsmon.c b/opticsmon/opticsmon.c
index c2f355e2..50dd8d7f 100644
--- a/opticsmon/opticsmon.c
+++ b/opticsmon/opticsmon.c
@@ -280,16 +280,15 @@ void on_link_change(struct zpci_netdev *netdev, void *arg)
if (!ctx->zpci_list || util_list_is_empty(ctx->zpci_list))
zpci_list_reload(&ctx->zpci_list);
-reload:
+find:
util_list_iterate(ctx->zpci_list, zdev) {
for (i = 0; i < zdev->num_netdevs; i++) {
if (!strcmp(zdev->netdevs[i].name, netdev->name)) {
- reloads--;
/* Skip data collection if operational state is
* unchanged
*/
if (zdev->netdevs[i].operstate == netdev->operstate)
- continue;
+ return;
/* Update operation state for VFs even though
* they are skipped just for a consistent view
*/
@@ -297,14 +296,15 @@ void on_link_change(struct zpci_netdev *netdev, void *arg)
/* Only collect optics data for PFs */
if (!zpci_is_vf(zdev))
dump_adapter_data(ctx, zdev);
+ return;
}
}
}
/* Might be a new device, reload list of devices and retry */
- if (reloads) {
+ if (reloads > 0) {
zpci_list_reload(&ctx->zpci_list);
reloads--;
- goto reload;
+ goto find;
}
}

View File

@ -0,0 +1,129 @@
From cf5560a100b5552e2eeeaac9c60a88ae77233530 Mon Sep 17 00:00:00 2001
From: Niklas Schnelle <schnelle@linux.ibm.com>
Date: Mon, 9 Dec 2024 15:08:03 +0100
Subject: [PATCH] libzpci: opticsmon: Refactor on_link_change() using new
zpci_find_by_netdev()
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Finding a PCI device given the name of a netdev seems generally useful
so pull this out into a new zpci_find_by_netdev() function in libzpci
and use this to simplify on_link_change() removing the need for
backwards goto.
Reviewed-by: Halil Pasic <pasic@linux.ibm.com>
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Niklas Schnelle <schnelle@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
---
include/lib/pci_list.h | 3 +++
libzpci/pci_list.c | 31 +++++++++++++++++++++++++++++++
opticsmon/opticsmon.c | 27 +++++++++++----------------
3 files changed, 45 insertions(+), 16 deletions(-)
diff --git a/include/lib/pci_list.h b/include/lib/pci_list.h
index 829ec244..5b2918bc 100644
--- a/include/lib/pci_list.h
+++ b/include/lib/pci_list.h
@@ -93,4 +93,7 @@ const char *zpci_pft_str(struct zpci_dev *zdev);
const char *zpci_operstate_str(operstate_t state);
operstate_t zpci_operstate_from_str(const char *oper_str);
+struct zpci_dev *zpci_find_by_netdev(struct util_list *zpci_list, char *netdev_name,
+ struct zpci_netdev **netdev);
+
#endif /* LIB_ZPCI_PCI_LIST_H */
diff --git a/libzpci/pci_list.c b/libzpci/pci_list.c
index 10f64e89..e0d56e44 100644
--- a/libzpci/pci_list.c
+++ b/libzpci/pci_list.c
@@ -356,3 +356,34 @@ void zpci_free_dev_list(struct util_list *zpci_list)
}
util_list_free(zpci_list);
}
+
+/**
+ * Find a PCI device given the name of a netdev
+ *
+ * This function allows finding a PCI device when only the name of one
+ * of its netdevs is known.
+ *
+ * @param[in] zpci_list The device list to search
+ * @param[in] netdev_name The name of the netdev
+ * @param[out] netdev Pointer to store the netdev or NULL if
+ * only the PCI device is needed
+ *
+ * @return The PCI device if one is found NULL otherwise
+ */
+struct zpci_dev *zpci_find_by_netdev(struct util_list *zpci_list, char *netdev_name,
+ struct zpci_netdev **netdev)
+{
+ struct zpci_dev *zdev = NULL;
+ int i;
+
+ util_list_iterate(zpci_list, zdev) {
+ for (i = 0; i < zdev->num_netdevs; i++) {
+ if (!strcmp(zdev->netdevs[i].name, netdev_name)) {
+ if (netdev)
+ *netdev = &zdev->netdevs[i];
+ return zdev;
+ }
+ }
+ }
+ return NULL;
+}
diff --git a/opticsmon/opticsmon.c b/opticsmon/opticsmon.c
index 50dd8d7f..7ecaa125 100644
--- a/opticsmon/opticsmon.c
+++ b/opticsmon/opticsmon.c
@@ -274,38 +274,33 @@ static int oneshot_mode(struct opticsmon_ctx *ctx)
void on_link_change(struct zpci_netdev *netdev, void *arg)
{
struct opticsmon_ctx *ctx = arg;
- struct zpci_dev *zdev;
- int i, reloads = 1;
-
- if (!ctx->zpci_list || util_list_is_empty(ctx->zpci_list))
- zpci_list_reload(&ctx->zpci_list);
+ struct zpci_netdev *found_netdev;
+ struct zpci_dev *zdev = NULL;
+ int reloads = 1;
-find:
- util_list_iterate(ctx->zpci_list, zdev) {
- for (i = 0; i < zdev->num_netdevs; i++) {
- if (!strcmp(zdev->netdevs[i].name, netdev->name)) {
+ do {
+ if (ctx->zpci_list) {
+ zdev = zpci_find_by_netdev(ctx->zpci_list, netdev->name, &found_netdev);
+ if (zdev) {
/* Skip data collection if operational state is
* unchanged
*/
- if (zdev->netdevs[i].operstate == netdev->operstate)
+ if (found_netdev->operstate == netdev->operstate)
return;
/* Update operation state for VFs even though
* they are skipped just for a consistent view
*/
- zdev->netdevs[i].operstate = netdev->operstate;
+ found_netdev->operstate = netdev->operstate;
/* Only collect optics data for PFs */
if (!zpci_is_vf(zdev))
dump_adapter_data(ctx, zdev);
return;
}
}
- }
- /* Might be a new device, reload list of devices and retry */
- if (reloads > 0) {
+ /* Could be uninitalized list or a new device, retry after reload */
zpci_list_reload(&ctx->zpci_list);
reloads--;
- goto find;
- }
+ } while (reloads > 0);
}
#define MAX_EVENTS 8

View File

@ -0,0 +1,147 @@
From 1e44ace41de3cbd744b22a8f9835473b091186e0 Mon Sep 17 00:00:00 2001
From: Steffen Eiden <seiden@linux.ibm.com>
Date: Thu, 18 Jul 2024 10:55:45 +0200
Subject: [PATCH] rust/pvsecret: Refactor writing secret
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Refactor the writing of secret-type dependent output files to ease
extensions.
Reviewed-by: Marc Hartmayer <marc@linux.ibm.com>
Reviewed-by: Christoph Schlameuss <schlameuss@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
---
rust/pv/src/uvsecret/guest_secret.rs | 2 +-
rust/pvsecret/src/cmd/create.rs | 89 +++++++++++++++-------------
2 files changed, 48 insertions(+), 43 deletions(-)
diff --git a/rust/pv/src/uvsecret/guest_secret.rs b/rust/pv/src/uvsecret/guest_secret.rs
index 509691fa..4f1db31c 100644
--- a/rust/pv/src/uvsecret/guest_secret.rs
+++ b/rust/pv/src/uvsecret/guest_secret.rs
@@ -68,7 +68,7 @@ impl GuestSecret {
}
/// Reference to the confidential data
- pub(crate) fn confidential(&self) -> &[u8] {
+ pub fn confidential(&self) -> &[u8] {
match &self {
Self::Null => &[],
Self::Association { secret, .. } => secret.value().as_slice(),
diff --git a/rust/pvsecret/src/cmd/create.rs b/rust/pvsecret/src/cmd/create.rs
index 808b29e1..9251c38c 100644
--- a/rust/pvsecret/src/cmd/create.rs
+++ b/rust/pvsecret/src/cmd/create.rs
@@ -62,7 +62,7 @@ pub fn create(opt: &CreateSecretOpt) -> Result<()> {
write_out(&opt.output, ser_asrbc, "add-secret request")?;
info!("Successfully wrote the request to '{}'", &opt.output);
- write_secret(&opt.secret, &asrcb, &opt.output)
+ write_secret(&opt.secret, asrcb.guest_secret(), &opt.output)
}
/// Read+parse the first key from the buffer.
@@ -206,54 +206,59 @@ fn read_cuid(asrcb: &mut AddSecretRequest, opt: &CreateSecretOpt) -> Result<()>
Ok(())
}
+// Write non confidential data (=name+id) to a yaml stdout
+fn write_yaml<P: AsRef<Path>>(
+ name: &str,
+ guest_secret: &GuestSecret,
+ stdout: &bool,
+ outp_path: P,
+) -> Result<()> {
+ debug!("Non-confidential secret information: {guest_secret:x?}");
+
+ let secret_info = serde_yaml::to_string(guest_secret)?;
+ if stdout.to_owned() {
+ println!("{secret_info}");
+ return Ok(());
+ }
+
+ let gen_name: String = name
+ .chars()
+ .map(|c| if c.is_whitespace() { '_' } else { c })
+ .collect();
+ let mut yaml_path = outp_path
+ .as_ref()
+ .parent()
+ .with_context(|| format!("Cannot open directory of {:?}", outp_path.as_ref()))?
+ .to_owned();
+ yaml_path.push(gen_name);
+ yaml_path.set_extension("yaml");
+ write_out(&yaml_path, secret_info, "secret information")?;
+ warn!(
+ "Successfully wrote secret info to '{}'",
+ yaml_path.display().to_string()
+ );
+ Ok(())
+}
+
/// Write the generated secret (if any) to the specified output stream
fn write_secret<P: AsRef<Path>>(
secret: &AddSecretType,
- asrcb: &AddSecretRequest,
+ guest_secret: &GuestSecret,
outp_path: P,
) -> Result<()> {
- if let AddSecretType::Association {
- name,
- stdout,
- output_secret: secret_out,
- ..
- } = secret
- {
- let gen_name: String = name
- .chars()
- .map(|c| if c.is_whitespace() { '_' } else { c })
- .collect();
- let mut gen_path = outp_path
- .as_ref()
- .parent()
- .with_context(|| format!("Cannot open directory of {:?}", outp_path.as_ref()))?
- .to_owned();
- gen_path.push(format!("{gen_name}.yaml"));
-
- // write non confidential data (=name+id) to a yaml
- let secret_info = serde_yaml::to_string(asrcb.guest_secret())?;
- if stdout.to_owned() {
- println!("{secret_info}");
- } else {
- write_out(&gen_path, secret_info, "association secret info")?;
- debug!(
- "Non-confidential secret information: {:x?}",
- asrcb.guest_secret()
- );
- warn!(
- "Successfully wrote association info to '{}'",
- gen_path.display()
- );
- }
-
- if let Some(path) = secret_out {
- if let GuestSecret::Association { secret, .. } = asrcb.guest_secret() {
- write_out(path, secret.value(), "Association secret")?
- } else {
- unreachable!("The secret type has to be `association` at this point (bug)!")
+ match secret {
+ AddSecretType::Association {
+ name,
+ stdout,
+ output_secret,
+ ..
+ } => {
+ write_yaml(name, guest_secret, stdout, outp_path)?;
+ if let Some(path) = output_secret {
+ write_out(path, guest_secret.confidential(), "Association secret")?
}
- info!("Successfully wrote generated association secret to '{path}'");
}
+ _ => (),
};
Ok(())
}

View File

@ -0,0 +1,467 @@
From d1636168b26cc842bc0766235c8a4f2da9663f20 Mon Sep 17 00:00:00 2001
From: Steffen Eiden <seiden@linux.ibm.com>
Date: Tue, 5 Mar 2024 10:46:29 +0100
Subject: [PATCH] rust/pv: Support for writing data in PEM format
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Use existing OpenSSL functionalities to create PEM files containing
arbitrary data.
Acked-by: Marc Hartmayer <marc@linux.ibm.com>
Acked-by: Christoph Schlameuss <schlameuss@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
---
rust/pv/src/error.rs | 3 +
rust/pv/src/lib.rs | 6 +
rust/pv/src/openssl_extensions/bio.rs | 85 +++++++
rust/pv/src/openssl_extensions/mod.rs | 2 +
.../src/openssl_extensions/stackable_crl.rs | 41 +---
rust/pv/src/pem_utils.rs | 222 ++++++++++++++++++
6 files changed, 321 insertions(+), 38 deletions(-)
create mode 100644 rust/pv/src/openssl_extensions/bio.rs
create mode 100644 rust/pv/src/pem_utils.rs
diff --git a/rust/pv/src/error.rs b/rust/pv/src/error.rs
index af85e93e..3ba808f2 100644
--- a/rust/pv/src/error.rs
+++ b/rust/pv/src/error.rs
@@ -106,6 +106,9 @@ pub enum Error {
)]
AddDataMissing(&'static str),
+ #[error("An ASCII string was expected, but non-ASCII characters were received.")]
+ NonAscii,
+
// errors from other crates
#[error(transparent)]
PvCore(#[from] pv_core::Error),
diff --git a/rust/pv/src/lib.rs b/rust/pv/src/lib.rs
index 7a33210c..ec31b9a4 100644
--- a/rust/pv/src/lib.rs
+++ b/rust/pv/src/lib.rs
@@ -37,6 +37,7 @@ mod brcb;
mod crypto;
mod error;
mod openssl_extensions;
+mod pem_utils;
mod req;
mod utils;
mod uvattest;
@@ -71,6 +72,11 @@ pub mod attest {
};
}
+/// Definitions and functions to write objects in PEM format
+pub mod pem {
+ pub use crate::pem_utils::Pem;
+}
+
/// Miscellaneous functions and definitions
pub mod misc {
pub use pv_core::misc::*;
diff --git a/rust/pv/src/openssl_extensions/bio.rs b/rust/pv/src/openssl_extensions/bio.rs
new file mode 100644
index 00000000..73528eed
--- /dev/null
+++ b/rust/pv/src/openssl_extensions/bio.rs
@@ -0,0 +1,85 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright IBM Corp. 2024
+
+use core::slice;
+use openssl::error::ErrorStack;
+use openssl_sys::BIO_new_mem_buf;
+use std::ffi::c_int;
+use std::{marker::PhantomData, ptr};
+
+pub struct BioMem(*mut openssl_sys::BIO);
+
+impl Drop for BioMem {
+ fn drop(&mut self) {
+ // SAFETY: Pointer is valid. The pointer value is dropped after the free.
+ unsafe {
+ openssl_sys::BIO_free_all(self.0);
+ }
+ }
+}
+
+impl BioMem {
+ pub fn new() -> Result<Self, ErrorStack> {
+ openssl_sys::init();
+
+ // SAFETY: Returns a valid pointer or null. null-case is tested right after this.
+ let bio = unsafe { openssl_sys::BIO_new(openssl_sys::BIO_s_mem()) };
+ match bio.is_null() {
+ true => Err(ErrorStack::get()),
+ false => Ok(Self(bio)),
+ }
+ }
+
+ pub fn as_ptr(&self) -> *mut openssl_sys::BIO {
+ self.0
+ }
+
+ /// Copies the content of this slice into a Vec
+ pub fn to_vec(&self) -> Vec<u8> {
+ let buf;
+ // SAFTEY: BIO provides a continuous memory that can be used to build a slice.
+ unsafe {
+ let mut ptr = ptr::null_mut();
+ let len = openssl_sys::BIO_get_mem_data(self.0, &mut ptr);
+ buf = slice::from_raw_parts(ptr as *const _ as *const _, len as usize)
+ }
+ buf.to_vec()
+ }
+}
+
+pub struct BioMemSlice<'a>(*mut openssl_sys::BIO, PhantomData<&'a [u8]>);
+impl Drop for BioMemSlice<'_> {
+ fn drop(&mut self) {
+ // SAFETY: Pointer is valid. The pointer value is dropped after the free.
+ unsafe {
+ openssl_sys::BIO_free_all(self.0);
+ }
+ }
+}
+
+impl<'a> BioMemSlice<'a> {
+ pub fn new(buf: &'a [u8]) -> Result<BioMemSlice<'a>, ErrorStack> {
+ openssl_sys::init();
+
+ // SAFETY: `buf` is a slice (i.e. pointer+size) pointing to a valid memory region.
+ // So the resulting bio is valid. Lifetime of the slice is connected by this Rust
+ // structure.
+ assert!(buf.len() <= c_int::MAX as usize);
+ let bio = unsafe {
+ {
+ let r = BIO_new_mem_buf(buf.as_ptr() as *const _, buf.len() as c_int);
+ match r.is_null() {
+ true => Err(ErrorStack::get()),
+ false => Ok(r),
+ }
+ }?
+ };
+
+ Ok(BioMemSlice(bio, PhantomData))
+ }
+
+ pub fn as_ptr(&self) -> *mut openssl_sys::BIO {
+ self.0
+ }
+}
diff --git a/rust/pv/src/openssl_extensions/mod.rs b/rust/pv/src/openssl_extensions/mod.rs
index fab26638..f6234e5d 100644
--- a/rust/pv/src/openssl_extensions/mod.rs
+++ b/rust/pv/src/openssl_extensions/mod.rs
@@ -6,8 +6,10 @@
/// Extensions to the rust-openssl crate
mod akid;
+mod bio;
mod crl;
mod stackable_crl;
pub use akid::*;
+pub use bio::*;
pub use crl::*;
diff --git a/rust/pv/src/openssl_extensions/stackable_crl.rs b/rust/pv/src/openssl_extensions/stackable_crl.rs
index aef7cf86..12a9f9de 100644
--- a/rust/pv/src/openssl_extensions/stackable_crl.rs
+++ b/rust/pv/src/openssl_extensions/stackable_crl.rs
@@ -2,16 +2,14 @@
//
// Copyright IBM Corp. 2023
-use std::{marker::PhantomData, ptr};
-
+use crate::openssl_extensions::bio::BioMemSlice;
use foreign_types::{ForeignType, ForeignTypeRef};
use openssl::{
error::ErrorStack,
stack::Stackable,
x509::{X509Crl, X509CrlRef},
};
-use openssl_sys::BIO_new_mem_buf;
-use std::ffi::c_int;
+use std::ptr;
#[derive(Debug)]
pub struct StackableX509Crl(*mut openssl_sys::X509_CRL);
@@ -62,44 +60,11 @@ impl Stackable for StackableX509Crl {
type StackType = openssl_sys::stack_st_X509_CRL;
}
-pub struct MemBioSlice<'a>(*mut openssl_sys::BIO, PhantomData<&'a [u8]>);
-impl Drop for MemBioSlice<'_> {
- fn drop(&mut self) {
- unsafe {
- openssl_sys::BIO_free_all(self.0);
- }
- }
-}
-
-impl<'a> MemBioSlice<'a> {
- pub fn new(buf: &'a [u8]) -> Result<MemBioSlice<'a>, ErrorStack> {
- openssl_sys::init();
-
- assert!(buf.len() <= c_int::MAX as usize);
- let bio = unsafe {
- {
- let r = BIO_new_mem_buf(buf.as_ptr() as *const _, buf.len() as c_int);
- if r.is_null() {
- Err(ErrorStack::get())
- } else {
- Ok(r)
- }
- }?
- };
-
- Ok(MemBioSlice(bio, PhantomData))
- }
-
- pub fn as_ptr(&self) -> *mut openssl_sys::BIO {
- self.0
- }
-}
-
impl StackableX509Crl {
pub fn stack_from_pem(pem: &[u8]) -> Result<Vec<X509Crl>, ErrorStack> {
unsafe {
openssl_sys::init();
- let bio = MemBioSlice::new(pem)?;
+ let bio = BioMemSlice::new(pem)?;
let mut crls = vec![];
loop {
diff --git a/rust/pv/src/pem_utils.rs b/rust/pv/src/pem_utils.rs
new file mode 100644
index 00000000..e6462519
--- /dev/null
+++ b/rust/pv/src/pem_utils.rs
@@ -0,0 +1,222 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright IBM Corp. 2024
+
+use crate::Result;
+use crate::{openssl_extensions::BioMem, Error};
+use openssl::error::ErrorStack;
+use pv_core::request::Confidential;
+use std::{
+ ffi::{c_char, CString},
+ fmt::Display,
+};
+
+mod ffi {
+ use openssl_sys::BIO;
+ use std::ffi::{c_char, c_int, c_long, c_uchar};
+ extern "C" {
+ pub fn PEM_write_bio(
+ bio: *mut BIO,
+ name: *const c_char,
+ header: *const c_char,
+ data: *const c_uchar,
+ len: c_long,
+ ) -> c_int;
+ }
+}
+
+/// Thin wrapper around [`CString`] only containing ASCII chars.
+#[derive(Debug)]
+struct AsciiCString(CString);
+
+impl AsciiCString {
+ /// Convert from string
+ ///
+ /// # Returns
+ /// Error if string is not ASCII or contains null chars
+ pub(crate) fn from_str(s: &str) -> Result<Self> {
+ match s.is_ascii() {
+ true => Ok(Self(CString::new(s).map_err(|_| Error::NonAscii)?)),
+ false => Err(Error::NonAscii),
+ }
+ }
+
+ fn as_ptr(&self) -> *const c_char {
+ self.0.as_ptr()
+ }
+}
+
+/// Helper struct to construct the PEM format
+#[derive(Debug)]
+struct InnerPem<'d> {
+ name: AsciiCString,
+ header: Option<AsciiCString>,
+ data: &'d [u8],
+}
+
+impl<'d> InnerPem<'d> {
+ fn new(name: &str, header: Option<String>, data: &'d [u8]) -> Result<Self> {
+ Ok(Self {
+ name: AsciiCString::from_str(name)?,
+ header: match header {
+ Some(h) => Some(AsciiCString::from_str(&h)?),
+ None => None,
+ },
+ data,
+ })
+ }
+
+ /// Generate PEM representation of the data
+ fn to_pem(&self) -> Result<Vec<u8>> {
+ let bio = BioMem::new()?;
+ let hdr_ptr = match self.header {
+ // avoid moving variable -> use reference
+ Some(ref h) => h.as_ptr(),
+ None => std::ptr::null(),
+ };
+
+ // SAFETY:
+ // All pointers point to valid C strings or memory regions
+ let rc = unsafe {
+ ffi::PEM_write_bio(
+ bio.as_ptr(),
+ self.name.as_ptr(),
+ hdr_ptr,
+ self.data.as_ptr(),
+ self.data.len() as std::ffi::c_long,
+ )
+ };
+
+ match rc {
+ 1 => Err(Error::InternalSsl("Could not write PEM", ErrorStack::get())),
+ _ => Ok(bio.to_vec()),
+ }
+ }
+}
+
+/// Data in PEM format
+///
+/// Displays into a printable PEM structure.
+/// Must be constructed from another structure in this library.
+///
+/// ```rust,ignore
+/// let pem: Pem = ...;
+/// println!("PEM {pem}");
+/// ```
+/// ```PEM
+///-----BEGIN <name>-----
+///<header>
+///
+///<Base64 formatted binary data>
+///-----END <name>-----
+
+#[derive(Debug)]
+pub struct Pem {
+ pem: Confidential<String>,
+}
+
+#[allow(unused)]
+impl Pem {
+ /// Create a new PEM structure.
+ ///
+ /// # Errors
+ ///
+ /// This function will return an error if name or header contain non-ASCII chars, or OpenSSL
+ /// could not generate the PEM (very likely due to OOM).
+ pub(crate) fn new<D, H>(name: &str, header: H, data: D) -> Result<Self>
+ where
+ D: AsRef<[u8]>,
+ H: Into<Option<String>>,
+ {
+ let mut header = header.into();
+ let header = match header {
+ Some(h) if h.ends_with('\n') => Some(h),
+ Some(h) if h.is_empty() => None,
+ Some(mut h) => {
+ h.push('\n');
+ Some(h)
+ }
+ None => None,
+ };
+
+ let inner_pem = InnerPem::new(name, header, data.as_ref())?;
+
+ // Create the PEM format eagerly so that to_string/display cannot fail because of ASCII or OpenSSL Errors
+ // Both error should be very unlikely
+ // OpenSSL should be able to create PEM if there is enough memory and produce a non-null
+ // terminated ASCII-string
+ // Unwrap succeeds it's all ASCII
+ // Std lib implements all the conversations without a copy
+ let pem = CString::new(inner_pem.to_pem()?)
+ .map_err(|_| Error::NonAscii)?
+ .into_string()
+ .unwrap()
+ .into();
+
+ Ok(Self { pem })
+ }
+
+ /// Converts the PEM-data into a byte vector.
+ ///
+ /// This consumes the `PEM`.
+ #[inline]
+ #[must_use = "`self` will be dropped if the result is not used"]
+ pub fn into_bytes(self) -> Confidential<Vec<u8>> {
+ self.pem.into_inner().into_bytes().into()
+ }
+}
+
+impl Display for Pem {
+ fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+ self.pem.value().fmt(f)
+ }
+}
+
+#[cfg(test)]
+mod test {
+ use super::*;
+
+ #[test]
+ fn no_data() {
+ const EXP: &str =
+ "-----BEGIN PEM test-----\ntest hdr value: 17\n\n-----END PEM test-----\n";
+ let test_pem = Pem::new("PEM test", "test hdr value: 17".to_string(), []).unwrap();
+ let pem_str = test_pem.to_string();
+ assert_eq!(pem_str, EXP);
+ }
+
+ #[test]
+ fn no_hdr() {
+ const EXP: &str =
+ "-----BEGIN PEM test-----\ndmVyeSBzZWNyZXQga2V5\n-----END PEM test-----\n";
+ let test_pem = Pem::new("PEM test", None, "very secret key").unwrap();
+ let pem_str = test_pem.to_string();
+ assert_eq!(pem_str, EXP);
+ }
+
+ #[test]
+ fn some_data() {
+ const EXP: &str= "-----BEGIN PEM test-----\ntest hdr value: 17\n\ndmVyeSBzZWNyZXQga2V5\n-----END PEM test-----\n";
+ let test_pem = Pem::new(
+ "PEM test",
+ "test hdr value: 17".to_string(),
+ "very secret key",
+ )
+ .unwrap();
+ let pem_str = test_pem.to_string();
+ assert_eq!(pem_str, EXP);
+ }
+
+ #[test]
+ fn data_linebreak() {
+ const EXP: &str= "-----BEGIN PEM test-----\ntest hdr value: 17\n\ndmVyeSBzZWNyZXQga2V5\n-----END PEM test-----\n";
+ let test_pem = Pem::new(
+ "PEM test",
+ "test hdr value: 17\n".to_string(),
+ "very secret key",
+ )
+ .unwrap();
+ let pem_str = test_pem.to_string();
+ assert_eq!(pem_str, EXP);
+ }
+}

View File

@ -0,0 +1,57 @@
From 69eb06f39e5134565babfe96c66a3786c0a571cf Mon Sep 17 00:00:00 2001
From: Steffen Eiden <seiden@linux.ibm.com>
Date: Tue, 20 Feb 2024 14:50:47 +0100
Subject: [PATCH] rust/pv_core: Update ffi.rs to linux/uvdevice.h v6.13
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
While at it, add a file global #[allow(dead_code)].
The file is a rustified copy of linux/arch/s390/include/uapi/asm/uvdevice.h
and there might be things that are not needed here but are defined in that header.
Acked-by: Marc Hartmayer <marc@linux.ibm.com>
Reviewed-by: Christoph Schlameuss <schlameuss@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
---
rust/pv_core/src/uvdevice/ffi.rs | 11 +++++++++--
1 file changed, 9 insertions(+), 2 deletions(-)
diff --git a/rust/pv_core/src/uvdevice/ffi.rs b/rust/pv_core/src/uvdevice/ffi.rs
index bbcc5867..3d9998db 100644
--- a/rust/pv_core/src/uvdevice/ffi.rs
+++ b/rust/pv_core/src/uvdevice/ffi.rs
@@ -2,6 +2,13 @@
//
// Copyright IBM Corp. 2023
+// This file is a rustified copy of linux/arch/s390/include/uapi/asm/uvdevice.h
+// There might be things that are not needed here but nontheless defined in that header.
+// Those two files should be in sync -> there might be unused/dead code.
+//
+// The `UVIO_IOCTL_*` and `UVIO_SUPP_*` macros
+#![allow(dead_code)]
+
use std::mem::size_of;
use crate::{assert_size, static_assert};
@@ -11,9 +18,8 @@ pub const UVIO_ATT_ARCB_MAX_LEN: usize = 0x100000;
pub const UVIO_ATT_MEASUREMENT_MAX_LEN: usize = 0x8000;
pub const UVIO_ATT_ADDITIONAL_MAX_LEN: usize = 0x8000;
pub const UVIO_ADD_SECRET_MAX_LEN: usize = 0x100000;
-#[allow(unused)]
-// here for completeness
pub const UVIO_LIST_SECRETS_LEN: usize = 0x1000;
+pub const UVIO_RETR_SECRET_MAX_LEN: usize = 0x2000;
// equal to ascii 'u'
pub const UVIO_TYPE_UVC: u8 = 117u8;
@@ -23,6 +29,7 @@ pub const UVIO_IOCTL_ATT_NR: u8 = 1;
pub const UVIO_IOCTL_ADD_SECRET_NR: u8 = 2;
pub const UVIO_IOCTL_LIST_SECRETS_NR: u8 = 3;
pub const UVIO_IOCTL_LOCK_SECRETS_NR: u8 = 4;
+pub const UVIO_IOCTL_RETR_SECRET_NR: u8 = 5;
/// Uvdevice IOCTL control block
/// Programs can use this struct to communicate with the uvdevice via IOCTLs

View File

@ -0,0 +1,204 @@
From 01cd81ecf5d1a7e1e504ae1b67692cf63cd4b51d Mon Sep 17 00:00:00 2001
From: Steffen Eiden <seiden@linux.ibm.com>
Date: Tue, 5 Mar 2024 11:56:57 +0100
Subject: [PATCH] rust/pv_core: Retrieve Secret UVC
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Create the uvdevice-IOCTL functionality for the new Retrieve Secret UVC.
Reviewed-by: Christoph Schlameuss <schlameuss@linux.ibm.com>
Acked-by: Marc Hartmayer <marc@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
---
rust/pv_core/src/error.rs | 7 ++
rust/pv_core/src/lib.rs | 2 +-
rust/pv_core/src/uvdevice/secret.rs | 97 +++++++++++++++++++++++-
rust/pv_core/src/uvdevice/secret_list.rs | 15 ++++
4 files changed, 118 insertions(+), 3 deletions(-)
diff --git a/rust/pv_core/src/error.rs b/rust/pv_core/src/error.rs
index 20fca24d..ba7b7e26 100644
--- a/rust/pv_core/src/error.rs
+++ b/rust/pv_core/src/error.rs
@@ -4,6 +4,8 @@
use std::path::PathBuf;
+use crate::uv::SecretId;
+
/// Result type for this crate
pub type Result<T, E = Error> = std::result::Result<T, E>;
@@ -70,6 +72,11 @@ pub enum Error {
#[error("The attestation request does not specify a measurement size or measurement data.")]
BinArcbNoMeasurement,
+ #[error(
+ "The secret with the ID {id} cannot be retrieved. The requested size is too large ({size})"
+ )]
+ InvalidRetrievableSecretType { id: SecretId, size: usize },
+
// errors from other crates
#[error(transparent)]
Io(#[from] std::io::Error),
diff --git a/rust/pv_core/src/lib.rs b/rust/pv_core/src/lib.rs
index 349c0b28..5922211f 100644
--- a/rust/pv_core/src/lib.rs
+++ b/rust/pv_core/src/lib.rs
@@ -32,7 +32,7 @@ pub mod misc {
/// [`crate::uv::UvCmd`]
pub mod uv {
pub use crate::uvdevice::attest::AttestationCmd;
- pub use crate::uvdevice::secret::{AddCmd, ListCmd, LockCmd};
+ pub use crate::uvdevice::secret::{AddCmd, ListCmd, LockCmd, RetrieveCmd};
pub use crate::uvdevice::secret_list::{ListableSecretType, SecretEntry, SecretId, SecretList};
pub use crate::uvdevice::{ConfigUid, UvCmd, UvDevice, UvDeviceInfo, UvFlags, UvcSuccess};
}
diff --git a/rust/pv_core/src/uvdevice/secret.rs b/rust/pv_core/src/uvdevice/secret.rs
index 6c22b6ed..263f17d5 100644
--- a/rust/pv_core/src/uvdevice/secret.rs
+++ b/rust/pv_core/src/uvdevice/secret.rs
@@ -3,8 +3,15 @@
// Copyright IBM Corp. 2023
use super::ffi;
-use crate::{request::MagicValue, uv::UvCmd, uvsecret::AddSecretMagic, Error, Result, PAGESIZE};
-use std::io::Read;
+use crate::{
+ request::{Confidential, MagicValue},
+ uv::{SecretEntry, UvCmd},
+ uvsecret::AddSecretMagic,
+ Error, Result, PAGESIZE,
+};
+use log::debug;
+use std::{io::Read, mem::size_of_val};
+use zerocopy::AsBytes;
/// _List Secrets_ Ultravisor command.
///
@@ -116,3 +123,89 @@ impl UvCmd for LockCmd {
}
}
}
+
+/// Retrieve a secret value from UV store
+#[derive(Debug)]
+pub struct RetrieveCmd {
+ entry: SecretEntry,
+ key: Confidential<Vec<u8>>,
+}
+
+impl RetrieveCmd {
+ /// Maximum size of a retrieved key (=2 pages)
+ pub const MAX_SIZE: usize = ffi::UVIO_RETR_SECRET_MAX_LEN;
+
+ /// Create a retrieve-secret UVC from a [`SecretEntry`].
+ ///
+ /// This uses the index of the secret entry for the UVC.
+ pub fn from_entry(entry: SecretEntry) -> Result<Self> {
+ entry.try_into()
+ }
+
+ /// Transform a [`RetrieveCmd`] into a key-vector.
+ ///
+ /// Only makes sense to call after a successful UVC execution.
+ pub fn into_key(self) -> Confidential<Vec<u8>> {
+ self.key
+ }
+
+ /// Get the secret entry
+ ///
+ /// Get the secret entry that is used as metadata to retrieve the secret
+ pub fn meta_data(&self) -> &SecretEntry {
+ &self.entry
+ }
+}
+
+impl TryFrom<SecretEntry> for RetrieveCmd {
+ type Error = Error;
+
+ fn try_from(entry: SecretEntry) -> Result<Self> {
+ let len = entry.secret_size() as usize;
+
+ // Next to impossible if the secret entry is a valid response from UV
+ if len > Self::MAX_SIZE {
+ return Err(Error::InvalidRetrievableSecretType {
+ id: entry.secret_id().to_owned(),
+ size: len,
+ });
+ }
+
+ // Ensure that an u16 fits into the buffer.
+ let size = std::cmp::max(size_of_val(&entry.index()), len);
+ debug!("Create a buf with {} elements", size);
+ let mut buf = vec![0; size];
+ // The IOCTL expects the secret index in the first two bytes of the buffer. They will be
+ // overwritten in the response
+ entry.index_be().write_to_prefix(&mut buf).unwrap();
+ Ok(Self {
+ entry,
+ key: buf.into(),
+ })
+ }
+}
+
+impl UvCmd for RetrieveCmd {
+ const UV_IOCTL_NR: u8 = ffi::UVIO_IOCTL_RETR_SECRET_NR;
+
+ fn rc_fmt(&self, rc: u16, _: u16) -> Option<&'static str> {
+ match rc {
+ // should not appear (TM), software creates request from a list item
+ 0x0009 => Some("the allocated buffer is to small to store the secret"),
+ // should not appear (TM), kernel allocates the memory
+ 0x0102 => {
+ Some("access exception recognized when accessing retrieved secret storage area")
+ }
+ // should not appear (TM), software creates request from a list item
+ 0x010f => Some("the Secret Store is empty"),
+ // should not appear (TM), software creates request from a list item
+ 0x0110 => Some("the Secret Store does not contain a secret with the specified index"),
+ 0x0111 => Some("the secret is not retrievable"),
+ _ => None,
+ }
+ }
+
+ fn data(&mut self) -> Option<&mut [u8]> {
+ Some(self.key.value_mut())
+ }
+}
diff --git a/rust/pv_core/src/uvdevice/secret_list.rs b/rust/pv_core/src/uvdevice/secret_list.rs
index d20928b5..0a8af504 100644
--- a/rust/pv_core/src/uvdevice/secret_list.rs
+++ b/rust/pv_core/src/uvdevice/secret_list.rs
@@ -110,6 +110,11 @@ impl SecretEntry {
self.index.get()
}
+ /// Returns the index of this [`SecretEntry`] in BE.
+ pub(crate) fn index_be(&self) -> &U16<BigEndian> {
+ &self.index
+ }
+
/// Returns the secret type of this [`SecretEntry`].
pub fn stype(&self) -> ListableSecretType {
self.stype.into()
@@ -127,6 +132,16 @@ impl SecretEntry {
pub fn id(&self) -> &[u8] {
self.id.as_ref()
}
+
+ /// Get the id as [`SecretId`] reference
+ pub(crate) fn secret_id(&self) -> &SecretId {
+ &self.id
+ }
+
+ /// Returns the secret size of this [`SecretEntry`].
+ pub fn secret_size(&self) -> u32 {
+ self.len.get()
+ }
}
impl Display for SecretEntry {

View File

@ -0,0 +1,710 @@
From 4af137f4fad8638169ccf0ddcb6dc4b0fe8fb1c1 Mon Sep 17 00:00:00 2001
From: Steffen Eiden <seiden@linux.ibm.com>
Date: Tue, 5 Mar 2024 12:16:44 +0100
Subject: [PATCH] rust/pv_core: Support for listing Retrievable Secrets
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Add support for listing retrievable secrets in the List Secrets UVC.
Acked-by: Marc Hartmayer <marc@linux.ibm.com>
Reviewed-by: Christoph Schlameuss <schlameuss@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
---
rust/pv_core/src/lib.rs | 2 +
rust/pv_core/src/uvdevice.rs | 1 +
rust/pv_core/src/uvdevice/retr_secret.rs | 399 +++++++++++++++++++++++
rust/pv_core/src/uvdevice/secret_list.rs | 157 +++++++--
4 files changed, 536 insertions(+), 23 deletions(-)
create mode 100644 rust/pv_core/src/uvdevice/retr_secret.rs
diff --git a/rust/pv_core/src/lib.rs b/rust/pv_core/src/lib.rs
index 5922211f..caebfcea 100644
--- a/rust/pv_core/src/lib.rs
+++ b/rust/pv_core/src/lib.rs
@@ -32,6 +32,8 @@ pub mod misc {
/// [`crate::uv::UvCmd`]
pub mod uv {
pub use crate::uvdevice::attest::AttestationCmd;
+ pub use crate::uvdevice::retr_secret::RetrievableSecret;
+ pub use crate::uvdevice::retr_secret::{AesSizes, AesXtsSizes, EcCurves, HmacShaSizes};
pub use crate::uvdevice::secret::{AddCmd, ListCmd, LockCmd, RetrieveCmd};
pub use crate::uvdevice::secret_list::{ListableSecretType, SecretEntry, SecretId, SecretList};
pub use crate::uvdevice::{ConfigUid, UvCmd, UvDevice, UvDeviceInfo, UvFlags, UvcSuccess};
diff --git a/rust/pv_core/src/uvdevice.rs b/rust/pv_core/src/uvdevice.rs
index d4176815..e9848243 100644
--- a/rust/pv_core/src/uvdevice.rs
+++ b/rust/pv_core/src/uvdevice.rs
@@ -25,6 +25,7 @@ mod info;
mod test;
pub(crate) use ffi::uv_ioctl;
pub mod attest;
+pub mod retr_secret;
pub mod secret;
pub mod secret_list;
diff --git a/rust/pv_core/src/uvdevice/retr_secret.rs b/rust/pv_core/src/uvdevice/retr_secret.rs
new file mode 100644
index 00000000..490152b4
--- /dev/null
+++ b/rust/pv_core/src/uvdevice/retr_secret.rs
@@ -0,0 +1,399 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright IBM Corp. 2024
+
+use crate::uv::{ListableSecretType, RetrieveCmd};
+use serde::{Deserialize, Serialize, Serializer};
+use std::fmt::Display;
+
+/// Allowed sizes for AES keys
+#[non_exhaustive]
+#[derive(PartialEq, Eq, Debug)]
+pub enum AesSizes {
+ /// 128 bit key
+ Bits128,
+ /// 192 bit key
+ Bits192,
+ /// 256 bit key
+ Bits256,
+}
+
+impl AesSizes {
+ /// Construct the key-size from the bit-size.
+ ///
+ /// Returns [`None`] if the bit-size is not supported.
+ pub fn from_bits(bits: u32) -> Option<Self> {
+ match bits {
+ 128 => Some(Self::Bits128),
+ 192 => Some(Self::Bits192),
+ 256 => Some(Self::Bits256),
+ _ => None,
+ }
+ }
+
+ /// Returns the bit-size for the key-type
+ const fn bit_size(&self) -> u32 {
+ match self {
+ Self::Bits128 => 128,
+ Self::Bits192 => 192,
+ Self::Bits256 => 256,
+ }
+ }
+}
+
+impl Display for AesSizes {
+ fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+ write!(f, "{}", self.bit_size())
+ }
+}
+
+/// Allowed sizes for AES-XTS keys
+#[non_exhaustive]
+#[derive(PartialEq, Eq, Debug)]
+pub enum AesXtsSizes {
+ /// Two AES 128 bit keys
+ Bits128,
+ /// Two AES 256 bit keys
+ Bits256,
+}
+
+impl AesXtsSizes {
+ /// Construct the key-size from the bit-size.
+ ///
+ /// It's a key containing two keys; bit-size is half the number of bits it has
+ /// Returns [`None`] if the bit-size is not supported.
+ pub fn from_bits(bits: u32) -> Option<Self> {
+ match bits {
+ 128 => Some(Self::Bits128),
+ 256 => Some(Self::Bits256),
+ _ => None,
+ }
+ }
+
+ /// Returns the bit-size for the key-type
+ ///
+ /// It's a key containing two keys: bit-size is half the number of bits it has
+ const fn bit_size(&self) -> u32 {
+ match self {
+ Self::Bits128 => 128,
+ Self::Bits256 => 256,
+ }
+ }
+}
+
+impl Display for AesXtsSizes {
+ fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+ write!(f, "{}", self.bit_size())
+ }
+}
+
+/// Allowed sizes for HMAC-SHA keys
+#[non_exhaustive]
+#[derive(PartialEq, Eq, Debug)]
+pub enum HmacShaSizes {
+ /// SHA 256 bit
+ Sha256,
+ /// SHA 512 bit
+ Sha512,
+}
+
+impl HmacShaSizes {
+ /// Construct the key-size from the sha-size.
+ ///
+ /// FW expects maximum resistance keys (double the SHA size).
+ /// The `sha_size` is half of the number of bits in the key
+ /// Returns [`None`] if the `sha_size` is not supported.
+ pub fn from_sha_size(sha_size: u32) -> Option<Self> {
+ match sha_size {
+ 256 => Some(Self::Sha256),
+ 512 => Some(Self::Sha512),
+ _ => None,
+ }
+ }
+
+ /// Returns the sha-size for the key-type
+ ///
+ /// FW expects maximum resistance keys (double the SHA size).
+ /// The `sha_size` is half of the number of bits in the key
+ const fn sha_size(&self) -> u32 {
+ match self {
+ Self::Sha256 => 256,
+ Self::Sha512 => 512,
+ }
+ }
+}
+
+impl Display for HmacShaSizes {
+ fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+ write!(f, "{}", self.sha_size())
+ }
+}
+
+/// Allowed curves for EC private keys
+#[non_exhaustive]
+#[derive(PartialEq, Eq, Debug)]
+pub enum EcCurves {
+ /// secp256r1 or prime256v1 curve
+ Secp256R1,
+ /// secp384p1 curve
+ Secp384R1,
+ /// secp521r1 curve
+ Secp521R1,
+ /// ed25519 curve
+ Ed25519,
+ /// ed448 curve
+ Ed448,
+}
+
+impl EcCurves {
+ const fn exp_size(&self) -> usize {
+ match self {
+ Self::Secp256R1 => 32,
+ Self::Secp384R1 => 48,
+ Self::Secp521R1 => 80,
+ Self::Ed25519 => 32,
+ Self::Ed448 => 64,
+ }
+ }
+
+ /// Resizes the raw key to the expected size.
+ ///
+ /// See [`Vec::resize`]
+ pub fn resize_raw_key(&self, mut raw: Vec<u8>) -> Vec<u8> {
+ raw.resize(self.exp_size(), 0);
+ raw
+ }
+}
+
+// The names have to stay constant, otherwise the PEM contains invalid types
+impl Display for EcCurves {
+ fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+ match self {
+ Self::Secp256R1 => write!(f, "SECP256R1"),
+ Self::Secp384R1 => write!(f, "SECP384R1"),
+ Self::Secp521R1 => write!(f, "SECP521R1"),
+ Self::Ed25519 => write!(f, "ED25519"),
+ Self::Ed448 => write!(f, "ED448"),
+ }
+ }
+}
+
+/// Retrievable Secret types
+#[non_exhaustive]
+#[derive(PartialEq, Eq, Debug)]
+pub enum RetrievableSecret {
+ /// Plain-text secret
+ PlainText,
+ /// Protected AES key
+ Aes(AesSizes),
+ /// Protected AES-XTS key
+ AesXts(AesXtsSizes),
+ /// Protected HMAC-SHA key
+ HmacSha(HmacShaSizes),
+ /// Protected EC-private key
+ Ec(EcCurves),
+}
+
+// The names have to stay constant, otherwise the PEM contains invalid/unknown types
+impl Display for RetrievableSecret {
+ fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+ // Alternate representation: Omit sizes/curves
+ if f.alternate() {
+ match self {
+ Self::PlainText => write!(f, "PLAINTEXT"),
+ Self::Aes(_) => write!(f, "AES-KEY"),
+ Self::AesXts(_) => write!(f, "AES-XTS-KEY"),
+ Self::HmacSha(_) => write!(f, "HMAC-SHA-KEY"),
+ Self::Ec(_) => write!(f, "EC-PRIVATE-KEY"),
+ }
+ } else {
+ match self {
+ Self::PlainText => write!(f, "PLAINTEXT"),
+ Self::Aes(s) => write!(f, "AES-{s}-KEY"),
+ Self::AesXts(s) => write!(f, "AES-XTS-{s}-KEY"),
+ Self::HmacSha(s) => write!(f, "HMAC-SHA-{s}-KEY"),
+ Self::Ec(c) => write!(f, "EC-{c}-PRIVATE-KEY"),
+ }
+ }
+ }
+}
+
+impl RetrievableSecret {
+ /// Report expected input types
+ pub fn expected(&self) -> String {
+ match self {
+ Self::PlainText => format!("less than {}", RetrieveCmd::MAX_SIZE),
+ Self::Aes(_) => "128, 192, or 256".to_string(),
+ Self::AesXts(_) => "128 or 256".to_string(),
+ Self::HmacSha(_) => "256 or 512".to_string(),
+ Self::Ec(_) => "secp256r1, secp384r1, secp521r1, ed25519, or ed448".to_string(),
+ }
+ }
+}
+
+impl From<&RetrievableSecret> for u16 {
+ fn from(value: &RetrievableSecret) -> Self {
+ match value {
+ RetrievableSecret::PlainText => ListableSecretType::PLAINTEXT,
+ RetrievableSecret::Aes(AesSizes::Bits128) => ListableSecretType::AES_128_KEY,
+ RetrievableSecret::Aes(AesSizes::Bits192) => ListableSecretType::AES_192_KEY,
+ RetrievableSecret::Aes(AesSizes::Bits256) => ListableSecretType::AES_256_KEY,
+ RetrievableSecret::AesXts(AesXtsSizes::Bits128) => ListableSecretType::AES_128_XTS_KEY,
+ RetrievableSecret::AesXts(AesXtsSizes::Bits256) => ListableSecretType::AES_256_XTS_KEY,
+ RetrievableSecret::HmacSha(HmacShaSizes::Sha256) => {
+ ListableSecretType::HMAC_SHA_256_KEY
+ }
+ RetrievableSecret::HmacSha(HmacShaSizes::Sha512) => {
+ ListableSecretType::HMAC_SHA_512_KEY
+ }
+ RetrievableSecret::Ec(EcCurves::Secp256R1) => ListableSecretType::ECDSA_P256_KEY,
+ RetrievableSecret::Ec(EcCurves::Secp384R1) => ListableSecretType::ECDSA_P384_KEY,
+ RetrievableSecret::Ec(EcCurves::Secp521R1) => ListableSecretType::ECDSA_P521_KEY,
+ RetrievableSecret::Ec(EcCurves::Ed25519) => ListableSecretType::ECDSA_ED25519_KEY,
+ RetrievableSecret::Ec(EcCurves::Ed448) => ListableSecretType::ECDSA_ED448_KEY,
+ }
+ }
+}
+
+// serializes to: <secret type nb> (String name)
+impl Serialize for RetrievableSecret {
+ fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error>
+ where
+ S: Serializer,
+ {
+ let id: u16 = self.into();
+ serializer.serialize_str(&format!("{id} ({self})"))
+ }
+}
+
+/// deserializes from the secret type nb only
+impl<'de> Deserialize<'de> for RetrievableSecret {
+ fn deserialize<D>(de: D) -> Result<Self, D::Error>
+ where
+ D: serde::Deserializer<'de>,
+ {
+ struct RetrSecretVisitor;
+ impl<'de> serde::de::Visitor<'de> for RetrSecretVisitor {
+ type Value = RetrievableSecret;
+
+ fn expecting(&self, fmt: &mut std::fmt::Formatter) -> std::fmt::Result {
+ fmt.write_str(
+ "a retrievable secret type: `<number> (String name)` number in [3,10]|[17,21]",
+ )
+ }
+ fn visit_str<E>(self, s: &str) -> Result<Self::Value, E>
+ where
+ E: serde::de::Error,
+ {
+ let (n, _) = s.split_once(' ').ok_or(serde::de::Error::invalid_value(
+ serde::de::Unexpected::Str(s),
+ &self,
+ ))?;
+ let id: u16 = n.parse().map_err(|_| {
+ serde::de::Error::invalid_value(serde::de::Unexpected::Str(n), &self)
+ })?;
+ let listable: ListableSecretType = id.into();
+ match listable {
+ ListableSecretType::Retrievable(r) => Ok(r),
+ _ => Err(serde::de::Error::invalid_value(
+ serde::de::Unexpected::Unsigned(id.into()),
+ &self,
+ )),
+ }
+ }
+ }
+ de.deserialize_str(RetrSecretVisitor)
+ }
+}
+
+#[cfg(test)]
+mod test {
+ use serde_test::{assert_tokens, Token};
+
+ use super::*;
+
+ #[test]
+ fn retr_serde_plain() {
+ let retr = RetrievableSecret::PlainText;
+ assert_tokens(&retr, &[Token::Str("3 (PLAINTEXT)")]);
+ }
+
+ #[test]
+ fn retr_serde_aes() {
+ let retr = RetrievableSecret::Aes(AesSizes::Bits192);
+ assert_tokens(&retr, &[Token::Str("5 (AES-192-KEY)")]);
+ }
+
+ #[test]
+ fn retr_serde_aes_xts() {
+ let retr = RetrievableSecret::AesXts(AesXtsSizes::Bits128);
+ assert_tokens(&retr, &[Token::Str("7 (AES-XTS-128-KEY)")]);
+ }
+
+ #[test]
+ fn retr_serde_hmac() {
+ let retr = RetrievableSecret::HmacSha(HmacShaSizes::Sha256);
+ assert_tokens(&retr, &[Token::Str("9 (HMAC-SHA-256-KEY)")]);
+ }
+
+ #[test]
+ fn retr_serde_es() {
+ let retr = RetrievableSecret::Ec(EcCurves::Secp521R1);
+ assert_tokens(&retr, &[Token::Str("19 (EC-SECP521R1-PRIVATE-KEY)")]);
+ }
+
+ // Ensure that the string representation of the retrievable types stay constant, or PEM will have
+ // different, incompatible types
+ #[test]
+ fn stable_type_names() {
+ assert_eq!("PLAINTEXT", RetrievableSecret::PlainText.to_string());
+ assert_eq!(
+ "AES-128-KEY",
+ RetrievableSecret::Aes(AesSizes::Bits128).to_string()
+ );
+ assert_eq!(
+ "AES-192-KEY",
+ RetrievableSecret::Aes(AesSizes::Bits192).to_string()
+ );
+ assert_eq!(
+ "AES-256-KEY",
+ RetrievableSecret::Aes(AesSizes::Bits256).to_string()
+ );
+ assert_eq!(
+ "AES-XTS-128-KEY",
+ RetrievableSecret::AesXts(AesXtsSizes::Bits128).to_string()
+ );
+ assert_eq!(
+ "AES-XTS-256-KEY",
+ RetrievableSecret::AesXts(AesXtsSizes::Bits256).to_string()
+ );
+ assert_eq!(
+ "HMAC-SHA-256-KEY",
+ RetrievableSecret::HmacSha(HmacShaSizes::Sha256).to_string()
+ );
+ assert_eq!(
+ "HMAC-SHA-512-KEY",
+ RetrievableSecret::HmacSha(HmacShaSizes::Sha512).to_string()
+ );
+ assert_eq!(
+ "EC-SECP256R1-PRIVATE-KEY",
+ RetrievableSecret::Ec(EcCurves::Secp256R1).to_string()
+ );
+ assert_eq!(
+ "EC-SECP384R1-PRIVATE-KEY",
+ RetrievableSecret::Ec(EcCurves::Secp384R1).to_string()
+ );
+ assert_eq!(
+ "EC-SECP521R1-PRIVATE-KEY",
+ RetrievableSecret::Ec(EcCurves::Secp521R1).to_string()
+ );
+ assert_eq!(
+ "EC-ED25519-PRIVATE-KEY",
+ RetrievableSecret::Ec(EcCurves::Ed25519).to_string()
+ );
+ assert_eq!(
+ "EC-ED448-PRIVATE-KEY",
+ RetrievableSecret::Ec(EcCurves::Ed448).to_string()
+ );
+ }
+}
diff --git a/rust/pv_core/src/uvdevice/secret_list.rs b/rust/pv_core/src/uvdevice/secret_list.rs
index 0a8af504..4e955010 100644
--- a/rust/pv_core/src/uvdevice/secret_list.rs
+++ b/rust/pv_core/src/uvdevice/secret_list.rs
@@ -2,9 +2,14 @@
//
// Copyright IBM Corp. 2024
-use crate::assert_size;
-use crate::{misc::to_u16, uv::ListCmd, uvdevice::UvCmd, Error, Result};
-use byteorder::{BigEndian, ReadBytesExt, WriteBytesExt};
+use crate::{
+ assert_size,
+ misc::to_u16,
+ uv::{AesSizes, AesXtsSizes, EcCurves, HmacShaSizes, ListCmd, RetrievableSecret},
+ uvdevice::UvCmd,
+ Error, Result,
+};
+use byteorder::{BigEndian, ByteOrder, ReadBytesExt, WriteBytesExt};
use serde::{Deserialize, Serialize, Serializer};
use std::{
fmt::Display,
@@ -18,7 +23,7 @@ use zerocopy::{AsBytes, FromBytes, FromZeroes, U16, U32};
///
/// (de)serializes itself in/from a hex-string
#[repr(C)]
-#[derive(PartialEq, Eq, AsBytes, FromZeroes, FromBytes, Debug, Clone)]
+#[derive(PartialEq, Eq, AsBytes, FromZeroes, FromBytes, Debug, Clone, Default)]
pub struct SecretId([u8; Self::ID_SIZE]);
assert_size!(SecretId, SecretId::ID_SIZE);
@@ -94,11 +99,11 @@ impl SecretEntry {
/// Create a new entry for a [`SecretList`].
///
/// The content of this entry will very likely not represent the status of the guest in the
- /// Ultravisor. Use of [`SecretList::decode`] in any non-test environments is encuraged.
+ /// Ultravisor. Use of [`SecretList::decode`] in any non-test environments is encouraged.
pub fn new(index: u16, stype: ListableSecretType, id: SecretId, secret_len: u32) -> Self {
Self {
index: index.into(),
- stype: stype.into(),
+ stype: U16::new(stype.into()),
len: secret_len.into(),
res_8: 0,
id,
@@ -117,7 +122,7 @@ impl SecretEntry {
/// Returns the secret type of this [`SecretEntry`].
pub fn stype(&self) -> ListableSecretType {
- self.stype.into()
+ self.stype.get().into()
}
/// Returns a reference to the id of this [`SecretEntry`].
@@ -146,7 +151,7 @@ impl SecretEntry {
impl Display for SecretEntry {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
- let stype: ListableSecretType = self.stype.into();
+ let stype: ListableSecretType = self.stype.get().into();
writeln!(f, "{} {}:", self.index, stype)?;
write!(f, " ")?;
for b in self.id.as_ref() {
@@ -298,51 +303,115 @@ fn ser_u16<S: Serializer>(v: &U16<BigEndian>, ser: S) -> Result<S::Ok, S::Error>
pub enum ListableSecretType {
/// Association Secret
Association,
+ /// Retrievable key
+ Retrievable(RetrievableSecret),
+
/// Invalid secret type, that should never appear in a list
///
/// 0 is reserved
- /// 1 is Null secret, with no id and not listable
+ /// 1 is Null secret, with no id and not list-able
Invalid(u16),
/// Unknown secret type
Unknown(u16),
}
impl ListableSecretType {
- /// UV type id for an association secret
- pub const ASSOCIATION: u16 = 0x0002;
- /// UV type id for a null secret
- pub const NULL: u16 = 0x0001;
const RESERVED_0: u16 = 0x0000;
+ /// UV secret-type id for a null secret
+ pub const NULL: u16 = 0x0001;
+ /// UV secret-type id for an association secret
+ pub const ASSOCIATION: u16 = 0x0002;
+ /// UV secret-type id for a plain text secret
+ pub const PLAINTEXT: u16 = 0x0003;
+ /// UV secret-type id for an aes-128-key secret
+ pub const AES_128_KEY: u16 = 0x0004;
+ /// UV secret-type id for an aes-192-key secret
+ pub const AES_192_KEY: u16 = 0x0005;
+ /// UV secret-type id for an aes-256-key secret
+ pub const AES_256_KEY: u16 = 0x0006;
+ /// UV secret-type id for an aes-xts-128-key secret
+ pub const AES_128_XTS_KEY: u16 = 0x0007;
+ /// UV secret-type id for an aes-xts-256-key secret
+ pub const AES_256_XTS_KEY: u16 = 0x0008;
+ /// UV secret-type id for an hmac-sha-256-key secret
+ pub const HMAC_SHA_256_KEY: u16 = 0x0009;
+ /// UV secret-type id for an hmac-sha-512-key secret
+ pub const HMAC_SHA_512_KEY: u16 = 0x000a;
+ // 0x000b - 0x0010 reserved
+ /// UV secret-type id for an ecdsa-p256-private-key secret
+ pub const ECDSA_P256_KEY: u16 = 0x0011;
+ /// UV secret-type id for an ecdsa-p384-private-key secret
+ pub const ECDSA_P384_KEY: u16 = 0x0012;
+ /// UV secret-type id for an ecdsa-p521-private-key secret
+ pub const ECDSA_P521_KEY: u16 = 0x0013;
+ /// UV secret-type id for an ed25519-private-key secret
+ pub const ECDSA_ED25519_KEY: u16 = 0x0014;
+ /// UV secret-type id for an ed448-private-key secret
+ pub const ECDSA_ED448_KEY: u16 = 0x0015;
}
impl Display for ListableSecretType {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
match self {
Self::Association => write!(f, "Association"),
- Self::Invalid(n) => write!(f, "Invalid({n})"),
- Self::Unknown(n) => write!(f, "Unknown({n})"),
+ Self::Invalid(n) => write!(f, "Invalid(0x{n:04x})"),
+ Self::Unknown(n) => write!(f, "Unknown(0x{n:04x})"),
+ Self::Retrievable(r) => write!(f, "{r}"),
}
}
}
-impl From<U16<BigEndian>> for ListableSecretType {
- fn from(value: U16<BigEndian>) -> Self {
- match value.get() {
+impl<O: ByteOrder> From<U16<O>> for ListableSecretType {
+ fn from(value: U16<O>) -> Self {
+ value.get().into()
+ }
+}
+
+impl From<u16> for ListableSecretType {
+ fn from(value: u16) -> Self {
+ match value {
Self::RESERVED_0 => Self::Invalid(Self::RESERVED_0),
Self::NULL => Self::Invalid(Self::NULL),
Self::ASSOCIATION => Self::Association,
+ Self::PLAINTEXT => Self::Retrievable(RetrievableSecret::PlainText),
+ Self::AES_128_KEY => Self::Retrievable(RetrievableSecret::Aes(AesSizes::Bits128)),
+ Self::AES_192_KEY => Self::Retrievable(RetrievableSecret::Aes(AesSizes::Bits192)),
+ Self::AES_256_KEY => Self::Retrievable(RetrievableSecret::Aes(AesSizes::Bits256)),
+ Self::AES_128_XTS_KEY => {
+ Self::Retrievable(RetrievableSecret::AesXts(AesXtsSizes::Bits128))
+ }
+ Self::AES_256_XTS_KEY => {
+ Self::Retrievable(RetrievableSecret::AesXts(AesXtsSizes::Bits256))
+ }
+ Self::HMAC_SHA_256_KEY => {
+ Self::Retrievable(RetrievableSecret::HmacSha(HmacShaSizes::Sha256))
+ }
+ Self::HMAC_SHA_512_KEY => {
+ Self::Retrievable(RetrievableSecret::HmacSha(HmacShaSizes::Sha512))
+ }
+ Self::ECDSA_P256_KEY => Self::Retrievable(RetrievableSecret::Ec(EcCurves::Secp256R1)),
+ Self::ECDSA_P384_KEY => Self::Retrievable(RetrievableSecret::Ec(EcCurves::Secp384R1)),
+ Self::ECDSA_P521_KEY => Self::Retrievable(RetrievableSecret::Ec(EcCurves::Secp521R1)),
+ Self::ECDSA_ED25519_KEY => Self::Retrievable(RetrievableSecret::Ec(EcCurves::Ed25519)),
+ Self::ECDSA_ED448_KEY => Self::Retrievable(RetrievableSecret::Ec(EcCurves::Ed448)),
n => Self::Unknown(n),
}
}
}
-impl From<ListableSecretType> for U16<BigEndian> {
+impl<O: ByteOrder> From<ListableSecretType> for U16<O> {
+ fn from(value: ListableSecretType) -> Self {
+ Self::new(value.into())
+ }
+}
+
+impl From<ListableSecretType> for u16 {
fn from(value: ListableSecretType) -> Self {
match value {
ListableSecretType::Association => ListableSecretType::ASSOCIATION,
ListableSecretType::Invalid(n) | ListableSecretType::Unknown(n) => n,
+ ListableSecretType::Retrievable(r) => (&r).into(),
}
- .into()
}
}
@@ -363,8 +432,8 @@ where
where
E: serde::de::Error,
{
- if s.len() != SecretId::ID_SIZE * 2 + 2 {
- return Err(serde::de::Error::invalid_length(s.len(), &self));
+ if s.len() != SecretId::ID_SIZE * 2 + "0x".len() {
+ return Err(serde::de::Error::invalid_length(s.len() - 2, &self));
}
let nb = s.strip_prefix("0x").ok_or_else(|| {
serde::de::Error::invalid_value(serde::de::Unexpected::Str(s), &self)
@@ -385,7 +454,6 @@ mod test {
use super::*;
use std::io::{BufReader, BufWriter, Cursor};
-
#[test]
fn dump_secret_entry() {
const EXP: &[u8] = &[
@@ -516,4 +584,47 @@ mod test {
)],
)
}
+
+ #[test]
+ fn secret_list_ser() {
+ let list = SecretList {
+ total_num_secrets: 0x112,
+ secrets: vec![SecretEntry {
+ index: 1.into(),
+ stype: 2.into(),
+ len: 32.into(),
+ res_8: 0,
+ id: SecretId::from([0; 32]),
+ }],
+ };
+
+ assert_ser_tokens(
+ &list,
+ &[
+ Token::Struct {
+ name: "SecretList",
+ len: 2,
+ },
+ Token::String("total_num_secrets"),
+ Token::U64(0x112),
+ Token::String("secrets"),
+ Token::Seq { len: Some(1) },
+ Token::Struct {
+ name: "SecretEntry",
+ len: (4),
+ },
+ Token::String("index"),
+ Token::U16(1),
+ Token::String("stype"),
+ Token::U16(2),
+ Token::String("len"),
+ Token::U32(32),
+ Token::String("id"),
+ Token::String("0x0000000000000000000000000000000000000000000000000000000000000000"),
+ Token::StructEnd,
+ Token::SeqEnd,
+ Token::StructEnd,
+ ],
+ )
+ }
}

View File

@ -0,0 +1,877 @@
From fd024387d710887bd2016658c44d4762a08c791c Mon Sep 17 00:00:00 2001
From: Steffen Eiden <seiden@linux.ibm.com>
Date: Tue, 5 Mar 2024 12:19:22 +0100
Subject: [PATCH] rust/pv: Retrievable secrets support
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Support retrievable secret for Add-Secret requests.
Acked-by: Marc Hartmayer <marc@linux.ibm.com>
Reviewed-by: Christoph Schlameuss <schlameuss@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
---
rust/pv/src/crypto.rs | 3 +-
rust/pv/src/error.rs | 8 +
rust/pv/src/lib.rs | 8 +-
rust/pv/src/uvsecret.rs | 1 +
rust/pv/src/uvsecret/guest_secret.rs | 399 +++++++++++++++++++++++++--
rust/pv/src/uvsecret/retr_secret.rs | 234 ++++++++++++++++
6 files changed, 631 insertions(+), 22 deletions(-)
create mode 100644 rust/pv/src/uvsecret/retr_secret.rs
diff --git a/rust/pv/src/crypto.rs b/rust/pv/src/crypto.rs
index 8f11d2b4..ebc85f72 100644
--- a/rust/pv/src/crypto.rs
+++ b/rust/pv/src/crypto.rs
@@ -29,7 +29,6 @@ pub type Aes256XtsKey = Confidential<[u8; SymKeyType::AES_256_XTS_KEY_LEN]>;
/// SHA-512 digest length (in bytes)
pub const SHA_512_HASH_LEN: usize = 64;
-
#[allow(dead_code)]
pub(crate) const SHA_256_HASH_LEN: u32 = 32;
#[allow(dead_code)]
@@ -60,6 +59,8 @@ impl SymKeyType {
pub const AES_256_XTS_KEY_LEN: usize = 64;
/// AES256-XTS tweak length (in bytes)
pub const AES_256_XTS_TWEAK_LEN: usize = 16;
+ /// AES256 GCM Block length
+ pub const AES_256_GCM_BLOCK_LEN: usize = 16;
/// Returns the tag length of the [`SymKeyType`] if it is an AEAD key
pub const fn tag_len(&self) -> Option<usize> {
diff --git a/rust/pv/src/error.rs b/rust/pv/src/error.rs
index 3ba808f2..601b40f0 100644
--- a/rust/pv/src/error.rs
+++ b/rust/pv/src/error.rs
@@ -109,6 +109,14 @@ pub enum Error {
#[error("An ASCII string was expected, but non-ASCII characters were received.")]
NonAscii,
+ #[error("Incorrect {what} for a {kind}. Is: {value}; expected: {exp}")]
+ RetrInvKey {
+ what: &'static str,
+ kind: String,
+ value: String,
+ exp: String,
+ },
+
// errors from other crates
#[error(transparent)]
PvCore(#[from] pv_core::Error),
diff --git a/rust/pv/src/lib.rs b/rust/pv/src/lib.rs
index ec31b9a4..43375669 100644
--- a/rust/pv/src/lib.rs
+++ b/rust/pv/src/lib.rs
@@ -104,7 +104,12 @@ pub mod request {
/// Reexports some useful OpenSSL symbols
pub mod openssl {
- pub use openssl::{error::ErrorStack, hash::DigestBytes, pkey, x509};
+ pub use openssl::{error::ErrorStack, hash::DigestBytes, nid::Nid, pkey, x509};
+ // rust-OpenSSL does not define these NIDs
+ #[allow(missing_docs)]
+ pub const NID_ED25519: Nid = Nid::from_raw(openssl_sys::NID_ED25519);
+ #[allow(missing_docs)]
+ pub const NID_ED448: Nid = Nid::from_raw(openssl_sys::NID_ED448);
}
pub use pv_core::request::*;
@@ -118,6 +123,7 @@ pub mod secret {
asrcb::{AddSecretFlags, AddSecretRequest, AddSecretVersion},
ext_secret::ExtSecret,
guest_secret::GuestSecret,
+ retr_secret::{IbmProtectedKey, RetrievedSecret},
user_data::verify_asrcb_and_get_user_data,
};
}
diff --git a/rust/pv/src/uvsecret.rs b/rust/pv/src/uvsecret.rs
index 343e4b05..c3b43bba 100644
--- a/rust/pv/src/uvsecret.rs
+++ b/rust/pv/src/uvsecret.rs
@@ -10,4 +10,5 @@
pub mod asrcb;
pub mod ext_secret;
pub mod guest_secret;
+pub mod retr_secret;
pub mod user_data;
diff --git a/rust/pv/src/uvsecret/guest_secret.rs b/rust/pv/src/uvsecret/guest_secret.rs
index 4f1db31c..3bad6d3c 100644
--- a/rust/pv/src/uvsecret/guest_secret.rs
+++ b/rust/pv/src/uvsecret/guest_secret.rs
@@ -4,20 +4,34 @@
#[allow(unused_imports)] // used for more convenient docstring
use super::asrcb::AddSecretRequest;
-use crate::assert_size;
use crate::{
- crypto::{hash, random_array},
- request::Confidential,
- Result,
+ assert_size,
+ crypto::{hash, random_array, SymKeyType},
+ request::{
+ openssl::{NID_ED25519, NID_ED448},
+ Confidential,
+ },
+ uv::{
+ AesSizes, AesXtsSizes, EcCurves, HmacShaSizes, ListableSecretType, RetrievableSecret,
+ RetrieveCmd, SecretId,
+ },
+ Error, Result,
};
use byteorder::BigEndian;
-use openssl::hash::MessageDigest;
-use pv_core::uv::{ListableSecretType, SecretId};
+use openssl::{
+ hash::MessageDigest,
+ nid::Nid,
+ pkey::{Id, PKey, Private},
+};
+use pv_core::static_assert;
use serde::{Deserialize, Serialize};
-use std::{convert::TryInto, fmt::Display};
+use std::fmt::Display;
use zerocopy::{AsBytes, U16, U32};
const ASSOC_SECRET_SIZE: usize = 32;
+/// Maximum size of a plain-text secret payload (8190)
+pub(crate) const MAX_SIZE_PLAIN_PAYLOAD: usize = RetrieveCmd::MAX_SIZE - 2;
+static_assert!(MAX_SIZE_PLAIN_PAYLOAD == 8190);
/// A Secret to be added in [`AddSecretRequest`]
#[derive(Debug, Serialize, Deserialize, PartialEq, Eq)]
@@ -36,13 +50,60 @@ pub enum GuestSecret {
#[serde(skip)]
secret: Confidential<[u8; ASSOC_SECRET_SIZE]>,
},
+ /// Retrievable key
+ ///
+ /// Create Retrievables using [`GuestSecret::retrievable`]
+ /// Secret size is always valid for the type/kind
+ Retrievable {
+ /// Retrievable secret type
+ kind: RetrievableSecret,
+ /// Name of the secret
+ name: String,
+ /// SHA256 hash of [`GuestSecret::RetrievableKey::name`]
+ id: SecretId,
+ /// Confidential actual retrievable secret (32 bytes)
+ #[serde(skip)]
+ secret: Confidential<Vec<u8>>,
+ },
+}
+
+macro_rules! retr_constructor {
+ ($(#[$err:meta])* | $(#[$kind:meta])* => $type: ty, $func: ident) => {
+ /// Create a new
+ $(#[$kind])*
+ /// [`GuestSecret::Retrievable`] secret.
+ ///
+ /// * `name` - Name of the secret. Will be hashed into a 32 byte id
+ /// * `secret` - the secret value
+ ///
+ /// # Errors
+ ///
+ $(#[$err])*
+ pub fn $func(name: &str, secret: $type) -> Result<Self> {
+ let (kind, secret) = $func(secret)?;
+ Ok(Self::Retrievable {
+ kind,
+ name: name.to_string(),
+ id: Self::name_to_id(name)?,
+ secret,
+ })
+ }
+ };
}
impl GuestSecret {
+ fn name_to_id(name: &str) -> Result<SecretId> {
+ let id: [u8; SecretId::ID_SIZE] = hash(MessageDigest::sha256(), name.as_bytes())?
+ .to_vec()
+ .try_into()
+ .unwrap();
+ Ok(id.into())
+ }
+
/// Create a new [`GuestSecret::Association`].
///
/// * `name` - Name of the secret. Will be hashed into a 32 byte id
- /// * `secret` - Value of the secret. Ranom if [`Option::None`]
+ /// * `secret` - Value of the secret. Random if [`Option::None`]
///
/// # Errors
///
@@ -51,10 +112,6 @@ impl GuestSecret {
where
O: Into<Option<[u8; ASSOC_SECRET_SIZE]>>,
{
- let id: [u8; SecretId::ID_SIZE] = hash(MessageDigest::sha256(), name.as_bytes())?
- .to_vec()
- .try_into()
- .unwrap();
let secret = match secret.into() {
Some(s) => s,
None => random_array()?,
@@ -62,16 +119,28 @@ impl GuestSecret {
Ok(Self::Association {
name: name.to_string(),
- id: id.into(),
+ id: Self::name_to_id(name)?,
secret: secret.into(),
})
}
+ retr_constructor!(#[doc = r"This function will return an error if the secret is larger than 8 pages"]
+ | #[doc = r"plaintext"] => Confidential<Vec<u8>>, plaintext);
+ retr_constructor!(#[doc = r"This function will return an error if OpenSSL cannot create a hash or the secret size is invalid"]
+ | #[doc = r"AES Key"] => Confidential<Vec<u8>>, aes);
+ retr_constructor!(#[doc = r"This function will return an error if OpenSSL cannot create a hash or the secret size is invalid"]
+ | #[doc = r"AES-XTS Key"] => Confidential<Vec<u8>>, aes_xts);
+ retr_constructor!(#[doc = r"This function will return an error if OpenSSL cannot create a hash or the secret size is invalid"]
+ | #[doc = r"HMAC-SHA Key"] => Confidential<Vec<u8>>, hmac_sha);
+ retr_constructor!(#[doc = r"This function will return an error if OpenSSL cannot create a hash or the curve is invalid"]
+ | #[doc = r"EC PRIVATE Key"] => PKey<Private>, ec);
+
/// Reference to the confidential data
pub fn confidential(&self) -> &[u8] {
match &self {
Self::Null => &[],
Self::Association { secret, .. } => secret.value().as_slice(),
+ Self::Retrievable { secret, .. } => secret.value(),
}
}
@@ -79,7 +148,7 @@ impl GuestSecret {
pub(crate) fn auth(&self) -> SecretAuth {
match &self {
Self::Null => SecretAuth::Null,
- // Panic: every non null secret type is listable -> no panic
+ // Panic: every non null secret type is list-able -> no panic
listable => {
SecretAuth::Listable(ListableSecretHdr::from_guest_secret(listable).unwrap())
}
@@ -92,6 +161,7 @@ impl GuestSecret {
// Null is not listable, but the ListableSecretType provides the type constant (1)
Self::Null => ListableSecretType::NULL,
Self::Association { .. } => ListableSecretType::ASSOCIATION,
+ Self::Retrievable { kind, .. } => kind.into(),
}
}
@@ -100,6 +170,7 @@ impl GuestSecret {
match self {
Self::Null => 0,
Self::Association { secret, .. } => secret.value().len() as u32,
+ Self::Retrievable { secret, .. } => secret.value().len() as u32,
}
}
@@ -107,18 +178,157 @@ impl GuestSecret {
fn id(&self) -> Option<SecretId> {
match self {
Self::Null => None,
- Self::Association { id, .. } => Some(id.to_owned()),
+ Self::Association { id, .. } | Self::Retrievable { id, .. } => Some(id.to_owned()),
}
}
}
+type RetrKeyInfo = (RetrievableSecret, Confidential<Vec<u8>>);
+
+fn extend_to_multiple(mut key: Vec<u8>, multiple: usize) -> Confidential<Vec<u8>> {
+ match key.len().checked_rem(multiple) {
+ Some(0) | None => key,
+ Some(m) => {
+ key.resize(key.len() + multiple - m, 0);
+ key
+ }
+ }
+ .into()
+}
+
+/// Get a plain-text key
+///
+/// ```none
+/// size U16<BigEndian> | payload (0-8190) bytes
+/// ```
+fn plaintext(inp: Confidential<Vec<u8>>) -> Result<RetrKeyInfo> {
+ let key_len = inp.value().len();
+ if key_len > RetrieveCmd::MAX_SIZE {
+ return Err(Error::RetrInvKey {
+ what: "key size",
+ value: key_len.to_string(),
+ kind: RetrievableSecret::PlainText.to_string(),
+ exp: RetrievableSecret::PlainText.expected(),
+ });
+ }
+ let mut key = Vec::with_capacity(2 + inp.value().len());
+ let key_len: U16<BigEndian> = (key_len as u16).into();
+ key.extend_from_slice(key_len.as_bytes());
+ key.extend_from_slice(inp.value());
+ let key = extend_to_multiple(key, SymKeyType::AES_256_GCM_BLOCK_LEN);
+
+ Ok((RetrievableSecret::PlainText, key))
+}
+
+/// Get an AES-key
+fn aes(key: Confidential<Vec<u8>>) -> Result<RetrKeyInfo> {
+ let key_len = key.value().len() as u32;
+ let bit_size = bitsize(key_len);
+ match AesSizes::from_bits(bit_size) {
+ Some(size) => Ok((RetrievableSecret::Aes(size), key)),
+ None => {
+ // Use some AES type to get exp sizes and name
+ let kind = RetrievableSecret::Aes(AesSizes::Bits128);
+ Err(Error::RetrInvKey {
+ what: "key size",
+ value: bit_size.to_string(),
+ kind: format!("{kind:#}"),
+ exp: kind.expected(),
+ })
+ }
+ }
+}
+
+/// Get an AES-XTS-key
+fn aes_xts(key: Confidential<Vec<u8>>) -> Result<RetrKeyInfo> {
+ let key_len = key.value().len() as u32;
+ let bit_size = bitsize(key_len / 2);
+ match AesXtsSizes::from_bits(bit_size) {
+ Some(size) => Ok((RetrievableSecret::AesXts(size), key)),
+ None => {
+ // Use some AES-XTS type to get exp sizes and name
+ let kind = RetrievableSecret::AesXts(AesXtsSizes::Bits128);
+ Err(Error::RetrInvKey {
+ what: "key size",
+ value: bit_size.to_string(),
+ kind: format!("{kind:#}"),
+ exp: kind.expected(),
+ })
+ }
+ }
+}
+
+/// Get an HMAC-SHA-key
+fn hmac_sha(key: Confidential<Vec<u8>>) -> Result<RetrKeyInfo> {
+ let key_len = key.value().len() as u32;
+ let size = bitsize(key_len / 2);
+ match HmacShaSizes::from_sha_size(size) {
+ Some(size) => Ok((RetrievableSecret::HmacSha(size), key)),
+ None => {
+ // Use some HMAC type to get exp sizes and name
+ let kind = RetrievableSecret::HmacSha(HmacShaSizes::Sha256);
+ Err(Error::RetrInvKey {
+ what: "key size",
+ value: size.to_string(),
+ kind: format!("{kind:#}"),
+ exp: kind.expected(),
+ })
+ }
+ }
+}
+
+/// Get an EC-private-key
+fn ec(key: PKey<Private>) -> Result<RetrKeyInfo> {
+ let (key, nid) = match key.id() {
+ Id::EC => {
+ let ec_key = key.ec_key()?;
+ let key = ec_key.private_key().to_vec();
+ let nid = ec_key.group().curve_name().unwrap_or(Nid::UNDEF);
+ (key, nid)
+ }
+ // ED keys are not handled via the EC struct in OpenSSL.
+ id @ (Id::ED25519 | Id::ED448) => {
+ let key = key.raw_private_key()?;
+ let nid = Nid::from_raw(id.as_raw());
+ (key, nid)
+ }
+ _ => (vec![], Nid::UNDEF),
+ };
+
+ let kind = match nid {
+ Nid::X9_62_PRIME256V1 => EcCurves::Secp256R1,
+ Nid::SECP384R1 => EcCurves::Secp384R1,
+ Nid::SECP521R1 => EcCurves::Secp521R1,
+ NID_ED25519 => EcCurves::Ed25519,
+ NID_ED448 => EcCurves::Ed448,
+ nid => {
+ // Use some EC type to get exp sizes and name
+ let ec = RetrievableSecret::Ec(EcCurves::Secp521R1);
+ return Err(Error::RetrInvKey {
+ what: "curve or format",
+ kind: format!("{ec:#}"),
+ value: nid.long_name()?.to_string(),
+ exp: ec.expected(),
+ });
+ }
+ };
+
+ let key = kind.resize_raw_key(key);
+ Ok((RetrievableSecret::Ec(kind), key.into()))
+}
+
+#[inline(always)]
+const fn bitsize(bytesize: u32) -> u32 {
+ bytesize * 8
+}
+
impl Display for GuestSecret {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
match self {
Self::Null => write!(f, "Meta"),
gs => {
let kind: U16<BigEndian> = gs.kind().into();
- let st: ListableSecretType = kind.into();
+ let st: ListableSecretType = kind.get().into();
write!(f, "{st}")
}
}
@@ -153,20 +363,24 @@ assert_size!(ListableSecretHdr, 0x30);
impl ListableSecretHdr {
fn from_guest_secret(gs: &GuestSecret) -> Option<Self> {
- let id = gs.id()?;
Some(Self {
res0: 0,
kind: gs.kind().into(),
secret_len: gs.secret_len().into(),
res8: 0,
- id,
+ id: gs.id()?,
})
}
}
#[cfg(test)]
mod test {
+
+ use super::HmacShaSizes as HmacSizes;
+ use super::RetrievableSecret::*;
use super::*;
+ use openssl::ec::{EcGroup, EcKey};
+ use pv_core::uv::AesSizes;
use serde_test::{assert_tokens, Token};
#[test]
@@ -187,8 +401,103 @@ mod test {
assert_eq!(secret, exp);
}
+ macro_rules! retr_test {
+ ($name: ident, $func: ident, $size: expr, $exp_kind: expr) => {
+ #[test]
+ fn $name() {
+ let secret_value = vec![0x11; $size];
+ let name = "test retr secret".to_string();
+ let secret = GuestSecret::$func(&name, secret_value.clone().into()).unwrap();
+ let exp_id = [
+ 0x61, 0x2c, 0xd6, 0x3e, 0xa8, 0xf2, 0xc1, 0x15, 0xc1, 0xe, 0x15, 0xb8, 0x8a,
+ 0x90, 0x16, 0xc1, 0x55, 0xef, 0x9c, 0x7c, 0x2c, 0x8e, 0x56, 0xd0, 0x78, 0x4c,
+ 0x8a, 0x1d, 0xc9, 0x3a, 0x80, 0xba,
+ ];
+ let exp = GuestSecret::Retrievable {
+ kind: $exp_kind,
+ name,
+ id: exp_id.into(),
+ secret: secret_value.into(),
+ };
+ assert_eq!(exp, secret);
+ }
+ };
+ }
+
+ retr_test!(retr_aes_128, aes, 16, Aes(AesSizes::Bits128));
+ retr_test!(retr_aes_192, aes, 24, Aes(AesSizes::Bits192));
+ retr_test!(retr_aes_256, aes, 32, Aes(AesSizes::Bits256));
+ retr_test!(retr_aes_xts_128, aes_xts, 32, AesXts(AesXtsSizes::Bits128));
+ retr_test!(retr_aes_xts_256, aes_xts, 64, AesXts(AesXtsSizes::Bits256));
+ retr_test!(retr_aes_hmac_256, hmac_sha, 64, HmacSha(HmacSizes::Sha256));
+ retr_test!(retr_aes_hmac_512, hmac_sha, 128, HmacSha(HmacSizes::Sha512));
+
+ #[test]
+ fn plaintext_no_pad() {
+ let key = vec![0, 14, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7];
+ let name = "PLAINTEXT_PAD".to_string();
+ let secret = GuestSecret::plaintext(&name, key[2..].to_vec().into()).unwrap();
+ let exp_id = [
+ 15, 123, 176, 210, 135, 231, 220, 232, 148, 93, 198, 195, 165, 212, 214, 129, 45, 1,
+ 94, 11, 167, 18, 151, 15, 120, 254, 13, 109, 173, 186, 37, 74,
+ ];
+ let exp = GuestSecret::Retrievable {
+ kind: PlainText,
+ name,
+ id: exp_id.into(),
+ secret: key.into(),
+ };
+
+ assert_eq!(secret, exp);
+ }
+
#[test]
- fn ap_asc_parse() {
+ fn plaintext_pad() {
+ let key = vec![0, 10, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 0, 0, 0, 0];
+ let name = "PLAINTEXT_PAD".to_string();
+ let secret = GuestSecret::plaintext(&name, key[2..12].to_vec().into()).unwrap();
+ let exp_id = [
+ 15, 123, 176, 210, 135, 231, 220, 232, 148, 93, 198, 195, 165, 212, 214, 129, 45, 1,
+ 94, 11, 167, 18, 151, 15, 120, 254, 13, 109, 173, 186, 37, 74,
+ ];
+ let exp = GuestSecret::Retrievable {
+ kind: PlainText,
+ name,
+ id: exp_id.into(),
+ secret: key.into(),
+ };
+
+ assert_eq!(secret, exp);
+ }
+
+ #[track_caller]
+ fn test_ec(grp: Nid, exp_kind: EcCurves, exp_len: usize) {
+ let key = match grp {
+ NID_ED25519 => PKey::generate_ed25519().unwrap(),
+ NID_ED448 => PKey::generate_ed448().unwrap(),
+ nid => {
+ let group = EcGroup::from_curve_name(nid).unwrap();
+ let key = EcKey::generate(&group).unwrap();
+ PKey::from_ec_key(key).unwrap()
+ }
+ };
+ let (kind, key) = ec(key).unwrap();
+
+ assert_eq!(kind, Ec(exp_kind));
+ assert_eq!(key.value().len(), exp_len);
+ }
+
+ #[test]
+ fn retr_ec() {
+ test_ec(Nid::X9_62_PRIME256V1, EcCurves::Secp256R1, 32);
+ test_ec(Nid::SECP384R1, EcCurves::Secp384R1, 48);
+ test_ec(Nid::SECP521R1, EcCurves::Secp521R1, 80);
+ test_ec(NID_ED25519, EcCurves::Ed25519, 32);
+ test_ec(NID_ED448, EcCurves::Ed448, 64);
+ }
+
+ #[test]
+ fn asc_parse() {
let id = [
0x01, 0x23, 0x45, 0x67, 0x89, 0xab, 0xcd, 0xef, 0x01, 0x23, 0x45, 0x67, 0x89, 0xab,
0xcd, 0xef, 0x01, 0x23, 0x45, 0x67, 0x89, 0xab, 0xcd, 0xef, 0x01, 0x23, 0x45, 0x67,
@@ -217,6 +526,39 @@ mod test {
);
}
+ #[test]
+ fn retrievable_parse() {
+ let id = [
+ 0x01, 0x23, 0x45, 0x67, 0x89, 0xab, 0xcd, 0xef, 0x01, 0x23, 0x45, 0x67, 0x89, 0xab,
+ 0xcd, 0xef, 0x01, 0x23, 0x45, 0x67, 0x89, 0xab, 0xcd, 0xef, 0x01, 0x23, 0x45, 0x67,
+ 0x89, 0xab, 0xcd, 0xef,
+ ];
+ let asc = GuestSecret::Retrievable {
+ kind: PlainText,
+ name: "test123".to_string(),
+ id: id.into(),
+ secret: vec![].into(),
+ };
+
+ assert_tokens(
+ &asc,
+ &[
+ Token::StructVariant {
+ name: "GuestSecret",
+ variant: "Retrievable",
+ len: 3,
+ },
+ Token::String("kind"),
+ Token::String("3 (PLAINTEXT)"),
+ Token::String("name"),
+ Token::String("test123"),
+ Token::String("id"),
+ Token::String("0x0123456789abcdef0123456789abcdef0123456789abcdef0123456789abcdef"),
+ Token::StructVariantEnd,
+ ],
+ );
+ }
+
#[test]
fn guest_secret_bin_null() {
let gs = GuestSecret::Null;
@@ -228,7 +570,7 @@ mod test {
}
#[test]
- fn guest_secret_bin_ap() {
+ fn guest_secret_bin_asoc() {
let gs = GuestSecret::Association {
name: "test".to_string(),
id: [1; 32].into(),
@@ -241,4 +583,21 @@ mod test {
assert_eq!(exp, gs_bytes_auth.get());
assert_eq!(&[2; 32], gs.confidential());
}
+
+ #[test]
+ fn guest_secret_bin_retr() {
+ let gs = GuestSecret::Retrievable {
+ kind: PlainText,
+ name: "test".to_string(),
+ id: [1; 32].into(),
+ secret: vec![2; 32].into(),
+ };
+ let auth = gs.auth();
+ let gs_bytes_auth = auth.get();
+ let mut exp = vec![0u8, 0, 0, 3, 0, 0, 0, 0x20, 0, 0, 0, 0, 0, 0, 0, 0];
+ exp.extend([1; 32]);
+
+ assert_eq!(exp, gs_bytes_auth);
+ assert_eq!(&[2; 32], gs.confidential());
+ }
}
diff --git a/rust/pv/src/uvsecret/retr_secret.rs b/rust/pv/src/uvsecret/retr_secret.rs
new file mode 100644
index 00000000..5fad016f
--- /dev/null
+++ b/rust/pv/src/uvsecret/retr_secret.rs
@@ -0,0 +1,234 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright IBM Corp. 2024
+
+use crate::{pem::Pem, uvsecret::guest_secret::MAX_SIZE_PLAIN_PAYLOAD, Result};
+
+use byteorder::BigEndian;
+use log::warn;
+use pv_core::{
+ request::Confidential,
+ uv::{ListableSecretType, RetrievableSecret, RetrieveCmd},
+};
+use zerocopy::{FromBytes, U16};
+
+/// An IBM Protected Key
+///
+/// A protected key, writeable as pem.
+///
+/// Will convert into PEM as:
+/// ```PEM
+///-----BEGIN IBM PROTECTED KEY-----
+///kind: <name>
+///
+///<protected key in base64>
+///-----END IBM PROTECTED KEY-----
+/// ```
+#[derive(Debug, PartialEq, Eq)]
+pub struct IbmProtectedKey {
+ kind: ListableSecretType,
+ key: Confidential<Vec<u8>>,
+}
+
+impl IbmProtectedKey {
+ /// Get the binary representation of the key.
+ pub fn data(&self) -> &[u8] {
+ self.key.value()
+ }
+
+ /// Converts a [`IbmProtectedKey`] into a vector.
+ pub fn into_bytes(self) -> Confidential<Vec<u8>> {
+ self.key
+ }
+
+ /// Get the data in PEM format.
+ ///
+ /// # Errors
+ ///
+ /// This function will return an error if the PEM conversion failed (very unlikely).
+ pub fn to_pem(&self) -> Result<Pem> {
+ Pem::new(
+ "IBM PROTECTED KEY",
+ format!("kind: {}", self.kind),
+ self.key.value(),
+ )
+ }
+
+ fn new<K>(kind: ListableSecretType, key: K) -> Self
+ where
+ K: Into<Confidential<Vec<u8>>>,
+ {
+ Self {
+ kind,
+ key: key.into(),
+ }
+ }
+}
+
+impl From<RetrieveCmd> for RetrievedSecret {
+ fn from(value: RetrieveCmd) -> Self {
+ let kind = value.meta_data().stype();
+ let key = value.into_key();
+
+ match kind {
+ ListableSecretType::Retrievable(RetrievableSecret::PlainText) => {
+ // Will not run into default, retrieve has a granularity of 16 bytes and 16 bytes is the
+ // minimum size
+ let len = U16::<BigEndian>::read_from_prefix(key.value())
+ .unwrap_or_default()
+ .get() as usize;
+
+ // Test if the plain text secret has a size:
+ // 1. len <= 8190
+ // 2. first two bytes are max 15 less than buffer-size+2
+ // 3. bytes after len + 2 are zero
+ match len <= MAX_SIZE_PLAIN_PAYLOAD
+ && key.value().len() - (len + 2) < 15
+ && key.value()[len + 2..].iter().all(|c| *c == 0)
+ {
+ false => Self::Plaintext(key),
+ true => Self::Plaintext(key.value()[2..len + 2].to_vec().into()),
+ }
+ }
+ kind => {
+ match kind {
+ ListableSecretType::Retrievable(_) => (),
+ _ => warn!("Retrieved an unretrievable Secret! Will continue; interpreting it as a protected key."),
+ }
+ Self::ProtectedKey(IbmProtectedKey::new(kind, key))
+ }
+ }
+ }
+}
+
+/// A retrieved Secret.
+#[derive(Debug, PartialEq, Eq)]
+pub enum RetrievedSecret {
+ /// A plaintext secret
+ Plaintext(Confidential<Vec<u8>>),
+ /// An [`IbmProtectedKey`]
+ ProtectedKey(IbmProtectedKey),
+}
+
+impl RetrievedSecret {
+ /// Create a new IBM PROTECTED KEY object
+ pub fn from_cmd(cmd: RetrieveCmd) -> Self {
+ cmd.into()
+ }
+
+ /// Get the binary representation of the key.
+ pub fn data(&self) -> &[u8] {
+ match self {
+ RetrievedSecret::Plaintext(p) => p.value(),
+ RetrievedSecret::ProtectedKey(p) => p.data(),
+ }
+ }
+
+ /// Converts a [`IbmProtectedKey`] into a vector.
+ pub fn into_bytes(self) -> Confidential<Vec<u8>> {
+ match self {
+ RetrievedSecret::Plaintext(p) => p,
+ RetrievedSecret::ProtectedKey(p) => p.into_bytes(),
+ }
+ }
+ /// Get the data in PEM format.
+ ///
+ /// # Errors
+ ///
+ /// This function will return an error if the PEM conversion failed (very unlikely).
+ pub fn to_pem(&self) -> Result<Pem> {
+ match self {
+ RetrievedSecret::Plaintext(p) => Pem::new("PLAINTEXT SECRET", None, p.value()),
+ RetrievedSecret::ProtectedKey(p) => p.to_pem(),
+ }
+ }
+}
+
+#[cfg(test)]
+mod test {
+ use super::*;
+ use pv_core::uv::*;
+
+ fn mk_retr(secret: &[u8]) -> RetrievedSecret {
+ let entry = SecretEntry::new(
+ 0,
+ ListableSecretType::Retrievable(RetrievableSecret::PlainText),
+ SecretId::default(),
+ secret.len() as u32,
+ );
+ let mut cmd = RetrieveCmd::from_entry(entry).unwrap();
+ cmd.data().unwrap().copy_from_slice(secret);
+ RetrievedSecret::from_cmd(cmd)
+ }
+
+ #[test]
+ fn from_retr_cmd() {
+ let secret = vec![0, 10, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0xa, 0, 0, 0, 0];
+ let prot_key = mk_retr(&secret);
+ let exp = RetrievedSecret::Plaintext(secret[2..12].to_vec().into());
+ assert_eq!(prot_key, exp);
+ }
+
+ #[test]
+ fn from_retr_inv_size() {
+ let secret = vec![0x20; 32];
+ let prot_key = mk_retr(&secret);
+ let exp = RetrievedSecret::Plaintext(secret.into());
+ assert_eq!(prot_key, exp);
+ }
+
+ #[test]
+ fn from_retr_inv_no_zero_after_end() {
+ let secret = vec![0, 10, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0xa, 1, 0, 0, 0];
+ let prot_key = mk_retr(&secret);
+ let exp = RetrievedSecret::Plaintext(secret.into());
+ assert_eq!(prot_key, exp);
+ }
+
+ #[test]
+ fn from_retr_inv_to_much_padding() {
+ let secret = vec![
+ 0, 10, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0xa, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+ 0, 0, 0, 0,
+ ];
+ let prot_key = mk_retr(&secret);
+ let exp = RetrievedSecret::Plaintext(secret.into());
+ assert_eq!(prot_key, exp);
+ }
+
+ #[test]
+ fn from_retr_0_size() {
+ let secret = vec![0x00; 32];
+ let prot_key = mk_retr(&secret);
+ let exp = RetrievedSecret::Plaintext(secret.into());
+ assert_eq!(prot_key, exp);
+ }
+
+ #[test]
+ fn plain_text_pem() {
+ let exp = "\
+ -----BEGIN PLAINTEXT SECRET-----\n\
+ ERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERER\n\
+ -----END PLAINTEXT SECRET-----\n";
+ let prot = RetrievedSecret::Plaintext(vec![17; 48].into());
+ let pem = prot.to_pem().unwrap();
+ let pem_str = pem.to_string();
+ assert_eq!(pem_str, exp);
+ }
+
+ #[test]
+ fn prot_key_pem() {
+ let exp = "\
+ -----BEGIN IBM PROTECTED KEY-----\n\
+ kind: AES-128-KEY\n\n\
+ ERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERERER\n\
+ -----END IBM PROTECTED KEY-----\n";
+ let prot = IbmProtectedKey::new(
+ ListableSecretType::Retrievable(RetrievableSecret::Aes(AesSizes::Bits128)),
+ vec![17; 48],
+ );
+ let pem = prot.to_pem().unwrap();
+ let pem_str = pem.to_string();
+ assert_eq!(pem_str, exp);
+ }
+}

View File

@ -0,0 +1,95 @@
From a14f9d4edcc5db0d54e4fbe3ec3d98c7c270bf8e Mon Sep 17 00:00:00 2001
From: Steffen Eiden <seiden@linux.ibm.com>
Date: Fri, 13 Dec 2024 15:04:02 +0100
Subject: [PATCH] rust/pvsecret: Improve CLI
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Improve the wording of the help/man text/
Acked-by: Marc Hartmayer <marc@linux.ibm.com>
Reviewed-by: Christoph Schlameuss <schlameuss@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
---
rust/pvsecret/src/cli.rs | 26 +++++++++++++-------------
1 file changed, 13 insertions(+), 13 deletions(-)
diff --git a/rust/pvsecret/src/cli.rs b/rust/pvsecret/src/cli.rs
index 6deaaebd..c4b9f2b3 100644
--- a/rust/pvsecret/src/cli.rs
+++ b/rust/pvsecret/src/cli.rs
@@ -37,8 +37,8 @@ pub struct CreateSecretOpt {
/// Specifies the header of the guest image.
///
- /// Can be an IBM Secure Execution image created by genprotimg or an extracted IBM Secure
- /// Execution header. The header must start at a page boundary.
+ /// Can be an IBM Secure Execution image created by 'pvimg/genprotimg' or an
+ /// extracted IBM Secure Execution header.
#[arg(long, value_name = "FILE", value_hint = ValueHint::FilePath)]
pub hdr: String,
@@ -150,12 +150,12 @@ pub enum AddSecretType {
/// Create an association secret.
///
- /// Use an association secret to connect a trusted I/O device to a guest. The `pvapconfig` tool
+ /// Use an association secret to connect a trusted I/O device to a guest. The 'pvapconfig' tool
/// provides more information about association secrets.
Association {
- /// String to identify the new secret.
+ /// String that identifies the new secret.
///
- /// The actual secret is set with --input-secret. The name is saved in `NAME.yaml` with
+ /// The actual secret is set with '--input-secret'. The name is saved in `NAME.yaml` with
/// white-spaces mapped to `_`.
name: String,
@@ -166,15 +166,15 @@ pub enum AddSecretType {
stdout: bool,
/// Path from which to read the plaintext secret. Uses a random secret if not specified.
- #[arg(long, value_name = "FILE", value_hint = ValueHint::FilePath, conflicts_with("output_secret"))]
+ #[arg(long, value_name = "SECRET-FILE", value_hint = ValueHint::FilePath, conflicts_with("output_secret"))]
input_secret: Option<String>,
- /// Save the generated secret as plaintext in FILE.
+ /// Save the generated secret as plaintext in SECRET-FILE.
///
/// The generated secret can be used to generate add-secret requests for a different guest
- /// with the same secret using --input-secret. Destroy the secret when it is not used
+ /// with the same secret using '--input-secret'. Destroy the secret when it is not used
/// anymore.
- #[arg(long, value_name = "FILE", value_hint = ValueHint::FilePath,)]
+ #[arg(long, value_name = "SECRET-FILE", value_hint = ValueHint::FilePath,)]
output_secret: Option<String>,
},
}
@@ -243,13 +243,13 @@ pub enum Command {
/// Create a new add-secret request.
///
/// Create add-secret requests for IBM Secure Execution guests. Only create these requests in a
- /// trusted environment, such as your workstation. The `pvattest create` command creates a
+ /// trusted environment, such as your workstation. The 'pvattest create' command creates a
/// randomly generated key to protect the request. The generated requests can then be added on
- /// an IBM Secure Execution guest using `pvsecret add`. The guest can then use the secrets with
+ /// an IBM Secure Execution guest using 'pvsecret add'. The guest can then use the secrets with
/// the use case depending on the secret type.
Create(Box<CreateSecretOpt>),
- /// Perform an add-secret request (s390x only).
+ /// Submit an add-secret request to the Ultravisor (s390x only).
///
/// Perform an add-secret request using a previously generated add-secret request. Only
/// available on s390x.
@@ -258,7 +258,7 @@ pub enum Command {
/// Lock the secret-store (s390x only).
///
/// Lock the secret store (s390x only). After this command executed successfully, all
- /// add-secret requests will fail. Only available on s390x.
+ /// subsequent add-secret requests will fail. Only available on s390x.
Lock,
/// List all ultravisor secrets (s390x only).

View File

@ -0,0 +1,423 @@
From 93da795520ca2f0a73cfbfc951a9b16437a1b95b Mon Sep 17 00:00:00 2001
From: Steffen Eiden <seiden@linux.ibm.com>
Date: Mon, 19 Feb 2024 15:15:16 +0100
Subject: [PATCH] rust/pvsecret: Add support for retrievable secrets
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Support for creating and retrieving retrievable secrets.
Acked-by: Marc Hartmayer <marc@linux.ibm.com>
Reviewed-by: Christoph Schlameuss <schlameuss@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
---
rust/pvsecret/src/cli.rs | 129 +++++++++++++++++++++++++++++++-
rust/pvsecret/src/cmd.rs | 6 +-
rust/pvsecret/src/cmd/create.rs | 30 +++++++-
rust/pvsecret/src/cmd/list.rs | 12 ++-
rust/pvsecret/src/cmd/retr.rs | 62 +++++++++++++++
rust/pvsecret/src/main.rs | 1 +
6 files changed, 230 insertions(+), 10 deletions(-)
create mode 100644 rust/pvsecret/src/cmd/retr.rs
diff --git a/rust/pvsecret/src/cli.rs b/rust/pvsecret/src/cli.rs
index c4b9f2b3..4e747682 100644
--- a/rust/pvsecret/src/cli.rs
+++ b/rust/pvsecret/src/cli.rs
@@ -1,7 +1,10 @@
// SPDX-License-Identifier: MIT
//
-// Copyright IBM Corp. 2023
+// Copyright IBM Corp. 2023, 2024
+use std::fmt::Display;
+
+use clap::error::ErrorKind::ValueValidation;
use clap::{ArgGroup, Args, CommandFactory, Parser, Subcommand, ValueEnum, ValueHint};
use utils::{CertificateOptions, DeprecatedVerbosityOptions, STDOUT};
@@ -177,6 +180,72 @@ pub enum AddSecretType {
#[arg(long, value_name = "SECRET-FILE", value_hint = ValueHint::FilePath,)]
output_secret: Option<String>,
},
+
+ /// Create a retrievable secret.
+ ///
+ /// A retrievable secret is stored in the per-guest storage of the Ultravisor. A SE-guest can
+ /// retrieve the secret at runtime and use it. All retrievable secrets, but the plaintext
+ /// secret, are retrieved as wrapped/protected key objects and only usable inside the current,
+ /// running SE-guest instance.
+ #[command(visible_alias = "retr")]
+ Retrievable {
+ /// String that identifies the new secret.
+ ///
+ /// The actual secret is set with '--secret'. The name is saved in `NAME.yaml` with
+ /// white-spaces mapped to `_`.
+ name: String,
+
+ /// Print the hashed name to stdout.
+ ///
+ /// The hashed name is not written to `NAME.yaml`
+ #[arg(long)]
+ stdout: bool,
+
+ /// Use SECRET-FILE as retrievable secret
+ #[arg(long, value_name = "SECRET-FILE", value_hint = ValueHint::FilePath)]
+ secret: String,
+
+ /// Specify the secret type.
+ ///
+ /// Limitations to the input data apply depending on the secret type.
+ #[arg(long = "type", value_name = "TYPE")]
+ kind: RetrieveableSecretInpKind,
+ },
+}
+
+#[derive(Copy, Clone, PartialEq, Eq, PartialOrd, Ord, ValueEnum, Debug)]
+pub enum RetrieveableSecretInpKind {
+ /// A plaintext secret.
+ /// Can be any file up to 8190 bytes long
+ Plain,
+ /// An AES key.
+ /// Must be a plain byte file 128, 192, or 256 bit long.
+ Aes,
+ /// An AES-XTS key.
+ /// Must be a plain byte file 512, or 1024 bit long.
+ AesXts,
+ /// A HMAC-SHA key.
+ /// Must be a plain byte file 512, or 1024 bit long.
+ HmacSha,
+ /// An elliptic curve private key.
+ /// Must be a PEM or DER file.
+ Ec,
+}
+
+impl Display for RetrieveableSecretInpKind {
+ fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+ write!(
+ f,
+ "{}",
+ match self {
+ Self::Plain => "PLAINTEXT",
+ Self::Aes => "AES KEY",
+ Self::AesXts => "AES-XTS KEY",
+ Self::HmacSha => "HMAC-SHA KEY",
+ Self::Ec => "EC PRIVATE KEY",
+ }
+ )
+ }
}
// all members s390x only
@@ -238,6 +307,56 @@ pub struct VerifyOpt {
pub output: String,
}
+// all members s390x only
+#[derive(Args, Debug)]
+pub struct RetrSecretOptions {
+ /// Specify the secret ID to be retrieved.
+ ///
+ /// Input type depends on '--inform'. If `yaml` (default) is specified, it must be a yaml
+ /// created by the create subcommand of this tool. If `hex` is specified, it must be a hex
+ /// 32-byte unsigned big endian number string. Leading zeros are required.
+ #[cfg(target_arch = "s390x")]
+ #[arg(value_name = "ID", value_hint = ValueHint::FilePath)]
+ pub input: String,
+
+ /// Specify the output path to place the secret value
+ #[cfg(target_arch = "s390x")]
+ #[arg(short, long, value_name = "FILE", default_value = STDOUT, value_hint = ValueHint::FilePath)]
+ pub output: String,
+
+ /// Define input type for the Secret ID
+ #[cfg(target_arch = "s390x")]
+ #[arg(long, value_enum, default_value_t)]
+ pub inform: RetrInpFmt,
+
+ /// Define the output format for the retrieved secret
+ #[cfg(target_arch = "s390x")]
+ #[arg(long, value_enum, default_value_t)]
+ pub outform: RetrOutFmt,
+}
+
+#[derive(Copy, Clone, PartialEq, Eq, PartialOrd, Ord, ValueEnum, Debug, Default)]
+pub enum RetrInpFmt {
+ /// Use a yaml file
+ #[default]
+ Yaml,
+ /// Use a hex string.
+ Hex,
+}
+
+#[derive(Copy, Clone, PartialEq, Eq, PartialOrd, Ord, ValueEnum, Debug, Default)]
+pub enum RetrOutFmt {
+ /// Write the secret as PEM.
+ ///
+ /// File starts with `-----BEGIN IBM PROTECTED KEY----` and `-----BEGIN
+ /// PLAINTEXT SECRET-----` for plaintext secrets it contains one header
+ /// line with the type information and the base64 protected key
+ #[default]
+ Pem,
+ /// Write the secret in binary.
+ Bin,
+}
+
#[derive(Subcommand, Debug)]
pub enum Command {
/// Create a new add-secret request.
@@ -274,6 +393,10 @@ pub enum Command {
/// provided key. Outputs the arbitrary user-data.
Verify(VerifyOpt),
+ /// Retrieve a secret from the UV secret store (s390x only).
+ #[command(visible_alias = "retr")]
+ Retrieve(RetrSecretOptions),
+
/// Print version information and exit.
#[command(aliases(["--version"]), hide(true))]
Version,
@@ -294,13 +417,13 @@ pub fn validate_cli(cli: &CliOptions) -> Result<(), clap::Error> {
}
if secret_out == &Some(format!("{name}.yaml")) {
return Err(CliOptions::command().error(
- clap::error::ErrorKind::ValueValidation,
+ ValueValidation,
format!("Secret output file and the secret name '{name}.yaml' are the same."),
));
}
if format!("{name}.yaml") == opt.output {
return Err(CliOptions::command().error(
- clap::error::ErrorKind::ValueValidation,
+ ValueValidation,
format!(
"output file and the secret name '{}' are the same.",
&opt.output
diff --git a/rust/pvsecret/src/cmd.rs b/rust/pvsecret/src/cmd.rs
index a826fb31..10d99a5b 100644
--- a/rust/pvsecret/src/cmd.rs
+++ b/rust/pvsecret/src/cmd.rs
@@ -16,6 +16,8 @@ mod add;
mod list;
#[cfg(target_arch = "s390x")]
mod lock;
+#[cfg(target_arch = "s390x")]
+mod retr;
// Commands (directly) related to UVCs are only available on s389x
#[cfg(target_arch = "s390x")]
@@ -24,12 +26,13 @@ mod uv_cmd {
pub use add::add;
pub use list::list;
pub use lock::lock;
+ pub use retr::retr;
pub const UV_CMD_FN: &[&str] = &["+add", "+lock", "+list"];
}
#[cfg(not(target_arch = "s390x"))]
mod uv_cmd {
- use crate::cli::{AddSecretOpt, ListSecretOpt};
+ use crate::cli::{AddSecretOpt, ListSecretOpt, RetrSecretOptions};
use anyhow::{bail, Result};
macro_rules! not_supp {
($name: ident $( ,$opt: ty )?) => {
@@ -40,6 +43,7 @@ mod uv_cmd {
}
not_supp!(add, AddSecretOpt);
not_supp!(list, ListSecretOpt);
+ not_supp!(retr, RetrSecretOptions);
not_supp!(lock);
pub const UV_CMD_FN: &[&str] = &[];
}
diff --git a/rust/pvsecret/src/cmd/create.rs b/rust/pvsecret/src/cmd/create.rs
index 9251c38c..73089a12 100644
--- a/rust/pvsecret/src/cmd/create.rs
+++ b/rust/pvsecret/src/cmd/create.rs
@@ -4,7 +4,6 @@
use std::path::Path;
-use crate::cli::{AddSecretType, CreateSecretFlags, CreateSecretOpt};
use anyhow::{anyhow, bail, Context, Error, Result};
use log::{debug, info, trace, warn};
use pv::{
@@ -22,6 +21,8 @@ use pv::{
use serde_yaml::Value;
use utils::get_writer_from_cli_file_arg;
+use crate::cli::{AddSecretType, CreateSecretFlags, CreateSecretOpt, RetrieveableSecretInpKind};
+
fn write_out<P, D>(path: &P, data: D, ctx: &str) -> pv::Result<()>
where
P: AsRef<Path>,
@@ -32,6 +33,23 @@ where
Ok(())
}
+fn retrievable(name: &str, secret: &str, kind: &RetrieveableSecretInpKind) -> Result<GuestSecret> {
+ let secret_data = read_file(secret, &format!("retrievable {kind}"))?.into();
+
+ match kind {
+ RetrieveableSecretInpKind::Plain => GuestSecret::plaintext(name, secret_data),
+ RetrieveableSecretInpKind::Aes => GuestSecret::aes(name, secret_data),
+ RetrieveableSecretInpKind::AesXts => GuestSecret::aes_xts(name, secret_data),
+ RetrieveableSecretInpKind::HmacSha => GuestSecret::hmac_sha(name, secret_data),
+ RetrieveableSecretInpKind::Ec => GuestSecret::ec(
+ name,
+ read_private_key(secret_data.value())
+ .with_context(|| format!("Cannot read {secret} as {kind} from PEM or DER"))?,
+ ),
+ }
+ .map_err(Error::from)
+}
+
/// Prepare an add-secret request
pub fn create(opt: &CreateSecretOpt) -> Result<()> {
if pv_guest_bit_set() {
@@ -88,6 +106,9 @@ fn build_asrcb(opt: &CreateSecretOpt) -> Result<AddSecretRequest> {
input_secret: None,
..
} => GuestSecret::association(name, None)?,
+ AddSecretType::Retrievable {
+ name, secret, kind, ..
+ } => retrievable(name, secret, kind)?,
};
trace!("AddSecret: {secret:x?}");
@@ -136,7 +157,9 @@ fn build_asrcb(opt: &CreateSecretOpt) -> Result<AddSecretRequest> {
.as_ref()
.map(|p| read_file(p, "User-signing key"))
.transpose()?
- .map(|buf| read_private_key(&buf))
+ .map(|buf| {
+ read_private_key(&buf).context("Cannot read {secret} as private key from PEM or DER")
+ })
.transpose()?;
if user_data.is_some() || user_key.is_some() {
@@ -258,6 +281,9 @@ fn write_secret<P: AsRef<Path>>(
write_out(path, guest_secret.confidential(), "Association secret")?
}
}
+ AddSecretType::Retrievable { name, stdout, .. } => {
+ write_yaml(name, guest_secret, stdout, outp_path)?
+ }
_ => (),
};
Ok(())
diff --git a/rust/pvsecret/src/cmd/list.rs b/rust/pvsecret/src/cmd/list.rs
index f7e3a72b..0bd9eca4 100644
--- a/rust/pvsecret/src/cmd/list.rs
+++ b/rust/pvsecret/src/cmd/list.rs
@@ -3,21 +3,25 @@
// Copyright IBM Corp. 2023
use crate::cli::{ListSecretOpt, ListSecretOutputType};
-use anyhow::{Context, Result};
+use anyhow::{Context, Error, Result};
use log::warn;
use pv::uv::{ListCmd, SecretList, UvDevice, UvcSuccess};
use utils::{get_writer_from_cli_file_arg, STDOUT};
/// Do a List Secrets UVC
-pub fn list(opt: &ListSecretOpt) -> Result<()> {
- let uv = UvDevice::open()?;
+pub fn list_uvc(uv: &UvDevice) -> Result<SecretList> {
let mut cmd = ListCmd::default();
match uv.send_cmd(&mut cmd)? {
UvcSuccess::RC_SUCCESS => (),
UvcSuccess::RC_MORE_DATA => warn!("There is more data available than expected"),
};
+ cmd.try_into().map_err(Error::new)
+}
- let secret_list: SecretList = cmd.try_into()?;
+/// Do a List Secrets UVC and output the list in the requested format
+pub fn list(opt: &ListSecretOpt) -> Result<()> {
+ let uv = UvDevice::open()?;
+ let secret_list = list_uvc(&uv)?;
let mut wr_out = get_writer_from_cli_file_arg(&opt.output)?;
match &opt.format {
diff --git a/rust/pvsecret/src/cmd/retr.rs b/rust/pvsecret/src/cmd/retr.rs
new file mode 100644
index 00000000..7f7704cc
--- /dev/null
+++ b/rust/pvsecret/src/cmd/retr.rs
@@ -0,0 +1,62 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright IBM Corp. 2024
+
+use super::list::list_uvc;
+use crate::cli::{RetrInpFmt, RetrOutFmt, RetrSecretOptions};
+use anyhow::{anyhow, bail, Context, Result};
+use log::{debug, info};
+use pv::{
+ misc::open_file,
+ misc::write,
+ secret::{GuestSecret, RetrievedSecret},
+ uv::{RetrieveCmd, SecretId, UvDevice},
+};
+use utils::get_writer_from_cli_file_arg;
+
+fn retrieve(id: &SecretId) -> Result<RetrievedSecret> {
+ let uv = UvDevice::open()?;
+ let secrets = list_uvc(&uv)?;
+ let secret = secrets
+ .into_iter()
+ .find(|s| s.id() == id.as_ref())
+ .ok_or(anyhow!(
+ "The UV secret-store has no secret with the ID {id}"
+ ))?;
+
+ info!("Try to retrieve secret at index: {}", secret.index());
+ debug!("Try to retrieve: {secret:?}");
+
+ let mut uv_cmd = RetrieveCmd::from_entry(secret)?;
+ uv.send_cmd(&mut uv_cmd)?;
+
+ Ok(RetrievedSecret::from_cmd(uv_cmd))
+}
+
+pub fn retr(opt: &RetrSecretOptions) -> Result<()> {
+ let mut output = get_writer_from_cli_file_arg(&opt.output)?;
+ let id = match &opt.inform {
+ RetrInpFmt::Yaml => match serde_yaml::from_reader(&mut open_file(&opt.input)?)? {
+ GuestSecret::Retrievable { id, .. } => id,
+ gs => bail!("The file contains a {gs}-secret, which is not retrievable."),
+ },
+ RetrInpFmt::Hex => {
+ serde_yaml::from_str(&opt.input).context("Cannot parse SecretId information")?
+ }
+ };
+
+ let retr_secret =
+ retrieve(&id).context("Could not retrieve the secret from the UV secret store.")?;
+
+ let out_data = match opt.outform {
+ RetrOutFmt::Bin => retr_secret.into_bytes(),
+ RetrOutFmt::Pem => retr_secret.to_pem()?.into_bytes(),
+ };
+ write(
+ &mut output,
+ out_data.value(),
+ &opt.output,
+ "IBM Protected Key",
+ )?;
+ Ok(())
+}
diff --git a/rust/pvsecret/src/main.rs b/rust/pvsecret/src/main.rs
index 502a6ea0..883a3ee2 100644
--- a/rust/pvsecret/src/main.rs
+++ b/rust/pvsecret/src/main.rs
@@ -45,6 +45,7 @@ fn main() -> ExitCode {
Command::Create(opt) => cmd::create(opt),
Command::Version => Ok(print_version!("2024", log_level; FEATURES.concat())),
Command::Verify(opt) => cmd::verify(opt),
+ Command::Retrieve(opt) => cmd::retr(opt),
};
match res {

View File

@ -0,0 +1,313 @@
From 256289a30aa5d3f6a4d2631dea69d1dc47205150 Mon Sep 17 00:00:00 2001
From: Steffen Eiden <seiden@linux.ibm.com>
Date: Wed, 12 Jun 2024 16:23:31 +0200
Subject: [PATCH] rust/pv_core: Refactor secret list
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Improve the secret list implementation. Use structs+{As,From}Bytes
instead of arbitrary seeks and reads/writes to parse the secret list.
Acked-by: Marc Hartmayer <marc@linux.ibm.com>
Reviewed-by: Christoph Schlameuss <schlameuss@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
---
rust/pv_core/src/uvdevice.rs | 10 ++
rust/pv_core/src/uvdevice/secret_list.rs | 124 ++++++++++++++---------
2 files changed, 86 insertions(+), 48 deletions(-)
diff --git a/rust/pv_core/src/uvdevice.rs b/rust/pv_core/src/uvdevice.rs
index e9848243..e701366d 100644
--- a/rust/pv_core/src/uvdevice.rs
+++ b/rust/pv_core/src/uvdevice.rs
@@ -163,6 +163,16 @@ pub enum UvcSuccess {
RC_MORE_DATA = UvDevice::RC_MORE_DATA,
}
+impl UvcSuccess {
+ /// Returns true if there is more data available
+ pub fn more_data(&self) -> bool {
+ match self {
+ Self::RC_SUCCESS => false,
+ Self::RC_MORE_DATA => true,
+ }
+ }
+}
+
/// The `UvDevice` is a (virtual) device on s390 machines to send Ultravisor commands(UVCs) from
/// userspace.
///
diff --git a/rust/pv_core/src/uvdevice/secret_list.rs b/rust/pv_core/src/uvdevice/secret_list.rs
index 4e955010..d7c268c9 100644
--- a/rust/pv_core/src/uvdevice/secret_list.rs
+++ b/rust/pv_core/src/uvdevice/secret_list.rs
@@ -4,16 +4,16 @@
use crate::{
assert_size,
- misc::to_u16,
uv::{AesSizes, AesXtsSizes, EcCurves, HmacShaSizes, ListCmd, RetrievableSecret},
uvdevice::UvCmd,
Error, Result,
};
-use byteorder::{BigEndian, ByteOrder, ReadBytesExt, WriteBytesExt};
+use byteorder::{BigEndian, ByteOrder};
use serde::{Deserialize, Serialize, Serializer};
use std::{
fmt::Display,
io::{Cursor, Read, Seek, Write},
+ mem::size_of,
slice::Iter,
vec::IntoIter,
};
@@ -31,7 +31,7 @@ impl SecretId {
/// Size in bytes of the [`SecretId`]
pub const ID_SIZE: usize = 32;
- /// Create a [`SecretId`] forom a buffer.
+ /// Create a [`SecretId`] from a buffer.
pub fn from(buf: [u8; Self::ID_SIZE]) -> Self {
buf.into()
}
@@ -120,7 +120,7 @@ impl SecretEntry {
&self.index
}
- /// Returns the secret type of this [`SecretEntry`].
+ /// Returns the secret type of this [`SecretEntry`]
pub fn stype(&self) -> ListableSecretType {
self.stype.get().into()
}
@@ -161,12 +161,45 @@ impl Display for SecretEntry {
}
}
+#[repr(C)]
+#[derive(Debug, FromBytes, AsBytes, FromZeroes, Clone, PartialEq, Eq, Default, Serialize)]
+struct SecretListHdr {
+ #[serde(skip)]
+ num_secrets_stored: U16<BigEndian>,
+ #[serde(serialize_with = "ser_u16")]
+ total_num_secrets: U16<BigEndian>,
+ #[serde(skip)]
+ next_secret_idx: U16<BigEndian>,
+ #[serde(skip)]
+ reserved_06: u16,
+ #[serde(skip)]
+ reserved_08: u64,
+}
+
+impl SecretListHdr {
+ fn new(num_secrets_stored: u16, total_num_secrets: u16, next_secret_idx: u16) -> Self {
+ Self {
+ num_secrets_stored: num_secrets_stored.into(),
+ total_num_secrets: total_num_secrets.into(),
+ next_secret_idx: next_secret_idx.into(),
+ reserved_06: 0,
+ reserved_08: 0,
+ }
+ }
+}
+assert_size!(SecretListHdr, 16);
+
/// List of secrets used to parse the [`crate::uv::ListCmd`] result.
///
-/// The list should not hold more than 0xffffffff elements
-#[derive(Debug, PartialEq, Eq, Serialize)]
+/// The list should ONLY be created from an UV-Call result using either:
+/// - [`TryInto::try_into`] from [`ListCmd`]
+/// - [`SecretList::decode`]
+/// Any other ways can create invalid lists that do not represent the UV secret store.
+/// The list must not hold more than [`u32::MAX`] elements
+#[derive(Debug, PartialEq, Eq, Serialize, Default)]
pub struct SecretList {
- total_num_secrets: usize,
+ #[serde(flatten)]
+ hdr: SecretListHdr,
secrets: Vec<SecretEntry>,
}
@@ -202,10 +235,14 @@ impl SecretList {
/// The content of this list will very likely not represent the status of the guest in the
/// Ultravisor. Use of [`SecretList::decode`] in any non-test environments is encuraged.
pub fn new(total_num_secrets: u16, secrets: Vec<SecretEntry>) -> Self {
- Self {
- total_num_secrets: total_num_secrets as usize,
+ Self::new_with_hdr(
+ SecretListHdr::new(total_num_secrets, total_num_secrets, 0),
secrets,
- }
+ )
+ }
+
+ fn new_with_hdr(hdr: SecretListHdr, secrets: Vec<SecretEntry>) -> Self {
+ Self { hdr, secrets }
}
/// Returns an iterator over the slice.
@@ -229,19 +266,12 @@ impl SecretList {
///
/// This number may be not equal to the provided number of [`SecretEntry`]
pub fn total_num_secrets(&self) -> usize {
- self.total_num_secrets
+ self.hdr.total_num_secrets.get() as usize
}
/// Encodes the list in the same binary format the UV would do
pub fn encode<T: Write>(&self, w: &mut T) -> Result<()> {
- let num_s = to_u16(self.secrets.len()).ok_or(Error::ManySecrets)?;
- w.write_u16::<BigEndian>(num_s)?;
- w.write_u16::<BigEndian>(
- self.total_num_secrets
- .try_into()
- .map_err(|_| Error::ManySecrets)?,
- )?;
- w.write_all(&[0u8; 12])?;
+ w.write_all(self.hdr.as_bytes())?;
for secret in &self.secrets {
w.write_all(secret.as_bytes())?;
}
@@ -250,19 +280,20 @@ impl SecretList {
/// Decodes the list from the binary format of the UV into this internal representation
pub fn decode<R: Read + Seek>(r: &mut R) -> std::io::Result<Self> {
- let num_s = r.read_u16::<BigEndian>()?;
- let total_num_secrets = r.read_u16::<BigEndian>()? as usize;
- let mut v: Vec<SecretEntry> = Vec::with_capacity(num_s as usize);
- r.seek(std::io::SeekFrom::Current(12))?; // skip reserved bytes
+ let mut buf = [0u8; size_of::<SecretListHdr>()];
+ r.read_exact(&mut buf)?;
+ let hdr = SecretListHdr::ref_from(&buf).unwrap();
+
let mut buf = [0u8; SecretEntry::STRUCT_SIZE];
- for _ in 0..num_s {
+ let mut v = Vec::with_capacity(hdr.num_secrets_stored.get() as usize);
+ for _ in 0..hdr.num_secrets_stored.get() {
r.read_exact(&mut buf)?;
// cannot fail. buffer has the same size as the secret entry
let secr = SecretEntry::read_from(buf.as_slice()).unwrap();
v.push(secr);
}
Ok(Self {
- total_num_secrets,
+ hdr: hdr.clone(),
secrets: v,
})
}
@@ -278,7 +309,7 @@ impl TryFrom<ListCmd> for SecretList {
impl Display for SecretList {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
- writeln!(f, "Total number of secrets: {}", self.total_num_secrets)?;
+ writeln!(f, "Total number of secrets: {}", self.total_num_secrets())?;
if !self.secrets.is_empty() {
writeln!(f)?;
}
@@ -481,8 +512,8 @@ mod test {
let buf = [
0x00u8, 0x01, // num secr stored
0x01, 0x12, // total num secrets
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, // reserved
+ 0x01, 0x01, // next valid idx
+ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, // reserved
// secret
0x00, 0x01, 0x00, 0x02, // idx + type
0x00, 0x00, 0x00, 0x20, // len
@@ -493,16 +524,16 @@ mod test {
0x00, 0x00, 0x00, 0x00,
];
- let exp = SecretList {
- total_num_secrets: 0x112,
- secrets: vec![SecretEntry {
+ let exp = SecretList::new_with_hdr(
+ SecretListHdr::new(0x001, 0x112, 0x101),
+ vec![SecretEntry {
index: 1.into(),
stype: 2.into(),
len: 32.into(),
res_8: 0,
id: SecretId::from([0; 32]),
}],
- };
+ );
let mut br = BufReader::new(Cursor::new(buf));
let sl = SecretList::decode(&mut br).unwrap();
@@ -514,8 +545,8 @@ mod test {
const EXP: &[u8] = &[
0x00, 0x01, // num secr stored
0x01, 0x12, // total num secrets
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, // reserved
+ 0x01, 0x01, // next valid idx
+ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, // reserved
// secret
0x00, 0x01, 0x00, 0x02, // idx + type
0x00, 0x00, 0x00, 0x20, // len
@@ -526,16 +557,16 @@ mod test {
0x00, 0x00, 0x00, 0x00,
];
- let sl = SecretList {
- total_num_secrets: 0x112,
- secrets: vec![SecretEntry {
+ let sl = SecretList::new_with_hdr(
+ SecretListHdr::new(0x001, 0x112, 0x101),
+ vec![SecretEntry {
index: 1.into(),
stype: 2.into(),
len: 32.into(),
res_8: 0,
id: SecretId::from([0; 32]),
}],
- };
+ );
let mut buf = [0u8; 0x40];
{
@@ -587,26 +618,23 @@ mod test {
#[test]
fn secret_list_ser() {
- let list = SecretList {
- total_num_secrets: 0x112,
- secrets: vec![SecretEntry {
+ let list = SecretList::new_with_hdr(
+ SecretListHdr::new(0x001, 0x112, 0x101),
+ vec![SecretEntry {
index: 1.into(),
stype: 2.into(),
len: 32.into(),
res_8: 0,
id: SecretId::from([0; 32]),
}],
- };
+ );
assert_ser_tokens(
&list,
&[
- Token::Struct {
- name: "SecretList",
- len: 2,
- },
+ Token::Map { len: None },
Token::String("total_num_secrets"),
- Token::U64(0x112),
+ Token::U16(0x112),
Token::String("secrets"),
Token::Seq { len: Some(1) },
Token::Struct {
@@ -623,7 +651,7 @@ mod test {
Token::String("0x0000000000000000000000000000000000000000000000000000000000000000"),
Token::StructEnd,
Token::SeqEnd,
- Token::StructEnd,
+ Token::MapEnd,
],
)
}

View File

@ -0,0 +1,111 @@
From 93216d916c479ee1292aa1d598ac9c0e7f585bd8 Mon Sep 17 00:00:00 2001
From: Steffen Eiden <seiden@linux.ibm.com>
Date: Wed, 12 Jun 2024 16:35:15 +0200
Subject: [PATCH] rust/pv*: Support longer secret lists
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Make use of the enhanced list secrets UAPI for the uvdevice in the latest kernel
version. This allows fetching secret lists with more than 85 entries via
reserving more userspace memory in the IOCTL argument.
While at it, move the errno readout next to the ioctl-syscall.
Acked-by: Marc Hartmayer <marc@linux.ibm.com>
Reviewed-by: Christoph Schlameuss <schlameuss@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
---
rust/pv_core/src/uvdevice.rs | 6 ++++--
rust/pv_core/src/uvdevice/secret.rs | 11 +++++++++++
rust/pvsecret/src/cmd/list.rs | 28 +++++++++++++++++++++-------
3 files changed, 36 insertions(+), 9 deletions(-)
diff --git a/rust/pv_core/src/uvdevice.rs b/rust/pv_core/src/uvdevice.rs
index e701366d..689748a1 100644
--- a/rust/pv_core/src/uvdevice.rs
+++ b/rust/pv_core/src/uvdevice.rs
@@ -59,11 +59,13 @@ fn ioctl_raw(raw_fd: RawFd, cmd: c_ulong, cb: &mut IoctlCb) -> Result<()> {
rc = ioctl(raw_fd, cmd, cb.as_ptr_mut());
}
+ // NOTE io::Error handles all errnos ioctl uses
+ let errno = std::io::Error::last_os_error();
+
debug!("ioctl resulted with {cb:?}");
match rc {
0 => Ok(()),
- // NOTE io::Error handles all errnos ioctl uses
- _ => Err(std::io::Error::last_os_error().into()),
+ _ => Err(errno.into()),
}
}
diff --git a/rust/pv_core/src/uvdevice/secret.rs b/rust/pv_core/src/uvdevice/secret.rs
index 263f17d5..cb5b7233 100644
--- a/rust/pv_core/src/uvdevice/secret.rs
+++ b/rust/pv_core/src/uvdevice/secret.rs
@@ -24,6 +24,17 @@ impl ListCmd {
Self(vec![0; size])
}
+ /// Create a new list secrets command with `pages` capacity.
+ ///
+ /// * `pages` - number pf pages to allocate for this IOCTL
+ ///
+ /// # Panic
+ /// This function will trigger a panic if the allocation size is larger than [`usize::MAX`].
+ /// Very likely an OOM situation occurs way before this!
+ pub fn with_pages(pages: usize) -> Self {
+ Self::with_size(pages * PAGESIZE)
+ }
+
/// Create a new list secrets command with a one page capacity
pub fn new() -> Self {
Self::with_size(PAGESIZE)
diff --git a/rust/pvsecret/src/cmd/list.rs b/rust/pvsecret/src/cmd/list.rs
index 0bd9eca4..56294cac 100644
--- a/rust/pvsecret/src/cmd/list.rs
+++ b/rust/pvsecret/src/cmd/list.rs
@@ -2,19 +2,33 @@
//
// Copyright IBM Corp. 2023
+use std::io::ErrorKind;
+
use crate::cli::{ListSecretOpt, ListSecretOutputType};
use anyhow::{Context, Error, Result};
-use log::warn;
-use pv::uv::{ListCmd, SecretList, UvDevice, UvcSuccess};
+use log::{info, warn};
+use pv::uv::{ListCmd, SecretList, UvDevice};
use utils::{get_writer_from_cli_file_arg, STDOUT};
+const SECRET_LIST_BUF_SIZE: usize = 4;
+
/// Do a List Secrets UVC
pub fn list_uvc(uv: &UvDevice) -> Result<SecretList> {
- let mut cmd = ListCmd::default();
- match uv.send_cmd(&mut cmd)? {
- UvcSuccess::RC_SUCCESS => (),
- UvcSuccess::RC_MORE_DATA => warn!("There is more data available than expected"),
- };
+ let mut cmd = ListCmd::with_pages(SECRET_LIST_BUF_SIZE);
+ let more_data = match uv.send_cmd(&mut cmd) {
+ Ok(v) => Ok(v),
+ Err(pv::PvCoreError::Io(e)) if e.kind() == ErrorKind::InvalidInput => {
+ info!("Uvdevice does not suport longer list. Fallback to one page list.");
+ cmd = ListCmd::default();
+ uv.send_cmd(&mut cmd)
+ }
+ Err(e) => Err(e),
+ }?
+ .more_data();
+ if more_data {
+ warn!("The secret list contains more data but the uvdevice cannot show all.");
+ }
+
cmd.try_into().map_err(Error::new)
}

View File

@ -0,0 +1,387 @@
From ff04f76257791593c8f92374f295a0c478e3b0f7 Mon Sep 17 00:00:00 2001
From: Steffen Eiden <seiden@linux.ibm.com>
Date: Mon, 5 Aug 2024 09:34:47 +0200
Subject: [PATCH] rust/pv*: Allow the use of non-hashes secret IDs
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Secret IDs identify a secret in the store. Tooling (pvsecret) calculates
them by hashing a user-defined string. With this patch it is now
possible to skip the hash step and directly use the input string as the
ID. Up to the first 31 bytes of the input ASCII-string are used. The last byte
is the NUL char. During list pvsecret tries to interpret the secret
as ASCII string and if possible displays the ASCII characters alongside
the hex number.
Also, use the Upper/Lower Hex formatters for the hexstring formatting of
SecretId. Display will, additionally show the ASCII representation if
applicable.
While at it, use Self wherever possible.
Acked-by: Marc Hartmayer <marc@linux.ibm.com>
Reviewed-by: Christoph Schlameuss <schlameuss@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
---
rust/pv/src/uvsecret/guest_secret.rs | 16 ++-
rust/pv_core/src/uvdevice/secret_list.rs | 163 ++++++++++++++++++++---
rust/pvsecret/src/cli.rs | 8 ++
rust/pvsecret/src/cmd/create.rs | 4 +-
rust/pvsecret/src/cmd/retr.rs | 14 +-
5 files changed, 184 insertions(+), 21 deletions(-)
diff --git a/rust/pv/src/uvsecret/guest_secret.rs b/rust/pv/src/uvsecret/guest_secret.rs
index 3bad6d3c..87b58bb8 100644
--- a/rust/pv/src/uvsecret/guest_secret.rs
+++ b/rust/pv/src/uvsecret/guest_secret.rs
@@ -92,7 +92,8 @@ macro_rules! retr_constructor {
}
impl GuestSecret {
- fn name_to_id(name: &str) -> Result<SecretId> {
+ /// Hashes the name with sha256
+ pub fn name_to_id(name: &str) -> Result<SecretId> {
let id: [u8; SecretId::ID_SIZE] = hash(MessageDigest::sha256(), name.as_bytes())?
.to_vec()
.try_into()
@@ -135,6 +136,19 @@ impl GuestSecret {
retr_constructor!(#[doc = r"This function will return an error if OpenSSL cannot create a hash or the curve is invalid"]
| #[doc = r"EC PRIVATE Key"] => PKey<Private>, ec);
+ /// Use the name as ID, do not hash it
+ pub fn no_hash_name(&mut self) {
+ match self {
+ Self::Null => (),
+ Self::Association {
+ name, ref mut id, ..
+ }
+ | Self::Retrievable {
+ name, ref mut id, ..
+ } => id.clone_from(&SecretId::from_string(name)),
+ }
+ }
+
/// Reference to the confidential data
pub fn confidential(&self) -> &[u8] {
match &self {
diff --git a/rust/pv_core/src/uvdevice/secret_list.rs b/rust/pv_core/src/uvdevice/secret_list.rs
index d7c268c9..7c7e63b5 100644
--- a/rust/pv_core/src/uvdevice/secret_list.rs
+++ b/rust/pv_core/src/uvdevice/secret_list.rs
@@ -11,7 +11,9 @@ use crate::{
use byteorder::{BigEndian, ByteOrder};
use serde::{Deserialize, Serialize, Serializer};
use std::{
- fmt::Display,
+ cmp::min,
+ ffi::CStr,
+ fmt::{Debug, Display, LowerHex, UpperHex},
io::{Cursor, Read, Seek, Write},
mem::size_of,
slice::Iter,
@@ -35,6 +37,33 @@ impl SecretId {
pub fn from(buf: [u8; Self::ID_SIZE]) -> Self {
buf.into()
}
+
+ /// Create a Id from a string
+ ///
+ /// Uses the first 31 bytes from `name` as id
+ /// Does not hash anything. Byte 32 is the NUL char
+ pub fn from_string(name: &str) -> Self {
+ let len = min(name.len(), Self::ID_SIZE - 1);
+ let mut res = Self::default();
+ res.0[0..len].copy_from_slice(&name.as_bytes()[0..len]);
+ res
+ }
+
+ /// Tries to represent the Id as printable-ASCII string
+ pub fn as_ascii(&self) -> Option<&str> {
+ if let Ok(t) = CStr::from_bytes_until_nul(&self.0) {
+ if let Ok(t) = t.to_str() {
+ if !t.is_empty()
+ && t.chars()
+ .all(|c| c.is_ascii_whitespace() | c.is_ascii_graphic())
+ && self.0[t.len()..].iter().all(|b| *b == 0)
+ {
+ return Some(t);
+ }
+ }
+ };
+ None
+ }
}
impl Serialize for SecretId {
@@ -42,8 +71,8 @@ impl Serialize for SecretId {
where
S: Serializer,
{
- // calls Display at one point
- ser.serialize_str(&self.to_string())
+ // calls LowerHex at one point
+ ser.serialize_str(&format!("{self:#x}"))
}
}
@@ -56,12 +85,36 @@ impl<'de> Deserialize<'de> for SecretId {
}
}
+impl UpperHex for SecretId {
+ fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+ if f.alternate() {
+ write!(f, "0x")?;
+ }
+ for b in self.0 {
+ write!(f, "{b:02X}")?;
+ }
+ Ok(())
+ }
+}
+
+impl LowerHex for SecretId {
+ fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+ if f.alternate() {
+ write!(f, "0x")?;
+ }
+ for b in self.0 {
+ write!(f, "{b:02x}")?;
+ }
+ Ok(())
+ }
+}
+
impl Display for SecretId {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
- let mut s = String::with_capacity(32 * 2 + 2);
- s.push_str("0x");
- let s = self.0.iter().fold(s, |acc, e| acc + &format!("{e:02x}"));
- write!(f, "{s}")
+ if let Some(s) = self.as_ascii() {
+ write!(f, "{s} | ")?;
+ }
+ write!(f, "{self:#x}")
}
}
@@ -79,7 +132,7 @@ impl AsRef<[u8]> for SecretId {
/// A secret in a [`SecretList`]
#[repr(C)]
-#[derive(Debug, PartialEq, Eq, AsBytes, FromZeroes, FromBytes, Serialize)]
+#[derive(Debug, Clone, PartialEq, Eq, AsBytes, FromZeroes, FromBytes, Serialize)]
pub struct SecretEntry {
#[serde(serialize_with = "ser_u16")]
index: U16<BigEndian>,
@@ -153,11 +206,7 @@ impl Display for SecretEntry {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
let stype: ListableSecretType = self.stype.get().into();
writeln!(f, "{} {}:", self.index, stype)?;
- write!(f, " ")?;
- for b in self.id.as_ref() {
- write!(f, "{b:02x}")?;
- }
- Ok(())
+ write!(f, " {}", self.id)
}
}
@@ -269,6 +318,11 @@ impl SecretList {
self.hdr.total_num_secrets.get() as usize
}
+ /// Find the first [`SecretEntry`] that has the provided [`SecretId`]
+ pub fn find(&self, id: &SecretId) -> Option<SecretEntry> {
+ self.iter().find(|e| e.id() == id.as_ref()).cloned()
+ }
+
/// Encodes the list in the same binary format the UV would do
pub fn encode<T: Write>(&self, w: &mut T) -> Result<()> {
w.write_all(self.hdr.as_bytes())?;
@@ -456,7 +510,7 @@ where
type Value = [u8; SecretId::ID_SIZE];
fn expecting(&self, formatter: &mut std::fmt::Formatter) -> std::fmt::Result {
- formatter.write_str("a `32 bytes long hexstring` prepended with 0x")
+ formatter.write_str("a `32 bytes (=64 character) long hexstring` prepended with 0x")
}
fn visit_str<E>(self, s: &str) -> Result<Self::Value, E>
@@ -464,7 +518,10 @@ where
E: serde::de::Error,
{
if s.len() != SecretId::ID_SIZE * 2 + "0x".len() {
- return Err(serde::de::Error::invalid_length(s.len() - 2, &self));
+ return Err(serde::de::Error::invalid_length(
+ s.len().saturating_sub("0x".len()),
+ &self,
+ ));
}
let nb = s.strip_prefix("0x").ok_or_else(|| {
serde::de::Error::invalid_value(serde::de::Unexpected::Str(s), &self)
@@ -655,4 +712,80 @@ mod test {
],
)
}
+
+ #[test]
+ fn secret_id_display() {
+ let text = "Fancy secret ID";
+ let id = SecretId::from_string(text);
+
+ let exp =
+ "Fancy secret ID | 0x46616e6379207365637265742049440000000000000000000000000000000000";
+ assert_eq!(id.to_string(), exp);
+ }
+
+ #[test]
+ fn secret_id_long_name() {
+ let text = "the most fanciest secret ID you ever seen in the time the universe exists";
+ let id = SecretId::from_string(text);
+ let exp =
+ "the most fanciest secret ID you | 0x746865206d6f73742066616e63696573742073656372657420494420796f7500";
+ assert_eq!(id.to_string(), exp);
+ }
+
+ #[test]
+ fn secret_id_no_ascii_name() {
+ let text = [0; 32];
+ let id = SecretId::from(text);
+
+ let exp = "0x0000000000000000000000000000000000000000000000000000000000000000";
+ assert_eq!(id.to_string(), exp);
+ }
+
+ #[test]
+ fn secret_id_no_ascii_name2() {
+ let text = [
+ 0x25, 0x55, 3, 4, 50, 0, 6, 0, 8, 0, 0, 0, 0, 0, 0, 0, 90, 0, 0xa, 0, 0, 0, 0, 0xf, 0,
+ 0, 0, 0, 0, 0, 0, 0,
+ ];
+ let id = SecretId::from(text);
+ assert_eq!(id.as_ascii(), None);
+ }
+
+ #[test]
+ fn secret_id_no_ascii_name3() {
+ let text = [
+ 0x25, 0x55, 0, 4, 50, 0, 6, 0, 8, 0, 0, 0, 0, 0, 0, 0, 90, 0, 0xa, 0, 0, 0, 0, 0xf, 0,
+ 0, 0, 0, 0, 0, 0, 0,
+ ];
+ let id = SecretId::from(text);
+ assert_eq!(id.as_ascii(), None);
+ }
+
+ #[test]
+ fn secret_id_hex() {
+ let id_str = "Nice Test 123";
+ let id = SecretId::from_string(id_str);
+
+ let s = format!("{id:#x}");
+ assert_eq!(
+ s,
+ "0x4e69636520546573742031323300000000000000000000000000000000000000"
+ );
+ let s = format!("{id:x}");
+ assert_eq!(
+ s,
+ "4e69636520546573742031323300000000000000000000000000000000000000"
+ );
+ let s = format!("{id:#X}");
+ assert_eq!(
+ s,
+ "0x4E69636520546573742031323300000000000000000000000000000000000000"
+ );
+
+ let s = format!("{id:X}");
+ assert_eq!(
+ s,
+ "4E69636520546573742031323300000000000000000000000000000000000000"
+ );
+ }
}
diff --git a/rust/pvsecret/src/cli.rs b/rust/pvsecret/src/cli.rs
index 4e747682..d858fc29 100644
--- a/rust/pvsecret/src/cli.rs
+++ b/rust/pvsecret/src/cli.rs
@@ -141,6 +141,12 @@ pub struct CreateSecretOpt {
/// by default.
#[arg(long, value_name = "FILE", value_hint = ValueHint::FilePath,)]
pub user_sign_key: Option<String>,
+
+ /// Do not hash the name, use it directly as secret ID.
+ ///
+ /// Ignored for meta-secrets.
+ #[arg(long)]
+ pub use_name: bool,
}
#[derive(Subcommand, Debug)]
@@ -342,6 +348,8 @@ pub enum RetrInpFmt {
Yaml,
/// Use a hex string.
Hex,
+ /// Use a name-string. Will hash it if no secret with the name found.
+ Name,
}
#[derive(Copy, Clone, PartialEq, Eq, PartialOrd, Ord, ValueEnum, Debug, Default)]
diff --git a/rust/pvsecret/src/cmd/create.rs b/rust/pvsecret/src/cmd/create.rs
index 73089a12..fab37e67 100644
--- a/rust/pvsecret/src/cmd/create.rs
+++ b/rust/pvsecret/src/cmd/create.rs
@@ -94,7 +94,7 @@ fn read_private_key(buf: &[u8]) -> Result<PKey<Private>> {
fn build_asrcb(opt: &CreateSecretOpt) -> Result<AddSecretRequest> {
debug!("Build add-secret request");
- let secret = match &opt.secret {
+ let mut secret = match &opt.secret {
AddSecretType::Meta => GuestSecret::Null,
AddSecretType::Association {
name,
@@ -112,6 +112,8 @@ fn build_asrcb(opt: &CreateSecretOpt) -> Result<AddSecretRequest> {
};
trace!("AddSecret: {secret:x?}");
+ opt.use_name.then(|| secret.no_hash_name());
+
let mut flags = match &opt.pcf {
Some(v) => (&try_parse_u64(v, "pcf")?).into(),
None => AddSecretFlags::default(),
diff --git a/rust/pvsecret/src/cmd/retr.rs b/rust/pvsecret/src/cmd/retr.rs
index 7f7704cc..ad3e91c3 100644
--- a/rust/pvsecret/src/cmd/retr.rs
+++ b/rust/pvsecret/src/cmd/retr.rs
@@ -17,12 +17,17 @@ use utils::get_writer_from_cli_file_arg;
fn retrieve(id: &SecretId) -> Result<RetrievedSecret> {
let uv = UvDevice::open()?;
let secrets = list_uvc(&uv)?;
- let secret = secrets
- .into_iter()
- .find(|s| s.id() == id.as_ref())
+ let secret = match secrets.find(id) {
+ Some(s) => s,
+ // hash it + try again if it is ASCII-representable
+ None => match id.as_ascii() {
+ Some(s) => secrets.find(&GuestSecret::name_to_id(s)?),
+ None => None,
+ }
.ok_or(anyhow!(
"The UV secret-store has no secret with the ID {id}"
- ))?;
+ ))?,
+ };
info!("Try to retrieve secret at index: {}", secret.index());
debug!("Try to retrieve: {secret:?}");
@@ -43,6 +48,7 @@ pub fn retr(opt: &RetrSecretOptions) -> Result<()> {
RetrInpFmt::Hex => {
serde_yaml::from_str(&opt.input).context("Cannot parse SecretId information")?
}
+ RetrInpFmt::Name => SecretId::from_string(&opt.input),
};
let retr_secret =

File diff suppressed because it is too large Load Diff

View File

@ -0,0 +1,334 @@
From cf51ac786095f2a1a17d04fea9ee73271438d247 Mon Sep 17 00:00:00 2001
From: Marc Hartmayer <mhartmay@linux.ibm.com>
Date: Wed, 11 Dec 2024 19:25:59 +0100
Subject: [PATCH] rust/pvimg: Add '--(enable|disable)-image-encryption' flags
to 'pvimg create'
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
With runtime attestation it might be useful to have non-encrypted Secure
Execution images. This patch adds the support for this to the 'pvimg
create' and 'genprotimg' commands.
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Acked-by: Hendrik Brueckner <brueckner@linux.ibm.com>
Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
---
rust/pvimg/man/genprotimg.1 | 26 +++++++++++++++++++++-----
rust/pvimg/man/pvimg-create.1 | 26 +++++++++++++++++++++-----
rust/pvimg/man/pvimg-info.1 | 10 +++++-----
rust/pvimg/man/pvimg-test.1 | 10 +++++-----
rust/pvimg/man/pvimg.1 | 10 +++++-----
rust/pvimg/src/cli.rs | 18 ++++++++++++++++++
rust/pvimg/src/cmd/create.rs | 10 ++++++++++
7 files changed, 85 insertions(+), 25 deletions(-)
diff --git a/rust/pvimg/man/genprotimg.1 b/rust/pvimg/man/genprotimg.1
index 46a91aa4..3f4949e9 100644
--- a/rust/pvimg/man/genprotimg.1
+++ b/rust/pvimg/man/genprotimg.1
@@ -3,11 +3,11 @@
.\" it under the terms of the MIT license. See LICENSE for details.
.\"
-.TH genprotimg 1 "2024-12-05" "s390-tools" "Genprotimg Manual"
+.TH genprotimg 1 "2024-12-11" "s390-tools" "Genprotimg Manual"
.nh
.ad l
.SH NAME
-\fBgenprotimg\fP - Create an IBM Secure Execution image
+\fBgenprotimg\fP \- Create an IBM Secure Execution image
\fB
.SH SYNOPSIS
.nf
@@ -196,6 +196,22 @@ Disable the support for backup target keys (default).
.RE
.RE
.PP
+\-\-enable\-image\-encryption
+.RS 4
+Enable encryption of the image components (default). The image components are:
+the kernel, ramdisk, and kernel command line.
+.RE
+.RE
+.PP
+\-\-disable\-image\-encryption
+.RS 4
+Disable encryption of the image components. The image components are: the
+kernel, ramdisk, and kernel command line. Use only if the components used do not
+contain any confidential content (for example, secrets like non\-public
+cryptographic keys).
+.RE
+.RE
+.PP
\-v, \-\-verbose
.RS 4
Provide more detailed output.
@@ -222,16 +238,16 @@ Print help (see a summary with \fB\-h\fR).
.SH EXIT STATUS
.TP 8
-.B 0 - Program finished successfully
+.B 0 \- Program finished successfully
The command was executed successfully.
.RE
.TP 8
-.B 1 - Generic error
+.B 1 \- Generic error
Something went wrong during the operation. Refer to the error
message.
.RE
.TP 8
-.B 2 - Usage error
+.B 2 \- Usage error
The command was used incorrectly, for example: unsupported command
line flag, or wrong number of arguments.
.RE
diff --git a/rust/pvimg/man/pvimg-create.1 b/rust/pvimg/man/pvimg-create.1
index aba197fa..dae1cf18 100644
--- a/rust/pvimg/man/pvimg-create.1
+++ b/rust/pvimg/man/pvimg-create.1
@@ -3,11 +3,11 @@
.\" it under the terms of the MIT license. See LICENSE for details.
.\"
-.TH pvimg-create 1 "2024-12-05" "s390-tools" "Pvimg Manual"
+.TH pvimg-create 1 "2024-12-11" "s390-tools" "Pvimg Manual"
.nh
.ad l
.SH NAME
-\fBpvimg create\fP - Create an IBM Secure Execution image
+\fBpvimg create\fP \- Create an IBM Secure Execution image
\fB
.SH SYNOPSIS
.nf
@@ -195,6 +195,22 @@ Disable the support for backup target keys (default).
.RE
.RE
.PP
+\-\-enable\-image\-encryption
+.RS 4
+Enable encryption of the image components (default). The image components are:
+the kernel, ramdisk, and kernel command line.
+.RE
+.RE
+.PP
+\-\-disable\-image\-encryption
+.RS 4
+Disable encryption of the image components. The image components are: the
+kernel, ramdisk, and kernel command line. Use only if the components used do not
+contain any confidential content (for example, secrets like non\-public
+cryptographic keys).
+.RE
+.RE
+.PP
\-h, \-\-help
.RS 4
Print help (see a summary with \fB\-h\fR).
@@ -203,16 +219,16 @@ Print help (see a summary with \fB\-h\fR).
.SH EXIT STATUS
.TP 8
-.B 0 - Program finished successfully
+.B 0 \- Program finished successfully
The command was executed successfully.
.RE
.TP 8
-.B 1 - Generic error
+.B 1 \- Generic error
Something went wrong during the operation. Refer to the error
message.
.RE
.TP 8
-.B 2 - Usage error
+.B 2 \- Usage error
The command was used incorrectly, for example: unsupported command
line flag, or wrong number of arguments.
.RE
diff --git a/rust/pvimg/man/pvimg-info.1 b/rust/pvimg/man/pvimg-info.1
index e88cbe49..d2726c35 100644
--- a/rust/pvimg/man/pvimg-info.1
+++ b/rust/pvimg/man/pvimg-info.1
@@ -3,11 +3,11 @@
.\" it under the terms of the MIT license. See LICENSE for details.
.\"
-.TH pvimg-info 1 "2024-12-05" "s390-tools" "Pvimg Manual"
+.TH pvimg-info 1 "2024-12-11" "s390-tools" "Pvimg Manual"
.nh
.ad l
.SH NAME
-\fBpvimg info\fP - Print information about the IBM Secure Execution image
+\fBpvimg info\fP \- Print information about the IBM Secure Execution image
\fB
.SH SYNOPSIS
.nf
@@ -51,16 +51,16 @@ Print help (see a summary with \fB\-h\fR).
.SH EXIT STATUS
.TP 8
-.B 0 - Program finished successfully
+.B 0 \- Program finished successfully
The command was executed successfully.
.RE
.TP 8
-.B 1 - Generic error
+.B 1 \- Generic error
Something went wrong during the operation. Refer to the error
message.
.RE
.TP 8
-.B 2 - Usage error
+.B 2 \- Usage error
The command was used incorrectly, for example: unsupported command
line flag, or wrong number of arguments.
.RE
diff --git a/rust/pvimg/man/pvimg-test.1 b/rust/pvimg/man/pvimg-test.1
index 901c7edb..4fb7d73f 100644
--- a/rust/pvimg/man/pvimg-test.1
+++ b/rust/pvimg/man/pvimg-test.1
@@ -3,11 +3,11 @@
.\" it under the terms of the MIT license. See LICENSE for details.
.\"
-.TH pvimg-test 1 "2024-12-05" "s390-tools" "Pvimg Manual"
+.TH pvimg-test 1 "2024-12-11" "s390-tools" "Pvimg Manual"
.nh
.ad l
.SH NAME
-\fBpvimg test\fP - Test different aspects of an existing IBM Secure Execution image
+\fBpvimg test\fP \- Test different aspects of an existing IBM Secure Execution image
\fB
.SH SYNOPSIS
.nf
@@ -54,16 +54,16 @@ Print help (see a summary with \fB\-h\fR).
.SH EXIT STATUS
.TP 8
-.B 0 - Program finished successfully
+.B 0 \- Program finished successfully
The command was executed successfully.
.RE
.TP 8
-.B 1 - Generic error
+.B 1 \- Generic error
Something went wrong during the operation. Refer to the error
message.
.RE
.TP 8
-.B 2 - Usage error
+.B 2 \- Usage error
The command was used incorrectly, for example: unsupported command
line flag, or wrong number of arguments.
.RE
diff --git a/rust/pvimg/man/pvimg.1 b/rust/pvimg/man/pvimg.1
index 37c8e978..5676b61d 100644
--- a/rust/pvimg/man/pvimg.1
+++ b/rust/pvimg/man/pvimg.1
@@ -3,11 +3,11 @@
.\" it under the terms of the MIT license. See LICENSE for details.
.\"
-.TH pvimg 1 "2024-12-05" "s390-tools" "Pvimg Manual"
+.TH pvimg 1 "2024-12-11" "s390-tools" "Pvimg Manual"
.nh
.ad l
.SH NAME
-\fBpvimg\fP - Create and inspect IBM Secure Execution images
+\fBpvimg\fP \- Create and inspect IBM Secure Execution images
\fB
.SH SYNOPSIS
.nf
@@ -69,16 +69,16 @@ Print help (see a summary with \fB\-h\fR).
.SH EXIT STATUS
.TP 8
-.B 0 - Program finished successfully
+.B 0 \- Program finished successfully
The command was executed successfully.
.RE
.TP 8
-.B 1 - Generic error
+.B 1 \- Generic error
Something went wrong during the operation. Refer to the error
message.
.RE
.TP 8
-.B 2 - Usage error
+.B 2 \- Usage error
The command was used incorrectly, for example: unsupported command
line flag, or wrong number of arguments.
.RE
diff --git a/rust/pvimg/src/cli.rs b/rust/pvimg/src/cli.rs
index 2ca4e901..12f0b764 100644
--- a/rust/pvimg/src/cli.rs
+++ b/rust/pvimg/src/cli.rs
@@ -140,6 +140,20 @@ pub struct CreateBootImageLegacyFlags {
/// Disable the support for backup target keys (default).
#[arg(long, action = clap::ArgAction::SetTrue, conflicts_with="enable_backup_keys", group="header-flags")]
pub disable_backup_keys: Option<bool>,
+
+ /// Enable encryption of the image components (default).
+ ///
+ /// The image components are: the kernel, ramdisk, and kernel command line.
+ #[arg(long, action = clap::ArgAction::SetTrue, group="header-flags")]
+ pub enable_image_encryption: Option<bool>,
+
+ /// Disable encryption of the image components.
+ ///
+ /// The image components are: the kernel, ramdisk, and kernel command line.
+ /// Use only if the components used do not contain any confidential content
+ /// (for example, secrets like non-public cryptographic keys).
+ #[arg(long, action = clap::ArgAction::SetTrue, conflicts_with="enable_image_encryption", group="header-flags")]
+ pub disable_image_encryption: Option<bool>,
}
#[non_exhaustive]
@@ -476,6 +490,8 @@ mod test {
flat_map_collect(insert(mvca.clone(), vec![CliOption::new("enable-pckmo", ["--enable-pckmo"])])),
flat_map_collect(insert(mvca.clone(), vec![CliOption::new("enable-pckmo-hmac", ["--enable-pckmo-hmac"])])),
flat_map_collect(insert(mvca.clone(), vec![CliOption::new("enable-backup-keys", ["--enable-backup-keys"])])),
+ flat_map_collect(insert(mvca.clone(), vec![CliOption::new("disable-image-encryption", ["--disable-image-encryption"])])),
+ flat_map_collect(insert(mvca.clone(), vec![CliOption::new("enable-image-encryption", ["--enable-image-encryption"])])),
];
let invalid_create_args = [
flat_map_collect(remove(mvcanv.clone(), "no-verify")),
@@ -501,6 +517,8 @@ mod test {
CliOption::new("x-pcf2", ["--x-pcf", "0x0"])])),
flat_map_collect(insert(mvca.clone(), vec![CliOption::new("enable-pckmo", ["--enable-pckmo"]),
CliOption::new("disable-pckmo", ["--disable-pckmo"])])),
+ flat_map_collect(insert(mvca.clone(), vec![CliOption::new("enable-image-encryption", ["--enable-image-encryption"]),
+ CliOption::new("disable-image-encryption", ["--disable-image-encryption"])])),
];
let mut genprotimg_valid_args = vec![
diff --git a/rust/pvimg/src/cmd/create.rs b/rust/pvimg/src/cmd/create.rs
index b696d790..475d3523 100644
--- a/rust/pvimg/src/cmd/create.rs
+++ b/rust/pvimg/src/cmd/create.rs
@@ -80,6 +80,12 @@ fn parse_flags(
lf.enable_backup_keys
.filter(|x| *x)
.and(Some(PcfV1::all_enabled([PcfV1::BackupTargetKeys]))),
+ lf.disable_image_encryption
+ .filter(|x| *x)
+ .and(Some(PcfV1::all_enabled([PcfV1::NoComponentEncryption]))),
+ lf.enable_image_encryption
+ .filter(|x| *x)
+ .and(Some(PcfV1::all_disabled([PcfV1::NoComponentEncryption]))),
]
.into_iter()
.flatten()
@@ -135,6 +141,10 @@ pub fn create(opt: &CreateBootImageArgs) -> Result<OwnExitCode> {
read_user_provided_keys(opt.comm_key.as_deref(), &opt.experimental_args)?;
let (plaintext_flags, secret_flags) = parse_flags(opt)?;
+ if plaintext_flags.is_set(PcfV1::NoComponentEncryption) {
+ warn!("The components encryption is disabled, make sure that the components do not contain any confidential content.");
+ }
+
let mut components = components(&opt.component_paths)?;
if opt.no_component_check {
warn!("The component check is turned off!");

View File

@ -0,0 +1,167 @@
From 5b6d7a467dc342c9c25a0af72b2d5546798cdc94 Mon Sep 17 00:00:00 2001
From: Marc Hartmayer <mhartmay@linux.ibm.com>
Date: Thu, 12 Dec 2024 20:19:56 +0100
Subject: [PATCH] rust/pvimg: Add '--cck <FILE>' command line option and make
'--comm-key' an alias
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Add '--cck <FILE>' as an command line option and make '--comm-key' an
alias of it. This makes the command line more similar to the other
Secure Execution related PV-tools (e.g. pvattest and pvsecret).
Suggested-by: Reinhard Bündgen <buendgen@de.ibm.com>
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
---
rust/pvimg/man/genprotimg.1 | 11 +++++------
rust/pvimg/man/pvimg-create.1 | 11 +++++------
rust/pvimg/src/cli.rs | 14 ++++++++------
rust/pvimg/src/cmd/create.rs | 3 +--
4 files changed, 19 insertions(+), 20 deletions(-)
Index: s390-tools-2.36.0/rust/pvimg/man/genprotimg.1
===================================================================
--- s390-tools-2.36.0.orig/rust/pvimg/man/genprotimg.1
+++ s390-tools-2.36.0/rust/pvimg/man/genprotimg.1
@@ -123,7 +123,7 @@ Overwrite an existing Secure Execution b
.RE
.RE
.PP
-\-\-comm\-key <FILE>
+\-\-cck, \-\-comm\-key <FILE>
.RS 4
Use the content of FILE as the customer\-communication key (CCK). The file must
contain exactly 32 bytes of data.
@@ -133,7 +133,7 @@ contain exactly 32 bytes of data.
\-\-enable\-dump
.RS 4
Enable Secure Execution guest dump support. This option requires the
-\fB\-\-comm\-key\fR option.
+\fB\-\-cck\fR option.
.RE
.RE
.PP
@@ -146,8 +146,7 @@ Disable Secure Execution guest dump supp
\-\-enable\-cck\-extension\-secret
.RS 4
Add\-secret requests must provide an extension secret that matches the
-CCK\-derived extension secret. This option requires the \fB\-\-comm\-key\fR
-option.
+CCK\-derived extension secret. This option requires the \fB\-\-cck\fR option.
.RE
.RE
.PP
@@ -268,7 +267,7 @@ Generate an IBM Secure Execution image:
Generate an IBM Secure Execution image with Secure Execution guest dump support:
.PP
-.B genprotimg \-i \fI\,/boot/vmlinuz\/\fR \-r \fI\,/boot/initrd.img\/\fR \-p \fI\,parmfile\/\fR \-k \fI\,host_key.crt\/\fR \-C \fI\,ibm-z-host-key-signing.crt\/\fR \-C \fI\,DigiCertCA.crt\fR \-o \fI\,/boot/secure-linux\/\fR \-\-enable\-dump \-\-comm\-key \fI\,comm-key\fR
+.B genprotimg \-i \fI\,/boot/vmlinuz\/\fR \-r \fI\,/boot/initrd.img\/\fR \-p \fI\,parmfile\/\fR \-k \fI\,host_key.crt\/\fR \-C \fI\,ibm-z-host-key-signing.crt\/\fR \-C \fI\,DigiCertCA.crt\fR \-o \fI\,/boot/secure-linux\/\fR \-\-enable\-dump \-\-cck \fI\,comm-key\fR
.SH NOTES
.IP "1." 4
The \fBgenprotimg\fR(1) command is a symbolic link to the \fBpvimg-create\fR(1) command.
Index: s390-tools-2.36.0/rust/pvimg/man/pvimg-create.1
===================================================================
--- s390-tools-2.36.0.orig/rust/pvimg/man/pvimg-create.1
+++ s390-tools-2.36.0/rust/pvimg/man/pvimg-create.1
@@ -122,7 +122,7 @@ Overwrite an existing Secure Execution b
.RE
.RE
.PP
-\-\-comm\-key <FILE>
+\-\-cck, \-\-comm\-key <FILE>
.RS 4
Use the content of FILE as the customer\-communication key (CCK). The file must
contain exactly 32 bytes of data.
@@ -132,7 +132,7 @@ contain exactly 32 bytes of data.
\-\-enable\-dump
.RS 4
Enable Secure Execution guest dump support. This option requires the
-\fB\-\-comm\-key\fR option.
+\fB\-\-cck\fR option.
.RE
.RE
.PP
@@ -145,8 +145,7 @@ Disable Secure Execution guest dump supp
\-\-enable\-cck\-extension\-secret
.RS 4
Add\-secret requests must provide an extension secret that matches the
-CCK\-derived extension secret. This option requires the \fB\-\-comm\-key\fR
-option.
+CCK\-derived extension secret. This option requires the \fB\-\-cck\fR option.
.RE
.RE
.PP
@@ -249,7 +248,7 @@ Generate an IBM Secure Execution image:
Generate an IBM Secure Execution image with Secure Execution guest dump support:
.PP
-.B pvimg create \-i \fI\,/boot/vmlinuz\/\fR \-r \fI\,/boot/initrd.img\/\fR \-p \fI\,parmfile\/\fR \-k \fI\,host_key.crt\/\fR \-C \fI\,ibm-z-host-key-signing.crt\/\fR \-C \fI\,DigiCertCA.crt\fR \-o \fI\,/boot/secure-linux\/\fR \-\-enable\-dump \-\-comm\-key \fI\,comm-key\fR
+.B pvimg create \-i \fI\,/boot/vmlinuz\/\fR \-r \fI\,/boot/initrd.img\/\fR \-p \fI\,parmfile\/\fR \-k \fI\,host_key.crt\/\fR \-C \fI\,ibm-z-host-key-signing.crt\/\fR \-C \fI\,DigiCertCA.crt\fR \-o \fI\,/boot/secure-linux\/\fR \-\-enable\-dump \-\-cck \fI\,comm-key\fR
.SH NOTES
.IP "1." 4
The \fBgenprotimg\fR(1) command is a symbolic link to the \fBpvimg-create\fR(1) command.
Index: s390-tools-2.36.0/rust/pvimg/src/cli.rs
===================================================================
--- s390-tools-2.36.0.orig/rust/pvimg/src/cli.rs
+++ s390-tools-2.36.0/rust/pvimg/src/cli.rs
@@ -96,8 +96,8 @@ pub struct ComponentPaths {
#[command(group(ArgGroup::new("header-flags").multiple(true).conflicts_with_all(["x_pcf", "x_scf"])))]
pub struct CreateBootImageLegacyFlags {
/// Enable Secure Execution guest dump support. This option requires the
- /// '--comm-key' option.
- #[arg(long, action = clap::ArgAction::SetTrue, requires="comm_key", group="header-flags")]
+ /// '--cck' option.
+ #[arg(long, action = clap::ArgAction::SetTrue, requires="cck", group="header-flags")]
pub enable_dump: Option<bool>,
/// Disable Secure Execution guest dump support (default).
@@ -105,9 +105,9 @@ pub struct CreateBootImageLegacyFlags {
pub disable_dump: Option<bool>,
/// Add-secret requests must provide an extension secret that matches the
- /// CCK-derived extension secret. This option requires the '--comm-key'
+ /// CCK-derived extension secret. This option requires the '--cck'
/// option.
- #[arg(long, action = clap::ArgAction::SetTrue, requires="comm_key", group="header-flags")]
+ #[arg(long, action = clap::ArgAction::SetTrue, requires="cck", group="header-flags")]
pub enable_cck_extension_secret: Option<bool>,
/// Add-secret requests don't have to provide the CCK-derived extension
@@ -328,8 +328,8 @@ pub struct CreateBootImageArgs {
/// Use the content of FILE as the customer-communication key (CCK).
///
/// The file must contain exactly 32 bytes of data.
- #[arg(long, value_name = "FILE")]
- pub comm_key: Option<PathBuf>,
+ #[arg(long, value_name = "FILE", visible_alias = "comm-key")]
+ pub cck: Option<PathBuf>,
#[clap(flatten)]
pub legacy_flags: CreateBootImageLegacyFlags,
@@ -482,6 +482,8 @@ mod test {
flat_map_collect(insert(mvca.clone(), vec![CliOption::new("enable-dump", ["--enable-dump"]),
CliOption::new("comm-key", ["--comm-key", "/dev/null"])])),
flat_map_collect(insert(mvca.clone(), vec![CliOption::new("enable-dump", ["--enable-dump"]),
+ CliOption::new("comm-key", ["--cck", "/dev/null"])])),
+ flat_map_collect(insert(mvca.clone(), vec![CliOption::new("enable-dump", ["--enable-dump"]),
CliOption::new("comm-key", ["--comm-key", "/dev/null"])])),
flat_map_collect(insert(mvca.clone(), vec![CliOption::new("x-pcf", ["--x-pcf", "0x0"]),
CliOption::new("x-scf", ["--x-scf", "0x0"])])),
Index: s390-tools-2.36.0/rust/pvimg/src/cmd/create.rs
===================================================================
--- s390-tools-2.36.0.orig/rust/pvimg/src/cmd/create.rs
+++ s390-tools-2.36.0/rust/pvimg/src/cmd/create.rs
@@ -137,8 +137,7 @@ pub fn create(opt: &CreateBootImageArgs)
let verified_host_keys = opt
.certificate_args
.get_verified_hkds("Secure Execution image")?;
- let user_provided_keys =
- read_user_provided_keys(opt.comm_key.as_deref(), &opt.experimental_args)?;
+ let user_provided_keys = read_user_provided_keys(opt.cck.as_deref(), &opt.experimental_args)?;
let (plaintext_flags, secret_flags) = parse_flags(opt)?;
if plaintext_flags.is_set(PcfV1::NoComponentEncryption) {

View File

@ -0,0 +1,58 @@
From 560b276f7e9938475af921c8ebd4cd05910dbf31 Mon Sep 17 00:00:00 2001
From: Marc Hartmayer <mhartmay@linux.ibm.com>
Date: Fri, 6 Dec 2024 20:45:36 +0100
Subject: [PATCH] rust/pvimg: Fix possible 'range start index out of range for
slice' error
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Fix possible 'range start index 16 out of range for slice of length 0'
error by adding a check of the slice data length.
Fixes: f4cf4ae6ebb1 ("rust: Add a new tool called 'pvimg'")
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
---
rust/pvimg/src/pv_utils/se_hdr/brb.rs | 23 +++++++++++++++++++++++
1 file changed, 23 insertions(+)
diff --git a/rust/pvimg/src/pv_utils/se_hdr/brb.rs b/rust/pvimg/src/pv_utils/se_hdr/brb.rs
index f7ae1bc9..ac3a2e6e 100644
--- a/rust/pvimg/src/pv_utils/se_hdr/brb.rs
+++ b/rust/pvimg/src/pv_utils/se_hdr/brb.rs
@@ -259,6 +259,10 @@ impl SeHdr {
return Err(Error::InvalidSeHdr);
}
+ if sehs <= common_size {
+ return Err(Error::InvalidSeHdr);
+ }
+
data.resize(sehs, 0);
reader.read_exact(&mut data[common_size..])?;
Self::try_from_data(&data)
@@ -366,3 +370,22 @@ impl AeadCipherTrait for SeHdrPlain {
self.data.aead_tag_size()
}
}
+
+#[cfg(test)]
+mod tests {
+ use std::io::Cursor;
+
+ use super::SeHdr;
+ use crate::error::Error;
+
+ #[test]
+ fn test_sehdr_try_from_io() {
+ // Invalid SeHdr as `sehs` is set to 0
+ assert!(matches!(
+ SeHdr::try_from_io(Cursor::new([
+ 73, 66, 77, 83, 101, 99, 69, 120, 0, 0, 1, 0, 0, 0, 0, 0, 2, 0, 8
+ ])),
+ Err(Error::InvalidSeHdr)
+ ));
+ }
+}

View File

@ -0,0 +1,51 @@
From 3f6572e901ddcc654021c4302cb2a99999acb87a Mon Sep 17 00:00:00 2001
From: Marc Hartmayer <mhartmay@linux.ibm.com>
Date: Wed, 18 Dec 2024 13:41:13 +0100
Subject: [PATCH] rust/utils: mkdtemp: fix memory leak
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Fix memory leak of @template_raw. The documentation of CString::into_raw
reads:
"Consumes the CString and transfers ownership of the string to a C
caller.
...
Failure to call CString::from_raw will lead to a memory leak." [1]
Let's fix the memory leak by always calling `CString::from_raw` and
therefore reclaim the ownership.
[1] https://doc.rust-lang.org/std/ffi/struct.CString.html#method.into_raw
Fixes: e56acf4f14b0 ("pv_core: add `TemporaryDirectory`")
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
---
rust/utils/src/tmpfile.rs | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/rust/utils/src/tmpfile.rs b/rust/utils/src/tmpfile.rs
index 07acdba8..883d5586 100644
--- a/rust/utils/src/tmpfile.rs
+++ b/rust/utils/src/tmpfile.rs
@@ -16,13 +16,14 @@ fn mkdtemp<P: AsRef<Path>>(template: P) -> Result<PathBuf, std::io::Error> {
// SAFETY: template_raw is a valid CString because it was generated by
// the `CString::new`.
let ret = libc::mkdtemp(template_raw);
+ // SAFETY: `template_raw` is still a valid CString because it was
+ // generated by `CString::new` and modified by `libc::mkdtemp`.
+ let path_cstr = std::ffi::CString::from_raw(template_raw);
if ret.is_null() {
+ drop(path_cstr);
Err(std::io::Error::last_os_error())
} else {
- // SAFETY: `template_raw` is still a valid CString because it was
- // generated by `CString::new` and modified by `libc::mkdtemp`.
- let path_cstr = std::ffi::CString::from_raw(template_raw);
let path = OsStr::from_bytes(path_cstr.as_bytes());
let path = std::path::PathBuf::from(path);

View File

@ -0,0 +1,334 @@
From 944581eaefe4c6887790f2b8ed39c9ee76146c55 Mon Sep 17 00:00:00 2001
From: Marc Hartmayer <mhartmay@linux.ibm.com>
Date: Tue, 17 Dec 2024 11:58:01 +0100
Subject: [PATCH] rust/pvimg: Add upper estimates for the Secure Execution
header
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
A Secure Execution header V1 can be at maximum two pages large, optional
items are not supported, and the size of the encrypted part cannot be
larger than the total size of the Secure Execution header add this as
Deku assertions and additional conditions to the code. In addition, add
a check for the number of key slots.
Fixes: f4cf4ae6ebb1 ("rust: Add a new tool called 'pvimg'")
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
---
rust/pvimg/src/pv_utils/error.rs | 3 +
rust/pvimg/src/pv_utils/se_hdr/brb.rs | 50 +++++++++++++---
rust/pvimg/src/pv_utils/se_hdr/builder.rs | 10 +++-
rust/pvimg/src/pv_utils/se_hdr/hdr_v1.rs | 71 ++++++++++++++++++++---
rust/pvimg/src/pv_utils/uvdata.rs | 18 ++++--
5 files changed, 130 insertions(+), 22 deletions(-)
diff --git a/rust/pvimg/src/pv_utils/error.rs b/rust/pvimg/src/pv_utils/error.rs
index 2a176276..a12c4a22 100644
--- a/rust/pvimg/src/pv_utils/error.rs
+++ b/rust/pvimg/src/pv_utils/error.rs
@@ -30,6 +30,9 @@ pub enum Error {
#[error("Invalid Secure Execution header")]
InvalidSeHdr,
+ #[error("Secure Execution header size {given} is larger than the maximum of {maximum} bytes")]
+ InvalidSeHdrTooLarge { given: usize, maximum: usize },
+
#[error("Invalid component metadata.")]
InvalidComponentMetadata,
diff --git a/rust/pvimg/src/pv_utils/se_hdr/brb.rs b/rust/pvimg/src/pv_utils/se_hdr/brb.rs
index ac3a2e6e..b8dadba1 100644
--- a/rust/pvimg/src/pv_utils/se_hdr/brb.rs
+++ b/rust/pvimg/src/pv_utils/se_hdr/brb.rs
@@ -171,8 +171,8 @@ impl AeadCipherTrait for SeHdr {
}
impl AeadDataTrait for SeHdr {
- fn aad(&self) -> Vec<u8> {
- [serialize_to_bytes(&self.common).unwrap(), self.data.aad()].concat()
+ fn aad(&self) -> Result<Vec<u8>> {
+ Ok([serialize_to_bytes(&self.common)?, self.data.aad()?].concat())
}
fn data(&self) -> Vec<u8> {
@@ -265,7 +265,7 @@ impl SeHdr {
data.resize(sehs, 0);
reader.read_exact(&mut data[common_size..])?;
- Self::try_from_data(&data)
+ Self::try_from_data(&data).map_err(|_| Error::InvalidSeHdr)
}
}
@@ -342,13 +342,13 @@ impl UvDataPlainTrait for SeHdrPlain {
}
impl AeadPlainDataTrait for SeHdrPlain {
- fn aad(&self) -> Vec<u8> {
- let data_aad = self.data.aad();
+ fn aad(&self) -> Result<Vec<u8>> {
+ let data_aad = self.data.aad()?;
- [serialize_to_bytes(&self.common).unwrap(), data_aad].concat()
+ Ok([serialize_to_bytes(&self.common)?, data_aad].concat())
}
- fn data(&self) -> Confidential<Vec<u8>> {
+ fn data(&self) -> Result<Confidential<Vec<u8>>> {
self.data.data()
}
@@ -387,5 +387,41 @@ mod tests {
])),
Err(Error::InvalidSeHdr)
));
+
+ // Invalid SeHdr as the `sehs` is too large.
+ assert!(matches!(
+ SeHdr::try_from_io(Cursor::new([
+ 73, 66, 77, 83, 101, 99, 69, 120, 0, 0, 1, 0, 0, 0, 1, 255, 65, 65, 65, 65, 67, 0,
+ 65, 17, 65, 0, 65, 65, 65, 65, 65, 65, 91, 91, 180, 91, 91, 91, 91, 91, 91, 91, 91,
+ 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
+ 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
+ 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
+ 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 241, 241,
+ 241, 241, 241, 91, 91, 91, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112,
+ 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 80,
+ 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112,
+ 112, 112, 112, 112, 91, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112,
+ 112, 112, 112, 112, 112, 112, 112, 0, 0, 0, 0, 101, 99, 255, 255, 255, 255, 255,
+ 255, 255, 255, 255, 255, 255, 65, 65, 65, 65, 67, 0, 65, 17, 65, 0, 65, 65, 65, 65,
+ 65, 65, 91, 91, 180, 91, 91, 91, 91, 91, 91, 91, 91, 255, 255, 255, 255, 255, 255,
+ 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
+ 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
+ 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
+ 255, 255, 255, 255, 255, 255, 255, 255, 241, 241, 241, 241, 241, 91, 91, 91, 112,
+ 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112,
+ 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 80, 112, 112, 112, 112, 112, 112,
+ 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 91, 112, 112,
+ 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 73, 66, 77, 83, 101, 99, 69, 120,
+ 0, 112, 112, 0, 1, 0, 0, 0, 0, 101, 99, 255, 255, 255, 255, 255, 255, 255, 255,
+ 255, 255, 255, 65, 65, 65, 65, 67, 0, 65, 17, 65, 0, 65, 65, 65, 65, 65, 65, 91,
+ 91, 180, 91, 91, 91, 91, 91, 91, 91, 91, 91, 91, 91, 91, 91, 91, 91, 91, 91, 91,
+ 91, 91, 112, 112, 112, 112, 112, 73, 66, 77, 83, 101, 99, 69, 120, 0, 0, 1, 0, 0,
+ 0, 0, 48, 53, 53, 53, 53, 53, 53, 53, 91, 91, 91, 241, 241, 46, 49, 49, 0, 49, 49,
+ 0, 0, 112, 112, 112, 91, 0, 0, 0, 0, 9, 0, 49, 50, 22, 241, 241, 241, 241, 241,
+ 241, 241, 241, 241, 241, 241, 91, 91, 91, 91, 91, 255, 251, 0, 0, 91, 91, 91, 91,
+ 91, 91, 91, 91, 91, 91, 91, 0, 0, 91, 0, 0, 10, 91, 91, 91, 65, 65, 65, 65
+ ])),
+ Err(Error::InvalidSeHdr)
+ ));
}
}
diff --git a/rust/pvimg/src/pv_utils/se_hdr/builder.rs b/rust/pvimg/src/pv_utils/se_hdr/builder.rs
index ba6de898..93bcc7af 100644
--- a/rust/pvimg/src/pv_utils/se_hdr/builder.rs
+++ b/rust/pvimg/src/pv_utils/se_hdr/builder.rs
@@ -230,8 +230,14 @@ mod tests {
let decrypted = bin.decrypt(&prot_key).expect("BUG");
assert_eq!(bin.common, decrypted.common);
- assert_eq!(bin.aad(), decrypted.aad());
- assert_ne!(&bin.data(), decrypted.data().value());
+ assert_eq!(
+ bin.aad().expect("should not fail"),
+ decrypted.aad().expect("should not fail")
+ );
+ assert_ne!(
+ &bin.data(),
+ decrypted.data().expect("should not fail").value()
+ );
let _decrypted_hdrv1: SeHdrDataV1 = decrypted.data.try_into().expect("BUG");
}
diff --git a/rust/pvimg/src/pv_utils/se_hdr/hdr_v1.rs b/rust/pvimg/src/pv_utils/se_hdr/hdr_v1.rs
index a7f2f609..b179d50d 100644
--- a/rust/pvimg/src/pv_utils/se_hdr/hdr_v1.rs
+++ b/rust/pvimg/src/pv_utils/se_hdr/hdr_v1.rs
@@ -19,6 +19,7 @@ use serde::{Serialize, Serializer};
use super::keys::phkh_v1;
use crate::{
error::Error,
+ misc::PAGESIZE,
pv_utils::{
error::Result,
se_hdr::{
@@ -51,11 +52,14 @@ struct HdrSizesV1 {
#[derive(Debug, Clone, PartialEq, Eq, DekuRead, DekuWrite, Serialize)]
#[deku(endian = "endian", ctx = "endian: Endian", ctx_default = "Endian::Big")]
struct SeHdrAadV1 {
+ #[deku(assert = "*sehs <= SeHdrDataV1::MAX_SIZE.try_into().unwrap()")]
sehs: u32,
#[serde(serialize_with = "ser_hex")]
iv: [u8; SymKeyType::AES_256_GCM_IV_LEN],
res1: u32,
+ #[deku(assert = "*nks <= (*sehs).into()", update = "self.keyslots.len()")]
nks: u64,
+ #[deku(assert = "*sea <= (*sehs).into()")]
sea: u64,
nep: u64,
#[serde(serialize_with = "ser_lower_hex")]
@@ -118,6 +122,7 @@ pub struct SeHdrConfV1 {
psw: PSW,
#[serde(serialize_with = "ser_lower_hex")]
scf: u64,
+ #[deku(assert_eq = "0")]
noi: u32,
res2: u32,
#[deku(count = "noi")]
@@ -200,6 +205,7 @@ where
}
impl SeHdrDataV1 {
+ const MAX_SIZE: usize = 2 * PAGESIZE;
const PCF_DEFAULT: u64 = 0x0;
const SCF_DEFAULT: u64 = 0x0;
@@ -241,7 +247,14 @@ impl SeHdrDataV1 {
tag: SeHdrTagV1::default(),
};
let hdr_size = ret.size()?;
- ret.aad.sehs = hdr_size.phs.try_into()?;
+ let phs = hdr_size.phs.try_into()?;
+ if phs > Self::MAX_SIZE {
+ return Err(Error::InvalidSeHdrTooLarge {
+ given: phs,
+ maximum: Self::MAX_SIZE,
+ });
+ }
+ ret.aad.sehs = phs.try_into()?;
ret.aad.sea = hdr_size.sea;
Ok(ret)
}
@@ -494,8 +507,8 @@ impl KeyExchangeTrait for SeHdrBinV1 {
}
impl AeadDataTrait for SeHdrBinV1 {
- fn aad(&self) -> Vec<u8> {
- serialize_to_bytes(&self.aad).unwrap()
+ fn aad(&self) -> Result<Vec<u8>> {
+ serialize_to_bytes(&self.aad)
}
fn data(&self) -> Vec<u8> {
@@ -508,12 +521,12 @@ impl AeadDataTrait for SeHdrBinV1 {
}
impl AeadPlainDataTrait for SeHdrDataV1 {
- fn aad(&self) -> Vec<u8> {
- serialize_to_bytes(&self.aad).unwrap()
+ fn aad(&self) -> Result<Vec<u8>> {
+ serialize_to_bytes(&self.aad)
}
- fn data(&self) -> Confidential<Vec<u8>> {
- serialize_to_bytes(self.data.value()).unwrap().into()
+ fn data(&self) -> Result<Confidential<Vec<u8>>> {
+ Ok(serialize_to_bytes(self.data.value())?.into())
}
fn tag(&self) -> Vec<u8> {
@@ -610,4 +623,48 @@ mod tests {
assert_eq!(psw, hdr_data_v1.data.value().psw);
assert_eq!(cck.value(), hdr_data_v1.data.value().cck.value());
}
+
+ #[test]
+ fn max_size_sehdr_test() {
+ const MAX_HOST_KEYS: usize = 95;
+
+ let (_, host_key) = get_test_key_and_cert();
+ let pub_key = host_key.public_key().unwrap();
+ let host_keys_max: Vec<_> = (0..MAX_HOST_KEYS).map(|_| pub_key.clone()).collect();
+ let too_many_host_keys: Vec<_> = (0..MAX_HOST_KEYS + 1).map(|_| pub_key.clone()).collect();
+ let xts_key = Confidential::new([0x3; SymKeyType::AES_256_XTS_KEY_LEN]);
+ let meta = ComponentMetadataV1 {
+ ald: [0x1; SHA_512_HASH_LEN],
+ pld: [0x2; SHA_512_HASH_LEN],
+ tld: [0x3; SHA_512_HASH_LEN],
+ nep: 3,
+ key: xts_key,
+ };
+ let psw = PSW {
+ addr: 1234,
+ mask: 5678,
+ };
+
+ let mut builder = SeHdrBuilder::new(SeHdrVersion::V1, psw.clone(), meta.clone())
+ .expect("should not fail");
+ builder
+ .add_hostkeys(&host_keys_max)
+ .expect("should not fail")
+ .with_components(meta.clone())
+ .expect("should not fail");
+ let bin = builder.build().expect("should not fail");
+ assert_eq!(bin.common.version, SeHdrVersion::V1);
+ let hdr_v1: SeHdrBinV1 = bin.data.try_into().expect("should not fail");
+ assert_eq!(hdr_v1.aad.sehs, 8160);
+
+ let mut builder = SeHdrBuilder::new(SeHdrVersion::V1, psw.clone(), meta.clone())
+ .expect("should not fail");
+
+ builder
+ .add_hostkeys(&too_many_host_keys)
+ .expect("should not fail")
+ .with_components(meta)
+ .expect("should not fail");
+ assert!(matches!(builder.build(), Err(Error::InvalidSeHdr)));
+ }
}
diff --git a/rust/pvimg/src/pv_utils/uvdata.rs b/rust/pvimg/src/pv_utils/uvdata.rs
index b0ec355a..c6ed9567 100644
--- a/rust/pvimg/src/pv_utils/uvdata.rs
+++ b/rust/pvimg/src/pv_utils/uvdata.rs
@@ -34,7 +34,7 @@ pub trait AeadCipherTrait {
#[enum_dispatch]
pub trait AeadDataTrait {
/// Returns the authenticated associated data.
- fn aad(&self) -> Vec<u8>;
+ fn aad(&self) -> Result<Vec<u8>>;
/// Returns the encrypted data.
fn data(&self) -> Vec<u8>;
@@ -47,10 +47,10 @@ pub trait AeadDataTrait {
#[enum_dispatch]
pub trait AeadPlainDataTrait {
/// Returns the authenticated associated data.
- fn aad(&self) -> Vec<u8>;
+ fn aad(&self) -> Result<Vec<u8>>;
/// Returns the unencrypted data.
- fn data(&self) -> Confidential<Vec<u8>>;
+ fn data(&self) -> Result<Confidential<Vec<u8>>>;
/// Returns the tag data.
fn tag(&self) -> Vec<u8>;
@@ -124,8 +124,14 @@ pub trait UvDataPlainTrait:
expected: self.aead_key_type().to_string(),
});
}
- let aad = self.aad();
- let unecrypted_data = self.data();
+ let aad = self.aad().map_err(|err| match err {
+ Error::Deku(_) => Error::InvalidSeHdr,
+ err => err,
+ })?;
+ let unecrypted_data = self.data().map_err(|err| match err {
+ Error::Deku(_) => Error::InvalidSeHdr,
+ err => err,
+ })?;
let iv = self.iv();
let result = encrypt_aead(key, iv, &aad, unecrypted_data.value())?;
Self::C::try_from_data(&result.into_buf())
@@ -169,7 +175,7 @@ pub trait UvDataTrait: AeadDataTrait + AeadCipherTrait + KeyExchangeTrait + Clon
}
let tag_size = self.aead_tag_size();
- let aad = self.aad();
+ let aad = self.aad()?;
let unecrypted_data = self.data();
let iv = self.iv();
let tag = self.tag();

View File

@ -0,0 +1,101 @@
From 6e48c5ebaa26c6bd2a1bc33ccf36ed8bd6946358 Mon Sep 17 00:00:00 2001
From: Marc Hartmayer <mhartmay@linux.ibm.com>
Date: Tue, 17 Dec 2024 18:13:31 +0100
Subject: [PATCH] pvimg: info: Rename '--key' into '--hdr-key' and use '--key'
as an alias
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Rename '--key' into '--hdr-key' and use '--key' as an (non-visible)
alias for '--hdr-key' in order to keep the command line backwards
compatible. The chances of someone using '--key' are very low, as this
version has not yet been released by any OS distribution.
This change makes the command line options for the different subcommands
more consistent and therefore easier to use.
Suggested-by: Reinhard Bündgen <buendgen@de.ibm.com>
Acked-by: Hendrik Brueckner <brueckner@linux.ibm.com>
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
---
rust/pvimg/man/pvimg-info.1 | 2 +-
rust/pvimg/src/cli.rs | 22 +++++++++++++++++++---
rust/pvimg/src/cmd/info.rs | 2 +-
3 files changed, 21 insertions(+), 5 deletions(-)
Index: s390-tools-2.36.0/rust/pvimg/man/pvimg-info.1
===================================================================
--- s390-tools-2.36.0.orig/rust/pvimg/man/pvimg-info.1
+++ s390-tools-2.36.0/rust/pvimg/man/pvimg-info.1
@@ -37,7 +37,7 @@ Possible values:
.RE
.RE
.PP
-\-\-key <FILE>
+\-\-hdr\-key <FILE>
.RS 4
Use the key in FILE to decrypt the Secure Execution header.
.RE
Index: s390-tools-2.36.0/rust/pvimg/src/cli.rs
===================================================================
--- s390-tools-2.36.0.orig/rust/pvimg/src/cli.rs
+++ s390-tools-2.36.0/rust/pvimg/src/cli.rs
@@ -192,8 +192,8 @@ pub struct InfoArgs {
pub format: OutputFormat,
/// Use the key in FILE to decrypt the Secure Execution header.
- #[arg(long, value_name = "FILE", value_hint = ValueHint::FilePath,)]
- pub key: Option<PathBuf>,
+ #[arg(long, value_name = "FILE", value_hint = ValueHint::FilePath, alias = "key")]
+ pub hdr_key: Option<PathBuf>,
}
#[derive(Args, Debug)]
@@ -710,6 +710,22 @@ mod test {
CliOption::new("image", ["/dev/null"]),
],
)),
+ flat_map_collect(insert(
+ args.clone(),
+ vec![
+ CliOption::new("hdr-key", ["--hdr-key", "/dev/null"]),
+ CliOption::new("format", ["--format=json"]),
+ CliOption::new("image", ["/dev/null"]),
+ ],
+ )),
+ flat_map_collect(insert(
+ args.clone(),
+ vec![
+ CliOption::new("hdr-key", ["--key", "/dev/null"]),
+ CliOption::new("format", ["--format=json"]),
+ CliOption::new("image", ["/dev/null"]),
+ ],
+ )),
// separation between keyword and positional args works
flat_map_collect(insert(
args.clone(),
@@ -750,7 +766,7 @@ mod test {
// Test for invalid combinations
// Input is missing
- let mut pvimg_invalid_args = vec![vec!["pvimg", "test"]];
+ let mut pvimg_invalid_args = vec![vec!["pvimg", "info"]];
for create_args in &valid_test_args {
pvimg_valid_args.push(
Index: s390-tools-2.36.0/rust/pvimg/src/cmd/info.rs
===================================================================
--- s390-tools-2.36.0.orig/rust/pvimg/src/cmd/info.rs
+++ s390-tools-2.36.0/rust/pvimg/src/cmd/info.rs
@@ -27,7 +27,7 @@ pub fn info(opt: &InfoArgs) -> Result<Ow
SeHdr::seek_sehdr(&mut input, None)?;
let hdr = SeHdr::try_from_io(input)?;
- if let Some(key_path) = &opt.key {
+ if let Some(key_path) = &opt.hdr_key {
let key =
SymKey::try_from_data(hdr.key_type(), read_file(key_path, "Reading key")?.into())?;
serde_json::to_writer_pretty(&mut output, &hdr.decrypt(&key)?)?;

6
s390-tools-rpmlintrc Normal file
View File

@ -0,0 +1,6 @@
addFilter("statically-linked-binary /usr/lib/s390-tools/.*")
addFilter("statically-linked-binary /usr/bin/read_values")
addFilter("systemd-service-without-service_.* *@.service")
addFilter("position-independent-executable-suggested ")
addFilter("non-etc-or-var-file-marked-as-conffile /boot/zipl/active_devices.txt")
addFilter("zero-length /boot/zipl/active_devices.txt")

View File

@ -0,0 +1,30 @@
---
etc/udev/rules.d/59-dasd.rules | 10 +++++++++-
1 file changed, 9 insertions(+), 1 deletion(-)
--- a/etc/udev/rules.d/59-dasd.rules
+++ b/etc/udev/rules.d/59-dasd.rules
@@ -15,7 +15,7 @@
LABEL="dasd_block_end"
-ACTION!="change|add", GOTO="dasd_symlinks_end"
+ACTION!="change|add", GOTO="dasd_partition_end"
# for partitions import parent information
KERNEL=="dasd*[0-9]", IMPORT{parent}=="ID_*"
@@ -24,6 +24,14 @@
KERNEL=="dasd*[0-9]", ENV{ID_UID}=="?*", SYMLINK+="disk/by-id/$env{ID_BUS}-$env{ID_UID}-part%n"
KERNEL=="dasd*[0-9]", ENV{ID_XUID}=="?*", SYMLINK+="disk/by-id/$env{ID_BUS}-$env{ID_XUID}-part%n"
+LABEL="dasd_partition_end"
+
+ENV{ID_SERIAL}!="?*", GOTO="dasd_symlinks_end"
+# by-label/by-uuid (filesystem properties)
+IMPORT{builtin}="blkid"
+ENV{ID_FS_USAGE}=="filesystem|other|crypto", ENV{ID_FS_UUID}=="?*", SYMLINK+="disk/by-uuid/$env{ID_FS_UUID}"
+ENV{ID_FS_USAGE}=="filesystem|other", ENV{ID_FS_LABEL_SAFE}=="?*", SYMLINK+="disk/by-label/$env{ID_FS_LABEL_SAFE}"
+
LABEL="dasd_symlinks_end"
# on device add set request queue scheduler to none

View File

@ -0,0 +1,33 @@
From d0c2ffc90b9ee0e7b741d1c4b644cdf79f1d922b Mon Sep 17 00:00:00 2001
From: Hannes Reinecke <hare@suse.de>
Date: Wed, 20 May 2015 11:57:11 +0200
Subject: [PATCH] fdasd: skip partition check and BLKRRPART ioctl for emulated
devices
If 'fdasd -f' is called we cannot rely on the partition detection
via a simple check of the minor number, so the check should be
suppressed.
Similarly, not every emulated device supports the BLKRRPART ioctl,
so we should be suppressing the error message for these devices, too.
Signed-off-by: Hannes Reinecke <hare@suse.de>
---
fdasd/fdasd.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
--- a/fdasd/fdasd.c
+++ b/fdasd/fdasd.c
@@ -1231,10 +1231,12 @@
*/
static void fdasd_reread_partition_table(fdasd_anchor_t *anc)
{
+ int rc = 0 ;
if (!anc->silent)
printf("rereading partition table...\n");
- if (dasd_reread_partition_table(options.device, 5) != 0) {
+ rc = dasd_reread_partition_table(options.device, 1);
+ if (rc == EINVAL && !anc->force_virtual) {
fdasd_error(anc, unable_to_ioctl, "Error while rereading "
"partition table.\nPlease reboot!");
}

View File

@ -0,0 +1,34 @@
From f7a0f391f2c4e8acc96b21ab5de54a178aa60088 Mon Sep 17 00:00:00 2001
From: Hannes Reinecke <hare@suse.de>
Date: Fri, 22 Nov 2013 15:39:38 +0100
Subject: [PATCH] 59-dasd.rules: generate by-id links on 'change' and 'add'
The by-id rules need to be triggered on both, 'change' and 'add',
to work correctly during restarting udev.
References: bnc#808042
Signed-off-by: Robert Milasan <rmilasan@suse.de>
---
etc/udev/rules.d/59-dasd.rules | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/etc/udev/rules.d/59-dasd.rules b/etc/udev/rules.d/59-dasd.rules
index 2b1435c..a08cb7c 100644
--- a/etc/udev/rules.d/59-dasd.rules
+++ b/etc/udev/rules.d/59-dasd.rules
@@ -6,9 +6,9 @@
SUBSYSTEM!="block", GOTO="dasd_symlinks_end"
KERNEL!="dasd*", GOTO="dasd_symlinks_end"
-ACTION!="change", GOTO="dasd_block_end"
+ACTION!="change|add", GOTO="dasd_block_end"
# by-id (hardware serial number)
-KERNEL=="dasd*[!0-9]", ATTRS{status}=="online", IMPORT{program}="/sbin/dasdinfo -a -e -b $kernel"
+KERNEL=="dasd*[!0-9]", ATTRS{status}=="online", IMPORT{program}="/usr/sbin/dasdinfo -a -e -b $kernel"
KERNEL=="dasd*[!0-9]", ENV{ID_SERIAL}=="?*", SYMLINK+="disk/by-id/$env{ID_BUS}-$env{ID_SERIAL}"
KERNEL=="dasd*[!0-9]", ENV{ID_UID}=="?*", SYMLINK+="disk/by-id/$env{ID_BUS}-$env{ID_UID}"
KERNEL=="dasd*[!0-9]", ENV{ID_XUID}=="?*", SYMLINK+="disk/by-id/$env{ID_BUS}-$env{ID_XUID}"
--
1.8.1.4

View File

@ -0,0 +1,26 @@
From f7a0f391f2c4e8acc96b21ab5de54a178aa60088 Mon Sep 17 00:00:00 2001
From: Hannes Reinecke <hare@suse.de>
Date: Fri, 22 Nov 2013 15:39:38 +0100
Subject: [PATCH] 59-dasd.rules: generate by-id links on 'change' and 'add'
The by-id rules need to be triggered on both, 'change' and 'add',
to work correctly during restarting udev.
References: bnc#808042
Signed-off-by: Robert Milasan <rmilasan@suse.de>
---
etc/udev/rules.d/59-dasd.rules | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
--- a/etc/udev/rules.d/59-dasd.rules
+++ b/etc/udev/rules.d/59-dasd.rules
@@ -6,7 +6,7 @@
SUBSYSTEM!="block", GOTO="dasd_symlinks_end"
KERNEL!="dasd*", GOTO="dasd_symlinks_end"
-ACTION!="change", GOTO="dasd_block_end"
+ACTION!="change|add", GOTO="dasd_block_end"
# by-id (hardware serial number)
KERNEL=="dasd*[!0-9]", ATTRS{status}=="online", IMPORT{program}="/sbin/dasdinfo -a -e -b $kernel"
KERNEL=="dasd*[!0-9]", ENV{ID_SERIAL}=="?*", SYMLINK+="disk/by-id/$env{ID_BUS}-$env{ID_SERIAL}"

View File

@ -0,0 +1,20 @@
---
zipl/boot/menu.c | 7 +++++--
1 file changed, 5 insertions(+), 2 deletions(-)
--- a/zipl/boot/menu.c
+++ b/zipl/boot/menu.c
@@ -168,8 +168,11 @@
/* print config list */
menu_list();
- if (is_zvm())
- printf("Note: VM users please use '#cp vi vmsg <input>'\n");
+ if (is_zvm()) {
+ printf(" \n");
+ printf("Note: VM users please use '#cp vi vmsg <input> <kernel-parameters>'\n");
+ printf(" \n");
+ }
value = menu_read();

View File

@ -0,0 +1,148 @@
---
etc/sysconfig/dumpconf | 133 +++++++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 133 insertions(+)
--- a/etc/sysconfig/dumpconf
+++ b/etc/sysconfig/dumpconf
@@ -1,3 +1,4 @@
+###########################################################################################
#
# s390 dump config
#
@@ -78,3 +79,135 @@
# dumpconf becomes active immediately during system startup.
#
# ON_PANIC=reipl
+
+############################ Begin Definitions ###########################################
+## Path: System/Dumpconf
+## Description: Configures the actions which should be performed after a kernel panic
+## Type: list(stop,dump,vmcmd,reipl,dump_reipl)
+## Default: "stop"
+## ServiceRestart: dumpconf
+#
+# Define the action that should be taken if a kernel panic happens.
+#
+ON_PANIC="stop"
+
+## Type: integer(0:300)
+## Default: 5
+## ServiceRestart: dumpconf
+#
+# Using reipl or dump_reipl actions with ON_PANIC can lead to the system
+# looping with alternating IPLs and crashes. Use DELAY_MINUTES to prevent
+# such a loop. DELAY_MINUTES delays activating the specified panic action
+# for a newly started system. When the specified time has elapsed, dumpconf
+# activates the specified panic action. This action is taken should the
+# system subsequently crash. If the system crashes before the time has
+# elapsed the previously defined action is taken. If no previous action has
+# been defined the default action (STOP) is performed.
+#
+DELAY_MINUTES="5"
+
+## Type: list(ccw,fcp,nvme)
+## Default: ""
+## ServiceRestart: dumpconf
+#
+# Define the type, ccw for DASD, fcp for zFCP, or nvme for NVMe Disk.
+#
+DUMP_TYPE=""
+
+## Type: string
+## Default: ""
+## ServiceRestart: dumpconf
+#
+# Define the device id for a DASD or SCSI over zFCP dump device.
+#
+# For example (DASD and SCSI over zFCP have the same structure): DEVICE=0.0.4711
+#
+DEVICE=""
+
+# Type: string
+## Default: ""
+## ServiceRestart: dumpconf
+#
+# Define the WWPN for a zFCP dump device.
+#
+# For example: WWPN=0x5005076303004711
+#
+WWPN=""
+
+## Type: string
+## Default: ""
+## ServiceRestart: dumpconf
+#
+# Define the LUN for a zFCP dump device.
+#
+# For example: LUN=0x4711000000000000
+#
+LUN=""
+
+## Type: integer(0:30)
+## Default: "0"
+## ServiceRestart: dumpconf
+#
+# Define the Boot program selector for a zFCP dump device.
+#
+# A decimal value between 0 and 30 specifying the program to be loaded from
+# the FCP-I/O device.
+#
+BOOTPROG="0"
+
+## Type: string
+## Default: "0"
+## ServiceRestart: dumpconf
+#
+# Define the Boot record logical block address for a zFCP dump device.
+#
+# The hexadecimal digits designating the logical-block address of the boot record of the FCP-I/O device.
+# It must be a value from 0-FFFFFFFF FFFFFFFF. For values longer than 8 hex characters at least one separator
+# blank is required after the 8th character.
+#
+BR_LBA="0"
+
+## Type: string
+## Default: ""
+## ServiceRestart: dumpconf
+#
+# Define the Function ID for NVMe dump device.
+#
+# The hexadecimal digits designating the Function ID for the NMVe disk.
+#
+# For example: FID=0x00000300
+#
+FID=""
+
+## Type: string
+## Default: ""
+## ServiceRestart: dumpconf
+#
+# Define the Namespace ID for the NVMe dump device
+#
+# The hexadecimal digits designating the Namespace ID for the NMVe disk.
+#
+# For example: NSID=0x00000001
+#
+NSID=""
+
+## Type: string
+## Default: ""
+## ServiceRestart: dumpconf
+#
+# VMCMD_<X>
+# Specifies a CP command, <X> is a number from one to eight. You can
+# specify up to eight CP commands that are executed in case of a kernel
+# panic. Note that VM commands, device adresses, and VM guest names
+# must be uppercase.
+#
+VMCMD_1=""
+VMCMD_2=""
+VMCMD_3=""
+VMCMD_4=""
+VMCMD_5=""
+VMCMD_6=""
+VMCMD_7=""
+VMCMD_8=""
+
+############################### End Definitions ##############################################
\ No newline at end of file

View File

@ -0,0 +1,50 @@
Subject: zdev: Add support for handling I/O configuration data
From: Peter Oberparleiter <oberpar@linux.ibm.com>
Summary: zdev: Add support for handling I/O configuration data
Description: LPARs that are running in IBM Dynamic Partition Manager (DPM) mode
can access a firmware-generated I/O configuration data file that
contains s390-specific information about available I/O devices
such as qeth device numbers and parameters, and FCP device IDs.
This data file is intended to remove the need for users to
manually enter the corresponding device data during installation.
Linux kernels with the corresponding support make the I/O
configuration data available at the following location:
/sys/firmware/sclp_sd/config/data
This patch set adds support for handling this data file using the
chzdev and lszdev tools:
- I/O configuration data can be applied using chzdev's --import
option
- Initial RAM-Disk scripts automatically apply the
I/O configuration data to the system configuration
- lszdev can be used to display the applied auto-configuration
data
- chzdev can be used to manually override the
auto-configuration data
Upstream-ID: -
Problem-ID: LS1604
Signed-off-by: Peter Oberparleiter <oberpar@linux.ibm.com>
---
zdev/src/zdev-root-update.dracut | 6 ------
1 file changed, 6 deletions(-)
--- a/zdev/src/zdev-root-update.dracut
+++ b/zdev/src/zdev-root-update.dracut
@@ -20,10 +20,4 @@
exit 1
}
-echo "Installing IPL record"
-zipl --noninteractive || {
- echo "${TOOLNAME}: Error: Could not install IPL record" >&2
- exit 1
-}
-
exit 0

View File

@ -0,0 +1,465 @@
From d6582bbaf0f3986a42f562046dc0caa9de89c75e Mon Sep 17 00:00:00 2001
From: Hannes Reinecke <hare@suse.de>
Date: Fri, 6 Oct 2017 08:58:17 +0200
Subject: [PATCH] dasdfmt: Allow multiple device arguments
Allow the user to specify several devices as arguments to dasdfmt.
Signed-off-by: Hannes Reinecke <hare@suse.com>
---
dasdfmt/dasdfmt.8 | 6 -
dasdfmt/dasdfmt.c | 197 +++++++++++++++++++++++++++++++-----------------------
2 files changed, 119 insertions(+), 84 deletions(-)
--- a/dasdfmt/dasdfmt.8
+++ b/dasdfmt/dasdfmt.8
@@ -11,14 +11,14 @@
.br
[\-r \fIcylinder\fR] [\-b \fIblksize\fR] [\-l \fIvolser\fR] [\-d \fIlayout\fR]
.br
- [\-L] [\-V] [\-F] [\-k] [\-C] [\-M \fImode\fR] \fIdevice\fR
+ [\-L] [\-V] [\-F] [\-k] [\-C] [\-M \fImode\fR] \fIdevice\fR [\fIdevice\fR]
.SH DESCRIPTION
-\fBdasdfmt\fR formats a DASD (ECKD) disk drive to prepare it
+\fBdasdfmt\fR formats one or several DASD (ECKD) disk drive(s) to prepare them
for usage with Linux for S/390.
The \fIdevice\fR is the node of the device (e.g. '/dev/dasda').
Any device node created by udev for kernel 2.6 can be used
-(e.g. '/dev/dasd/0.0.b100/disc').
+(e.g. '/dev/dasd/0.0.b100/disc'). It is possible to specify up to 512 devices.
.br
\fBWARNING\fR: Careless usage of \fBdasdfmt\fR can result in
--- a/dasdfmt/dasdfmt.c
+++ b/dasdfmt/dasdfmt.c
@@ -25,6 +25,8 @@
#include "dasdfmt.h"
+#define MAX_DEVICES 512
+#define MAX_LENGTH 256
#define BUSIDSIZE 8
#define SEC_PER_DAY (60 * 60 * 24)
#define SEC_PER_HOUR (60 * 60)
@@ -57,7 +59,9 @@
static struct dasdfmt_globals {
dasd_information2_t dasd_info;
char *dev_path; /* device path entered by user */
+ char dev_path_array[MAX_DEVICES][MAX_LENGTH]; /* Array of device paths entered by user */
char *dev_node; /* reliable device node determined by dasdfmt */
+ char dev_node_array[MAX_DEVICES][MAX_LENGTH]; /* Array of reliable device nodes determined by dasdfmt */
int verbosity;
int testmode;
int withoutprompt;
@@ -484,15 +488,15 @@
program_interrupt_in_progress = 1;
if (disk_disabled) {
- printf("Re-accessing the device...\n");
+ printf("Re-accessing %s...\n", g.dev_path);
disk_enable();
}
- printf("Rereading the partition table...\n");
+ printf("Rereading the partition table for %s...\n", g.dev_path);
rc = dasd_reread_partition_table(g.dev_node, 5);
if (rc) {
ERRMSG("%s: (signal handler) Re-reading partition table "
- "failed. (%s)\n", prog_name, strerror(rc));
+ "for %s failed. (%s)\n", prog_name, g.dev_path, strerror(rc));
} else {
printf("Exiting...\n");
}
@@ -512,9 +516,6 @@
unsigned int maj, min;
struct stat dev_stat;
- if (optind + 1 < argc)
- error("More than one device specified!");
-
if (optind >= argc)
error("No device specified!");
@@ -610,10 +611,10 @@
error("the ioctl call to retrieve read/write status information failed: %s",
strerror(err));
if (ro)
- error("Disk is read only!");
+ error("Disk %s is read only!", g.dev_path);
if (!g.force) {
if (g.dasd_info.open_count > 1)
- error("Disk in use!");
+ error("Disk %s is in use!", g.dev_path);
}
if (strncmp(g.dasd_info.type, "ECKD", 4) != 0) {
warnx("Unsupported disk type");
@@ -700,7 +701,7 @@
struct dasd_eckd_characteristics *characteristics;
if (g.verbosity > 0)
- printf("Retrieving disk geometry...\n");
+ printf("Retrieving disk geometry for %s...\n", g.dev_path);
characteristics = (struct dasd_eckd_characteristics *)
&g.dasd_info.characteristics;
@@ -728,13 +729,13 @@
"Cylinders above this limit will not be"
" accessible as a linux partition!\n"
"Type \"yes\" to continue, no will leave"
- " the disk untouched: ", LV_COMPAT_CYL);
+ " the %s disk untouched: ", LV_COMPAT_CYL, g.dev_path);
if (fgets(inp_buffer, sizeof(inp_buffer), stdin) == NULL)
return;
if (strcasecmp(inp_buffer, "yes") &&
strcasecmp(inp_buffer, "yes\n")) {
- printf("Omitting ioctl call (disk will "
- "NOT be formatted).\n");
+ printf("Omitting ioctl call (disk %s will "
+ "NOT be formatted).\n", g.dev_path);
return;
}
}
@@ -872,7 +873,7 @@
check_params->start_unit = 0;
check_params->stop_unit = (cylinders * heads) - 1;
- printf("Checking format of the entire disk...\n");
+ printf("Checking format of the entire %s disk...\n", g.dev_path);
if (g.testmode) {
printf("Test mode active, omitting ioctl.\n");
@@ -896,7 +897,7 @@
if (process_tracks(cylinders, heads, check_params))
error("Use --mode=full to perform a clean format.");
- printf("Done. Disk is fine.\n");
+ printf("Done. Disk %s is fine.\n", g.dev_path);
}
/*
@@ -946,8 +947,8 @@
printf("Device Type: %s Provisioned\n",
g.ese ? "Thinly" : "Fully");
- printf("\nI am going to format the device ");
- printf("%s in the following way:\n", g.dev_path);
+ printf("\nI am going to format %s ", g.dev_path);
+ printf("in the following way:\n");
printf(" Device number of device : 0x%x\n", g.dasd_info.devno);
printf(" Labelling device : %s\n",
(g.writenolabel) ? "no" : "yes");
@@ -1012,7 +1013,7 @@
int ipl1_record_len, ipl2_record_len;
if (g.verbosity > 0)
- printf("Retrieving dasd information... ");
+ printf("Retrieving dasd information for %s... ", g.dev_path);
get_blocksize(&blksize);
@@ -1030,7 +1031,7 @@
/* write empty bootstrap (initial IPL records) */
if (g.verbosity > 0)
- printf("Writing empty bootstrap...\n");
+ printf("Writing empty bootstrap to %s...\n", g.dev_path);
/*
* Note: ldl labels do not contain the key field
@@ -1089,7 +1090,7 @@
label_position = g.dasd_info.label_block * blksize;
if (g.verbosity > 0)
- printf("Writing label...\n");
+ printf("Writing label to %s...\n", g.dev_path);
rc = lseek(fd, label_position, SEEK_SET);
if (rc != label_position) {
@@ -1120,7 +1121,7 @@
}
if (g.verbosity > 0)
- printf("Writing VTOC... ");
+ printf("Writing VTOC to %s... ", g.dev_path);
label_position = (VTOC_START_CC * heads + VTOC_START_HH) *
geo.sectors * blksize;
@@ -1242,7 +1243,7 @@
if (!g.ese || g.no_discard)
return;
- printf("Releasing space for the entire device...\n");
+ printf("Releasing space for the entire %s device...\n", g.dev_path);
err = dasd_release_space(g.dev_node, &r);
if (err)
error("Could not release space: %s", strerror(err));
@@ -1261,20 +1262,21 @@
int err;
if (!(g.withoutprompt && g.verbosity < 1))
- printf("Formatting the device. This may take a while "
- "(get yourself a coffee).\n");
+ printf("Formatting the %s device. This may take a while "
+ "(get yourself a coffee).\n", g.dev_path);
if (g.verbosity > 0)
- printf("Detaching the device...\n");
+ printf("Detaching the %s device...\n", g.dev_path);
disk_disable(g.dev_node);
if (g.verbosity > 0)
- printf("Invalidate first track...\n");
+ printf("Invalidate first track on %s...\n", g.dev_path);
err = dasd_format_disk(filedes, &temp);
if (err != 0)
- error("(invalidate first track) IOCTL BIODASDFMT failed: %s", strerror(err));
+ error("(invalidate first track) IOCTL BIODASDFMT failed for %s: %s",
+ g.dev_path, strerror(err));
/* except track 0 from standard formatting procss */
p->start_unit = 1;
@@ -1282,19 +1284,19 @@
process_tracks(cylinders, heads, p);
if (g.verbosity > 0)
- printf("formatting tracks complete...\n");
+ printf("formatting tracks for %s complete...\n", g.dev_path);
temp.intensity = p->intensity;
if (g.verbosity > 0)
- printf("Revalidate first track...\n");
+ printf("Revalidate first track on %s...\n", g.dev_path);
err = dasd_format_disk(filedes, &temp);
if (err != 0)
error("(re-validate first track) IOCTL BIODASDFMT failed: %s", strerror(err));
if (g.verbosity > 0)
- printf("Re-accessing the device...\n");
+ printf("Re-accessing the %s device...\n", g.dev_path);
disk_enable();
}
@@ -1306,18 +1308,18 @@
format_data_t *p)
{
if (!(g.withoutprompt && g.verbosity < 1))
- printf("Formatting the device. This may take a while "
- "(get yourself a coffee).\n");
+ printf("Formatting the %s device. This may take a while "
+ "(get yourself a coffee).\n", g.dev_path);
if (g.verbosity > 0)
- printf("Detaching the device...\n");
+ printf("Detaching the %s device...\n", g.dev_path);
disk_disable(g.dev_node);
process_tracks(cylinders, heads, p);
if (g.verbosity > 0)
- printf("Formatting tracks complete...\n");
+ printf("formatting tracks for %s complete...\n", g.dev_path);
if (g.verbosity > 0)
printf("Re-accessing the device...\n");
@@ -1426,16 +1428,16 @@
if (!g.withoutprompt) {
printf("\n");
if (mode != EXPAND)
- printf("--->> ATTENTION! <<---\nAll data of "
- "that device will be lost.\n");
+ printf("--->> ATTENTION! <<---\nAll data on "
+ "the %s device will be lost.\n", g.dev_path);
printf("Type \"yes\" to continue, no will leave the "
"disk untouched: ");
if (fgets(inp_buffer, sizeof(inp_buffer), stdin) == NULL)
return;
if (strcasecmp(inp_buffer, "yes") &&
strcasecmp(inp_buffer, "yes\n")) {
- printf("Omitting ioctl call (disk will "
- "NOT be formatted).\n");
+ printf("Omitting ioctl call (disk %s will "
+ "NOT be formatted).\n", g.dev_path);
return;
}
}
@@ -1453,12 +1455,12 @@
break;
}
- printf("Finished formatting the device.\n");
+ printf("Finished formatting the %s device.\n", g.dev_path);
if (!(g.writenolabel || mode == EXPAND))
dasdfmt_write_labels(vlabel, cylinders, heads);
- printf("Rereading the partition table... ");
+ printf("Rereading the partition table for %s... ", g.dev_path);
err = dasd_reread_partition_table(g.dev_node, 5);
if (err != 0) {
ERRMSG("%s: error during rereading the partition "
@@ -1472,7 +1474,7 @@
static void eval_format_mode(void)
{
if (!g.force && g.mode_specified && g.ese && mode == EXPAND) {
- warnx("WARNING: The specified device is thin-provisioned");
+ warnx("WARNING: The specified device, %s, is thin-provisioned", g.dev_path);
warnx("Format mode 'expand' is not feasible.");
error("Use --mode=full or --mode=quick to perform a clean format");
}
@@ -1495,20 +1497,70 @@
prog_name = p + 1;
}
-int main(int argc, char *argv[])
+void process_dasd(volume_label_t *orig_vlabel, format_data_t format_params)
{
volume_label_t vlabel;
char old_volser[7];
-
char str[ERR_LENGTH];
+ unsigned int cylinders, heads; int rc;
+
+ rc = dasd_get_info(g.dev_node, &g.dasd_info);
+ if (rc != 0)
+ error("the ioctl call to retrieve device information failed: %s", strerror(rc));
+
+ g.ese = dasd_sys_ese(g.dev_node);
+ eval_format_mode();
+
+ /* Not sure this next line is needed in the new version of the code. */
+ memcpy(&vlabel, orig_vlabel, sizeof(vlabel));
+
+ /* Either let the user specify the blksize or get it from the kernel */
+ if (!g.blksize_specified) {
+ if (!(mode == FULL ||
+ g.dasd_info.format == DASD_FORMAT_NONE) || g.check)
+ get_blocksize(&format_params.blksize);
+ else
+ format_params = ask_user_for_blksize(format_params);
+ }
+
+ if (g.keep_volser) {
+ if (g.labelspec)
+ error("The -k and -l options are mutually exclusive");
+ if (!(format_params.intensity & DASD_FMT_INT_COMPAT))
+ error("WARNING: VOLSER cannot be kept when using the ldl format!");
+
+ if (dasdfmt_get_volser(old_volser) == 0)
+ vtoc_volume_label_set_volser(&vlabel, old_volser);
+ else
+ error("VOLSER not found on device %s", g.dev_path);
+ }
+
+ check_disk();
+
+ if (check_param(str, ERR_LENGTH, &format_params) < 0)
+ error("%s", str);
+
+ set_geo(&cylinders, &heads);
+ set_label(&vlabel, &format_params, cylinders);
+
+ if (g.check)
+ check_disk_format(cylinders, heads, &format_params);
+ else
+ do_format_dasd(&vlabel, &format_params, cylinders, heads);
+
+}
+
+int main(int argc, char *argv[])
+{
+ volume_label_t vlabel;
+
char buf[7];
char *blksize_param_str = NULL;
char *reqsize_param_str = NULL;
char *hashstep_str = NULL;
- int rc;
- unsigned int cylinders, heads;
+ int rc, numdev = 0, i;
/* Establish a handler for interrupt signals. */
signal(SIGTERM, program_interrupt_signal);
@@ -1644,6 +1696,9 @@
break; /* exit loop if finished */
}
+ /* Reset the value of rc since we're going to use it again later. */
+ rc = 0;
+
CHECK_SPEC_MAX_ONCE(g.blksize_specified, "blocksize");
CHECK_SPEC_MAX_ONCE(g.labelspec, "label");
CHECK_SPEC_MAX_ONCE(g.writenolabel, "omit-label-writing flag");
@@ -1662,48 +1717,28 @@
if (g.print_hashmarks)
PARSE_PARAM_INTO(g.hashstep, hashstep_str, 10, "hashstep");
- get_device_name(optind, argc, argv);
-
- rc = dasd_get_info(g.dev_node, &g.dasd_info);
- if (rc != 0)
- error("the ioctl call to retrieve device information failed: %s", strerror(rc));
-
- g.ese = dasd_sys_ese(g.dev_node);
- eval_format_mode();
+ while (optind < argc) {
+ get_device_name(optind, argc, argv);
+ strncpy(g.dev_path_array[numdev], g.dev_path, strlen(g.dev_path));
+ strncpy(g.dev_node_array[numdev], g.dev_node, strlen(g.dev_node));
- /* Either let the user specify the blksize or get it from the kernel */
- if (!g.blksize_specified) {
- if (!(mode == FULL ||
- g.dasd_info.format == DASD_FORMAT_NONE) || g.check)
- get_blocksize(&format_params.blksize);
- else
- format_params = ask_user_for_blksize(format_params);
+ optind++;
+ numdev++;
}
- if (g.keep_volser) {
- if (g.labelspec)
- error("The -k and -l options are mutually exclusive");
- if (!(format_params.intensity & DASD_FMT_INT_COMPAT))
- error("WARNING: VOLSER cannot be kept when using the ldl format!");
-
- if (dasdfmt_get_volser(old_volser) == 0)
- vtoc_volume_label_set_volser(&vlabel, old_volser);
- else
- error("VOLSER not found on device %s", g.dev_path);
- }
+ if (!numdev)
+ error("%s: No device specified!\n",
+ prog_name);
- check_disk();
+ if (numdev > 1 && g.labelspec)
+ error("Specifying a volser to be written doesn't make sense when formatting multiple DASD volumes.");
- if (check_param(str, ERR_LENGTH, &format_params) < 0)
- error("%s", str);
-
- set_geo(&cylinders, &heads);
- set_label(&vlabel, &format_params, cylinders);
-
- if (g.check)
- check_disk_format(cylinders, heads, &format_params);
- else
- do_format_dasd(&vlabel, &format_params, cylinders, heads);
+ for (i = 0; i < numdev; i++)
+ {
+ strncpy(g.dev_path, g.dev_path_array[i], strlen(g.dev_path_array[i])+1);
+ strncpy(g.dev_node, g.dev_node_array[i], strlen(g.dev_node_array[i])+1);
+ process_dasd(&vlabel, format_params);
+ }
free(g.dev_path);
free(g.dev_node);

View File

@ -0,0 +1,176 @@
From a61154fd93122f5a0f2b74f21c3ac29eb437f150 Mon Sep 17 00:00:00 2001
From: Hannes Reinecke <hare@suse.de>
Date: Fri, 6 Oct 2017 09:39:36 +0200
Subject: [PATCH] dasdfmt: Format devices in parallel
Allow dasdfmt to run in parallel when several devices are specified.
Signed-off-by: Hannes Reinecke <hare@suse.com>
---
dasdfmt/dasdfmt.8 | 16 +++++++++++++-
dasdfmt/dasdfmt.c | 58 ++++++++++++++++++++++++++++++++++++++++++------------
2 files changed, 60 insertions(+), 14 deletions(-)
--- a/dasdfmt/dasdfmt.8
+++ b/dasdfmt/dasdfmt.8
@@ -7,7 +7,7 @@
dasdfmt \- formatting of DASD (ECKD) disk drives.
.SH SYNOPSIS
-\fBdasdfmt\fR [\-h] [\-t] [\-v] [\-y] [\-p] [\-P] [\-m \fIstep\fR]
+\fBdasdfmt\fR [\-h] [\-t] [\-v] [\-y] [\-p] [\-Q] [\-P] [\-m \fIstep\fR]
.br
[\-r \fIcylinder\fR] [\-b \fIblksize\fR] [\-l \fIvolser\fR] [\-d \fIlayout\fR]
.br
@@ -95,7 +95,7 @@
running in background or redirecting the output to a file.
.TP
-\fB\-P\fR or \fB\-\-percentage\fR
+\fB\-Q\fR or \fB\-\-percentage\fR
Print one line for each formatted cylinder showing the number of the
cylinder and percentage of formatting process.
Intended to be used by higher level interfaces.
@@ -164,6 +164,18 @@
.TP
\fB\-l\fR \fIvolser\fR or \fB\-\-label\fR=\fIvolser\fR
+\fB-P\fR \fInumdisks\fR or \fB--max_parallel\fR=\fInumdisks\fR
+Specify the number of disks to be formatted in parallel.
+\fInumdisks\fR specifies the number of formatting processed,
+independent on the overall number of disks to be formatted.
+The maximum value for \fInumdisks\fR is 512. Default is 1.
+.br
+Using this option can decrease overall processing time when formatting
+several disks. Please note that the I/O throughput will dramatically
+increase when using this option. Use with care.
+.br
+
+.TP
Specify the volume serial number or volume identifier to be written
to disk after formatting. If no label is specified, a sensible default
is used. \fIvolser\fR is interpreted as ASCII string and is automatically
--- a/dasdfmt/dasdfmt.c
+++ b/dasdfmt/dasdfmt.c
@@ -13,6 +13,7 @@
#include <sys/sysmacros.h>
#include <sys/time.h>
#include <sys/utsname.h>
+#include <sys/wait.h>
#include "lib/dasd_base.h"
#include "lib/dasd_sys.h"
@@ -81,6 +82,7 @@
int mode_specified;
int ese;
int no_discard;
+ int procnum;
} g = {
.dasd_info = { 0 },
};
@@ -105,6 +107,11 @@
.desc = "Perform complete format check on device",
.flags = UTIL_OPT_FLAG_NOSHORT,
},
+ {
+ .option = { "max_parallel", required_argument, NULL, 'P' },
+ .desc = "Format devices in parallel",
+ .flags = UTIL_OPT_FLAG_NOLONG,
+ },
UTIL_OPT_SECTION("FORMAT OPTIONS"),
{
.option = { "blocksize", required_argument, NULL, 'b' },
@@ -162,7 +169,7 @@
.desc = "Show a progressbar",
},
{
- .option = { "percentage", no_argument, NULL, 'P' },
+ .option = { "percentage", no_argument, NULL, 'Q' },
.desc = "Show progress in percent",
},
UTIL_OPT_SECTION("MISC"),
@@ -311,7 +318,7 @@
}
if (g.print_hashmarks && (cyl / g.hashstep - hashcount) != 0) {
- printf("#");
+ printf("%d|", g.procnum);
fflush(stdout);
hashcount++;
}
@@ -1560,7 +1567,11 @@
char *reqsize_param_str = NULL;
char *hashstep_str = NULL;
- int rc, numdev = 0, i;
+ int rc, numdev = 0, numproc = 0, status;
+ int max_parallel =1 ;
+ int running = 0;
+ int chpid;
+ int tmp;
/* Establish a handler for interrupt signals. */
signal(SIGTERM, program_interrupt_signal);
@@ -1623,7 +1634,7 @@
g.print_hashmarks = 1;
}
break;
- case 'P':
+ case 'Q':
if (!(g.print_hashmarks || g.print_progressbar))
g.print_percentage = 1;
break;
@@ -1682,6 +1693,9 @@
case OPT_NODISCARD:
g.no_discard = 1;
break;
+ case 'P':
+ max_parallel = atoi(optarg);
+ break;
case OPT_CHECK:
g.check = 1;
break;
@@ -1733,15 +1747,35 @@
if (numdev > 1 && g.labelspec)
error("Specifying a volser to be written doesn't make sense when formatting multiple DASD volumes.");
- for (i = 0; i < numdev; i++)
- {
- strncpy(g.dev_path, g.dev_path_array[i], strlen(g.dev_path_array[i])+1);
- strncpy(g.dev_node, g.dev_node_array[i], strlen(g.dev_node_array[i])+1);
- process_dasd(&vlabel, format_params);
+ for (numproc = 0; numproc < numdev; numproc++) {
+ chpid = fork();
+ if (chpid == -1 )
+ ERRMSG_EXIT(EXIT_FAILURE,
+ "%s: Unable to create child process: %s\n",
+ prog_name, strerror(errno));
+ if (!chpid) {
+ g.procnum = numproc;
+ strncpy(g.dev_path, g.dev_path_array[numproc], strlen(g.dev_path_array[numproc])+1);
+ strncpy(g.dev_node, g.dev_node_array[numproc], strlen(g.dev_node_array[numproc])+1);
+ process_dasd(&vlabel, format_params);
+
+ free(g.dev_path);
+ free(g.dev_node);
+ exit(0);
+ } else {
+ running++;
+ if (running >= max_parallel) {
+ if (wait(&tmp) > 0 && WEXITSTATUS(tmp))
+ rc = WEXITSTATUS(tmp);
+ running--;
+ }
+ }
}
- free(g.dev_path);
- free(g.dev_node);
+ /* wait until all formatting children have finished */
+ while(wait(&status) > 0)
+ if (WEXITSTATUS(status))
+ rc = WEXITSTATUS(status);
- return 0;
+ return rc;
}

View File

@ -0,0 +1,196 @@
From eabcb26fa4a91d410a6f75a9915a9ebb9f702c6b Mon Sep 17 00:00:00 2001
From: Hannes Reinecke <hare@suse.de>
Date: Fri, 6 Oct 2017 09:55:40 +0200
Subject: [PATCH] dasdfmt: Implement '-Y/--yast_mode'
Implement an option '-Y' to suppress most output.
Signed-off-by: Hannes Reinecke <hare@suse.com>
---
dasdfmt/dasdfmt.8 | 7 ++++-
dasdfmt/dasdfmt.c | 73 ++++++++++++++++++++++++++++++++++++++++++++++++------
2 files changed, 72 insertions(+), 8 deletions(-)
--- a/dasdfmt/dasdfmt.8
+++ b/dasdfmt/dasdfmt.8
@@ -7,7 +7,7 @@
dasdfmt \- formatting of DASD (ECKD) disk drives.
.SH SYNOPSIS
-\fBdasdfmt\fR [\-h] [\-t] [\-v] [\-y] [\-p] [\-Q] [\-P] [\-m \fIstep\fR]
+\fBdasdfmt\fR [\-h] [\-t] [\-v] [\-y] [\-p] [\-Q] [\-P] [\-Y] [\-m \fIstep\fR]
.br
[\-r \fIcylinder\fR] [\-b \fIblksize\fR] [\-l \fIvolser\fR] [\-d \fIlayout\fR]
.br
@@ -112,6 +112,11 @@
.br
.TP
+\fB-Y\fR or \fB--yast_mode\fR
+YaST mode; suppress most output.
+.br
+
+.TP
\fB\-M\fR \fImode\fR or \fB\-\-mode\fR=\fImode\fR
Specify the \fImode\fR to be used to format the device. Valid modes are:
.RS
--- a/dasdfmt/dasdfmt.c
+++ b/dasdfmt/dasdfmt.c
@@ -83,6 +83,7 @@
int ese;
int no_discard;
int procnum;
+ int yast_mode;
} g = {
.dasd_info = { 0 },
};
@@ -172,6 +173,10 @@
.option = { "percentage", no_argument, NULL, 'Q' },
.desc = "Show progress in percent",
},
+ {
+ .option = { "yast_mode", no_argument, NULL, 'Y' },
+ .desc = "YaST mode",
+ },
UTIL_OPT_SECTION("MISC"),
{
.option = { "check_host_count", no_argument, NULL, 'C' },
@@ -318,7 +323,9 @@
}
if (g.print_hashmarks && (cyl / g.hashstep - hashcount) != 0) {
- printf("%d|", g.procnum);
+ if (g.yast_mode)
+ printf("%d|", g.procnum);
+ else printf("#");
fflush(stdout);
hashcount++;
}
@@ -392,7 +399,7 @@
unsigned int kl = 0;
int blksize = cdata->expect.blksize;
- if (g.print_progressbar || g.print_hashmarks)
+ if ((g.print_progressbar || g.print_hashmarks) && !g.yast_mode)
printf("\n");
/*
@@ -780,8 +787,9 @@
g.hashstep = 10;
}
- printf("Printing hashmark every %d cylinders.\n",
- g.hashstep);
+ if (!g.yast_mode)
+ printf("Printing hashmark every %d cylinders.\n",
+ g.hashstep);
}
}
@@ -1462,17 +1470,19 @@
break;
}
- printf("Finished formatting the %s device.\n", g.dev_path);
+ if (!g.yast_mode)
+ printf("Finished formatting the %s device.\n", g.dev_path);
if (!(g.writenolabel || mode == EXPAND))
dasdfmt_write_labels(vlabel, cylinders, heads);
- printf("Rereading the partition table for %s... ", g.dev_path);
+ if (!g.yast_mode)
+ printf("Rereading the partition table for %s... ", g.dev_path);
err = dasd_reread_partition_table(g.dev_node, 5);
if (err != 0) {
ERRMSG("%s: error during rereading the partition "
"table: %s.\n", prog_name, strerror(err));
- } else {
+ } else if (!g.yast_mode) {
printf("ok\n");
}
}
@@ -1548,6 +1558,7 @@
error("%s", str);
set_geo(&cylinders, &heads);
+
set_label(&vlabel, &format_params, cylinders);
if (g.check)
@@ -1557,6 +1568,29 @@
}
+static void yast_print_cylinfo(const char *dev_filename)
+{
+ unsigned int cylinders = -1u;
+ int fd;
+ dasd_information2_t dasd_info;
+ struct dasd_eckd_characteristics *characteristics;
+
+ fd = open(dev_filename, O_RDONLY);
+ if ((fd != -1) && ( ! ioctl(fd, BIODASDINFO2, &dasd_info))) {
+
+ characteristics = (struct dasd_eckd_characteristics *) &dasd_info.characteristics;
+ if (characteristics->no_cyl == LV_COMPAT_CYL && characteristics->long_no_cyl)
+ cylinders = characteristics->long_no_cyl;
+ else
+ cylinders = characteristics->no_cyl;
+ }
+
+ if (fd != -1)
+ close(fd);
+ printf("%u\n", cylinders);
+ fflush(stdout);
+}
+
int main(int argc, char *argv[])
{
volume_label_t vlabel;
@@ -1693,6 +1727,10 @@
case OPT_NODISCARD:
g.no_discard = 1;
break;
+ case 'Y':
+ /* YaST mode */
+ g.yast_mode = 1;
+ break;
case 'P':
max_parallel = atoi(optarg);
break;
@@ -1728,6 +1766,21 @@
reqsize = DEFAULT_REQUESTSIZE;
}
+/* If -Y (YaST mode) was specified by the caller, then we need to suppress
+ * most of all the other output that might be generated. But, we _do_ want
+ * hashmarks printed so that YaST can track what's going on. If it wasn't
+ * specified on the command line, set it to a default of 10 cylinders.
+ */
+ if (g.yast_mode) {
+ g.verbosity = 0;
+ g.print_progressbar = 0;
+ g.print_percentage = 0;
+ if (! g.print_hashmarks) {
+ g.print_hashmarks = 1;
+ hashstep_str = "10";
+ }
+ }
+
if (g.print_hashmarks)
PARSE_PARAM_INTO(g.hashstep, hashstep_str, 10, "hashstep");
@@ -1747,6 +1800,12 @@
if (numdev > 1 && g.labelspec)
error("Specifying a volser to be written doesn't make sense when formatting multiple DASD volumes.");
+ if (g.yast_mode) {
+ for (numproc = 0; numproc < numdev; numproc++)
+ yast_print_cylinfo(g.dev_path_array[numproc]);
+
+ }
+
for (numproc = 0; numproc < numdev; numproc++) {
chpid = fork();
if (chpid == -1 )

View File

@ -0,0 +1,62 @@
From 8f05578d90df49dce6e13ee850fdc8bab84916ba Mon Sep 17 00:00:00 2001
From: Hannes Reinecke <hare@suse.de>
Date: Fri, 6 Oct 2017 12:23:32 +0200
Subject: [PATCH] dasdfmt: Implement '-f' for backwards compability
YaST is calling dasdfmt with '-f device', which used to be the old
calling convention. So to not keel over when used with an older
version of YaST we should accept this option, too.
Signed-off-by: Hannes Reinecke <hare@suse.com>
---
dasdfmt/dasdfmt.8 | 5 ++++-
dasdfmt/dasdfmt.c | 10 ++++++++++
2 files changed, 14 insertions(+), 1 deletion(-)
--- a/dasdfmt/dasdfmt.8
+++ b/dasdfmt/dasdfmt.8
@@ -11,7 +11,7 @@
.br
[\-r \fIcylinder\fR] [\-b \fIblksize\fR] [\-l \fIvolser\fR] [\-d \fIlayout\fR]
.br
- [\-L] [\-V] [\-F] [\-k] [\-C] [\-M \fImode\fR] \fIdevice\fR [\fIdevice\fR]
+ [\-L] [\-V] [\-F] [\-k] [\-C] [\-M \fImode\fR] [-f \fIdevice\fR] [\fIdevice\fR]
.SH DESCRIPTION
\fBdasdfmt\fR formats one or several DASD (ECKD) disk drive(s) to prepare them
@@ -39,6 +39,9 @@
.TP
\fB\-v\fR
Increases verbosity.
+.TP
+\fB-f\fR \fIdevice\fR or \fB--device\fR=\fIdevice\fR
+Specify device to format. For backwards compability only.
.TP
\fB\-y\fR
--- a/dasdfmt/dasdfmt.c
+++ b/dasdfmt/dasdfmt.c
@@ -113,6 +113,10 @@
.desc = "Format devices in parallel",
.flags = UTIL_OPT_FLAG_NOLONG,
},
+ {
+ .option = { "device", required_argument, NULL, 'f' },
+ .desc = "Specify device to format",
+ },
UTIL_OPT_SECTION("FORMAT OPTIONS"),
{
.option = { "blocksize", required_argument, NULL, 'b' },
@@ -1649,6 +1653,12 @@
}
g.layout_specified = 1;
break;
+ case 'f':
+ get_device_name(optind-1, argc, argv);
+ strncpy(g.dev_path_array[numdev], g.dev_path, strlen(g.dev_path));
+ strncpy(g.dev_node_array[numdev], g.dev_node, strlen(g.dev_node));
+ numdev++;
+ break;
case 'y':
g.withoutprompt = 1;
break;

View File

@ -0,0 +1,56 @@
From 943e577440d74ad7f8787af2590c8ab4579a459b Mon Sep 17 00:00:00 2001
From: Hannes Reinecke <hare@suse.de>
Date: Thu, 5 Nov 2015 10:57:38 +0100
Subject: [PATCH] dasdfmt: retry BIODASDINFO if device is busy
Modern udev have the wonderful 'feature' to sending a 'change'
event whenever a device opened with O_RDWR is closed again.
The reasoning is that the said program _might_ have changed
the partition table and hence we _might_ have missed a partition
update.
But in doing so it not only generated tons of pointless events
but also confused the hell out of other programs.
Idiots.
References: bsc#937340
Signed-off-by: Hannes Reinecke <hare@suse.de>
---
dasdfmt/dasdfmt.c | 19 ++++++++++++++++++-
1 file changed, 18 insertions(+), 1 deletion(-)
--- a/dasdfmt/dasdfmt.c
+++ b/dasdfmt/dasdfmt.c
@@ -621,7 +621,7 @@
*/
static void check_disk(void)
{
- int err;
+ int err, index = 0 ;
bool ro;
err = dasd_is_ro(g.dev_node, &ro);
@@ -631,6 +631,23 @@
if (ro)
error("Disk %s is read only!", g.dev_path);
if (!g.force) {
+ /*
+ * udev strikes again.
+ * Modern udev will issue a 'change' event whenever
+ * a device opened with O_RDWR is closed again.
+ * On the grounds that program _might_ have changed
+ * the partition table.
+ * And confusing the hell out ouf anyone else.
+ * Bah.
+ */
+ for ( index = 0 ; index < 6 ; index++ ) {
+ if (g.dasd_info.open_count > 1) {
+ dasd_get_info(g.dev_node, &g.dasd_info);
+ sleep(1);
+ }
+ else break;
+
+ }
if (g.dasd_info.open_count > 1)
error("Disk %s is in use!", g.dev_path);
}

View File

@ -0,0 +1,286 @@
Index: s390-tools-service/rust/pv/src/verify.rs
===================================================================
--- s390-tools-service.orig/rust/pv/src/verify.rs
+++ s390-tools-service/rust/pv/src/verify.rs
@@ -3,10 +3,11 @@
// Copyright IBM Corp. 2023
use core::slice;
-use log::debug;
+use log::{debug, trace};
+use openssl::error::ErrorStack;
use openssl::stack::Stack;
use openssl::x509::store::X509Store;
-use openssl::x509::{CrlStatus, X509Ref, X509StoreContext, X509};
+use openssl::x509::{CrlStatus, X509NameRef, X509Ref, X509StoreContext, X509StoreContextRef, X509};
use openssl_extensions::crl::StackableX509Crl;
use openssl_extensions::crl::X509StoreContextExtension;
@@ -82,8 +83,8 @@ impl HkdVerifier for CertVerifier {
if verified_crls.is_empty() {
bail_hkd_verify!(NoCrl);
}
- for crl in &verified_crls {
- match crl.get_by_cert(&hkd.to_owned()) {
+ for crl in verified_crls {
+ match crl.get_by_serial(hkd.serial_number()) {
CrlStatus::NotRevoked => (),
_ => bail_hkd_verify!(HdkRevoked),
}
@@ -94,21 +95,54 @@ impl HkdVerifier for CertVerifier {
}
impl CertVerifier {
+ fn quirk_crls(
+ ctx: &mut X509StoreContextRef,
+ subject: &X509NameRef,
+ ) -> Result<Stack<StackableX509Crl>, ErrorStack> {
+ match ctx.crls(subject) {
+ Ok(ret) if !ret.is_empty() => return Ok(ret),
+ _ => (),
+ }
+
+ // Armonk/Poughkeepsie fixup
+ trace!("quirk_crls: Try Locality");
+ if let Some(locality_subject) = helper::armonk_locality_fixup(subject) {
+ match ctx.crls(&locality_subject) {
+ Ok(ret) if !ret.is_empty() => return Ok(ret),
+ _ => (),
+ }
+
+ // reorder
+ trace!("quirk_crls: Try Locality+Reorder");
+ if let Ok(locality_ordered_subject) = helper::reorder_x509_names(&locality_subject) {
+ match ctx.crls(&locality_ordered_subject) {
+ Ok(ret) if !ret.is_empty() => return Ok(ret),
+ _ => (),
+ }
+ }
+ }
+
+ // reorder unchanged loaciliy subject
+ trace!("quirk_crls: Try Reorder");
+ if let Ok(ordered_subject) = helper::reorder_x509_names(subject) {
+ match ctx.crls(&ordered_subject) {
+ Ok(ret) if !ret.is_empty() => return Ok(ret),
+ _ => (),
+ }
+ }
+ // nothing found, return empty stack
+ Stack::new()
+ }
+
///Download the CLRs that a HKD refers to.
pub fn hkd_crls(&self, hkd: &X509Ref) -> Result<Stack<StackableX509Crl>> {
let mut ctx = X509StoreContext::new()?;
// Unfortunately we cannot use a dedicated function here and have to use a closure (E0434)
// Otherwise, we cannot refer to self
+ // Search for local CRLs
let mut crls = ctx.init_opt(&self.store, None, None, |ctx| {
let subject = self.ibm_z_sign_key.subject_name();
- match ctx.crls(subject) {
- Ok(crls) => Ok(crls),
- _ => {
- // reorder the name and try again
- let broken_subj = helper::reorder_x509_names(subject)?;
- ctx.crls(&broken_subj).or_else(helper::stack_err_hlp)
- }
- }
+ Self::quirk_crls(ctx, subject)
})?;
if !self.offline {
Index: s390-tools-service/rust/pv/src/verify/helper.rs
===================================================================
--- s390-tools-service.orig/rust/pv/src/verify/helper.rs
+++ s390-tools-service/rust/pv/src/verify/helper.rs
@@ -13,7 +13,7 @@ use openssl::{
error::ErrorStack,
nid::Nid,
ssl::SslFiletype,
- stack::{Stack, Stackable},
+ stack::Stack,
x509::{
store::{File, X509Lookup, X509StoreBuilder, X509StoreBuilderRef, X509StoreRef},
verify::{X509VerifyFlags, X509VerifyParam},
@@ -25,6 +25,7 @@ use openssl_extensions::{
akid::{AkidCheckResult, AkidExtension},
crl::X509StoreExtension,
};
+use std::str::from_utf8;
use std::{cmp::Ordering, ffi::c_int, time::Duration, usize};
/// Minimum security level for the keys/certificates used to establish a chain of
@@ -39,7 +40,6 @@ const SECURITY_CHAIN_MAX_LEN: c_int = 2;
/// verifies that the HKD
/// * has enough security bits
/// * is inside its validity period
-/// * issuer name is the subject name of the [`sign_key`]
/// * the Authority Key ID matches the Signing Key ID of the [`sign_key`]
pub fn verify_hkd_options(hkd: &X509Ref, sign_key: &X509Ref) -> Result<()> {
let hk_pkey = hkd.public_key()?;
@@ -53,9 +53,6 @@ pub fn verify_hkd_options(hkd: &X509Ref,
// verify that the hkd is still valid
check_validity_period(hkd.not_before(), hkd.not_after())?;
- // check if hkd.issuer_name == issuer.subject
- check_x509_name_equal(sign_key.subject_name(), hkd.issuer_name())?;
-
// verify that the AKID of the hkd matches the SKID of the issuer
if let Some(akid) = hkd.akid() {
if akid.check(sign_key) != AkidCheckResult::OK {
@@ -75,9 +72,6 @@ pub fn verify_crl(crl: &X509CrlRef, issu
return None;
}
}
-
- check_x509_name_equal(crl.issuer_name(), issuer.subject_name()).ok()?;
-
match crl.verify(issuer.public_key().ok()?.as_ref()).ok()? {
true => Some(()),
false => None,
@@ -207,7 +201,8 @@ pub fn download_crls_into_store(store: &
//Asn1StringRef::as_slice aka ASN1_STRING_get0_data gives a string without \0 delimiter
const IBM_Z_COMMON_NAME: &[u8; 43usize] = b"International Business Machines Corporation";
const IBM_Z_COUNTRY_NAME: &[u8; 2usize] = b"US";
-const IBM_Z_LOCALITY_NAME: &[u8; 12usize] = b"Poughkeepsie";
+const IBM_Z_LOCALITY_NAME_POUGHKEEPSIE: &[u8; 12usize] = b"Poughkeepsie";
+const IBM_Z_LOCALITY_NAME_ARMONK: &[u8; 6usize] = b"Armonk";
const IBM_Z_ORGANIZATIONAL_UNIT_NAME_SUFFIX: &str = "Key Signing Service";
const IBM_Z_ORGANIZATION_NAME: &[u8; 43usize] = b"International Business Machines Corporation";
const IBM_Z_STATE: &[u8; 8usize] = b"New York";
@@ -226,7 +221,8 @@ fn is_ibm_signing_cert(cert: &X509) -> b
if subj.entries().count() != IMB_Z_ENTRY_COUNT
|| !name_data_eq(subj, Nid::COUNTRYNAME, IBM_Z_COUNTRY_NAME)
|| !name_data_eq(subj, Nid::STATEORPROVINCENAME, IBM_Z_STATE)
- || !name_data_eq(subj, Nid::LOCALITYNAME, IBM_Z_LOCALITY_NAME)
+ || !(name_data_eq(subj, Nid::LOCALITYNAME, IBM_Z_LOCALITY_NAME_POUGHKEEPSIE)
+ || name_data_eq(subj, Nid::LOCALITYNAME, IBM_Z_LOCALITY_NAME_ARMONK))
|| !name_data_eq(subj, Nid::ORGANIZATIONNAME, IBM_Z_ORGANIZATION_NAME)
|| !name_data_eq(subj, Nid::COMMONNAME, IBM_Z_COMMON_NAME)
{
@@ -367,24 +363,6 @@ fn check_validity_period(not_before: &As
}
}
-fn check_x509_name_equal(lhs: &X509NameRef, rhs: &X509NameRef) -> Result<()> {
- if lhs.entries().count() != rhs.entries().count() {
- bail_hkd_verify!(IssuerMismatch);
- }
-
- for l in lhs.entries() {
- // search for the matching value in the rhs names
- // found none? -> names are not equal
- if !rhs
- .entries()
- .any(|r| l.data().as_slice() == r.data().as_slice())
- {
- bail_hkd_verify!(IssuerMismatch);
- }
- }
- Ok(())
-}
-
const NIDS_CORRECT_ORDER: [Nid; 6] = [
Nid::COUNTRYNAME,
Nid::ORGANIZATIONNAME,
@@ -407,13 +385,28 @@ pub fn reorder_x509_names(subject: &X509
Ok(correct_subj.build())
}
-pub fn stack_err_hlp<T: Stackable>(
- e: ErrorStack,
-) -> std::result::Result<Stack<T>, openssl::error::ErrorStack> {
- match e.errors().len() {
- 0 => Stack::<T>::new(),
- _ => Err(e),
+/**
+* Workaround for potential locality mismatches between CRLs and Certs
+* # Return
+* fixed subject or none if locality was not Armonk or any OpenSSL error
+*/
+pub fn armonk_locality_fixup(subject: &X509NameRef) -> Option<X509Name> {
+ if !name_data_eq(subject, Nid::LOCALITYNAME, IBM_Z_LOCALITY_NAME_ARMONK) {
+ return None;
+ }
+
+ let mut ret = X509Name::builder().ok()?;
+ for entry in subject.entries() {
+ match entry.object().nid() {
+ nid @ Nid::LOCALITYNAME => ret
+ .append_entry_by_nid(nid, from_utf8(IBM_Z_LOCALITY_NAME_POUGHKEEPSIE).ok()?)
+ .ok()?,
+ _ => {
+ ret.append_entry(entry).ok()?;
+ }
+ }
}
+ Some(ret.build())
}
#[cfg(test)]
@@ -451,20 +444,6 @@ mod test {
));
}
- #[test]
- fn x509_name_equal() {
- let sign_crt = load_gen_cert("ibm.crt");
- let hkd = load_gen_cert("host.crt");
- let other = load_gen_cert("inter_ca.crt");
-
- assert!(super::check_x509_name_equal(sign_crt.subject_name(), hkd.issuer_name()).is_ok(),);
-
- assert!(matches!(
- super::check_x509_name_equal(other.subject_name(), hkd.subject_name()),
- Err(Error::HkdVerify(IssuerMismatch))
- ));
- }
-
#[test]
fn is_ibm_z_sign_key() {
let ibm_crt = load_gen_cert("ibm.crt");
Index: s390-tools-service/rust/pv/src/verify/test.rs
===================================================================
--- s390-tools-service.orig/rust/pv/src/verify/test.rs
+++ s390-tools-service/rust/pv/src/verify/test.rs
@@ -84,7 +84,6 @@ fn verify_online() {
let inter_crt = get_cert_asset_path_string("inter_ca.crt");
let ibm_crt = get_cert_asset_path_string("ibm.crt");
let hkd_revoked = load_gen_cert("host_rev.crt");
- let hkd_inv = load_gen_cert("host_invalid_signing_key.crt");
let hkd_exp = load_gen_cert("host_crt_expired.crt");
let hkd = load_gen_cert("host.crt");
@@ -112,11 +111,6 @@ fn verify_online() {
));
assert!(matches!(
- verifier.verify(&hkd_inv),
- Err(Error::HkdVerify(IssuerMismatch))
- ));
-
- assert!(matches!(
verifier.verify(&hkd_exp),
Err(Error::HkdVerify(AfterValidity))
));
@@ -130,7 +124,6 @@ fn verify_offline() {
let ibm_crt = get_cert_asset_path_string("ibm.crt");
let ibm_crl = get_cert_asset_path_string("ibm.crl");
let hkd_revoked = load_gen_cert("host_rev.crt");
- let hkd_inv = load_gen_cert("host_invalid_signing_key.crt");
let hkd_exp = load_gen_cert("host_crt_expired.crt");
let hkd = load_gen_cert("host.crt");
@@ -149,11 +142,6 @@ fn verify_offline() {
));
assert!(matches!(
- verifier.verify(&hkd_inv),
- Err(Error::HkdVerify(IssuerMismatch))
- ));
-
- assert!(matches!(
verifier.verify(&hkd_exp),
Err(Error::HkdVerify(AfterValidity))
));

View File

@ -0,0 +1,17 @@
---
common.mak | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
--- a/common.mak
+++ b/common.mak
@@ -338,8 +338,8 @@
LDFLAGS CPPFLAGS ALL_CFLAGS ALL_CXXFLAGS ALL_LDFLAGS ALL_CPPFLAGS
ifneq ($(shell $(CC_SILENT) -dumpspecs 2>/dev/null | grep -e '[^f]no-pie'),)
- NO_PIE_CFLAGS := -fno-pie
- NO_PIE_LDFLAGS := -no-pie
+ NO_PIE_CFLAGS :=
+ NO_PIE_LDFLAGS :=
else
NO_PIE_CFLAGS :=
NO_PIE_LDFLAGS :=

View File

@ -0,0 +1,97 @@
From 3ea6d6dfd2eb120ffee4c44ff51b7e9e7a9097a6 Mon Sep 17 00:00:00 2001
From: Thomas Blume <Thomas.Blume@suse.com>
Date: Thu, 28 Mar 2024 13:32:46 +0100
Subject: [PATCH] parse ipl device for activation
ported from dracut modules
---
zdev/dracut/95zdev/parse-dasd.sh | 15 ++++++++---
zdev/dracut/95zdev/parse-zfcp.sh | 46 +++++++++++++++++++-------------
2 files changed, 39 insertions(+), 22 deletions(-)
diff --git a/zdev/dracut/95zdev/parse-dasd.sh b/zdev/dracut/95zdev/parse-dasd.sh
index a97801f..eb2fa64 100644
--- a/zdev/dracut/95zdev/parse-dasd.sh
+++ b/zdev/dracut/95zdev/parse-dasd.sh
@@ -27,9 +27,18 @@ zdev_vinfo() {
zdev_parse_rd_dasd() {
local _zdev_dasd _zdev_dasd_list
- for _zdev_dasd in $(getargs rd.dasd -d 'rd_DASD='); do
- _zdev_dasd_list="${_zdev_dasd_list:+${_zdev_dasd_list},}$_zdev_dasd"
- done
+ # autodetect active bootdev from zipl device
+ if ! getargbool 0 'rd.dasd' \
+ && [[ -f /sys/firmware/ipl/ipl_type ]] \
+ && [[ $(< /sys/firmware/ipl/ipl_type) == "ccw" ]]; then
+ read -r _ccw < /sys/firmware/ipl/device
+
+ chzdev --offline --existing --enable --active dasd "$_ccw"
+ else
+ for _zdev_dasd in $(getargs rd.dasd -d 'rd_DASD='); do
+ _zdev_dasd_list="${_zdev_dasd_list:+${_zdev_dasd_list},}$_zdev_dasd"
+ done
+ fi
echo "$_zdev_dasd_list"
}
diff --git a/zdev/dracut/95zdev/parse-zfcp.sh b/zdev/dracut/95zdev/parse-zfcp.sh
index 715aa00..6279beb 100644
--- a/zdev/dracut/95zdev/parse-zfcp.sh
+++ b/zdev/dracut/95zdev/parse-zfcp.sh
@@ -12,25 +12,33 @@
zdev_zfcp_base_args="--no-settle --yes --no-root-update --force"
-for zdev_zfcp_arg in $(getargs rd.zfcp -d 'rd_ZFCP='); do
- (
- IFS_SAVED="$IFS"
- IFS="," # did not work in front of built-in set command below
- # shellcheck disable=SC2086
- set -- $zdev_zfcp_arg
- IFS=":" args="$*"
- IFS="$IFS_SAVED"
- echo "rd.zfcp ${zdev_zfcp_arg} :" | zdev_vinfo
- if [ "$#" -eq 1 ]; then
+# autodetect active bootdev from zipl device
+if ! getargbool 0 'rd.zfcp' \
+ && [[ -f /sys/firmware/ipl/ipl_type ]] \
+ && [[ $(< /sys/firmware/ipl/ipl_type) == "fcp" ]]; then
+ chzdev --offline --existing --enable --active zfcp-host 2>&1 | zdev_vinfo
+else
+ for zdev_zfcp_arg in $(getargs rd.zfcp -d 'rd_ZFCP='); do
+ (
+ IFS_SAVED="$IFS"
+ IFS="," # did not work in front of built-in set command below
# shellcheck disable=SC2086
- chzdev --enable --persistent $zdev_zfcp_base_args \
- zfcp-host "$args" 2>&1 | zdev_vinfo
- else
- # shellcheck disable=SC2086
- chzdev --enable --persistent $zdev_zfcp_base_args \
- zfcp-lun "$args" 2>&1 | zdev_vinfo
- fi
- )
-done
+ set -- $zdev_zfcp_arg
+ IFS=":" args="$*"
+ IFS="$IFS_SAVED"
+ echo "rd.zfcp ${zdev_zfcp_arg} :" | zdev_vinfo
+ if [ "$#" -eq 1 ]; then
+ # shellcheck disable=SC2086
+ chzdev --enable --persistent $zdev_zfcp_base_args \
+ zfcp-host "$args" 2>&1 | zdev_vinfo
+ else
+ # shellcheck disable=SC2086
+ chzdev --enable --persistent $zdev_zfcp_base_args \
+ zfcp-lun "$args" 2>&1 | zdev_vinfo
+ fi
+ )
+ done
+fi
+
unset zdev_zfcp_arg
unset zdev_zfcp_base_args
--
2.44.0

View File

@ -0,0 +1,304 @@
Index: s390-tools-service/genprotimg/src/include/pv_crypto_def.h
===================================================================
--- s390-tools-service.orig/genprotimg/src/include/pv_crypto_def.h
+++ s390-tools-service/genprotimg/src/include/pv_crypto_def.h
@@ -17,7 +17,8 @@
/* IBM signing key subject */
#define PV_IBM_Z_SUBJECT_COMMON_NAME "International Business Machines Corporation"
#define PV_IBM_Z_SUBJECT_COUNTRY_NAME "US"
-#define PV_IBM_Z_SUBJECT_LOCALITY_NAME "Poughkeepsie"
+#define PV_IBM_Z_SUBJECT_LOCALITY_NAME_POUGHKEEPSIE "Poughkeepsie"
+#define PV_IBM_Z_SUBJECT_LOCALITY_NAME_ARMONK "Armonk"
#define PV_IBM_Z_SUBJECT_ORGANIZATIONONAL_UNIT_NAME_SUFFIX "Key Signing Service"
#define PV_IBM_Z_SUBJECT_ORGANIZATION_NAME "International Business Machines Corporation"
#define PV_IBM_Z_SUBJECT_STATE "New York"
Index: s390-tools-service/genprotimg/src/utils/crypto.c
===================================================================
--- s390-tools-service.orig/genprotimg/src/utils/crypto.c
+++ s390-tools-service/genprotimg/src/utils/crypto.c
@@ -664,62 +664,9 @@ static gboolean x509_name_data_by_nid_eq
return memcmp(data, y, data_len) == 0;
}
-static gboolean own_X509_NAME_ENTRY_equal(const X509_NAME_ENTRY *x,
- const X509_NAME_ENTRY *y)
-{
- const ASN1_OBJECT *x_obj = X509_NAME_ENTRY_get_object(x);
- const ASN1_STRING *x_data = X509_NAME_ENTRY_get_data(x);
- const ASN1_OBJECT *y_obj = X509_NAME_ENTRY_get_object(y);
- const ASN1_STRING *y_data = X509_NAME_ENTRY_get_data(y);
- gint x_len = ASN1_STRING_length(x_data);
- gint y_len = ASN1_STRING_length(y_data);
-
- if (x_len < 0 || x_len != y_len)
- return FALSE;
-
- /* ASN1_STRING_cmp(x_data, y_data) == 0 doesn't work because it also
- * compares the type, which is sometimes different.
- */
- return OBJ_cmp(x_obj, y_obj) == 0 &&
- memcmp(ASN1_STRING_get0_data(x_data),
- ASN1_STRING_get0_data(y_data),
- (unsigned long)x_len) == 0;
-}
-
-static gboolean own_X509_NAME_equal(const X509_NAME *x, const X509_NAME *y)
-{
- gint x_count = X509_NAME_entry_count(x);
- gint y_count = X509_NAME_entry_count(y);
-
- if (x != y && (!x || !y))
- return FALSE;
-
- if (x_count != y_count)
- return FALSE;
-
- for (gint i = 0; i < x_count; i++) {
- const X509_NAME_ENTRY *entry_i = X509_NAME_get_entry(x, i);
- gboolean entry_found = FALSE;
-
- for (gint j = 0; j < y_count; j++) {
- const X509_NAME_ENTRY *entry_j =
- X509_NAME_get_entry(y, j);
-
- if (own_X509_NAME_ENTRY_equal(entry_i, entry_j)) {
- entry_found = TRUE;
- break;
- }
- }
-
- if (!entry_found)
- return FALSE;
- }
- return TRUE;
-}
-
/* Checks whether the subject of @cert is a IBM signing key subject. For this we
* must check that the subject is equal to: 'C = US, ST = New York, L =
- * Poughkeepsie, O = International Business Machines Corporation, CN =
+ * Poughkeepsie or Armonk, O = International Business Machines Corporation, CN =
* International Business Machines Corporation' and the organization unit (OUT)
* must end with the suffix ' Key Signing Service'.
*/
@@ -743,8 +690,10 @@ static gboolean has_ibm_signing_subject(
PV_IBM_Z_SUBJECT_STATE))
return FALSE;
- if (!x509_name_data_by_nid_equal(subject, NID_localityName,
- PV_IBM_Z_SUBJECT_LOCALITY_NAME))
+ if (!(x509_name_data_by_nid_equal(subject, NID_localityName,
+ PV_IBM_Z_SUBJECT_LOCALITY_NAME_POUGHKEEPSIE) ||
+ x509_name_data_by_nid_equal(subject, NID_localityName,
+ PV_IBM_Z_SUBJECT_LOCALITY_NAME_ARMONK)))
return FALSE;
if (!x509_name_data_by_nid_equal(subject, NID_organizationName,
@@ -806,6 +755,39 @@ static X509_NAME *x509_name_reorder_attr
return g_steal_pointer(&ret);
}
+/** Replace locality 'Armonk' with 'Pougkeepsie'. If Armonk was not set return
+ * `NULL`.
+ */
+static X509_NAME *x509_armonk_locality_fixup(const X509_NAME *name)
+{
+ g_autoptr(X509_NAME) ret = NULL;
+ int pos;
+
+ /* Check if ``L=Armonk`` */
+ if (!x509_name_data_by_nid_equal((X509_NAME *)name, NID_localityName,
+ PV_IBM_Z_SUBJECT_LOCALITY_NAME_ARMONK))
+ return NULL;
+
+ ret = X509_NAME_dup(name);
+ if (!ret)
+ g_abort();
+
+ pos = X509_NAME_get_index_by_NID(ret, NID_localityName, -1);
+ if (pos == -1)
+ return NULL;
+
+ X509_NAME_ENTRY_free(X509_NAME_delete_entry(ret, pos));
+
+ /* Create a new name entry at the same position as before */
+ if (X509_NAME_add_entry_by_NID(
+ ret, NID_localityName, MBSTRING_UTF8,
+ (const unsigned char *)&PV_IBM_Z_SUBJECT_LOCALITY_NAME_POUGHKEEPSIE,
+ sizeof(PV_IBM_Z_SUBJECT_LOCALITY_NAME_POUGHKEEPSIE) - 1, pos, 0) != 1)
+ return NULL;
+
+ return g_steal_pointer(&ret);
+}
+
/* In RFC 5280 the attributes of a (subject/issuer) name is not mandatory
* ordered. The problem is that our certificates are not consistent in the order
* (see https://tools.ietf.org/html/rfc5280#section-4.1.2.4 for details).
@@ -828,24 +810,10 @@ X509_NAME *c2b_name(const X509_NAME *nam
return X509_NAME_dup((X509_NAME *)name);
}
-/* Verify that: subject(issuer) == issuer(crl) and SKID(issuer) == AKID(crl) */
+/* Verify that SKID(issuer) == AKID(crl) if available */
static gint check_crl_issuer(X509_CRL *crl, X509 *issuer, GError **err)
{
- const X509_NAME *crl_issuer = X509_CRL_get_issuer(crl);
- const X509_NAME *issuer_subject = X509_get_subject_name(issuer);
- AUTHORITY_KEYID *akid = NULL;
-
- if (!own_X509_NAME_equal(issuer_subject, crl_issuer)) {
- g_autofree char *issuer_subject_str = X509_NAME_oneline(issuer_subject,
- NULL, 0);
- g_autofree char *crl_issuer_str = X509_NAME_oneline(crl_issuer, NULL, 0);
-
- g_set_error(err, PV_CRYPTO_ERROR,
- PV_CRYPTO_ERROR_CRL_SUBJECT_ISSUER_MISMATCH,
- _("issuer mismatch:\n%s\n%s"),
- issuer_subject_str, crl_issuer_str);
- return -1;
- }
+ g_autoptr(AUTHORITY_KEYID) akid = NULL;
/* If AKID(@crl) is specified it must match with SKID(@issuer) */
akid = X509_CRL_get_ext_d2i(crl, NID_authority_key_identifier, NULL, NULL);
@@ -881,7 +849,6 @@ gint check_crl_valid_for_cert(X509_CRL *
return -1;
}
- /* check that the @crl issuer matches with the subject name of @cert*/
if (check_crl_issuer(crl, cert, err) < 0)
return -1;
@@ -910,6 +877,60 @@ gint check_crl_valid_for_cert(X509_CRL *
return 0;
}
+/* This function contains work-arounds for some known subject(CRT)<->issuer(CRL)
+ * issues.
+ */
+static STACK_OF_X509_CRL *quirk_X509_STORE_ctx_get1_crls(X509_STORE_CTX *ctx,
+ const X509_NAME *subject, GError **err)
+{
+ g_autoptr(X509_NAME) fixed_subject = NULL;
+ g_autoptr(STACK_OF_X509_CRL) ret = NULL;
+
+ ret = Pv_X509_STORE_CTX_get1_crls(ctx, subject);
+ if (ret && sk_X509_CRL_num(ret) > 0)
+ return g_steal_pointer(&ret);
+
+ /* Workaround to fix the mismatch between issuer name of the * IBM
+ * signing CRLs and the IBM signing key subject name. Locality name has
+ * changed from Poughkeepsie to Armonk.
+ */
+ fixed_subject = x509_armonk_locality_fixup(subject);
+ /* Was the locality replaced? */
+ if (fixed_subject) {
+ X509_NAME *tmp;
+
+ sk_X509_CRL_free(ret);
+ ret = Pv_X509_STORE_CTX_get1_crls(ctx, fixed_subject);
+ if (ret && sk_X509_CRL_num(ret) > 0)
+ return g_steal_pointer(&ret);
+
+ /* Workaround to fix the ordering mismatch between issuer name
+ * of the IBM signing CRLs and the IBM signing key subject name.
+ */
+ tmp = fixed_subject;
+ fixed_subject = c2b_name(fixed_subject);
+ X509_NAME_free(tmp);
+ sk_X509_CRL_free(ret);
+ ret = Pv_X509_STORE_CTX_get1_crls(ctx, fixed_subject);
+ if (ret && sk_X509_CRL_num(ret) > 0)
+ return g_steal_pointer(&ret);
+ X509_NAME_free(fixed_subject);
+ fixed_subject = NULL;
+ }
+
+ /* Workaround to fix the ordering mismatch between issuer name of the
+ * IBM signing CRLs and the IBM signing key subject name.
+ */
+ fixed_subject = c2b_name(subject);
+ sk_X509_CRL_free(ret);
+ ret = Pv_X509_STORE_CTX_get1_crls(ctx, fixed_subject);
+ if (ret && sk_X509_CRL_num(ret) > 0)
+ return g_steal_pointer(&ret);
+
+ g_set_error(err, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_NO_CRL, _("no CRL found"));
+ return NULL;
+}
+
/* Given a certificate @cert try to find valid revocation lists in @ctx. If no
* valid CRL was found NULL is returned.
*/
@@ -927,20 +948,9 @@ STACK_OF_X509_CRL *store_ctx_find_valid_
return NULL;
}
- ret = X509_STORE_CTX_get1_crls(ctx, subject);
- if (!ret) {
- /* Workaround to fix the mismatch between issuer name of the
- * IBM Z signing CRLs and the IBM Z signing key subject name.
- */
- g_autoptr(X509_NAME) broken_subject = c2b_name(subject);
-
- ret = X509_STORE_CTX_get1_crls(ctx, broken_subject);
- if (!ret) {
- g_set_error(err, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_NO_CRL,
- _("no CRL found"));
- return NULL;
- }
- }
+ ret = quirk_X509_STORE_ctx_get1_crls(ctx, subject, err);
+ if (!ret)
+ return NULL;
/* Filter out non-valid CRLs for @cert */
for (gint i = 0; i < sk_X509_CRL_num(ret); i++) {
@@ -1328,32 +1338,14 @@ gint check_chain_parameters(const STACK_
/* It's almost the same as X509_check_issed from OpenSSL does except that we
* don't check the key usage of the potential issuer. This means we check:
- * 1. issuer_name(cert) == subject_name(issuer)
- * 2. Check whether the akid(cert) (if available) matches the issuer skid
- * 3. Check that the cert algrithm matches the subject algorithm
- * 4. Verify the signature of certificate @cert is using the public key of
+ * 1. Check whether the akid(cert) (if available) matches the issuer skid
+ * 2. Check that the cert algrithm matches the subject algorithm
+ * 3. Verify the signature of certificate @cert is using the public key of
* @issuer.
*/
static gint check_host_key_issued(X509 *cert, X509 *issuer, GError **err)
{
- const X509_NAME *issuer_subject = X509_get_subject_name(issuer);
- const X509_NAME *cert_issuer = X509_get_issuer_name(cert);
- AUTHORITY_KEYID *akid = NULL;
-
- /* We cannot use X509_NAME_cmp() because it considers the order of the
- * X509_NAME_Entries.
- */
- if (!own_X509_NAME_equal(issuer_subject, cert_issuer)) {
- g_autofree char *issuer_subject_str =
- X509_NAME_oneline(issuer_subject, NULL, 0);
- g_autofree char *cert_issuer_str =
- X509_NAME_oneline(cert_issuer, NULL, 0);
- g_set_error(err, PV_CRYPTO_ERROR,
- PV_CRYPTO_ERROR_CERT_SUBJECT_ISSUER_MISMATCH,
- _("Subject issuer mismatch:\n'%s'\n'%s'"),
- issuer_subject_str, cert_issuer_str);
- return -1;
- }
+ g_autoptr(AUTHORITY_KEYID) akid = NULL;
akid = X509_get_ext_d2i(cert, NID_authority_key_identifier, NULL, NULL);
if (akid && X509_check_akid(issuer, akid) != X509_V_OK) {
Index: s390-tools-service/genprotimg/src/utils/crypto.h
===================================================================
--- s390-tools-service.orig/genprotimg/src/utils/crypto.h
+++ s390-tools-service/genprotimg/src/utils/crypto.h
@@ -75,6 +75,7 @@ void x509_pair_free(x509_pair *pair);
/* Register auto cleanup functions */
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(ASN1_INTEGER, ASN1_INTEGER_free)
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(ASN1_OCTET_STRING, ASN1_OCTET_STRING_free)
+WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(AUTHORITY_KEYID, AUTHORITY_KEYID_free)
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(BIGNUM, BN_free)
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(BIO, BIO_free_all)
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(BN_CTX, BN_CTX_free)

View File

@ -0,0 +1,224 @@
Index: s390-tools-service/include/libpv/cert.h
===================================================================
--- s390-tools-service.orig/include/libpv/cert.h
+++ s390-tools-service/include/libpv/cert.h
@@ -16,7 +16,8 @@
#define PV_IBM_Z_SUBJECT_COMMON_NAME "International Business Machines Corporation"
#define PV_IBM_Z_SUBJECT_COUNTRY_NAME "US"
-#define PV_IBM_Z_SUBJECT_LOCALITY_NAME "Poughkeepsie"
+#define PV_IBM_Z_SUBJECT_LOCALITY_NAME_POUGHKEEPSIE "Poughkeepsie"
+#define PV_IBM_Z_SUBJECT_LOCALITY_NAME_ARMONK "Armonk"
#define PV_IBM_Z_SUBJECT_ORGANIZATIONAL_UNIT_NAME_SUFFIX "Key Signing Service"
#define PV_IBM_Z_SUBJECT_ORGANIZATION_NAME "International Business Machines Corporation"
#define PV_IBM_Z_SUBJECT_STATE "New York"
Index: s390-tools-service/libpv/cert.c
===================================================================
--- s390-tools-service.orig/libpv/cert.c
+++ s390-tools-service/libpv/cert.c
@@ -857,7 +857,7 @@ static gboolean x509_name_data_by_nid_eq
/* Checks whether the subject of @cert is a IBM signing key subject. For this we
* must check that the subject is equal to: 'C = US, ST = New York, L =
- * Poughkeepsie, O = International Business Machines Corporation, CN =
+ * Poughkeepsie or Armonk, O = International Business Machines Corporation, CN =
* International Business Machines Corporation' and the organization unit (OUT)
* must end with the suffix ' Key Signing Service'.
*/
@@ -879,7 +879,10 @@ static gboolean has_ibm_signing_subject(
if (!x509_name_data_by_nid_equal(subject, NID_stateOrProvinceName, PV_IBM_Z_SUBJECT_STATE))
return FALSE;
- if (!x509_name_data_by_nid_equal(subject, NID_localityName, PV_IBM_Z_SUBJECT_LOCALITY_NAME))
+ if (!(x509_name_data_by_nid_equal(subject, NID_localityName,
+ PV_IBM_Z_SUBJECT_LOCALITY_NAME_POUGHKEEPSIE) ||
+ x509_name_data_by_nid_equal(subject, NID_localityName,
+ PV_IBM_Z_SUBJECT_LOCALITY_NAME_ARMONK)))
return FALSE;
if (!x509_name_data_by_nid_equal(subject, NID_organizationName,
@@ -1085,10 +1088,9 @@ static int check_signature_algo_match(co
/* It's almost the same as X509_check_issed from OpenSSL does except that we
* don't check the key usage of the potential issuer. This means we check:
- * 1. issuer_name(cert) == subject_name(issuer)
- * 2. Check whether the akid(cert) (if available) matches the issuer skid
- * 3. Check that the cert algrithm matches the subject algorithm
- * 4. Verify the signature of certificate @cert is using the public key of
+ * 1. Check whether the akid(cert) (if available) matches the issuer skid
+ * 2. Check that the cert algrithm matches the subject algorithm
+ * 3. Verify the signature of certificate @cert is using the public key of
* @issuer.
*/
static int check_host_key_issued(X509 *cert, X509 *issuer, GError **error)
@@ -1097,19 +1099,6 @@ static int check_host_key_issued(X509 *c
const X509_NAME *cert_issuer = X509_get_issuer_name(cert);
g_autoptr(AUTHORITY_KEYID) akid = NULL;
- /* We cannot use X509_NAME_cmp() because it considers the order of the
- * X509_NAME_Entries.
- */
- if (!own_X509_NAME_equal(issuer_subject, cert_issuer)) {
- g_autofree char *issuer_subject_str = pv_X509_NAME_oneline(issuer_subject);
- g_autofree char *cert_issuer_str = pv_X509_NAME_oneline(cert_issuer);
-
- g_set_error(error, PV_CERT_ERROR, PV_CERT_ERROR_CERT_SUBJECT_ISSUER_MISMATCH,
- _("Subject issuer mismatch:\n'%s'\n'%s'"), issuer_subject_str,
- cert_issuer_str);
- return -1;
- }
-
akid = X509_get_ext_d2i(cert, NID_authority_key_identifier, NULL, NULL);
if (akid && X509_check_akid(issuer, akid) != X509_V_OK) {
g_set_error(error, PV_CERT_ERROR, PV_CERT_ERROR_SKID_AKID_MISMATCH,
@@ -1286,21 +1275,10 @@ int pv_verify_cert(X509_STORE_CTX *ctx,
return 0;
}
-/* Verify that: subject(issuer) == issuer(crl) and SKID(issuer) == AKID(crl) */
+/* Verify that SKID(issuer) == AKID(crl) */
static int check_crl_issuer(X509_CRL *crl, X509 *issuer, GError **error)
{
- const X509_NAME *crl_issuer = X509_CRL_get_issuer(crl);
- const X509_NAME *issuer_subject = X509_get_subject_name(issuer);
- AUTHORITY_KEYID *akid = NULL;
-
- if (!own_X509_NAME_equal(issuer_subject, crl_issuer)) {
- g_autofree char *issuer_subject_str = pv_X509_NAME_oneline(issuer_subject);
- g_autofree char *crl_issuer_str = pv_X509_NAME_oneline(crl_issuer);
-
- g_set_error(error, PV_CERT_ERROR, PV_CERT_ERROR_CRL_SUBJECT_ISSUER_MISMATCH,
- _("issuer mismatch:\n%s\n%s"), issuer_subject_str, crl_issuer_str);
- return -1;
- }
+ g_autoptr(AUTHORITY_KEYID) akid = NULL;
/* If AKID(@crl) is specified it must match with SKID(@issuer) */
akid = X509_CRL_get_ext_d2i(crl, NID_authority_key_identifier, NULL, NULL);
@@ -1325,7 +1303,6 @@ int pv_verify_crl(X509_CRL *crl, X509 *c
return -1;
}
- /* check that the @crl issuer matches with the subject name of @cert*/
if (check_crl_issuer(crl, cert, error) < 0)
return -1;
@@ -1393,6 +1370,93 @@ int pv_check_chain_parameters(const STAC
return 0;
}
+/** Replace locality 'Armonk' with 'Pougkeepsie'. If Armonk was not set return
+ * `NULL`.
+ */
+static X509_NAME *x509_armonk_locality_fixup(const X509_NAME *name)
+{
+ g_autoptr(X509_NAME) ret = NULL;
+ int pos;
+
+ /* Check if ``L=Armonk`` */
+ if (!x509_name_data_by_nid_equal((X509_NAME *)name, NID_localityName,
+ PV_IBM_Z_SUBJECT_LOCALITY_NAME_ARMONK))
+ return NULL;
+
+ ret = X509_NAME_dup(name);
+ if (!ret)
+ g_abort();
+
+ pos = X509_NAME_get_index_by_NID(ret, NID_localityName, -1);
+ if (pos == -1)
+ return NULL;
+
+ X509_NAME_ENTRY_free(X509_NAME_delete_entry(ret, pos));
+
+ /* Create a new name entry at the same position as before */
+ if (X509_NAME_add_entry_by_NID(
+ ret, NID_localityName, MBSTRING_UTF8,
+ (const unsigned char *)&PV_IBM_Z_SUBJECT_LOCALITY_NAME_POUGHKEEPSIE,
+ sizeof(PV_IBM_Z_SUBJECT_LOCALITY_NAME_POUGHKEEPSIE) - 1, pos, 0) != 1)
+ return NULL;
+
+ return g_steal_pointer(&ret);
+}
+
+/* This function contains work-arounds for some known subject(CRT)<->issuer(CRL)
+ * issues.
+ */
+static STACK_OF_X509_CRL *quirk_X509_STORE_ctx_get1_crls(X509_STORE_CTX *ctx,
+ const X509_NAME *subject, GError **err)
+{
+ g_autoptr(X509_NAME) fixed_subject = NULL;
+ g_autoptr(STACK_OF_X509_CRL) ret = NULL;
+
+ ret = pv_X509_STORE_CTX_get1_crls(ctx, subject);
+ if (ret && sk_X509_CRL_num(ret) > 0)
+ return g_steal_pointer(&ret);
+
+ /* Workaround to fix the mismatch between issuer name of the * IBM
+ * signing CRLs and the IBM signing key subject name. Locality name has
+ * changed from Poughkeepsie to Armonk.
+ */
+ fixed_subject = x509_armonk_locality_fixup(subject);
+ /* Was the locality replaced? */
+ if (fixed_subject) {
+ X509_NAME *tmp;
+
+ sk_X509_CRL_free(ret);
+ ret = pv_X509_STORE_CTX_get1_crls(ctx, fixed_subject);
+ if (ret && sk_X509_CRL_num(ret) > 0)
+ return g_steal_pointer(&ret);
+
+ /* Workaround to fix the ordering mismatch between issuer name
+ * of the IBM signing CRLs and the IBM signing key subject name.
+ */
+ tmp = fixed_subject;
+ fixed_subject = pv_c2b_name(fixed_subject);
+ X509_NAME_free(tmp);
+ sk_X509_CRL_free(ret);
+ ret = pv_X509_STORE_CTX_get1_crls(ctx, fixed_subject);
+ if (ret && sk_X509_CRL_num(ret) > 0)
+ return g_steal_pointer(&ret);
+ X509_NAME_free(fixed_subject);
+ fixed_subject = NULL;
+ }
+
+ /* Workaround to fix the ordering mismatch between issuer name of the
+ * IBM signing CRLs and the IBM signing key subject name.
+ */
+ fixed_subject = pv_c2b_name(subject);
+ sk_X509_CRL_free(ret);
+ ret = pv_X509_STORE_CTX_get1_crls(ctx, fixed_subject);
+ if (ret && sk_X509_CRL_num(ret) > 0)
+ return g_steal_pointer(&ret);
+
+ g_set_error(err, PV_CERT_ERROR, PV_CERT_ERROR_NO_CRL, _("no CRL found"));
+ return NULL;
+}
+
/* Given a certificate @cert try to find valid revocation lists in @ctx. If no
* valid CRL was found NULL is returned.
*/
@@ -1412,21 +1476,9 @@ STACK_OF_X509_CRL *pv_store_ctx_find_val
return NULL;
}
- ret = pv_X509_STORE_CTX_get1_crls(ctx, subject);
- if (!ret) {
- /* Workaround to fix the mismatch between issuer name of the
- * IBM Z signing CRLs and the IBM Z signing key subject name.
- */
- g_autoptr(X509_NAME) broken_subject = pv_c2b_name(subject);
-
- ret = pv_X509_STORE_CTX_get1_crls(ctx, broken_subject);
- if (!ret) {
- g_set_error(error, PV_CERT_ERROR, PV_CERT_ERROR_NO_CRL, _("no CRL found"));
- g_info("ERROR: %s", (*error)->message);
- return NULL;
- }
- }
-
+ ret = quirk_X509_STORE_ctx_get1_crls(ctx, subject, error);
+ if (!ret)
+ return NULL;
/* Filter out non-valid CRLs for @cert */
for (int i = 0; i < sk_X509_CRL_num(ret); i++) {
X509_CRL *crl = sk_X509_CRL_value(ret, i);

View File

@ -0,0 +1,25 @@
Index: s390-tools-service/pvattest/src/argparse.c
===================================================================
--- s390-tools-service.orig/pvattest/src/argparse.c
+++ s390-tools-service/pvattest/src/argparse.c
@@ -190,13 +190,13 @@ static gboolean hex_str_toull(const char
}
/* NOTE REQUIRED */
-#define _entry_root_ca(__arg_data, __indent) \
- { \
- .long_name = "root-ca", .short_name = 0, .flags = G_OPTION_FLAG_NONE, \
- .arg = G_OPTION_ARG_FILENAME_ARRAY, .arg_data = __arg_data, \
- .description = "Use FILE as the trusted root CA instead the\n" __indent \
- "root CAs that are installed on the system (optional).\n", \
- .arg_description = "FILE", \
+#define _entry_root_ca(__arg_data, __indent) \
+ { \
+ .long_name = "root-ca", .short_name = 0, .flags = G_OPTION_FLAG_NONE, \
+ .arg = G_OPTION_ARG_FILENAME, .arg_data = __arg_data, \
+ .description = "Use FILE as the trusted root CA instead the\n" __indent \
+ "root CAs that are installed on the system (optional).\n", \
+ .arg_description = "FILE", \
}
/* NOTE REQUIRED */

View File

@ -0,0 +1,92 @@
From 0748d365a60477c96cb9f6a12e9dbe547d549e1f Mon Sep 17 00:00:00 2001
From: Marc Hartmayer <mhartmay@linux.ibm.com>
Date: Tue, 12 Mar 2024 09:33:19 +0000
Subject: [PATCH] genprotimg/**/Makefile: Fix staged installs
Fix the support for staged installs. The Makefile variable `PKGDATADIR`
uses `DESTDIR` for all Makefile target, but actually it should only be
used for the `install*` and `uninstall*` targets. [1] Fix this by using
`DESTDIR` only for `install*` targets - uninstall* targets are not
supported by s390-tools.
Before this change, if `DESTDIR` was set for staged installs,
`genprotimg` has tried to find the bootloader binaries at the temporary
installation path `$DESTDIR$(TOOLS_DATADIR)/genprotimg/` instead of
`$(TOOLS_DATADIR)/genprotimg`.
[1] https://www.gnu.org/prep/standards/html_node/DESTDIR.html
Fixes: 65b9fc442c1a ("genprotimg: introduce new tool for the creation of PV images")
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
---
genprotimg/Makefile | 6 +++---
genprotimg/boot/Makefile | 8 ++++----
genprotimg/src/Makefile | 2 +-
3 files changed, 8 insertions(+), 8 deletions(-)
diff --git a/genprotimg/Makefile b/genprotimg/Makefile
index 8c9f7048..6a2e37e4 100644
--- a/genprotimg/Makefile
+++ b/genprotimg/Makefile
@@ -3,7 +3,7 @@ include ../common.mak
.DEFAULT_GOAL := all
-PKGDATADIR := "$(DESTDIR)$(TOOLS_DATADIR)/genprotimg"
+PKGDATADIR := "$(TOOLS_DATADIR)/genprotimg"
TESTS :=
SUBDIRS := boot src man
RECURSIVE_TARGETS := all-recursive install-recursive clean-recursive
@@ -11,8 +11,8 @@ RECURSIVE_TARGETS := all-recursive install-recursive clean-recursive
all: all-recursive
install: install-recursive
- $(INSTALL) -d -m 755 "$(PKGDATADIR)"
- $(INSTALL) -g $(GROUP) -o $(OWNER) -m 755 samples/check_hostkeydoc "$(PKGDATADIR)"
+ $(INSTALL) -d -m 755 "$(DESTDIR)$(PKGDATADIR)"
+ $(INSTALL) -g $(GROUP) -o $(OWNER) -m 755 samples/check_hostkeydoc "$(DESTDIR)$(PKGDATADIR)"
clean: clean-recursive
diff --git a/genprotimg/boot/Makefile b/genprotimg/boot/Makefile
index 799df9cc..73f3c9a8 100644
--- a/genprotimg/boot/Makefile
+++ b/genprotimg/boot/Makefile
@@ -7,7 +7,7 @@ DEBUG_FILES := $(addsuffix .debug,$(FILES))
ifeq ($(HOST_ARCH),s390x)
ZIPL_DIR := $(rootdir)/zipl
ZIPL_BOOT_DIR := $(ZIPL_DIR)/boot
-PKGDATADIR := $(DESTDIR)$(TOOLS_DATADIR)/genprotimg
+PKGDATADIR := $(TOOLS_DATADIR)/genprotimg
INCLUDE_PATHS := $(ZIPL_BOOT_DIR) $(ZIPL_DIR)/include $(rootdir)/include
INCLUDE_PARMS := $(addprefix -I,$(INCLUDE_PATHS))
@@ -86,9 +86,9 @@ stage3b.elf: head.o $(ZIPL_OBJS)
@chmod a-x $@
install: stage3a.bin stage3b_reloc.bin
- $(INSTALL) -d -m 755 "$(PKGDATADIR)"
- $(INSTALL) -g $(GROUP) -o $(OWNER) -m 644 stage3a.bin "$(PKGDATADIR)"
- $(INSTALL) -g $(GROUP) -o $(OWNER) -m 644 stage3b_reloc.bin "$(PKGDATADIR)"
+ $(INSTALL) -d -m 755 "$(DESTDIR)$(PKGDATADIR)"
+ $(INSTALL) -g $(GROUP) -o $(OWNER) -m 644 stage3a.bin "$(DESTDIR)$(PKGDATADIR)"
+ $(INSTALL) -g $(GROUP) -o $(OWNER) -m 644 stage3b_reloc.bin "$(DESTDIR)$(PKGDATADIR)"
else
# Don't generate the dependency files (see `common.mak` for the
diff --git a/genprotimg/src/Makefile b/genprotimg/src/Makefile
index 08734bff..d447e6cf 100644
--- a/genprotimg/src/Makefile
+++ b/genprotimg/src/Makefile
@@ -3,7 +3,7 @@ include ../../common.mak
bin_PROGRAM = genprotimg
-PKGDATADIR ?= "$(DESTDIR)$(TOOLS_DATADIR)/genprotimg"
+PKGDATADIR ?= "$(TOOLS_DATADIR)/genprotimg"
SRC_DIR := $(dir $(realpath $(firstword $(MAKEFILE_LIST))))
TOP_SRCDIR := $(SRC_DIR)/../
ROOT_DIR = $(TOP_SRC_DIR)/../../

View File

@ -0,0 +1,28 @@
From: Jiri Bohac <jbohac@suse.cz>
References: bsc#1219471
Subject: include 59-zfcp-compat.rules in kdump initrd
kdump uses a random one of the by-path symlinks to refer to the target
partition.
With 59-zfcp-compat.rules added to the SUSE package, symlinks in the form
/dev/disk/by-path/ccw-*.*.*-zfcp-*:*-part* are created. If kdump uses this symlink when generating
the kdump initrd it will fail on boot beacause the udev rule is missing in the kdump initrd
and the symlink not created in the kdump environment.
Fix this by including 59-zfcp-compat.rules in the kdump initrd.
---
zdev/dracut/95zdev-kdump/module-setup.sh | 1 +
1 file changed, 1 insertion(+)
--- a/zdev/dracut/95zdev-kdump/module-setup.sh
+++ b/zdev/dracut/95zdev-kdump/module-setup.sh
@@ -78,6 +78,7 @@
inst_multiple /lib/s390-tools/zdev-from-dasd_mod.dasd
inst_rules "59-dasd.rules"
+ inst_rules "59-zfcp-compat.rules"
# Obtain kdump target device configuration

View File

@ -0,0 +1,112 @@
From 001c5aa5d40ffa7a40d64416c43c67004de29b8f Mon Sep 17 00:00:00 2001
From: Thomas Blume <Thomas.Blume@suse.com>
Date: Thu, 28 Mar 2024 13:32:46 +0100
Subject: [PATCH] parse ipl device for activation
ported from dracut modules
---
zdev/dracut/95zdev/parse-dasd.sh | 20 +++++++++++--
zdev/dracut/95zdev/parse-zfcp.sh | 56 +++++++++++++++++++++++++--------------
2 files changed, 54 insertions(+), 22 deletions(-)
--- a/zdev/dracut/95zdev/parse-dasd.sh
+++ b/zdev/dracut/95zdev/parse-dasd.sh
@@ -10,6 +10,8 @@
# parameters are evaluated and used to configure dasd devices.
#
+zdev_dasd_base_args="--no-settle --yes --no-root-update --force"
+
# shellcheck source=/dev/null
type zdev_parse_dasd_list > /dev/null 2>&1 || . /lib/s390-tools/zdev-from-dasd_mod.dasd
@@ -27,9 +29,21 @@
zdev_parse_rd_dasd() {
local _zdev_dasd _zdev_dasd_list
- for _zdev_dasd in $(getargs rd.dasd -d 'rd_DASD='); do
- _zdev_dasd_list="${_zdev_dasd_list:+${_zdev_dasd_list},}$_zdev_dasd"
- done
+ # autodetect active bootdev from zipl device
+ if ! getargbool 0 'rd.dasd' \
+ && [[ -f /sys/firmware/ipl/ipl_type ]] \
+ && [[ $(< /sys/firmware/ipl/ipl_type) == "ccw" ]]; then
+ read -r _ccw < /sys/firmware/ipl/device
+
+ if lszdev --offline "$_ccw" &>/dev/null; then
+ chzdev --offline --existing --enable --active $zdev_dasd_base_args \
+ dasd "$_ccw"
+ fi
+ else
+ for _zdev_dasd in $(getargs rd.dasd -d 'rd_DASD='); do
+ _zdev_dasd_list="${_zdev_dasd_list:+${_zdev_dasd_list},}$_zdev_dasd"
+ done
+ fi
echo "$_zdev_dasd_list"
}
--- a/zdev/dracut/95zdev/parse-zfcp.sh
+++ b/zdev/dracut/95zdev/parse-zfcp.sh
@@ -12,25 +12,43 @@
zdev_zfcp_base_args="--no-settle --yes --no-root-update --force"
-for zdev_zfcp_arg in $(getargs rd.zfcp -d 'rd_ZFCP='); do
- (
- IFS_SAVED="$IFS"
- IFS="," # did not work in front of built-in set command below
- # shellcheck disable=SC2086
- set -- $zdev_zfcp_arg
- IFS=":" args="$*"
- IFS="$IFS_SAVED"
- echo "rd.zfcp ${zdev_zfcp_arg} :" | zdev_vinfo
- if [ "$#" -eq 1 ]; then
- # shellcheck disable=SC2086
- chzdev --enable --persistent $zdev_zfcp_base_args \
- zfcp-host "$args" 2>&1 | zdev_vinfo
- else
+zdev_vinfo() {
+ local _zdev_vinfo_line
+ while read -r _zdev_vinfo_line || [ -n "$_zdev_vinfo_line" ]; do
+ # Prefix "<30>" represents facility LOG_DAEMON 3 and loglevel INFO 6:
+ # (facility << 3) | level.
+ echo "<30>dracut: $_zdev_vinfo_line" > /dev/kmsg
+ done
+}
+
+# autodetect active bootdev from zipl device
+if ! getargbool 0 'rd.zfcp' \
+ && [[ -f /sys/firmware/ipl/ipl_type ]] \
+ && [[ $(< /sys/firmware/ipl/ipl_type) == "fcp" ]]; then
+ chzdev --offline --existing --enable --active $zdev_zfcp_base_args \
+ zfcp-host 2>&1 | zdev_vinfo
+else
+ for zdev_zfcp_arg in $(getargs rd.zfcp -d 'rd_ZFCP='); do
+ (
+ IFS_SAVED="$IFS"
+ IFS="," # did not work in front of built-in set command below
# shellcheck disable=SC2086
- chzdev --enable --persistent $zdev_zfcp_base_args \
- zfcp-lun "$args" 2>&1 | zdev_vinfo
- fi
- )
-done
+ set -- $zdev_zfcp_arg
+ IFS=":" args="$*"
+ IFS="$IFS_SAVED"
+ echo "rd.zfcp ${zdev_zfcp_arg} :" | zdev_vinfo
+ if [ "$#" -eq 1 ]; then
+ # shellcheck disable=SC2086
+ chzdev --enable --persistent $zdev_zfcp_base_args \
+ zfcp-host "$args" 2>&1 | zdev_vinfo
+ else
+ # shellcheck disable=SC2086
+ chzdev --enable --persistent $zdev_zfcp_base_args \
+ zfcp-lun "$args" 2>&1 | zdev_vinfo
+ fi
+ )
+ done
+fi
+
unset zdev_zfcp_arg
unset zdev_zfcp_base_args

View File

@ -0,0 +1,19 @@
We strongly recommend that you get your z/OS support teams involved before installing this package.
The zdsfs command is a new feature provided by IBM with the s390-tools package in SLES12. The zdsfs command allows Linux systems to mount z/OS DASD volumes as a Linux file system. The zdsfs file system translates the z/OS data sets into Linux semantics.
Through the zdsfs file system, applications on Linux can read z/OS physical sequential data sets (PS) and partitioned data sets (PDS) on the DASD. If implemented improperly, or without the knowledge and cooperation of the systems programmers and information security professionals responsible for the z/OS system, the zdsfs command represents a potentially very serious security and data integrity exposure.
There are a number of factors to consider if you choose to install this package. A necessarily incomplete list of these would be:
- Through the zdsfs file system, whole DASD volumes are accessible to Linux
- This access is not controlled or detectable by any z/OS security or auditing mechanisms.
- This access is not controlled by any z/OS "locking" facility such as provided by ENQ, GRS, etc.
- To avoid data inconsistencies, ensure the DASD volumes are offline to z/OS before you mount them in Linux.
- To minimize security problems, you should dedicate the z/OS DASD volumes for the sole purpose of providing data to Linux.
- To share z/OS data with Linux, copy it to a dataset on that separate volume.
- Because the datasets will be accessed outside of z/OS, they will appear to have never been read after creation.
- You should ensure the datasets that Linux is to access are on a separate volume that is not used for automatic dataset allocation and that is not under System Managed Storage (SMS) control. This prevents dataset migration since they will appear to never be used (except when you update them), and it avoids unaudited access to datasets that are not intended for access by the Linux server.
- When running Linux native in an LPAR, ensure that the LPAR has access only to the specific z/OS volumes that contain the data to be accessed by Linux.
- By default, only the Linux user who mounts the zdsfs file system has access to it.
By confirming this caution, you are acknowledging that you are aware there are potential data security and integrity exposures involved in the use of this package, and that you want to install it anyway.

Some files were not shown because too many files have changed in this diff Show More