Compare commits

...

90 Commits

Author SHA1 Message Date
Jan Höppner
a6f844d642 New release s390-tools-2.22.0
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 14:08:53 +02:00
Marc Hartmayer
7f059bef13 common.mak: fix -Wnewline-eof in check_header_prereq and check_dep
printf "#include <%s>\n int main(void) {return 0;}"  "glib.h" | ( clang -DS390_TOOLS_RELEASE=2.21.0-build-20220609 -DS390_TOOLS_LIBDIR=/lib/s390-tools -DS390_TOOLS_DATADIR=/usr/share/s390-tools -DS390_TOOLS_SYSCONFDIR=/etc -DS390_TOOLS_BINDIR=/sbin -g -fstack-protector-all -W -Wall -Wformat-security -O3 -std=gnu11 -DOPENSSL_API_COMPAT=0x10101000L -I/usr/include/glib-2.0 -I/usr/lib64/glib-2.0/include -I/usr/include/sysprof-4 -pthread -Wall -Wextra -Wshadow -Wcast-align -Wwrite-strings -Wmissing-prototypes -Wmissing-declarations -Wredundant-decls -Wnested-externs -Wno-long-long -Wuninitialized -Wconversion -Wstrict-prototypes -Wpointer-arith -Wno-error=inline -Wpedantic -Werror -fPIC -I ../include -D_GNU_SOURCE  -o /dev/null -xc - )
<stdin>:2:28: error: no newline at end of file [-Werror,-Wnewline-eof]
 int main(void) {return 0;}
                           ^
1 error generated.

Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Reviewed-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:05 +02:00
Steffen Eiden
4543d23dcf Makefile: remove trailing whitespace
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Reviewed-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Reviewed-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:05 +02:00
Ingo Franzki
12f7dbbb3d zkey: Fix EP11 host library version checking
Extract the minor version and modification level separately.
Previously only the modification level has been extracted, and was
reported as minor version.

Currently no one is checking the minor version or modification level,
so it does not hurt. But maybe in the future one will check, so report
it correctly.

Signed-off-by: Ingo Franzki <ifranzki@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:05 +02:00
Ingo Franzki
f7c048d0eb zkey_kmip: Setup ext-lib once the APQNs have been configured
During plugin initialization, the external libraries such as the
EP11 or CCA host libraries are set up, if the configuration is
appropriate.

A secure identity key may be generated once the APQNs are configured,
but before the server connection is configured. Trying to re-encipher
the plugin's secure keys to a new HSM master key at that stage fails
with 'ERROR: Invalid ext lib type: 0' because the external libraries
have not been setup yet.

Change the code to setup the libraries once the APQNs have been
configured, and not only after the server connection has been
configured.

Signed-off-by: Ingo Franzki <ifranzki@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:05 +02:00
Mete Durlu
80e54ac888 hyptop: increase initial update interval
Increase initial update interval from 200ms to 1 seconds to avoid
fluctuations on the initial data output.

Signed-off-by: Mete Durlu <meted@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:05 +02:00
Steffen Eiden
26148740df pvattest/tools: Add tool for attestation
Adds:
  * extract_hdr_from_image
    a bash script to extract the SE header from an SE image.

Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Acked-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:05 +02:00
Steffen Eiden
3ab06d77fb pvattest: Create, perform, and verify attestation measurements
pvattest is a tool to attest an IBM Secure Execution guest.

In a trusted environment, one can create a request using
`pvattest create`. To get a measurement of an untrusted
IBM Secure Execution guest call 'pvattest perform'.
Again in a trusted environment, call 'pvattest verify'
to verify that the measurement is the expected one.

The tool runs on s390 and x86.
It has the same requirements like libpv and therefore
requires openssl v1.1.1+, glib2.56+, and libcurl.
Additionally, to measure, the linux kernel must provide
the Ultravisor userspace interface `uvdevice` at /dev/uv
and must be executed  on an IBM Secure Execution guest on
hardware with Ultravisor attestation support, like IBM z16 or later.

Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Reviewed-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:05 +02:00
Steffen Eiden
386392690d libpv: New library for PV tools
libpv is a collection of definitions and functions related to
Protected Virtualization (PV).
The functions cover mainly encryption (e.g. AES-GCM)
and certificates (X509). There are also helping functions for glib2.

Most of the code is extracted+refactored from `genprotimg`, which
will use this library in future.

Requires openssl v1.1.1+, glib2.56+, and libcurl.

libpv is not designed or intended to be dynamically linked or used
outside of this project. Its purpose is to avoid code duplication
as PV tools do very similar things regarding cryptography.

Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Acked-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:05 +02:00
Alexander Egorenkov
72df42ea4e zdump: Adapt man page for NGDump dump tool
Document how to use zgetdump with NGDump and NVMe.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Reviewed-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:05 +02:00
Alexander Egorenkov
6c0f429805 zdump: Implement DFI interface for NGDump
The NGDump DFI interface enables zgetdump to read and mount dumps created
with NGDump stand-alone dump.

Under the hood, the NGDump DFI delegates the task of reading and
parsing of ELF dump files to the ELF DFI interface dfi_elf.

Usage example 1:
----------------

$ zgetdump -i /dev/nvme0n1p1
General dump info:
  Dump format........: elf
  Version............: 1
  UTS node name......: t83lp49.lnxne.boe
  UTS kernel release.: 5.14.0-20210819.rc6.git0.efb8a921eec7.300.fc34.s390x
  UTS kernel version.: #1 SMP Thu Aug 19 00:22:04 CEST 2021
  System arch........: s390x (64 bit)
  CPU count (online).: 32
  Dump memory range..: 16384 MB

Memory map:
  0000000000000000 - 00000003ffffffff (16384 MB)

Usage example 2:
----------------

$ zgetdump /dev/nvme0n1p1 > dump.elf

Usage example 3:
----------------

$ zgetdump -m /dev/nvme0n1p1 /mnt
$ ls -l /mnt/dump.elf
$ zgetdump -u /mnt

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Reviewed-by: Alexander Gordeev <agordeev@linux.ibm.com>
Reviewed-by: Mikhail Zaslonko <zaslonko@linux.ibm.com>
Tested-by: Mikhail Zaslonko <zaslonko@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:05 +02:00
Alexander Egorenkov
0466760ec1 zdump: Implement DT interface for NGDump
The NGDump DT interface enables zgetdump to display various meta
information contained within a dump partition prepared for NGDump
stand-alone dump.

Usage example:
--------------

No dump yet made
----------------

$ zgetdump -d /dev/nvme0n1
Dump device info:
  Dump tool.........: Next Generation (NGDump) dump tool
  Version...........: 1
  Architecture......: s390x (64 bit)

Partition info:
  Partition number..: 1

Dump present
------------

$ zgetdump -d /dev/nvme0n1
Dump device info:
  Dump tool.........: Next Generation (NGDump) dump tool
  Version...........: 1
  Architecture......: s390x (64 bit)

Partition info:
  Partition number..: 1
Meta info:
  File..............: dump.elf

Alternative disk path
---------------------

$ zgetdump -d /dev/disk/by-id/nvme-eui.01000000010000005cd2e4c5bc845051
Dump device info:
  Dump tool.........: Next Generation (NGDump) dump tool
  Version...........: 1
  Architecture......: s390x (64 bit)

Partition info:
  Partition number..: 1
Meta info:
  File..............: dump.elf

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Reviewed-by: Alexander Gordeev <agordeev@linux.ibm.com>
Reviewed-by: Mikhail Zaslonko <zaslonko@linux.ibm.com>
Tested-by:  Mikhail Zaslonko <zaslonko@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:05 +02:00
Alexander Egorenkov
8ddc31ea77 zdump: Implement NGDump helpers
This commit introduces a new module containing various helpers for
NGDump stand-alone dump. The purpose of these helpers is to facilitate
the implementation of DT and DFI interfaces for NGDump, share code
between both, reduce duplication and enable unit testing later on.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Reviewed-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:05 +02:00
Alexander Egorenkov
60f862b21c libutil/util_part: return partition containing given block range
Return the partition number which contains the given block range,
before it was tested for exact match between a partition block range
and the one provided by user. The old behavior with exact match
should still work, this change just relaxes the partition matching
algorithm and allows one to find a partition which contains the given
block range.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Reviewed-by: Mikhail Zaslonko <zaslonko@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:05 +02:00
Alexander Egorenkov
a66f6bb075 zipl: Adapt man pages for NVMe stand-alone dump
From user perspective, there is no difference between SCSI stand-alone
dump and NVMe one.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Reviewed-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:05 +02:00
Alexander Egorenkov
2b015183aa zipl: Implement NGDump
This commit finalizes the implementation of the new stand-alone dump -
Next Gen Dump (NGDump).

NGDump stand-alone dump is a universal stand-alone dump which works for both
SCSI and NVMe disks. But currently can be used only with NVMe disks
and SCSI stand-alone dump remains the default for SCSI disks.

Currently, this stand-alone dump can be used only on IBM z15 or newer
machine generations because it requires larger amount of HSA memory offered
by firmware only on IBM z15 machines. Whereas SCSI stand-alone dump is able
to work with HSA memory of 32M, the new stand-alone dump requires 512M HSA
memory at the moment.

NGDump stand-alone dump installation is initiated by passing a path to
a NVMe disk partition to zipl via the command-line -d, similar to SCSI
stand-alone dump. zipl will then:
- build a dumper initramfs with either dracut (Fedora/RHEL/SLES) or
  initramfs-tools (Ubuntu/Debian)
- format the given NVMe partition with ext4 file system
- create a bootmap file on the newly created file system using the built
  initramfs and the currently active kernel image
- install a boot loader on the disk the given dump partition belongs to

The operations described above are destructive for the given NVMe dump
partition and the boot record(s) of its disk.

After the installation step, users can configure the dumpconf to IPL
the dumper automatically on panic or trigger a dump manually via HMC
interface.

When activated, the dumper will create a dump ELF file named "dump.elf"
on the given NVMe dump partition by using the makedumpfile tool. The kdump
compressed file format is not supported yet due to zgetdump not being able
to read such a file format. Therefore, the dumper is restricted to write
the dump only in ELF format but only kernel pages which are in use.
This will usually make the dump smaller than the original size of
/proc/vmcore and the whole dump process faster as well.

Example configuration for dumpconf:

ON_PANIC=dump   # or dump_reipl
DUMP_TYPE=nvme
FID=0x00000001
NSID=0x00000001
BOOTPROG=0
BR_LBA=0

$ systemctl enable --now dumpconf

To install the dracut support on Fedora/RHEL/SUSE:

$ make -C zipl/dracut HAVE_DRACUT=1 install

To install the initramfs-tools support on Ubuntu/Debian:

$ make -C zipl/initramfs-tools HAVE_INITRAMFS=1 install

Example of NGDump console output on SLES during dump
----------------------------------------------------

         Starting NGDump...
[    8.932343] ngdump.sh[327]: NGDump started
[    8.934739] ngdump.sh[336]: Checking for memory holes                         : [  0.0 %] /
[    8.967536] ngdump.sh[336]: Checking for memory holes                         : [100.0 %] |
[    9.076976] ngdump.sh[336]: Excluding unnecessary pages                       : [100.0 %] \
[   11.721157] ngdump.sh[336]: Copying data                                      : [  0.0 %] -
[   12.317319] ngdump.sh[336]: Copying data                                      : [ 22.5 %] /           eta: 2s
[   13.273000] ngdump.sh[336]: Copying data                                      : [100.0 %] |           eta: 0s
[   13.273244] ngdump.sh[336]: The kernel version is not supported.
[   13.273263] ngdump.sh[336]: The makedumpfile operation may be incomplete.
[   13.273281] ngdump.sh[336]: The dumpfile is saved to /ngdump/dump.elf.
[   13.273299] ngdump.sh[336]: makedumpfile Completed.

Example of NGDump console output on Ubuntu during dump
-------------------------------------------------------

[    3.240078] zdump: The dump process started for a 64-bit operating system
Loading, please wait...
Starting version 245.4-4ubuntu3.17
Begin: Starting firmware auto-configuration ... done.
Begin: Loading essential drivers ... done.
Begin: Running /scripts/init-premount ...
Begin: NGDump ...
Checking for memory holes                         : [  0.0 %] /
Checking for memory holes                         : [100.0 %] |
Excluding unnecessary pages                       : [100.0 %] \
Copying data                                      : [  0.0 %] -
Copying data                                      : [ 40.7 %] /           eta: 1s
Copying data                                      : [100.0 %] |           eta: 0s
The kernel version is not supported.
The makedumpfile operation may be incomplete.
The dumpfile is saved to /ngdump/dump.elf.
makedumpfile Completed.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Acked-by: Alexander Gordeev <agordeev@linux.ibm.com>
Tested-by: Alexander Gordeev <agordeev@linux.ibm.com>
Reviewed-by: Mikhail Zaslonko <zaslonko@linux.ibm.com>
Tested-by: Mikhail Zaslonko <zaslonko@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:05 +02:00
Alexander Egorenkov
b4142a9966 zipl: Implement zipl helper script for NGDump
The purpose of the new zipl helper script is to build a dumper initramfs
for NGDump stand-alone dump. zipl executes the helper script when
preparing a NVMe dump partition for stand-alone dump.

The zipl helper script expects a single parameter - a path to the dump
partition to store a dump on. The helper script terminates with 0 on
success and a positive value otherwise.

The zipl helper script performs the following steps:
- It builds an initramfs image suitable for dumping. To perform this step,
  the script employs the dracut tool on Fedora/RHEL/SLES and
  the initramfs-tools on Ubuntu/Debian.
- It outputs the path to the newly built initramfs image and the currently
  active kernel bzImage on the standard output.
- It prints the kernel command-line to be used by the dumper on
  the standard output.

On success, zipl expects the helper script to return at least three lines
containing:
- initrd=<path to an initramfs image>
- kernel=<path to a kernel bzImage>
- cmdline=<kernel command-line parameters>

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Acked-by: Alexander Gordeev <agordeev@linux.ibm.com>
Tested-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:05 +02:00
Alexander Egorenkov
c1d08e1380 zipl: Add initramfs-tools module for NGDump
The NGDump initramfs-tools module is required to build an initramfs image
to be used with NGDump stand-alone dump. NGDump stand-alone dump does not
use a pre-built initramfs image in contrast to SCSI stand-alone dump.
Instead, an initramfs image is built with initramfs-tools if NGDump
stand-alone dump is installed on a NVMe partition.

The NGDump initramfs-tools module ensures that all necessary tools are
present within the built initramfs and a dump of /proc/vmcore is initiated
to the chosen NVMe partition when the installed dumper is IPLed.

The NGDump intramfs-tools module installs a configuration file and
a dump script into dumper initramfs. This dump script starts at boot
shortly after the initialization of the dump target device. It reads
the aforementioned configuration file that contains the name of a dump
partition to store a dump on. The dump script reads the configuration file,
mounts then the dump target device, creates a copy of /proc/vmcore with
the makedumpfile tool on it, and then shuts down the system.

Ubuntu and Debian are the main Linux distributions targeted by this
initramfs-tools module.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Acked-by: Alexander Gordeev <agordeev@linux.ibm.com>
Tested-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:05 +02:00
Alexander Egorenkov
85c9a49e63 zipl: Add dracut module for NGDump
The NGDump dracut module is required to build an initramfs image to be used
with NGDump stand-alone dump. NGDump stand-alone dump does not
use a pre-built initramfs image in contrast to SCSI stand-alone dump.
Instead, an initramfs image is built with dracut if NGDump stand-alone dump
is installed on a NVMe partition.

The NGDump dracut module ensures that all necessary tools are present
within the built initramfs and a dump of /proc/vmcore is initiated
to the chosen NVMe partition when the installed dumper is IPLed.

The NGDump dracut module installs a new systemd service and a dump script
into dumper initramfs. The systemd init process starts the NGDump service
at boot shortly after the initialization of the dump target device.
The NGDump systemd service, in its turn, starts the dump script provided
by the new dracut module. The dump script mounts then the dump target
device, creates a copy of /proc/vmcore with the makedumpfile tool on it,
and then shuts down the system.

The NGDump dracut module can be used on any Linux distribution which
supports dracut. But the primary targets are:
- Fedora
- RHEL
- SLES

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Acked-by: Alexander Gordeev <agordeev@linux.ibm.com>
Tested-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:05 +02:00
Alexander Egorenkov
4939531fe5 zipl: Add bootmap directory parameter to bootmap_create_file()
This change adds a new parameter to the function bootmap_create_file()
that requires the caller to specify a path to the directory where
a bootmap file will be created. The primary motivation for this change is
to support NVMe stand-alone dump, in that case the job's bootmap directory
is not set.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Acked-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:05 +02:00
Alexander Egorenkov
84abb27077 zipl: Extract SCSI dump partition check from bootmap_create_device()
The purpose of this change is to simplify the function
bootmap_create_device() and make it more readable. Furthermore,
the dump partition check can be reused for NVMe dump partitions later on.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Acked-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:05 +02:00
Alexander Egorenkov
6826316350 zipl: Extract SCSI dump size estimation from bootmap_create_device()
The purpose of this change is to simplify the function
bootmap_create_device() and to make it more readable.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Reviewed-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Acked-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:05 +02:00
Alexander Egorenkov
b95a3c24b8 zipl: Extract SCSI superblock writing from bootmap_create_device()
The purpose of this change is to simplify the function
bootmap_create_device() and make it more readable.
Furthermore, make bootmap_create_device() more robust by checking the return
value of disk_write_block_aligned().

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Reviewed-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Acked-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:05 +02:00
Alexander Egorenkov
1f8496aa92 zipl: Extract SCSI dump partition handling from bootmap_create()
The purpose of this change is to simplify the function bootmap_create(),
make it more readable and prepare for new changes that will follow.
The original function bootmap_create() tried to handle two different cases
and, therefore, contained many conditional statements which had a negative
effect on readability.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Acked-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:05 +02:00
Alexander Egorenkov
9aa59bfc4b zipl: Fix stage2 disk block memory leaks in bootmap_install_stages()
This commit fixes a bug that fails to release stage2 disk blocks in case
the installation of stage1b disk blocks has failed.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Suggested-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Reviewed-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Acked-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:05 +02:00
Alexander Egorenkov
fa111404cb zipl: Extract stage2 loader handling from bootmap_create()
The purpose of this change is to simplify the function bootmap_create(),
make it more readable and prepare for new changes that will follow.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Acked-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:05 +02:00
Alexander Egorenkov
5094354c39 zipl: Implement helper function disk_is_nvme()
The new function disk_is_nvme() is a convenience function that indicates
whether the given path to a device file represents a NVMe disk.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Reviewed-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Acked-by: Stefan Haberland <sth@linux.ibm.com>
Acked-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:05 +02:00
Alexander Egorenkov
844058bf4d zipl: Extend disk_get_info() to recognize NVMe disks
Until now, NVMe disks were handled by zipl as SCSI disks. To support
NVMe stand-alone dump, it became necessary to further differentiate between
both types of disks.

There are two cases that must be handled:
1. A non-device-mapper device is a NVMe disk if it is assigned to the blkext
   device driver.
2. A device-mapper disk is a NVMe disk if the ioctl NVME_IOCTL_ID succeeds.
   This case is necessary to properly recognize NVMe disks which are
   DM devices and, therefore, not directly handled by the blkext device
   driver.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Acked-by: Stefan Haberland <sth@linux.ibm.com>
Acked-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:04 +02:00
Alexander Egorenkov
dc49b4fcf0 zdump/scsi: reuse zipl magic string definitions from boot_defs.h
To remove duplication of definitions.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Reviewed-by: Mikhail Zaslonko <zaslonko@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:04 +02:00
Alexander Egorenkov
647e8b5c37 zipl/boot: move zipl magic string definitions to boot_defs.h for reuse
To reduce duplication of definitions and facilitate reuse between zipl
and zgetdump.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Reviewed-by: Mikhail Zaslonko <zaslonko@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:04 +02:00
Alexander Egorenkov
1ae4fd18fe zfcpdump: reuse SCSI struct definitions from boot_defs.h
To remove duplication of definitions.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Reviewed-by: Mikhail Zaslonko <zaslonko@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:04 +02:00
Alexander Egorenkov
561883b1d8 zdump/scsi: reuse SCSI MBR struct definitions from boot_defs.h
To remove duplication of definitions.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Reviewed-by: Mikhail Zaslonko <zaslonko@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:04 +02:00
Alexander Egorenkov
953af2cc67 zipl/boot: move SCSI MBR struct definition to boot_defs.h for reuse
To reduce duplication of definitions and facilitate reuse between zipl
and zgetdump.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Reviewed-by: Mikhail Zaslonko <zaslonko@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:04 +02:00
Alexander Egorenkov
8f37000137 zdump/scsi: reuse boot info struct definitions from boot_defs.h
To remove duplication of definitions.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Reviewed-by: Mikhail Zaslonko <zaslonko@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:04 +02:00
Alexander Egorenkov
ba7bf2f6d3 zipl/boot: move boot info struct definitions to boot_defs.h for reuse
To reduce duplication of definitions and facilitate reuse between zipl
and zgetdump.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Reviewed-by: Mikhail Zaslonko <zaslonko@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:04 +02:00
Alexander Egorenkov
4a8e52a019 zdump/scsi: reuse SCSI superblock struct definitions from boot_defs.h
To remove duplication of definitions.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Reviewed-by: Mikhail Zaslonko <zaslonko@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:04 +02:00
Alexander Egorenkov
446ace09ad zipl/boot: move SCSI superblock struct definitions to boot_defs.h for reuse
To reduce duplication of definitions and facilitate reuse between zipl
and zgetdump.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Reviewed-by: Mikhail Zaslonko <zaslonko@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:04 +02:00
Alexander Egorenkov
012b842652 zdump/scsi: reuse boot component struct definitions from boot_defs.h
To remove duplication of definitions.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Reviewed-by: Mikhail Zaslonko <zaslonko@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:04 +02:00
Alexander Egorenkov
fed12a71cf zipl/boot: move boot component struct definitions to boot_defs.h for reuse
To reduce duplication of definitions and facilitate reuse between zipl
and zgetdump.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Reviewed-by: Mikhail Zaslonko <zaslonko@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:04 +02:00
Thomas Richter
5cd9ad13e5 cpumf/lshwc: Use util_file_read_va function
Use util_file_read_va() function to read out the event number
given the event file. Also use util_strdup() function to save
the event name. This function terminates on Out Of Memory condition.

Signed-off-by: Thomas Richter <tmricht@linux.ibm.com>
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:04 +02:00
Frank Heimes
f2d5b649e1 README.md: Add 70-chreipl-fcp-mpath.rules to the list of udev rule descriptions
All udev rules that are part of the s390-tools package are listed in the main
README.md with a brief description.
This commit adds information about the newly added rule for chreipl-fcp-mpath,
called '70-chreipl-fcp-mpath.rules'.

Closes: https://github.com/ibm-s390-linux/s390-tools/pull/136
Signed-off-by: Frank Heimes <frank.heimes@canonical.com>
Acked-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-20 13:14:04 +02:00
Jan Höppner
a957e3efd6 zcryptstats: Fix man page section in title line
The title line of the zcryptstats man page specifies section 1 (user
command) while the tool is a system command.
Change it to section 8 for system command.

Fixes: https://github.com/ibm-s390-linux/s390-tools/issues/135
Reviewd-by Ingo Franzki <ifranzki@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-03 18:40:36 +02:00
Jan Höppner
463189bf60 cpumf: Move man pages to System commands section
lscpumf, lshwc, and pai are system commands but the corresponding man
pages are located in section 1 for user commands.

Move the man pages to section 8 for system commands.

Fixes: https://github.com/ibm-s390-linux/s390-tools/issues/135
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-03 18:40:36 +02:00
Jan Höppner
aa2e8e1caa scripts: Move man pages to System commands section
dbginfo.sh, zfcpdbf, and zipl-switch-to-blscfg are system commands but
the corresponding man pages are located in section 1 for user commands.

Move the man pages to section 8 for system commands.

Fixes: https://github.com/ibm-s390-linux/s390-tools/issues/135
Reviewed-by: Steffen Maier <maier@linux.ibm.com> [zfcpdbf]
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-03 18:40:36 +02:00
Jan Höppner
7bc15537c8 genprotimg: Move man page to section 1 for user commands
genprotimg is a user command and as such installed into /usr/bin. Adapt
the man page accordingly and move it from section 8 (system commnds) to
section 1 (user commands).

Reported-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Reviewed-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-03 18:40:36 +02:00
Frank Heimes
901f082f1a s390-tools: Fix typos that were detected by lintian as 'typo-in-manual-page'
The static analysis tool for Debian packages 'lintian', especially if called
like 'lintian -EvIL +pedantic', checks manual pages for correctness and typos.
This commit fixes typos that were identified by lintian and marked with
'typo-in-manual-page' while s390-tools version 2.20.0 was packaged.

Closes: https://github.com/ibm-s390-linux/s390-tools/pull/134
Signed-off-by: Frank Heimes <frank.heimes@canonical.com>
Acked-by: Steffen Maier maier@linux.ibm.com [ziomon]
Acked-by: Ingo Franzki ifranzki@linux.ibm.com [zkey]
[hoeppner@linux.ibm.com: fix commit message]
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-03 18:40:36 +02:00
Ingo Franzki
cd2647ab78 libseckey: make secure key provider functions more type save
The provider functions are passed to OpenSSL in the dispatch array
casted to 'void (*)(void)', so they can not be type checked by the
compiler. Add function prototypes using the OpenSSL provided
function typedefs to allow the compiler to check the function
signatures.

No functional change.

Signed-off-by: Ingo Franzki <ifranzki@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-03 18:40:36 +02:00
Ingo Franzki
48bcfc3366 zkey-kmip: Fix possible use after free
Signed-off-by: Ingo Franzki <ifranzki@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-03 18:40:36 +02:00
Ingo Franzki
57f3527799 libkmipclient: Fix -Wmaybe-uninitialized warning with gcc 12
Gcc 12 produces a -Wmaybe-uninitialized warning that the content
of the tmp buffer may be uninitialized. This is a false positive.
Silence the warning by using calloc instead of malloc to ensure
that the allocated buffer is initialized.

Signed-off-by: Ingo Franzki <ifranzki@linux.ibm.com>
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-03 18:40:36 +02:00
Eduard Shishkin
f51dc05f7f zipl/Makefile: create an empty environment file /etc/ziplenv
Create an empty environment file /etc/ziplenv at make install
time if that file doesn't exist

Reviewed-by: Stefan Haberland <sth@linux.ibm.com>
Signed-off-by: Eduard Shishkin <edward6@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-03 18:40:36 +02:00
Eduard Shishkin
3f9ead5731 ipl-tools/man: update a man page with LOADPARM specifications
Reviewed-by: Stefan Haberland <sth@linux.ibm.com>
Signed-off-by: Eduard Shishkin <edward6@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-03 18:40:36 +02:00
Eduard Shishkin
b4e4cd6cd5 zipl/man: update zipl-editenv man page
Add definitions of sites, namespaces and sections.
Document the new options --site (-S) and --effective-site (-E)

Reviewed-by: Stefan Haberland <sth@linux.ibm.com>
Signed-off-by: Eduard Shishkin <edward6@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-03 18:40:36 +02:00
Eduard Shishkin
e97c85bb80 zipl-editenv: added zIPL multienvironment support
. Support multiple namespaces in the installed environment block;
. Add option --site (-S) to specify a particular namespace
  when operating on the installed environment block;
. Add option --effective-site (-E) to specify a particular
  namespace. When using in a combination with the option -l
  (--list), it displays zIPL environment that would take place if
  the specified namespace was activated at boot time;
. Modify set/unset/reset/list operations over the installed
  environment block in the case when the option --site is specified

Reviewed-by: Stefan Haberland <sth@linux.ibm.com>
Signed-off-by: Eduard Shishkin <edward6@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-03 18:40:36 +02:00
Eduard Shishkin
c6443cf977 zipl/boot: added zIPL multienvironment support
. Parse LOADPARM specified by user to find out site-ID;
. Support multiple namespaces in the environment block. Use hash
  function on "extended" strings (i.e. strings with logical
  prefixes) for the key-value store implementation;
. Added logic to activate a particular namespace by site-ID passed
  by user via LOADPARM;
. Modify the logic of evaluation of zIPL environment variables in
  the kernel command line. Use the activated namespace for the
  evaluation

Reviewed-by: Stefan Haberland <sth@linux.ibm.com>
Signed-off-by: Eduard Shishkin <edward6@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-03 18:40:36 +02:00
Eduard Shishkin
6a439e6771 zipl/src: added zIPL multienvironment support
Support multiple namespaces in environment block. Use hash
function on "extended" strings for key-value store.
Support multiple sections in environment file.
Modify the import environment precedure to be "section-aware"

Reviewed-by: Stefan Haberland <sth@linux.ibm.com>
Signed-off-by: Eduard Shishkin <edward6@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-03 18:40:36 +02:00
Marc Hartmayer
2134aff491 zipl/boot: kdump: remove static struct
Makes the code easier to read since it can have less side effects.

Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Reviewed-by: Stefan Haberland <sth@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-03 18:40:36 +02:00
Marc Hartmayer
b41ac66f36 zipl: consolidate boot and dump type definitions
Consolidate type definitions in order to avoid duplicated code, getting
better compiler support, and to avoid confusing namings for the same
thing - e.g. `blk_end` vs. `blockcnt` field name used in the old `struct
fba_dump_param` definitions. It also allows us to get rid of two
`uint64_t` casts.

While at it, fix the documentation of `struct eckd_blockptr`.

Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Reviewed-by: Stefan Haberland <sth@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-03 18:40:36 +02:00
Marc Hartmayer
9aa368af54 zipl/boot: eckd2/fba: remove two superfluous casts
Remove two superfluous casts.

Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Reviewed-by: Stefan Haberland <sth@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-03 18:40:36 +02:00
Marc Hartmayer
2e6cf81c2a zipl/boot: stage3.lds.S: put notes section explicitly at the end
...otherwise `.note.gnu.build-id` will be put at 0x0 (location for s390x
lowcore) if the linker option `--build-id` is used (which is the default
if gcc is used for linking).

Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Reviewed-by: Stefan Haberland <sth@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-03 18:40:36 +02:00
Marc Hartmayer
ae27066c15 zipl/boot: add struct tpi_info to the lowcore struct
Use the `tpi_info` struct definition, similar to the Linux kernel, in
the lowcore struct. This change allows us to use simple assignments
instead of using casts. Additionally, there is the advantage that the
lowcore definition from the s390-tools now looks more similar to the
lowcore definition used by the Linux kernel.

Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Reviewed-by: Stefan Haberland <sth@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-03 18:40:36 +02:00
Marc Hartmayer
47fcc71f96 include/s390.h: add STATIC_ASSERTION for size of lowcore
This helps to find possible bugs earlier, e.g. possible bugs in the
changes in the upcoming patches.

Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Reviewed-by: Stefan Haberland <sth@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-03 18:40:36 +02:00
Marc Hartmayer
5ac171d224 genprotimg/zipl: boot: Makefile: small refactoring
Move the common `-nostdlib` and `$(NO_PIE_LINKFLAGS) linker options to
the actual linker call. This makes the code easier to maintain.

Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Reviewed-by: Stefan Haberland <sth@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-03 18:40:36 +02:00
Marc Hartmayer
67dbf094b2 genprotimg/zipl: boot: Makefile: do not link against shared libraries
It's pretty obvious that we don't want to link against shared libraries
therefore use `-static` for linking.

`ld` call before this change:

  /usr/bin/ld -v -plugin /usr/lib/gcc/s390x-linux-gnu/9/liblto_plugin.so -plugin-opt=/usr/lib/gcc/s390x-linux-gnu/9/lto-wrapper -plugin-opt=-fresolution=/tmp/ccLfGZu4.res --build-id --eh-frame-hdr -m elf64_s390 --hash-style=gnu --as-needed -dynamic-linker /lib/ld64.so.1 -z relro -o stage3.exec -L/usr/lib/gcc/s390x-linux-gnu/9 -L/usr/lib/gcc/s390x-linux-gnu/9/../../../../s390x-linux-gnu/lib/../lib -L/usr/lib/gcc/s390x-linux-gnu/9/../../../s390x-linux-gnu -L/usr/lib/gcc/s390x-linux-gnu/9/../../../../lib -L/lib/s390x-linux-gnu -L/lib/../lib -L/usr/lib/s390x-linux-gnu -L/usr/lib/../lib -L/usr/lib/gcc/s390x-linux-gnu/9/../../../../s390x-linux-gnu/lib -L/usr/lib/gcc/s390x-linux-gnu/9/../../.. -T stage3.lds stage3.o head.o kdump3.o libc.o ebcdic.o ebcdic_conv.o sclp.o sclp_stage3.o kdump.o entry.o

`ld` call after this change:

  /usr/bin/ld -v -plugin /usr/lib/gcc/s390x-linux-gnu/9/liblto_plugin.so -plugin-opt=/usr/lib/gcc/s390x-linux-gnu/9/lto-wrapper -plugin-opt=-fresolution=/tmp/ccw0ELAp.res --build-id -m elf64_s390 --hash-style=gnu --as-needed -static -z relro -o stage3.exec -L/usr/lib/gcc/s390x-linux-gnu/9 -L/usr/lib/gcc/s390x-linux-gnu/9/../../../../s390x-linux-gnu/lib/../lib -L/usr/lib/gcc/s390x-linux-gnu/9/../../../s390x-linux-gnu -L/usr/lib/gcc/s390x-linux-gnu/9/../../../../lib -L/lib/s390x-linux-gnu -L/lib/../lib -L/usr/lib/s390x-linux-gnu -L/usr/lib/../lib -L/usr/lib/gcc/s390x-linux-gnu/9/../../../../s390x-linux-gnu/lib -L/usr/lib/gcc/s390x-linux-gnu/9/../../.. -T stage3.lds stage3.o head.o kdump3.o libc.o ebcdic.o ebcdic_conv.o sclp.o sclp_stage3.o kdump.o entry.o

Fixes: b627b8d8e1 ("Initial s390-tools-2.0.0 import")
Fixes: 3356d6f4fa ("genprotimg: boot: initial bootloader support")
Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Reviewed-by: Halil Pasic <pasic@linux.ibm.com>
Reviewed-by: Stefan Haberland <sth@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-06-03 18:40:36 +02:00
Dimitri John Ledkov
3669fd4d8a zkey: Add initramfs hook
Add hook script to allow zkey utilities to be used in initramfs.

Closes: https://github.com/ibm-s390-linux/s390-tools/pull/42
Signed-off-by: Dimitri John Ledkov <xnox@ubuntu.com>
Reviewd-by Ingo Franzki <ifranzki@linux.ibm.com>
[hoeppner@linux.ibm.com: removed / in Makefile and updated commit message]
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-05-17 13:20:27 +02:00
Jan Höppner
b01ef782f0 scripts/dumpconf: Bring installation rules in line with other scripts
Until recently dumpconf might have been installed in two different
locations, /etc/init.d/ and /lib/s390-tools/. As dumpconf is now part of
the helper scripts in the scripts directory, extend the installation
rule and add the %S390_TOOLS_VERSION% wildcard to the script along the
way for correct version output.

The installation rule in the systemd/Makefile is removed, bringing it in
line with cpictl.

Acked-by: Stefan Haberland <sth@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-05-17 13:20:27 +02:00
Jan Höppner
a177ae2439 scripts/dumpconf: Update header of the dumpconf script
Remove the leftovers of the SysV init configuration and replace it with
proper information about the tool. Also add the missing Copyright
statement and remove DUMPCONF_BIN as there is no user.

Fixes: https://github.com/ibm-s390-linux/s390-tools/issues/132
Acked-by: Stefan Haberland <sth@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-05-17 13:20:27 +02:00
Jan Höppner
13fef6dbe0 dumpconf: Move dumpconf tool to scripts directory
Modern distributions use systemd and a corresponding systemd unit file
is provided that handles the dumpconf tool. The tool is also the last
script left in etc/init.d. Move it to scripts and get rid of the init.d
directory altogether.

Acked-by: Stefan Haberland <sth@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-05-17 13:20:27 +02:00
Jan Höppner
508253c666 etc/init.d: Remove SysV related daemon scripts
For cpacfstatsd, cpuplugd, and mon_statd there are still SysV daemon
scripts available. However, modern distributions use systemd and
corresponding systemd unit files are provided and in use for a long time
already.

Remove the daemon scripts for the obsolete SysV init system as they're
not used by modern systems anymore.

Acked-by: Stefan Haberland <sth@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-05-17 13:20:27 +02:00
Matthew Rosato
773d01e674 zdev: Introduce the ap device type
The AP device type initially allows only the setting of type
attributes 'apmask' and 'aqmask'.

Reviewed-by: Tony Krowiak <akrowiak@linux.ibm.com>
Reviewed-by: Peter Oberparleiter <oberpar@linux.ibm.com>
Signed-off-by: Matthew Rosato <mjrosato@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-05-17 13:20:27 +02:00
Matthew Rosato
2da206f5a6 ap_tools: Introduce ap_tools and the ap-check tool
The ap_tools utilities are intended to be used in conjunction with
the mdevctl utility for safely managing and inspecting vfio-ap
mediated devices.  For now, this will consist of the ap-check tool
which will be driven via a call-out from mdevctl to validate a
proposed vfio-ap mediated device change.

Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Reviewed-by: Tony Krowiak <akrowiak@linux.ibm.com>
Reviewed-by: Jason J. Herne <jjherne@linux.ibm.com>
Signed-off-by: Matthew Rosato <mjrosato@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-05-17 13:20:27 +02:00
Matthew Rosato
bf71e3bb1f libap: Add library for ap / vfio-ap management tools
libap is intended to provide utility functions to be used by
tooling supporting the ap bus and vfio-ap mediated devices.

Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Reviewed-by: Tony Krowiak <akrowiak@linux.ibm.com>
Signed-off-by: Matthew Rosato <mjrosato@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-05-17 13:20:27 +02:00
Matthew Rosato
6f7982dae5 libutil: Introduce util_udev
Move code to read udev files from zdev into a utility library for use by
other tools.

Acked-by: Jan Höppner <hoeppner@linux.ibm.com>
Reviewed-by: Peter Oberparleiter <oberpar@linux.ibm.com>
Reviewed-by: Tony Krowiak <akrowiak@linux.ibm.com>
Signed-off-by: Matthew Rosato <mjrosato@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-05-17 13:20:27 +02:00
Matthew Rosato
f64b4087dd libutil: Move some zdev file reading functions into util
In preparation for sharing some zdev udev code with other libraries, move
some file operation code from zdev into util_file.

Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Reviewed-by: Tony Krowiak <akrowiak@linux.ibm.com>
Reviewed-by: Peter Oberparleiter <oberpar@linux.ibm.com>
Signed-off-by: Matthew Rosato <mjrosato@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-05-17 13:20:27 +02:00
Matthew Rosato
a739584d6f libutil: add util_exit_code
zdev uses a particular set of exit codes -- In preparation for sharing
some zdev udev code with other libraries, let's also create a libutil
include to specify reusable exit codes.  For now, let's just initialize
it with the codes from zdev we care about.

Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Reviewed-by: Jason J. Herne <jjherne@linux.ibm.com>
Reviewed-by: Peter Oberparleiter <oberpar@linux.ibm.com>
Signed-off-by: Matthew Rosato <mjrosato@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-05-17 13:20:27 +02:00
Thomas Richter
cce5f510c3 cpumf/lscpumf: Add IBM z16 extended counter set definitions
Add IBM z16 extended counter set. For each counter in this set
add the counter number, the short and long description and the
counter number.

Signed-off-by: Thomas Richter <tmricht@linux.ibm.com>
Acked-by: Sumanth Korikkar <sumanthk@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-05-17 13:20:27 +02:00
Thomas Richter
2515832469 util_arch: Add IBM z16 as known machine
Add IBM z16 as known machine to the architecture definitions.

Signed-off-by: Thomas Richter <tmricht@linux.ibm.com>
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-05-17 13:20:27 +02:00
Ingo Franzki
6c5c5f7e55 libseckey: Adapt keymgmt_match() implementation to OpenSSL
OpenSSL commit ee22a3741e3fc27c981e7f7e9bcb8d3342b0c65a changed the
OpenSSL provider's keymgmt_match() function to be not so strict with
the selector bits in regards to matching different key parts.

Adapt the secure key provider's match function accordingly.
This means, that if the public key is selected to be matched, and
the public key matches (together with any also selected parameters),
then the private key is no longer checked, although it may also be
selected to be matched. This is according to how the OpenSSL function
EVP_PKEY_eq() is supposed to behave.

OpenSSL function SSL_CTX_use_PrivateKey() calls the providers match
function to check if the private key specified matches the public key
of the certificate using EVP_PKEY_eq(). EVP_PKEY_eq() includes the
private key into the selector bits here, although the certificate
only contains the public key part.

Signed-off-by: Ingo Franzki <ifranzki@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-05-17 13:20:26 +02:00
Thomas Richter
736c69379d cpumf/pai: Add support for PAI extension 1 NNPA counters
The Processor Activity Instrumentation facility (PAI) Extension 1
adds support for the counter set for Neural Network Processing Assist
(NNPA) counters.

NNPA counter values are appended as raw data to the data report.

Signed-off-by: Thomas Richter <tmricht@linux.ibm.com>
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-05-13 11:42:54 +02:00
Thomas Richter
130880159a libcpumf: Detect PMU pai_nnpa
Add a check to detect Processor Activity Instrumentation (PAI)
Extension 1 facility unit. This covers the counter set for
Neural Network Processing Assist (NNPA) counters.

Signed-off-by: Thomas Richter <tmricht@linux.ibm.com>
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-05-13 11:42:54 +02:00
Thomas Richter
d7b1cbad8b cpumf/pai: Add Processor Activity Instrumentation tool
The Processor Activity Instrumentation facility (PAI) uses the
perf_event_open system call to report data and the crypto counter
values. The data can be inspected with the perf tool in a very
generic way using raw data report and a hex dump utility.

The pai tool also uses the perf_event_open system call to record the
data, but uses specialized event attributes for recording detailed
information such as the recording of fork, exit, exec system call
events and context switch events.
Cryptography counter values are appended as raw data attachments.

The pai command can be used for recording and reporting.
Also a list of CPUs can be specified.
The recording ring buffers are created per CPU
and are named paicrypto.XXX where XXX stands for the CPU number.
The numbering scheme has leading zeros, such as 007 for CPU number 7.

For invocation and possible command line options see the man page.

Signed-off-by: Thomas Richter <tmricht@linux.ibm.com>
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-05-13 11:42:51 +02:00
Thomas Richter
a58460b5c7 libcpumf: Function to return PMU name
Return the PMU name as found in the sysfs tree given a PMU type number.

Signed-off-by: Thomas Richter <tmricht@linux.ibm.com>
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-05-13 11:41:56 +02:00
Thomas Richter
acdce2a7a9 libcpumf: Detect PMU pai_crypto
Add a check to detect Processor Activity Instrumentation (PAI)
facility unit.

Signed-off-by: Thomas Richter <tmricht@linux.ibm.com>
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-05-13 11:41:56 +02:00
Sven Schnelle
0981df62c0 cmsfs-fuse: fix enabling of hard_remove option
Since the switch to fuse3 setting the hard_remove option with the
FUSE_LIB_OPT() macro is no longer supported. See libfuse commit
8ee553dac029 ("fuse_new(): don't accept options that don't make sense
for end-users") for details. To fix this, add an appropriate init
function which sets this option.

Fixes: https://github.com/ibm-s390-linux/s390-tools/issues/131
Fixes: e679a88d88 ("Switch from fuse2 to fuse3")
Signed-off-by: Sven Schnelle <svens@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-05-10 13:33:24 +02:00
Jan Höppner
4e28f6749b Add missing copyright header
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-05-10 13:33:24 +02:00
Thomas Richter
5f08033bb8 libcpumf: Fix doxygen comments
Fix the doxygen comments and add a description on the return values.

Signed-off-by: Thomas Richter <tmricht@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-05-10 13:33:24 +02:00
Marc Hartmayer
8231ec5f38 zipl/boot/Makefile: fix cleanup of linker script dependency files
By default, the Bash file globbing pattern '*' doesn't list files prefixed by a
dot (see https://linux.die.net/man/1/bash). Let's fix this by using the pattern
`.*` instead. While at it, add `--` option to remove possible problems with
leading dashes in filenames.

Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-05-10 13:33:24 +02:00
Marc Hartmayer
9e62005818 genprotimg/boot: disable -Warray-bounds for now
This work around fixes the gcc-12 false positive by disabling `Warray-bounds`:

  CC      genprotimg/boot/stage3a.o
  In file included from stage3a.c:14:
  In function ‘__test_facility’,
      inlined from ‘test_facility’ at ../../include/boot/s390.h:428:9,
      inlined from ‘start’ at stage3a.c:42:7:
  ../../include/boot/s390.h:418:17: error: array subscript 0 is outside array bounds of ‘void[0]’ [-Werror=array-bounds]
    418 |         return (*ptr & (0x80 >> (nr & 7))) != 0;
	|                 ^~~~

Unfortunately, there is currently no better fix available that doesn't result
in larger boot loader code sizes. Given the importancy of the boot loader file
sizes the other fixes aren't acceptable. The Linux kernel shares the
problem (but for performance reasons), take a look at the discussion
https://lore.kernel.org/lkml/yt9dzgkelelc.fsf@linux.ibm.com/ for details.

Fixes: https://github.com/ibm-s390-linux/s390-tools/issues/130
Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-05-10 13:33:24 +02:00
Eduard Shishkin
a0dba6bfdb zipl/src: Implement sorting bls entries by versions
This patch implements an additional method of sorting bls entries
by version field and uses it as default one. If sorting by version
fails, then it falls back to the most robust method of sorting by
file names

Signed-off-by: Eduard Shishkin <edward6@linux.ibm.com>
Reviewed-by: Stefan Haberland <sth@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-05-10 13:33:24 +02:00
Eduard Shishkin
ebc3384b9f zipl/src: make blsdir variable global instead of local
It is used by the next patch.
Specifically, ->filter() and ->sort() callbacks of the scandir(3)
accept only short names of directory entries to be filtered/sorted,
while we need to know absolute names to perform filtering/sorting.

Signed-off-by: Eduard Shishkin <edward6@linux.ibm.com>
Reviewed-by: Stefan Haberland <sth@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-05-10 13:33:24 +02:00
Jan Höppner
e9bf141264 Prepare for next release
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2022-04-20 13:29:04 +02:00
192 changed files with 17671 additions and 2010 deletions

7
.gitignore vendored
View File

@@ -16,12 +16,14 @@ TAGS
#
# Ignore generated executables and other generated files
#
ap_tools/ap-check
cmsfs-fuse/cmsfs-fuse
cpacfstats/cpacfstats
cpacfstats/cpacfstatsd
cpumf/chcpumf
cpumf/lscpumf
cpumf/lshwc
cpumf/pai
cpuplugd/cpuplugd
dasdfmt/dasdfmt
dasdinfo/dasdinfo
@@ -45,6 +47,8 @@ iucvterm/src/iucvconn
iucvterm/src/iucvtty
iucvterm/src/ttyrun
iucvterm/test/test_afiucv
libap/check-dep-lock
libap/check-dep-json
libekmfweb/check-dep-libekmfweb
libekmfweb/detect-openssl-version.dep
libekmfweb/libekmfweb.so
@@ -64,6 +68,8 @@ lsstp/lsstp
mon_tools/mon_fsstatd
mon_tools/mon_procd
osasnmpd/osasnmpd
pvattest/src/pvattest
pvattest/tools/exchange_info
qetharp/qetharp
qethqoat/qethqoat
systemd/cpacfstatsd.service
@@ -117,3 +123,4 @@ zkey/kmip/zkey-kmip.so
zkey/zkey
zkey/zkey-cryptsetup
zpcictl/zpcictl
**/.detect-openssl.dep.c

View File

@@ -34,6 +34,7 @@ List of all individuals having contributed content to s390-tools
- Fedor Loshakov
- Felix Beck
- Frank Blaschka
- Frank Heimes
- Frank Munzert
- Frank Pavlic
- Fritz Elfert
@@ -79,6 +80,7 @@ List of all individuals having contributed content to s390-tools
- Matthew Rosato
- Maxim Shchetynin
- Melissa Howland
- Mete Durlu
- Michael Ernst
- Michael Holzheu
- Michael Mueller

View File

@@ -1,5 +1,37 @@
Release history for s390-tools (MIT version)
--------------------------------------------
* __v2.22.0 (2022-06-20)__
For Linux kernel version: 5.18
Add new tools / libraries:
- ap_tools: Introduce ap_tools and the ap-check tool
- cpumf/pai: Add Processor Activity Instrumentation tool
- libpv: New library for PV tools
- pvattest: Add new tool to create, perform, and verify attestation measurements
- zipl/zdump: Add Next Gen Dump (NGDump) support
Changes of existing tools:
- Move man pages to System commands section (lscpumf, lshwc, pai, dbginfo.sh, zfcpdbf, zipl-switch-to-blscfg)
- README.md: Add 70-chreipl-fcp-mpath.rules to the list of udev rule descriptions
- Remove SysV related daemon scripts (cpacfstatsd, cpuplugd, mon_statd)
- genprotimg: Move man page to section 1 for user commands
- hyptop: increase initial update interval
- libseckey: Adapt keymgmt_match() implementation to OpenSSL
- libutil: Add util_exit_code
- libutil: Introduce util_udev
- zdev: Introduce the ap device type
- zipl-editenv: Add zIPL multienvironment support
- zipl: Implement sorting BLS entries by versions
- zkey: Add initramfs hook
Bug Fixes:
- cmsfs-fuse: Fix enabling of hard_remove option
- s390-tools: Fix typos that were detected by lintian as 'typo-in-manual-page'
- zkey-kmip: Fix possible use after free
- zkey: Fix EP11 host library version checking
- zkey_kmip: Setup ext-lib once the APQNs have been configured
* __v2.21.0 (2022-04-20)__
For Linux kernel version: 5.17

View File

@@ -10,13 +10,13 @@ include common.mak
#
BASELIB_DIRS = libutil libseckey
LIB_DIRS = libvtoc libzds libdasd libvmdump libccw libvmcp libekmfweb \
libkmipclient libcpumf
libkmipclient libcpumf libap libpv
TOOL_DIRS = zipl zdump fdasd dasdfmt dasdview tunedasd \
tape390 osasnmpd qetharp ip_watcher qethconf scripts zconf \
vmconvert vmcp man mon_tools dasdinfo vmur cpuplugd ipl_tools \
ziomon iucvterm hyptop cmsfs-fuse qethqoat zfcpdump zdsfs cpumf \
systemd hmcdrvfs cpacfstats zdev dump2tar zkey netboot etc zpcictl \
genprotimg lsstp hsci hsavmcore chreipl-fcp-mpath
genprotimg lsstp hsci hsavmcore chreipl-fcp-mpath ap_tools pvattest
SUB_DIRS = $(BASELIB_DIRS) $(LIB_DIRS) $(TOOL_DIRS)
@@ -33,7 +33,7 @@ endif
#
# We have to build the base libraries before the other libraries are built,
# and then build the other libraries before the tools are built. Otherwise the
# and then build the other libraries before the tools are built. Otherwise the
# other libraries and tools would trigger parallel "make -C" builds for the
# base libraries and the other libraries in case of "make -j".
#

View File

@@ -33,6 +33,9 @@ Package contents
* genprotimg:
Create a protected virtualization image.
* pvattest:
Create, perform, and verify protected virtualization attestation measurements.
* udev rules:
- 59-dasd.rules: rules for unique DASD device nodes created in /dev/disk/.
- 57-osasnmpd.rules: udev rules for osasnmpd.
@@ -40,6 +43,9 @@ Package contents
- 40-z90crypt.rules: udev rules for z90crypt driver
- 90-cpi.rules: udev rule to update Control-Program-Information when KVM is
used.
- 70-chreipl-fcp-mpath.rules: udev rules to monitor multipath events for
re-IPL path failover and to adjust the re-IPL
device in case needed.
* systemd units:
- cpi.service: Unit to apply CPI settings
@@ -50,10 +56,10 @@ Package contents
- ttyrun-getty@.service: Instance unit to manage ttyrun
* zipl:
Make DASDs or tapes bootable for system IPL or system dump.
Make DASDs, SCSIs, NVMes or tapes bootable for system IPL or system dump.
* zgetdump:
Retrieve system dumps from either tapes or DASDs.
Retrieve system dumps from either tapes, DASDs, SCSIs or NVMes.
* qetharp:
Read and flush the ARP cache on OSA Express network cards.
@@ -274,6 +280,10 @@ Package contents
Use multipath information to change the configured FCP re-IPL path on
detecting issues with the current path.
* ap-check:
A utility called by mdevctl to assist in managing vfio_ap-passthrough
devices.
For more information refer to the following publications:
* "Device Drivers, Features, and Commands" chapter "Useful Linux commands"
@@ -300,22 +310,25 @@ build options:
| net-snmp | `HAVE_SNMP` | osasnmpd |
| glibc-static | `HAVE_LIBC_STATIC` | zfcpdump |
| openssl | `HAVE_OPENSSL` | genprotimg, zkey, libekmfweb, |
| | | libkmipclient |
| | | libkmipclient, pvattest |
| cryptsetup | `HAVE_CRYPTSETUP2` | zkey-cryptsetup |
| json-c | `HAVE_JSONC` | zkey-cryptsetup, libekmfweb, |
| | | libkmipclient |
| glib2 | `HAVE_GLIB2` | genprotimg |
| libcurl | `HAVE_LIBCURL` | genprotimg, libekmfweb, libkmipclient |
| glib2 | `HAVE_GLIB2` | genprotimg, pvattest |
| libcurl | `HAVE_LIBCURL` | genprotimg, libekmfweb, libkmipclient,|
| | | pvattest |
| libxml2 | `HAVE_LIBXML2` | libkmipclient |
| systemd | `HAVE_SYSTEMD` | hsavmcore |
| liblockfile | `HAVE_LOCKFILE` | ap-check |
This table lists additional build or install options:
| __COMPONENT__ | __OPTION__ | __TOOLS__ |
|------------------|:----------------------------:|:-----------------------:|
| dracut | `HAVE_DRACUT` | zdev, chreipl-fcp-mpath |
| initramfs-tools | `HAVE_INITRAMFS` | zdev |
| | `ZDEV_ALWAYS_UPDATE_INITRD` | zdev |
| __COMPONENT__ | __OPTION__ | __TOOLS__ |
|------------------|:----------------------------:|:------------------------:|
| dracut | `HAVE_DRACUT` | zdev, chreipl-fcp-mpath, |
| | | zipl |
| initramfs-tools | `HAVE_INITRAMFS` | zdev, zipl |
| | `ZDEV_ALWAYS_UPDATE_INITRD` | zdev |
The s390-tools build process uses "pkg-config" if available and hard-coded
compiler and linker options otherwise.
@@ -337,6 +350,16 @@ the different tools are provided:
The runtime requirements are: openssl-libs (>= 1.1.0) and glib2.
* pvattest:
For building pvattest you need OpenSSL version 1.1.1 or newer
installed (openssl-devel.rpm). Also required is glib2.56 or newer
(glib2-devel.rpm) and libcurl.
Tip: you may skip the pvattest build by adding
`HAVE_OPENSSL=0`, `HAVE_LIBCURL=0`, or `HAVE_GLIB2=0`.
The runtime requirements are: openssl-libs (>= 1.1.1) and
glib2.56 or newer.
* osasnmpd:
You need at least the NET-SNMP 5.1.x package (net-snmp-devel.rpm)
installed, before building the osasnmpd subagent.
@@ -481,3 +504,10 @@ the different tools are provided:
make invocation to build a fresh man page (instead of using the pre-cooked
version) and render the README.md as HTML, make further requires pandoc and
GNU awk for the build process.
* ap-check:
For building the ap-check mdevctl callout utility you need liblockfile
version 1.14 or newer installed (liblockfile-devel.rpm). Also required is
json-c version 0.13 or newer (json-c-devel.rpm).
Tip: you may skip ap-check build by adding `HAVE_LOCKFILE=0` or `HAVE_JSONC=0`
to the make invocation.

47
ap_tools/Makefile Normal file
View File

@@ -0,0 +1,47 @@
include ../common.mak
MDEVCTL_DIR = /etc/mdevctl.d/
MDEVCTL_SCRIPTS = /etc/mdevctl.d/scripts.d/
MDEVCTL_CALLOUTS = /etc/mdevctl.d/scripts.d/callouts/
libs = $(rootdir)/libap/libap.a \
$(rootdir)/libutil/libutil.a
ifeq (${HAVE_LOCKFILE},0)
all:
$(SKIP) HAVE_LOCKFILE=0
install:
$(SKIP) HAVE_LOCKFILE=0
else ifeq (${HAVE_JSONC},0)
all:
$(SKIP) HAVE_JSONC=0
install:
$(SKIP) HAVE_JSONC=0
else
LDLIBS += -llockfile -ljson-c
all: ap-check
ap-check: ap-check.o $(libs)
install: all
@if [ ! -d $(DESTDIR)$(MDEVCTL_CALLOUTS) ]; then \
mkdir -p $(DESTDIR)$(MDEVCTL_CALLOUTS); \
chown $(OWNER).$(GROUP) $(DESTDIR)$(MDEVCTL_DIR); \
chown $(OWNER).$(GROUP) $(DESTDIR)$(MDEVCTL_SCRIPTS); \
chown $(OWNER).$(GROUP) $(DESTDIR)$(MDEVCTL_CALLOUTS); \
chmod 755 $(DESTDIR)$(MDEVCTL_DIR); \
chmod 755 $(DESTDIR)$(MDEVCTL_SCRIPTS); \
chmod 755 $(DESTDIR)$(MDEVCTL_CALLOUTS); \
fi; \
$(INSTALL) -g $(GROUP) -o $(OWNER) -m 755 ap-check \
$(DESTDIR)$(MDEVCTL_CALLOUTS)
endif
clean:
rm -f *.o *~ ap-check core
.PHONY: all install clean

911
ap_tools/ap-check.c Normal file
View File

@@ -0,0 +1,911 @@
/*
* ap-check - Validate vfio-ap mediated device configuration changes
*
* This tool in intended to be driven via the callout API of the mdevctl
* utility (https://github.com/mdevctl/mdevctl/)
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#include <dirent.h>
#include <err.h>
#include <errno.h>
#include <fcntl.h>
#include <getopt.h>
#include <stdio.h>
#include <string.h>
#include <unistd.h>
#include "lib/ap.h"
#include "lib/util_base.h"
#include "lib/util_libc.h"
#include "lib/util_opt.h"
#include "lib/util_path.h"
#include "ap-check.h"
static const struct mdevctl_action mdevctl_action_table[NUM_MDEVCTL_ACTIONS] = {
{MDEVCTL_ACTION_DEFINE, "define"},
{MDEVCTL_ACTION_LIST, "list"},
{MDEVCTL_ACTION_MODIFY, "modify"},
{MDEVCTL_ACTION_START, "start"},
{MDEVCTL_ACTION_STOP, "stop"},
{MDEVCTL_ACTION_TYPES, "types"},
{MDEVCTL_ACTION_UNDEFINE, "undefine"},
{MDEVCTL_ACTION_ATTRIBUTES, "attributes"}
};
static const struct mdevctl_event mdevctl_event_table[NUM_MDEVCTL_EVENTS] = {
{MDEVCTL_EVENT_PRE, "pre"},
{MDEVCTL_EVENT_POST, "post"},
{MDEVCTL_EVENT_GET, "get"}
};
/*
* Convert mdevctl action string to an enumerated value
*/
static enum mdevctl_action_id validate_action(char *action)
{
int i;
for (i = 0; i < NUM_MDEVCTL_ACTIONS; i++) {
if (strcmp(action, mdevctl_action_table[i].action) == 0)
return mdevctl_action_table[i].id;
}
return MDEVCTL_ACTION_UNKNOWN;
}
/*
* Convert mdevctl event string to an enumerated value
*/
static enum mdevctl_event_id validate_event(char *event)
{
int i;
for (i = 0; i < NUM_MDEVCTL_EVENTS; i++) {
if (strcmp(event, mdevctl_event_table[i].event) == 0)
return mdevctl_event_table[i].id;
}
return MDEVCTL_EVENT_UNKNOWN;
}
static struct util_opt opt_vec[] = {
UTIL_OPT_SECTION("DEVICE"),
{
.option = { "e", required_argument, NULL, 'e' },
.argument = "EVENT",
.desc = "The type of callout being issued",
},
{
.option = { "a", required_argument, NULL, 'a' },
.argument = "ACTION",
.desc = "The action being performed on the specified device",
},
{
.option = { "s", required_argument, NULL, 's' },
.argument = "STATE",
.desc = "The state of the associated mdevctl command",
},
{
.option = { "u", required_argument, NULL, 'u' },
.argument = "UUID",
.desc = "Universally Unique ID for the mediated device",
},
{
.option = { "p", required_argument, NULL, 'p' },
.argument = "PDEV",
.desc = "Parent device name, e.g. matrix",
},
{
.option = { "t", required_argument, NULL, 't' },
.argument = "TYPE",
.desc = "Mediated device type, e.g. vfio_ap-passthrough",
},
UTIL_OPT_END
};
/*
* Initialize the ap_check anchor struct.
*/
static void ap_check_init(struct ap_check_anchor *anc)
{
anc->uuid = anc->parent = anc->type = NULL;
anc->dev = vfio_ap_device_new();
anc->cleanup_lock = false;
}
/*
* Free memory of ap_check anchor struct.
*/
static void ap_check_cleanup(struct ap_check_anchor *anc)
{
if (anc->uuid)
free(anc->uuid);
if (anc->parent)
free(anc->parent);
if (anc->type)
free(anc->type);
if (anc->dev)
vfio_ap_device_free(anc->dev);
if (anc->cleanup_lock)
ap_release_lock();
}
/*
* Exit ap_check
*/
static void __noreturn ap_check_exit(struct ap_check_anchor *anc, int rc)
{
ap_check_cleanup(anc);
exit(rc);
}
/*
* parses the command line
*/
static void ap_check_parse(struct ap_check_anchor *anc,
int argc, char *argv[])
{
bool action = false, event = false, state = false, bad_opts = false;
int opt;
util_opt_init(opt_vec, NULL);
while (1) {
opt = util_opt_getopt_long(argc, argv);
if (opt == -1)
break;
switch (opt) {
case 'e':
if (event) {
bad_opts = true;
} else {
anc->event = validate_event(optarg);
event = true;
}
break;
case 'a':
if (action) {
bad_opts = true;
} else {
anc->action = validate_action(optarg);
action = true;
}
break;
case 's':
if (state) {
bad_opts = true;
} else {
/* Ignore the state */
state = true;
}
break;
case 'u':
if (anc->uuid)
bad_opts = true;
else
anc->uuid = util_strdup(optarg);
break;
case 'p':
if (anc->parent)
bad_opts = true;
else
anc->parent = util_strdup(optarg);
break;
case 't':
if (anc->type)
bad_opts = true;
else
anc->type = util_strdup(optarg);
break;
default:
fprintf(stderr, "Unknown operand\n");
ap_check_exit(anc, EXIT_FAILURE);
}
}
/* Make sure we got all expected input values */
if (!(action && event && state && anc->uuid && anc->parent &&
anc->type) || bad_opts) {
fprintf(stderr, "Duplicate or missing operand\n");
ap_check_exit(anc, EXIT_FAILURE);
}
/* Check for invalid UUID */
if (!is_valid_uuid(anc->uuid)) {
fprintf(stderr, "Invalid UUID specified\n");
ap_check_exit(anc, EXIT_FAILURE);
}
anc->dev->uuid = util_strdup(anc->uuid);
/* Check for valid type */
if (strcmp(anc->type, VFIO_AP_TYPE) != 0)
ap_check_exit(anc, APC_EXIT_UNKNOWN_TYPE);
/* Check for invalid parent - currently only 'matrix' supported */
if (strcmp(anc->parent, "matrix") != 0) {
fprintf(stderr, "Invalid parent specified\n");
ap_check_exit(anc, EXIT_FAILURE);
}
}
/*
* Call a function for each entry in a directory:
* int callback(const char *abs_path, const char *rel_path, void *data)
* Continues for all entries in the directory regardless of callback return
* code. Will return 0 or, if one or more callbacks failed, the first nonzero
* rc received.
*/
static int path_for_each(const char *path,
int (*callback)(const char *, const char *, void *),
void *data)
{
struct dirent *de;
int rc = 0;
int r = 0;
DIR *dir;
char *p;
dir = opendir(path);
if (!dir)
return -1;
while ((de = readdir(dir))) {
if (strcmp(de->d_name, ".") == 0 ||
strcmp(de->d_name, "..") == 0)
continue;
util_asprintf(&p, "%s/%s", path, de->d_name);
r = callback(p, de->d_name, data);
/* Save first nonzero return code for caller */
if (rc == 0 && r != 0)
rc = r;
free(p);
}
closedir(dir);
return rc;
}
/*
* Report an error message when the specified configuration will conflict
* with an existing device
*/
static void conflict_error(const char *uuid, unsigned int a, unsigned int d,
bool persistent)
{
if (uuid) {
if (persistent) {
fprintf(stderr,
"APQN %u.%u is defined for autostart by %s\n",
a, d, uuid);
} else {
fprintf(stderr, "APQN %u.%u already in use by %s\n",
a, d, uuid);
}
} else {
if (persistent) {
fprintf(stderr, "AQPN %u.%u is not defined for "
"vfio_ap-passthrough use by the persistent "
"ap bus mask settings\n", a, d);
} else {
fprintf(stderr, "AQPN %u.%u is not allowed for "
"vfio_ap-passthrough use by the active ap "
"bus mask settings\n", a, d);
}
}
}
/*
* Compare the list of adapters and domains for two devices, reporting error
* messages for any conflicts that occur. A conflict occurs when both devices
* have the same adapter + domain pair.
* The function below takes advantage of the fact that the lists are known to
* be sorted in numeric order; therefore we can use this information to run
* the lists in parallel rather than always starting from the beginning.
*/
static int find_apqn_conflicts(const char *uuid,
struct util_list *adapters,
struct util_list *domains,
struct util_list *adapters2,
struct util_list *domains2,
bool persistent)
{
struct vfio_ap_node *a, *a2, *d, *d2;
int rc = 0;
/* Checks for conflicts with the device */
a = util_list_start(adapters);
a2 = util_list_start(adapters2);
while ((a != NULL) && (a2 != NULL)) {
if (a->id == a2->id) {
d = util_list_start(domains);
d2 = util_list_start(domains2);
while ((d != NULL) && (d2 != NULL)) {
if (d->id == d2->id) {
/* Report error, look for more */
conflict_error(uuid, a->id, d->id,
persistent);
rc = -1;
d = util_list_next(domains, d);
d2 = util_list_next(domains2, d2);
} else if (d->id > d2->id) {
d2 = util_list_next(domains2, d2);
} else {
d = util_list_next(domains, d);
}
}
a = util_list_next(adapters, a);
a2 = util_list_next(adapters2, a2);
} else if (a->id > a2->id) {
a2 = util_list_next(adapters2, a2);
} else {
a = util_list_next(adapters, a);
}
}
return rc;
}
/*
* If the provided path maps to a valid vfio-ap device configuration,
* determine if its current configuration will conflict with the proposed
* changes.
*/
static int check_other_mdev_cfg_cb(const char *path,
const char *filename,
void *data)
{
struct other_mdev_cb_data *cbdata = data;
struct vfio_ap_device *dev = cbdata->dev;
struct vfio_ap_device *dev2 = NULL;
int rc = 0;
/* Skip anything that isn't an mdev config */
if (!is_valid_uuid(filename))
goto out;
/* Skip if this is the input device */
if (strcasecmp(cbdata->uuid, filename) == 0)
goto out;
/* Read the device config */
dev2 = vfio_ap_device_new();
if (vfio_ap_read_device_config(path, dev2) != 0)
goto out;
/* If wrong device type, skip */
if (strcmp(dev2->type, VFIO_AP_TYPE) != 0)
goto out;
/* If not AUTO device, skip */
if (dev2->manual)
goto out;
/* Perform mdev-to-mdev apqn conflict analysis */
rc = find_apqn_conflicts(filename, dev->adapters, dev->domains,
dev2->adapters, dev2->domains, true);
out:
if (dev2 != NULL)
vfio_ap_device_free(dev2);
return rc;
}
/*
* Perform conflict analysis against all other vfio-ap persistent
* configurations.
*/
int check_other_mdevs_cfg(struct ap_check_anchor *anc)
{
struct other_mdev_cb_data cb_data;
if (!util_path_is_dir(VFIO_AP_CONFIG_PATH))
return 0;
cb_data.uuid = anc->uuid;
cb_data.dev = anc->dev;
return path_for_each(VFIO_AP_CONFIG_PATH, check_other_mdev_cfg_cb,
&cb_data);
}
/*
* If the provided path maps to a valid device, determine if its current
* configuration will conflict with the proposed changes.
*/
static int check_other_mdev_sysfs_cb(const char *path, const char *filename,
void *data)
{
struct other_mdev_cb_data *cbdata = data;
struct vfio_ap_device *dev = cbdata->dev;
struct vfio_ap_device *dev2;
char *matrix_path;
char buf[80];
int rc = 0;
FILE *f;
if (!is_valid_uuid(filename) || path == NULL ||
strcasecmp(filename, cbdata->uuid) == 0)
return 0;
dev2 = vfio_ap_device_new();
matrix_path = path_get_vfio_ap_attr(filename, "matrix");
f = fopen(matrix_path, "r");
while (fgets(buf, sizeof(buf), f))
vfio_ap_parse_matrix(dev2, buf);
vfio_ap_sort_matrix_results(dev2);
fclose(f);
free(matrix_path);
/* Look for conflicts between target device and this device */
rc = find_apqn_conflicts(filename, dev->adapters, dev->domains,
dev2->adapters, dev2->domains, false);
vfio_ap_device_free(dev2);
return rc;
}
/* Run conflict analysis against all other active vfio-ap devices */
static int check_other_mdevs_sysfs(struct ap_check_anchor *anc)
{
struct other_mdev_cb_data cb_data;
char *root;
int rc = 0;
cb_data.uuid = anc->uuid;
cb_data.dev = anc->dev;
root = path_get_vfio_ap_mdev("");
if (util_path_is_dir(root))
rc = path_for_each(root, check_other_mdev_sysfs_cb, &cb_data);
free(root);
return rc;
}
/*
* Determine if there are any conflicts between the specified device and
* the active apmask/aqmask settings. This is done by treating the masks
* as a temporary vfio_ap_device with all of the associated APQNs owned by
* the system.
*/
static int check_sysfs_mask_conflicts(struct ap_check_anchor *anc)
{
struct vfio_ap_device *sysdev = vfio_ap_device_new();
char *apmask = util_zalloc(AP_MASK_SIZE);
char *aqmask = util_zalloc(AP_MASK_SIZE);
int rc = 0;
if (ap_read_sysfs_masks(apmask, aqmask, AP_MASK_SIZE) != 0) {
fprintf(stderr, "Error reading system AP settings\n");
rc = -1;
goto out;
}
/* Convert the masks to a device with the associated APQNs */
ap_mask_to_list(apmask, sysdev->adapters);
ap_mask_to_list(aqmask, sysdev->domains);
/* Perform conflict analysis */
rc = find_apqn_conflicts(NULL, anc->dev->adapters,
anc->dev->domains, sysdev->adapters,
sysdev->domains, false);
out:
free(apmask);
free(aqmask);
vfio_ap_device_free(sysdev);
return rc;
}
/*
* Determine if there are any conflicts between the specified device and
* the apmask/aqmask settings stored in udev. This is done by treating
* the masks as a temporary vfio_ap_device with all of the associated
* AQPNs owned by the system.
*/
static int check_cfg_mask_conflicts(struct ap_check_anchor *anc)
{
struct vfio_ap_device *sysdev = vfio_ap_device_new();
char *apmask = util_zalloc(AP_MASK_SIZE);
char *aqmask = util_zalloc(AP_MASK_SIZE);
bool read_ap = false, read_aq = false;
char *path;
int rc = 0;
path = path_get_ap_udev();
if (!ap_read_udev_masks(path, apmask, aqmask, &read_ap, &read_aq)) {
fprintf(stderr, "Error reading system AP settings\n");
rc = -1;
goto out;
}
/* Convert the masks to a device with the associated APQNs */
ap_mask_to_list(apmask, sysdev->adapters);
ap_mask_to_list(aqmask, sysdev->domains);
/* Perform conflict analysis */
rc = find_apqn_conflicts(NULL, anc->dev->adapters,
anc->dev->domains, sysdev->adapters,
sysdev->domains, true);
out:
free(apmask);
free(aqmask);
free(path);
vfio_ap_device_free(sysdev);
return rc;
}
/* Subroutine to handle checking shared between DEFINE and MODIFY actions. */
static int ap_check_changes(struct ap_check_anchor *anc)
{
int rc = 0, rc2;
rc = ap_get_lock_callout();
if (rc) {
fprintf(stderr, "Failed to acquire configuration lock %d\n",
rc);
rc = -1;
goto out;
}
anc->cleanup_lock = true;
if (vfio_ap_read_device_config(NULL, anc->dev) != 0) {
fprintf(stderr, "Failed to read device config\n");
rc = -1;
goto out;
}
if (strcmp(anc->dev->type, anc->type) != 0) {
fprintf(stderr, "Invalid mdev_type: %s\n", anc->dev->type);
rc = -1;
goto out;
}
if (!anc->dev->manual) {
/* Check against all other AUTO config files */
rc = check_other_mdevs_cfg(anc);
/* Check against the system UDEV rule for apmask/aqmask */
rc2 = check_cfg_mask_conflicts(anc);
/* If either hit an error, reflect this */
rc = rc != 0 ? rc : rc2;
}
/* If successful, lock must remain held until post callout */
if (rc == 0)
anc->cleanup_lock = false;
out:
return rc;
}
/*
* Determine if defining the specified device is a valid operation.
* mdevctl can reach us for a DEFINE under the following circumstances:
* 1) the device does not exist
* 2) the device is active but does not have a config file, so this action
* would be to generate a config file based upon the active device.
* DEFINE has no effect on an active device (if one exists) it only creates
* the configuration file. The config file might be empty or may have various
* attributes if being fed by --jsonfile or an active device.
*/
static int ap_check_handle_define(struct ap_check_anchor *anc)
{
char *path = path_get_vfio_ap_mdev_config(anc->uuid);
if (util_path_is_readable(path)) {
fprintf(stderr, "Config already exists\n");
free(path);
return -1;
}
free(path);
return ap_check_changes(anc);
}
/*
* Determine if modifying the specified device is a valid operation.
* mdevctl can reach us for a MODIFY under the following circumstances:
* 1) Modifying a MANUAL device
* 2) Modifying an AUTO device
* In the case of MANUAL, we don't take any action because changes made via
* MODIFY don't take affect on the active mdev until a STOP/START cycle.
* In the case of AUTO, we must compare the contents of the proposed device
* with the contents of stashed AUTO mdev configurations + the system.
*/
static int ap_check_handle_modify(struct ap_check_anchor *anc)
{
char *path = path_get_vfio_ap_mdev_config(anc->uuid);
FILE *fd = fopen(path, "r");
/* Determine if a base config file already exists for UUID */
free(path);
if (fd == NULL) {
fprintf(stderr, "Config doesn't exist\n");
return -1;
}
fclose(fd);
return ap_check_changes(anc);
}
/*
* Determine if starting the specified device is a valid operation.
* mdevctl can reach us for a START under the following circumstances:
* 1) STARTing a MANUAL device
* 1a) Where the MANUAL device is defined (has a config file)
* 1b) Where the MANUAL device is NOT defined (no config file).
* For vfio-ap this case provides an mdev with no adapters/domains.
* 1c) Where the MANUAL device is NOT defined but a full configuration is
* provided via --jsonfile
* 2) STARTing an AUTO device
* 2a) Where the AUTO device is defined (has a config file)
* 2b) Where the AUTO device is NOT defined but a full configuration is
* provided via --jsonfile
* In each case, we must compare the proposed device with the contents of
* active mdevs + the system.
*/
static int ap_check_handle_start(struct ap_check_anchor *anc)
{
int rc = 0, rc2;
/* Can only start a device if vfio_ap is built-in or loaded */
if (!util_path_is_dir(VFIO_AP_PATH)) {
fprintf(stderr, "vfio_ap module is not loaded\n");
ap_check_exit(anc, EXIT_FAILURE);
}
rc = ap_get_lock_callout();
if (rc) {
fprintf(stderr, "Failed to acquire configuration lock %d\n",
rc);
rc = -1;
goto out;
}
anc->cleanup_lock = true;
if (vfio_ap_read_device_config(NULL, anc->dev) != 0) {
fprintf(stderr, "Failed to read device config\n");
rc = -1;
goto out;
}
if (strcmp(anc->dev->type, anc->type) != 0) {
fprintf(stderr, "Invalid mdev_type: %s\n", anc->dev->type);
rc = -1;
goto out;
}
/* Check against all other active vfio-ap devices */
rc = check_other_mdevs_sysfs(anc);
/* Check against the system sysfs values for apmask/aqmask */
rc2 = check_sysfs_mask_conflicts(anc);
/* If either hit an error, reflect this */
rc = rc != 0 ? rc : rc2;
/* If successful, lock must remain held until post callout */
if (rc == 0)
anc->cleanup_lock = false;
out:
return rc;
}
/*
* Acquire the appropriate serialization so that the specified device can be
* STOPped.
*/
static int ap_check_handle_stop(void)
{
int rc;
rc = ap_get_lock_callout();
if (rc) {
fprintf(stderr, "Failed to acquire configuration lock %d\n",
rc);
return -1;
}
/* The lock must remain held until post callout */
return 0;
}
/*
* Determine if UNDEFINEing the specified device is a valid operation.
* mdevctl can reach us for an UNDEFINE under the following circumstances:
* 1) UNDEFINEing an active device
* 2) UNDEFINEing an inactive device
* UNDEFINE has no effect on the active device, it only removes the config
* file.
*/
static int ap_check_handle_undefine(struct ap_check_anchor *anc)
{
char *path = path_get_vfio_ap_mdev_config(anc->uuid);
int rc = 0;
rc = ap_get_lock_callout();
if (rc) {
fprintf(stderr, "Failed to acquire configuration lock %d\n",
rc);
rc = -1;
goto out;
}
anc->cleanup_lock = true;
if (vfio_ap_read_device_config(path, anc->dev) != 0) {
fprintf(stderr, "Failed to read device config\n");
rc = -1;
goto out;
}
if (strcmp(anc->dev->type, anc->type) != 0) {
fprintf(stderr, "Invalid mdev_type: %s\n", anc->dev->type);
rc = -1;
goto out;
}
/* Success: lock must remain held until post callout */
anc->cleanup_lock = false;
out:
free(path);
return rc;
}
/*
* For callouts where the "pre" callout would have acquired the lock, it is
* now safe to remove the lock as all changes have been committed.
*/
static int ap_check_handle_post(void)
{
return ap_release_lock();
}
/* For the specified device, print the attributes to stdout in JSON format */
static int ap_check_handle_get_attributes(struct ap_check_anchor *anc)
{
struct vfio_ap_device *dev = anc->dev;
struct vfio_ap_node *node;
bool has_attr = false;
char buf[80];
char *path;
FILE *f;
path = path_get_vfio_ap_attr(anc->uuid, "matrix");
f = fopen(path, "r");
while (fgets(buf, sizeof(buf), f))
vfio_ap_parse_matrix(dev, buf);
vfio_ap_sort_matrix_results(dev);
fclose(f);
free(path);
path = path_get_vfio_ap_attr(anc->uuid, "control_domains");
f = fopen(path, "r");
while (fgets(buf, sizeof(buf), f))
vfio_ap_parse_control(dev, buf);
fclose(f);
free(path);
printf("[{");
if (!util_list_is_empty(dev->adapters)) {
util_list_iterate(dev->adapters, node) {
if (has_attr)
printf("},{");
printf("\"assign_adapter\": \"%u\"", node->id);
has_attr = true;
}
}
if (!util_list_is_empty(dev->domains)) {
util_list_iterate(dev->domains, node) {
if (has_attr)
printf("},{");
printf("\"assign_domain\": \"%u\"", node->id);
has_attr = true;
}
}
if (!util_list_is_empty(dev->controls)) {
util_list_iterate(dev->controls, node) {
if (has_attr)
printf("},{");
printf("\"assign_control_domain\": \"%u\"", node->id);
has_attr = true;
}
}
printf("}]\n");
return 0;
}
/*
* Determine which mdevctl action is being checked and handle accordingly.
*/
static int ap_check_handle_action(struct ap_check_anchor *anc)
{
int rc = 0;
switch (anc->event) {
case MDEVCTL_EVENT_PRE:
switch (anc->action) {
case MDEVCTL_ACTION_DEFINE:
rc = ap_check_handle_define(anc);
break;
case MDEVCTL_ACTION_MODIFY:
rc = ap_check_handle_modify(anc);
break;
case MDEVCTL_ACTION_START:
rc = ap_check_handle_start(anc);
break;
case MDEVCTL_ACTION_STOP:
rc = ap_check_handle_stop();
break;
case MDEVCTL_ACTION_UNDEFINE:
rc = ap_check_handle_undefine(anc);
break;
case MDEVCTL_ACTION_LIST:
case MDEVCTL_ACTION_TYPES:
default:
/* Ignore some actions including unknown ones */
break;
}
break;
case MDEVCTL_EVENT_POST:
switch (anc->action) {
case MDEVCTL_ACTION_DEFINE:
case MDEVCTL_ACTION_MODIFY:
case MDEVCTL_ACTION_START:
case MDEVCTL_ACTION_STOP:
case MDEVCTL_ACTION_UNDEFINE:
ap_check_handle_post();
break;
default:
/* Ignore other post events */
break;
}
break;
case MDEVCTL_EVENT_GET:
switch (anc->action) {
case MDEVCTL_ACTION_ATTRIBUTES:
rc = ap_check_handle_get_attributes(anc);
break;
default:
/* Ignore some actions including unknown ones */
break;
}
break;
default:
/* Ignore any unknown events */
break;
}
return rc;
}
/*
*
*/
int main(int argc, char *argv[])
{
struct ap_check_anchor anchor;
int rc;
ap_check_init(&anchor);
ap_check_parse(&anchor, argc, argv);
rc = ap_check_handle_action(&anchor);
ap_check_exit(&anchor, rc);
}

74
ap_tools/ap-check.h Normal file
View File

@@ -0,0 +1,74 @@
/*
* ap-check - Validate vfio-ap mediated device configuration changes
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#ifndef AP_CHECK_H
#define AP_CHECK_H
#include <stdbool.h>
/*
* List of all of the supported mdevctl actions
*/
enum mdevctl_action_id {
MDEVCTL_ACTION_DEFINE = 0,
MDEVCTL_ACTION_LIST,
MDEVCTL_ACTION_MODIFY,
MDEVCTL_ACTION_START,
MDEVCTL_ACTION_STOP,
MDEVCTL_ACTION_TYPES,
MDEVCTL_ACTION_UNDEFINE,
MDEVCTL_ACTION_ATTRIBUTES,
/* UNKNOWN must always be the last in the list */
MDEVCTL_ACTION_UNKNOWN,
};
#define NUM_MDEVCTL_ACTIONS MDEVCTL_ACTION_UNKNOWN
struct mdevctl_action {
enum mdevctl_action_id id;
const char action[32];
};
enum mdevctl_event_id {
MDEVCTL_EVENT_PRE = 0,
MDEVCTL_EVENT_POST,
MDEVCTL_EVENT_GET,
MDEVCTL_EVENT_UNKNOWN,
};
#define NUM_MDEVCTL_EVENTS MDEVCTL_EVENT_UNKNOWN
struct mdevctl_event {
enum mdevctl_event_id id;
const char event[32];
};
/* ap-check special exit codes */
#define APC_EXIT_UNKNOWN_TYPE 2
struct ap_check_anchor {
enum mdevctl_event_id event;
enum mdevctl_action_id action;
char *uuid;
char *parent;
char *type;
struct vfio_ap_device *dev;
/* Active Masks */
char apmask[80];
char aqmask[80];
/* Persistent Masks */
char p_apmask[80];
char p_aqmask[80];
bool cleanup_lock;
};
struct other_mdev_cb_data {
const char *uuid;
struct vfio_ap_device *dev;
};
#endif /* AP_CHECK_H */

View File

@@ -4492,6 +4492,15 @@ static void destroy_file_object(struct file *f)
free(f);
}
static void *cmsfs_oper_init(struct fuse_conn_info *conn, struct fuse_config *cfg)
{
(void)conn;
/* force immediate file removal */
cfg->hard_remove = 1;
return NULL;
}
static struct file_operations fops_fixed = {
.cache_data = cache_file_fixed,
.write_data = extend_block_fixed,
@@ -4507,6 +4516,7 @@ static struct file_operations fops_variable = {
};
static struct fuse_operations cmsfs_oper = {
.init = cmsfs_oper_init,
.getattr = cmsfs_getattr,
.statfs = cmsfs_statfs,
.readdir = cmsfs_readdir,
@@ -4683,8 +4693,6 @@ int main(int argc, char *argv[])
fuse_opt_add_arg(&args, "-oro");
/* force single threaded mode which requires no locking */
fuse_opt_add_arg(&args, "-s");
/* force immediate file removal */
fuse_opt_add_arg(&args, "-ohard_remove");
if (cmsfs.mode == BINARY_MODE &&
(cmsfs.codepage_from != NULL || cmsfs.codepage_to != NULL))

View File

@@ -5,7 +5,7 @@ COMMON_INCLUDED = true
# The variable "DISTRELEASE" should be overwritten in rpm spec files with:
# "make DISTRELEASE=%{release}" and "make install DISTRELEASE=%{release}"
VERSION = 2
RELEASE = 21
RELEASE = 22
PATCHLEVEL = 0
DISTRELEASE = build-$(shell date +%Y%m%d)
S390_TOOLS_RELEASE = $(VERSION).$(RELEASE).$(PATCHLEVEL)-$(DISTRELEASE)
@@ -118,7 +118,7 @@ endif
# Returns "yes" on success and nothing otherwise
#
define check_header_prereq
$(shell printf "#include <%s>\n int main(void) {return 0;}" $1 | \
$(shell printf "#include <%s>\n int main(void) {return 0;}\n" $1 | \
( $(CC) $(filter-out --coverage, $(ALL_CFLAGS)) $(ALL_CPPFLAGS) \
$2 -o /dev/null -xc - ) >/dev/null 2>&1 && echo -n yes)
endef
@@ -133,7 +133,7 @@ endef
# $5: Additional compiler & linker options (optional)
#
check_dep=\
printf "\#include <%s>\n int main(void) {return 0;}" $2 | ( $(CC) $(filter-out --coverage, $(ALL_CFLAGS)) $(ALL_CPPFLAGS) $5 -o /dev/null -xc - ) > /dev/null 2>&1; \
printf "\#include <%s>\n int main(void) {return 0;}\n" $2 | ( $(CC) $(filter-out --coverage, $(ALL_CFLAGS)) $(ALL_CPPFLAGS) $5 -o /dev/null -xc - ) > /dev/null 2>&1; \
if [ $$? != 0 ]; \
then \
printf " REQCHK %s (%s)\n" $1 $2; \
@@ -383,6 +383,14 @@ $(rootdir)/libkmipclient/libkmipclient.so: $(rootdir)/libkmipclient
$(MAKE) -C $(rootdir)/libkmipclient/ libkmipclient.so
.PHONY: $(rootdir)/libkmipclient
$(rootdir)/libap/libap.a: $(rootdir)/libap
$(MAKE) -C $(rootdir)/libap/ libap.a
.PHONY: $(rootdir)/libap
$(rootdir)/libpv/libpv.a: $(rootdir)/libpv
$(MAKE) -C $(rootdir)/libpv libpv.a
.PHONY: $(rootdir)/libpv
$(rootdir)/zipl/boot/data.o:
$(MAKE) -C $(rootdir)/zipl/boot/ data.o

View File

@@ -1,15 +1,16 @@
include ../common.mak
BIN_FILES = lscpumf chcpumf lshwc
MAN_FILES = lscpumf.1 chcpumf.8 lshwc.1
BIN_FILES = lscpumf chcpumf lshwc pai
MAN_FILES = lscpumf.8 chcpumf.8 lshwc.8 pai.8
all: $(BIN_FILES)
libs = $(rootdir)/libutil/libutil.a $(rootdir)/libcpumf/libcpumf.a
libs = $(rootdir)/libcpumf/libcpumf.a $(rootdir)/libutil/libutil.a
lscpumf: lscpumf.o $(libs)
chcpumf: chcpumf.o $(libs)
lshwc: lshwc.o $(libs)
pai: pai.o $(libs)
install: all install-man
$(INSTALL) -d -m 755 $(DESTDIR)$(BINDIR) $(DESTDIR)$(MANDIR)/man8
@@ -22,8 +23,8 @@ clean:
install-man:
for man in $(MAN_FILES); do \
msection=`echo $$man |sed 's/.*\.\([1-9]\)$$/man\1/'` ; \
$(INSTALL) -g $(GROUP) -o $(OWNER) -m 644 -D man/$$man $(DESTDIR)$(MANDIR)/$$msection/$$man ; \
$(INSTALL) -g $(GROUP) -o $(OWNER) -m 644 man/$$man \
$(DESTDIR)$(MANDIR)/man8 ; \
done
.PHONY: all install clean

View File

@@ -2555,6 +2555,586 @@ static struct counters cpumcf_z15_counters[] = {
},
};
static struct counters cpumcf_z16_counters[] = {
{
.ctrnum = 128,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "L1D_RO_EXCL_WRITES",
.desc = "A directory write to the Level-1 Data cache where"
"\n\t\tthe line was originally in a Read-Only state in the"
"\n\t\tcache but has been updated to be in the Exclusive"
"\n\t\tstate that allows stores to the cache line.",
},
{
.ctrnum = 129,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "DTLB2_WRITES",
.desc = "A translation has been written into The Translation"
"\n\t\tLookaside Buffer 2 (TLB2) and the request was made"
"\n\t\tby the Level-1 Data cache. This is a replacement"
"\n\t\tfor what was provided for the DTLB on z13 and prior"
"\n\t\tmachines.",
},
{
.ctrnum = 130,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "DTLB2_MISSES",
.desc = "A TLB2 miss is in progress for a request made by"
"\n\t\tthe Level-1 Data cache. Incremented by one for"
"\n\t\tevery TLB2 miss in progress for the Level-1 Data"
"\n\t\tcache on this cycle. This is a replacement for what"
"\n\t\twas provided for the DTLB on z13 and prior"
"\n\t\tmachines.",
},
{
.ctrnum = 131,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "CRSTE_1MB_WRITES",
.desc = "A translation entry was written into the Combined"
"\n\t\tRegion and Segment Table Entry array in the Level-2"
"\n\t\tTLB for a one-megabyte page.",
},
{
.ctrnum = 132,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "DTLB2_GPAGE_WRITES",
.desc = "A translation entry for a two-gigabyte page was"
"\n\t\twritten into the Level-2 TLB.",
},
{
.ctrnum = 134,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "ITLB2_WRITES",
.desc = "A translation entry has been written into the"
"\n\t\tTranslation Lookaside Buffer 2 (TLB2) and the"
"\n\t\trequest was made by the instruction cache. This is"
"\n\t\ta replacement for what was provided for the ITLB on"
"\n\t\tz13 and prior machines.",
},
{
.ctrnum = 135,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "ITLB2_MISSES",
.desc = "A TLB2 miss is in progress for a request made by"
"\n\t\tthe Level-1 Instruction cache. Incremented by one"
"\n\t\tfor every TLB2 miss in progress for the Level-1"
"\n\t\tInstruction cache in a cycle. This is a replacement"
"\n\t\tfor what was provided for the ITLB on z13 and prior"
"\n\t\tmachines.",
},
{
.ctrnum = 137,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "TLB2_PTE_WRITES",
.desc = "A translation entry was written into the Page Table"
"\n\t\tEntry array in the Level-2 TLB.",
},
{
.ctrnum = 138,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "TLB2_CRSTE_WRITES",
.desc = "Translation entries were written into the Combined"
"\n\t\tRegion and Segment Table Entry array and the Page"
"\n\t\tTable Entry array in the Level-2 TLB.",
},
{
.ctrnum = 139,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "TLB2_ENGINES_BUSY",
.desc = "The number of Level-2 TLB translation engines busy"
"\n\t\tin a cycle.",
},
{
.ctrnum = 140,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "TX_C_TEND",
.desc = "A TEND instruction has completed in a constrained"
"\n\t\ttransactional-execution mode.",
},
{
.ctrnum = 141,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "TX_NC_TEND",
.desc = "A TEND instruction has completed in a non-"
"\n\t\tconstrained transactional-execution mode.",
},
{
.ctrnum = 143,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "L1C_TLB2_MISSES",
.desc = "Increments by one for any cycle where a level-1"
"\n\t\tcache or level-2 TLB miss is in progress.",
},
{
.ctrnum = 145,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "DCW_REQ",
.desc = "A directory write to the Level-1 Data cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tfrom the requestors Level-2 cache.",
},
{
.ctrnum = 146,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "DCW_REQ_IV",
.desc = "A directory write to the Level-1 Data cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tfrom the requestors Level-2 cache with"
"\n\t\tintervention.",
},
{
.ctrnum = 147,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "DCW_REQ_CHIP_HIT",
.desc = "A directory write to the Level-1 Data cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tfrom the requestors Level-2 cache after using"
"\n\t\tchip level horizontal persistence, Chip-HP hit.",
},
{
.ctrnum = 148,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "DCW_REQ_DRAWER_HIT",
.desc = "A directory write to the Level-1 Data cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tfrom the requestors Level-2 cache after using"
"\n\t\tdrawer level horizontal persistence, Drawer-HP hit.",
},
{
.ctrnum = 149,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "DCW_ON_CHIP",
.desc = "A directory write to the Level-1 Data cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tfrom an On-Chip Level-2 cache.",
},
{
.ctrnum = 150,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "DCW_ON_CHIP_IV",
.desc = "A directory write to the Level-1 Data cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tfrom an On-Chip Level-2 cache with intervention.",
},
{
.ctrnum = 151,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "DCW_ON_CHIP_CHIP_HIT",
.desc = "A directory write to the Level-1 Data cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tfrom an On-Chip Level-2 cache after using chip"
"\n\t\tlevel horizontal persistence, Chip-HP hit.",
},
{
.ctrnum = 152,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "DCW_ON_CHIP_DRAWER_HIT",
.desc = "A directory write to the Level-1 Data cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tfrom an On-Chip Level-2 cache using drawer level"
"\n\t\thorizontal persistence, Drawer-HP hit.",
},
{
.ctrnum = 153,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "DCW_ON_MODULE",
.desc = "A directory write to the Level-1 Data cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tfrom an On-Module Level-2 cache.",
},
{
.ctrnum = 154,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "DCW_ON_DRAWER",
.desc = "A directory write to the Level-1 Data cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tfrom an On-Drawer Level-2 cache.",
},
{
.ctrnum = 155,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "DCW_OFF_DRAWER",
.desc = "A directory write to the Level-1 Data cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tfrom an Off-Drawer Level-2 cache.",
},
{
.ctrnum = 156,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "DCW_ON_CHIP_MEMORY",
.desc = "A directory write to the Level-1 Data cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tfrom On-Chip memory.",
},
{
.ctrnum = 157,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "DCW_ON_MODULE_MEMORY",
.desc = "A directory write to the Level-1 Data cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tfrom On-Module memory.",
},
{
.ctrnum = 158,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "DCW_ON_DRAWER_MEMORY",
.desc = "A directory write to the Level-1 Data cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tfrom On-Drawer memory.",
},
{
.ctrnum = 159,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "DCW_OFF_DRAWER_MEMORY",
.desc = "A directory write to the Level-1 Data cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tfrom Off-Drawer memory.",
},
{
.ctrnum = 160,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "IDCW_ON_MODULE_IV",
.desc = "A directory write to the Level-1 Data or Level-1"
"\n\t\tInstruction cache directory where the returned"
"\n\t\tcache line was sourced from an On-Module Level-2"
"\n\t\tcache with intervention.",
},
{
.ctrnum = 161,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "IDCW_ON_MODULE_CHIP_HIT",
.desc = "A directory write to the Level-1 Data or Level-1"
"\n\t\tInstruction cache directory where the returned"
"\n\t\tcache line was sourced from an On-Module Level-2"
"\n\t\tcache using chip horizontal persistence, Chip-HP"
"\n\t\thit.",
},
{
.ctrnum = 162,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "IDCW_ON_MODULE_DRAWER_HIT",
.desc = "A directory write to the Level-1 Data or Level-1"
"\n\t\tInstruction cache directory where the returned"
"\n\t\tcache line was sourced from an On-Module Level-2"
"\n\t\tcache using drawer level horizontal persistence,"
"\n\t\tDrawer-HP hit.",
},
{
.ctrnum = 163,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "IDCW_ON_DRAWER_IV",
.desc = "A directory write to the Level-1 Data or Level-1"
"\n\t\tInstruction cache directory where the returned"
"\n\t\tcache line was sourced from an On-Drawer Level-2"
"\n\t\tcache with intervention.",
},
{
.ctrnum = 164,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "IDCW_ON_DRAWER_CHIP_HIT",
.desc = "A directory write to the Level-1 Data or Level-1"
"\n\t\tinstruction cache directory where the returned"
"\n\t\tcache line was sourced from an On-Drawer Level-2"
"\n\t\tcache using chip level horizontal persistence, Chip-"
"\n\t\tHP hit.",
},
{
.ctrnum = 165,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "IDCW_ON_DRAWER_DRAWER_HIT",
.desc = "A directory write to the Level-1 Data or Level-1"
"\n\t\tinstruction cache directory where the returned"
"\n\t\tcache line was sourced from an On-Drawer Level-2"
"\n\t\tcache using drawer level horizontal persistence,"
"\n\t\tDrawer-HP hit.",
},
{
.ctrnum = 166,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "IDCW_OFF_DRAWER_IV",
.desc = "A directory write to the Level-1 Data or Level-1"
"\n\t\tinstruction cache directory where the returned"
"\n\t\tcache line was sourced from an Off-Drawer Level-2"
"\n\t\tcache with intervention.",
},
{
.ctrnum = 167,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "IDCW_OFF_DRAWER_CHIP_HIT",
.desc = "A directory write to the Level-1 Data or Level-1"
"\n\t\tinstruction cache directory where the returned"
"\n\t\tcache line was sourced from an Off-Drawer Level-2"
"\n\t\tcache using chip level horizontal persistence, Chip-"
"\n\t\tHP hit.",
},
{
.ctrnum = 168,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "IDCW_OFF_DRAWER_DRAWER_HIT",
.desc = "A directory write to the Level-1 Data or Level-1"
"\n\t\tInstruction cache directory where the returned"
"\n\t\tcache line was sourced from an Off-Drawer Level-2"
"\n\t\tcache using drawer level horizontal persistence,"
"\n\t\tDrawer-HP hit.",
},
{
.ctrnum = 169,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "ICW_REQ",
.desc = "A directory write to the Level-1 Instruction cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tthe requestors Level-2 cache.",
},
{
.ctrnum = 170,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "ICW_REQ_IV",
.desc = "A directory write to the Level-1 Instruction cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tfrom the requestors Level-2 cache with"
"\n\t\tintervention.",
},
{
.ctrnum = 171,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "ICW_REQ_CHIP_HIT",
.desc = "A directory write to the Level-1 Instruction cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tfrom the requestors Level-2 cache using chip level"
"\n\t\thorizontal persistence, Chip-HP hit.",
},
{
.ctrnum = 172,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "ICW_REQ_DRAWER_HIT",
.desc = "A directory write to the Level-1 Instruction cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tfrom the requestors Level-2 cache using drawer"
"\n\t\tlevel horizontal persistence, Drawer-HP hit.",
},
{
.ctrnum = 173,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "ICW_ON_CHIP",
.desc = "A directory write to the Level-1 Instruction cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tfrom an On-Chip Level-2 cache.",
},
{
.ctrnum = 174,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "ICW_ON_CHIP_IV",
.desc = "A directory write to the Level-1 Instruction cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tan On-Chip Level-2 cache with intervention.",
},
{
.ctrnum = 175,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "ICW_ON_CHIP_CHIP_HIT",
.desc = "A directory write to the Level-1 Instruction cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tfrom an On-Chip Level-2 cache using chip level"
"\n\t\thorizontal persistence, Chip-HP hit.",
},
{
.ctrnum = 176,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "ICW_ON_CHIP_DRAWER_HIT",
.desc = "A directory write to the Level-1 Instruction cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tfrom an On-Chip level 2 cache using drawer level"
"\n\t\thorizontal persistence, Drawer-HP hit.",
},
{
.ctrnum = 177,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "ICW_ON_MODULE",
.desc = "A directory write to the Level-1 Instruction cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tfrom an On-Module Level-2 cache.",
},
{
.ctrnum = 178,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "ICW_ON_DRAWER",
.desc = "A directory write to the Level-1 Instruction cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tan On-Drawer Level-2 cache.",
},
{
.ctrnum = 179,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "ICW_OFF_DRAWER",
.desc = "A directory write to the Level-1 Instruction cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tan Off-Drawer Level-2 cache.",
},
{
.ctrnum = 180,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "ICW_ON_CHIP_MEMORY",
.desc = "A directory write to the Level-1 Instruction cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tfrom On-Chip memory.",
},
{
.ctrnum = 181,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "ICW_ON_MODULE_MEMORY",
.desc = "A directory write to the Level-1 Instruction cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tfrom On-Module memory.",
},
{
.ctrnum = 182,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "ICW_ON_DRAWER_MEMORY",
.desc = "A directory write to the Level-1 Instruction cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tfrom On-Drawer memory.",
},
{
.ctrnum = 183,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "ICW_OFF_DRAWER_MEMORY",
.desc = "A directory write to the Level-1 Instruction cache"
"\n\t\tdirectory where the returned cache line was sourced"
"\n\t\tfrom Off-Drawer memory.",
},
{
.ctrnum = 224,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "BCD_DFP_EXECUTION_SLOTS",
.desc = "Count of floating point execution slots used for"
"\n\t\tfinished Binary Coded Decimal to Decimal Floating"
"\n\t\tPoint conversions. Instructions: CDZT, CXZT, CZDT,"
"\n\t\tCZXT.",
},
{
.ctrnum = 225,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "VX_BCD_EXECUTION_SLOTS",
.desc = "Count of floating point execution slots used for"
"\n\t\tfinished vector arithmetic Binary Coded Decimal"
"\n\t\tinstructions. Instructions: VAP, VSP, VMP, VMSP,"
"\n\t\tVDP, VSDP, VRP, VLIP, VSRP, VPSOP, VCP, VTP, VPKZ,"
"\n\t\tVUPKZ, VCVB, VCVBG, VCVD, VCVDG.",
},
{
.ctrnum = 226,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "DECIMAL_INSTRUCTIONS",
.desc = "Decimal instruction dispatched. Instructions: CVB,"
"\n\t\tCVD, AP, CP, DP, ED, EDMK, MP, SRP, SP, ZAP.",
},
{
.ctrnum = 232,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "LAST_HOST_TRANSLATIONS",
.desc = "Last Host Translation done",
},
{
.ctrnum = 244,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "TX_NC_TABORT",
.desc = "A transaction abort has occurred in a non-"
"\n\t\tconstrained transactional-execution mode.",
},
{
.ctrnum = 245,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "TX_C_TABORT_NO_SPECIAL",
.desc = "A transaction abort has occurred in a constrained"
"\n\t\ttransactional-execution mode and the CPU is not"
"\n\t\tusing any special logic to allow the transaction to"
"\n\t\tcomplete.",
},
{
.ctrnum = 246,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "TX_C_TABORT_SPECIAL",
.desc = "A transaction abort has occurred in a constrained"
"\n\t\ttransactional-execution mode and the CPU is using"
"\n\t\tspecial logic to allow the transaction to complete.",
},
{
.ctrnum = 248,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "DFLT_ACCESS",
.desc = "Cycles CPU spent obtaining access to Deflate unit",
},
{
.ctrnum = 253,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "DFLT_CYCLES",
.desc = "Cycles CPU is using Deflate unit",
},
{
.ctrnum = 256,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "SORTL",
.desc = "Increments by one for every SORT LISTS instruction"
"\n\t\texecuted.",
},
{
.ctrnum = 265,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "DFLT_CC",
.desc = "Increments by one for every DEFLATE CONVERSION CALL"
"\n\t\tinstruction executed.",
},
{
.ctrnum = 266,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "DFLT_CCFINISH",
.desc = "Increments by one for every DEFLATE CONVERSION CALL"
"\n\t\tinstruction executed that ended in Condition Codes"
"\n\t\t0, 1 or 2.",
},
{
.ctrnum = 267,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "NNPA_INVOCATIONS",
.desc = "Increments by one for every Neural Network"
"\n\t\tProcessing Assist instruction executed.",
},
{
.ctrnum = 268,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "NNPA_COMPLETIONS",
.desc = "Increments by one for every Neural Network"
"\n\t\tProcessing Assist instruction executed that ended"
"\n\t\tin Condition Codes 0, 1 or 2.",
},
{
.ctrnum = 269,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "NNPA_WAIT_LOCK",
.desc = "Cycles CPU spent obtaining access to IBM Z"
"\n\t\tIntegrated Accelerator for AI.",
},
{
.ctrnum = 270,
.ctrset = CPUMF_CTRSET_EXTENDED,
.name = "NNPA_HOLD_LOCK",
.desc = "Cycles CPU is using IBM Z Integrated Accelerator"
"\n\t\tfor AI.",
},
{
.ctrnum = 448,
.ctrset = CPUMF_CTRSET_MT_DIAG,
.name = "MT_DIAG_CYCLES_ONE_THR_ACTIVE",
.desc = "Cycle count with one thread active",
},
{
.ctrnum = 449,
.ctrset = CPUMF_CTRSET_MT_DIAG,
.name = "MT_DIAG_CYCLES_TWO_THR_ACTIVE",
.desc = "Cycle count with two threads active",
},
};
/* Return the type number of the CPU Measurement facility from the sysfs file.
* If the type number is equal to PERF_TYPE_RAW, then the prefix is 'r' to
* specify the raw counter number by the perf tool.
@@ -2890,6 +3470,9 @@ static struct counters *get_counter(int ctrset, size_t *len)
*len = ARRAY_SIZE(cpumcf_z15_counters);
read_ccerror(cp, *len);
break;
case UTIL_ARCH_MACHINE_TYPE_Z16:
cp = cpumcf_z16_counters;
*len = ARRAY_SIZE(cpumcf_z16_counters);
}
break;
}

View File

@@ -38,6 +38,7 @@
#include "lib/util_path.h"
#include "lib/util_scandir.h"
#include "lib/util_libc.h"
#include "lib/util_file.h"
#include "lib/libcpumf.h"
#include "lshwc.h"
@@ -62,25 +63,6 @@ static struct ctrname { /* List of defined counters */
unsigned long *ccv; /* Per CPU counter value */
} ctrname[MAXCTRS];
/* Open file and extract counter number */
static int read_counter(const char *p)
{
FILE *fp = fopen(p, "r");
int rc = 0, ctr;
if (fp) {
rc = fscanf(fp, "event=%x", &ctr);
fclose(fp);
}
return rc == 1 ? ctr : -EINVAL;
}
static int add_countername(char *name, int nr)
{
ctrname[nr].name = strdup(name);
return ctrname[nr].name ? 0 : -ENOMEM;
}
static bool read_counternames(void)
{
struct dirent **namelist = NULL;
@@ -96,13 +78,12 @@ static bool read_counternames(void)
}
for (i = 0; i < count && ctr >= 0; i++) {
util_asprintf(&ctrpath, "%s/%s", path, namelist[i]->d_name);
ctr = read_counter(ctrpath);
if (util_file_read_va(ctrpath, "event=%x", &ctr) == 1)
ctrname[ctr].name = util_strdup(namelist[i]->d_name);
else
warnx("Cannot parse %s", ctrpath);
free(ctrpath);
if (ctr >= 0)
ctr = add_countername(namelist[i]->d_name, ctr);
}
if (ctr < 0)
warnx("Cannot parse %s", path);
util_scandir_free(namelist, count);
free(path);
return ctr < 0 ? false : true;

View File

@@ -1,11 +1,11 @@
\" lscpumf.1
\" lscpumf.8
.\"
.\"
.\" Copyright IBM Corp. 2014, 2020
.\" s390-tools is free software; you can redistribute it and/or modify
.\" it under the terms of the MIT license. See LICENSE for details.
.\" ----------------------------------------------------------------------
.TH lscpumf "1" "May 2020" "s390-tools" "CPU-MF management programs"
.TH lscpumf "8" "May 2022" "s390-tools" "CPU-MF management programs"
.
.ds c \fBlscpumf\fP
.

View File

@@ -1,4 +1,4 @@
.\" lshwc.1
.\" lshwc.8
.\"
.\"
.\" Copyright IBM Corp. 2021
@@ -7,7 +7,7 @@
.\" ----------------------------------------------------------------------
.ds c \fBlshwc\fP
.
.TH \*c "1" "February 2021" "s390-tools" "CPU-MF management programs"
.TH \*c "8" "May 2022" "s390-tools" "CPU-MF management programs"
.
.SH NAME
\*c \- extract CPU Measurement Facilities counter sets
@@ -130,5 +130,5 @@ Date,Time,CPU,PROBLEM_STATE_CPU_CYCLES(32),PROBLEM_STATE_INSTRUCTIONS(33)
.ft
.fi
.SH "SEE ALSO"
.BR lscpumf (1)
.BR lscpumf (8)
.BR chcpumf (8)

268
cpumf/man/pai.8 Normal file
View File

@@ -0,0 +1,268 @@
.\" pai.8
.\"
.\"
.\" Copyright IBM Corp. 2022
.\" s390-tools is free software; you can redistribute it and/or modify
.\" it under the terms of the MIT license. See LICENSE for details.
.\" ----------------------------------------------------------------------
.ds c \fBpai\fP
.
.TH \*c "8" "May 2022" "s390-tools" "PAI Management Programs"
.
.SH NAME
\*c \- record and report Processor Activity Instrumentation (PAI) counters
.
.SH SYNOPSIS
\*c
.RB [ \-V ][ \-m | \-\-mapsize
.IR size ]
.RB [ \-i | \-\-interval
.IR ms ]
.BR \-c | \-\-crypto [ \fIcpulist ][: \fIdata\fR "] [" \fIloops\fP ]
.br
\*c
.RB [ \-V ][ \-m | \-\-mapsize
.IR size ]
.RB [ \-i | \-\-interval
.IR ms ]
.BR \-n | \-\-nnpa [ \fIcpulist ][: \fIdata\fR "] [" \fIloops\fP ]
.br
\*c
.RB [ \-V ][ \-H | \-\-humantime ][ \-S | \-\-summary "] " \-r | \-\-report " [" \fIfiles\fP ]
.br
\*c
.BR \-h | \-\-help
.br
\*c
.BR \-v | \-\-version
.
.
.SH DESCRIPTION
\*c
counters tally calls for specific CPU instructions.
The \*c command records PAI counters in a ring buffer.
\*c can record counter data for all CPUs or for selected CPUs.
The main command options are
.B \-c
for recording cryptographic CPU instructions,
.B \-n
for recording NNPA CPU instructions
and
.B \-r
for reporting.
If all three options are omitted, option
.B \-r
is assumed and a message is printed.
Recording stores data, by CPU, in files
.I paicrypto.<XXX>,
for option
.B \-c
or
.I painnpa.<XXX>,
for option
.B \-n
where <XXX> specifies the CPU number with leading
zeros.
The files are created in the working directory,
existing files are overwritten.
Reporting evaluates files that are created by recording.
.SH OPTIONS
.TP
.BR \-c ", " \-\-crypto "\fR[\fIcpulist\fR][:\fIdata\fR]"
Records data for all (default) or a specified list of CPUs.
The CPU list is a comma-separated list of CPU numbers and ranges.
In a range, a hyphen separates the first CPU number
from the last CPU number.
By default \*c lists all CPUs.
.RS
The optional data specification
follows the colon
and determines additional collection of data.
The specification consists of alphabetic
characters that can be upper or lower case:
.IP c|C
Include task rename system calls
.B exec
and
.BR prctl .
.IP f|F
Include task creation and deletion system calls
.B fork
and
.BR exit .
.IP s|S
Include context switch records created by the kernel scheduler.
.IP k|K
Include only PAI cryptographic counters changed during system call
execution.
.IP u|U
Include only PAI cryptographic counters changed during user space
execution.
.RE
.
.TP
.BR \-n ", " \-\-nnpa "\fR[\fIcpulist\fR][:\fIdata\fR]"
Records data for all (default) or a specified list of CPUs.
The CPU list is a comma-separated list of CPU numbers and ranges.
In a range, a hyphen separates the first CPU number
from the last CPU number.
By default \*c lists all CPUs.
.RS
The optional data specification
follows the colon
and determines additional collection of data.
The specification consists of alphabetic
characters that can be upper or lower case:
.IP c|C
Include task rename system calls
.B exec
and
.BR prctl .
.IP f|F
Include task creation and deletion system calls
.B fork
and
.BR exit .
.IP s|S
Include context switch records created by the kernel scheduler.
.RE
.
.TP
.BR \-r ", " \-\-report
Generates a report from the specified files.
Files is a list of blank-separated file names.
If no files are specified,
uses all files named
.I paicrypto.<X>,
in the working directory, where
.I <X>
is a CPU number that
identifies the CPU for which the data was recorded.
.
.TP
.BR \-h ", " \-\-help
Displays help information, then exits.
.
.TP
.BR \-v ", " \-\-version
Displays version information, then exits.
.
.TP
.BR \-V ", " \-\-verbose
Displays the following information during reporting:
the file name and the hexadecimal offset for each sample header.
Displays the following information during recording:
the read position in the ring buffer,
the file name, and the hexadecimal offset for each sample header.
.
.TP
.BR \-H ", " \-\-humantime
Changes the time stamp format for sample entries from a hexadecimal
number to a human readable <seconds>.<nanoseconds> format.
The time stamp specifies the elapsed time since Linux was booted.
.
.TP
.BR \-S ", " \-\-summary
Prints a summary of counter names and
counter numbers with non-zero values.
The summary is the last line in the command output.
It shows the sum of the counter values of all processed files.
.
.TP
.BR \-i ", " \-\-interval "\ ms"
Specifies the waiting time,
in milliseconds,
between ring buffer read operations during recording.
The default is 1000 milliseconds.
Argument
.B loops
is an integer that specifies the number of
read operations during recording. The default is 1.
.
.TP
.BR \-m ", " \-\-mapsize "\ size"
Specifies the size of the ring buffer
that holds the collected data.
The value specifies the number of 4 KB pages
to be allocated and must be a power of 2.
The default size is 512 pages.
The ring buffer is created with the
.IR mmap (2)
system call.
.
.SH ARGUMENT
The command line options determine how command line
arguments are interpreted.
For option
.B \-c
to start recording, the argument specifies
the number of read operations.
If omitted, the default is one.
For option
.B \-r
to start reporting, the argument specifies
the file names to be read.
If omitted all files in the current directory
with file name
.I paicryto.<X>
are read.
.SH "Concurrency with perf tool"
The following concurrency restrictions apply
for \*c and the
.B perf
tool.
Both tools use the
.IR perf_event_open ()
system call and use the same device driver.
The device driver supports limited concurrency:
.IP Counting:
Counting pai events can run in parallel.
.IP Sampling:
Only one sampling event pai_crypto/CRYPTO_ALL can be active at any one time.
If a sampling event is active, no pai counting event can be active.
Both tools stop with an error message if they detect a collision
reported by the device driver.
.SH "EXAMPLES"
This example collects cryptographic counter on CPU 0.
The program runs for 10 seconds
(10 intervals of 1000 milliseconds).
.sp 1
.nf
.ft CW
# \*c -c0 10
.ft R
.fi
.sp 1
This example displays the data
that is collected in the first example.
Each data line shows the time stamp in jiffies,
followed by the CPU number,
the event number,
process and thread identifier (pid/tid) separated by slash (/),
and a comma-separated list of counter number, colon (:),
and the counter value n hexadecimal notation.
.sp 1
.nf
.ft CW
# \*c -r
0x62a668f2fa 0 event 4096 sample pid 4956/4956 9:0xa7,73:0x8,74:0x18
0x6319c75653 0 event 4096 sample pid 4972/4972 32:0x1
0x6319e2ddee 0 event 4096 sample pid 4972/4972 32:0x1
0x631d3e44f5 0 event 4096 sample pid 4972/4972 32:0x2
0x631d4cfc2e 0 event 4096 sample pid 4972/4972 32:0x3
0x631d529fdd 0 event 4096 sample pid 4972/4972 32:0x2
0x631d84cfa0 0 event 4096 sample pid 4972/4972 73:0x1
0x636e9826bc 0 event 4096 sample pid 4984/4984 9:0xa7,73:0x8,74:0x18
0x636f81a137 0 event 4096 sample pid 4984/4984 9:0x2,74:0x7
0x6378026e54 0 event 4096 sample pid 4984/4984 73:0x2
0x637bcdc8da 0 event 4096 sample pid 4984/4984 73:0x2
0x637bd426cd 0 event 4096 sample pid 4984/4984 73:0x6
0x637c503384 0 event 4096 sample pid 4984/4984 73:0x2
0x64991d83ba 0 event 4096 sample pid 5026/5026 9:0xb0,73:0x8,74:0x18
#
.ft R
.fi
.sp 1
.SH "SEE ALSO"
perf(1)

1121
cpumf/pai.c Normal file

File diff suppressed because it is too large Load Diff

105
cpumf/pai.h Normal file
View File

@@ -0,0 +1,105 @@
/*
* pai - Extract CPU Processor Activity Instrumentation (PAI) facility data.
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#ifndef PAI_H
#define PAI_H
#define S390_EVT_PAI_CRYPTO 0x1000
enum { /* Event attribute specifications */
S390_EVTATTR_CTX_SWITCH = 1,
S390_EVTATTR_COMM = 2,
S390_EVTATTR_FORK = 4,
S390_EVTATTR_USERSPACE = 8,
S390_EVTATTR_KERNELSPACE = 16
};
struct pai_event {
struct util_list_node node; /* List node */
struct perf_event_attr attr; /* Perf_event_open(2) attr */
int fd; /* Perf event file descriptor */
void *map_addr; /* Sampling data mapping address */
size_t map_size; /* Sampling size mapping size */
char file_name[16]; /* File name of sampled data */
int file_fd; /* Map data output file descriptor */
int cpu; /* Perf_event_open(2) CPU */
unsigned long flags; /* Perf_event_open(2) flags */
};
struct pai_event_out { /* Output for CRYPTO_ALL event */
__u32 type; /* Header type, see PERF_RECORD_xxx */
__u16 misc; /* Header misc, value depends on type */
__u64 time; /* Time stamp valid for all entries */
__u32 cpu; /* CPU number valid for all entries */
union {
/* Fields from PERF_RECORD_FORK|EXIT */
struct {
__u32 pid, ppid;
__u32 tid, ptid;
__u64 time;
} s_fork;
/* Fields from PERF_RECORD_COMM */
struct {
__u32 pid, tid;
__u8 cmd[16];
} s_comm;
/* Fields from PERF_RECORD_SWITCH_CPU_WIDE */
struct {
__u32 next_prev_pid, next_prev_tid;
} s_cs;
/* Fields from PERF_RECORD_LOST_SAMPLES */
struct {
__u64 lost;
} s_lost;
/* Relevant fields from PERF_RECORD_SAMPLE, time and cpu
* are stored above
*/
struct {
__u64 ip;
__u64 period;
__u32 pid, tid;
} s_sample;
/* Fields from PERF_RECORD_[UN]THROTTLE */
struct {
__u64 time;
__u64 id;
__u64 stream_id;
} s_throttle;
} u;
/* Information on last context switch out */
struct cs_switch {
unsigned char valid;
__u32 topid, totid;
__u32 frompid, fromtid;
} cs_switch;
void *raw; /* Pointer to key/value array for crypto counters */
};
struct data_pos { /* Perf event mapped ring buffer */
__u64 data_head; /* Head in the data section */
__u64 data_tail; /* User-space written tail */
__u64 data_offset; /* Where the buffer starts */
__u64 data_size; /* Data buffer size */
};
struct event_name { /* Event list for number to name xlate */
char *name; /* Event name */
__u64 config; /* Event config value */
__u64 total; /* Total counter value */
};
struct pmu_events { /* Event list for PMU number to name xlate */
struct util_list_node node; /* List node */
char *name; /* PMU name */
int type; /* PMU type */
int lstlen; /* # of entries in lst */
unsigned long base; /* Base event number */
struct event_name *lst; /* List of event names */
};
#endif /* PAI_H */

View File

@@ -47,7 +47,7 @@ Print DASD uid without z/VM minidisk token
z/VM may provide an additional token that can be used to distinguish
between different minidisks (see --extended-uid option). To remain
compatibile with systems that were installed on older Linux or z/VM
compatible with systems that were installed on older Linux or z/VM
levels, the -u option will print the uid excluding any z/VM-provided
minidisk token.

View File

@@ -1,89 +0,0 @@
#!/bin/bash
### BEGIN INIT INFO
# Provides: cpacfstatsd
# Required-Start: $local_fs
# Required-Stop: $local_fs
# Should-Start:
# Should-Stop:
# Default-Start: 2 3 5
# Default-Stop: 0 1 6
# Short-Description: Start the cpacfstatsd daemon for Linux on System z
# Description: CPACF statistics collection daemon process for Linux on System z
### END INIT INFO
DAEMON=cpacfstatsd
DAEMON_PATH=/usr/sbin/cpacfstatsd
RUN_PID_FILE=/run/cpacfstatsd.pid
RETVAL=0
OPTIONS=""
# source function library
. /lib/lsb/init-functions
start()
{
if [ ! -f $RUN_PID_FILE ]; then
echo -n $"Starting $DAEMON:"
$DAEMON_PATH $OPTIONS
if [ $? == "0" ]; then
log_success_msg
else
log_failure_msg
fi
echo
else
echo "$DAEMON (pid $(cat $RUN_PID_FILE)) is already running..."
echo
fi
}
stop()
{
echo -n $"Stopping $DAEMON:"
if [ -f $RUN_PID_FILE ]; then
killproc $DAEMON_PATH -TERM
log_success_msg
rm -f $RUN_PID_FILE
else
log_failure_msg
fi
echo
}
restart() {
stop
sleep 1
start
}
status()
{
if [ ! -f $RUN_PID_FILE ]; then
echo "$DAEMON is not running."
echo
else
echo "$DAEMON (pid $(cat $RUN_PID_FILE)) is running."
echo
fi
}
# How are we called?
case "$1" in
start)
start
;;
stop)
stop
;;
status)
status
;;
restart)
restart
;;
*)
echo "Usage: $DAEMON {start|stop|status|restart}"
RETVAL=1
esac
exit $RETVAL

View File

@@ -1,116 +0,0 @@
#!/bin/bash
### BEGIN INIT INFO
# Provides: cpuplugd
# Required-Start: $local_fs $remote_fs
# Required-Stop: $local_fs $remote_fs
# Should-Start:
# Should-Stop:
# Default-Start: 2 3 5
# Default-Stop: 0 1 6
# Short-Description: Start the cpu hotplug daemon for Linux on System z
# Description: Starts the cpuplugd. It uses the configuration
# file /etc/cpuplugd.conf
### END INIT INFO
# chkconfig: 235 01 99
DAEMON=cpuplugd
DAEMON_PATH=/usr/sbin/cpuplugd
CONFIG_FILE=/etc/cpuplugd.conf
RUN_PID_FILE=/run/cpuplugd.pid
RETVAL=0
OPTIONS="-c $CONFIG_FILE"
# source function library
. /lib/lsb/init-functions
start()
{
if [ ! -f $RUN_PID_FILE ]; then
echo -n $"Starting $DAEMON:"
$DAEMON_PATH $OPTIONS
if [ $? == "0" ]; then
log_success_msg
else
log_failure_msg
fi
echo
else
echo "$DAEMON (pid $(cat $RUN_PID_FILE)) is already running..."
echo
fi
}
stop()
{
echo -n $"Stopping $DAEMON:"
if [ -f $RUN_PID_FILE ]; then
killproc $DAEMON_PATH -TERM
log_success_msg
rm -f $RUN_PID_FILE
else
log_failure_msg
fi
echo
}
restart() {
stop
#
# We have to wait 2-3 seconds here. When the daemon is stopped it takes
# the time we sleep to reactivate cpus. If we restart to fast and
# cpuplugd wasn't able to restore some settings we may get a undesired
# online cpu count after cpuplugd shutdown
#
sleep 4
start
}
status()
{
if [ ! -f $RUN_PID_FILE ]; then
echo "$DAEMON is not running."
echo
else
echo "$DAEMON (pid $(cat $RUN_PID_FILE), options: $OPTIONS) is running."
echo
fi
}
reload()
{
echo -n $"Reloading $DAEMON: "
if [ -f $RUN_PID_FILE ]; then
killproc $DAEMON_PATH -HUP
log_success_msg
else
log_failure_msg
fi
RETVAL=$?
echo
}
# How are we called?
case "$1" in
start)
start
;;
stop)
stop
;;
status)
status
;;
restart)
restart
;;
reload|force-reload)
reload
;;
*)
echo "Usage: $DAEMON {start|stop|status|restart|reload}"
RETVAL=1
esac
exit $RETVAL

View File

@@ -1,161 +0,0 @@
#!/bin/bash
### BEGIN INIT INFO
# Provides: mon_statd
# Required-Start: $local_fs $remote_fs
# Required-Stop: $local_fs $remote_fs
# Should-Start:
# Should-Stop:
# Default-Start: 2 3 5
# Default-Stop: 0 1 6
# Short-Description: Configure the mon_fsstatd and mon_procd daemons.
# Description: Configures the mon_fsstatd and mon_procd daemons. It uses the
# configuration file /etc/sysconfig/mon_statd.
### END INIT INFO
# chkconfig: 235 01 99
DAEMON=mon_statd
FSSTATD=mon_fsstatd
PROCD=mon_procd
FSSTATD_PATH=/usr/sbin/$FSSTATD
PROCD_PATH=/usr/sbin/$PROCD
CONFIG_FILE=/etc/sysconfig/$DAEMON
FSSTATD_PID_FILE=/run/$FSSTATD.pid
PROCD_PID_FILE=/run/$PROCD.pid
# source function library
. /lib/lsb/init-functions
# Source config file
if [ -f $CONFIG_FILE ]; then
. $CONFIG_FILE
fi
UDEVSETTLE=/usr/bin/udevadm
if [ ! -e $UDEVSETTLE ]
then
UDEVSETTLE=/sbin/udevsettle
UDEVSETTLE_CALL="$UDEVSETTLE --timeout=10"
else
UDEVSETTLE_CALL="$UDEVSETTLE settle --timeout=10"
fi
load_kernel_module()
{
if [ ! -e /dev/monwriter ]; then
echo "Loading monwriter module..."
modprobe monwriter 2>&1
if [ $? -ne 0 ]; then
exit 1
fi
if [ -e $UDEVSETTLE ]
then
$UDEVSETTLE_CALL
fi
fi
}
start_daemon()
{
local daemon_name=$1
local daemon_interval=$2
local daemon_pid_file=$3
local daemon_path=$4
if [ ! -f $daemon_pid_file ]; then
load_kernel_module
echo -n "Starting $daemon_name:"
$daemon_path -i $daemon_interval
if [ $? -eq 0 ]; then
log_success_msg
else
log_failure_msg
fi
else
echo "$daemon_name (pid $(cat $daemon_pid_file)) is already running..."
fi
}
start()
{
if [ "$FSSTAT" = "yes" ]; then
start_daemon $FSSTATD $FSSTAT_INTERVAL $FSSTATD_PID_FILE \
$FSSTATD_PATH
fi
if [ "$PROC" = "yes" ]; then
start_daemon $PROCD $PROC_INTERVAL $PROCD_PID_FILE \
$PROCD_PATH
fi
}
stop_daemon()
{
local daemon_name=$1
local daemon_pid_file=$2
local daemon_path=$3
echo -n "Stopping $daemon_name:"
if [ -f $daemon_pid_file ]; then
killproc $daemon_path -TERM
log_success_msg
rm -f $daemon_pid_file
else
log_failure_msg
fi
}
stop()
{
if [ "$FSSTAT" = "yes" ]; then
stop_daemon $FSSTATD $FSSTATD_PID_FILE $FSSTATD_PATH
fi
if [ "$PROC" = "yes" ]; then
stop_daemon $PROCD $PROCD_PID_FILE $PROCD_PATH
fi
}
restart() {
stop
start
}
status_daemon()
{
local daemon_name=$1
local daemon_pid_file=$2
local daemon_interval=$3
if [ ! -f $daemon_pid_file ]; then
echo "$daemon_name is not running."
else
echo "$daemon_name (pid $(cat $daemon_pid_file), interval: $daemon_interval) is running."
fi
}
status()
{
status_daemon $FSSTATD $FSSTATD_PID_FILE $FSSTAT_INTERVAL
status_daemon $PROCD $PROCD_PID_FILE $PROC_INTERVAL
}
# How are we called?
case "$1" in
start)
start
;;
stop)
stop
;;
status)
status
;;
restart|reload|force-reload)
restart
;;
*)
echo "Usage: $DAEMON {start|stop|status|restart|reload}"
exit 1
esac
exit 0

View File

@@ -15,7 +15,8 @@ ALL_CFLAGS := $(NO_PIE_CFLAGS) -Os -g \
-fno-delete-null-pointer-checks -fno-stack-protector \
-fexec-charset=IBM1047 -m64 -mpacked-stack \
-mstack-size=4096 -mstack-guard=128 -msoft-float \
-Wall -Wformat-security -Wextra -Werror
-Wall -Wformat-security -Wextra -Werror \
-Wno-array-bounds
FILES := stage3a.bin stage3b.bin stage3b_reloc.bin
@@ -74,11 +75,11 @@ stage3b_reloc.elf:
%.elf: %.o
case $* in \
stage3a) SFLAGS="$(NO_PIE_LDFLAGS) -nostdlib -Wl,-T,stage3a.lds";; \
stage3b) SFLAGS="$(NO_PIE_LDFLAGS) -nostdlib -Wl,-T,stage3b.lds";; \
stage3b_reloc) SFLAGS="$(NO_PIE_LDFLAGS) -nostdlib -Wl,-estage3b_reloc_start,-Ttext,0";; \
stage3a) SFLAGS="-Wl,-T,stage3a.lds";; \
stage3b) SFLAGS="-Wl,-T,stage3b.lds";; \
stage3b_reloc) SFLAGS="-Wl,-estage3b_reloc_start,-Ttext,0";; \
esac; \
$(LINK) $$SFLAGS -m64 $(filter %.o, $^) -o $@
$(LINK) $$SFLAGS $(NO_PIE_LDFLAGS) -m64 -static -nostdlib $(filter %.o, $^) -o $@
@chmod a-x $@
%.bin: %.elf

View File

@@ -4,8 +4,8 @@ include ../../common.mak
all:
install:
$(INSTALL) -d -m 755 $(DESTDIR)$(MANDIR)/man8
$(INSTALL) -m 644 -c genprotimg.8 $(DESTDIR)$(MANDIR)/man8
$(INSTALL) -d -m 755 $(DESTDIR)$(MANDIR)/man1
$(INSTALL) -m 644 -c genprotimg.1 $(DESTDIR)$(MANDIR)/man1
clean:

View File

@@ -2,7 +2,7 @@
.\" s390-tools is free software; you can redistribute it and/or modify
.\" it under the terms of the MIT license. See LICENSE for details.
.\"
.TH GENPROTIMG 8 "November 2020" "s390-tools"
.TH GENPROTIMG 1 "May 2022" "s390-tools"
.SH NAME
genprotimg \- Create a protected virtualization image

View File

@@ -1,3 +1,10 @@
/*
* Copyright IBM Corp. 2020, 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#ifndef COMMON_H
#define COMMON_H

View File

@@ -17,7 +17,7 @@
#include "helper.h"
#include "table.h"
#define SD_DG_INIT_INTERVAL_MS 200
#define SD_DG_INIT_INTERVAL_SEC 1
#define SD_SYS_ID_SIZE 9
/*

View File

@@ -150,7 +150,7 @@ void sd_update(void)
*/
void sd_dg_register(struct sd_dg *dg, int has_core_data)
{
struct timespec ts = {0, SD_DG_INIT_INTERVAL_MS * 1000000};
struct timespec ts = {SD_DG_INIT_INTERVAL_SEC, 0};
struct sd_sys_item *sys_item;
struct sd_cpu_item *cpu_item;
unsigned int i;

185
include/boot/boot_defs.h Normal file
View File

@@ -0,0 +1,185 @@
/*
* Boot and dump related definitions
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#ifndef BOOT_DEFS_H
#define BOOT_DEFS_H
#include <stdint.h>
#include "lib/zt_common.h"
#define ZIPL_MAGIC "zIPL"
#define ZIPL_MAGIC_SIZE 4
/*
* ECKD dump parameter
*/
struct eckd_dump_param {
uint32_t blk_start;
uint32_t blk_end;
uint16_t blk_size;
uint8_t num_heads;
uint8_t bpt;
char reserved[4];
} __packed;
/*
* FBA dump parameter
*/
struct fba_dump_param {
uint32_t res1;
uint32_t blk_start;
uint32_t res2;
uint32_t blk_end;
} __packed;
/*
* SCSI dump parameter
*/
struct scsi_dump_param {
uint64_t block;
uint64_t reserved;
} __packed;
/*
* Layout of block pointer for linear devices
* e.g. SCSI
*/
struct linear_blockptr {
uint64_t blockno;
uint16_t size;
uint16_t blockct;
uint8_t reserved[4];
} __packed;
/*
* Layout of block pointer for cylinder/head/sector devices
* e.g. ECKD
*/
struct eckd_blockptr {
uint16_t cyl;
uint16_t head;
uint8_t sec;
uint16_t size;
uint8_t blockct;
uint8_t reserved[8];
} __packed;
typedef enum {
COMPONENT_EXECUTE = 0x01,
COMPONENT_LOAD = 0x02,
COMPONENT_SIGNATURE = 0x03
} component_type;
typedef enum {
COMPONENT_HEADER_IPL = 0x00,
COMPONENT_HEADER_DUMP = 0x01
} component_header_type;
struct component_header {
uint8_t magic[4];
uint8_t type;
uint8_t reserved[27];
} __packed;
struct signature_header {
uint8_t format;
uint8_t reserved[3];
uint32_t length;
} __packed;
typedef union {
uint64_t load_address;
uint64_t load_psw;
struct signature_header sig_head;
} component_data;
struct component_entry {
uint8_t data[23];
uint8_t type;
component_data compdat;
} __packed;
/* SCSI dump super block */
struct scsi_dump_sb {
uint64_t magic;
uint64_t version;
uint64_t part_start;
uint64_t part_size;
uint64_t dump_offset;
uint64_t dump_size;
uint64_t csum_offset;
uint64_t csum_size;
uint64_t csum;
} __packed;
#define SCSI_DUMP_SB_MAGIC 0x5a46435044554d50ULL /* ZFCPDUMP */
/* To avoid a csum entry of 0 a seed is used */
#define SCSI_DUMP_SB_SEED 0x12345678
#define SCSI_DUMP_SB_CSUM_SIZE 4096
/* Boot info */
#define BOOT_INFO_VERSION 1
#define BOOT_INFO_MAGIC "zIPL"
#define BOOT_INFO_DEV_TYPE_ECKD 0x00
#define BOOT_INFO_DEV_TYPE_FBA 0x01
#define BOOT_INFO_DEV_TYPE_SCSI 0x02
#define BOOT_INFO_BP_TYPE_IPL 0x00
#define BOOT_INFO_BP_TYPE_DUMP 0x01
#ifdef __s390x__
#define BOOT_INFO_FLAGS_ARCH 0x01
#else
#define BOOT_INFO_FLAGS_ARCH 0x00
#endif
struct boot_info_bp_dump {
union {
struct eckd_dump_param eckd;
struct fba_dump_param fba;
struct scsi_dump_param scsi;
} param;
uint8_t unused[16];
} __packed;
struct boot_info_bp_ipl {
union {
struct eckd_blockptr eckd;
struct linear_blockptr lin;
} bm_ptr;
uint8_t unused[16];
} __packed;
struct boot_info {
char magic[4];
uint8_t version;
uint8_t bp_type;
uint8_t dev_type;
uint8_t flags;
union {
struct boot_info_bp_dump dump;
struct boot_info_bp_ipl ipl;
} bp;
} __packed;
#define DISK_LAYOUT_ID 0x00000001
struct scsi_mbr {
uint8_t magic[4];
uint32_t version_id;
uint8_t reserved[8];
struct linear_blockptr program_table_pointer;
uint8_t reserved2[0x50];
struct boot_info boot_info;
} __packed;
#endif /* BOOT_DEFS_H */

View File

@@ -16,6 +16,8 @@
#define __LC_IPLDEV 0x0c6c
#define __LC_OS_INFO 0x0e18
#define LOWCORE_SIZE _AC(0x2000, UL)
#define PAGE_SIZE _AC(4096, UL)
/* Minimum size of a stack frame in bytes */
@@ -56,6 +58,26 @@ struct psw32_t {
void load_wait_psw(uint64_t, struct psw_t *);
struct subchannel_id {
uint32_t cssid:8;
uint32_t:4;
uint32_t m:1;
uint32_t ssid:2;
uint32_t one:1;
uint32_t sch_no:16;
} __packed __aligned(4);
struct tpi_info {
struct subchannel_id schid;
uint32_t intparm;
uint32_t adapter_IO:1;
uint32_t directed_irq:1;
uint32_t isc:3;
uint32_t:12;
uint32_t type:3;
uint32_t:12;
} __packed __aligned(4);
struct _lowcore {
uint8_t pad_0x0000[0x0014-0x0000]; /* 0x0000 */
uint32_t ipl_parmblock_ptr; /* 0x0014 */
@@ -78,10 +100,15 @@ struct _lowcore {
uint8_t pad_0x00a4[0x00a8-0x00a4]; /* 0x00a4 */
uint64_t trans_exc_code; /* 0x00a8 */
uint64_t monitor_code; /* 0x00b0 */
uint16_t subchannel_id; /* 0x00b8 */
uint16_t subchannel_nr; /* 0x00ba */
uint32_t io_int_parm; /* 0x00bc */
uint32_t io_int_word; /* 0x00c0 */
union {
struct {
uint16_t subchannel_id; /* 0x00b8 */
uint16_t subchannel_nr; /* 0x00ba */
uint32_t io_int_parm; /* 0x00bc */
uint32_t io_int_word; /* 0x00c0 */
};
struct tpi_info tpi_info; /* 0x00b8 */
};
uint8_t pad_0x00c4[0x00c8-0x00c4]; /* 0x00c4 */
uint32_t stfl_fac_list; /* 0x00c8 */
uint8_t pad_0x00cc[0x00e8-0x00cc]; /* 0x00cc */
@@ -210,6 +237,7 @@ struct _lowcore {
/* align to the top of the prefix area */
uint8_t pad_0x1900[0x2000-0x1900]; /* 0x1900 */
} __packed __aligned(8192);
STATIC_ASSERT(sizeof(struct _lowcore) == LOWCORE_SIZE)
#define S390_lowcore (*((struct _lowcore *) 0))

94
include/lib/ap.h Normal file
View File

@@ -0,0 +1,94 @@
/*
* libap - A collection of tools for ap/vfio-ap management
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#ifndef LIB_AP_H
#define LIB_AP_H
#include <stdbool.h>
#include "lib/util_list.h"
#define VFIO_AP_PATH "/sys/devices/vfio_ap"
#define VFIO_AP_PARENT_PATH "devices/vfio_ap/matrix"
#define VFIO_AP_CONFIG_PATH "/etc/mdevctl.d/matrix"
#define VFIO_AP_TYPE "vfio_ap-passthrough"
#define AP_UDEV_FILE "/etc/udev/rules.d/41-ap.rules"
#define AP_LOCKFILE "/run/lock/s390apconfig.lock"
#define AP_LOCK_RETRIES 15
/* apmask and aqmask are each represented as 67 character strings with:
* '0x' leading characters
* 64 hex digits (to represent 256 bits)
* terminating character
*/
#define AP_MASK_SIZE 67
#define AP_MAX_MASK_VALUE 255
/* List structure used for keeping track of lists of adapter/domain IDs */
struct vfio_ap_node {
struct util_list_node node; /* prev/next list info */
unsigned int id; /* list entry (adapter, domain, etc) */
};
/*
* Structure used to represent a vfio_ap-passthrough device configuration.
* The list of adapters and domains can be used to derive the APQNs for the
* device. The type value is used to verify that the device (when read from
* a mdevctl config file) is of type vfio_ap-passthrough. The manual value
* represents whether the device is started on-demand after boot (true) or
* automatically during boot (false).
*/
struct vfio_ap_device {
char *uuid; /* Unique ID for this mdev */
struct util_list *adapters; /* List of adapters for device */
struct util_list *domains; /* List of usage domains for device */
struct util_list *controls; /* List of control domains for device */
char *type; /* mdev type string */
bool manual; /* manual/auto start setting for mdev */
};
/* General Utility Functions */
void print_ap_device(struct vfio_ap_device *dev);
bool is_valid_uuid(const char *uuid);
int ap_test_bit(int n, const char *hexbytestr);
void ap_set_bit(int n, char *hexbytestr, bool val);
/* Path-related functions */
char *path_get_vfio_ap_mdev(const char *uuid);
char *path_get_vfio_ap_mdev_config(const char *uuid);
char *path_get_vfio_ap_attr(const char *uuid, const char *attr);
char *path_get_ap_udev(void);
/* Functions for manipulating sysfs to read active device info */
void vfio_ap_parse_matrix(struct vfio_ap_device *dev, char *matrix);
void vfio_ap_sort_matrix_results(struct vfio_ap_device *dev);
void vfio_ap_parse_control(struct vfio_ap_device *dev, char *control);
/* Functions for reading JSON device config */
int vfio_ap_read_device_config(const char *path, struct vfio_ap_device *dev);
/* Functions for managing vfio_ap device structures */
struct vfio_ap_device *vfio_ap_device_new(void);
void vfio_ap_device_clear(struct vfio_ap_device *dev);
void vfio_ap_device_free(struct vfio_ap_device *dev);
/* Functions for acquiring current vfio_ap device info */
int ap_read_sysfs_masks(char *ap, char *aq, int size);
bool ap_read_udev_masks(char *path, char *ap, char *aq, bool *read_ap,
bool *read_aq);
void ap_mask_to_list(char *mask, struct util_list *list);
void ap_list_remove_all(struct util_list *list);
/* Lock Functions */
int ap_get_lock(void);
int ap_get_lock_callout(void);
int ap_release_lock(void);
#endif /* LIB_AP_H */

View File

@@ -16,30 +16,47 @@
#define S390_CPUS_POSSIBLE "/sys/devices/system/cpu/possible"
#define S390_CPUS_ONLINE "/sys/devices/system/cpu/online"
#define S390_CPUMSF_BUFFERSZ "/sys/module/kernel/parameters/cpum_sfb_size"
#define S390_SYSFS_PAI_CRYPTO "/sys/devices/pai_crypto/"
#define S390_SYSFS_PAI_NNPA "/sys/devices/pai_nnpa/"
/**
* Read out the PMU type from a given file.
*
* Return the PMU type number assigned to this PMU by the kernel. This is
* a non zero number.
* If the PMU does not exist return -1 and set errno.
*
* @param[in] dirname Name of the event directory in sysfs
* @retval >=0 Number of PMU assigned by the kernel
* @retval -1 PMU unknown to kernel
*/
int libcpumf_pmutype(const char *dirname);
/**
* Read out the PMU name from a given type.
*
* Return the PMU name this PMU was registered in kernel. If the PMU was
* registered without a name, it is not listed in the directory.
* The caller must free the memory returned by name.
*
* @param[in] wanted_type Type number of the PMU
* @param[out] name Name of the PMU (when retval is 0)
* @retval 0 PMU wanted_type detected and PMU name valid
* @retval -1 No PMU with wanted_type
*/
int libcpumf_pmuname(unsigned int wanted_type, char **name);
/**
* Read out the CPU list from a given file name, for example from files
* /sys/devices/system/cpu/online or /sys/devices/system/cpu/possible.
*
* Return the cpu_set_t created from parsing the CPU list in the second
* parameter.
* Return code of zero indicates proper conversion and -1 indicates an
* error.
*
* @param[in] buffer Comma separated string of a CPU list
* @param[in] filename Name of a sysfs CPU list file name
* @param[out] mask Converted buffer into cpu_set_t mask structure
* @retval 0 Successful conversion of cpulist
* @retval -1 Unsuccessful conversion of cpulist
*/
int libcpumf_cpuset(const char *buffer, cpu_set_t *mask);
int libcpumf_cpuset_fn(const char *filename, cpu_set_t *mask);
@@ -55,11 +72,16 @@ int libcpumf_cpuset_fn(const char *filename, cpu_set_t *mask);
* @param[out] cfvn Contains CPUMF counter first version number
* @param[out] csvn Contains CPUMF counter second version number
* @param[out] auth Contains CPUMF counter set authorization level
* @retval true Information returned in parameters is valid
* @retval false Information could not be retrieved
*/
bool libcpumf_cpumcf_info(int *cfvn, int *csvn, int *auth);
/**
* Return true if CPU Measurement Counter Facility is available.
*
* @retval true CPU Measurement Counter Facility is available
* @retval false CPU Measurement Counter Facility is not available
*/
bool libcpumf_have_cpumcf(void);
@@ -77,18 +99,28 @@ bool libcpumf_have_cpumcf(void);
* microsecond
* @param[out] basic_sz Basic sample size in bytes
* @param[out] diag_sz Diagnostic sample size in bytes
* @retval true Information returned in parameters is valid
* @retval false Information could not be retrieved
*/
bool libcpumf_cpumsf_info(unsigned long *min, unsigned long *max,
unsigned long *speed, int *basic_sz, int *diag_sz);
/**
* Return true if CPU Measurement Sampling Facility is available.
*
* @retval true CPU Measurement Sampling Facility is available
* @retval false CPU Measurement Sampling Facility is not available
*/
bool libcpumf_have_cpumsf(void);
/**
* Return true if CPU Measurement Sampling Facility buffer sizes are
* available.
*
* @retval true CPU Measurement Sampling Facility buffer sizes are
* available
* @retval false CPU Measurement Sampling Facility buffer sizes are
* not available
*/
bool libcpumf_have_sfb(void);
@@ -101,6 +133,25 @@ bool libcpumf_have_sfb(void);
*
* @param[out] min Minimum supported sampling buffer size
* @param[out] max Maximum supported sampling buffer size
* @retval true Information returned in parameters is valid
* @retval false Information could not be retrieved
*/
bool libcpumf_sfb_info(unsigned long *min, unsigned long *max);
/**
* Return true if PAI_CRYPTO counter Facility is supported.
*
* @retval true PAI_CRYPTO counter Facility is available
* @retval false PAI_CRYPTO counter Facility is not available
*/
bool libcpumf_have_pai_crypto(void);
/**
* Return true if PAI_NNPA counter Facility is supported. This PMU facility
* supports the Neural Network Processing Assist (NNPA) counter set.
*
* @retval true PAI_NNPA counter Facility is available
* @retval false PAI_NNPA counter Facility is not available
*/
bool libcpumf_have_pai_nnpa(void);
#endif

View File

@@ -26,6 +26,7 @@ enum util_arch_machine_type {
UTIL_ARCH_MACHINE_TYPE_Z14_ZR1 = 3907,
UTIL_ARCH_MACHINE_TYPE_Z15 = 8561,
UTIL_ARCH_MACHINE_TYPE_Z15_T02 = 8562,
UTIL_ARCH_MACHINE_TYPE_Z16 = 3931,
};
int util_arch_machine_type(void);

View File

@@ -0,0 +1,21 @@
/**
* @defgroup util_exit_code_h util_exit_code: General purpose exit codes
* @{
* @brief General purpose exit codes
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#ifndef LIB_UTIL_EXIT_CODE_H
#define LIB_UTIL_EXIT_CODE_H
typedef enum {
UTIL_EXIT_OK = 0, /* Program finished successfully */
UTIL_EXIT_RUNTIME_ERROR = 15, /* A run-time error occurred */
UTIL_EXIT_OUT_OF_MEMORY = 22, /* Not enough available memory */
} util_exit_code_t;
#endif /** LIB_UTIL_EXIT_CODE_H @} */

View File

@@ -11,6 +11,8 @@
#ifndef LIB_UTIL_FILE_H
#define LIB_UTIL_FILE_H
#include "lib/util_exit_code.h"
int util_file_read_line(char *str, size_t size, const char *fmt, ...);
int util_file_read_i(int *val, int base, const char *fmt, ...);
int util_file_read_l(long *val, int base, const char *fmt, ...);
@@ -26,4 +28,10 @@ int util_file_write_ul(unsigned long val, int base, const char *fmt, ...);
int util_file_write_ull(unsigned long long val, int base, const char *fmt, ...);
int util_file_read_va(const char *path, const char *fmt, ...);
util_exit_code_t util_file_read_fd_buf(FILE *fd, void **buffer_ptr,
size_t *size_ptr);
char *util_file_read_fd(FILE *fd, int chomp);
char *util_file_read_text_file(const char *path, int chomp);
#endif /** LIB_UTIL_FILE_H @} */

44
include/lib/util_udev.h Normal file
View File

@@ -0,0 +1,44 @@
/*
* @defgroup util_udev_h util_udev: UDEV interface
* @{
* @brief Work with UDEV files
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#ifndef LIB_UTIL_UDEV_H
#define LIB_UTIL_UDEV_H
#include <stdbool.h>
#include "lib/util_exit_code.h"
#include "lib/util_list.h"
/* Single key-operator-value entry in a udev rule line.*/
struct util_udev_entry_node {
struct util_list_node node;
char *key;
char *op;
char *value;
};
/* Single udev line in a udev rule file. */
struct util_udev_line_node {
struct util_list_node node;
struct util_list entries;
char *line;
};
/* Udev rule file. */
struct util_udev_file {
struct util_list lines;
};
util_exit_code_t util_udev_read_file(const char *path,
struct util_udev_file **file_ptr);
void util_udev_free_file(struct util_udev_file *file);
void util_udev_file_print(struct util_udev_file *file);
#endif /** LIB_UTIL_UDEV_H @} */

438
include/libpv/cert.h Normal file
View File

@@ -0,0 +1,438 @@
/*
* Certificate functions and definitions.
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#ifndef LIBPV_CERT_H
#define LIBPV_CERT_H
#include <openssl/x509v3.h>
#include <openssl/err.h>
#include "libpv/common.h"
#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_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"
#define PV_IMB_Z_SUBJECT_ENTRY_COUNT 6
/* Minimum security level for the keys/certificates used to establish a chain of
* trust (see https://www.openssl.org/docs/man1.1.1/man3/X509_VERIFY_PARAM_set_auth_level.html
* for details).
*/
#define PV_CERTS_SECURITY_LEVEL 2
/** pv_cert_init:
*
* Should not be called by user.
* Use pv_init() instead which
* calls this function during creation.
*
* Sets up data structures for caching CRLs.
*/
void pv_cert_init(void);
/** pv_cert_cleanup:
*
* Should not be called by user.
* Use pv_cleanup() instead which
* calls this function during creation.
*
* Cleans up data structures for caching CRLs.
*/
void pv_cert_cleanup(void);
#define PV_CERT_ERROR g_quark_from_static_string("pv-cert-error-quark")
typedef enum {
PV_CERT_ERROR_CERT_REVOKED,
PV_CERT_ERROR_CERT_SIGNATURE_INVALID,
PV_CERT_ERROR_CERT_SUBJECT_ISSUER_MISMATCH,
PV_CERT_ERROR_CRL_DOWNLOAD_FAILED,
PV_CERT_ERROR_CRL_SIGNATURE_INVALID,
PV_CERT_ERROR_CRL_SUBJECT_ISSUER_MISMATCH,
PV_CERT_ERROR_FAILED_DOWNLOAD_CRL,
PV_CERT_ERROR_INTERNAL,
PV_CERT_ERROR_INVALID_PARM,
PV_CERT_ERROR_INVALID_SIGNATURE_ALGORITHM,
PV_CERT_ERROR_INVALID_VALIDITY_PERIOD,
PV_CERT_ERROR_LOAD_CRL,
PV_CERT_ERROR_LOAD_DEFAULT_CA,
PV_CERT_ERROR_LOAD_ROOT_CA,
PV_CERT_ERROR_MALFORMED_CERTIFICATE,
PV_CERT_ERROR_MALFORMED_ROOT_CA,
PV_CERT_ERROR_NO_CRL,
PV_CERT_ERROR_NO_CRLDP,
PV_CERT_ERROR_NO_ISSUER_IBM_Z_FOUND,
PV_CERT_ERROR_NO_PUBLIC_KEY,
PV_CERT_ERROR_READ_CERTIFICATE,
PV_CERT_ERROR_READ_CRL,
PV_CERT_ERROR_SIGNATURE_ALGORITHM_MISMATCH,
PV_CERT_ERROR_SKID_AKID_MISMATCH,
PV_CERT_ERROR_VERIFICATION_FAILED,
PV_CERT_ERROR_WRONG_CA_USED,
} PvCertErrors;
/** PvX509WithPath - X509 certificate associated with a path
*/
typedef struct {
X509 *cert;
char *path;
} PvX509WithPath;
/** pv_x509_with_path_new:
*
* @cert: X509 certificate
* @path: Path of that X509 certificate
*
* Returns: (nullable) (transfer full): new X509 with path
*/
PvX509WithPath *pv_x509_with_path_new(X509 *cert, const char *path);
/** pv_x509_with_path_free:
*
* Frees the path and the PvX509WithPath; Decreases the refcount of the X509
*/
void pv_x509_with_path_free(PvX509WithPath *cert);
typedef STACK_OF(DIST_POINT) STACK_OF_DIST_POINT;
typedef STACK_OF(X509) STACK_OF_X509;
typedef STACK_OF(X509_CRL) STACK_OF_X509_CRL;
typedef GSList PvCertWithPathList;
typedef struct {
X509 *cert;
STACK_OF_X509_CRL *crls;
} PvX509Pair;
/** pv_x509_pair_new_take:
* @cert: ptr to X509
* @crls: ptr to CRLs
*
* Takes a X509 and the associated CRLs and builds a pair.
* Both, *cert and *crls will be NULL afterwards, and owned by the pair.
*
* Returns: (nullable) (transfer full): New PvX509Pair
*/
PvX509Pair *pv_x509_pair_new_take(X509 **cert, STACK_OF_X509_CRL **crls);
/** pv_x509_pair_free:
*
* Decreases the refcount of the X509 and crls.
* Frees the PvX509Pair.
*/
void pv_x509_pair_free(PvX509Pair *pair);
void STACK_OF_DIST_POINT_free(STACK_OF_DIST_POINT *stack);
void STACK_OF_X509_free(STACK_OF_X509 *stack);
void STACK_OF_X509_CRL_free(STACK_OF_X509_CRL *stack);
/** pv_x509_from_pem_der_data:
*
* @data: GBytes containing the cert in PEM format
* @error: return location for a #GError
*
* Returns: (nullable) (transfer full): X509 cert
*/
X509 *pv_x509_from_pem_der_data(GBytes *data, GError **error);
/** pv_x509_get_ec_pubkey:
*
* @cert: X509 to extract elliptic curve pubkey from
* @nid: numerical identifier of the expected curve
* @error: return location for a #GError
*
* Returns: (nullable) (transfer full): corresponding pupkey for the given certificate
*/
EVP_PKEY *pv_x509_get_ec_pubkey(X509 *cert, int nid, GError **error);
/** pv_get_ec_pubkeys:
*
* @certs_with_path: List of PvX509WithPath
* @nid: numerical identifier of the expected curve
* @error: return location for a #GError
*
* Returns: (nullable) (transfer full): List of corresponding public keys for the given certificate
*/
GSList *pv_get_ec_pubkeys(PvCertWithPathList *certs_with_path, int nid, GError **error);
/* pv_load_certificates:
*
* @cert_paths: list of cert paths.
* @error: return location for a #GError
*
* @cert_paths must contain at least one element, otherwise an error is
* reported.
*
* Returns: (nullable) (transfer full): List of PvX509WithPath corresponding to the given paths
*/
PvCertWithPathList *pv_load_certificates(char **cert_paths, GError **error);
/* pv_load_first_cert_from_file:
*
* @path: location of the x509
* @error: return location for a #GError
*
* This function reads in only the first certificate and ignores all other. This
* is only relevant for the PEM file format. For the host-key document and the
* root CA this behavior is expected.
*
* Returns: (nullable) (transfer full): PvX509WithPath corresponding to the given path
*/
X509 *pv_load_first_cert_from_file(const char *path, GError **error);
/* pv_load_first_crl_from_file:
*
* @path: location of the x509 CRL
* @error: return location for a #GError
*
* This function reads in only the first CRL and ignores all other. This
* is only relevant for the PEM file format.
*
* Returns: (nullable) (transfer full): X509_CRL corresponding to the given path
*/
X509_CRL *pv_load_first_crl_from_file(const char *path, GError **error);
/** pv_store_setup_crl_download:
*
* @st: X509_STORE
*/
void pv_store_setup_crl_download(X509_STORE *st);
/** pv_load_first_crl_by_cert:
* @cert: X509 to specify the download location.
* @error: return location for a #GError
*
* This function returns the first X509_CRL found from the CRL distribution
* points specified in @cert.
*
* Returns: (nullable) (transfer full): x509 CRL corresponding to the given X509
*/
X509_CRL *pv_load_first_crl_by_cert(X509 *cert, GError **error);
/** pv_try_load_crls_by_certs:
*
* @certs_with_path: List of PvX509WithPath
*
* Returns: (nullable) (transfer full): Stack of CRLs corresponding to the given X509
*/
STACK_OF_X509_CRL *pv_try_load_crls_by_certs(PvCertWithPathList *certs_with_path);
/** pv_store_setup:
*
* @root_ca_path: Location of the rootCA or NULL if SystemRoot CA shall be used
* @crl_paths: List of CRL paths or NULL
* @cert_with_crl_paths: List of (untrusted) X509 paths
* @error: return location for a #GError
*
* The untrusted certs need to be verified before actually verifying a Host Key Document.
*
* Returns: (nullable) (transfer full): X509_store with given input data.
*
*/
X509_STORE *pv_store_setup(char *root_ca_path, char **crl_paths, char **cert_with_crl_paths,
GError **error);
/** pv_get_x509_stack:
*
* x509_with_path_list: list of PvX509WithPath
*
* Returns: (nullable) (transfer full): Stack of X509 corresponding to the given x509 with path
*/
STACK_OF_X509 *pv_get_x509_stack(const GSList *x509_with_path_list);
/** pv_init_store_ctx:
*
* @ctx: a uninitialized Store CTX
* @trusted: X509_STORE with a trusted rootCA
* @chain: untrusted X509s
* @error: return location for a #GError
*
* Can be called multiple times on the same context if X509_STORE_CTX_cleanup(ctx)
* was called before.
*
* Returns:
* 0 on success
* -1 in failure
*/
int pv_init_store_ctx(X509_STORE_CTX *ctx, X509_STORE *trusted, STACK_OF_X509 *chain,
GError **error) PV_NONNULL(1, 2, 3);
/** pv_init_store_ctx:
*
* @trusted: X509_STORE with a trusted rootCA
* @chain: untrusted X509s
* @error: return location for a #GError
*
* Returns: (nullable) (transfer full): X509_STORE_CTX setup with the input data
*/
X509_STORE_CTX *pv_create_store_ctx(X509_STORE *trusted, STACK_OF_X509 *chain, GError **error)
PV_NONNULL(1, 2);
/** pv_remove_ibm_signing_certs:
*
* @certs: Stack of X509s
*
* Returns: (transfer full):
* List of all IBM Z signing key certificates in @certs and remove them
* from the chain.
* Empty stack if no IBM Z signing key is found.
*/
STACK_OF_X509 *pv_remove_ibm_signing_certs(STACK_OF_X509 *certs);
/** pv_c2b_name:
*
* Workaround to fix the mismatch between issuer name of the
* IBM Z signing CRLs and the IBM Z signing key subject name.
*
* 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).
*
* This function tries to reorder the name attributes such that
* further OpenSSL calls can work with it. The caller is
* responsible to free the returned value.
*/
X509_NAME *pv_c2b_name(const X509_NAME *name);
/** pv_verify_host_key:
*
* @host_key: X509 to be verified
* @issuer_pairs: IBM signing key X509+CRLs Pairs used for verification
* @level: Security level. see PV_CERTS_SECURITY_LEVEL
* @error: return location for a #GError
*
* Returns:
* 0 if Host key could be verified with one of the IBM signing keys
* -1 if no IBM signing key could verify the authenticity of the given host key
*
*/
int pv_verify_host_key(X509 *host_key, GSList *issuer_pairs, int verify_flags, int level,
GError **error);
/** pv_verify_cert:
*
* @ctx: trusted store ctx used for verification
* @cert: X509 to be verified
* @error: return location for a #GError
*
* Cannot be used to verify host keys with IBM signing keys, as IBM signing
* keys are no intermediate CAs. Use pv_verify_host_key() instead.
*
* Returns:
* 0 if @cert could be verified
* -1 if @cert could not be verified
*/
int pv_verify_cert(X509_STORE_CTX *ctx, X509 *cert, GError **error) PV_NONNULL(1, 2);
/** pv_check_crl_valid_for_cert:
*
* @crl: CRL to be verified
* @cert: Cert that probably issued the given CRL
* @verify_flags: X509 Verification flags (X509_V_FLAG_<TYPE>)
* @error: return location for a #GError
*
* Verify whether a revocation list @crl is valid and is issued by @cert. For
* this multiple steps must be done:
*
* 1. verify issuer of the CRL matches with the suject name of @cert
* 2. verify the validity period of the CRL
* 3. verify the signature of the CRL
*
* Important: This function does not verify whether @cert is allowed to issue a
* CRL.
*
* Returns:
* 0 if @crl is valid and issued by @cert
* -1 otherwise
*/
int pv_verify_crl(X509_CRL *crl, X509 *cert, int verify_flags, GError **error);
/** pv_check_chain_parameters:
*
* @chain: chain of trust to be validated
* @error: return location for a #GError
*
* Verifies that chain has at least a RootCA ans intermediate CA
* and logs the used ROD CA subject
*
* Returns:
* 0 @chain is valid
* -1 otherwise
*/
int pv_check_chain_parameters(const STACK_OF_X509 *chain, GError **error);
/** pv_store_set_verify_param:
*
* @store: X509_STORE to set parameters
* @error: return location for a #GError
*
* Returns:
* 0 on success
* -1 on failure
*/
int pv_store_set_verify_param(X509_STORE *store, GError **error);
/** pv_store_ctx_find_valid_crls:
*
* @ctx: STORE_CTX for searching CRLs
* @cert: X509 to match CRLs aggainst
* @error: return location for a #GError
*
* Returns: (nullable) (transfer full): STACK of CRLs related to given @crl fin @ctx
*/
STACK_OF_X509_CRL *pv_store_ctx_find_valid_crls(X509_STORE_CTX *ctx, X509 *cert, GError **error)
PV_NONNULL(1, 2);
/** pv_verify_host_key_doc:
*
* @host_key_certs_with_path: X509s to be verified
* @trusted. X509_STORE with a rusted RootCA
* @untrusted_certs: STACK OF untrusted X509s
* @online: true if CRLs shall be downloaded
* @error: return location for a #GError
*
* Returns:
* 0 if all given HKDs could be verified using the chain of trust.
* -1 otherwise
*/
int pv_verify_host_key_doc(PvCertWithPathList *host_key_certs_with_path, X509_STORE *trusted,
STACK_OF_X509 *untrusted_certs, gboolean online, GError **error)
PV_NONNULL(1, 2, 3);
/** pv_verify_host_key_docs_by_path:
*
* @host_key_paths: locations of X509 to be verified
* @optional_root_ca_path: rootCA location or NULL if Default shall be used
* @optional_crl_paths: locations of CRLs or NULL
* @untrusted_cert_paths: locations of IntermediateCAs including the IBM signing key
* @online: true if CRLs shall be downloaded
* @error: return location for a #GError
*
* Returns:
* 0 if all given HKDs could be verfied using the chain of trust.
* -1 otherwise
*/
int pv_verify_host_key_docs_by_path(char **host_key_paths, char *optional_root_ca_path,
char **optional_crl_paths, char **untrusted_cert_paths,
gboolean online, GError **error) PV_NONNULL(1, 4);
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(AUTHORITY_KEYID, AUTHORITY_KEYID_free)
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(PvX509WithPath, pv_x509_with_path_free)
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(STACK_OF_DIST_POINT, STACK_OF_DIST_POINT_free)
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(STACK_OF_X509, STACK_OF_X509_free)
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(STACK_OF_X509_CRL, STACK_OF_X509_CRL_free)
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(X509, X509_free)
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(X509_CRL, X509_CRL_free)
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(X509_LOOKUP, X509_LOOKUP_free)
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(X509_NAME, X509_NAME_free)
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(X509_VERIFY_PARAM, X509_VERIFY_PARAM_free)
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(PvX509Pair, pv_x509_pair_free)
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(X509_STORE, X509_STORE_free)
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(X509_STORE_CTX, X509_STORE_CTX_free)
#endif /* LIBPV_CERT_H */

35
include/libpv/common.h Normal file
View File

@@ -0,0 +1,35 @@
/*
* Libpv common definitions.
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*
*/
#ifndef LIBPV_COMMON_H
#define LIBPV_COMMON_H
/* must be included before any (other) glib header to verify that
* the glib version is supported
*/
#include "libpv/glib-helper.h"
#include <glib/gi18n-lib.h>
#include "libpv/openssl-compat.h"
#include "libpv/macros.h"
/** pv_init:
*
* Must be called before any libpv call.
*/
int pv_init(void);
/** pv_cleanup:
*
* Must be called when done with using libpv.
*/
void pv_cleanup(void);
#endif /* LIBPV_COMMON_H */

200
include/libpv/crypto.h Normal file
View File

@@ -0,0 +1,200 @@
/*
* General cryptography helper functions and definitions
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#ifndef LIBPV_CRYPTO_H
#define LIBPV_CRYPTO_H
#include <openssl/bio.h>
#include <openssl/bn.h>
#include <openssl/ec.h>
#include <openssl/evp.h>
#include "libpv/common.h"
typedef struct pv_cipher_parms {
const EVP_CIPHER *cipher;
size_t tag_size;
GBytes *key;
union {
GBytes *iv;
GBytes *tweak;
};
} PvCipherParms;
typedef union {
struct {
uint8_t x[80];
uint8_t y[80];
};
uint8_t data[160];
} PvEcdhPubKey;
G_STATIC_ASSERT(sizeof(PvEcdhPubKey) == 160);
typedef GSList PvEvpKeyList;
enum PvCryptoMode {
PV_ENCRYPT,
PV_DECRYPT,
};
/** pv_get_openssl_error:
*
* Returns: (transfer full): String representing the error.
*/
const char *pv_get_openssl_error(void);
/**
* pv_BIO_reset:
* @b: BIO to reset
*
* Resets a BIO to its initial state.
*
* Returns: 0 in case of success, -1 otherwise.
*/
int pv_BIO_reset(BIO *b);
/**
* pv_generate_rand_data:
* @size: number of generated random bytes using a crypographically secure pseudo random generator
* @error: return location for a #GError
*
* Creates a new #GBytes with @size random bytes using a cryptographically
* secure pseudo random generator.
*
* Returns: (nullable) (transfer full): a new #GBytes, or %NULL in case of an error
*/
GBytes *pv_generate_rand_data(size_t size, GError **error);
/**
* pv_generate_key:
* @cipher: specifies the OpenSSL cipher for which a cryptographically secure key should be generated
* @error: return location for a #GError
*
* Creates a random key for @cipher using a cryptographically secure pseudo
* random generator.
*
* Returns: (nullable) (transfer full): a new #GBytes, or %NULL in case of an error
*/
GBytes *pv_generate_key(const EVP_CIPHER *cipher, GError **error) PV_NONNULL(1);
/**
* pv_generate_iv:
* @cipher: specifies the OpenSSL cipher for which a cryptographically secure IV should be generated
* @error: return location for a #GError
*
* Creates a random IV for @cipher using a cryptographically secure pseudo
* random generator.
*
* Returns: (nullable) (transfer full): a new #GBytes, or %NULL in case of an error
*/
GBytes *pv_generate_iv(const EVP_CIPHER *cipher, GError **error) PV_NONNULL(1);
/* Symmetric en/decryption functions */
/**
* pv_gcm_encrypt:
* @plain: data to encrypt
* @aad: (optional): additional data that should be authenticated with the key
* @parms:
* @cipher: (out): location to store the ciphertext
* @tag: (out): location to store the generated GCM tag
* @error: return location for a #GError
*
* Encrypts the @plain data and authenticates @aad data.
*
* Returns: number of bytes, or -1 in case of an error
*/
int64_t pv_gcm_encrypt(GBytes *plain, GBytes *aad, const PvCipherParms *parms, GBytes **cipher,
GBytes **tag, GError **error) PV_NONNULL(1, 3, 4, 5);
/**
* pv_gcm_decrypt:
* @cipher: ciphertext to decrypt
* @aad: (optional): additional date to authenticate
* @tag: the GCM tag
* @parms:
* @plain: (out): location to store the decrypted data
* @error: return location for a #GError
*
* Decrypts the @cipher data and authenticates the @aad data.
*
* Returns: number of bytes, or -1 in case of an error
*/
int64_t pv_gcm_decrypt(GBytes *cipher, GBytes *aad, GBytes *tag, const PvCipherParms *parms,
GBytes **plain, GError **error) PV_NONNULL(1, 3, 4, 5);
/** pv_hkdf_extract_and_expand:
* @derived_key_len: size of the output key
* @key: input key
* @salt: salt for the extraction
* @info: infor for the expansion
* @md: EVP mode of operation
* @error: return location for a #GError
*
* Performs a RFC 5869 HKDF.
*
* Returns: (nullable) (transfer full): Result of RFC 5869 HKDF
*
*/
GBytes *pv_hkdf_extract_and_expand(size_t derived_key_len, GBytes *key, GBytes *salt, GBytes *info,
const EVP_MD *md, GError **error) PV_NONNULL(2, 3, 4, 5);
/** pv_generate_ec_key:
*
* @nid: Numerical identifier of the curve
* @error: return location for a #GError
*
* Returns: (nullable) (transfer full): new random key based on the given curve
*/
EVP_PKEY *pv_generate_ec_key(int nid, GError **error);
/** pv_evp_pkey_to_ecdh_pub_key:
*
* @key: input key in EVP_PKEY format
* @error: return location for a #GError
*
* Returns: the public part of the input @key in ECDH format.
*/
PvEcdhPubKey *pv_evp_pkey_to_ecdh_pub_key(EVP_PKEY *key, GError **error) PV_NONNULL(1);
/** pv_derive_exchange_key:
* @cust: Customer Key
* @host: Host key
* @error: return location for a #GError
*
* Returns: (nullable) (transfer full): Shared Secret of @cust and @host
*/
GBytes *pv_derive_exchange_key(EVP_PKEY *cust, EVP_PKEY *host, GError **error) PV_NONNULL(1, 2);
GQuark pv_crypto_error_quark(void);
#define PV_CRYPTO_ERROR pv_crypto_error_quark()
typedef enum {
PV_CRYPTO_ERROR_DERIVE,
PV_CRYPTO_ERROR_HKDF_FAIL,
PV_CRYPTO_ERROR_INTERNAL,
PV_CRYPTO_ERROR_INVALID_KEY_SIZE,
PV_CRYPTO_ERROR_KEYGENERATION,
PV_CRYPTO_ERROR_RANDOMIZATION,
PV_CRYPTO_ERROR_READ_FILE,
PV_CRYPTO_ERROR_NO_IBM_Z_SIGNING_KEY,
PV_CRYPTO_ERROR_NO_MATCH_TAG,
} PvCryptoErrors;
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(BIO, BIO_free_all)
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(BIGNUM, BN_free)
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(BN_CTX, BN_CTX_free)
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(EC_GROUP, EC_GROUP_free)
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(EC_KEY, EC_KEY_free)
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(EC_POINT, EC_POINT_free)
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(EVP_CIPHER_CTX, EVP_CIPHER_CTX_free)
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(EVP_PKEY, EVP_PKEY_free)
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(EVP_PKEY_CTX, EVP_PKEY_CTX_free)
#endif /* LIBPV_CRYPTO_H */

53
include/libpv/curl.h Normal file
View File

@@ -0,0 +1,53 @@
/*
* Libcurl utils
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#ifndef LIBPV_CURL_H
#define LIBPV_CURL_H
#include <curl/curl.h>
#include "libpv/common.h"
#define CRL_DOWNLOAD_TIMEOUT_MS 3000
#define CRL_DOWNLOAD_MAX_SIZE 0x100000
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(CURL, curl_easy_cleanup)
/** curl_download:
* @url: URL to specify location of data
* @timeout_ms: time to wait until fail
* @max_size: Maximum size of the downloaded data
* @error: return location for a GError
*
* Returns: (nullable) (transfer full): Downloaded data as #GByteArray
*/
GByteArray *curl_download(const char *url, long timeout_ms, uint max_size, GError **err);
/** pv_curl_init:
*
* Should not be called by user.
* Use pv_init() instead which
* calls this function during creation.
*/
int pv_curl_init(void);
/** pv_curl_cleanup:
*
* Should not be called by user.
* Use pv_cleanup() instead which
* calls this function during creation.
*/
void pv_curl_cleanup(void);
#define PV_CURL_ERROR g_quark_from_static_string("pv-curl-error-quark")
typedef enum {
PV_CURL_ERROR_CURL_INIT_FAILED,
PV_CURL_ERROR_DOWNLOAD_FAILED,
} PvCurlErrors;
#endif /* LIBPV_CURL_H */

117
include/libpv/glib-helper.h Normal file
View File

@@ -0,0 +1,117 @@
/*
* Glib convenience functions
* Shall be used instead of manually including glib.
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#ifndef LIBPV_GLIB_HELPER_H
#define LIBPV_GLIB_HELPER_H
#if defined(GLIB_VERSION_MIN_REQUIRED)
#if GLIB_VERSION_MIN_REQUIRED < GLIB_VERSION_2_56
#error "GLIB_VERSION must be at least 2.56"
#endif
#else
#define GLIB_VERSION_MIN_REQUIRED GLIB_VERSION_2_56
#endif
#include <glib.h>
#include <gmodule.h>
#include <stdio.h>
#include "libpv/macros.h"
#ifdef __clang__
#define WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(...) \
DO_PRAGMA(clang diagnostic push) \
DO_PRAGMA(clang diagnostic ignored "-Wunused-function") \
G_DEFINE_AUTOPTR_CLEANUP_FUNC(__VA_ARGS__) \
DO_PRAGMA(clang diagnostic pop)
#else
#define WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(...) G_DEFINE_AUTOPTR_CLEANUP_FUNC(__VA_ARGS__)
#endif
#define pv_wrapped_g_assert(__expr) g_assert(__expr)
/** pv_sec_gbytes_new_take:
*
* #g_bytes_new_take() with secure cleanup
*/
GBytes *pv_sec_gbytes_new_take(void *data, size_t size);
/** pv_sec_gbytes_new:
*
* #g_bytes_new() with secure cleanup
*/
GBytes *pv_sec_gbytes_new(const void *data, size_t size);
/** pv_file_get_content_as_secure_bytes:
*
* @filename: path to file for reading in
*
* read file and save as secure gbytes
*
* Return:
* Content of file as #GBytes with secure cleanup
*/
GBytes *pv_file_get_content_as_secure_bytes(const char *filename);
/** pv_file_get_content_as_g_bytes:
*
* @filename: path to file for reading in
*
* read file and save as gbytes
*
* Return:
* Content of file as #GBytes
*/
GBytes *pv_file_get_content_as_g_bytes(const char *filename, GError **error);
/** pv_file_seek:
*
* fseek with error reporting
*/
int pv_file_seek(FILE *file, long offset, int whence, GError **error);
/** pv_file_write:
*
* fwrite with error reporting
*/
size_t pv_file_write(FILE *file, const void *ptr, size_t size, GError **error);
/** pv_file_close:
*
* fclose with error reporting
*/
long pv_file_close(FILE *file, GError **error);
/** pv_file_tell:
*
* ftell with error reporting
*/
long pv_file_tell(FILE *file, GError **error);
/** pv_file_open:
*
* fopen with error reporting
*/
FILE *pv_file_open(const char *filename, const char *mode, GError **error);
/** pv_gbytes_memcpy:
*
* memcpy with size check.
*/
void *pv_gbytes_memcpy(void *dst, size_t dst_size, GBytes *src);
#define PV_GLIB_HELPER_ERROR g_quark_from_static_string("pv-glib-helper_error-quark")
typedef enum {
PV_GLIB_HELPER_FILE_ERROR,
} pv_glib_helper_error_e;
void pv_auto_close_file(FILE *file);
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(FILE, pv_auto_close_file)
#endif /* LIBPV_GLIB_HELPER_H */

119
include/libpv/hash.h Normal file
View File

@@ -0,0 +1,119 @@
/*
* Hashing definitions.
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#ifndef LIBPV_HASH_H
#define LIBPV_HASH_H
#include <openssl/hmac.h>
#include "libpv/common.h"
/** pv_digest_ctx_new:
* @md: mode of digest, e.g. #EVP_sha256()
* @error: return location for a #GError
*
* Returns: (nullable) (transfer full): a new #EVP_MD_CTX, or %NULL in case of an error
*/
EVP_MD_CTX *pv_digest_ctx_new(const EVP_MD *md, GError **error);
/** pv_digest_ctx_update:
* @ctx: EVP_MD_CTX to add data
* @data: #GBytes to add to the context
* @error: return location for a #GError
*
* Adds @data to the digest context. Can be called multiple times.
*
* Returns: 0 in case of success, -1 otherwise.
*/
int pv_digest_ctx_update(EVP_MD_CTX *ctx, GBytes *data, GError **error);
/** pv_digest_ctx_update_raw:
* @ctx: #EVP_MD_CTX to add data
* @buf: data to add to the context
* @size: size of @buf
* @error: return location for a #GError
*
* Adds @buf to the digest context. Can be called multiple times.
*
* Returns: 0 in case of success, -1 otherwise.
*/
int pv_digest_ctx_update_raw(EVP_MD_CTX *ctx, const uint8_t *buf, size_t size, GError **error);
/** pv_digest_ctx_finalize:
* @ctx: #EVP_MD_CTX with data to digest
* @error: return location for a #GError
*
* Calculates the digest of all previously added data. Do not use @ctx afterwards.
*
* Returns: (nullable) (transfer full): Digest of all data added before as #GBytes, or NULL in case of error.
*/
GBytes *pv_digest_ctx_finalize(EVP_MD_CTX *ctx, GError **error);
/** pv_sha256_hash:
* @buf: data for which a sha256 hash sould be calculated
* @size: size of @buf
* @error: return location for a #GError
*
* Shorthand for initializing a sha256-digest ctx, updating, and finalizing.
*
* Returns: (nullable) (transfer full): SHA256 of @buf as #GBytes, or NULL in case of error.
*/
GBytes *pv_sha256_hash(uint8_t *buf, size_t size, GError **error);
/** pv_hmac_ctx_new:
* @key: key used for the HMAC
* @md: mode of digest, e.g. #EVP_sha512()
* @error: return location for a #GError
*
* Returns: (nullable) (transfer full): New #HMAC_CTX or NULL in case of error
*/
HMAC_CTX *pv_hmac_ctx_new(GBytes *key, const EVP_MD *md, GError **error);
/** pv_hmac_ctx_update_raw:
* @ctx: #HMAC_CTX to add data
* @buf: data to add to the context
* @size: size of @buf
* @error: return location for a #GError
*
* Adds @buf to the HMAC context. Can be called multiple times.
*
* Returns: 0 in case of success, -1 otherwise.
*/
int pv_hmac_ctx_update_raw(HMAC_CTX *ctx, const void *data, size_t size, GError **error);
/** pv_hmac_ctx_update:
* @ctx: #HMAC_CTX to add data
* @data: #GBytes to add to the context
* @error: return location for a #GError
*
* Adds @data to the HMAC context. Can be called multiple times.
*
* Returns: 0 in case of success, -1 otherwise.
*/
int pv_hmac_ctx_update(HMAC_CTX *ctx, GBytes *data, GError **error);
/** pv_hmac_ctx_finalize:
* @ctx: #HMAC_CTX with data to digest
* @error: return location for a #GError
*
* Calculates the HMAC of all previously added data. Do not use @ctx afterwards.
*
* Returns: (nullable) (transfer full): HMAC of all data added before as #GBytes, or NULL in case of error.
*/
GBytes *pv_hamc_ctx_finalize(HMAC_CTX *ctx, GError **error);
#define PV_HASH_ERROR g_quark_from_static_string("pv-crypro-error-quark")
typedef enum {
PV_HASH_ERROR_INTERNAL,
} PvHashErrors;
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(EVP_MD_CTX, EVP_MD_CTX_free)
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(HMAC_CTX, HMAC_CTX_free)
#endif /* LIBPV_HASH_H */

16
include/libpv/macros.h Normal file
View File

@@ -0,0 +1,16 @@
/*
* Libpv common macro definitions.
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*
*/
#ifndef LIBPV_MACROS_H
#define LIBPV_MACROS_H
#define PV_NONNULL(...)
#define DO_PRAGMA(x) _Pragma(#x)
#endif /* LIBPV_MACROS_H */

View File

@@ -0,0 +1,29 @@
/*
* OpenSSL compatibility utils
*
* Copyright IBM Corp. 2021
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#ifndef LIBPV_OPENSSL_COMPAT_H
#define LIBPV_OPENSSL_COMPAT_H
#include <openssl/opensslv.h>
#include <openssl/x509.h>
#include <openssl/x509_vfy.h>
#if OPENSSL_VERSION_NUMBER >= 0x30000000L
#define pv_X509_STORE_CTX_get_current_cert(ctx) X509_STORE_CTX_get_current_cert(ctx)
#define pv_X509_STORE_CTX_get1_crls(ctx, nm) X509_STORE_CTX_get1_crls((ctx), (nm))
#define pv_X509_STORE_set_lookup_crls(st, cb) X509_STORE_set_lookup_crls(st, cb)
#elif OPENSSL_VERSION_NUMBER >= 0x10100000L
#define pv_X509_STORE_CTX_get_current_cert(ctx) \
X509_STORE_CTX_get_current_cert((X509_STORE_CTX *)(ctx))
#define pv_X509_STORE_CTX_get1_crls(ctx, nm) \
X509_STORE_CTX_get1_crls((X509_STORE_CTX *)(ctx), (X509_NAME *)(nm))
#define pv_X509_STORE_set_lookup_crls(st, cb) \
X509_STORE_set_lookup_crls(st, (X509_STORE_CTX_lookup_crls_fn)(cb))
#endif
#endif /* LIBPV_OPENSSL_COMPAT_H */

98
include/libpv/se-hdr.h Normal file
View File

@@ -0,0 +1,98 @@
/*
* PV/SE header definitions
*
* Copyright IBM Corp. 2020
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#ifndef LIBPV_SE_HDR_H
#define LIBPV_SE_HDR_H
#include "libpv/common.h"
#include <openssl/sha.h>
#include "boot/s390.h"
#include "libpv/crypto.h"
/* Magic number which is used to identify the file containing the PV
* header
*/
#define PV_MAGIC_NUMBER 0x49424d5365634578ULL
#define PV_VERSION_1 0x00000100U
/* Internal helper macro */
#define __PV_BIT(nr) (1ULL << (63 - (nr)))
/* Plaintext control flags */
/* dumping of the configuration is allowed */
#define PV_PCF_ALLOW_DUMPING __PV_BIT(34)
/* prevent Ultravisor decryption during unpack operation */
#define PV_PCF_NO_DECRYPTION __PV_BIT(35)
/* PCKMO encrypt-DEA/TDEA-key functions allowed */
#define PV_PCF_PCKMO_DEA_TDEA __PV_BIT(56)
/* PCKMO encrypt-AES-key functions allowed */
#define PV_PCF_PCKMO_AES __PV_BIT(57)
/* PCKMO encrypt-ECC-key functions allowed */
#define PV_PCF_PCKM_ECC __PV_BIT(58)
/* maxima for the PV version 1 */
#define PV_V1_IPIB_MAX_SIZE PAGE_SIZE
#define PV_V1_PV_HDR_MIN_SIZE \
(sizeof(struct pv_hdr_head) + sizeof(struct pv_hdr_encrypted) + \
sizeof(((struct pv_hdr *)0)->tag) + 1 * sizeof(struct pv_hdr_key_slot))
#define PV_V1_PV_HDR_MAX_SIZE (2 * PAGE_SIZE)
#define PV_IMAGE_ENCR_KEY_SIZE 64
typedef struct pv_hdr_key_slot {
uint8_t digest_key[SHA256_DIGEST_LENGTH];
uint8_t wrapped_key[32];
uint8_t tag[16];
} __packed PvHdrKeySlot;
typedef struct pv_hdr_opt_item {
uint32_t otype;
uint8_t ibk[32];
uint8_t data[];
} __packed PvHdrOptItem;
/* integrity protected data (by GCM tag), but non-encrypted */
struct pv_hdr_head {
uint64_t magic;
uint32_t version;
uint32_t phs;
uint8_t iv[12];
uint32_t res1;
uint64_t nks;
uint64_t sea;
uint64_t nep;
uint64_t pcf;
PvEcdhPubKey cust_pub_key;
uint8_t pld[SHA512_DIGEST_LENGTH];
uint8_t ald[SHA512_DIGEST_LENGTH];
uint8_t tld[SHA512_DIGEST_LENGTH];
} __packed;
/* Must not have any padding */
struct pv_hdr_encrypted {
uint8_t cust_comm_key[32];
uint8_t img_enc_key_1[PV_IMAGE_ENCR_KEY_SIZE / 2];
uint8_t img_enc_key_2[PV_IMAGE_ENCR_KEY_SIZE / 2];
struct psw_t psw;
uint64_t scf;
uint32_t noi;
uint32_t res2;
};
G_STATIC_ASSERT(sizeof(struct pv_hdr_encrypted) == 32 + 32 + 32 + sizeof(struct psw_t) + 8 + 4 + 4);
typedef struct pv_hdr {
struct pv_hdr_head head;
struct pv_hdr_key_slot *slots;
struct pv_hdr_encrypted *encrypted;
struct pv_hdr_opt_item **optional_items;
uint8_t tag[16];
} PvHdr;
#endif /* LIBPV_SE_HDR_H */

View File

@@ -139,8 +139,20 @@ the same as 0.0.5000).
.BR "\-L" " or " "\-\-loadparm"
Specifies an entry in the
.BR zipl (8)
boot menu. If this option is omitted, the default menu entry is used.
boot menu, and an active namespace in the installed zIPL environment block
for evaluation of zIPL environment variables in the kernel command line.
This option requires an argument, which has to be a string of the form
"nSX", where optional number 'n' indicates the boot menu entry, and optional
token 'SX' specifies the active namespace. This token consists of a mandatory
leading capital 'S' and of an optional trailing character 'X', which is
either any decimal digit, indicating ID of the active site, or a capital 'S'.
In the last case, ID of the active namespace is determined as the SSID
(Subchannel Set ID).
If the menu entry is omitted in the argument, then the default menu entry
is used. If the 'X' is omitted, or the whole token 'SX' is omitted, then the
common namespace of the installed zIPL environment block will be activated.
.TP
.BR "\-c" " or " "\-\-clear"
Specify whether memory should be cleared on re-IPL. Possible values are 0 to
@@ -158,9 +170,11 @@ loadparm:
\fB# chreipl ccw 0.0.7e78\fP
2. Next time reboot from the CCW device with the bus-ID 0.0.7e78
using the first entry of the zipl boot menu:
using the first entry of the zipl boot menu and the second site namespace
in the installed zIPL environment block for evaluation of zIPL environment
variables in the kernel command line:
\fB# chreipl ccw -d 0.0.7e78 -L 1\fP
\fB# chreipl ccw -d 0.0.7e78 -L 1S2\fP
.SH fcp
Use the fcp re-IPL target for SCSI disks that are accessed by the hardware
using Fibre Channel Protocol (FCP) channels.
@@ -306,4 +320,5 @@ that you would otherwise have to specify with the ccw or fcp target.
.BR lsreipl (8),
.BR zipl (8),
.BR zipl.conf (5),
.BR zipl-editenv (8)
.BR reboot (8)

37
libap/Makefile Normal file
View File

@@ -0,0 +1,37 @@
include ../common.mak
lib = libap.a
check-dep-lock:
touch check-dep-lock
ifneq (${HAVE_LOCKFILE},0)
$(call check_dep, \
"libap", \
"lockfile.h", \
"liblockfile-devel", \
"HAVE_LOCKFILE=0")
ALL_CPPFLAGS += -DHAVE_LOCKFILE
endif
check-dep-json:
touch check-dep-json
ifneq (${HAVE_JSONC},0)
$(call check_dep, \
"libap", \
"json-c/json.h", \
"json-c-devel", \
"HAVE_JSONC=0")
ALL_CPPFLAGS += -DHAVE_JSONC
endif
all: $(lib)
objects = ap.o
$(lib): $(objects)
$(objects): check-dep-lock check-dep-json
install: all
clean:
rm -f *.o check-dep-lock check-dep-json $(lib)

760
libap/ap.c Normal file
View File

@@ -0,0 +1,760 @@
/*
* libap - A collection of tools for ap/vfio-ap management
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#include <ctype.h>
#include <dirent.h>
#include <err.h>
#include <errno.h>
#include <stdarg.h>
#include <stdint.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#ifdef HAVE_JSONC
#include <json-c/json.h>
#endif /* HAVE_JSONC */
#ifdef HAVE_LOCKFILE
#include <lockfile.h>
#endif /* HAVE_LOCKFILE */
#include "lib/ap.h"
#include "lib/util_file.h"
#include "lib/util_libc.h"
#include "lib/util_panic.h"
#include "lib/util_path.h"
#include "lib/util_udev.h"
/*
* Return sysfs path to a bus attribute
* Note: caller is responsible for freeing the returned string
*/
static char *path_get_bus_attr(const char *bus, const char *attr)
{
return util_path_sysfs("bus/%s/%s", bus, attr);
}
/*
* Compare two vfio_ap nodes based upon their id value. Return:
* -1: a < b
* 1: a > b
* 0: a == b
*/
static int vfio_ap_node_cmp(void *a, void *b, void *UNUSED(data))
{
struct vfio_ap_node *a_node = a, *b_node = b;
if (a_node->id < b_node->id)
return -1;
return (a_node->id > b_node->id);
}
static void vfio_ap_node_add_tail(struct util_list *list, unsigned int val)
{
struct vfio_ap_node *node = util_zalloc(sizeof(struct vfio_ap_node));
node->id = val;
util_list_add_tail(list, node);
}
/* Remove duplicate entries from a sorted vfio_ap_node list */
static void vfio_ap_node_remove_dupes(struct util_list *list)
{
struct vfio_ap_node *node, *check, *next;
util_list_iterate(list, node) {
/* Remove any subsequent duplicates */
check = util_list_next(list, node);
while (check) {
next = util_list_next(list, check);
if (node->id == check->id) {
util_list_remove(list, check);
free(check);
}
check = next;
}
}
}
static bool starts_with(const char *str, const char *s)
{
size_t len = strlen(s);
return (strncmp(str, s, len) == 0);
}
/*
* Pass a comma-delimited string of IDs (adapter, domains or control domains)
* and add these IDs to the input list. The resulting list will be sorted
* and duplicates removed before returning.
*/
static void modify_device_attr(struct util_list *list, char *value)
{
unsigned int val;
char *curr;
curr = strtok(value, ",");
if (curr == NULL)
return;
/* Create list from input setting */
while (curr != NULL) {
val = strtol(curr, NULL, 0);
vfio_ap_node_add_tail(list, val);
curr = strtok(NULL, ",");
}
/* Cleanup the list */
util_list_sort(list, vfio_ap_node_cmp, NULL);
vfio_ap_node_remove_dupes(list);
}
static void load_attr_to_device(struct vfio_ap_device *dev, char *attr,
const char *value)
{
char *v = util_strdup(value);
if (strcmp(attr, "assign_adapter") == 0)
modify_device_attr(dev->adapters, v);
else if (strcmp(attr, "assign_domain") == 0)
modify_device_attr(dev->domains, v);
else if (strcmp(attr, "assign_control_domain") == 0)
modify_device_attr(dev->controls, v);
free(v);
}
/**
* Print the contents of the vfio-ap device struct to stderr. Used for
* debugging.
*
* @param[in] dev vfio-ap device strcucture to print
*/
void print_ap_device(struct vfio_ap_device *dev)
{
struct vfio_ap_node *node;
warnx("Device %s:", dev->uuid);
warnx("Type: %s", dev->type);
if (dev->manual)
warnx("Start: MANUAL");
else
warnx("Start: AUTO");
if (util_list_is_empty(dev->adapters)) {
warnx("Adapters: (none)");
} else {
warnx("Adapters:");
util_list_iterate(dev->adapters, node) {
warnx(" %u", node->id);
}
}
if (util_list_is_empty(dev->domains)) {
warnx("Domains: (none)");
} else {
warnx("Domains:");
util_list_iterate(dev->domains, node) {
warnx(" %u", node->id);
}
}
if (util_list_is_empty(dev->controls)) {
warnx("Controls: (none)");
} else {
warnx("Controls:");
util_list_iterate(dev->controls, node) {
warnx(" %u", node->id);
}
}
}
/**
* Determine if the input string is a valid Universally Unique Identifier
*
* @param[in] uuid Character string to inspect
*
* @retval true Specified string is a valid UUID
* @retval false Specified string is not a valid UUID
*/
bool is_valid_uuid(const char *uuid)
{
uint32_t s1, s2, s3, s4;
uint64_t s5;
char d;
return (strlen(uuid) == 36 && sscanf(uuid, "%8x-%4x-%4x-%4x-%12x %c",
&s1, &s2, &s3, &s4,
(unsigned int *) &s5, &d) == 5);
}
/*
* For an input string of hex characters, find the nth character and return its
* numeric value to the caller.
*/
static int get_hexbyte_value(int n, const char *hexbytestr)
{
int i = 0, v = 0;
char c;
if (strncmp(hexbytestr, "0x", 2) == 0)
i = 2;
/*
* The specified mask is only valid if it includes at least 64 hex
* digits. The specified mask may optionally include a leading '0x'
*/
util_assert((strlen(hexbytestr) >= (size_t)(AP_MASK_SIZE - i - 1)),
"Invalid hex string provided for mask: %s", hexbytestr);
c = hexbytestr[i + n / 4];
if (c >= '0' && c <= '9')
v = c - '0';
else if (c >= 'a' && c <= 'f')
v = 10 + c - 'a';
else if (c >= 'A' && c <= 'F')
v = 10 + c - 'A';
else
util_assert(false, "Could not parse hex digit '%c'", c);
return v;
}
/**
* For an input hex string, determine if the specified nth bit is ON or OFF.
*
* @param[in] n Bit number to test
* @param[in, out] hexbytestr Character string of hex characters
*
* @retval 0 Specified bit is OFF
* @retval != 0 Specified bit is ON
*/
int ap_test_bit(int n, const char *hexbytestr)
{
int v;
v = get_hexbyte_value(n, hexbytestr);
return v & (1 << (3 - (n % 4)));
}
/**
* For an input hex string, set the nth bit true/false
*
* @param[in] n Bit number to set
* @param[in, out] hexbytestr Character string of hex characters
* @param[in] val Bit is to be set ON (true) or OFF (false)
*/
void ap_set_bit(int n, char *hexbytestr, bool val)
{
char c = 0;
int v, m, i = 0;
v = get_hexbyte_value(n, hexbytestr);
/* Calculate the bit mask */
m = (1 << (3 - (n % 4)));
/* Return if bit already at correct value */
if (((val) && ((v & m) != 0)) || (!val && ((v & m) == 0)))
return;
if (val)
v = v + m;
else
v = v - m;
if (v < 10)
c = '0' + v;
else if (v >= 10 && v <= 15)
c = 'a' + (v - 10);
else
util_assert(false, "Could not set bit value '%d'", v);
if (strncmp(hexbytestr, "0x", 2) == 0)
i = 2;
/* Set the new value */
hexbytestr[i + n / 4] = c;
}
/**
* Return sysfs path to vfio_ap mdev
* Note: caller is responsible for freeing the returned string
*
* @param[in] uuid Character string containing an mdev UUID
*
* @retval != 0 sysfs path for a vfio-ap device with this UUID
*/
char *path_get_vfio_ap_mdev(const char *uuid)
{
return util_path_sysfs("%s/%s", VFIO_AP_PARENT_PATH, uuid);
}
/**
* Return path to mdevctl config file for specified UUID
* Note: caller is responsible for freeing the returned string
*
* @param[in] uuid Character string containing an mdev UUID
*
* @retval != 0 config file path for vfio-ap device with this UUID
*/
char *path_get_vfio_ap_mdev_config(const char *uuid)
{
char *path;
util_asprintf(&path, "%s/%s", VFIO_AP_CONFIG_PATH, uuid);
return path;
}
/**
* Return sysfs path to vfio_ap mdev attribute (matrix, remove, ...)
* Note: caller is responsible for freeing the returned string
*
* @param[in] uuid Character string containing an mdev UUID
* @param[in] attr Character string containing device attribute name
*
* @retval != 0 sysfs path to specified attribute for this device
*/
char *path_get_vfio_ap_attr(const char *uuid, const char *attr)
{
return util_path_sysfs("%s/%s/%s", VFIO_AP_PARENT_PATH, uuid, attr);
}
/**
* Return path to ap udev config file
* Note: caller is responsible for freeing the returned string
*
* @retval != 0 path to the ap udev config file
*/
char *path_get_ap_udev(void)
{
char *path;
util_asprintf(&path, "%s", AP_UDEV_FILE);
return path;
}
/**
* Take one line from the active 'matrix' attribute and parse it
* into a list of adapters and domains. Each line of the 'matrix'
* attribute is presented as a "adapter.domain" (e.g. "03.0005") where the
* numeric values are always hexadecimal.
* In a case where no adapters are assigned, a valid string might be ".0005"
* In a case where no domains are assigned, a valid string might be "03."
*
* @param[in, out] dev Vfio-ap struct that will be updated
* @param[in] matrix Character string to parse
*/
void vfio_ap_parse_matrix(struct vfio_ap_device *dev, char *matrix)
{
char *curr;
int val;
if (!matrix)
return;
if (*matrix != '.') {
/* Handle a device with adapters */
curr = strtok(matrix, ".");
val = strtol(curr, NULL, 16);
vfio_ap_node_add_tail(dev->adapters, val);
curr = strtok(NULL, "\n");
} else {
/* Handle a device with no adapters */
curr = strtok(matrix + 1, "\n");
}
/* Leave now if the device has no domains */
if (!curr)
return;
/* Get the domain */
val = strtol(curr, NULL, 16);
vfio_ap_node_add_tail(dev->domains, val);
}
/**
* Function to sort the results of repeated vfio_ap_parse_matrix calls
*
* @param[in, out] dev Vfio-ap struct whose lists will be sorted
*/
void vfio_ap_sort_matrix_results(struct vfio_ap_device *dev)
{
/* Sort the lists for later use */
util_list_sort(dev->adapters, vfio_ap_node_cmp, NULL);
util_list_sort(dev->domains, vfio_ap_node_cmp, NULL);
/* Run the lists and delete duplicates */
vfio_ap_node_remove_dupes(dev->adapters);
vfio_ap_node_remove_dupes(dev->domains);
}
/**
* Take the string provided by the active 'control_domains' attribute and
* parse it into a list of control domains
*
* @param[in, out] dev Vfio-ap struct that will be updated
* @param[in] control Character string to parse
*/
void vfio_ap_parse_control(struct vfio_ap_device *dev, char *control)
{
char *curr;
int val;
curr = strtok(control, "\n");
while (curr != NULL) {
val = strtol(curr, NULL, 16);
vfio_ap_node_add_tail(dev->controls, val);
curr = strtok(NULL, "\n");
}
}
#ifdef HAVE_JSONC
/**
* For a given path, read in the contents. If no path is provided, get the
* input from stdin instead.
*
* @param[in] path Path to mdevctl config file
* @param[in, out] dev Vfio-ap struct that will be updated
*
* @retval 0 Config read successfully, dev updated
* @retval -1 Failed to read config, dev may have partial info
*/
int vfio_ap_read_device_config(const char *path, struct vfio_ap_device *dev)
{
json_object *root, *type, *start, *attrs, *attr;
int i, len, rc = 0;
const char *val;
if (path == NULL)
root = json_object_from_fd(STDIN_FILENO);
else
root = json_object_from_file(path);
if (root == NULL)
return -1;
if (json_object_object_get_ex(root, "mdev_type", &type)) {
val = json_object_get_string(type);
if (!val)
goto err;
dev->type = util_strdup(val);
if (strcmp(val, "vfio_ap-passthrough") != 0)
goto err;
}
if (json_object_object_get_ex(root, "start", &start)) {
val = json_object_get_string(start);
if (!val)
goto err;
if (strcmp(val, "auto") == 0)
dev->manual = false;
else if (strcmp(val, "manual") == 0)
dev->manual = true;
else
goto err;
}
if (json_object_object_get_ex(root, "attrs", &attrs)) {
len = json_object_array_length(attrs);
for (i = 0; i < len; i++) {
attr = json_object_array_get_idx(attrs, i);
json_object_object_foreach(attr, key, setting) {
val = json_object_get_string(setting);
load_attr_to_device(dev, key, val);
}
}
}
out:
json_object_put(root);
return rc;
err:
rc = -1;
goto out;
}
#else
int vfio_ap_read_device_config(const char *path, struct vfio_ap_device *dev)
{
return -1;
}
#endif /* HAVE_JSONC */
/**
* Allocate and initialize a vfio-ap device structure.
*
* @retval !=0 Address of the new vfio-ap device structure
*/
struct vfio_ap_device *vfio_ap_device_new(void)
{
struct vfio_ap_device *dev;
dev = util_zalloc(sizeof(struct vfio_ap_device));
dev->manual = false;
dev->type = NULL;
dev->uuid = NULL;
dev->adapters = util_list_new(struct vfio_ap_node, node);
dev->domains = util_list_new(struct vfio_ap_node, node);
dev->controls = util_list_new(struct vfio_ap_node, node);
return dev;
}
/**
* Re-initialize a vfio-ap device structure, leaving the structure allocated.
*
* @param[in, out] dev Vfio-ap struct that will be re-initialized
*/
void vfio_ap_device_clear(struct vfio_ap_device *dev)
{
if (dev == NULL)
return;
dev->manual = false;
if (dev->type) {
free(dev->type);
dev->type = NULL;
}
if (dev->uuid) {
free(dev->uuid);
dev->uuid = NULL;
}
ap_list_remove_all(dev->adapters);
ap_list_remove_all(dev->domains);
ap_list_remove_all(dev->controls);
}
/**
* Clear and release a vfio-ap device structure.
*
* @param[in, out] dev Vfio-ap struct that will be freed
*/
void vfio_ap_device_free(struct vfio_ap_device *dev)
{
if (dev == NULL)
return;
vfio_ap_device_clear(dev);
util_list_free(dev->adapters);
util_list_free(dev->domains);
util_list_free(dev->controls);
free(dev);
}
static int read_sysfs_mask(const char *path, char *mask, int size)
{
return util_file_read_line(mask, size, "%s", path);
}
/**
* Get the apmask and aqmask from sysfs
*
* @param[in, out] ap Buffer to hold apmask contents
* @param[in, out] aq Buffer to hold aqmask contents
* @param[in] size Size of the mask buffers
*
* @retval 0 Both mask values read successfully
* @retval != 0 Failed to read one or both mask values
*/
int ap_read_sysfs_masks(char *ap, char *aq, int size)
{
char *path;
int rc = 0;
path = path_get_bus_attr("ap", "apmask");
rc = read_sysfs_mask(path, ap, size);
free(path);
if (rc != 0)
goto out;
path = path_get_bus_attr("ap", "aqmask");
rc = read_sysfs_mask(path, aq, size);
free(path);
out:
return rc;
}
/**
* Get the apmask and aqmask from udev, falling back to sysfs if a udev rule
* is not available or does not provide values for both masks.
* The values in read_ap and read_aq tell the caller whether each mask was
* successfully loaded from udev or if the sysfs value was substituted
* Note: both ap and aq must point to a string that is at least AP_MASK_SIZE
* in length.
*
* @param[in] path Path to the ap udev file
* @param[in, out] ap Buffer to hold apmask contents
* @param[in, out] aq Buffer to hold aqmask contents
* @param[out] read_ap Specifies if an apmask value was read from udev
* @param[out] read_aq Specifies if an aqmask value was read from udev
*
* @retval true Udev file read successfully or did not exist
* @retval false Udev file exists but error was encountered
*/
bool ap_read_udev_masks(char *path, char *ap, char *aq, bool *read_ap,
bool *read_aq)
{
struct util_udev_entry_node *entry;
struct util_udev_file *file = NULL;
struct util_udev_line_node *line;
char sysap[AP_MASK_SIZE];
char sysaq[AP_MASK_SIZE];
int rc;
/* Assume we fail to read both masks */
*read_ap = *read_aq = false;
/* If a udev file doesn't exist, quietly use the active masks */
if (!util_path_exists(path))
goto out;
rc = util_udev_read_file(path, &file);
/* If errors were encountered reading the udev file, exit now */
if (rc)
return false;
util_list_iterate(&file->lines, line) {
entry = util_list_start(&line->entries);
/* Skip comments and empty lines. */
if (!entry)
continue;
if (starts_with(entry->key, "ATTR{")) {
if (strstr(entry->key, "apmask")) {
util_strlcpy(ap, entry->value, AP_MASK_SIZE);
*read_ap = true;
} else if (strstr(entry->key, "aqmask")) {
util_strlcpy(aq, entry->value, AP_MASK_SIZE);
*read_aq = true;
}
}
}
util_udev_free_file(file);
out:
/* If we didn't read in masks, use current sysfs values */
if ((!*read_ap) || (!*read_aq)) {
rc = ap_read_sysfs_masks(sysap, sysaq, AP_MASK_SIZE);
if (rc != 0)
return false;
if (!*read_ap)
strcpy(ap, sysap);
if (!*read_aq)
strcpy(aq, sysaq);
}
return true;
}
/**
* For a given bitmask, create a list of vfio_ap_node entries corresponding
* to the ON bits in the mask.
*
* @param[in] mask Character string of hex characters
* @param[in, out] list List to be updated with entries for each ON bit
*/
void ap_mask_to_list(char *mask, struct util_list *list)
{
int i;
if (mask == NULL || list == NULL)
return;
for (i = 0; i <= AP_MAX_MASK_VALUE; i++) {
if (ap_test_bit(i, mask))
vfio_ap_node_add_tail(list, i);
}
/* Could have duplicates if the input list was not empty */
util_list_sort(list, vfio_ap_node_cmp, NULL);
vfio_ap_node_remove_dupes(list);
}
/**
* For the specified list, remove all elements and free each node
*
* @param[in, out] list List that will have all entries removed
*/
void ap_list_remove_all(struct util_list *list)
{
struct ap_node *node;
while (!util_list_is_empty(list)) {
node = util_list_start(list);
util_list_remove(list, node);
free(node);
}
}
#ifdef HAVE_LOCKFILE
static int ap_lockfile_create(int flags)
{
return lockfile_create(AP_LOCKFILE, AP_LOCK_RETRIES, flags);
}
/**
* Acquire the ap config lock using this process PID (L_PID)
*
* @retval 0 Lock successfully acquired on behalf of L_PID
* @retval != 0 Error, lock was not obtained
*/
int ap_get_lock(void)
{
return ap_lockfile_create(L_PID);
}
/**
* Acquire the ap config lock using the parent process PID (L_PPID) -- intended
* for use by the mdevctl callout ap-check utility
*
* @retval 0 Lock successfully acquired on behalf of L_PPID
* @retval != 0 Error, lock was not obtained
*/
int ap_get_lock_callout(void)
{
return ap_lockfile_create(L_PPID);
}
/**
* Release the ap config lock
*
* @retval 0 Lock successfully released or file didn't exist
* @retval != 0 Error removing the lockfile
*/
int ap_release_lock(void)
{
return lockfile_remove(AP_LOCKFILE);
}
#else
/* If no liblockfile, actions are performed without acquiring the file lock */
int ap_get_lock(void)
{
return 0;
}
int ap_get_lock_callout(void)
{
return 0;
}
int ap_release_lock(void)
{
return 0;
}
#endif /* HAVE_LOCKFILE */

View File

@@ -13,7 +13,7 @@ $(lib): $(objects)
install: all
libcpumf_example: libcpumf_example.o $(lib)
libcpumf_example: libcpumf_example.o $(lib) $(rootdir)/libutil/libutil.a
clean:
rm -f *.o $(lib) $(examples)

View File

@@ -14,6 +14,7 @@ int main(void)
{
unsigned long min, max, speed, sfb_min, sfb_max;
int rc, pmu, cfvn, csvn, auth;
char *pmuname;
cpu_set_t set;
pmu = libcpumf_pmutype(S390_CPUMF_CF);
@@ -72,5 +73,17 @@ int main(void)
if (rc)
printf(" sfb_min %lu sfb_max %lu", sfb_min, sfb_max);
putchar('\n');
printf("PAI crypto support %d\n", libcpumf_have_pai_crypto());
rc = libcpumf_pmuname(10, &pmuname);
if (rc) {
printf("PMU type 10 PMU name lookup error %d\n", rc);
} else {
printf("PMU type 10 PMU name %s\n", pmuname);
free(pmuname);
}
printf("PAI NNPA support %d\n", libcpumf_have_pai_nnpa());
return EXIT_SUCCESS;
}

View File

@@ -4,10 +4,16 @@
* it under the terms of the MIT license. See LICENSE for details.
*/
#include <err.h>
#include <errno.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include "lib/util_file.h"
#include "lib/util_libc.h"
#include "lib/util_path.h"
#include "lib/util_scandir.h"
#include "lib/libcpumf.h"
int libcpumf_pmutype(const char *dirname)
@@ -31,3 +37,36 @@ int libcpumf_pmutype(const char *dirname)
}
return ret;
}
int libcpumf_pmuname(unsigned int wanted_type, char **name)
{
struct dirent **de_vec;
unsigned int type;
char *dirname;
char *path;
int count;
int rc;
path = util_path_sysfs("devices");
count = util_scandir(&de_vec, alphasort, path, "(pai|cpum_).*");
free(path);
*name = NULL;
for (int i = 0; i < count; i++) {
if (de_vec[i]->d_type == DT_DIR) {
dirname = de_vec[i]->d_name;
path = util_path_sysfs("devices/%s/type", dirname);
rc = util_file_read_ui(&type, 10, path);
if (rc)
warn("Failed to open %s", path);
free(path);
if (!rc && type == wanted_type) {
*name = util_strdup(dirname);
goto out;
}
}
}
out:
util_scandir_free(de_vec, count);
return *name ? 0 : -1;
}

View File

@@ -139,3 +139,17 @@ bool libcpumf_sfb_info(unsigned long *min, unsigned long *max)
fclose(fp);
return rc;
}
bool libcpumf_have_pai_crypto(void)
{
struct stat statbuf;
return (stat(S390_SYSFS_PAI_CRYPTO, &statbuf) == -1) ? false : true;
}
bool libcpumf_have_pai_nnpa(void)
{
struct stat statbuf;
return (stat(S390_SYSFS_PAI_NNPA, &statbuf) == -1) ? false : true;
}

View File

@@ -259,7 +259,7 @@ int kmip_decode_bignum(const unsigned char *data, uint32_t length, BIGNUM **bn)
if (data[0] & 0x80) {
neg = 1;
tmp = malloc(length);
tmp = calloc(1, length);
if (tmp == NULL)
return -ENOMEM;

101
libpv/Makefile Normal file
View File

@@ -0,0 +1,101 @@
# Common definitions
include ../common.mak
.DEFAULT_GOAL := all
LIB := libpv.a
ifneq ($(shell sh -c 'command -v pkg-config'),)
GLIB2_CFLAGS := $(shell pkg-config --silence-errors --cflags glib-2.0)
GLIB2_LIBS := $(shell pkg-config --silence-errors --libs glib-2.0)
LIBCRYPTO_CFLAGS := $(shell pkg-config --silence-errors --cflags libcrypto openssl)
LIBCRYPTO_LIBS := $(shell pkg-config --silence-errors --libs libcrypto openssl)
LIBCURL_CFLAGS := $(shell pkg-config --silence-errors --cflags libcurl)
LIBCURL_LIBS := $(shell pkg-config --silence-errors --libs libcurl)
else
GLIB2_CFLAGS := -I/usr/include/glib-2.0 -I/usr/lib64/glib-2.0/include
GLIB2_LIBS := -lglib-2.0
LIBCRYPTO_CFLAGS :=
LIBCRYPTO_LIBS := -lcrypto -lssl
LIBCURL_CFLAGS := -I/usr/include/s390x-linux-gnu
LIBCURL_LIBS := -lcurl
endif
LDLIBS += $(GLIB2_LIBS) $(LIBCRYPTO_LIBS) $(LIBCURL_LIBS)
WARNINGS := -Wall -Wextra -Wshadow \
-Wcast-align -Wwrite-strings -Wmissing-prototypes \
-Wmissing-declarations -Wredundant-decls -Wnested-externs \
-Wno-long-long -Wuninitialized -Wconversion -Wstrict-prototypes \
-Wpointer-arith -Wno-error=inline \
-Wno-unused-function -Wno-unused-parameter -Wno-unused-variable \
-Werror \
$(NULL)
ALL_CFLAGS += -std=gnu11 \
-DOPENSSL_API_COMPAT=0x10101000L \
$(GLIB2_CFLAGS) \
$(LIBCRYPTO_CFLAGS) \
$(LIBCURL_CFLAGS) \
$(WARNINGS) \
$(NULL)
BUILD_TARGETS := skip-$(LIB)
ifneq (${HAVE_OPENSSL},0)
ifneq (${HAVE_GLIB2},0)
ifneq (${HAVE_LIBCURL},0)
BUILD_TARGETS := $(LIB)
endif
endif
endif
sources := $(wildcard *.c)
objects := $(patsubst %.c,%.o,$(sources))
all: $(BUILD_TARGETS) .check_dep-$(LIB)
$(LIB): $(objects)
$(LIB): ALL_CFLAGS += -fPIC
$(objects): .check-dep-$(LIB)
install: all
clean:
rm -f -- $(objects)
rm -f -- $(LIB)
rm -f -- .check-dep-$(LIB) .detect-openssl.dep.c
skip-$(LIB):
echo " SKIP $(LIB) due to unresolved dependencies"
.PHONY: all install clean skip-$(LIB) install-$(LIB)
.detect-openssl.dep.c:
echo "#include <openssl/evp.h>" > $@
echo "#if OPENSSL_VERSION_NUMBER < 0x10101000L" >> $@
echo " #error openssl version 1.1.0 is required" >> $@
echo "#endif" >> $@
echo "static void __attribute__((unused)) test(void) {" >> $@
echo " EVP_MD_CTX *ctx = EVP_MD_CTX_new();" >> $@
echo " EVP_MD_CTX_free(ctx);" >> $@
echo "}" >> $@
.check-dep-$(LIB): .detect-openssl.dep.c
$(call check_dep, \
"$(LIB)", \
"glib.h", \
"glib2-devel / libglib2.0-dev", \
"HAVE_GLIB2=0")
$(call check_dep, \
"$(LIB)", \
$^, \
"openssl-devel / libssl-dev version >= 1.1.0", \
"HAVE_OPENSSL=0", \
"-I.")
$(call check_dep, \
"$(LIB)", \
"curl/curl.h", \
"libcurl-devel", \
"HAVE_LIBCURL=0")
touch $@

1645
libpv/cert.c Normal file

File diff suppressed because it is too large Load Diff

45
libpv/common.c Normal file
View File

@@ -0,0 +1,45 @@
/*
* Libpv common functions.
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*
*/
/* Must be included before any other header */
#include "config.h"
#include "libpv/common.h"
#include "libpv/cert.h"
#include "libpv/curl.h"
/* setup and tear down */
int pv_init(void)
{
static size_t openssl_initalized;
if (g_once_init_enter(&openssl_initalized)) {
if (OPENSSL_VERSION_NUMBER < 0x1000100fL)
g_assert_not_reached();
#if OPENSSL_VERSION_NUMBER < 0x10100000L
SSL_library_init();
SSL_load_error_strings();
#else
OPENSSL_init_crypto(0, NULL);
#endif
if (pv_curl_init() != 0)
return -1;
pv_cert_init();
g_once_init_leave(&openssl_initalized, 1);
}
return 0;
}
void pv_cleanup(void)
{
pv_cert_cleanup();
pv_curl_cleanup();
}

15
libpv/config.h Normal file
View File

@@ -0,0 +1,15 @@
/*
* Config file.
* Must be include before any other header.
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*
*/
#ifndef LIBPV_CONFIG_H
#define LIBPV_CONFIG_H
#define GETTEXT_PACKAGE "libpv"
#endif /* LIBPV_CONFIG_H */

529
libpv/crypto.c Normal file
View File

@@ -0,0 +1,529 @@
/*
* Cryptography functions
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
/* Must be included before any other header */
#include "config.h"
#include <openssl/bio.h>
#include <openssl/err.h>
#include <openssl/rand.h>
#include <openssl/kdf.h>
#include "lib/zt_common.h"
#include "libpv/crypto.h"
#include "libpv/glib-helper.h"
#include "libpv/hash.h"
const char *pv_get_openssl_error(void)
{
const char *ret;
BIO *bio;
char *buf;
long len;
bio = BIO_new(BIO_s_mem());
ERR_print_errors(bio);
len = BIO_get_mem_data(bio, &buf);
if (len < 0)
ret = "Cannot receive OpenSSL error message.";
else
ret = g_strndup(buf, (size_t)len);
BIO_free(bio);
return ret;
}
int pv_BIO_reset(BIO *b)
{
int rc = BIO_reset(b);
if (rc != 1 && (BIO_method_type(b) == BIO_TYPE_FILE && rc != 0))
return -1;
return 1;
}
GBytes *pv_generate_rand_data(size_t size, GError **error)
{
g_autofree uint8_t *data = NULL;
if (size > INT_MAX) {
g_set_error_literal(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_RANDOMIZATION,
"Too many random data requested. Split it up");
OPENSSL_clear_free(data, size);
return NULL;
}
data = g_malloc(size);
if (RAND_bytes(data, (int)size) != 1) {
g_set_error_literal(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_RANDOMIZATION,
"The required amount of random data is not available");
return NULL;
}
return pv_sec_gbytes_new_take(g_steal_pointer(&data), size);
}
GBytes *pv_generate_key(const EVP_CIPHER *cipher, GError **error)
{
int size;
pv_wrapped_g_assert(cipher);
size = EVP_CIPHER_key_length(cipher);
if (size <= 0) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_KEYGENERATION,
"Unknown cipher");
return NULL;
}
return pv_generate_rand_data((guint)size, error);
}
GBytes *pv_generate_iv(const EVP_CIPHER *cipher, GError **error)
{
int size;
pv_wrapped_g_assert(cipher);
size = EVP_CIPHER_iv_length(cipher);
if (size <= 0) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_KEYGENERATION,
"Unknown cipher");
return NULL;
}
return pv_generate_rand_data((guint)size, error);
}
static int64_t pv_gcm_encrypt_decrypt(GBytes *input, GBytes *aad, const PvCipherParms *parms,
GBytes **output, GBytes **tagp, enum PvCryptoMode mode,
GError **error)
{
const uint8_t *in_data, *aad_data = NULL, *iv_data, *key_data;
size_t in_size, aad_size = 0, iv_size, key_size, out_size;
const EVP_CIPHER *cipher = parms->cipher;
const size_t tag_size = parms->tag_size;
gboolean encrypt = mode == PV_ENCRYPT;
g_autoptr(EVP_CIPHER_CTX) ctx = NULL;
g_autofree uint8_t *out_data = NULL;
g_autofree uint8_t *tag_data = NULL;
const GBytes *key = parms->key;
const GBytes *iv = parms->iv;
int cipher_block_size;
int64_t ret = -1;
int len = -1;
GBytes *tag;
g_assert(tagp);
g_assert(cipher);
g_assert(key);
g_assert(iv);
tag = *tagp;
in_data = g_bytes_get_data((GBytes *)input, &in_size);
if (aad)
aad_data = g_bytes_get_data((GBytes *)aad, &aad_size);
iv_data = g_bytes_get_data((GBytes *)iv, &iv_size);
key_data = g_bytes_get_data((GBytes *)key, &key_size);
out_size = in_size;
cipher_block_size = EVP_CIPHER_block_size(cipher);
/* Checks for later casts */
g_assert(aad_size <= INT_MAX);
g_assert(in_size <= INT_MAX);
g_assert(iv_size <= INT_MAX);
g_assert(cipher_block_size > 0);
ctx = EVP_CIPHER_CTX_new();
if (!ctx)
g_abort();
if (tag_size == 0 || (tag_size % (size_t)cipher_block_size != 0)) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_INTERNAL,
"Passed tag size is incorrect");
return -1;
}
/* Has the passed key the correct size? */
if (EVP_CIPHER_key_length(cipher) != (int)key_size) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_INTERNAL,
"Passed key has incorrect size: %ld != %d", key_size,
EVP_CIPHER_key_length(cipher));
return -1;
}
/* First, set the cipher algorithm so we can verify our key/IV lengths
*/
if (EVP_CipherInit_ex(ctx, cipher, NULL, NULL, NULL, encrypt) != 1) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_INTERNAL,
"EVP_CIPHER_CTX_new failed");
return -1;
}
/* Set IV length */
if (EVP_CIPHER_CTX_ctrl(ctx, EVP_CTRL_GCM_SET_IVLEN, (int)iv_size, NULL) != 1) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_INTERNAL,
"EVP_CIPHER_CTX_ex failed");
return -1;
}
/* Initialise key and IV */
if (EVP_CipherInit_ex(ctx, NULL, NULL, key_data, iv_data, encrypt) != 1) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_INTERNAL,
"EVP_CipherInit_ex failed");
return -1;
}
/* Allocate output data */
out_data = g_malloc0(out_size);
if (encrypt)
tag_data = g_malloc0(tag_size);
if (aad_size > 0) {
/* Provide any AAD data */
if (EVP_CipherUpdate(ctx, NULL, &len, aad_data, (int)aad_size) != 1) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_INTERNAL,
"EVP_CipherUpdate failed");
return -1;
}
g_assert(len == (int)aad_size);
}
/* Provide data to be en/decrypted */
if (EVP_CipherUpdate(ctx, out_data, &len, in_data, (int)in_size) != 1) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_INTERNAL,
"EVP_CipherUpdate failed");
return -1;
}
ret = len;
if (!encrypt) {
const uint8_t *tmp_tag_data = NULL;
size_t tmp_tag_size = 0;
if (tag)
tmp_tag_data = g_bytes_get_data(tag, &tmp_tag_size);
if (tag_size != tmp_tag_size) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_INTERNAL,
"Getting the GCM tag failed");
return -1;
}
/* Set expected tag value */
if (EVP_CIPHER_CTX_ctrl(ctx, EVP_CTRL_GCM_SET_TAG, (int)tmp_tag_size,
(uint8_t *)tmp_tag_data) != 1) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_INTERNAL,
"Setting the GCM tag failed");
return -1;
}
}
/* Finalize the en/decryption */
if (EVP_CipherFinal_ex(ctx, (uint8_t *)out_data + len, &len) != 1) {
if (encrypt)
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_INTERNAL,
"Encrypting failed (EVP_CipherFinal_ex)");
else
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_NO_MATCH_TAG,
"Verifying the GCM tag failed");
return -1;
}
ret += len;
if (encrypt) {
/* Get the tag */
if (EVP_CIPHER_CTX_ctrl(ctx, EVP_CTRL_GCM_GET_TAG, (int)tag_size, tag_data) != 1) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_INTERNAL,
"Getting the GCM tag failed");
return -1;
}
g_assert(!*tagp);
*tagp = g_bytes_new_take(g_steal_pointer(&tag_data), tag_size);
}
g_assert(ret == (int)out_size);
g_assert(out_size == in_size);
g_assert(!*output);
*output = pv_sec_gbytes_new_take(g_steal_pointer(&out_data), out_size);
return ret;
}
int64_t pv_gcm_encrypt(GBytes *plain, GBytes *aad, const PvCipherParms *parms, GBytes **cipher,
GBytes **tag, GError **error)
{
pv_wrapped_g_assert(plain);
pv_wrapped_g_assert(parms);
pv_wrapped_g_assert(cipher);
pv_wrapped_g_assert(tag);
return pv_gcm_encrypt_decrypt(plain, aad, parms, cipher, tag, PV_ENCRYPT, error);
}
int64_t pv_gcm_decrypt(GBytes *cipher, GBytes *aad, GBytes *tag, const PvCipherParms *parms,
GBytes **plain, GError **error)
{
pv_wrapped_g_assert(cipher);
pv_wrapped_g_assert(tag);
pv_wrapped_g_assert(parms);
pv_wrapped_g_assert(plain);
return pv_gcm_encrypt_decrypt(cipher, aad, parms, plain, &tag, PV_DECRYPT, error);
}
GBytes *pv_hkdf_extract_and_expand(size_t derived_key_len, GBytes *key, GBytes *salt, GBytes *info,
const EVP_MD *md, G_GNUC_UNUSED GError **error)
{
const unsigned char *salt_data, *key_data, *info_data;
g_autoptr(EVP_PKEY_CTX) ctx = NULL;
size_t salt_len, key_len, info_len;
g_autofree unsigned char *derived_key = NULL;
g_assert(derived_key_len > 0);
pv_wrapped_g_assert(key);
pv_wrapped_g_assert(salt);
pv_wrapped_g_assert(info);
pv_wrapped_g_assert(md);
ctx = EVP_PKEY_CTX_new_id(EVP_PKEY_HKDF, NULL);
if (!ctx)
g_abort();
if (EVP_PKEY_derive_init(ctx) != 1) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_HKDF_FAIL,
"FAILED to derive key via HKDF");
return NULL;
}
if (EVP_PKEY_CTX_hkdf_mode(ctx, EVP_PKEY_HKDEF_MODE_EXTRACT_AND_EXPAND) != 1) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_HKDF_FAIL,
"FAILED to derive key via HKDF");
return NULL;
}
if (EVP_PKEY_CTX_set_hkdf_md(ctx, md) != 1) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_HKDF_FAIL,
"FAILED to derive key via HKDF");
return NULL;
}
salt_data = g_bytes_get_data(salt, &salt_len);
if (salt_len > INT_MAX) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_HKDF_FAIL,
"FAILED to derive key via HKDF");
return NULL;
}
if (EVP_PKEY_CTX_set1_hkdf_salt(ctx, salt_data, (int)salt_len) != 1) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_HKDF_FAIL,
"FAILED to derive key via HKDF");
return NULL;
}
key_data = g_bytes_get_data(key, &key_len);
if (key_len > INT_MAX) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_HKDF_FAIL,
"FAILED to derive key via HKDF");
return NULL;
}
if (EVP_PKEY_CTX_set1_hkdf_key(ctx, key_data, (int)key_len) != 1) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_HKDF_FAIL,
"FAILED to derive key via HKDF");
return NULL;
}
info_data = g_bytes_get_data(info, &info_len);
if (info_len > INT_MAX) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_HKDF_FAIL,
"FAILED to derive key via HKDF");
return NULL;
}
if (EVP_PKEY_CTX_add1_hkdf_info(ctx, (unsigned char *)info_data, (int)info_len) != 1) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_HKDF_FAIL,
"FAILED to derive key via HKDF");
return NULL;
}
derived_key = g_malloc0(derived_key_len);
if (EVP_PKEY_derive(ctx, derived_key, &derived_key_len) != 1) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_HKDF_FAIL,
"FAILED to derive key via HKDF");
printf("%s\n", pv_get_openssl_error());
return NULL;
}
return pv_sec_gbytes_new_take(g_steal_pointer(&derived_key), derived_key_len);
}
EVP_PKEY *pv_generate_ec_key(int nid, GError **error)
{
g_autoptr(EVP_PKEY_CTX) ctx = EVP_PKEY_CTX_new_id(EVP_PKEY_EC, NULL);
g_autoptr(EVP_PKEY) ret = NULL;
g_assert(ctx);
if (EVP_PKEY_keygen_init(ctx) != 1) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_KEYGENERATION,
_("EC key could not be auto-generated"));
return NULL;
}
if (EVP_PKEY_CTX_set_ec_paramgen_curve_nid(ctx, nid) != 1) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_KEYGENERATION,
_("EC key could not be auto-generated"));
return NULL;
}
if (EVP_PKEY_keygen(ctx, &ret) != 1) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_KEYGENERATION,
_("EC key could not be auto-generated"));
return NULL;
}
return g_steal_pointer(&ret);
}
/* Convert a EVP_PKEY to the key format used in the PV header */
PvEcdhPubKey *pv_evp_pkey_to_ecdh_pub_key(EVP_PKEY *key, GError **error)
{
g_autofree PvEcdhPubKey *ret = g_new0(PvEcdhPubKey, 1);
g_autoptr(BIGNUM) pub_x_big = NULL, pub_y_big = NULL;
g_autoptr(EC_KEY) ec_key = NULL;
const EC_POINT *pub_key;
const EC_GROUP *grp;
pv_wrapped_g_assert(key);
ec_key = EVP_PKEY_get1_EC_KEY(key);
if (!ec_key) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_INTERNAL,
_("Key has the wrong type"));
return NULL;
}
pub_key = EC_KEY_get0_public_key(ec_key);
if (!pub_key) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_INTERNAL,
_("Failed to get public key"));
return NULL;
}
grp = EC_KEY_get0_group(ec_key);
if (!grp) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_INTERNAL,
_("Failed to get EC group"));
return NULL;
}
pub_x_big = BN_new();
if (!pub_x_big)
g_abort();
pub_y_big = BN_new();
if (!pub_y_big)
g_abort();
if (EC_POINT_get_affine_coordinates_GFp(grp, pub_key, pub_x_big, pub_y_big, NULL) != 1) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_INTERNAL,
_("Cannot convert key to internal format"));
return NULL;
}
if (BN_bn2binpad(pub_x_big, ret->x, sizeof(ret->x)) < 0) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_INTERNAL,
_("Cannot convert key to internal format"));
return NULL;
}
if (BN_bn2binpad(pub_y_big, ret->y, sizeof(ret->y)) < 0) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_INTERNAL,
_("Cannot convert key to internal format"));
return NULL;
}
return g_steal_pointer(&ret);
}
static GBytes *derive_key(EVP_PKEY *key1, EVP_PKEY *key2, GError **error)
{
g_autoptr(EVP_PKEY_CTX) ctx = NULL;
uint8_t *data = NULL;
size_t data_size, key_size;
ctx = EVP_PKEY_CTX_new(key1, NULL);
if (!ctx)
g_abort();
if (EVP_PKEY_derive_init(ctx) != 1) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_INTERNAL,
_("Key derivation failed"));
return NULL;
}
if (EVP_PKEY_derive_set_peer(ctx, key2) != 1) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_INTERNAL,
_("Key derivation failed"));
return NULL;
}
/* Determine buffer length */
if (EVP_PKEY_derive(ctx, NULL, &key_size) != 1) {
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_DERIVE,
_("Key derivation failed"));
return NULL;
}
data_size = key_size;
data = OPENSSL_malloc(data_size);
if (!data)
g_abort();
if (EVP_PKEY_derive(ctx, data, &data_size) != 1) {
OPENSSL_clear_free(data, data_size);
g_set_error(error, PV_CRYPTO_ERROR, PV_CRYPTO_ERROR_DERIVE,
_("Key derivation failed"));
return NULL;
}
g_assert(data_size == key_size);
return pv_sec_gbytes_new_take(g_steal_pointer(&data), data_size);
}
GBytes *pv_derive_exchange_key(EVP_PKEY *cust, EVP_PKEY *host, GError **error)
{
const guint8 append[] = { 0x00, 0x00, 0x00, 0x01 };
g_autoptr(GBytes) derived_key = NULL, ret = NULL;
g_autoptr(GByteArray) der_key_ga = NULL;
g_autofree uint8_t *raw = NULL;
size_t raw_len;
pv_wrapped_g_assert(cust);
pv_wrapped_g_assert(host);
derived_key = derive_key(cust, host, error);
if (!derived_key)
return NULL;
der_key_ga = g_bytes_unref_to_array(g_steal_pointer(&derived_key));
/* ANSI X.9.63-2011: 66 bytes x with leading 7 bits and
* concatenate 32 bit int '1'
*/
der_key_ga = g_byte_array_append(der_key_ga, append, sizeof(append));
/* free GBytesArray and get underlying data */
raw_len = der_key_ga->len;
raw = g_byte_array_free(g_steal_pointer(&der_key_ga), FALSE);
ret = pv_sha256_hash(raw, raw_len, error);
OPENSSL_cleanse(raw, raw_len);
return g_steal_pointer(&ret);
}
GQuark pv_crypto_error_quark(void)
{
return g_quark_from_static_string("pv-crypto-error-quark");
}

116
libpv/curl.c Normal file
View File

@@ -0,0 +1,116 @@
/*
* Libcurl utils
*
* Copyright IBM Corp. 2020
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
/* Must be included before any other header */
#include "config.h"
#include <curl/curl.h>
#include <stdio.h>
#include "lib/zt_common.h"
#include "libpv/curl.h"
struct UserData {
GByteArray *buffer;
uint max_size;
};
static size_t write_callback(char *ptr, size_t size, size_t nmemb, void *userdata)
{
g_assert(userdata);
struct UserData *data = (struct UserData *)userdata;
GByteArray *buffer = data->buffer;
uint64_t actual_size;
size_t err;
g_assert(buffer);
if (!g_uint64_checked_mul(&actual_size, size, nmemb))
g_abort();
/* Signal an error condition by returning a amount that differs
* from the amount passed to the callback. This results in a
* CURLE_WRITE_ERROR.
*/
err = actual_size + 1;
if (actual_size > G_MAXUINT)
return err;
data->buffer = g_byte_array_append(buffer, (uint8_t *)ptr, (uint)actual_size);
if (data->buffer->len > data->max_size)
return err;
return actual_size;
}
int pv_curl_init(void)
{
if (curl_global_init(CURL_GLOBAL_ALL) != 0)
return -1;
return 0;
}
void pv_curl_cleanup(void)
{
curl_global_cleanup();
}
GByteArray *curl_download(const char *url, long timeout_ms, uint max_size, GError **err)
{
g_autoptr(GByteArray) ret = NULL;
g_autoptr(CURL) handle = NULL;
g_autofree char *agent = NULL;
struct UserData userdata;
CURLcode rc;
/* set up curl session */
handle = curl_easy_init();
if (!handle)
g_abort();
/* follow redirection */
rc = curl_easy_setopt(handle, CURLOPT_FOLLOWLOCATION, 1L);
if (rc != CURLE_OK)
goto curl_err;
rc = curl_easy_setopt(handle, CURLOPT_TIMEOUT_MS, timeout_ms);
if (rc != CURLE_OK)
goto curl_err;
rc = curl_easy_setopt(handle, CURLOPT_NOSIGNAL, 1L);
if (rc != CURLE_OK)
goto curl_err;
agent = g_strdup_printf("%s/%s", GETTEXT_PACKAGE, RELEASE_STRING);
rc = curl_easy_setopt(handle, CURLOPT_USERAGENT, agent);
if (rc != CURLE_OK)
goto curl_err;
rc = curl_easy_setopt(handle, CURLOPT_WRITEFUNCTION, write_callback);
if (rc != CURLE_OK)
goto curl_err;
ret = g_byte_array_new();
userdata.buffer = ret;
userdata.max_size = max_size;
rc = curl_easy_setopt(handle, CURLOPT_WRITEDATA, (void *)&userdata);
if (rc != CURLE_OK)
goto curl_err;
rc = curl_easy_setopt(handle, CURLOPT_URL, url);
if (rc != CURLE_OK)
goto curl_err;
rc = curl_easy_perform(handle);
if (rc != CURLE_OK) {
g_set_error(err, PV_CURL_ERROR, PV_CURL_ERROR_DOWNLOAD_FAILED,
_("download failed: %s"), curl_easy_strerror(rc));
return NULL;
}
return g_steal_pointer(&ret);
curl_err:
g_set_error(err, PV_CURL_ERROR, PV_CURL_ERROR_CURL_INIT_FAILED,
_("cURL initialization failed: %s"), curl_easy_strerror(rc));
return NULL;
}

179
libpv/glib-helper.c Normal file
View File

@@ -0,0 +1,179 @@
/*
* Glib convenience functions
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
/* Must be included before any other header */
#include "config.h"
#include <errno.h>
#include <openssl/crypto.h>
#include "libpv/glib-helper.h"
struct __data {
void *data;
size_t size;
GFreeFunc free_func;
};
static void __data_clear_and_free(void *p)
{
struct __data *ptr = p;
if (!ptr)
return;
if (ptr->data) {
OPENSSL_cleanse(ptr->data, ptr->size);
ptr->free_func(ptr->data);
}
g_free(ptr);
}
static GBytes *pv_sec_gbytes_new_take_func(void *data, size_t size, GFreeFunc free_func)
{
struct __data *tmp = g_new(struct __data, 1);
tmp->data = data;
tmp->size = size;
tmp->free_func = free_func;
return g_bytes_new_with_free_func(data, size, __data_clear_and_free, tmp);
}
GBytes *pv_sec_gbytes_new_take(void *data, size_t size)
{
return pv_sec_gbytes_new_take_func(data, size, g_free);
}
GBytes *pv_sec_gbytes_new(const void *data, size_t size)
{
g_autofree void *tmp_data = NULL;
g_return_val_if_fail(data || size != 0, NULL);
tmp_data = g_malloc(size);
memcpy(tmp_data, data, size);
return pv_sec_gbytes_new_take(g_steal_pointer(&tmp_data), size);
}
int pv_file_seek(FILE *file, long offset, int whence, GError **error)
{
int cached_errno;
int ret = fseek(file, offset, whence);
if (ret) {
cached_errno = errno;
g_set_error(error, PV_GLIB_HELPER_ERROR, PV_GLIB_HELPER_FILE_ERROR,
"Cannot seek: %s", g_strerror(cached_errno));
}
return ret;
}
size_t pv_file_write(FILE *file, const void *ptr, size_t size, GError **error)
{
int cached_errno;
size_t n = fwrite(ptr, 1, size, file);
if (n != size) {
cached_errno = errno;
g_set_error(error, PV_GLIB_HELPER_ERROR, PV_GLIB_HELPER_FILE_ERROR,
"Cannot write: %s", g_strerror(cached_errno));
}
return n;
}
long pv_file_close(FILE *file, GError **error)
{
int cached_errno;
int ret = fclose(file);
if (ret) {
cached_errno = errno;
g_set_error(error, PV_GLIB_HELPER_ERROR, PV_GLIB_HELPER_FILE_ERROR,
"Cannot close: %s", g_strerror(cached_errno));
}
return ret;
}
void pv_auto_close_file(FILE *file)
{
if (!file)
return;
(void)pv_file_close(file, NULL);
}
long pv_file_tell(FILE *file, GError **error)
{
int cached_errno;
long n = ftell(file);
if (n < 0) {
cached_errno = errno;
g_set_error(error, PV_GLIB_HELPER_ERROR, PV_GLIB_HELPER_FILE_ERROR,
"Cannot tell: %s", g_strerror(cached_errno));
}
return n;
}
FILE *pv_file_open(const char *filename, const char *mode, GError **error)
{
FILE *file = fopen(filename, mode);
int cached_errno;
if (!file) {
cached_errno = errno;
g_set_error(error, PV_GLIB_HELPER_ERROR, PV_GLIB_HELPER_FILE_ERROR,
"Cannot open '%s'. %s", filename, g_strerror(cached_errno));
return NULL;
}
return file;
}
GBytes *pv_file_get_content_as_g_bytes(const char *filename, GError **error)
{
g_autofree char *data = NULL;
size_t data_size;
if (!g_file_get_contents(filename, &data, &data_size, error))
return NULL;
return g_bytes_new_take(g_steal_pointer(&data), data_size);
}
GBytes *pv_file_get_content_as_secure_bytes(const char *filename)
{
g_autoptr(FILE) f = fopen(filename, "rb");
g_autofree char *data = NULL;
ssize_t file_size;
size_t data_size;
if (!f)
return NULL;
fseek(f, 0, SEEK_END);
file_size = ftell(f);
if (file_size < 0)
return NULL;
data_size = (size_t)file_size;
fseek(f, 0, SEEK_SET);
data = g_malloc0(data_size);
if (data_size != fread(data, 1, data_size, f))
return NULL;
return pv_sec_gbytes_new_take(g_steal_pointer(&data), data_size);
}
void *pv_gbytes_memcpy(void *dst, size_t dst_size, GBytes *src)
{
size_t src_size;
const void *src_data = g_bytes_get_data(src, &src_size);
if (dst_size < src_size)
return NULL;
return memcpy(dst, src_data, src_size);
}

147
libpv/hash.c Normal file
View File

@@ -0,0 +1,147 @@
/*
* Hashing functions.
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
/* Must be included before any other header */
#include "config.h"
#include "libpv/crypto.h"
#include "libpv/hash.h"
GBytes *pv_sha256_hash(uint8_t *buf, size_t size, GError **error)
{
g_autoptr(EVP_MD_CTX) ctx = NULL;
ctx = pv_digest_ctx_new(EVP_sha256(), error);
if (!ctx)
return NULL;
if (pv_digest_ctx_update_raw(ctx, buf, size, error) != 0)
return NULL;
return pv_digest_ctx_finalize(ctx, error);
}
EVP_MD_CTX *pv_digest_ctx_new(const EVP_MD *md, GError **error)
{
g_autoptr(EVP_MD_CTX) ctx = EVP_MD_CTX_new();
if (!ctx) {
g_set_error(error, PV_HASH_ERROR, PV_HASH_ERROR_INTERNAL,
_("Hash context generation failed"));
return NULL;
}
if (EVP_DigestInit_ex(ctx, md, NULL) != 1) {
g_set_error(error, PV_HASH_ERROR, PV_HASH_ERROR_INTERNAL,
_("EVP_DigestInit_ex failed"));
return NULL;
}
return g_steal_pointer(&ctx);
}
int pv_digest_ctx_update_raw(EVP_MD_CTX *ctx, const uint8_t *buf, size_t size, GError **error)
{
if (!buf || size == 0)
return 0;
if (EVP_DigestUpdate(ctx, buf, size) != 1) {
g_set_error(error, PV_HASH_ERROR, PV_HASH_ERROR_INTERNAL,
_("EVP_DigestUpdate failed"));
return -1;
}
return 0;
}
int pv_digest_ctx_update(EVP_MD_CTX *ctx, GBytes *data, GError **error)
{
const uint8_t *buf;
size_t buf_size;
if (!data)
return 0;
buf = g_bytes_get_data((GBytes *)data, &buf_size);
return pv_digest_ctx_update_raw(ctx, buf, buf_size, error);
}
GBytes *pv_digest_ctx_finalize(EVP_MD_CTX *ctx, GError **error)
{
int md_size = EVP_MD_size(EVP_MD_CTX_md(ctx));
g_autofree uint8_t *digest = NULL;
unsigned int digest_size;
g_assert(md_size > 0);
digest = g_malloc0((uint)md_size);
if (EVP_DigestFinal_ex(ctx, digest, &digest_size) != 1) {
g_set_error(error, PV_HASH_ERROR, PV_HASH_ERROR_INTERNAL,
_("EVP_DigestFinal_ex failed"));
return NULL;
}
g_assert(digest_size == (uint)md_size);
return g_bytes_new_take(g_steal_pointer(&digest), digest_size);
}
HMAC_CTX *pv_hmac_ctx_new(GBytes *key, const EVP_MD *md, GError **error)
{
g_autoptr(HMAC_CTX) ctx = HMAC_CTX_new();
const uint8_t *key_data;
size_t key_size;
key_data = g_bytes_get_data(key, &key_size);
if (HMAC_Init_ex(ctx, key_data, (int)key_size, md, NULL) != 1) {
g_set_error(error, PV_HASH_ERROR, PV_HASH_ERROR_INTERNAL,
"unable to create HMAC context: %s", pv_get_openssl_error());
return NULL;
}
return g_steal_pointer(&ctx);
}
int pv_hmac_ctx_update_raw(HMAC_CTX *ctx, const void *buf, size_t size, GError **error)
{
if (!buf || size == 0)
return 0;
if (HMAC_Update(ctx, buf, size) != 1) {
g_set_error(error, PV_HASH_ERROR, PV_HASH_ERROR_INTERNAL,
"unable to add data to HMAC context: %s", pv_get_openssl_error());
return -1;
}
return 0;
}
int pv_hmac_ctx_update(HMAC_CTX *ctx, GBytes *data, GError **error)
{
const uint8_t *buf;
size_t buf_size;
if (!data)
return 0;
buf = g_bytes_get_data((GBytes *)data, &buf_size);
return pv_hmac_ctx_update_raw(ctx, buf, buf_size, error);
}
GBytes *pv_hamc_ctx_finalize(HMAC_CTX *ctx, GError **error)
{
int md_size = EVP_MD_size(HMAC_CTX_get_md(ctx));
g_autofree uint8_t *hmac = NULL;
unsigned int hmac_size = 0;
g_assert(md_size > 0);
hmac = g_malloc0((unsigned int)md_size);
if (HMAC_Final(ctx, hmac, &hmac_size) != 1) {
g_set_error(error, PV_HASH_ERROR, PV_HASH_ERROR_INTERNAL,
"unable to calculate HMAC: %s", pv_get_openssl_error());
return NULL;
}
return g_bytes_new_take(g_steal_pointer(&hmac), hmac_size);
}

File diff suppressed because it is too large Load Diff

View File

@@ -93,6 +93,8 @@ const char *util_arch_machine_type_to_str(int type)
return "IBM z15";
case UTIL_ARCH_MACHINE_TYPE_Z15_T02:
return "IBM z15 Model T02";
case UTIL_ARCH_MACHINE_TYPE_Z16:
return "IBM z16";
default:
return "Unknown machine type";
}

View File

@@ -9,6 +9,7 @@
* it under the terms of the MIT license. See LICENSE for details.
*/
#include <ctype.h>
#include <errno.h>
#include <limits.h>
#include <linux/limits.h>
@@ -21,11 +22,16 @@
#include <unistd.h>
#include "lib/util_base.h"
#include "lib/util_exit_code.h"
#include "lib/util_file.h"
#include "lib/util_libc.h"
#include "lib/util_panic.h"
#include "lib/util_prg.h"
#define READ_CHUNK_SIZE 4096
extern const char *toolname;
/*
* Read the first line of a file into given buffer
*/
@@ -536,3 +542,129 @@ int util_file_read_va(const char *path, const char *fmt, ...)
return -1;
return ret;
}
/**
* Print an error message indicating an out-of-memory situation and exit.
*/
static void oom(void)
{
fprintf(stderr, "Out of memory\n");
/* We can't rely on our clean-up routines to work reliably during an
* OOM situation, so just exit here.
*/
exit(UTIL_EXIT_OUT_OF_MEMORY);
}
/**
* Read all data from @fd and return address of resulting buffer in
* @buffer_ptr. If @size_ptr is non-zero, use it to store the size of the
* resulting buffer. Return %UTIL_EXIT_OK on success.
*
* @param[in] fd File descriptor to read data from
* @param[in, out] buffer_ptr Buffer to read data into
* @param[in, out] size_ptr Buffer to save size of data read into buffer_ptr
*
* @retval 0 read was successful
* @retval UTIL_EXIT_RUNTIME_ERROR error while reading file
*/
util_exit_code_t util_file_read_fd_buf(FILE *fd, void **buffer_ptr,
size_t *size_ptr)
{
char *buffer = NULL;
size_t done = 0;
while (!feof(fd)) {
buffer = realloc(buffer, done + READ_CHUNK_SIZE);
if (!buffer)
oom();
done += fread(&buffer[done], 1, READ_CHUNK_SIZE, fd);
if (ferror(fd)) {
free(buffer);
return UTIL_EXIT_RUNTIME_ERROR;
}
}
buffer = realloc(buffer, done);
if (!buffer && done > 0)
oom();
*buffer_ptr = buffer;
if (size_ptr)
*size_ptr = done;
return UTIL_EXIT_OK;
}
/**
* Read text from @fd and return resulting NULL-terminated text buffer.
* If @chomp is non-zero, remove trailing newline character. Return %NULL
* on error or when unprintable characters are read.
*
* @param[in] fd File descriptor to read data from
* @param[in] chomp Flag to indicate trailing newlines should be removed
*
* @retval NULL Error reading from fd
* @retval != 0 Data read successfully, buffer returned
*/
char *util_file_read_fd(FILE *fd, int chomp)
{
char *buffer;
size_t done, i;
if (util_file_read_fd_buf(fd, (void **) &buffer, &done))
return NULL;
/* Check if this is a text file at all (required to filter out
* binary sysfs attributes).
*/
for (i = 0; i < done; i++) {
if (!isgraph(buffer[i]) && !isspace(buffer[i])) {
free(buffer);
return NULL;
}
}
/* Remove trailing new-line character if requested. */
if (chomp && done > 0 && buffer[done - 1] == '\n')
done--;
/* NULL-terminate. */
buffer = realloc(buffer, done + 1);
if (!buffer)
oom();
buffer[done] = 0;
return buffer;
}
/**
* Read file as text and return NULL-terminated contents. Remove trailing
* newline if CHOMP is specified.
*
* @param[in] path Path to the file that will be read from
* @param[in] chomp Flag to indicate trailing newlines should be removed
*
* @retval NULL Error reading from file
* @retval != 0 File read successfully, contents returned in buffer
*/
char *util_file_read_text_file(const char *path, int chomp)
{
char *buffer = NULL;
FILE *fd;
fd = fopen(path, "r");
if (!fd)
goto out;
buffer = util_file_read_fd(fd, chomp);
fclose(fd);
out:
if (!buffer) {
fprintf(stderr, "Could not read file %s: %s\n", path,
strerror(errno));
}
return buffer;
}

View File

@@ -128,7 +128,7 @@ static int mbr_table_ext_search(int fh, size_t blk_start_mbr,
start = blk_start_mbr + le32toh(mbr.part_entry_vec[0].blk_start);
cnt = le32toh(mbr.part_entry_vec[0].blk_cnt);
if ((start == blk_start) && (cnt == blk_cnt))
if ((start <= blk_start) && (cnt >= (blk_start - start) + blk_cnt))
return part_num;
/* Second entry contains relative offset for next logical volume */
@@ -162,11 +162,11 @@ static int mbr_table_search(int fh, struct mbr *mbr, size_t blk_start,
* The kernel sets count for extended partitions explicitly.
* Therefore we do not check count here.
*/
if (mbr_part_is_ext(type) && (start == blk_start)) {
if (mbr_part_is_ext(type) && (start <= blk_start)) {
*part_ext = 1;
return part_num;
}
if ((start == blk_start) && (cnt == blk_cnt))
if ((start <= blk_start) && (cnt >= (blk_start - start) + blk_cnt))
return part_num;
if (!mbr_part_is_ext(type))
continue;
@@ -207,7 +207,7 @@ static int gpt_table_search(int fh, struct gpt *gpt, size_t blk_start,
end = le64toh(part_entry->blk_end);
if (start == 0) /* Empty slot */
continue;
if ((start == blk_start) && (end == blk_end))
if ((start <= blk_start) && (end >= blk_end))
return part_num;
}
return 0;

362
libutil/util_udev.c Normal file
View File

@@ -0,0 +1,362 @@
/*
* util - Utility function library
*
* UDEV helper functions
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#include <ctype.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include "lib/util_exit_code.h"
#include "lib/util_file.h"
#include "lib/util_libc.h"
#include "lib/util_list.h"
#include "lib/util_path.h"
#include "lib/util_udev.h"
/* Create a newly allocated udev entry. */
static struct util_udev_entry_node *util_udev_entry_node_new(const char *key,
const char *op,
const char *value)
{
struct util_udev_entry_node *entry;
entry = util_zalloc(sizeof(struct util_udev_entry_node));
entry->key = util_strdup(key);
entry->op = util_strdup(op);
entry->value = util_strdup(value);
return entry;
}
/* Release resources associated with udev entry. */
static void util_udev_entry_node_free(struct util_udev_entry_node *entry)
{
if (!entry)
return;
free(entry->key);
free(entry->op);
free(entry->value);
free(entry);
}
/* Create a newly allocated udev line. */
static struct util_udev_line_node *util_udev_line_node_new(void)
{
struct util_udev_line_node *line;
line = util_zalloc(sizeof(struct util_udev_line_node));
util_list_init(&line->entries, struct util_udev_entry_node, node);
return line;
}
/* Release resources associated with udev line. */
static void util_udev_line_node_free(struct util_udev_line_node *line)
{
struct util_udev_entry_node *e, *n;
if (!line)
return;
util_list_iterate_safe(&line->entries, e, n) {
util_list_remove(&line->entries, e);
util_udev_entry_node_free(e);
}
free(line->line);
free(line);
}
/* Create a newly allocated udev file. */
static struct util_udev_file *util_udev_file_new(void)
{
struct util_udev_file *file;
file = util_zalloc(sizeof(struct util_udev_file));
util_list_init(&file->lines, struct util_udev_line_node, node);
return file;
}
/**
* Release resources associated with udev file.
*
* @param[in, out] file Udev file structure to be freed
*/
void util_udev_free_file(struct util_udev_file *file)
{
struct util_udev_line_node *l, *n;
if (!file)
return;
util_list_iterate_safe(&file->lines, l, n) {
util_list_remove(&file->lines, l);
util_udev_line_node_free(l);
}
free(file);
}
/**
* Print the contents of a udev file to stdout. Used for debugging.
*
* @param[in] file Udev file structure to print
*/
void util_udev_file_print(struct util_udev_file *file)
{
struct util_udev_line_node *l;
struct util_udev_entry_node *e;
printf("util_udev_file at %p\n", (void *) file);
if (!file)
return;
util_list_iterate(&file->lines, l) {
printf(" util_udev_line_node at %p\n", (void *) l);
printf(" line='%s'\n", l->line);
util_list_iterate(&l->entries, e) {
printf(" util_udev_entry at %p\n", (void *) e);
printf(" '%s' '%s' '%s'\n", e->key, e->op,
e->value);
}
}
}
static void skip_whitespace(const char **s_ptr)
{
const char *s = *s_ptr;
while (*s && isspace(*s))
s++;
*s_ptr = s;
}
static char *parse_key(const char **s_ptr)
{
const char *s, *e;
char *key;
s = *s_ptr;
/* Parse \w+(\{[^\}]*\})? */
e = s;
while (*e && (isalnum(*e) || *e == '_'))
e++;
if (*e == '{') {
while (*e && *e != '}')
e++;
if (*e == '}')
e++;
}
if (e == s)
return NULL;
/* s points to key start, e to character after key end. */
key = util_zalloc(e - s + 1);
memcpy(key, s, e - s);
*s_ptr = e;
return key;
}
static char *parse_op(const char **s_ptr)
{
const char *ops[] = { "==", "!=", "=", "+=", ":=", NULL };
const char *entry;
size_t len;
int i;
entry = *s_ptr;
for (i = 0; ops[i]; i++) {
len = strlen(ops[i]);
if (strncmp(entry, ops[i], len) == 0) {
*s_ptr += len;
return util_strdup(ops[i]);
}
}
return NULL;
}
static char *parse_value(const char **s_ptr)
{
const char *s, *e;
char *value;
/* Parse: ^\s*(.*)\s*$ */
s = *s_ptr;
skip_whitespace(&s);
e = s;
while (*e)
e++;
e--;
while (e > s && isspace(*e))
e--;
e++;
*s_ptr = e;
/* Remove quotes. */
if ((*s == '"' && *(e - 1) == '"') ||
(*s == '\'' && *(e - 1) == '\'')) {
s++;
e--;
}
/* s points to value start, e to character after value end. */
value = util_zalloc(e - s + 1);
memcpy(value, s, e - s);
return value;
}
static bool parse_util_udev_entry(struct util_udev_line_node *line,
const char *entry)
{
char *key = NULL, *op = NULL, *value = NULL;
struct util_udev_entry_node *e;
bool rc = false;
/* Parse: ^\s*(\w+)\s*(==|!=|=|\+=|:=)\s*"?([^"]*)"\s*$ */
/* Parse key. */
skip_whitespace(&entry);
key = parse_key(&entry);
if (!key)
goto out;
/* Parse operator. */
skip_whitespace(&entry);
op = parse_op(&entry);
if (!op)
goto out;
/* Parse value. */
skip_whitespace(&entry);
value = parse_value(&entry);
if (!value)
goto out;
skip_whitespace(&entry);
/* Check for unrecognized characters at end of entry. */
if (*entry != 0)
goto out;
/* Add entry to list. */
e = util_udev_entry_node_new(key, op, value);
util_list_add_tail(&line->entries, e);
rc = true;
out:
free(key);
free(op);
free(value);
return rc;
}
static void replace_unquoted(char *s, char from, char to)
{
char quoted = 0;
for (; *s; s++) {
if (quoted) {
/* Skip until quote end is found. */
if (*s == quoted)
quoted = 0;
continue;
}
if (*s == '"' || *s == '\'') {
quoted = *s;
continue;
}
if (*s == from)
*s = to;
}
}
static bool parse_util_udev_line(struct util_udev_file *file, const char *line)
{
char *copy, *curr, *next;
struct util_udev_line_node *l;
int i;
bool result = true;
l = util_udev_line_node_new();
l->line = util_strdup(line);
/* Check for empty lines and comment lines. */
for (i = 0; line[i] && isspace(line[i]); i++);
if (line[i] == 0 || line[i] == '#')
goto ok;
/* Parse each comma-separated entry. */
copy = util_strdup(line);
/* A hack to differentiate between quoted and unquoted commas. */
replace_unquoted(copy, ',', 1);
next = copy;
while ((curr = strsep(&next, "\1"))) {
if (!parse_util_udev_entry(l, curr)) {
result = false;
break;
}
}
free(copy);
ok:
if (result)
util_list_add_tail(&file->lines, l);
else
util_udev_line_node_free(l);
return result;
}
/**
* Create a new util_udev_file structure and read the contents of a specified
* udev file into that structure.
*
* @param[in] path Path to the udev file that will be read in
* @param[in, out] file_ptr A buffer to store resulting udev file structure
*
* @retval 0 Udev file read successfully
* @retval UTIL_EXIT_RUNTIME_ERROR Error reading the udev file
*/
util_exit_code_t util_udev_read_file(const char *path,
struct util_udev_file **file_ptr)
{
char *text, *curr, *next;
struct util_udev_file *file;
int once = 0;
text = util_file_read_text_file(path, 0);
if (!text)
return UTIL_EXIT_RUNTIME_ERROR;
file = util_udev_file_new();
/* Iterate over each line. */
next = text;
while ((curr = strsep(&next, "\n"))) {
if (parse_util_udev_line(file, curr))
continue;
if (!once) {
once = 1;
fprintf(stderr, "Unrecognized udev rule in %s:\n",
path);
}
fprintf(stderr, "%s\n", curr);
}
free(text);
*file_ptr = file;
return UTIL_EXIT_OK;
}

View File

@@ -1,3 +1,12 @@
/*
* lsstp - Display STP system information
*
* Copyright IBM Corp. 2020, 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#include <stdio.h>
#include <stdlib.h>
#include <stdarg.h>

22
pvattest/Makefile Normal file
View File

@@ -0,0 +1,22 @@
# Common definitions
include ../common.mak
.DEFAULT_GOAL := all
PKGDATADIR := "$(DESTDIR)$(TOOLS_DATADIR)/pvattest"
SUBDIRS := src man tools
RECURSIVE_TARGETS := all-recursive clean-recursive install-recursive
all: all-recursive
install: all install-recursive
clean: clean-recursive
$(RECURSIVE_TARGETS):
@target=`echo $@ |sed s/-recursive//`; \
for d in $(SUBDIRS); do \
$(MAKE) -C $$d $$target || exit 1; \
done
.PHONY: all install clean $(RECURSIVE_TARGETS)

100
pvattest/README.md Normal file
View File

@@ -0,0 +1,100 @@
# pvattest
Use `pvattest` to attest an IBM Secure Execution guest running on z16 and later.
With `pvattest` you can create attestation requests in a trusted environment and attest
an IBM Secure Execution for Linux guest to verify that a provider is running the correct image.
To achieve this, use the following commands:
* `create` On a trusted system, creates an attestation request.
* `perform` Performs an attestation measurement on the SE-guest to be attested. For this a
attestation request is sent to the Ultravisor (UV) and the answer received. The `perform`
command requires IBM z16 or later z/Architecture hardware.
* `verify` On a trusted system, compares the answer from the Ultravisor to the
expected answer. If they differ, the Secure Execution guest might be a different guest
than expected, or not secure at all.
For meaningful results, run `create` and `verify` only in a trusted environment,
like your workstation or a previously attested IBM Secure Execution guest.
Otherwise, the attestation can be compromised.
For all certificates, revocation lists, and host-key documents, both the PEM and DER input
formats are supported. If you run this program on a non S390 System, 'perform' is not be available.
## Getting started
If all dependencies are met (see the s390-tools README) issue `make` in the source tree to build `pvattest`.
## Details
### create
`pvattest create` needs the host-key-document, a location to store the
attestation request protection key, and a location to store the request data.
Unless the `--no-verify` flag is set it additionally requires the IBM signing key
and the intermediate CA. The output contains the request in binary form which serves as input
to `pvattest perform`. Must be run in a trusted environment. Especially, do not create the request
on a system you want to attest. The attestation request protection key is valid for this request only,
must be kept until the verification is completed and must be destroyed afterwards
Keep the key secret.
### perform
`pvattest perform` needs a request in binary form generated by `pvattest create`and
a location to store the output. It will send the request to the device at `/dev/uv`
which passes the request to the Ultravisor.
Kernel will then send the request to the Ultravisor which will calculate the answer.
The Answer is then passed back to userspace and handled by `pvattest`
The output includes the original request and the answer from the Ultravisor.
### verify
`pvattest verify` needs the SE-guest header, the attestation request protection key,
and the attestation request and the response to the `pvattest perform` command from the Ultravisor.
It calculates the measurement in the trusted environment and compares it to the response from
the Ultravisor in the previous step.
The following return codes are possible:
0. successful verification: The calculated measurement matches the response from the Ultravisor
1. failed verification: The command ended with an error, for example, because of incorrect input or an invalid SE header
2. failed verification: The calculated measurement does not match the response from the Ultravisor
Run `pvattest verify` in a trusted environment. Especially, do not verify on the system you want to attest.
## Measurement
The measurement is a cryptographic measurement of the following block.
Only HMAC-SHA512 is supported.
| Start | Size | Content |
|---------|------------|---------------------------------------------------------------|
| 0x0 | 0x40 | Page List Digest (from SE header) |
| 0x40 | 0x40 | Address List Digest (from SE header) |
| 0x80 | 0x40 | Tweak List Digest (from SE header) |
| 0xc0 | 0x10 | SE Header Tag (from SE header) |
| 0xd0 | 0x10 | Configuration UID (generated by UV, included in the answer) |
| 0xe0 | 0x02 | User Data Length (defined during measurement on the SE-guest) |
| 0xe2 | 0x02 | Zeros |
| 0xe4 | 0x04 | Additional Data Length (set by UV, included in the answer) |
| 0xe8 | 0 - 0x100 | User Data (generated during measurement on the SE-guest) |
| ... | 0 or 0x10 | Optional Nonce (generated during request creation) |
| ... | 0 - 0x8000 | Additional Data (generated by UV, included in the answer) |
### User Data
By default `pvattest` does not include any User Data, therefore the length is zero.
`User Data` is data generated by the SE guest and passed to UV during the measurement.
The `User Data` must be known to or be replicable by the verifier to verify the correctness of the User Data.
The addition of user data is currently an experimental setting.
### Additional Data
`Additional data` is data known to the Ultravisor. By default UV will not include any `Additional Data`.
Adding `Additional Data` is currently an experimental setting.
## Example
Create an attestation request in a trusted environment:
`pvattest create -k hkd.crt --arpk arp.key -o arcb.bin --cert IntermediateCA.crt --cert IbmSigningKey.crt`
Perform an attestation measurement on an IBM Secure Execution guest:
`pvattest perform --input arcb.bin --output measurement.bin`
Verify the response from the Ultravisor against the attestation request in a trusted environment:
`pvattest verify --input measurement.bin --arpk arp.key --hdr se_guest.hdr`

9
pvattest/man/Makefile Normal file
View File

@@ -0,0 +1,9 @@
include ../../common.mak
all:
install:
$(INSTALL) -d -m 755 $(DESTDIR)$(MANDIR)/man1
$(INSTALL) -m 644 -c *.1 -t $(DESTDIR)$(MANDIR)/man1
.PHONY: all install clean

View File

@@ -0,0 +1,73 @@
.\" Copyright 2022 IBM Corp.
.\" s390-tools is free software; you can redistribute it and/or modify
.\" it under the terms of the MIT license. See LICENSE for details.
.\"
.TH pvattest-create 1 "07 June 2022" "s390-tools" "Attestation Manual"
.nh
.ad l
.SH NAME
\fBpvattest [OPTION?] create [OPTIONS] \fP- create an attestation measurement request
\fB
.SH DESCRIPTION
Prepare attestation measurement requests for an IBM Secure Execution guest. Only prepare attestation requests in a trusted environment, such as your workstation. The 'pvattest create' command creates a randomly generated key to protect the attestation request. This key is only valid for this specific request.In order to avoid compromising the attestation, do not publish the protection key and delete it after verification. Every 'create' command generates a new, random protection key.
.SH OPTIONS
.TP
.B
\fB-h\fP, \fB--help\fP
Show help options
.TP
.B
\fB-k\fP, \fB--host-key-document\fP=\fBFILE\fP
Specify one or more host key documents.
.TP
.B
\fB-C\fP, \fB--cert\fP=\fBFILE\fP
Specifies the certificate that is used to establish a chain of trust for the verification of the host-key documents. Specify this option twice to specify the IBM Z signing key and the intermediate CA certificate (signed by the root CA). Required. Ignored when \fB--no-verify\fP is specified.
.TP
.B
\fB--crl\fP=\fBFILE\fP
Specify \fBFILE\fP to be a certificate revocation list (optional).
.TP
.B
\fB--root-ca\fP=\fBFILE\fP
Use \fBFILE\fP as the trusted root CA instead the root CAs that are installed on the system (optional).
.TP
.B
\fB-o\fP, \fB--output\fP=\fBFILE\fP
\fBFILE\fP specifies the output for the attestation request control block.
.TP
.B
\fB-a\fP, \fB--arpk\fP=\fBFILE\fP
Save the protection key as GCM-AES256 key in \fBFILE\fP Do not publish this key, otherwise your attestation is compromised.
.TP
.B
\fB--no-verify\fP
Disable the host-key-document verification. Does not require the host-key documents to be valid. For testing purposes, do not use for a production image. (Optional)
.TP
.B
\fB--offline\fP
Don't download CRLs (optional).
.TP
.B
\fB-V\fP, \fB--verbose\fP
Provide more detailed output (optional)
.SH EXAMPLE
Create an attestation request with the protection key 'arp.key', write the request to 'arcb.bin', and verify the host-key document using the CA-signed key 'DigiCertCA.crt' and the intermediate key 'IbmSigningKey.crt'.
.PP
.nf
.fam C
pvattest create -k hkd.crt --arpk arp.key -o attreq.bin --cert DigiCertCA.crt --cert IbmSigningKey.crt
.fam T
.fi
Create an attestation request with the protection key 'arp.key', write the request to 'arcb.bin', verify the host-key document using the CA-signed key 'DigiCertCA.crt' and the intermediate key 'IbmSigningKey.crt', and instead of downloading the certificate revocation list use certificate revocation lists 'DigiCertCA.crl', 'IbmSigningKey.crl', and 'rootCA.crl'.
.PP
.nf
.fam C
pvattest create -k hkd.crt --arpk arp.key -o attreq.bin --cert DigiCertCA.crt --cert IbmSigningKey.crt --offline --crl DigiCertCA.crl --crl IbmSigningKey.crl --crl rootCA.crl
.fam T
.fi
.SH SEE ALSO
\fBpvattest\fP(1), \fBpvattest-verify\fP(1), \fBpvattest-perform\fP(1)

View File

@@ -0,0 +1,47 @@
.\" Copyright 2022 IBM Corp.
.\" s390-tools is free software; you can redistribute it and/or modify
.\" it under the terms of the MIT license. See LICENSE for details.
.\"
.TH pvattest-perform 1 "07 June 2022" "s390-tools" "Attestation Manual"
.nh
.ad l
.SH NAME
\fBpvattest [OPTION?] perform [OPTIONS] \fP- execute an attestation measurement request
\fB
.SH DESCRIPTION
Run a measurement of this system using '/dev/uv'. Works only if this device is available and the attestation Ultravisor facility is present. The input must be an attestation request created with 'pvattest create'. Output will contain the original request, the attestation measurement result, the configuration UID, and if requested in the request Additional Data.
.RE
.PP
.SH OPTIONS
.TP
.B
\fB-h\fP, \fB--help\fP
Show help options
.TP
.B
\fB-i\fP, \fB--input\fP=\fBFILE\fP
\fBFILE\fP specifies the attestation request as input.
.TP
.B
\fB-o\fP, \fB--output\fP=\fBFILE\fP
\fBFILE\fP specifies the output for the attestation result.
.TP
.B
\fB-V\fP, \fB--verbose\fP
Provide more detailed output (optional)
.RE
.PP
.SH EXAMPLE
Perform an attestation measurement with the attestation request 'arcb.bin' and write the output to 'measurement.bin'.
.PP
.nf
.fam C
pvattest perform --input attreq.bin --output attresp.bin
.fam T
.fi
.SH SEE ALSO
\fBpvattest\fP(1), \fBpvattest-create\fP(1), \fBpvattest-verify\fP(1)

View File

@@ -0,0 +1,60 @@
.\" Copyright 2022 IBM Corp.
.\" s390-tools is free software; you can redistribute it and/or modify
.\" it under the terms of the MIT license. See LICENSE for details.
.\"
.TH pvattest-verify 1 "07 June 2022" "s390-tools" "Attestation Manual"
.nh
.ad l
.SH NAME
\fBpvattest [OPTION?] verify [OPTIONS] \fP- verify an attestation measurement
\fB
.SH DESCRIPTION
Verify that a previously generated attestation measurement of an IBM Secure Execution guest is as expected. Only verify attestation requests in a trusted environment, such as your workstation. Input must contain the response as produced by 'pvattest perform'. The protection key must be the one that was used to create the request by 'pvattest create'. Please delete it after verification. The header must be the IBM Secure Execution header of the image that was attested during 'pvattest perform'
.RE
.PP
.SH OPTIONS
.TP
.B
\fB-h\fP, \fB--help\fP
Show help options
.TP
.B
\fB-i\fP, \fB--input\fP=\fBFILE\fP
\fBFILE\fP specifies the attestation result as input.
.TP
.B
\fB--hdr\fP=\fBFILE\fP
Specify the header of the guest image. Exactly one is required.
.TP
.B
\fB-a\fP, \fB--arpk\fP=\fBFILE\fP
Use \fBFILE\fP to specify the GCM-AES256 key to decrypt the attestation request. Delete this key after verification.
.TP
.B
\fB-V\fP, \fB--verbose\fP
Provide more detailed output (optional)
.RE
.PP
.SH EXAMPLE
To verify a measurement in 'measurement.bin' with the protection key 'arp.kep' and SE-guest header 'se_guest.hdr'.
.PP
.nf
.fam C
pvattest verify --input attresp.bin --arpk arp.key --hdr se_guest.hdr
.fam T
.fi
If the verification was successful the program exists with zero.
If the verification failed it exists with 2 and prints the following to stderr:
.PP
.nf
.fam C
ERROR: Attestation measurement verification failed:
Calculated and received attestation measurement are not the same.
.fam T
.fi
.SH SEE ALSO
\fBpvattest\fP(1), \fBpvattest-create\fP(1), \fBpvattest-perform\fP(1)

104
pvattest/man/pvattest.1 Normal file
View File

@@ -0,0 +1,104 @@
.\" Copyright 2022 IBM Corp.
.\" s390-tools is free software; you can redistribute it and/or modify
.\" it under the terms of the MIT license. See LICENSE for details.
.\"
.TH pvattest 1 "07 June 2022" "s390-tools" "Attestation Manual"
.nh
.ad l
.SH NAME
\fBpvattest [OPTION?] COMMAND [OPTIONS] \fP- create, perform, and verify attestation measurements
\fB
.RE
\fB
.SH SYNOPSIS
.nf
.fam C
\fBpvattest\fP \fIcreate\fP [\fIOPTIONS\fP]
\fBpvattest\fP \fIperform\fP [\fIOPTIONS\fP]
\fBpvattest\fP \fIverify\fP [\fIOPTIONS\fP]
.fam T
.fi
.fam T
.fi
.SH DESCRIPTION
Use \fBpvattest\fP to attest that an IBM Secure Execution guest is the correct guest, and that it was started in a secure manner.
Run '\fBpvattest\fP \fIcreate\fP' and '\fBpvattest\fP \fIverify\fP' in a trusted environment only.
.PP
.nf
.fam C
create On a trusted system, creates an attestation request.
perform On the SE-guest to be attested, sends the attestation request to the Ultravisor and receives the answer.
verify On a trusted system, compares the answer from the Ultravisor to the one from your trusted environment. If they differ, the Secure Execution guest might be compromised.
.fam T
.fi
For meaningful results, run '\fIcreate\fP' and '\fIverify\fP' in a trusted environment, like your workstation or a previously attested IBM Secure Execution guest. Otherwise, the attestation might be tampered with. For all certificates, revocation lists, and host-key documents, both the PEM and DER input formats are supported. If you run \fBpvattest\fP on a machine architecture other than z/Architecture, 'measure' is not available.
.PP
Use '\fBpvattest\fP [COMMAND] \fB-h\fP' to get detailed help
.RE
.PP
.SH OPTIONS
.TP
.B
\fB-h\fP, \fB--help\fP
Show help options
.TP
.B
\fB-v\fP, \fB--version\fP
Print the version and exit.
.TP
.B
\fB-V\fP, \fB--verbose\fP
Provide more detailed output (optional)
.RE
.PP
.SH EXAMPLE
For details refer to the man page of the command.
.PP
Create the request on a trusted system.
.PP
.nf
.fam C
trusted:~$ pvattest create -k hkd.crt --cert CA.crt --cert ibmsk.crt --arpk arp.key -o attreq.bin
.fam T
.fi
On the SE-guest, \fIperform\fP the attestation.
.PP
.nf
.fam C
seguest:~$ pvattest perform -i attreq.bin -o attresp.bin
.fam T
.fi
On a trusted system, \fIverify\fP that the response is correct. Here, the protection key from the creation and the SE-guests header is used to \fIverify\fP the measurement.
.PP
.nf
.fam C
trusted:~$ pvattest verify -i attresp.bin --arpk arp.key --hdr se_guest.hdr
trusted:~$ echo $?
0
.fam T
.fi
If the measurements do not match \fBpvattest\fP exits with code 2 and emits an error message. The SE-guest attestation failed.
.PP
.nf
.fam C
trusted:~$ pvattest verify -i wrongresp.bin --arpk arp.key --hdr se_guest.hdr
ERROR: Attestation measurement verification failed:
Calculated and received attestation measurement are not the same.
trusted:~$ echo $?
2
.fam T
.fi
.SH SEE ALSO
\fBpvattest\fP-\fIcreate\fP(1), \fBpvattest-\fIverify\fP\fP(1), \fBpvattest\fP-\fIperform\fP(1)

2
pvattest/src/.gitignore vendored Normal file
View File

@@ -0,0 +1,2 @@
.check-dep-pvattest
.detect-openssl.dep.c

124
pvattest/src/Makefile Normal file
View File

@@ -0,0 +1,124 @@
include ../../common.mak
BIN_PROGRAM = pvattest
PKGDATADIR ?= "$(DESTDIR)$(TOOLS_DATADIR)/$(BIN_PROGRAM)"
SRC_DIR := $(dir $(realpath $(firstword $(MAKEFILE_LIST))))
ROOT_DIR = $(rootdir)
PVATTESTDIR := $(ROOT_DIR)/pvattest
INCLUDE_PATHS = "$(SRC_DIR)" "$(ROOT_DIR)/include"
INCLUDE_PARMS = $(addprefix -I,$(INCLUDE_PATHS))
LIBPV_DIR = $(ROOT_DIR)/libpv
LIBPV = $(LIBPV_DIR)/libpv.a
WARNINGS := -Wall -Wextra -Wshadow \
-Wcast-align -Wwrite-strings -Wmissing-prototypes \
-Wmissing-declarations -Wredundant-decls -Wnested-externs \
-Wno-long-long -Wuninitialized -Wconversion -Wstrict-prototypes \
-Wpointer-arith -Wno-error=inline \
-Wno-unused-function -Wno-unused-parameter -Wno-unused-variable \
-Werror \
$(NULL)
PVATTEST_SRCS := $(wildcard *.c) \
$(NULL)
$(BIN_PROGRAM)_SRCS := \
$(PVATTEST_SRCS) \
$(NULL)
$(BIN_PROGRAM)_OBJS := $($(BIN_PROGRAM)_SRCS:.c=.o)
ifneq ($(shell sh -c 'command -v pkg-config'),)
GLIB2_CFLAGS := $(shell pkg-config --silence-errors --cflags glib-2.0)
GLIB2_LIBS := $(shell pkg-config --silence-errors --libs glib-2.0)
LIBCRYPTO_CFLAGS := $(shell pkg-config --silence-errors --cflags libcrypto openssl)
LIBCRYPTO_LIBS := $(shell pkg-config --silence-errors --libs libcrypto openssl)
LIBCURL_CFLAGS := $(shell pkg-config --silence-errors --cflags libcurl)
LIBCURL_LIBS := $(shell pkg-config --silence-errors --libs libcurl)
else
GLIB2_CFLAGS := -I/usr/include/glib-2.0 -I/usr/lib64/glib-2.0/include
GLIB2_LIBS := -lglib-2.0
LIBCRYPTO_CFLAGS :=
LIBCRYPTO_LIBS := -lcrypto -lssl
LIBCURL_CFLAGS := -I/usr/include/s390x-linux-gnu
LIBCURL_LIBS := -lcurl
endif
ALL_CFLAGS += -std=gnu11 \
-DPKGDATADIR=$(PKGDATADIR) \
-DOPENSSL_API_COMPAT=0x10101000L \
$(GLIB2_CFLAGS) \
$(LIBCRYPTO_CFLAGS) \
$(LIBCURL_CFLAGS) \
$(WARNINGS) \
$(NULL)
ifneq ($(call check_header_prereq,"asm/uvdevice.h"),yes)
ALL_CFLAGS += -DPVATTEST_NO_PERFORM
endif
ALL_CPPFLAGS += $(INCLUDE_PARMS)
LDLIBS += $(GLIB2_LIBS) $(LIBCRYPTO_LIBS) $(LIBCURL_LIBS)
BUILD_TARGETS := skip-$(BIN_PROGRAM)
INSTALL_TARGETS := skip-$(BIN_PROGRAM)
ifneq (${HAVE_OPENSSL},0)
ifneq (${HAVE_GLIB2},0)
ifneq (${HAVE_LIBCURL}, 0)
BUILD_TARGETS := $(BIN_PROGRAM)
INSTALL_TARGETS := install-$(BIN_PROGRAM)
endif
endif
endif
all: $(BUILD_TARGETS) .check-dep-$(BIN_PROGRAM)
install: $(INSTALL_TARGETS)
$(BIN_PROGRAM): $($(BIN_PROGRAM)_OBJS) $(LIBPV)
skip-$(BIN_PROGRAM):
echo " SKIP $(BIN_PROGRAM) due to unresolved dependencies"
clean:
$(RM) -f -- $($(BIN_PROGRAM)_OBJS) $(BIN_PROGRAM) .check-dep-$(BIN_PROGRAM) .detect-openssl.dep.c
install-$(BIN_PROGRAM): $(BIN_PROGRAM)
$(INSTALL) -d -m 755 $(DESTDIR)$(USRBINDIR)
$(INSTALL) -c $^ $(DESTDIR)$(USRBINDIR)
.PHONY: all install clean skip-$(BIN_PROGRAM) install-$(BIN_PROGRAM)
$($(BIN_PROGRAM)_OBJS): .check-dep-$(BIN_PROGRAM)
.detect-openssl.dep.c:
echo "#include <openssl/evp.h>" > $@
echo "#if OPENSSL_VERSION_NUMBER < 0x10101000L" >> $@
echo " #error openssl version 1.1.0 is required" >> $@
echo "#endif" >> $@
echo "static void __attribute__((unused)) test(void) {" >> $@
echo " EVP_MD_CTX *ctx = EVP_MD_CTX_new();" >> $@
echo " EVP_MD_CTX_free(ctx);" >> $@
echo "}" >> $@
.check-dep-$(BIN_PROGRAM): .detect-openssl.dep.c
$(call check_dep, \
"$(BIN_PROGRAM)", \
"glib.h", \
"glib2-devel / libglib2.0-dev", \
"HAVE_GLIB2=0")
$(call check_dep, \
"$(BIN_PROGRAM)", \
"openssl/evp.h", \
"openssl-devel / libssl-dev version >= 1.1.0", \
"HAVE_OPENSSL=0", \
"-I.")
$(call check_dep, \
"$(BIN_PROGRAM)", \
"curl/curl.h", \
"libcurl-devel", \
"HAVE_LIBCURL=0")
touch $@

423
pvattest/src/arcb.c Normal file
View File

@@ -0,0 +1,423 @@
/*
* Attestation Request Control Block related functions
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
/* Must be included before any other header */
#include "config.h"
#include <openssl/evp.h>
#include <stdlib.h>
#include "libpv/crypto.h"
#include "libpv/hash.h"
#include "arcb.h"
#include "common.h"
#include "log.h"
#define ARVN_VERSION_1 0x0100
#define MAX_ARL 0x2000
typedef struct arcb_v1_hdr {
uint64_t reserved0; /* 0x0000 */
be32_t arvn; /* 0x0008 */
be32_t arl; /* 0x000c */
uint8_t iv[ARCB_V1_IV_SIZE]; /* 0x0010 */
uint32_t reserved1c; /* 0x001c */
uint8_t reserved20[7]; /* 0x0020 */
uint8_t nks; /* 0x0027 */
uint32_t reserved28; /* 0x0028 */
be32_t sea; /* 0x002c */
be64_t paf; /* 0x0030 */
be32_t mai; /* 0x0038 */
uint32_t reserved3c; /* 0x003c */
PvEcdhPubKey cpk; /* 0x0040 */
} __packed arcb_v1_hdr_t;
G_STATIC_ASSERT(sizeof(arcb_v1_hdr_t) == 0xe0);
typedef struct arcb_v1_key_slot {
uint8_t phkh[ARCB_V1_PHKH_SIZE];
uint8_t warpk[ARCB_V1_ATTEST_PROT_KEY_SIZE];
uint8_t kst[ARCB_V1_TAG_SIZE];
} __packed arcb_v1_key_slot_t;
G_STATIC_ASSERT(sizeof(arcb_v1_key_slot_t) == 0x50);
struct arcb_v1 {
/* authenticated data */
uint32_t arvn;
uint32_t mai;
uint64_t paf;
GBytes *iv;
EVP_PKEY *evp_cust_pub_key;
GSList *host_key_slots;
/* confidential data */
GBytes *confidential_measurement_key;
GBytes *confidential_optional_nonce;
GBytes *confidential_att_req_prot_key;
};
void arcb_v1_clear_free(arcb_v1_t *arcb)
{
if (!arcb)
return;
g_slist_free_full(arcb->host_key_slots, g_free);
g_bytes_unref(arcb->confidential_measurement_key);
g_bytes_unref(arcb->confidential_optional_nonce);
g_bytes_unref(arcb->confidential_att_req_prot_key);
g_bytes_unref(arcb->iv);
EVP_PKEY_free(arcb->evp_cust_pub_key);
g_free(arcb);
}
static void arcb_v1_set_paf(arcb_v1_t *arcb, const uint64_t paf, GError **error)
{
const uint64_t known_flags = ARCB_V1_PAF_ALL & ~ARCB_V1_PAF_NONCE;
if ((paf & ARCB_V1_PAF_NONCE) != 0) {
g_set_error(error, ARCB_ERROR, ARCB_ERR_INVALID_PAF,
_("The given paf (%#.16lx) specifies the NONCE flag (%#.16lx)."), paf,
ARCB_V1_PAF_NONCE);
return;
}
if ((paf & ~known_flags) != 0)
pvattest_log_warning(
_("The given paf (%#.16lx) specifies unknown flags. Use at your own risk!"),
paf, known_flags);
arcb->paf = paf;
}
arcb_v1_t *arcb_v1_new(GBytes *arpk, GBytes *iv, uint32_t mai, EVP_PKEY *evp_cpk, GBytes *mkey,
uint64_t paf, GError **error)
{
g_autoptr(arcb_v1_t) arcb = g_new0(arcb_v1_t, 1);
g_assert(g_bytes_get_size(iv) == ARCB_V1_IV_SIZE);
g_assert(g_bytes_get_size(arpk) == ARCB_V1_ATTEST_PROT_KEY_SIZE);
g_assert(g_bytes_get_size(mkey) == HMAC_SHA512_KEY_SIZE);
pv_wrapped_g_assert(arpk);
pv_wrapped_g_assert(iv);
pv_wrapped_g_assert(evp_cpk);
pv_wrapped_g_assert(mkey);
arcb->arvn = ARVN_VERSION_1;
arcb->mai = mai;
arcb_v1_set_paf(arcb, paf, error);
if (*error)
return NULL;
arcb->iv = g_bytes_ref(iv);
if (EVP_PKEY_up_ref(evp_cpk) != 1)
g_abort();
arcb->evp_cust_pub_key = evp_cpk;
arcb->confidential_att_req_prot_key = g_bytes_ref(arpk);
arcb->confidential_measurement_key = g_bytes_ref(mkey);
return g_steal_pointer(&arcb);
}
int arcb_v1_add_key_slot(arcb_v1_t *arcb, EVP_PKEY *evp_host, GError **error)
{
g_autoptr(GBytes) warpk = NULL, tag = NULL, phkh = NULL;
g_autoptr(GBytes) exchange_key = NULL, iv = NULL;
g_autofree arcb_v1_key_slot_t *key_slot = NULL;
g_autofree PvEcdhPubKey *ecdh_host = NULL;
g_autofree uint8_t *iv_raw = NULL;
PvCipherParms parms;
int64_t gcm_rc;
g_assert(arcb->confidential_att_req_prot_key);
pv_wrapped_g_assert(arcb);
pv_wrapped_g_assert(evp_host);
/* encrypt (=wrap) attestation request protection key, store warpk + tag */
exchange_key = pv_derive_exchange_key(arcb->evp_cust_pub_key, evp_host, error);
if (!exchange_key)
return -1;
iv_raw = g_malloc0(ARCB_V1_IV_SIZE);
iv = g_bytes_new_take(g_steal_pointer(&iv_raw), ARCB_V1_IV_SIZE);
if (!iv)
g_abort();
parms.key = exchange_key;
parms.iv = iv;
parms.cipher = EVP_aes_256_gcm();
parms.tag_size = ARCB_V1_TAG_SIZE;
gcm_rc = pv_gcm_encrypt(arcb->confidential_att_req_prot_key, NULL, &parms, &warpk, &tag,
error);
if (gcm_rc != ARCB_V1_ATTEST_PROT_KEY_SIZE)
return -1;
/* calculate public host key hash */
ecdh_host = pv_evp_pkey_to_ecdh_pub_key(evp_host, error);
if (!ecdh_host)
return -1;
phkh = pv_sha256_hash(ecdh_host->data, sizeof(ecdh_host->data), error);
if (!phkh)
return -1;
/* copy to list */
g_assert(g_bytes_get_size(warpk) == sizeof(key_slot->warpk));
g_assert(g_bytes_get_size(tag) == sizeof(key_slot->kst));
g_assert(g_bytes_get_size(phkh) == sizeof(key_slot->phkh));
key_slot = g_malloc0(sizeof(*key_slot));
pv_gbytes_memcpy(key_slot->warpk, sizeof(key_slot->warpk), warpk);
pv_gbytes_memcpy(key_slot->kst, sizeof(key_slot->warpk), tag);
pv_gbytes_memcpy(key_slot->phkh, sizeof(key_slot->warpk), phkh);
arcb->host_key_slots = g_slist_prepend(arcb->host_key_slots, g_steal_pointer(&key_slot));
return 0;
}
void arcb_v1_set_nonce(arcb_v1_t *arcb, GBytes *nonce)
{
pv_wrapped_g_assert(arcb);
pv_wrapped_g_assert(nonce);
arcb_v1_rm_nonce(arcb);
g_assert(!arcb->confidential_optional_nonce);
g_assert(g_bytes_get_size(nonce) == ARCB_V1_NONCE_SIZE);
arcb->confidential_optional_nonce = g_bytes_ref(nonce);
arcb->paf |= ARCB_V1_PAF_NONCE;
}
void arcb_v1_rm_nonce(arcb_v1_t *arcb)
{
pv_wrapped_g_assert(arcb);
if (!arcb->confidential_optional_nonce)
return;
g_bytes_unref(arcb->confidential_optional_nonce);
arcb->confidential_optional_nonce = NULL;
arcb->paf &= ~ARCB_V1_PAF_NONCE;
}
GBytes *arcb_v1_serialize(const arcb_v1_t *arcb, GError **error)
{
pv_wrapped_g_assert(arcb);
g_autoptr(GByteArray) arcb_gba = NULL;
g_autoptr(GBytes) confidential_area = NULL;
g_autoptr(GBytes) aad = NULL;
g_autoptr(GBytes) art = NULL;
g_autoptr(GBytes) encrypted_area = NULL;
g_autoptr(GBytes) result = NULL;
g_autofree PvEcdhPubKey *ecdh_cpk = NULL;
PvCipherParms parms = {
.cipher = EVP_aes_256_gcm(),
.tag_size = AES_256_GCM_TAG_SIZE,
};
size_t att_req_len = 0, nks = 0, sea = 0;
arcb_v1_hdr_t hdr = {
.arvn = GUINT32_TO_BE(arcb->arvn),
.paf = GUINT64_TO_BE(arcb->paf),
.mai = GUINT32_TO_BE(arcb->mai),
};
g_assert(arcb->host_key_slots);
/* calculate sizes */
nks = g_slist_length(arcb->host_key_slots);
g_assert(nks < 0xFF);
sea = g_bytes_get_size(arcb->confidential_measurement_key);
if (arcb->confidential_optional_nonce)
sea += g_bytes_get_size(arcb->confidential_optional_nonce);
g_assert(sea == HMAC_SHA512_KEY_SIZE || sea == HMAC_SHA512_KEY_SIZE + ARCB_V1_NONCE_SIZE);
att_req_len = sizeof(hdr) + nks * sizeof(arcb_v1_key_slot_t) + HMAC_SHA512_KEY_SIZE +
ARCB_V1_TAG_SIZE;
if (arcb->confidential_optional_nonce)
att_req_len += ARCB_V1_NONCE_SIZE;
g_assert(att_req_len <= MAX_ARL);
/* copy plain data to contiguous memory */
hdr.arl = GUINT32_TO_BE((uint32_t)att_req_len);
pv_gbytes_memcpy(hdr.iv, ARCB_V1_IV_SIZE, arcb->iv);
hdr.nks = (uint8_t)nks;
hdr.sea = GUINT32_TO_BE((uint32_t)sea);
ecdh_cpk = pv_evp_pkey_to_ecdh_pub_key(arcb->evp_cust_pub_key, error);
memcpy(&hdr.cpk, ecdh_cpk, sizeof(*ecdh_cpk));
arcb_gba = g_byte_array_sized_new((guint)att_req_len);
g_byte_array_append(arcb_gba, (const uint8_t *)&hdr, sizeof(hdr));
for (GSList *elem = arcb->host_key_slots; elem; elem = elem->next)
g_byte_array_append(arcb_gba, elem->data, sizeof(arcb_v1_key_slot_t));
/* encrypt the confidential data */
confidential_area = secure_gbytes_concat(arcb->confidential_measurement_key,
arcb->confidential_optional_nonce);
parms.key = arcb->confidential_att_req_prot_key;
parms.iv = arcb->iv;
aad = g_bytes_new(arcb_gba->data, arcb_gba->len);
pv_gcm_encrypt(confidential_area, aad, &parms, &encrypted_area, &art, error);
if (*error)
return NULL;
g_byte_array_append(arcb_gba, g_bytes_get_data(encrypted_area, NULL), (guint)sea);
g_byte_array_append(arcb_gba, g_bytes_get_data(art, NULL), ARCB_V1_TAG_SIZE);
result = g_byte_array_free_to_bytes(arcb_gba);
arcb_gba = NULL;
return g_steal_pointer(&result);
}
uint32_t arcb_v1_get_required_measurement_size(const arcb_v1_t *arcb, GError **error)
{
pv_wrapped_g_assert(arcb);
switch (arcb->mai) {
case MAI_HMAC_SHA512:
return HMAC_SHA512_KEY_SIZE;
default:
g_set_error(error, ARCB_ERROR, ARCB_ERR_INVALID_MAI,
_("Unknown measurement algorithm ID specified (%#x)."), arcb->mai);
return 0;
}
}
uint32_t arcb_v1_get_required_additional_size(const arcb_v1_t *arcb)
{
uint32_t size = 0;
pv_wrapped_g_assert(arcb);
if (arcb_v1_additional_has_phkh_image(arcb))
size += ARCB_V1_PHKH_SIZE;
if (arcb_v1_additional_has_phkh_attest(arcb))
size += ARCB_V1_PHKH_SIZE;
return size;
}
gboolean arcb_v1_use_nonce(const arcb_v1_t *arcb)
{
pv_wrapped_g_assert(arcb);
return arcb->confidential_optional_nonce != NULL;
}
gboolean arcb_v1_additional_has_phkh_image(const arcb_v1_t *arcb)
{
pv_wrapped_g_assert(arcb);
return (arcb->paf & ARCB_V1_PAF_AAD_PHKH_HEADER) != 0;
}
gboolean arcb_v1_additional_has_phkh_attest(const arcb_v1_t *arcb)
{
pv_wrapped_g_assert(arcb);
return (arcb->paf & ARCB_V1_PAF_AAD_PHKH_ATTEST) != 0;
}
GBytes *arcb_v1_get_measurement_key(const arcb_v1_t *arcb)
{
pv_wrapped_g_assert(arcb);
return g_bytes_ref(arcb->confidential_measurement_key);
}
GBytes *arcb_v1_get_nonce(const arcb_v1_t *arcb)
{
pv_wrapped_g_assert(arcb);
if (arcb->confidential_optional_nonce)
return g_bytes_ref(arcb->confidential_optional_nonce);
return NULL;
}
GBytes *arcb_v1_get_arp_key(const arcb_v1_t *arcb)
{
pv_wrapped_g_assert(arcb);
return g_bytes_ref(arcb->confidential_att_req_prot_key);
}
static gboolean is_v1_arcb(size_t aad_size, size_t sea, size_t arl, size_t serialized_arcb_size,
uint32_t arcb_version, gboolean has_nonce)
{
gboolean result = aad_size + sea + ARCB_V1_TAG_SIZE == arl;
result &= arl <= serialized_arcb_size;
result &= arcb_version == ARVN_VERSION_1;
result &= has_nonce ? sea == HMAC_SHA512_KEY_SIZE + ARCB_V1_NONCE_SIZE :
sea == HMAC_SHA512_KEY_SIZE;
return result;
}
gboolean arcb_v1_verify_serialized_arcb(GBytes *serialized_arcb, GBytes *arpk,
GBytes **measurement_key, GBytes **optional_nonce,
GError **error)
{
g_autoptr(GBytes) encr = NULL, decr = NULL, aad = NULL, tag = NULL, iv = NULL;
const struct arcb_v1_hdr *serialized_arcb_hdr;
const uint8_t *encr_u8, *aad_u8, *tag_u8;
const uint8_t *serialized_arcb_u8;
size_t serialized_arcb_size;
uint32_t arcb_version, mai;
size_t aad_size, arl, sea;
PvCipherParms parms;
gboolean has_nonce;
uint64_t paf;
pv_wrapped_g_assert(serialized_arcb);
pv_wrapped_g_assert(arpk);
serialized_arcb_u8 = g_bytes_get_data(serialized_arcb, &serialized_arcb_size);
serialized_arcb_hdr = (const arcb_v1_hdr_t *)serialized_arcb_u8;
arl = GUINT32_FROM_BE(serialized_arcb_hdr->arl);
arcb_version = GUINT32_FROM_BE(serialized_arcb_hdr->arvn);
mai = GUINT32_FROM_BE(serialized_arcb_hdr->mai);
aad_u8 = serialized_arcb_u8;
aad_size = sizeof(*serialized_arcb_hdr) +
serialized_arcb_hdr->nks * sizeof(arcb_v1_key_slot_t);
encr_u8 = aad_u8 + aad_size;
sea = GUINT32_FROM_BE(serialized_arcb_hdr->sea);
tag_u8 = encr_u8 + sea;
paf = GUINT64_FROM_BE(serialized_arcb_hdr->paf);
has_nonce = (paf & ARCB_V1_PAF_NONCE) != 0;
if (!is_v1_arcb(aad_size, sea, arl, serialized_arcb_size, arcb_version, has_nonce)) {
g_set_error(error, ARCB_ERROR, ARCB_ERR_INVALID_ARCB,
_("The provided attestation request is not valid"));
return FALSE;
}
if (mai != MAI_HMAC_SHA512) {
g_set_error(error, ARCB_ERROR, ARCB_ERR_INVALID_MAI,
_("Unsupported measurement argument ID (%#x)"), mai);
return FALSE;
}
aad = g_bytes_new(aad_u8, aad_size);
encr = g_bytes_new(encr_u8, sea);
tag = g_bytes_new(tag_u8, ARCB_V1_TAG_SIZE);
iv = g_bytes_new(serialized_arcb_hdr->iv, sizeof(serialized_arcb_hdr->iv));
parms.cipher = EVP_aes_256_gcm();
parms.tag_size = AES_256_GCM_TAG_SIZE;
parms.key = arpk;
parms.iv = iv;
pv_gcm_decrypt(encr, aad, tag, &parms, &decr, error);
if (*error) {
GError *tmp_error = NULL;
g_set_error(&tmp_error, ARCB_ERROR, ARCB_ERR_INVALID_ARCB,
_("Cannot verify the attestation request: %s"), (*error)->message);
g_clear_error(error);
g_propagate_error(error, tmp_error);
return FALSE;
}
if (measurement_key)
*measurement_key = g_bytes_new(g_bytes_get_data(decr, NULL), HMAC_SHA512_KEY_SIZE);
if (optional_nonce && has_nonce)
*optional_nonce =
g_bytes_new((uint8_t *)g_bytes_get_data(decr, NULL) + HMAC_SHA512_KEY_SIZE,
ARCB_V1_NONCE_SIZE);
return TRUE;
}

152
pvattest/src/arcb.h Normal file
View File

@@ -0,0 +1,152 @@
/*
* Attestation Request Control Block related functions
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#ifndef PVATTEST_ARCB_H
#define PVATTEST_ARCB_H
/* Must be included before any other header */
#include "config.h"
#include "libpv/glib-helper.h"
#include "lib/zt_common.h"
#include "libpv/crypto.h"
#include "types.h"
#define MAI_HMAC_RESERVED_INVALID 0
#define MAI_HMAC_SHA512 0x1
#define HMAC_SHA512_KEY_SIZE 64
#define ARCB_V1_ATTEST_PROT_KEY_SIZE 32
#define ARCB_V1_NONCE_SIZE 16
#define ARCB_V1_TAG_SIZE 16
#define ARCB_V1_IV_SIZE 12
#define ARCB_V1_PHKH_SIZE 32
#define BIT(bit) ((uint64_t)1 << (63 - (bit)))
/* Optional nonce in ARCB */
#define ARCB_V1_PAF_NONCE BIT(1)
/* Public host key hash used to unseal SE header added to additional data to be measured */
#define ARCB_V1_PAF_AAD_PHKH_HEADER BIT(2)
/* Public host key hash used to unseal this attestation added to additional data to be measured */
#define ARCB_V1_PAF_AAD_PHKH_ATTEST BIT(3)
/* Temporary backup-host-key use allowed */
#define ARCB_V1_PAF_TMP_BACKUP_ALLOWED BIT(62)
/* Global not-host-specific key allowed */
#define ARCB_V1_PAF_GLOBAL_NHS_KEY_ALLOWED BIT(63)
#define ARCB_V1_PAF_ALL \
(ARCB_V1_PAF_NONCE | ARCB_V1_PAF_AAD_PHKH_HEADER | ARCB_V1_PAF_AAD_PHKH_ATTEST | \
ARCB_V1_PAF_TMP_BACKUP_ALLOWED | ARCB_V1_PAF_GLOBAL_NHS_KEY_ALLOWED)
typedef struct arcb_v1 arcb_v1_t;
/** arcb_v1_new:
*
* @arpk: Attestation Request Protection key. AES-GCM-256 key to
* protect Measurement Key and Nonce.
* Must be ´ARCB_V1_ATTEST_PROT_KEY_SIZE´ bytes long.
* @iv: IV for protecting Measuremt Key and Nonce.
* Should be random for each new ARPK.
* Must be ´ARCB_V1_IV_SIZE´ bytes long.
* @mai: Measurement Algorithm Identifier for the attestation measurement.
* See ´enum mai´
* @evp_cpk: Customer key in EVP_PKEY format. Must contain private and public key pair.
* @mkey: Measurement key to calculate the Measurement.
* Must be ´HMAC_SHA512_KEY_SIZE´ bytes long.
* @paf: Plain text Attestation Flags. See ´enum plaintext_attestattion_flags´.
* ´ARCB_V1_PAF_NONCE´ must not be set.
* @error: GError. *error will != NULL if error occours.
*
* arpk, mkey, and iv must me correct size
* If not this is considered as a programming error (No warning;
* Results in Assertion or undefined behavior).
*
* GBytes will be ref'ed.
*
* All numbers must be in system byte order and will be converted to big endian
* if needed.
*
* Returns: (nullable) (transfer full): new ARCB context.
*/
arcb_v1_t *arcb_v1_new(GBytes *arpk, GBytes *iv, uint32_t mai, EVP_PKEY *evp_cpk, GBytes *mkey,
uint64_t paf, GError **error) PV_NONNULL(1, 2, 4, 5);
void arcb_v1_clear_free(arcb_v1_t *arcb);
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(arcb_v1_t, arcb_v1_clear_free)
/** arcb_v1_add_key_slot:
*
* @arcb: ARCB context.
* @evp_host: Host public key.
* @error: GError. *error will != NULL if error occours.
*
* Builds a key slot. Calculates exchange key, wraps ARPK with the exchange key.
* Calculates the public host key hash. Calculates the key slot tag.
* Adds it to the ARCB.
*
* Returns: 0 in case of success, -1 otherwise
*/
int arcb_v1_add_key_slot(arcb_v1_t *arcb, EVP_PKEY *evp_host, GError **error) PV_NONNULL(1, 2);
void arcb_v1_set_nonce(arcb_v1_t *arcb, GBytes *nonce) PV_NONNULL(1, 2);
void arcb_v1_rm_nonce(arcb_v1_t *arcb) PV_NONNULL(1);
/** arcb_v1_serialize:
*
* @arcb: ARCB context.
* @error: GError. *error will != NULL if error occurs.
*
* Will create a valid ARCB for the UV. Including encrypting confidential data.
* At least one key_slot must be added beforehand.
*
* Returns: (nullable) (transfer full): The serialized ARCB which can be added to the
* Retrieve Attestation Measurement UVC as GBytes.
*/
GBytes *arcb_v1_serialize(const arcb_v1_t *arcb, GError **error) PV_NONNULL(1, 2);
uint32_t arcb_v1_get_required_measurement_size(const arcb_v1_t *arcb, GError **error)
PV_NONNULL(1, 2);
uint32_t arcb_v1_get_required_additional_size(const arcb_v1_t *arcb) PV_NONNULL(1);
gboolean arcb_v1_use_nonce(const arcb_v1_t *arcb) PV_NONNULL(1);
gboolean arcb_v1_additional_has_phkh_image(const arcb_v1_t *arcb) PV_NONNULL(1);
gboolean arcb_v1_additional_has_phkh_attest(const arcb_v1_t *arcb) PV_NONNULL(1);
GBytes *arcb_v1_get_measurement_key(const arcb_v1_t *arcb) PV_NONNULL(1);
GBytes *arcb_v1_get_nonce(const arcb_v1_t *arcb) PV_NONNULL(1);
GBytes *arcb_v1_get_arp_key(const arcb_v1_t *arcb) PV_NONNULL(1);
/** arcb_v1_verify_serialized_arcb:
*
* @serialized_arcb: binary ARCB in UV readable format.
* @arpk: Attestation Request Protection key that was used to create serialized_arpk
* @measurement_key: Output parameter: decrypted measurement key if no error.
* May be NULL if not interested for this output.
* @optional_nonce: Output parameter: decrypted nonce if no error.
* May be NULL if not interested for this output.
* @error: GError. *error will != NULL if error occurs.
*
*
* Checks if sizes are sound and flags are known by this implementation.
* Decrypts Measurement key and nonce (if given) and verifies ARCB tag.
*
* Returns: TRUE if ARCB is valid, including matching ARCB tag. Otherwise FALSE.
*
*/
gboolean arcb_v1_verify_serialized_arcb(GBytes *serialized_arcb, GBytes *arpk,
GBytes **measurement_key, GBytes **optional_nonce,
GError **error) PV_NONNULL(1, 2);
#define ARCB_ERROR g_quark_from_static_string("pv-arcb_error-quark")
typedef enum arcb_error {
ARCB_ERR_INVALID_ARCB,
ARCB_ERR_INVALID_PAF,
ARCB_ERR_INVALID_MAI,
ARCB_ERR_UNABLE_ENCR_ARPK,
} arcb_error_e;
#endif /* PVATTEST_ARCB_H */

649
pvattest/src/argparse.c Normal file
View File

@@ -0,0 +1,649 @@
/*
* Definitions used for parsing arguments.
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
/* Must be included before any other header */
#include "config.h"
#include <stdio.h>
#include <fcntl.h>
#include <errno.h>
#include <unistd.h>
#include "argparse.h"
#include "log.h"
#include "common.h"
#define DEFAULT_OUTPUT_FILE_NAME "attest.bin"
#define DEFAULT_OPTION_PHKH_IMG FALSE
#define DEFAULT_OPTION_PHKH_ATT FALSE
#define DEFAULT_OPTION_NO_VERIFY FALSE
#define DEFAULT_OPTION_ONLINE TRUE
#define DEFAULT_OPTION_NONCE TRUE
static pvattest_config_t pvattest_config = {
.general = {
.log_level = PVATTEST_LOG_LVL_DEFAULT,
},
.create = {
.output_path = NULL,
.host_key_document_paths = NULL,
.crl_paths = NULL,
.root_ca_path = NULL,
.certificate_paths = NULL,
.arp_key_out_path = NULL,
.phkh_img = DEFAULT_OPTION_PHKH_IMG,
.phkh_att = DEFAULT_OPTION_PHKH_ATT,
.online = DEFAULT_OPTION_ONLINE,
.use_nonce = DEFAULT_OPTION_NONCE,
.paf = 0,
.x_aad_size = -1,
},
.perform = {
.output_path = NULL,
.input_path = NULL,
},
.verify = {
.input_path = NULL,
.hdr_path = NULL,
.arp_key_in_path = NULL,
},
};
typedef gboolean (*verify_options_fn_t)(GError **);
static gboolean check_for_non_null(const void *ptr, const char *msg, GError **error)
{
if (!ptr) {
g_set_error(error, PVATTEST_ERROR, PVATTEST_ERR_INV_ARG, "%s", msg);
return FALSE;
}
return TRUE;
}
static gboolean _check_for_invalid_path(const char *path, gboolean must_exist, GError **error)
{
int cached_errno = 0;
g_assert(path);
if (must_exist) {
if (access(path, F_OK | R_OK) != 0)
cached_errno = errno;
}
if (cached_errno) {
g_set_error(error, PVATTEST_ERROR, PVATTEST_ERR_INV_ARG, "Cannot access '%s': %s",
path, g_strerror(cached_errno));
return FALSE;
}
return TRUE;
}
static gboolean check_for_optional_invalid_path(const char *path, gboolean must_exist,
GError **error)
{
if (!path)
return TRUE;
return _check_for_invalid_path(path, must_exist, error);
}
static gboolean check_for_invalid_path(const char *path, gboolean must_exist, const char *null_msg,
GError **error)
{
if (!check_for_non_null(path, null_msg, error))
return FALSE;
return _check_for_invalid_path(path, must_exist, error);
}
static gboolean _check_file_list(char **path_list, gboolean must_exist, GError **error)
{
char *path = NULL;
for (char **path_it = path_list; path_it != NULL && *path_it != NULL; path_it++) {
path = *path_it;
if (!_check_for_invalid_path(path, must_exist, error))
return FALSE;
}
return TRUE;
}
static gboolean check_optional_file_list(char **path_list, gboolean must_exist, GError **error)
{
if (!path_list)
return TRUE;
return _check_file_list(path_list, must_exist, error);
}
static gboolean check_file_list(char **path_list, gboolean must_exist, const char *null_msg,
GError **error)
{
if (!check_for_non_null(path_list, null_msg, error))
return FALSE;
return _check_file_list(path_list, must_exist, error);
}
static gboolean hex_str_toull(const char *nptr, uint64_t *dst, GError **error)
{
uint64_t value;
gchar *end;
g_assert(dst);
if (!g_str_is_ascii(nptr)) {
g_set_error(
error, PVATTEST_ERROR, PVATTEST_ERR_INV_ARG,
_("Invalid value: '%s'. A hexadecimal value is required, for example '0xcfe'"),
nptr);
return FALSE;
}
value = g_ascii_strtoull(nptr, &end, 16);
if ((value == G_MAXUINT64 && errno == ERANGE) || (end && *end != '\0')) {
g_set_error(
error, PVATTEST_ERROR, PVATTEST_ERR_INV_ARG,
_("Invalid value: '%s'. A hexadecimal value is required, for example '0xcfe'"),
nptr);
return FALSE;
}
*dst = value;
return TRUE;
}
#pragma GCC diagnostic push
#pragma GCC diagnostic ignored "-Wmissing-field-initializers"
/************************* SHARED OPTIONS *************************************/
/* NOTE REQUIRED */
#define _entry_host_key_document(__arg_data, __indent) \
{ \
.long_name = "host-key-document", .short_name = 'k', .flags = G_OPTION_FLAG_NONE, \
.arg = G_OPTION_ARG_FILENAME_ARRAY, .arg_data = __arg_data, \
.description = "Specify one or more host key documents.\n", \
.arg_description = "FILE", \
}
/* NOTE REQUIRED */
#define _entry_certs(__arg_data, __indent) \
{ \
.long_name = "cert", .short_name = 'C', .flags = G_OPTION_FLAG_NONE, \
.arg = G_OPTION_ARG_FILENAME_ARRAY, .arg_data = __arg_data, \
.description = \
"Specifies the certificate that is used to establish a chain\n" __indent \
"of trust for the verification of the host-key documents. Specify\n" __indent \
"this option twice to specify the IBM Z signing key and the\n" __indent \
"intermediate CA certificate (signed by the root CA). Required.\n" __indent \
"Ignored when --no-verify is specified.\n", \
.arg_description = "FILE", \
}
/* NOTE REQUIRED */
#define _entry_crls(__arg_data, __indent) \
{ \
.long_name = "crl", .short_name = 0, .flags = G_OPTION_FLAG_NONE, \
.arg = G_OPTION_ARG_FILENAME_ARRAY, .arg_data = __arg_data, \
.description = "Specify FILE to be a certificate revocation list\n" __indent \
"(optional).", \
.arg_description = "FILE", \
}
/* 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).", \
.arg_description = "FILE", \
}
/* NOTE REQUIRED */
#define _entry_guest_hdr(__arg_data, __indent) \
{ \
.long_name = "hdr", .short_name = 0, .flags = G_OPTION_FLAG_NONE, \
.arg = G_OPTION_ARG_FILENAME, .arg_data = __arg_data, \
.description = \
"Specify the header of the guest image. Exactly one is required.\n", \
.arg_description = "FILE", \
}
/* NOTE REQUIRED */
#define _entry_input(__arg_data, __additional_text, __indent) \
{ \
.long_name = "input", .short_name = 'i', .flags = G_OPTION_FLAG_NONE, \
.arg = G_OPTION_ARG_FILENAME, .arg_data = __arg_data, \
.description = "FILE specifies the " __additional_text "\n" __indent \
" as input.\n", \
.arg_description = "FILE", \
}
/* NOTE REQUIRED */
#define _entry_output(__arg_data, __additional_text, __indent) \
{ \
.long_name = "output", .short_name = 'o', .flags = G_OPTION_FLAG_NONE, \
.arg = G_OPTION_ARG_FILENAME, .arg_data = __arg_data, \
.description = "FILE specifies the output for the\n" __indent __additional_text \
".\n", \
.arg_description = "FILE", \
}
/* NOTE REQUIRED */
#define _entry_att_prot_key_save(__arg_data, __indent) \
{ \
.long_name = "arpk", .short_name = 'a', .flags = G_OPTION_FLAG_NONE, \
.arg = G_OPTION_ARG_FILENAME, .arg_data = __arg_data, \
.description = \
"Save the protection key as GCM-AES256 key in FILE\n" __indent \
"Do not publish this key, otherwise your attestation is compromised.\n", \
.arg_description = "FILE", \
}
/* NOTE REQUIRED */
#define _entry_att_prot_key_load(__arg_data, __indent) \
{ \
.long_name = "arpk", .short_name = 'a', .flags = G_OPTION_FLAG_NONE, \
.arg = G_OPTION_ARG_FILENAME, .arg_data = __arg_data, \
.description = "Use FILE to specify the GCM-AES256 key to decrypt\n" __indent \
"the attestation request.\n" __indent \
"Delete this key after verification.\n", \
.arg_description = "FILE", \
}
#define _entry_phkh_img(__arg_data, __indent) \
{ \
.long_name = "x-phkh-img", .short_name = 0, .flags = G_OPTION_FLAG_NONE, \
.arg = G_OPTION_ARG_NONE, .arg_data = __arg_data, \
.description = "add the public host key hash of the\n" __indent \
"image header used to decrypt\n" __indent \
"the secure guest to the measurement. (optional)\n" \
}
#define _entry_phkh_att(__arg_data, __indent) \
{ \
.long_name = "x-phkh-att", .short_name = 0, .flags = G_OPTION_FLAG_NONE, \
.arg = G_OPTION_ARG_NONE, .arg_data = __arg_data, \
.description = "add the public host key hash of the\n" __indent \
"attestation header used to decrypt\n" __indent \
"the attestation request to the measurement. (optional)\n" \
}
#define _entry_no_verify(__arg_data, __indent) \
{ \
.long_name = "no-verify", .short_name = 0, .flags = G_OPTION_FLAG_NONE, \
.arg = G_OPTION_ARG_NONE, .arg_data = __arg_data, \
.description = \
"Disable the host-key-document verification.\n" __indent \
"Does not require the host-key documents to be valid.\n" __indent \
"For testing purposes, do not use for a production image.\n" __indent \
"(optional)\n", \
}
#define _entry_offline_maps_to_online(__arg_data, __indent) \
{ \
.long_name = "offline", .short_name = 0, .flags = G_OPTION_FLAG_REVERSE, \
.arg = G_OPTION_ARG_NONE, .arg_data = __arg_data, \
.description = "Don't download CRLs (optional).\n", \
}
#define _entry_verbose(__indent) \
{ \
.long_name = "verbose", .short_name = 'V', .flags = G_OPTION_FLAG_NO_ARG, \
.arg = G_OPTION_ARG_CALLBACK, .arg_data = &increase_log_lvl, \
.description = "Provide more detailed output (optional)\n", \
.arg_description = NULL, \
}
#define _entry_x_paf(__arg_data, __indent) \
{ \
.long_name = "x-paf", .short_name = 0, .flags = G_OPTION_FLAG_NONE, \
.arg = G_OPTION_ARG_CALLBACK, .arg_data = __arg_data, \
.description = "Specify the Plain text Attestation Flags\n" __indent \
"as a hexadecimal value. Flags that change\n" __indent \
"the paf (--phkh-*) take precedence over\n" __indent \
"this flag.\n" __indent \
"Setting the nonce paf is not allowed here.\n" __indent \
"(optional, default 0x0)\n", \
.arg_description = "HEX", \
}
#define _entry_x_no_nonce(__arg_data, __indent) \
{ \
.long_name = "x-no-nonce", .short_name = 0, .flags = G_OPTION_FLAG_REVERSE, \
.arg = G_OPTION_ARG_NONE, .arg_data = __arg_data, \
.description = "Do not use a nonce in the request.\n" __indent \
"(optional, not recommended)\n" \
}
#define _entry_x_aad_size(__arg_data, __indent) \
{ \
.long_name = "x-add-size", .short_name = 0, .flags = G_OPTION_FLAG_NONE, \
.arg = G_OPTION_ARG_INT, .arg_data = __arg_data, \
.description = "Specify the size of the additional area\n" __indent \
"Overwrite every flag that changes\n" __indent \
"this size implicitly. No verification is performed!\n" __indent \
"Ignored if negative.\n" __indent "(optional, default ignored)\n", \
.arg_description = "INT" \
}
#define _entry_x_user_data(__arg_data, __indent) \
{ \
.long_name = "x-user-data", .short_name = 0, .flags = G_OPTION_FLAG_NONE, \
.arg = G_OPTION_ARG_FILENAME, .arg_data = __arg_data, \
.description = "Use FILE to specify the user data.\n", .arg_description = "FILE", \
}
static gboolean increase_log_lvl(G_GNUC_UNUSED const char *option_name,
G_GNUC_UNUSED const char *value, G_GNUC_UNUSED void *data,
G_GNUC_UNUSED GError **error)
{
pvattest_log_increase_log_lvl(&pvattest_config.general.log_level);
return TRUE;
}
static gboolean create_set_paf(G_GNUC_UNUSED const char *option_name, const char *value,
G_GNUC_UNUSED void *data, GError **error)
{
return hex_str_toull(value, &pvattest_config.create.paf, error);
}
/***************************** GENERAL OPTIONS ********************************/
static gboolean print_version = FALSE;
static GOptionEntry general_options[] = {
{
.long_name = "version",
.short_name = 'v',
.flags = G_OPTION_FLAG_NONE,
.arg = G_OPTION_ARG_NONE,
.arg_data = &print_version,
.description = "Print the version and exit.\n",
.arg_description = NULL,
},
_entry_verbose(""),
{ NULL },
};
/************************* CREATE ATTESTATION OPTIONS *************************/
#define create_indent " "
static GOptionEntry create_options[] = {
_entry_host_key_document(&pvattest_config.create.host_key_document_paths, create_indent),
_entry_certs(&pvattest_config.create.certificate_paths, create_indent),
_entry_crls(&pvattest_config.create.crl_paths, create_indent),
_entry_root_ca(&pvattest_config.create.root_ca_path, create_indent),
_entry_output(&pvattest_config.create.output_path, "attestation request", create_indent),
_entry_att_prot_key_save(&pvattest_config.create.arp_key_out_path, create_indent),
_entry_no_verify(&pvattest_config.create.no_verify, create_indent),
_entry_offline_maps_to_online(&pvattest_config.create.online, create_indent),
_entry_verbose(create_indent),
{ NULL }
};
static GOptionEntry experimental_create_options[] = {
_entry_x_no_nonce(&pvattest_config.create.use_nonce, create_indent),
_entry_x_paf(&create_set_paf, create_indent),
_entry_x_aad_size(&pvattest_config.create.x_aad_size, create_indent),
_entry_phkh_img(&pvattest_config.create.phkh_img, create_indent),
_entry_phkh_att(&pvattest_config.create.phkh_att, create_indent),
{ NULL }
};
static gboolean verify_create(GError **error)
{
if (!check_file_list(pvattest_config.create.host_key_document_paths, TRUE,
_("Specify --host-key-document at least once."), error))
return FALSE;
if (!pvattest_config.create.no_verify) {
if (!check_file_list(
pvattest_config.create.certificate_paths, TRUE,
_("Either specify the IBM Z signing key and"
" intermediate CA certificate\nby using the '--cert' option, or"
" use the '--no-verify' flag to disable the\nhost-key document"
" verification completely (at your own risk).\n"
"Only use this option in test environments or if"
" you trust the unverified document."),
error))
return FALSE;
}
if (!check_for_invalid_path(pvattest_config.create.arp_key_out_path, FALSE,
_("Missing argument for --arpk."), error))
return FALSE;
if (!check_for_invalid_path(pvattest_config.create.output_path, FALSE,
_("Missing argument for --output."), error))
return FALSE;
if (!check_optional_file_list(pvattest_config.create.crl_paths, TRUE, error))
return FALSE;
if (!check_for_optional_invalid_path(pvattest_config.create.root_ca_path, TRUE, error))
return FALSE;
return TRUE;
};
/************************* MEASUREMENT OPTIONS ********************************/
#define perform_indent " "
static GOptionEntry perform_options[] = {
_entry_input(&pvattest_config.perform.input_path, "attestation request", perform_indent),
_entry_output(&pvattest_config.perform.output_path, "attestation result", perform_indent),
_entry_verbose(perform_indent),
{ NULL },
};
static GOptionEntry experimental_perform_options[] = {
_entry_x_user_data(&pvattest_config.perform.user_data_path, perform_indent),
{ NULL },
};
static gboolean verify_perform(GError **error)
{
if (!check_for_invalid_path(pvattest_config.perform.input_path, TRUE,
_("Missing argument for --input."), error))
return FALSE;
if (!check_for_invalid_path(pvattest_config.perform.output_path, FALSE,
_("Missing argument for --output."), error))
return FALSE;
if (!check_for_optional_invalid_path(pvattest_config.perform.user_data_path, TRUE, error))
return FALSE;
return TRUE;
}
/************************* VERIFY OPTIONS ************************************/
#define verify_indent " "
static GOptionEntry verify_options[] = {
_entry_input(&pvattest_config.verify.input_path, "attestation result", verify_indent),
_entry_guest_hdr(&pvattest_config.verify.hdr_path, verify_indent),
_entry_att_prot_key_load(&pvattest_config.verify.arp_key_in_path, verify_indent),
_entry_verbose(verify_indent),
{ NULL },
};
static gboolean verify_verify(GError **error)
{
if (!check_for_invalid_path(pvattest_config.verify.input_path, TRUE,
_("Missing argument for --input."), error))
return FALSE;
if (!check_for_invalid_path(pvattest_config.verify.hdr_path, TRUE,
_("Missing argument for --hdr."), error))
return FALSE;
if (!check_for_invalid_path(pvattest_config.verify.arp_key_in_path, TRUE,
_("Missing argument for --arpk."), error))
return FALSE;
return TRUE;
}
/************************** OPTIONS END ***************************************/
#pragma GCC diagnostic pop
static char summary[] =
"\n"
"Create, perform, and verify attestation measurements for IBM Secure Execution guest"
" systems.\n"
"\n"
"COMMANDS\n"
" create On a trusted system, creates an attestation request.\n"
" perform On the SE-guest to be attested, sends the attestation request\n"
" to the Ultravisor and receives the answer.\n"
#ifndef PVATTEST_COMPILE_PERFORM
" (not supported on this platform)\n"
#endif /* PVATTEST_COMPILE_PERFORM */
" verify On a trusted system, compares the one from your trusted system.\n"
" If they differ, the Secure Execution guest might not be compromised\n"
"\n"
"Use '" GETTEXT_PACKAGE " [COMMAND] -h' to get detailed help\n";
static char create_summary[] =
"Create attestation measurement requests to attest an\n"
"IBM Secure Execution guest. Only build attestation requests in a trusted\n"
"environment such as your Workstation.\n"
"To avoid compromising the attestation do not publish the\n"
"protection key and delete it after verification.\n"
"Every 'create' will generate a new, random protection key.\n";
static char perform_summary[] =
#ifndef PVATTEST_COMPILE_PERFORM
"This system does NOT support 'perform'.\n"
#endif /* PVATTEST_COMPILE_PERFORM */
"Perform a measurement of this IBM Secure Execution guest using '/dev/uv'.\n";
static char verify_summary[] =
"Verify that a previously generated attestation measurement of an\n"
"IBM Secure Execution guest yielded the expected results.\n"
"Verify attestation requests only in a trusted environment, such as your workstation.";
static void print_version_and_exit(void)
{
printf("%s version %s\n", GETTEXT_PACKAGE, RELEASE_STRING);
printf("%s\n", COPYRIGHT_NOTICE);
exit(EXIT_SUCCESS);
}
static GOptionContext *create_ctx(GOptionEntry *options, GOptionEntry *experimental_options,
const char *param_name, const char *opt_summary)
{
GOptionContext *ret = g_option_context_new(param_name);
GOptionGroup *x_group = NULL;
g_option_context_add_main_entries(ret, options, NULL);
g_option_context_set_summary(ret, opt_summary);
if (experimental_options) {
x_group = g_option_group_new(
"experimental",
"Experimental Options; Do not use in a production environment",
"Show experimental options", NULL, NULL);
g_option_group_add_entries(x_group, experimental_options);
g_option_context_add_group(ret, x_group);
}
return ret;
}
enum pvattest_command pvattest_parse(int *argc, char **argvp[], pvattest_config_t **config,
GError **error)
{
g_autoptr(GOptionContext) main_context = NULL, subc_context = NULL;
char **argv = *argvp;
enum pvattest_command subc = PVATTEST_SUBC_INVALID;
verify_options_fn_t verify_options_fn = NULL;
pv_wrapped_g_assert(argc);
pv_wrapped_g_assert(argvp);
pv_wrapped_g_assert(config);
/*
* First parse until the first non dash argument. This must be one of the commands.
* (strict POSIX parsing)
*/
main_context = g_option_context_new(
"COMMAND [OPTIONS] - create, perform, and verify attestation measurements");
g_option_context_set_strict_posix(main_context, TRUE);
g_option_context_add_main_entries(main_context, general_options, NULL);
g_option_context_set_summary(main_context, summary);
if (!g_option_context_parse(main_context, argc, argvp, error))
return PVATTEST_SUBC_INVALID;
if (print_version)
print_version_and_exit();
/*
* Parse depending on the specified command
*/
else if (g_strcmp0(argv[1], PVATTEST_SUBC_STR_CREATE) == 0) {
subc_context =
create_ctx(create_options, experimental_create_options,
"create [OPTIONS] - create an attestation measurement request",
create_summary);
subc = PVATTEST_SUBC_CREATE;
verify_options_fn = &verify_create;
} else if (g_strcmp0(argv[1], PVATTEST_SUBC_STR_PERFORM) == 0) {
subc_context =
create_ctx(perform_options, experimental_perform_options,
"perform [OPTIONS] - perform an attestation measurement request",
perform_summary);
subc = PVATTEST_SUBC_PERFORM;
verify_options_fn = &verify_perform;
#ifndef PVATTEST_COMPILE_PERFORM
g_set_error(error, PVATTEST_ERROR, PVATTEST_ERR_INV_ARG,
_("This system does not support the 'perform' command."));
return PVATTEST_SUBC_INVALID;
#endif /* PVATTEST_COMPILE_PERFORM */
} else if (g_strcmp0(argv[1], PVATTEST_SUBC_STR_VERIFY) == 0) {
subc_context = create_ctx(verify_options, NULL,
"verify [OPTIONS] - verify an attestation measurement",
verify_summary);
subc = PVATTEST_SUBC_VERIFY;
verify_options_fn = &verify_verify;
} else {
if (argv[1])
g_set_error(error, PVATTEST_ERROR, PVATTEST_ERR_INV_ARGV,
_("Invalid command specified: %s."), argv[1]);
else
g_set_error(error, PVATTEST_ERROR, PVATTEST_ERR_INV_ARGV,
_("No command specified."));
return PVATTEST_SUBC_INVALID;
}
g_assert(verify_options_fn);
if (!g_option_context_parse(subc_context, argc, argvp, error))
return PVATTEST_SUBC_INVALID;
if (!verify_options_fn(error))
return PVATTEST_SUBC_INVALID;
*config = &pvattest_config;
return subc;
}
static void pvattest_parse_clear_create_config(pvattest_create_config_t *config)
{
if (!config)
return;
g_strfreev(config->host_key_document_paths);
g_strfreev(config->certificate_paths);
g_free(config->arp_key_out_path);
g_free(config->output_path);
}
static void pvattest_parse_clear_perform_config(pvattest_perform_config_t *config)
{
if (!config)
return;
g_free(config->input_path);
g_free(config->output_path);
}
static void pvattest_parse_clear_verify_config(pvattest_verify_config_t *config)
{
if (!config)
return;
g_free(config->input_path);
g_free(config->hdr_path);
g_free(config->arp_key_in_path);
}
void pvattest_parse_clear_config(pvattest_config_t *config)
{
if (!config)
return;
pvattest_parse_clear_create_config(&config->create);
pvattest_parse_clear_perform_config(&config->perform);
pvattest_parse_clear_verify_config(&config->verify);
}

106
pvattest/src/argparse.h Normal file
View File

@@ -0,0 +1,106 @@
/*
* Definitions used for parsing arguments.
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#ifndef PVATTEST_ARGPARSE_H
#define PVATTEST_ARGPARSE_H
/* Must be included before any other header */
#include "config.h"
#include <stdint.h>
#include "libpv/glib-helper.h"
#include "libpv/macros.h"
#define PVATTEST_SUBC_STR_CREATE "create"
#define PVATTEST_SUBC_STR_PERFORM "perform"
#define PVATTEST_SUBC_STR_VERIFY "verify"
enum pvattest_command {
PVATTEST_SUBC_INVALID,
PVATTEST_SUBC_CREATE,
PVATTEST_SUBC_PERFORM,
PVATTEST_SUBC_VERIFY,
};
typedef struct {
int log_level;
} pvattest_general_config_t;
typedef struct {
char **host_key_document_paths;
char **certificate_paths;
char **crl_paths;
char *root_ca_path;
char *arp_key_out_path;
char *output_path;
gboolean phkh_img;
gboolean phkh_att;
gboolean no_verify;
gboolean online;
/* experimental flags */
gboolean use_nonce; /* default TRUE */
uint64_t paf; /* default 0 */
int x_aad_size; /* default -1 -> ignore */
} pvattest_create_config_t;
typedef struct {
char *output_path;
char *input_path;
/* experimental flags */
char *user_data_path; /* default NULL */
} pvattest_perform_config_t;
typedef struct {
char *input_path;
char *hdr_path;
char *arp_key_in_path;
} pvattest_verify_config_t;
typedef struct {
pvattest_general_config_t general;
pvattest_create_config_t create;
pvattest_perform_config_t perform;
pvattest_verify_config_t verify;
} pvattest_config_t;
/**
* pvattest_parse_clear_config:
*
* @config: struct to be cleared
*
* clears but not frees all config.
* all non config members such like char* will be freed.
*/
void pvattest_parse_clear_config(pvattest_config_t *config);
/**
* pvattest_parse:
*
* @argc: ptr to argument count
* @argv: ptr to argument vector
* @config: output: ptr to parsed config. Target is statically allocated.
* You are responsible for freeing all non config ptrs.
* use #pvattest_parse_clear_config for that.
*
* Will not return if verbose or help parsed.
*
* Returns: selected command as enum
*/
enum pvattest_command pvattest_parse(int *argc, char **argvp[], pvattest_config_t **config,
GError **error) PV_NONNULL(1, 2, 3);
#define PVATTEST_ERROR g_quark_from_static_string("pv-pvattest_error-quark")
typedef enum {
PVATTEST_ERR_INV_ARGV,
PVATTEST_ERR_INV_ARG,
} pv_pvattest_error_e;
#endif /* PVATTEST_ARGPARSE_H */

148
pvattest/src/attestation.c Normal file
View File

@@ -0,0 +1,148 @@
/*
* Attestation related functions
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
/* Must be included before any other header */
#include "config.h"
#include "libpv/cert.h"
#include "libpv/hash.h"
#include "libpv/se-hdr.h"
#include "exchange_format.h"
#include "attestation.h"
G_STATIC_ASSERT(sizeof(((att_meas_ctx_t *)0)->pld) == sizeof(((struct pv_hdr_head *)0)->pld));
G_STATIC_ASSERT(sizeof(((att_meas_ctx_t *)0)->ald) == sizeof(((struct pv_hdr_head *)0)->ald));
G_STATIC_ASSERT(sizeof(((att_meas_ctx_t *)0)->tld) == sizeof(((struct pv_hdr_head *)0)->tld));
G_STATIC_ASSERT(sizeof(((att_meas_ctx_t *)0)->tag) == sizeof(((struct pv_hdr *)0)->tag));
struct att_meas_sizes {
uint16_t user_data_len;
uint16_t zeros;
uint32_t additional_data_len;
} __packed;
G_STATIC_ASSERT(sizeof(struct att_meas_sizes) == 8);
/*
* All optional arguments may be NULL
* user_data is up to 256 bytes long, or NULL.
* nonce is 16 bytes long or NULL.
* additional_data is up to 32768 bytes long or NULL.
*/
GBytes *att_gen_measurement_hmac_sha512(const att_meas_ctx_t *meas_ctx, GBytes *measurement_key,
GBytes *optional_user_data, GBytes *optional_nonce,
GBytes *optional_additional_data, GError **error)
{
struct att_meas_sizes meas_sizes = {};
g_autoptr(HMAC_CTX) hmac_ctx = NULL;
size_t additional_data_size = 0;
size_t user_data_size = 0;
size_t nonce_size = 0;
pv_wrapped_g_assert(meas_ctx);
pv_wrapped_g_assert(measurement_key);
if (optional_user_data)
user_data_size = g_bytes_get_size(optional_user_data);
if (optional_additional_data)
additional_data_size = g_bytes_get_size(optional_additional_data);
if (optional_nonce)
nonce_size = g_bytes_get_size(optional_nonce);
/* checks for these sizes resulting in GErrors are done before */
g_assert(user_data_size <= PVATTEST_USER_DATA_MAX_SIZE);
g_assert(additional_data_size <= PVATTEST_ADDITIONAL_MAX_SIZE);
g_assert(nonce_size == 0 || nonce_size == ARCB_V1_NONCE_SIZE);
pv_wrapped_g_assert(meas_ctx);
pv_wrapped_g_assert(measurement_key);
hmac_ctx = pv_hmac_ctx_new(measurement_key, EVP_sha512(), error);
if (!hmac_ctx)
return NULL;
meas_sizes.user_data_len = GUINT16_TO_BE((uint16_t)user_data_size);
meas_sizes.zeros = 0;
meas_sizes.additional_data_len = GUINT32_TO_BE((uint32_t)additional_data_size);
if (pv_hmac_ctx_update_raw(hmac_ctx, meas_ctx, sizeof(*meas_ctx), error) != 0)
return NULL;
/* add the sizes of user and additional data. */
if (pv_hmac_ctx_update_raw(hmac_ctx, &meas_sizes, sizeof(meas_sizes), error))
return NULL;
/* update optional data. if NULL passed (or size = 0) nothing will happen to the HMAC_CTX */
if (pv_hmac_ctx_update(hmac_ctx, optional_user_data, error) != 0)
return NULL;
if (pv_hmac_ctx_update(hmac_ctx, optional_nonce, error) != 0)
return NULL;
if (pv_hmac_ctx_update(hmac_ctx, optional_additional_data, error) != 0)
return NULL;
return pv_hamc_ctx_finalize(hmac_ctx, error);
}
att_meas_ctx_t *att_extract_from_hdr(GBytes *se_hdr, GError **error)
{
g_autofree att_meas_ctx_t *meas = NULL;
const struct pv_hdr *hdr = NULL;
size_t se_hdr_tag_offset;
size_t se_hdr_size;
uint8_t *hdr_u8;
pv_wrapped_g_assert(se_hdr);
hdr = g_bytes_get_data(se_hdr, &se_hdr_size);
hdr_u8 = (uint8_t *)hdr;
if (se_hdr_size < PV_V1_PV_HDR_MIN_SIZE) {
g_set_error(error, ATT_ERROR, ATT_ERR_INVALID_HDR,
_("Invalid SE header provided."));
return NULL;
}
if (GUINT32_FROM_BE(hdr->head.phs) != se_hdr_size ||
GUINT64_FROM_BE(hdr->head.magic) != PV_MAGIC_NUMBER) {
g_set_error(error, ATT_ERROR, ATT_ERR_INVALID_HDR,
_("Invalid SE header provided."));
return NULL;
}
se_hdr_tag_offset = GUINT32_FROM_BE(hdr->head.phs) - sizeof(hdr->tag);
meas = g_new0(att_meas_ctx_t, 1);
memcpy(meas->pld, hdr->head.pld, sizeof(meas->pld));
memcpy(meas->ald, hdr->head.ald, sizeof(meas->ald));
memcpy(meas->tld, hdr->head.tld, sizeof(meas->tld));
memcpy(meas->tag, hdr_u8 + se_hdr_tag_offset, sizeof(meas->tag));
return g_steal_pointer(&meas);
}
void att_add_uid(att_meas_ctx_t *meas_ctx, GBytes *config_uid)
{
pv_wrapped_g_assert(meas_ctx);
pv_wrapped_g_assert(config_uid);
g_assert(g_bytes_get_size(config_uid) == ATT_CONFIG_UID_SIZE);
pv_gbytes_memcpy(meas_ctx->config_uid, ATT_CONFIG_UID_SIZE, config_uid);
}
gboolean att_verify_measurement(const GBytes *calculated_measurement,
const GBytes *uvio_measurement, GError **error)
{
pv_wrapped_g_assert(calculated_measurement);
pv_wrapped_g_assert(uvio_measurement);
if (g_bytes_compare(calculated_measurement, uvio_measurement) != 0) {
g_set_error(error, ATT_ERROR, ATT_ERR_MEASUREMENT_VERIFICATION_FAILED,
_("Calculated and received attestation measurement are not equal."));
return FALSE;
}
return TRUE;
}

View File

@@ -0,0 +1,99 @@
/*
* Attestation related functions and definitions.
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#ifndef PVATTEST_ATTESTATION_H
#define PVATTEST_ATTESTATION_H
/* Must be included before any other header */
#include "config.h"
#include <openssl/sha.h>
#include "libpv/glib-helper.h"
#include "libpv/crypto.h"
#include "libpv/cert.h"
#include "common.h"
#include "types.h"
#include "arcb.h"
#define ATT_CONFIG_UID_SIZE 16
typedef struct {
uint8_t pld[SHA512_DIGEST_LENGTH];
uint8_t ald[SHA512_DIGEST_LENGTH];
uint8_t tld[SHA512_DIGEST_LENGTH];
uint8_t tag[AES_256_GCM_TAG_SIZE];
uint8_t config_uid[ATT_CONFIG_UID_SIZE];
} __packed att_meas_ctx_t;
G_STATIC_ASSERT(sizeof(att_meas_ctx_t) == 224);
/**
* att_gen_measurement_hmac_sha512:
*
* @meas_ctx: measurement context.
* @measurement_key: AES-256-GCM key for generating the measurement calculation.
* @optional_user_data: NULL or up to 256 bytes GBytes.
* @optional_nonce: NULL or a nonce of exactly `ARCB_V1_NONCE_SIZE` bytes
* @optional_additional_data: NULL or up to 0x8000 bytes of GBytes.
* @error: GError. *error will != NULL if error occurs.
*
* Calculates the measurement value.
* If the input data is the same which the UV used in the Retrieve Attestation Measurement
* the result should be identical to the data in the ´Measurement Data Address´ UVC.
*
* Returns: (nullable) (transfer full): a hmac_sha512 of the given data.
*/
GBytes *att_gen_measurement_hmac_sha512(const att_meas_ctx_t *meas_ctx, GBytes *measurement_key,
GBytes *optional_user_data, GBytes *optional_nonce,
GBytes *optional_additional_data, GError **error)
PV_NONNULL(1, 2);
/**
* att_extract_from_hdr:
*
* @se_hdr: binary SE guest header.
* @error: GError. *error will != NULL if error occurs.
*
* Verifies that SE header size and magic ,but no cryptographical verification.
* Then, find and extracts pld, ald, tld, and SE tagi and adds it to the context.
*
* Returns: new attestation measurement context.
*/
att_meas_ctx_t *att_extract_from_hdr(GBytes *se_hdr, GError **error) PV_NONNULL(1);
/** att_add_uid:
*
* @meas_ctx: measurement context.
* @config_uid: pointer to config UID. Must be `ATT_CONFIG_UID_SIZE` bytes long.
*
* Copies the config UID to the measurement context.
* Wrong size is considered as a Programming error.
*/
void att_add_uid(att_meas_ctx_t *meas_ctx, GBytes *config_uid) PV_NONNULL(1, 2);
/** att_verify_measurement:
*
* @calculated_measurement: measurement calculated by a trusted system
* @uvio_measurement: measurement generated by an UV
* @error: GError. *error will != NULL if error occurs.
*
* Returns: TRUE if measurements are identical, otherwise FALSE
*/
gboolean att_verify_measurement(const GBytes *calculated_measurement,
const GBytes *uvio_measurement, GError **error) PV_NONNULL(1, 2);
#define ATT_ERROR g_quark_from_static_string("pv-att_error-quark")
typedef enum att_error {
ATT_ERR_INVALID_HDR,
ATT_ERR_INVALID_USER_DATA,
ATT_ERR_MEASUREMENT_VERIFICATION_FAILED,
ATT_ERR_PHKH_NO_FIT_IN_USER_DATA,
ATT_ERR_PHKH_NO_MATCH,
} att_error_e;
#endif /* PVATTEST_ATTESTATION_H */

47
pvattest/src/common.c Normal file
View File

@@ -0,0 +1,47 @@
/*
* Common functions for pvattest.
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
/* Must be included before any other header */
#include "config.h"
#include <stdio.h>
#include <sys/stat.h>
#include "libpv/glib-helper.h"
#include "types.h"
#include "common.h"
gboolean wrapped_g_file_set_content(const char *filename, GBytes *bytes, mode_t mode,
GError **error)
{
const void *data;
size_t size;
gboolean rc;
data = g_bytes_get_data(bytes, &size);
rc = g_file_set_contents(filename, data, (ssize_t)size, error);
if (rc && mode != 0666)
chmod(filename, mode);
return rc;
}
GBytes *secure_gbytes_concat(GBytes *lh, GBytes *rh)
{
g_autoptr(GByteArray) lha = NULL;
if (!lh && !rh)
return NULL;
if (!lh)
return g_bytes_ref(rh);
if (!rh)
return g_bytes_ref(lh);
lha = g_bytes_unref_to_array(g_bytes_ref(lh));
g_byte_array_append(lha, g_bytes_get_data(rh, NULL), (guint)g_bytes_get_size(rh));
return pv_sec_gbytes_new(lha->data, lha->len);
}

37
pvattest/src/common.h Normal file
View File

@@ -0,0 +1,37 @@
/*
* Common functions for pvattest.
*
* IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#ifndef PVATTEST_COMMON_H
#define PVATTEST_COMMON_H
/* Must be included before any other header */
#include "config.h"
#include <glib/gi18n-lib.h>
#include <stdio.h>
#include "libpv/glib-helper.h"
#include "libpv/macros.h"
#include "lib/zt_common.h"
#include "types.h"
#define COPYRIGHT_NOTICE "Copyright IBM Corp. 2022"
#define AES_256_GCM_TAG_SIZE 16
gboolean wrapped_g_file_set_content(const char *filename, GBytes *bytes, mode_t mode,
GError **error);
/**
* just ref's up if one of them is NULL.
* If both NULL returns NULL.
* Otherwise returns lh ++ rh
*/
GBytes *secure_gbytes_concat(GBytes *lh, GBytes *rh);
#endif /* PVATTEST_COMMON_H */

31
pvattest/src/config.h Normal file
View File

@@ -0,0 +1,31 @@
/*
* Config file.
* Must be include before any other header.
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*
*/
#ifndef PVATTEST_CONFIG_H
#define PVATTEST_CONFIG_H
#define GETTEXT_PACKAGE "pvattest"
#ifdef __GNUC__
#ifdef __s390x__
#ifndef PVATTEST_NO_PERFORM
#define PVATTEST_COMPILE_PERFORM
#endif
#endif
#endif
#ifdef __clang__
#ifdef __zarch__
#ifndef PVATTEST_NO_PERFORM
#define PVATTEST_COMPILE_PERFORM
#endif
#endif
#endif
#endif /* PVATTEST_CONFIG_H */

View File

@@ -0,0 +1,480 @@
/*
* Functions for the pvattest exchange format to send attestation requests and responses between
* machines .
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
/* Must be included before any other header */
#include "config.h"
#include <stdio.h>
#include <string.h>
#include <errno.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <unistd.h>
#include "common.h"
#include "exchange_format.h"
#include "log.h"
struct exchange_shared_hdr {
be64_t magic;
be32_t version;
be32_t size;
} __packed;
/*
* If size == 0
* offset ignored.
* (part does not exist)
* if offset >0 and <0x50 -> invalid format
* if offset == 0 and size > 0 no data saved, however the request will need this amount of memory to
* succeed.
* Only makes sense for measurement and additional data. This however, is not enforced.
*/
struct entry {
be32_t size;
be32_t offset;
} __packed;
G_STATIC_ASSERT(sizeof(struct entry) == 8);
struct _exchange_format_v1_hdr {
be64_t magic;
be32_t version;
be32_t size;
uint64_t reserved;
struct entry serialized_arcb;
struct entry measurement;
struct entry additional_data;
struct entry user_data;
struct entry config_uid;
} __packed;
G_STATIC_ASSERT(sizeof(exchange_format_v1_hdr_t) == 0x40);
struct _exchange_format_ctx {
uint32_t version;
uint32_t req_meas_size;
uint32_t req_add_size;
GBytes *serialized_arcb;
GBytes *measurement;
GBytes *additional_data;
GBytes *user_data;
GBytes *config_uid;
};
/* Use a byte array to avoid any byteorder issues while checking */
#define PVATTEST_EXCHANGE_MAGIC 0x7076617474657374 /* pvattest */
static const uint8_t exchange_magic[] = { 0x70, 0x76, 0x61, 0x74, 0x74, 0x65, 0x73, 0x74 };
exchange_format_ctx_t *exchange_ctx_new(uint32_t version, GBytes *serialized_arcb,
uint32_t req_measurement_size, uint32_t req_additional_size,
GError **error)
{
g_autoptr(exchange_format_ctx_t) ctx = NULL;
pv_wrapped_g_assert(serialized_arcb);
if (version != PVATTEST_EXCHANGE_VERSION_1_00) {
g_set_error(error, EXCHANGE_FORMAT_ERROR, EXCHANGE_FORMAT_ERROR_UNSUPPORTED_VERSION,
_("'%d' unsupported version."), version);
return NULL;
}
ctx = g_malloc0(sizeof(*ctx));
ctx->version = version;
exchange_set_serialized_arcb(ctx, serialized_arcb);
ctx->req_meas_size = req_measurement_size;
ctx->req_add_size = req_additional_size;
return g_steal_pointer(&ctx);
}
static GBytes *get_content(GBytes *file_content, const struct entry *entry, const size_t max_size,
GError **error)
{
uint64_t size = GUINT32_FROM_BE(entry->size);
uint64_t offset = GUINT32_FROM_BE(entry->offset);
size_t file_size = 0;
const uint8_t *file_content_u8 = g_bytes_get_data(file_content, &file_size);
if (size == 0 || offset == 0)
return NULL;
if (offset < sizeof(exchange_format_v1_hdr_t) || offset + size > file_size ||
size > max_size) {
g_set_error(error, EXCHANGE_FORMAT_ERROR, EXCHANGE_FORMAT_ERROR_INVALID_FORMAT,
_("Input file is not in a valid format."));
return NULL;
}
return g_bytes_new(file_content_u8 + offset, size);
}
static gboolean check_format(const struct exchange_shared_hdr *hdr)
{
if (memcmp(exchange_magic, &hdr->magic, sizeof(exchange_magic)) == 0)
return TRUE;
return FALSE;
}
exchange_format_ctx_t *exchange_ctx_from_file(const char *filename, GError **error)
{
g_autoptr(exchange_format_ctx_t) ctx = g_malloc0(sizeof(*ctx));
const struct exchange_shared_hdr *hdr = NULL;
const exchange_format_v1_hdr_t *hdr_v1 = NULL;
g_autoptr(GBytes) file_content = NULL;
size_t config_uid_size = 0;
size_t file_size;
pv_wrapped_g_assert(filename);
file_content = pv_file_get_content_as_g_bytes(filename, error);
if (!file_content)
return NULL;
hdr = (const struct exchange_shared_hdr *)g_bytes_get_data(file_content, &file_size);
if (file_size < sizeof(*hdr) || !check_format(hdr)) {
g_set_error(error, EXCHANGE_FORMAT_ERROR, EXCHANGE_FORMAT_ERROR_INVALID_FORMAT,
_("'%s' is not in a valid format."), filename);
return NULL;
}
if (GUINT32_FROM_BE(hdr->version) != PVATTEST_EXCHANGE_VERSION_1_00) {
g_set_error(error, EXCHANGE_FORMAT_ERROR, EXCHANGE_FORMAT_ERROR_INVALID_FORMAT,
_("The version (%#x) of '%s' is not supported"),
GUINT32_FROM_BE(hdr->version), filename);
return NULL;
}
/* get the header */
if (file_size < sizeof(exchange_format_v1_hdr_t)) {
g_set_error(error, EXCHANGE_FORMAT_ERROR, EXCHANGE_FORMAT_ERROR_INVALID_FORMAT,
_("'%s' is not in a valid format."), filename);
return NULL;
}
hdr_v1 = (const exchange_format_v1_hdr_t *)hdr;
/* get entries if present */
ctx->serialized_arcb =
get_content(file_content, &hdr_v1->serialized_arcb, PVATTEST_ARCB_MAX_SIZE, error);
if (*error)
return NULL;
ctx->measurement = get_content(file_content, &hdr_v1->measurement,
PVATTEST_MEASUREMENT_MAX_SIZE, error);
if (*error)
return NULL;
ctx->additional_data = get_content(file_content, &hdr_v1->additional_data,
PVATTEST_ADDITIONAL_MAX_SIZE, error);
if (*error)
return NULL;
ctx->user_data =
get_content(file_content, &hdr_v1->user_data, PVATTEST_USER_DATA_MAX_SIZE, error);
if (*error)
return NULL;
ctx->config_uid = get_content(file_content, &hdr_v1->config_uid, PVATTEST_UID_SIZE, error);
if (*error)
return NULL;
if (ctx->config_uid)
config_uid_size = g_bytes_get_size(ctx->config_uid);
if (config_uid_size != PVATTEST_UID_SIZE && config_uid_size != 0) {
g_set_error(error, EXCHANGE_FORMAT_ERROR, EXCHANGE_FORMAT_ERROR_INVALID_FORMAT,
_("'%s' is not in a valid format."), filename);
return NULL;
}
ctx->req_meas_size = GUINT32_FROM_BE(hdr_v1->measurement.size);
ctx->req_add_size = GUINT32_FROM_BE(hdr_v1->additional_data.size);
ctx->version = GUINT32_TO_BE(hdr->version);
return g_steal_pointer(&ctx);
}
void clear_free_exchange_ctx(exchange_format_ctx_t *ctx)
{
if (!ctx)
return;
if (ctx->serialized_arcb)
g_bytes_unref(ctx->serialized_arcb);
if (ctx->measurement)
g_bytes_unref(ctx->measurement);
if (ctx->additional_data)
g_bytes_unref(ctx->additional_data);
if (ctx->user_data)
g_bytes_unref(ctx->user_data);
if (ctx->config_uid)
g_bytes_unref(ctx->config_uid);
g_free(ctx);
}
void exchange_set_serialized_arcb(exchange_format_ctx_t *ctx, GBytes *serialized_arcb)
{
pv_wrapped_g_assert(ctx);
pv_wrapped_g_assert(serialized_arcb);
g_bytes_ref(serialized_arcb);
g_bytes_unref(ctx->serialized_arcb);
ctx->serialized_arcb = serialized_arcb;
}
void exchange_set_measurement(exchange_format_ctx_t *ctx, GBytes *measurement)
{
pv_wrapped_g_assert(ctx);
pv_wrapped_g_assert(measurement);
g_bytes_ref(measurement);
g_bytes_unref(ctx->measurement);
ctx->measurement = measurement;
}
void exchange_set_additional_data(exchange_format_ctx_t *ctx, GBytes *additional_data)
{
pv_wrapped_g_assert(ctx);
pv_wrapped_g_assert(additional_data);
g_bytes_ref(additional_data);
g_bytes_unref(ctx->additional_data);
ctx->additional_data = additional_data;
}
void exchange_set_user_data(exchange_format_ctx_t *ctx, GBytes *user_data)
{
pv_wrapped_g_assert(ctx);
pv_wrapped_g_assert(user_data);
g_bytes_ref(user_data);
g_bytes_unref(ctx->user_data);
ctx->user_data = user_data;
}
void exchange_set_config_uid(exchange_format_ctx_t *ctx, GBytes *config_uid)
{
pv_wrapped_g_assert(ctx);
pv_wrapped_g_assert(config_uid);
g_bytes_ref(config_uid);
g_bytes_unref(ctx->config_uid);
ctx->config_uid = config_uid;
}
static GBytes *gbytes_ref0(GBytes *bytes)
{
if (!bytes)
return NULL;
return g_bytes_ref(bytes);
}
GBytes *exchange_get_serialized_arcb(const exchange_format_ctx_t *ctx)
{
pv_wrapped_g_assert(ctx);
return gbytes_ref0(ctx->serialized_arcb);
}
GBytes *exchange_get_measurement(const exchange_format_ctx_t *ctx)
{
pv_wrapped_g_assert(ctx);
return gbytes_ref0(ctx->measurement);
}
GBytes *exchange_get_additional_data(const exchange_format_ctx_t *ctx)
{
pv_wrapped_g_assert(ctx);
return gbytes_ref0(ctx->additional_data);
}
GBytes *exchange_get_user_data(const exchange_format_ctx_t *ctx)
{
pv_wrapped_g_assert(ctx);
return gbytes_ref0(ctx->user_data);
}
GBytes *exchange_get_config_uid(const exchange_format_ctx_t *ctx)
{
pv_wrapped_g_assert(ctx);
return gbytes_ref0(ctx->config_uid);
}
uint32_t exchange_get_requested_measurement_size(const exchange_format_ctx_t *ctx)
{
pv_wrapped_g_assert(ctx);
return ctx->req_meas_size;
}
uint32_t exchange_get_requested_additional_data_size(const exchange_format_ctx_t *ctx)
{
pv_wrapped_g_assert(ctx);
return ctx->req_add_size;
}
static struct entry add_g_bytes(GBytes *bytes, FILE *file, GError **error)
{
struct entry result = {};
long offset;
size_t size;
const void *data = g_bytes_get_data(bytes, &size);
g_assert(size <= G_MAXUINT32);
offset = pv_file_tell(file, error);
g_assert(offset <= G_MAXUINT32);
if (offset < 0)
return result;
result.offset = GUINT32_TO_BE((uint32_t)offset);
result.size = GUINT32_TO_BE((uint32_t)size);
pv_file_write(file, data, size, error);
return result;
}
int exchange_write_to_file(const exchange_format_ctx_t *ctx, const char *filename, GError **error)
{
exchange_format_v1_hdr_t hdr = {
.magic = GUINT64_TO_BE(PVATTEST_EXCHANGE_MAGIC),
.version = GUINT32_TO_BE(ctx->version),
};
size_t file_size = sizeof(hdr);
g_autoptr(FILE) file = NULL;
struct stat file_stat;
long actual_file_size;
size_t tmp_size;
pv_wrapped_g_assert(ctx);
pv_wrapped_g_assert(filename);
file = pv_file_open(filename, "w", error);
if (!file)
return -1;
if (fstat(fileno(file), &file_stat) != 0 || !S_ISREG(file_stat.st_mode)) {
g_set_error(error, EXCHANGE_FORMAT_ERROR,
EXCHANGE_FORMAT_ERROR_UNSUPPORTED_FILE_TYPE,
"Only regular files are supported: '%s'", filename);
return -1;
}
if (pv_file_seek(file, sizeof(exchange_format_v1_hdr_t), SEEK_SET, error))
return -1;
if (ctx->serialized_arcb) {
hdr.serialized_arcb = add_g_bytes(ctx->serialized_arcb, file, error);
if (*error)
return -1;
file_size += g_bytes_get_size(ctx->serialized_arcb);
}
if (ctx->measurement) {
hdr.measurement = add_g_bytes(ctx->measurement, file, error);
if (*error)
return -1;
file_size += g_bytes_get_size(ctx->measurement);
} else {
hdr.measurement.size = GUINT32_TO_BE(ctx->req_meas_size);
}
if (ctx->additional_data) {
hdr.additional_data = add_g_bytes(ctx->additional_data, file, error);
if (*error)
return -1;
file_size += g_bytes_get_size(ctx->additional_data);
} else {
hdr.additional_data.size = GUINT32_TO_BE(ctx->req_add_size);
}
if (ctx->user_data) {
tmp_size = g_bytes_get_size(ctx->user_data);
g_assert(tmp_size <= PVATTEST_USER_DATA_MAX_SIZE);
tmp_size = MIN(tmp_size, PVATTEST_USER_DATA_MAX_SIZE); /* should be a noop */
hdr.user_data = add_g_bytes(ctx->user_data, file, error);
if (*error)
return -1;
file_size += g_bytes_get_size(ctx->user_data);
}
if (ctx->config_uid) {
tmp_size = g_bytes_get_size(ctx->config_uid);
g_assert(tmp_size == PVATTEST_UID_SIZE);
tmp_size = MIN(tmp_size, PVATTEST_UID_SIZE); /* should be a noop */
hdr.config_uid = add_g_bytes(ctx->config_uid, file, error);
if (*error)
return -1;
file_size += g_bytes_get_size(ctx->config_uid);
}
/*
* This case should never happen. It could be seen as a programming error as:
* ARCB is restricted by kernel (and this tool) to be max 1M, Additional+meas to max 8pages
* userdata to 256B and config uid to 16b this is way less than 4G.
*
* However, lets be conservative and trow an error instead of an assertion.
*/
if (file_size > UINT32_MAX) {
g_set_error(
error, EXCHANGE_FORMAT_ERROR, EXCHANGE_FORMAT_ERROR_INVALID_FORMAT,
"The exchange file format cannot handle this much data in one blob. (%#lx bytes)",
file_size);
return -1;
}
hdr.size = GUINT32_TO_BE((uint32_t)file_size);
if (pv_file_seek(file, 0, SEEK_SET, error) != 0)
return -1;
if (sizeof(hdr) != pv_file_write(file, &hdr, sizeof(hdr), error))
return -1;
if (pv_file_seek(file, 0, SEEK_END, error) != 0)
return -1;
actual_file_size = pv_file_tell(file, error);
if (actual_file_size < 0)
return -1;
if (actual_file_size != (uint32_t)file_size) {
g_set_error(
error, EXCHANGE_FORMAT_ERROR, EXCHANGE_FORMAT_ERROR_WRITE,
"The exchange file size doesn't match the expectations: %ld bytes != %lu bytes",
actual_file_size, file_size);
return -1;
}
return 0;
}
static void print_entry(const char *name, GBytes *data, const gboolean print_data, FILE *stream)
{
if (!data)
return;
fprintf(stream, _("%s (%#lx bytes)"), name, g_bytes_get_size(data));
if (print_data) {
fprintf(stream, ":\n");
printf_hexdump(g_bytes_get_data(data, NULL), g_bytes_get_size(data), 16, " ",
stream);
}
fprintf(stream, "\n");
}
void exchange_info_print(const exchange_format_ctx_t *ctx, const gboolean print_data, FILE *stream)
{
pv_wrapped_g_assert(ctx);
pv_wrapped_g_assert(stream);
fprintf(stream, _("Version: %#x\n"), ctx->version);
fprintf(stream, _("Sections:\n"));
print_entry(_(" ARCB"), ctx->serialized_arcb, print_data, stream);
print_entry(_(" Measurement"), ctx->measurement, print_data, stream);
print_entry(_(" Additional Data"), ctx->additional_data, print_data, stream);
print_entry(_(" User Data"), ctx->user_data, print_data, stream);
print_entry(_(" Config UID"), ctx->config_uid, print_data, stream);
if (!ctx->measurement)
fprintf(stream, _("Required measurement size: %#x\n"), ctx->req_meas_size);
if (!ctx->additional_data)
fprintf(stream, _("Required additional data size: %#x\n"), ctx->req_add_size);
}

View File

@@ -0,0 +1,166 @@
/*
* Definitions for the pvattest exchange format to send attestation requests and responses between
* machines. The "exchange format" is a simple file format to send labeled binary blobs between
* pvattest instances on different machines. All sizes, etc are in big endian.
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#ifndef PVATTEST_EXCHANGE_FORMAT_H
#define PVATTEST_EXCHANGE_FORMAT_H
/* Must be included before any other header */
#include "config.h"
#include "libpv/glib-helper.h"
#include "types.h"
#include "common.h"
/* Similar to linux/arch/s390x/include/uapi/uvdevice.h as this part needs to be
* architecture independent.
*/
#define PVATTEST_UID_SIZE 0x10UL
#define PVATTEST_USER_DATA_MAX_SIZE 0x100UL
#define PVATTEST_ARCB_MAX_SIZE 0x100000
#define PVATTEST_MEASUREMENT_MAX_SIZE 0x8000
#define PVATTEST_ADDITIONAL_MAX_SIZE 0x8000
#define PVATTEST_EXCHANGE_V_INVALID 0
#define PVATTEST_EXCHANGE_VERSION_1_00 0x0100
typedef struct _exchange_format_v1_hdr exchange_format_v1_hdr_t;
typedef struct _exchange_format_ctx exchange_format_ctx_t;
/**
* exchange_ctx_new:
*
* @version: Format version. Currently, only version 1 supported.
* @serialized_arcb: ARCB as #GBytes
* @req_measurement_size: Measurement size the given ARCB needs.
* @req_measurement_size: Additional Data size the given ARCB needs.
* @error: GError. *error will != NULL if error occurs.
*
* Returns: (nullable) (transfer full): new, empty exchange format context
*
*/
exchange_format_ctx_t *exchange_ctx_new(const uint32_t version, GBytes *serialized_arcb,
const uint32_t req_measurement_size,
const uint32_t req_additional_size, GError **error)
PV_NONNULL(2);
/**
* exchange_ctx_from_file:
*
* @filename: name of the file to be loaded
* @error: GError. *error will != NULL if error occurs.
*
* Loads all blobs from file and caches them in the context structure.
*
* Returns: (nullable) (transfer full): exchange format context filled with data from file
*
*/
exchange_format_ctx_t *exchange_ctx_from_file(const char *filename, GError **error) PV_NONNULL(1);
void clear_free_exchange_ctx(exchange_format_ctx_t *ctx);
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(exchange_format_ctx_t, clear_free_exchange_ctx)
/**
* exchange_set_serialized_arcb:
*
* @ctx: exchange format context
* @serialized_arcb: blob to add.
*
* Adds blob to the exchange format. Unreferences old data if already set.
*/
void exchange_set_serialized_arcb(exchange_format_ctx_t *ctx, GBytes *serialized_arcb)
PV_NONNULL(1, 2);
/**
* exchange_set_measurement:
*
* @ctx: exchange format context
* @measurement: blob to add.
*
* Adds blob to the exchange format. Unreferences old data if already set.
*/
void exchange_set_measurement(exchange_format_ctx_t *ctx, GBytes *measurement) PV_NONNULL(1, 2);
/**
* exchange_set_additional_data:
*
* @ctx: exchange format context
* @additional_data: blob to add.
*
* Adds blob to the exchange format. Unreferences old data if already set.
*/
void exchange_set_additional_data(exchange_format_ctx_t *ctx, GBytes *additional_data)
PV_NONNULL(1, 2);
/**
* exchange_set_user_data:
*
* @ctx: exchange format context
* @user_data: blob to add.
*
* Adds blob to the exchange format. Unreferences old data if already set.
*/
void exchange_set_user_data(exchange_format_ctx_t *ctx, GBytes *user_data) PV_NONNULL(1, 2);
/**
* exchange_set_config_uid:
*
* @ctx: exchange format context
* @config_uid: blob to add.
*
* Adds blob to the exchange format. Unreferences old data if already set.
*/
void exchange_set_config_uid(exchange_format_ctx_t *ctx, GBytes *config_uid) PV_NONNULL(1, 2);
GBytes *exchange_get_serialized_arcb(const exchange_format_ctx_t *ctx) PV_NONNULL(1);
GBytes *exchange_get_measurement(const exchange_format_ctx_t *ctx) PV_NONNULL(1);
GBytes *exchange_get_additional_data(const exchange_format_ctx_t *ctx) PV_NONNULL(1);
GBytes *exchange_get_user_data(const exchange_format_ctx_t *ctx) PV_NONNULL(1);
GBytes *exchange_get_config_uid(const exchange_format_ctx_t *ctx) PV_NONNULL(1);
uint32_t exchange_get_requested_measurement_size(const exchange_format_ctx_t *ctx) PV_NONNULL(1);
uint32_t exchange_get_requested_additional_data_size(const exchange_format_ctx_t *ctx)
PV_NONNULL(1);
/**
* exchange_write_to_file:
*
* @ctx: exchange format context
* @filename: name of the file to be loaded
* @error: GError. *error will != NULL if error occours.
*
* Takes all Data in the context and writes them into a file.
* Places the exchange format header before the data.
*
* Returns: 0 in case of success, -1 otherwise.
*/
int exchange_write_to_file(const exchange_format_ctx_t *ctx, const char *filename, GError **error)
PV_NONNULL(1, 2);
/**
* exchange_info_print:
*
* @ctx: exchange format context
* @print_data: TRUE: print present data + label names
* FALSE: just print label names of present data
* @stream: FILE* stream to print data
*
* Prints the content of @ctx to @stream.
*/
void exchange_info_print(const exchange_format_ctx_t *ctx, const gboolean print_data, FILE *stream)
PV_NONNULL(1, 3);
#define EXCHANGE_FORMAT_ERROR g_quark_from_static_string("pv-exchange-format_error-quark")
typedef enum {
EXCHANGE_FORMAT_ERROR_INVALID_FORMAT,
EXCHANGE_FORMAT_ERROR_UNSUPPORTED_VERSION,
EXCHANGE_FORMAT_ERROR_WRITE,
EXCHANGE_FORMAT_ERROR_UNSUPPORTED_FILE_TYPE,
} exchange_error_e;
#endif /* PVATTEST_EXCHANGE_FORMAT_H */

181
pvattest/src/log.c Normal file
View File

@@ -0,0 +1,181 @@
/*
* Functions used for logging.
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
/* Must be included before any other header */
#include "config.h"
#include <stdio.h>
#include <stdarg.h>
#include "types.h"
#include "common.h"
#include "log.h"
void pvattest_log_increase_log_lvl(int *log_lvl)
{
if (*log_lvl >= PVATTEST_LOG_LVL_MAX)
return;
*log_lvl = *log_lvl << 1;
}
void pvattest_log_error(const char *format, ...)
{
va_list argp;
va_start(argp, format);
g_logv(NULL, PVATTEST_LOG_LVL_ERROR, format, argp);
va_end(argp);
}
void pvattest_log_warning(const char *format, ...)
{
va_list argp;
va_start(argp, format);
g_logv(NULL, PVATTEST_LOG_LVL_WARNING, format, argp);
va_end(argp);
}
void pvattest_log_info(const char *format, ...)
{
va_list argp;
va_start(argp, format);
g_logv(NULL, PVATTEST_LOG_LVL_INFO, format, argp);
va_end(argp);
}
void pvattest_log_debug(const char *format, ...)
{
va_list argp;
va_start(argp, format);
g_logv(NULL, PVATTEST_LOG_LVL_DEBUG, format, argp);
va_end(argp);
}
static void _log_print(FILE *stream, const char *prefix, const char *message, const char *postfix)
{
g_autofree char *prefix_empty = NULL, *new_msg = NULL;
size_t prefix_len = strlen(prefix);
char **message_v;
if (!prefix || prefix_len == 0) {
printf("%s%s", message, postfix);
return;
}
message_v = g_strsplit(message, "\n", 0);
prefix_empty = g_malloc0(prefix_len + 2);
snprintf(prefix_empty, prefix_len + 2, "\n%*c\b", (int)prefix_len, ' ');
new_msg = g_strjoinv(prefix_empty, message_v);
fprintf(stream, "%s%s%s", prefix, new_msg, postfix);
g_strfreev(message_v);
}
static void _log_logger(GLogLevelFlags level, const char *message, int log_level,
gboolean use_prefix, const char *postfix)
{
const char *prefix = "";
/* filter out messages depending on debugging level */
if ((level & PVATTEST_LOG_LVL_DEBUG) && log_level < PVATTEST_LOG_LVL_DEBUG)
return;
if ((level & PVATTEST_LOG_LVL_INFO) && log_level < PVATTEST_LOG_LVL_INFO)
return;
if (use_prefix && level & (G_LOG_LEVEL_WARNING | PVATTEST_LOG_LVL_WARNING))
prefix = _("WARNING: ");
if (use_prefix && level & (G_LOG_LEVEL_ERROR | PVATTEST_LOG_LVL_ERROR))
prefix = _("ERROR: ");
if (use_prefix && level & (G_LOG_LEVEL_DEBUG | PVATTEST_LOG_LVL_DEBUG))
prefix = _("DEBUG: ");
if (level & (G_LOG_LEVEL_WARNING | G_LOG_LEVEL_ERROR | PVATTEST_LOG_LVL_WARNING |
PVATTEST_LOG_LVL_ERROR))
_log_print(stderr, prefix, message, postfix);
else
_log_print(stdout, prefix, message, postfix);
}
/**
* prefixes type. and adds a "\n" ad the end.
*/
void pvattest_log_default_logger(const char *log_domain G_GNUC_UNUSED, GLogLevelFlags level,
const char *message, void *user_data)
{
int log_level = *(int *)user_data;
_log_logger(level, message, log_level, TRUE, "\n");
}
/*
* writes message as it is if log level is high enough.
*/
void pvattest_log_plain_logger(const char *log_domain G_GNUC_UNUSED, GLogLevelFlags level,
const char *message, void *user_data)
{
int log_level = *(int *)user_data;
_log_logger(level, message, log_level, FALSE, "");
}
void hexdump(const void *data, size_t size, size_t len, const char *prefix, GLogLevelFlags log_lvl)
{
const uint8_t *data_b = data;
pv_wrapped_g_assert(data);
g_log(PVATTEST_HEXDUMP_LOG_DOMAIN, log_lvl, "%s0x0000 ", prefix);
for (size_t i = 0; i < size; i++) {
g_log(PVATTEST_HEXDUMP_LOG_DOMAIN, log_lvl, "%02x", data_b[i]);
if (i % 2 == 1)
g_log(PVATTEST_HEXDUMP_LOG_DOMAIN, log_lvl, " ");
if (i == size - 1)
break;
if (i % len == len - 1)
g_log(PVATTEST_HEXDUMP_LOG_DOMAIN, log_lvl, "\n%s0x%04lx ", prefix, i + 1);
}
g_log(PVATTEST_HEXDUMP_LOG_DOMAIN, log_lvl, "\n");
}
void printf_hexdump(const void *data, size_t size, size_t len, const char *prefix, FILE *stream)
{
const uint8_t *data_b = data;
pv_wrapped_g_assert(data);
pv_wrapped_g_assert(stream);
fprintf(stream, "%s0x0000 ", prefix);
for (size_t i = 0; i < size; i++) {
fprintf(stream, "%02x", data_b[i]);
if (i % 2 == 1)
fprintf(stream, " ");
if (i == size - 1)
break;
if (i % len == len - 1)
fprintf(stream, "\n%s0x%04lx ", prefix, i + 1);
}
fprintf(stream, "\n");
}
void pvattest_log_GError(const char *info, GError *error)
{
pv_wrapped_g_assert(info);
if (!error)
return;
pvattest_log_error("%s:\n%s", info, error->message);
}

67
pvattest/src/log.h Normal file
View File

@@ -0,0 +1,67 @@
/*
* Definitions used for logging.
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#ifndef PVATTEST_LOG_H
#define PVATTEST_LOG_H
/* Must be included before any other header */
#include "config.h"
#include "libpv/glib-helper.h"
#include "libpv/macros.h"
#define PVATTEST_LOG_LVL_TOOL_ALL (1 << (G_LOG_LEVEL_USER_SHIFT))
#define PVATTEST_LOG_LVL_ERROR (1 << (G_LOG_LEVEL_USER_SHIFT))
#define PVATTEST_LOG_LVL_WARNING (1 << (G_LOG_LEVEL_USER_SHIFT + 1))
#define PVATTEST_LOG_LVL_INFO (1 << (G_LOG_LEVEL_USER_SHIFT + 2))
#define PVATTEST_LOG_LVL_DEBUG (1 << (G_LOG_LEVEL_USER_SHIFT + 3))
#define PVATTEST_LOG_LVL_DEFAULT PVATTEST_LOG_LVL_WARNING
#define PVATTEST_LOG_LVL_MAX PVATTEST_LOG_LVL_DEBUG
#define PVATTEST_HEXDUMP_LOG_DOMAIN "pvattest_hdump"
void pvattest_log_increase_log_lvl(int *log_lvl);
void pvattest_log_error(const char *format, ...);
void pvattest_log_warning(const char *format, ...);
void pvattest_log_info(const char *format, ...);
void pvattest_log_debug(const char *format, ...);
/** pvattest_log_default_logger:
*
* A #GLogFunc implementation.
* Prefixes log level and adds a "\n" ad the end.
*/
void pvattest_log_default_logger(const char *log_domain, GLogLevelFlags level, const char *message,
void *user_data);
/* pvattest_log_plain_logger:
*
* A #GLogFunc implementation.
* Writes message as it is if log level is high enough.
*/
void pvattest_log_plain_logger(const char *log_domain, GLogLevelFlags level, const char *message,
void *user_data);
#define dhexdump(v, s) \
{ \
pvattest_log_debug("%s (%li byte):", #v, s); \
hexdump(v, s, 16L, " ", PVATTEST_LOG_LVL_DEBUG); \
g_log(PVATTEST_HEXDUMP_LOG_DOMAIN, PVATTEST_LOG_LVL_DEBUG, "\n"); \
}
#define gbhexdump(v) \
{ \
pvattest_log_debug("%s:(%li byte):", #v, g_bytes_get_size(v)); \
hexdump(g_bytes_get_data(v, NULL), g_bytes_get_size(v), 16L, " ", \
PVATTEST_LOG_LVL_DEBUG); \
g_log(PVATTEST_HEXDUMP_LOG_DOMAIN, PVATTEST_LOG_LVL_DEBUG, "\n"); \
}
void hexdump(const void *data, size_t size, size_t len, const char *prefix, GLogLevelFlags log_lvl)
PV_NONNULL(1);
void printf_hexdump(const void *data, size_t size, size_t len, const char *prefix, FILE *stream)
PV_NONNULL(1, 5);
void pvattest_log_GError(const char *info, GError *error) PV_NONNULL(1);
#endif /* PVATTEST_LOG_H */

392
pvattest/src/pvattest.c Normal file
View File

@@ -0,0 +1,392 @@
/*
* Entry point for the pvattest tool.
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
/* Must be included before any other header */
#include "config.h"
#include <stdio.h>
#include <unistd.h>
#include <openssl/evp.h>
#include "libpv/crypto.h"
#include "libpv/cert.h"
#include "uvio.h"
#include "common.h"
#include "attestation.h"
#include "arcb.h"
#include "argparse.h"
#include "exchange_format.h"
#include "log.h"
#define PVATTEST_NID NID_secp521r1
#define PVATTEST_UV_PATH "/dev/uv"
#define PVATTEST_EXIT_MEASURE_NOT_VERIFIED 2
enum pvattest_error {
PVATTEST_ERROR_INVAL_ATT_RESULT,
};
static arcb_v1_t *create_arcb(char **host_key_paths, const gboolean use_nonce,
const gboolean phkh_img, const gboolean phkh_att,
const uint64_t user_paf, GError **error)
{
g_autoptr(GBytes) arpk = NULL, meas_key = NULL, nonce = NULL, iv = NULL;
g_autoslist(PvX509WithPath) host_keys_with_path = NULL;
g_autoslist(EVP_PKEY) evp_host_keys = NULL;
const uint32_t mai = MAI_HMAC_SHA512;
g_autoptr(EVP_PKEY) evp_cpk = NULL;
g_autoptr(arcb_v1_t) arcb = NULL;
uint64_t paf = user_paf;
g_assert(host_key_paths);
arpk = pv_generate_key(EVP_aes_256_gcm(), error);
if (!arpk)
return NULL;
iv = pv_generate_iv(EVP_aes_256_gcm(), error);
if (!iv)
return NULL;
evp_cpk = pv_generate_ec_key(PVATTEST_NID, error);
if (!evp_cpk)
return NULL;
meas_key = pv_generate_rand_data(HMAC_SHA512_KEY_SIZE, error);
if (!meas_key)
return NULL;
if (phkh_img)
paf |= ARCB_V1_PAF_AAD_PHKH_HEADER;
if (phkh_att)
paf |= ARCB_V1_PAF_AAD_PHKH_ATTEST;
arcb = arcb_v1_new(arpk, iv, mai, evp_cpk, meas_key, paf, error);
if (!arcb)
return NULL;
if (use_nonce) {
nonce = pv_generate_rand_data(ARCB_V1_NONCE_SIZE, error);
if (!nonce)
return NULL;
arcb_v1_set_nonce(arcb, nonce);
}
host_keys_with_path = pv_load_certificates(host_key_paths, error);
if (!host_keys_with_path)
return NULL;
/* Extract EVP_PKEY structures and verify that the correct elliptic
* curve is used.
*/
evp_host_keys = pv_get_ec_pubkeys(host_keys_with_path, PVATTEST_NID, error);
if (!evp_host_keys)
return NULL;
for (GSList *iter = evp_host_keys; iter; iter = iter->next) {
EVP_PKEY *host_key = iter->data;
if (arcb_v1_add_key_slot(arcb, host_key, error) < 0)
return NULL;
}
return g_steal_pointer(&arcb);
}
#define __PVATTEST_CREATE_ERROR_MSG _("Creating the attestation request failed")
static int do_create(const pvattest_create_config_t *create_config)
{
g_autoptr(exchange_format_ctx_t) output_ctx = NULL;
uint32_t measurement_size, additional_data_size;
g_autoptr(GBytes) serialized_arcb = NULL;
g_autoptr(arcb_v1_t) arcb = NULL;
g_autoptr(GError) error = NULL;
g_autoptr(GBytes) arpk = NULL;
if (!create_config->use_nonce)
pvattest_log_warning(_("No nonce used. (Experimental setting)"));
if (create_config->no_verify) {
pvattest_log_warning(_("Host-key document verification is disabled.\n"
"The attestation result could be compromised!"));
pvattest_log_debug(_("Verification skipped."));
} else {
if (pv_verify_host_key_docs_by_path(
create_config->host_key_document_paths, create_config->root_ca_path,
create_config->crl_paths, create_config->certificate_paths,
create_config->online, &error) < 0)
goto err_exit;
pvattest_log_debug(_("Verification passed."));
}
/* build attestation request */
arcb = create_arcb(create_config->host_key_document_paths, create_config->use_nonce,
create_config->phkh_img, create_config->phkh_att, create_config->paf,
&error);
if (!arcb)
goto err_exit;
additional_data_size = arcb_v1_get_required_additional_size(arcb);
if (create_config->x_aad_size >= 0) {
g_assert_cmpint(create_config->x_aad_size, <=, UINT32_MAX);
additional_data_size = (uint32_t)create_config->x_aad_size;
}
measurement_size = arcb_v1_get_required_measurement_size(arcb, &error);
if (error)
goto err_exit;
serialized_arcb = arcb_v1_serialize(arcb, &error);
if (!serialized_arcb)
goto err_exit;
/* write attestation request data to file */
output_ctx = exchange_ctx_new(PVATTEST_EXCHANGE_VERSION_1_00, serialized_arcb,
measurement_size, additional_data_size, &error);
if (!output_ctx)
goto err_exit;
if (exchange_write_to_file(output_ctx, create_config->output_path, &error) < 0)
goto err_exit;
pvattest_log_debug(_("ARCB written to file."));
/* write attestation request protection key to file */
arpk = arcb_v1_get_arp_key(arcb);
wrapped_g_file_set_content(create_config->arp_key_out_path, arpk, 0600, &error);
if (error)
goto err_exit;
pvattest_log_debug(_("ARPK written to file."));
return EXIT_SUCCESS;
err_exit:
pvattest_log_GError(__PVATTEST_CREATE_ERROR_MSG, error);
return EXIT_FAILURE;
}
#ifdef PVATTEST_COMPILE_PERFORM
#define __PVATTEST_MEASURE_ERROR_MSG _("Performing the attestation measurement failed")
static int do_perform(pvattest_perform_config_t *perform_config)
{
g_autoptr(GBytes) serialized_arcb = NULL, user_data = NULL, measurement = NULL,
additional_data = NULL, config_uid = NULL;
size_t uv_measurement_data_size, uv_addidtional_data_size;
g_autoptr(exchange_format_ctx_t) exchange_ctx = NULL;
uint32_t measurement_size, additional_data_size;
g_autoptr(uvio_attest_t) uvio_attest = NULL;
g_autoptr(GError) error = NULL;
be16_t uv_rc;
int uv_fd;
exchange_ctx = exchange_ctx_from_file(perform_config->input_path, &error);
if (!exchange_ctx)
goto err_exit;
serialized_arcb = exchange_get_serialized_arcb(exchange_ctx);
if (!serialized_arcb) {
g_set_error(&error, PVATTEST_ERROR, ARCB_ERR_INVALID_ARCB,
_("The input does not provide an attestation request."));
goto err_exit;
}
measurement_size = exchange_get_requested_measurement_size(exchange_ctx);
additional_data_size = exchange_get_requested_additional_data_size(exchange_ctx);
pvattest_log_debug(_("Input data loaded."));
if (perform_config->user_data_path) {
user_data = pv_file_get_content_as_g_bytes(perform_config->user_data_path, &error);
if (!user_data)
goto err_exit;
pvattest_log_debug(_("Added user data from '%s'"), perform_config->user_data_path);
}
uvio_attest = build_attestation_v1_ioctl(serialized_arcb, user_data, measurement_size,
additional_data_size, &error);
if (!uvio_attest)
goto err_exit;
pvattest_log_debug(_("attestation context generated."));
/* execute attestation */
uv_fd = uvio_open(PVATTEST_UV_PATH, &error);
if (uv_fd < 0)
goto err_exit;
uv_rc = uvio_ioctl_attest(uv_fd, uvio_attest, &error);
close(uv_fd);
if (uv_rc != UVC_EXECUTED)
goto err_exit;
pvattest_log_debug(_("attestation measurement successful. rc = %#x"), uv_rc);
/* write to file */
measurement = uvio_get_measurement(uvio_attest);
additional_data = uvio_get_additional_data(uvio_attest);
config_uid = uvio_get_config_uid(uvio_attest);
uv_measurement_data_size = measurement == NULL ? 0 : g_bytes_get_size(measurement);
if (uv_measurement_data_size != measurement_size) {
g_set_error(&error, PVATTEST_ERROR, PVATTEST_ERROR_INVAL_ATT_RESULT,
"The measurement size returned by Ultravisor is not as expected.");
goto err_exit;
}
uv_addidtional_data_size = additional_data == NULL ? 0 : g_bytes_get_size(additional_data);
if (uv_addidtional_data_size != additional_data_size) {
g_set_error(&error, PVATTEST_ERROR, PVATTEST_ERROR_INVAL_ATT_RESULT,
"The additional data size returned by Ultravisor is not as expected.");
goto err_exit;
}
exchange_set_measurement(exchange_ctx, measurement);
if (additional_data)
exchange_set_additional_data(exchange_ctx, additional_data);
exchange_set_config_uid(exchange_ctx, config_uid);
if (user_data)
exchange_set_user_data(exchange_ctx, user_data);
if (exchange_write_to_file(exchange_ctx, perform_config->output_path, &error) < 0)
goto err_exit;
pvattest_log_debug(_("Output written to file."));
return EXIT_SUCCESS;
err_exit:
pvattest_log_GError(__PVATTEST_MEASURE_ERROR_MSG, error);
return EXIT_FAILURE;
}
#endif /* PVATTEST_COMPILE_PERFORM */
#define __PVATTEST_VERIFY_ERROR_MSG _("Attestation measurement verification failed")
static int do_verify(pvattest_verify_config_t *verify_config)
{
g_autoptr(GBytes) user_data = NULL, uv_measurement = NULL, additional_data = NULL,
image_hdr = NULL, calc_measurement = NULL, config_uid = NULL,
meas_key = NULL, arp_key = NULL, nonce = NULL, serialized_arcb = NULL;
g_autofree att_meas_ctx_t *measurement_hdr = NULL;
g_autoptr(exchange_format_ctx_t) input_ctx = NULL;
g_autoptr(GError) error = NULL;
gboolean rc;
image_hdr = pv_file_get_content_as_g_bytes(verify_config->hdr_path, &error);
if (!image_hdr)
goto err_exit;
measurement_hdr = att_extract_from_hdr(image_hdr, &error);
if (!measurement_hdr)
goto err_exit;
pvattest_log_debug(_("Image header loaded."));
input_ctx = exchange_ctx_from_file(verify_config->input_path, &error);
if (!input_ctx)
goto err_exit;
config_uid = exchange_get_config_uid(input_ctx);
uv_measurement = exchange_get_measurement(input_ctx);
user_data = exchange_get_user_data(input_ctx);
additional_data = exchange_get_additional_data(input_ctx);
serialized_arcb = exchange_get_serialized_arcb(input_ctx);
if (!uv_measurement || !serialized_arcb) {
g_set_error(&error, PVATTEST_ERROR, PVATTEST_SUBC_INVALID,
_("Input data has no measurement"));
goto err_exit;
}
pvattest_log_debug(_("Input data loaded."));
att_add_uid(measurement_hdr, config_uid);
arp_key = pv_file_get_content_as_g_bytes(verify_config->arp_key_in_path, &error);
if (!arp_key)
goto err_exit;
pvattest_log_debug(_("ARPK loaded."));
rc = arcb_v1_verify_serialized_arcb(serialized_arcb, arp_key, &meas_key, &nonce, &error);
if (!rc)
goto err_exit;
pvattest_log_debug(_("Input ARCB verified."));
calc_measurement = att_gen_measurement_hmac_sha512(measurement_hdr, meas_key, user_data,
nonce, additional_data, &error);
if (!calc_measurement)
goto err_exit;
pvattest_log_debug(_("Measurement calculated."));
if (!att_verify_measurement(calc_measurement, uv_measurement, &error)) {
pvattest_log_GError(__PVATTEST_VERIFY_ERROR_MSG, error);
pvattest_log_debug(_("Measurement values:"));
gbhexdump(uv_measurement);
gbhexdump(calc_measurement);
return PVATTEST_EXIT_MEASURE_NOT_VERIFIED;
}
pvattest_log_debug(_("Measurement verified."));
return EXIT_SUCCESS;
err_exit:
pvattest_log_GError(__PVATTEST_VERIFY_ERROR_MSG, error);
return EXIT_FAILURE;
}
/*
* Will not free the config structs, but the nested char* etc.
* that's what we need to do as we will receive a statically allocated config_t
* Not defined in the parse header as someone might incorrectly assume
* that the config pointers will be freed.
*/
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(pvattest_config_t, pvattest_parse_clear_config)
int main(int argc, char *argv[])
{
int appl_log_lvl = PVATTEST_LOG_LVL_DEFAULT;
g_autoptr(pvattest_config_t) config = NULL;
g_autoptr(GError) error = NULL;
enum pvattest_command command;
int rc;
/* setting up the default log handler to filter messages based on the
* log level specified by the user.
*/
g_log_set_handler(NULL, G_LOG_LEVEL_MASK | G_LOG_FLAG_FATAL | G_LOG_FLAG_RECURSION,
&pvattest_log_default_logger, &appl_log_lvl);
/* setting up the log handler for hexdumps (no prefix and '\n' at end of
* message)to filter messages based on the log level specified by the
* user.
*/
g_log_set_handler(PVATTEST_HEXDUMP_LOG_DOMAIN,
G_LOG_LEVEL_MASK | G_LOG_FLAG_FATAL | G_LOG_FLAG_RECURSION,
&pvattest_log_plain_logger, &appl_log_lvl);
command = pvattest_parse(&argc, &argv, &config, &error);
if (command == PVATTEST_SUBC_INVALID) {
pvattest_log_error(_("%s\nTry '%s --help' for more information"), error->message,
GETTEXT_PACKAGE);
exit(EXIT_FAILURE);
}
g_assert(config);
appl_log_lvl = config->general.log_level;
pv_init();
switch (command) {
case PVATTEST_SUBC_CREATE:
rc = do_create(&config->create);
break;
#ifdef PVATTEST_COMPILE_PERFORM
case PVATTEST_SUBC_PERFORM:
rc = do_perform(&config->perform);
break;
#endif /* PVATTEST_COMPILE_PERFORM */
case PVATTEST_SUBC_VERIFY:
rc = do_verify(&config->verify);
break;
default:
g_return_val_if_reached(EXIT_FAILURE);
}
pv_cleanup();
return rc;
}

18
pvattest/src/types.h Normal file
View File

@@ -0,0 +1,18 @@
/*
* Common data type definitions and functions
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#ifndef PVATTEST_TYPES_H
#define PVATTEST_TYPES_H
#include <stdint.h>
/* Types to mark values as big endian. */
typedef uint16_t be16_t;
typedef uint32_t be32_t;
typedef uint64_t be64_t;
#endif /* PVATTEST_TYPES_H */

177
pvattest/src/uvio.c Normal file
View File

@@ -0,0 +1,177 @@
/*
* UV device (uvio) related functions and definitions.
* uses s390 only (kernel) features.
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
/* Must be included before any other header */
#include "config.h"
#ifdef PVATTEST_COMPILE_PERFORM
#include <stdio.h>
#include <errno.h>
#include <fcntl.h>
#include <unistd.h>
#include <sys/ioctl.h>
#include "attestation.h"
#include "uvio.h"
#include "common.h"
#include "log.h"
/* some helper macros */
#define U64_TO_PTR(v) ((void *)(v))
#define PTR_TO_U64(ptr) ((uint64_t)(ptr))
uvio_attest_t *build_attestation_v1_ioctl(GBytes *serialized_arcb, GBytes *user_data,
const uint32_t measurement_size,
const uint32_t add_data_size, GError **error)
{
g_autoptr(uvio_attest_t) uvio_attest = NULL;
size_t arcb_size;
void *arcb;
pv_wrapped_g_assert(serialized_arcb);
g_bytes_ref(serialized_arcb);
arcb = g_bytes_unref_to_data(serialized_arcb, &arcb_size);
uvio_attest = g_malloc0(sizeof(*uvio_attest));
uvio_attest->arcb_addr = PTR_TO_U64(g_steal_pointer(&arcb));
g_assert_cmpuint(arcb_size, <, UINT32_MAX);
uvio_attest->arcb_len = GUINT32_TO_BE((uint32_t)arcb_size);
/* transferred the local ownership of the arcb from this function to uvio_attest; nullify pointer */
g_steal_pointer(&serialized_arcb);
if (user_data) {
if (g_bytes_get_size(user_data) > sizeof(uvio_attest->user_data)) {
g_set_error(error, ATT_ERROR, ATT_ERR_INVALID_USER_DATA,
_("User data larger than %li bytes"),
sizeof(uvio_attest->user_data));
return NULL;
}
uvio_attest->user_data_len = GUINT16_TO_BE((uint16_t)g_bytes_get_size(user_data));
pv_gbytes_memcpy(uvio_attest->user_data, uvio_attest->user_data_len, user_data);
}
uvio_attest->meas_len = GUINT32_TO_BE(measurement_size);
uvio_attest->meas_addr = PTR_TO_U64(g_malloc0(uvio_attest->meas_len));
uvio_attest->add_data_len = GUINT32_TO_BE(add_data_size);
uvio_attest->add_data_addr = PTR_TO_U64(g_malloc0(uvio_attest->add_data_len));
return g_steal_pointer(&uvio_attest);
}
void uvio_attest_free(uvio_attest_t *attest)
{
if (!attest)
return;
g_free(U64_TO_PTR(attest->arcb_addr));
g_free(U64_TO_PTR(attest->meas_addr));
g_free(U64_TO_PTR(attest->add_data_addr));
g_free(attest);
}
GBytes *uvio_get_measurement(const uvio_attest_t *attest)
{
pv_wrapped_g_assert(attest);
if (attest->meas_addr == (__u64)0)
return NULL;
return g_bytes_new(U64_TO_PTR(attest->meas_addr), GUINT32_FROM_BE(attest->meas_len));
}
GBytes *uvio_get_additional_data(const uvio_attest_t *attest)
{
pv_wrapped_g_assert(attest);
if (attest->add_data_addr == (__u64)0)
return NULL;
return g_bytes_new(U64_TO_PTR(attest->add_data_addr),
GUINT32_FROM_BE(attest->add_data_len));
}
GBytes *uvio_get_config_uid(const uvio_attest_t *attest)
{
pv_wrapped_g_assert(attest);
return g_bytes_new(attest->config_uid, sizeof(attest->config_uid));
}
uint16_t uvio_ioctl(const int uv_fd, const unsigned int cmd, const uint32_t flags,
const void *argument, const uint32_t argument_size, GError **error)
{
g_autofree struct uvio_ioctl_cb *uv_ioctl = g_malloc0(sizeof(*uv_ioctl));
int rc, cached_errno;
pv_wrapped_g_assert(argument);
uv_ioctl->flags = flags;
uv_ioctl->argument_addr = PTR_TO_U64(argument);
uv_ioctl->argument_len = argument_size;
rc = ioctl(uv_fd, cmd, uv_ioctl);
cached_errno = errno;
if (rc < 0) {
g_set_error(error, UVIO_ERROR, UVIO_ERR_UV_IOCTL, _("ioctl failed: %s "),
g_strerror(cached_errno));
return 0;
}
if (uv_ioctl->uv_rc != UVC_EXECUTED)
g_set_error(error, UVIO_ERROR, UVIO_ERR_UV_NOT_OK,
_("Ultravisor call returned '%#x' (%s)"), uv_ioctl->uv_rc,
uvio_uv_rc_to_str(uv_ioctl->uv_rc));
return GUINT16_FROM_BE(uv_ioctl->uv_rc);
}
uint16_t uvio_ioctl_attest(const int uv_fd, uvio_attest_t *attest, GError **error)
{
pv_wrapped_g_assert(attest);
return uvio_ioctl(uv_fd, UVIO_IOCTL_ATT, 0, attest, sizeof(*attest), error);
}
int uvio_open(const char *uv_path, GError **error)
{
pv_wrapped_g_assert(uv_path);
int uv_fd;
int cached_errno;
uv_fd = open(uv_path, O_RDWR);
cached_errno = errno;
if (uv_fd < 0)
g_set_error(error, UVIO_ERROR, UVIO_ERR_UV_OPEN,
_("Cannot open uv driver at %s: %s"), uv_path,
g_strerror(cached_errno));
return uv_fd;
}
const char *uvio_uv_rc_to_str(const int rc)
{
switch (rc) {
case 0x106:
return _("Unsupported attestation request version");
case 0x108:
return _("Number of key slots is greater than the maximum number supported");
case 0x10a:
return _("Unsupported plaintext attestation flags");
case 0x10c:
return _(
"Unable to decrypt attestation request control block. No valid host-key was provided");
case 0x10d:
return _("Measurement data length is too small to store measurement");
case 0x10e:
return _("Additional data length is too small to store measurement");
default:
return _("Unknown code");
}
}
#endif /* PVATTEST_COMPILE_PERFORM */

110
pvattest/src/uvio.h Normal file
View File

@@ -0,0 +1,110 @@
/*
* UV device (uvio) related functions and definitions.
*
* Copyright IBM Corp. 2022
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#ifndef PVATTEST_UVIO_H
#define PVATTEST_UVIO_H
#include "config.h"
#ifdef PVATTEST_COMPILE_PERFORM
#include <sys/ioctl.h>
#include <asm/uvdevice.h>
#include "libpv/glib-helper.h"
#include "arcb.h"
#include "common.h"
#define UVC_EXECUTED 0x0001
typedef struct uvio_attest uvio_attest_t;
G_STATIC_ASSERT(sizeof(uvio_attest_t) == 0x138);
G_STATIC_ASSERT(sizeof(struct uvio_ioctl_cb) == 0x40);
/**
* build_attestation_v1_ioctl:
* @serialized_arcb: A ARCB in binary format
* @user_data (optional): up to 256 bytes of user data to be added to the measurement
* @measurement_size: Size of the measurement result to be allocated
* @add_data_size: Size of the additional data to be allocated
* @error: return location for a #GError
*
* Builds the structure to be passed to `/dev/uv` for attestation IOCTLs and
* allocates any required memory.
*
* Returns: (nullable) (transfer full): Pointer to a uvio_attest_t to be passed to `/dev/uv`
*/
uvio_attest_t *build_attestation_v1_ioctl(GBytes *serialized_arcb, GBytes *user_data,
const uint32_t measurement_size,
const uint32_t add_data_size, GError **error)
PV_NONNULL(1);
GBytes *uvio_get_measurement(const uvio_attest_t *attest) PV_NONNULL(1);
GBytes *uvio_get_additional_data(const uvio_attest_t *attest) PV_NONNULL(1);
GBytes *uvio_get_config_uid(const uvio_attest_t *attest) PV_NONNULL(1);
void uvio_attest_free(uvio_attest_t *attest);
WRAPPED_G_DEFINE_AUTOPTR_CLEANUP_FUNC(uvio_attest_t, uvio_attest_free)
/**
* uvio_ioctl:
* @uv_fd: file descriptor to the UV-device
* @cmd: IOCTL cmd
* @flags: flags for the uv IOCTL
* @argument: pointer to the payload
* @argument_size: size of #argument
* @error: return location for a #GError
*
* Builds the IOCTL structure using, flags and argument, performs the IOCTL, and returns the UV rc in big endian.
* If the device driver emits an error code, a corresponding #GError will be created.
* Use the specialized calls (uvio_ioctl_*).
*
* Returns: UV rc if no device error occurred (>0)
* 0 on #GError
*/
uint16_t uvio_ioctl(const int uv_fd, const unsigned int cmd, const uint32_t flags,
const void *argument, const uint32_t argument_size, GError **error)
PV_NONNULL(4);
/**
* uvio_ioctl_attest:
* @uv_fd: file descriptor to the UV-device
* @attest: pointer to the attestation request
* @error: return location for a #GError
*
* Wraps 'uvio_ioctl' for attestation.
*
* Returns: UV rc if no device error occurred (>0)
* 0 on #GError
*/
uint16_t uvio_ioctl_attest(const int uv_fd, uvio_attest_t *attest, GError **error) PV_NONNULL(2);
/**
* uvio_open:
* @uv_path: path of the UV-device usually at /dev/uv
* @error: return location for a #GError
*
* Returns: File descriptor for the UV-device
* 0 on #GError
*/
int uvio_open(const char *uv_path, GError **error) PV_NONNULL(1);
/**
* uvio_uv_rc_to_str:
* @rc: UV return code
*
* Returns: Pointer to an error string corresponding to the given UV-rc.
*/
const char *uvio_uv_rc_to_str(const int rc);
#define UVIO_ERROR g_quark_from_static_string("pv-uvio_error-quark")
typedef enum {
UVIO_ERR_UV_IOCTL,
UVIO_ERR_UV_OPEN,
UVIO_ERR_UV_NOT_OK,
} uvio_error_e;
#endif /* PVATTEST_COMPILE_PERFORM */
#endif /* PVATTEST_UVIO_H */

5
pvattest/tools/Makefile Normal file
View File

@@ -0,0 +1,5 @@
include ../../common.mak
install:
$(INSTALL) -d -m 755 $(DESTDIR)$(USRBINDIR)
$(INSTALL) -g $(GROUP) -o $(OWNER) -m 755 pvextract-hdr "$(DESTDIR)$(USRBINDIR)"

89
pvattest/tools/pvextract-hdr Executable file
View File

@@ -0,0 +1,89 @@
#!/bin/bash
#
# pvextract_hdr - extract an IBM Secure Execution header from the Image
#
# Sample:
# ./pvextract-hdr -o sehdr.bin se-image.bin
#
# Copyright IBM Corp. 2022
#
# s390-tools is free software; you can redistribute it and/or modify
# it under the terms of the MIT license. See LICENSE for details.
set -o pipefail
set -o nounset
set -e
def_output="sehdr.bin"
def_skip=0x14
def_len=0x4
usage() {
cat <<-EOF
Usage: $(basename "$0") [-o ${def_output}] [-s ${def_skip}] [-l ${def_len}] FILE
Extract the header of the SE-image located in FILE.
By default ${def_skip} pages will be skipped until starting to search
for the header. By default the search will be stopped after ${def_len} pages.
'${def_output}' is the default output file name.
EOF
}
function check_file() {
[ -e "$1" ] ||
{ echo "ERROR: File '$1' not found" >&2 && exit 1; }
}
function check_hdr_ver() {
local hdr_start="$1"
local input="$2"
xxd -s $((hdr_start + 8)) -l 4 "$input" | grep -q "000 0100" ||
{ echo -n "WARNING: unknown hdr version " &&
xxd -s $((hdr_start + 8)) -l 4 "$input" | awk '{print "0x" $2 $3}'; }
}
output=${def_output}
parsed_skip=${def_skip}
parsed_len=${def_len}
# the last argument must be the input file
input="${*: -1}"
while getopts 'o:s:l:h' OPTION; do
case "$OPTION" in
o) output="$OPTARG" ;;
s) parsed_skip="$OPTARG" ;;
l) parsed_len="$OPTARG" ;;
h)
usage
exit 0
;;
:)
echo "ERROR: Must supply an argument to -$OPTARG." >&2
exit 1
;;
*)
usage
exit 1
;;
esac
done
#argument specify pages; convert to bytes
skip=$((parsed_skip * 0x1000))
len=$((parsed_len * 0x1000))
if [ $# -eq 0 ]; then
echo "ERROR: Input not set. Use '$(basename "$0") [FILE]' to specify the Input file" >&2
exit 1
fi
check_file "$input"
hdr_start=$(xxd -s $((skip)) -l $((len)) "${input}" | grep IBMSecEx || { echo ERROR: "${input} does not contain an SE header." >&2 && exit 1; })
hdr_start=$(echo "${hdr_start}" | awk '{print "0x" $1}' | cut -c 1-10)
echo "SE header found at offset ${hdr_start}"
check_hdr_ver "$hdr_start" "$input"
size=$(xxd -s $((hdr_start + 12)) -l 4 "${input}" | awk 'NR==1 {print "0x" $2 $3}')
dd if="${input}" of="${output}" bs=1 count=$((size)) skip=$((hdr_start)) status=none
echo "SE header written to '${output}' ($((size)) bytes)"

View File

@@ -2,12 +2,14 @@ include ../common.mak
SCRIPTS = dbginfo.sh zfcpdbf zipl-switch-to-blscfg scsi_logging_level
SCRIPTS += sclpdbf
MAN_PAGES = dbginfo.sh.1 zfcpdbf.1 zipl-switch-to-blscfg.1
# Helper scripts controlled by corresponding systemd services
SD_HELPER_SCRIPTS = cpictl dumpconf
MAN_PAGES = dbginfo.sh.8 zfcpdbf.8 zipl-switch-to-blscfg.8
all:
install:
$(INSTALL) -d -m 755 $(DESTDIR)$(BINDIR) $(DESTDIR)$(MANDIR)/man1
$(INSTALL) -d -m 755 $(DESTDIR)$(BINDIR) $(DESTDIR)$(MANDIR)/man8
@for i in $(SCRIPTS); \
do \
cat $$i | \
@@ -17,15 +19,19 @@ install:
chmod 755 $(DESTDIR)$(BINDIR)/$$i; \
done
sed -e 's/%S390_TOOLS_VERSION%/$(S390_TOOLS_RELEASE)/' cpictl > \
$(DESTDIR)$(TOOLS_LIBDIR)/cpictl
chown $(OWNER):$(GROUP) $(DESTDIR)$(TOOLS_LIBDIR)/cpictl
chmod 755 $(DESTDIR)$(TOOLS_LIBDIR)/cpictl
@for i in $(SD_HELPER_SCRIPTS); \
do \
cat $$i | \
sed -e 's/%S390_TOOLS_VERSION%/$(S390_TOOLS_RELEASE)/' \
>$(DESTDIR)$(TOOLS_LIBDIR)/$$i; \
chown $(OWNER):$(GROUP) $(DESTDIR)$(TOOLS_LIBDIR)/$$i; \
chmod 755 $(DESTDIR)$(TOOLS_LIBDIR)/$$i; \
done
@for i in $(MAN_PAGES); \
do \
$(INSTALL) -g $(GROUP) -o $(OWNER) -m 644 $$i \
$(DESTDIR)$(MANDIR)/man1; \
$(DESTDIR)$(MANDIR)/man8; \
done
clean:

View File

@@ -1,4 +1,4 @@
.TH DBGINFO.SH 1 "September 2021" "s390-tools"
.TH DBGINFO.SH 8 "May 2022" "s390-tools"
.SH NAME
dbginfo.sh \- collect runtime, configuration and trace information

View File

@@ -1,23 +1,18 @@
#!/bin/bash
### BEGIN INIT INFO
# Provides: dumpconf
# Required-Start: $local_fs
# Required-Stop: $local_fs
# Should-Start:
# Should-Stop:
# Default-Start: 0 1 2 3 5 6
# Default-Stop:
# Short-Description: Configure s390 dump feature
# Description: Configures the s390 dump feature. It uses the configuration file
# /etc/sysconfig/dumpconf
# X-Systemd-RemainAfterExit: true
### END INIT INFO
#
# dumpconf - dump configuration script
#
# This script is called by the corresponding systemd unit file dumpconf.service
# but can also be used stand-alone.
#
# The configuration file is located in /etc/sysconfig/dumpconf
#
# Copyright IBM Corp. 2006, 2022
#
# s390-tools is free software; you can redistribute it and/or modify
# it under the terms of the MIT license. See LICENSE for details.
#
# chkconfig: 012356 01 99
# This script can be either used stand-alone or as System V init script.
DUMPCONF_BIN=/etc/init.d/dumpconf
DUMP_CONFIG_FILE=/etc/sysconfig/dumpconf
CMDFULL=$0
CMD="dumpconf"
@@ -106,8 +101,8 @@ EOF
printversion()
{
cat <<EOF
dumpconf: zSeries dump configuration script version 1.1
Copyright IBM Corp. 2006, 2009
dumpconf: zSeries dump configuration script version %S390_TOOLS_VERSION%
Copyright IBM Corp. 2006, 2022
EOF
}

View File

@@ -2,7 +2,7 @@
.\" s390-tools is free software; you can redistribute it and/or modify
.\" it under the terms of the MIT license. See LICENSE for details.
.\"
.TH ZFCPDBF 1 "DEC 2010, JUL 2016" "s390-tools"
.TH ZFCPDBF 8 "May 2022" "s390-tools"
.SH NAME
zfcpdbf \- zfcp trace data sort/analyse tool.

View File

@@ -1,4 +1,4 @@
.TH ZIPL-SWITCH-TO-BLSCFG 1 "April 2018" "s390-tools"
.TH ZIPL-SWITCH-TO-BLSCFG 8 "May 2022" "s390-tools"
.SH NAME
zipl-switch-to-blscfg \- Switch zipl to use BootLoaderSpec configuration

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