Fix two off-by-one errors:
+ One caused an out-of-bounds read when the loadparm value was 63
+ The other made a boot entry unselectable.
Currently, these bugs have no practical impact because:
1. The memory area beyond __stage2_params.config[63] is empty.
2. BOOT_MENU_ENTRIES is set to 63, which exceeds the number of boot menu
entries that can be written to disk.
Reviewed-by: Mikhail Zaslonko <zaslonko@linux.ibm.com>
Reviewed-by: Eduard Shishkin <edward6@linux.ibm.com>
Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Add functionality for generating autocompletion scripts,
allowing for tab completion of tool options for bash and zsh.
This functionality relies on reading the available options
from the util_opt struct at runtime.
The script generation happens on the build system.
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Szabina Korbai <szkorbai@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
When a loadparm is presented by the user it's not checked whether it's
too large. Add this missing check to fix an out-of-bounds read.
The problem did not show up because the area behind the buffer is empty.
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Reviewed-by: Eduard Shishkin <edward6@linux.ibm.com>
Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Currently, 'sclp_setup(SCLP_INIT)' fails if no SCLP line-mode console is
available. As a result 'menu_param()' is never called, even though it is
required to retrieve the 'loadparm' value.
However, reading the loadparm via SCLP remains useful even when a SCLP
line-mode console is absent, because this value determines which boot
entry should be selected. Therefore, the boot process should continue by
retrieving the loadparm without requiring an SCLP line-mode console.
It's safe to continue without a SCLP console as 'printf' and
'menu_param' tolerates the absence of a SCLP line-mode console.
Fixes: https://github.com/ibm-s390-linux/s390-tools/issues/196
Reported-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Reviewed-by: Eduard Shishkin <edward6@linux.ibm.com>
Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Remove extern keyword from function signatures in header files.
Function signatures in headers are already extern implicitly.
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Mete Durlu <meted@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
When machine parseable format is specified via hyptop --format csv
the summary row for csv does not have any indicator or and a blank
system name. Avoid this confusion by adding "SUM" as system name.
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Mete Durlu <meted@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Fields without a valid value are normally excluded from formatted output
of hyptop. Add "--all" option to force hyptop to display those fields
with null values.
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Mete Durlu <meted@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Hyptop uses underscore("_") in between words for options, this breaks
consistency. Use hyphens ("-") in between words in options moving
forward. Underscore formats are still supported for compatibility.
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Mete Durlu <meted@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Use the common helpers in lib/util_time instead of using hardcoded
values for time calculations.
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Mete Durlu <meted@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Introduce util_time as a place for time related macros, definitions,
and functions that are commonly used. Add NSEC_PER_USEC, NSEC_PER_MSEC,
USEC_PER_SEC and NSEC_PER_SEC definitions that are used for time
conversion calculations as a start. Add time conversion helper functions.
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Mete Durlu <meted@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Physical machine information does not include any monotonically
increasing time field like other per CPU information blocks. This
Prevents the percentage calculation for the physical information fields
as the divisor(time_delta) is missing;
field% = (value_current - value_previous) / time_delta
To circumvent that, use the current partition's per CPU online time
values to calculate the time_delta. Integrate time_delta as a new
field(phys_delta_us) specific to the physical systems.
Explicitly CPU0 online time is used, since CPU0 is always online
and cannot be deconfigured on s390x. Its online-time deltas would
match those of the physical CPUs.
Since a new field has to be used for physical systems a new
column in the table also has to be created but instead map the
physical system field to the corresponding regular system fields
for a nicer table view.
Reviewed-by: Vasily Gorbik <gor@linux.ibm.com>
Signed-off-by: Mete Durlu <meted@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Introduce a new row to represent the underlying physical machine
and the physical information hyptop gathers from diag.
Reviewed-by: Vasily Gorbik <gor@linux.ibm.com>
Signed-off-by: Mete Durlu <meted@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Many distros and buildsystems use '-Wl,--as-needed' default-wise anyway.
For example, Ubuntu, Fedora, and Meson default to '-Wl,--as-needed'.
Considering some of the projects that use/support Meson, e.g. util-linux,
systemd, libvirt, or git, are all built with '-Wl,--as-needed' as well.
So add '-Wl,--as-needed' to DEFAULT_LDFLAGS to also default to that.
Note that with '-Wl,--as-needed' the order of flags, objects, and libraries
matters for the link step. The link rule in common.mak already has the
correct order, so using '-Wl,--as-needed' will not hurt.
%: %.o
$(LINK) $(ALL_LDFLAGS) $^ $(LDLIBS) -o $@
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>
The check_dep macro allows to pass additional compiler & linker
options as 5th argument. This argument might contain shared
libraries to link against (i.e. -lsomething). To ensure that
the check_dep macro always attempts to link to these libraries
'-Wl,--no-as-needed' is needed.
At least on Ubuntu '-Wl,--as-needed' is the default, and with
that, the library might get skipped if it is not really needed
by the program built by the check_dep macro. Furthermore, with
'-Wl,--as-needed' the order of the arguments matters, and thus
a library specified in the 5th argument is at the wrong position,
which leads to link errors and thus the check_dep macro will
report that the dependency is not fulfilled, although it might
be fulfilled.
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>
Add an additional trigger to the zdev-generated 41-ap.rules file. This
trigger will check the bindings_complete_count value during coldplug
replay of the ap bus, which allows the udev rule to handle the case
where zcrypt has already fired the BINDINGS==complete with
COMPLETECOUNT==1 before the udev rule was accessible (example: before
rootfs has been mounted). Without this, it's possible for the rule to
never fire despite the initial bindings having completed.
While it was always possible for zcrypt to be forced onto initrd, it
becomes far more likely with changes like 14c977768 ("zkey/dracut: Add
a dracut config file for zkey").
Reviewed-by: Boris Fiuczynski <fiuczy@linux.ibm.com>
Reviewed-by: Peter Oberparleiter <oberpar@linux.ibm.com>
Acked-by: Vineeth Vijayan <vneethv@linux.ibm.com>
Signed-off-by: Matthew Rosato <mjrosato@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Remove unused header files from these sources.
This simplifies maintenance and slightly reduces compile time.
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Jan Polensky <japo@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Drop unused headers from the internal s390-tools library. These headers
were only indirectly pulling in stdlib.h, which is now included
explicitly to support memory deallocation via free().
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Jan Polensky <japo@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Remove ctype.h and errno.h since they are not referenced in this
file.
Add err.h to explicitly declare the dependency on warnx().
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Jan Polensky <japo@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Remove lib/util_base.h and errno.h since they are not referenced in this
file.
Add err.h to explicitly declare the dependency on err() and warn().
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Jan Polensky <japo@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
For s390-tools util_strlcpy() would be a safer replacement. However,
no code is changed, only .checkpatch.conf is updated to avoid false
positives in userspace.
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Jan Polensky <japo@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Introduce a .clang-tidy file to provide consistent code quality checks
across the project.
The configuration is not integrated into .pre-commit-config.yaml because
it depends on a recently generated compile_commands.json file.
Example usage:
clang-tidy --fix-errors **/*.[ch]
Note: Running clang-tidy should be considered optional, not enforced.
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Jan Polensky <japo@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Ignore the removal of nonexistent $image file. This removes the
following error message:
$ mk-s390image /var/lib/libvirt/images/hades/vmlinux-s390x out.img
...
rm: cannot remove 'out.img': No such file or directory
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
If an input file is not readable return a proper error message indicating
this.
Before:
$ mk-s390image /var/lib/libvirt/images/hades/vmlinux-s390x out.img
mk-s390image: Unrecognized file format for /var/lib/libvirt/images/hades/vmlinux-s390x
After:
$ mk-s390image /var/lib/libvirt/images/hades/vmlinux-s390x out.img
mk-s390image: File /var/lib/libvirt/images/hades/vmlinux-s390x cannot be read, no read permission
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Since Linux commit 5ecb2da660ab ("s390: support command lines longer
than 896 bytes") the s390x kernel supports longer command lines than 896
bytes. The indication of the maximum size is stored in a new field at
address 0x10430, older kernels without the support store a value of 0
there so in that case fallback to the old maximum length of 896.
In addition, use the checked size as limit to copy in the 'dd' call.
This prevents Linux kernel corruption in case the parmline has changed
in between.
Fixes: https://github.com/ibm-s390-linux/s390-tools/issues/194
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Most of the fixes were auto-generated using the following command:
$ shellcheck --format=diff mk-s390image | git apply -
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
