The current handling of invalid command line option arguments is
inconsistent:

- opt_llong_callback() treats non-numerical input the same as "-1",
  which might or might not be rejected later.
- opt_int_callback() returns -1 on non-numerical input without an error
  message, which makes parsing fail silently (libsubcmd will
  automatically print the usage of the option only, no error message).
- custom callbacks abort command line parsing using fatal(), which
  displays an error message and exits, without libsubcmd printing the
  usage.

Unify this such that all invalid options, regardless of the format,
print an error message similar to the out of range case:

Error: --opt: 'value' is not a valid XY

followed by the usage of the option, e.g.:

$ rtla timerlat hist --period=1us
 Error: --period: '1us' is not a valid number

 Usage: rtla timerlat hist [<options>] [-h|--help]

    -p, --period <us>     timerlat period in us

As this is a libsubcmd help path, all option parsing failures now return
the exit code of 129 (help).

The unified handling is implemented using a new error message helper,
opt_err(), which is called from two new CLI-specific parsing functions,
strtoll_safe() and strtoi_safe(), as well as from custom helpers.

Option callback tests are updated to cover the new behavior.

Assisted-by: Claude:claude-opus-4-6
Reviewed-by: Wander Lairson Costa <wander@redhat.com>
Link: https://lore.kernel.org/r/20260716144901.1187474-1-tglozar@redhat.com
Signed-off-by: Tomas Glozar <tglozar@redhat.com>

Several RTLA options do range validation inside the CLI parser layer
(e.g. -p/--period). When RTLA migrated CLI parsing to libsubcmd, this
logic was moved unchanged inside opt_*() callbacks.

Unify range validation so that all options use two newly added
functions, check_llong_range() and check_int_range(), to validate the
range.

The new range validation returns -1 from opt_*() callbacks rather than
hard-exit with fatal(), allowing the help message for the specific
option to be automatically displayed by libsubcmd logic.

Many options no longer need a custom callback, as they use the unified
range validation of opt_llong_callback() and opt_int_callback().
Validation for several other options is improved:

- timerlat -p/--period: lower bound raised from 1 to 100 us to match
  the kernel's timerlat_min_period in trace_osnoise.c.
- timerlat -A/--aligned: reject negative values.
- timerlat --deepest-idle-state: add range [-1, INT_MAX]; previously,
  values <= -2 were read as "option not set".
- timerlat -p/--period, -A/--aligned, -b/--bucket-size: properly reject
  negative values instead of passing them to the tracer.

Remove unit tests for removed callbacks and test the new range
validation functionality of opt_llong_callback() and opt_int_callback().

Update runtime tests for histogram options to account for the new error
messages and exit value.

Assisted-by: Claude:claude-opus-4-6
Link: https://lore.kernel.org/r/20260710131554.338335-1-tglozar@redhat.com
Signed-off-by: Tomas Glozar <tglozar@redhat.com>

libsubcmd implicitly allows the user to unset already set options using
a "no-" prefix for long options. For example, if I set the period like
this:

$ rtla timerlat -D
Loading BPF program
reading osnoise/timerlat_period_us returned 1000
setting osnoise/timerlat_period_us to 1000
reading osnoise/print_stack returned 0
setting osnoise/print_stack to 0
...
<timerlat top>

it can be unset by a subsequent --no-debug:

$ rtla timerlat -D --no-debug
...
<timerlat top>

Currently, this works only for boolean options. Extend the feature for
all options by implementing handling of the "unset" argument in opt_*()
callbacks defined in cli_p.h, except for list options, i.e. options that
can be passed multiple times (--event, --filter, --trigger,
--on-threshold, --on-end).

This allows, for example, unsetting of int/long long options, e.g. "-p":

$ rtla timerlat -D -p100 --no-period
...
setting osnoise/timerlat_period_us to 1000
...

By default, options in params struct are reset to zero. A constant is
added for every parameter with a different default value, which is then
used both in <tool>_hist_args() while setting the initial value and in
opt_*() when unsetting the option. This refactoring ensures there is no
duplicate "magic number".

The default value for opt_llong_callback() and opt_int_callback() is
passed in struct option's defval field; new macros
RTLA_OPT_{LLONG,INT}{,_DEFVAL} are added to define the field
conveniently. The default value for other callbacks is hardcoded inside
each callback's unset logic.

Reviewed-by: Wander Lairson Costa <wander@redhat.com>
Link: https://lore.kernel.org/r/20260629083654.1548925-1-tglozar@redhat.com
Signed-off-by: Tomas Glozar <tglozar@redhat.com>

A bug was reported when rtla-timerlat-top tool performs on-threshold
actions, even though no threshold was hit. This is reproduced even if no
threshold is set at all:

$ rtla timerlat top -q -c 0 --on-threshold shell,command='echo BAD'
BAD
                                     Timer Latency
...

The bug is due to incorrect logic in timerlat_top_bpf_main_loop().
The loop uses timerlat_bpf_wait(), the return values of which are:

- > 0 (number of ringbuffer entries): at least 1 CPU hit threshold
- = 0: time out
- < 0: wait was interrupted by a signal

Commit 3138df6f0cd0 ("rtla/timerlat: Exit top main loop on any non-zero
wait_retval") changed the condition for "threshold hit" from
"wait_reval == 1" (exactly 1 CPU hit threshold) to "wait_retval != 0",
to fix a race where multiple CPUs hit the threshold at the same time.

That also made it incorrectly include a signal (< 0), coming from either
duration expired (SIGALRM) or user interrupt (SIGINT).

Check for wait_retval greater than zero in the if condition to cover all
return values correctly.

Fixes: 3138df6f0cd0 ("rtla/timerlat: Exit top main loop on any non-zero wait_retval")
Reported-by: Attila Fazekas <afazekas@redhat.com>
Reviewed-by: Wander Lairson Costa <wander@redhat.com>
Link: https://lore.kernel.org/r/20260713141047.687877-1-tglozar@redhat.com
Signed-off-by: Tomas Glozar <tglozar@redhat.com>

Compiling RTLA 7.1.x with GCC 16 and uClibc as standard library fails
with these errors:

src/common.c: In function ‘set_signals’:
src/common.c:40:17: error: implicit declaration of function ‘alarm’ [-Wimplicit-function-declaration]
   40 |                 alarm(params->duration);
      |                 ^~~~~
src/common.c: In function ‘common_apply_config’:
src/common.c:187:44: error: implicit declaration of function ‘getpid’; did you mean ‘getpt’? [-Wimplicit-function-declaration]
  187 |                 retval = sched_setaffinity(getpid(), sizeof(params->hk_cpu_set),
      |                                            ^~~~~~
      |                                            getpt
In file included from src/common.c:9:
src/common.c: In function ‘run_tool’:
src/common.c:262:19: error: implicit declaration of function ‘sysconf’; did you mean ‘sscanf’? [-Wimplicit-function-declaration]
  262 |         nr_cpus = get_nprocs_conf();
      |                   ^~~~~~~~~~~~~~~
src/common.c:262:19: error: ‘_SC_NPROCESSORS_CONF’ undeclared (first use in this function)
  262 |         nr_cpus = get_nprocs_conf();
      |                   ^~~~~~~~~~~~~~~
src/common.c:262:19: note: each undeclared identifier is reported only once for each function it appears in
src/common.c:370:17: error: implicit declaration of function ‘sleep’ [-Wimplicit-function-declaration]
  370 |                 sleep(1);
      |                 ^~~~~

Restore the missing unistd.h include.

Fixes: 115b06a00875 ("tools/rtla: Consolidate nr_cpus usage across all tools")
Signed-off-by: Andreas Ziegler <br025@umbiko.net>
Link: https://lore.kernel.org/r/20260624033342.129970-1-br025@umbiko.net
[ capital letter in patch subject, remove EOL after Fixes ]
Signed-off-by: Tomas Glozar <tglozar@redhat.com>

Add a new option, -A/--aligned, that enables timerlat thread alignment
implemented on the kernel-side in commit 4245bf4dc58f ("tracing/osnoise:
Add option to align tlat threads"). The option takes an argument,
representing alignment between timerlat threads in microseconds.

The feature is modeled after the option of the same name in the
cyclictest tool.

Link: https://lore.kernel.org/r/20260527144928.2944472-1-tglozar@redhat.com
Signed-off-by: Tomas Glozar <tglozar@redhat.com>

In addition to testing all tool_parse_args() functions, test also all
callbacks used for parsing custom option formats.

The callbacks represent a middle layer between the parsing functions
and utility functions dedicated to checking specific argument formats,
for example, scheduling class and duration. Callback tests are run
before parsing functions to make sure any issue in the former is
reported before it is encountered through the latter.

Tests verify both successful parsing and proper rejection of invalid
inputs (via exit tests). To enable testing static callbacks, a pragma
once guard is added to timerlat.h for safe inclusion by cli_p.h.

Add dependency of UNIT_TESTS_IN on LIBSUBCMD_INCLUDES, as the new test
file tests/unit/cli_opt_callback.c includes cli_p.h which includes
subcmd/parse-options.h.

Link: https://lore.kernel.org/r/20260528103254.2990068-7-tglozar@redhat.com
Signed-off-by: Tomas Glozar <tglozar@redhat.com>

Add a test suite for the _parse_args() function of each tool that checks
the params structures (struct common_params, struct osnoise_params,
struct timerlat_params) returned by them for correctness.

One test case is added per option, as well as a few special cases for
tricky combinations of options. Test cases are ordered the same as the
option arrays and help message to allow easy checking of whether all
options are covered.

This should help clarify what the proper command line behavior of RTLA
is in case there are holes in the documentation and verify that the
intended behavior is implemented correctly.

A few necessary changes to the unit tests were done as part of this
commit:

- Unit tests now also link to libsubcmd and its dependencies.
- A new global variable in_unit_test is added to RTLA's CLI interface,
  causing it to skip check for root if running in unit tests. This
  allows the CLI unit tests to run as non-root, like existing unit
  tests.

There is quite a lot of duplication, some of it is mitigated with macros,
but partially it is intentional so that future changes in behavior are
tracked across tools.

Link: https://lore.kernel.org/r/20260528103254.2990068-6-tglozar@redhat.com
Signed-off-by: Tomas Glozar <tglozar@redhat.com>

Instead of using getopt_long() directly to parse the command line
arguments given to an RTLA tool, use libsubcmd's parse_options().

Utilizing libsubcmd for parsing command line arguments has several
benefits:

- A help message is automatically generated by libsubcmd from the
  specification, removing the need of writing it by hand.
- Options are sorted into groups based on which part of tracing (CPU,
  thread, auto-analysis, tuning, histogram) they relate to.
- Common parsing patterns for numerical and boolean values now share
  code, with the target variable being stored in the option array.

To avoid duplication of the option parsing logic, RTLA-specific
macros defining struct option values are created:

- RTLA_OPT_* for options common to all tools
- OSNOISE_OPT_* and TIMERLAT_OPT_* for options specific to
  osnoise/timerlat tools
- HIST_OPT_* macros for options specific to histogram-based tools.

Individual *_parse_args() functions then construct an array out of
these macros that is then passed to libsubcmd's parse_options().

All code specific to command line options parsing is moved out of the
individual tool files into a new file, cli.c, which also contains the
contents of the rtla.c file. A private header, cli_p.h, is added
alongside the public header cli.h, so that unit tests are able to test
statically declared option callbacks.

Minor changes:

- The return value of tool-level help option changes to 129, as this is
  the value set by libsubcmd; this is reflected in affected test cases.
  The implementation of help for command-level and tracer-level help
  is set to 129 as well for consistency, and the change is reflected in
  exit value documentation.
- Related to the above, {rtla,osnoise,timerlat}_usage() are marked
  __noreturn and exit() is removed from after they are called for
  cleaner code.
- The error messages for invalid argument for options --dma-latency and
  -E/--entries were corrected, fixing off-by-one in the limits.

Note that unsetting options (using --no-<opt> syntax) is currently not
implemented for options that use custom callbacks. For --irq and
--thread, it will never be implemented, as they conflict with already
existing --no-irq and --no-thread with a different meaning.

Assisted-by: Composer:composer-1.5
Link: https://lore.kernel.org/r/20260528103254.2990068-5-tglozar@redhat.com
Signed-off-by: Tomas Glozar <tglozar@redhat.com>

In preparation for migrating RTLA to libsubcmd, build libsubcmd from the
appropriate directory next to the RTLA build proper, and link the
resulting object to RTLA.

libsubcmd uses str_error_r() and strlcpy() at several places. To support
these, also link the respective libraries from tools/lib.

For completeness, also add tools/include to include path. This will
allow other userspace functions and macros shipped with the kernel to be
used in RTLA; perf and bpftool, two other users of libsubcmd, already do
that.

To prevent a name conflict, rename RTLA's run_command() function to
run_tool_command(), and replace RTLA's own container_of implementation
with the one in tools/include/linux/container_of.h.

Assisted-by: Composer:composer-1
Link: https://lore.kernel.org/r/20260528103254.2990068-2-tglozar@redhat.com
Signed-off-by: Tomas Glozar <tglozar@redhat.com>