test: rely on EAL detection for core list

Cores count has a direct impact on the time needed to complete unit
tests.

Currently, the core list used for unit test is enforced to "all cores on
the system" with no way for (CI) users to adapt it.
On the other hand, EAL default behavior (when no -c/-l option gets passed)
is to start threads on as many cores available in the process cpu
affinity.

Remove logic from meson: users can then select where to run the tests by
either running meson with a custom cpu affinity (using taskset/cpuset
depending on OS) or by passing a --test-args option to meson.

Example:
$ sudo meson test -C build --suite fast-tests -t 3 --test-args "-l 0-3"

Signed-off-by: David Marchand <david.marchand@redhat.com>
Tested-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: Aaron Conole <aconole@redhat.com>

MAINTAINERS

@@ -1622,7 +1622,6 @@ Test Applications
 
 Unit tests framework
 F: app/test/commands.c
-F: app/test/get-coremask.sh
 F: app/test/has-hugepage.sh
 F: app/test/packet_burst_generator.c
 F: app/test/packet_burst_generator.h

app/test/get-coremask.sh

@@ -1,13 +0,0 @@
-#! /bin/sh -e
-# SPDX-License-Identifier: BSD-3-Clause
-# Copyright(c) 2019 Intel Corporation
-
-if [ "$(uname)" = "Linux" ] ; then
-	cat /sys/devices/system/cpu/present
-elif [ "$(uname)" = "FreeBSD" ] ; then
-	ncpus=$(/sbin/sysctl -n hw.ncpu)
-	echo 0-$(expr $ncpus - 1)
-else
-# fallback
-	echo 0-3
-fi

app/test/meson.build

@@ -465,13 +465,8 @@ message('hugepage availability: @0@'.format(has_hugepage))
 timeout_seconds = 600
 timeout_seconds_fast = 10
 
-get_coremask = find_program('get-coremask.sh')
-num_cores_arg = '-l ' + run_command(get_coremask).stdout().strip()
-
-default_test_args = [num_cores_arg]
-
 foreach arg : fast_tests
-    test_args = default_test_args
+    test_args = []
     run_test = true
     if not has_hugepage
         if arg[1]
@@ -504,7 +499,6 @@ endforeach
 foreach arg : perf_test_names
     test(arg, dpdk_test,
             env : ['DPDK_TEST=' + arg],
-            args : default_test_args,
             timeout : timeout_seconds,
             is_parallel : false,
             suite : 'perf-tests')
@@ -513,7 +507,6 @@ endforeach
 foreach arg : driver_test_names
     test(arg, dpdk_test,
             env : ['DPDK_TEST=' + arg],
-            args : default_test_args,
             timeout : timeout_seconds,
             is_parallel : false,
             suite : 'driver-tests')
@@ -522,7 +515,6 @@ endforeach
 foreach arg : dump_test_names
     test(arg, dpdk_test,
             env : ['DPDK_TEST=' + arg],
-            args : default_test_args,
             timeout : timeout_seconds,
             is_parallel : false,
             suite : 'debug-tests')

doc/guides/prog_guide/meson_ut.rst

@@ -37,6 +37,14 @@ command::
 
     $ meson test --suite fast-tests
 
+If desired, additional arguments can be passed to the test run via the meson
+``--test-args`` option.
+For example, tests will by default run on as many available cores as is needed
+for the test, starting with the lowest number core - generally core 0.
+To run the fast-tests suite using only cores 8 through 16, one can use::
+
+    $ meson test --suite fast-tests --test-args="-l 8-16"
+
 The meson command to list all available tests::
 
     $ meson test --list
@@ -47,9 +55,16 @@ Arguments of ``test()`` that can be provided in meson.build are as below:
 
 * ``is_parallel`` is used to run test case either in parallel or non-parallel mode.
 * ``timeout`` is used to specify the timeout of test case.
-* ``args`` is used to specify test specific parameters.
+* ``args`` is used to specify test specific parameters (see note below).
 * ``env`` is used to specify test specific environment parameters.
 
+Note: the content of meson ``--test-args`` option and the content of ``args``
+are appended when invoking the DPDK test binary.
+Because of this, it is recommended not to set any default coremask or memory
+configuration in per test ``args`` and rather let users select what best fits
+their environment. If a test can't run, then it should be skipped, as described
+below.
+
 
 Dealing with skipped test cases
 -------------------------------