From dae3e20b8d1833803058974c4a023e812a26701c Mon Sep 17 00:00:00 2001
From: Kienan Stewart <kstewart@efficios.com>
Date: Tue, 15 Aug 2023 11:50:16 -0400
Subject: [PATCH] Tests: use CPU ids from online ranges
MIME-Version: 1.0
Content-Type: text/plain; charset=utf8
Content-Transfer-Encoding: 8bit

test_tracefile_count could fail randomly on systems where there are CPUs
present but not online. For example:

  $ cat /sys/devices/system/cpu/online
  0-7
  $ cat /sys/devices/system/cpu/present
  0-39

When a CPU is present, it will have an entry in
/sys/devices/system/cpu/cpuX for it's ID, and thus the test may pick
that CPU's ID. However, a present CPU which is not online is not a valid
target for taskset.

In cases where `get_any_available_cpu` is used with task set, the tests
could fail for a similar reason. This case can be somewhat less common,
because it would return the numerically lowest CPU first; however, with
online as follows cpu 0 isn't available and taskset fails.

  $ cat /sys/devices/system/cpu/online
  18-19,135,142
  $ cat /sys/devices/system/cpu/present
  0-167

Change-Id: I06ac2e67495552f54765794b154acc9e8e9990ec
Signed-off-by: Kienan Stewart <kstewart@efficios.com>
Signed-off-by: Jérémie Galarneau <jeremie.galarneau@efficios.com>
---
 tests/regression/tools/snapshots/ust_test     |  6 ++---
 .../tracefile-limits/test_tracefile_count     | 21 +++-------------
 tests/utils/utils.sh                          | 25 +++++++++++++++++--
 3 files changed, 29 insertions(+), 23 deletions(-)

diff --git a/tests/regression/tools/snapshots/ust_test b/tests/regression/tools/snapshots/ust_test
index 24f43aec6..0a68e2b73 100755
--- a/tests/regression/tools/snapshots/ust_test
+++ b/tests/regression/tools/snapshots/ust_test
@@ -212,7 +212,7 @@ function test_ust_local_snapshot_small_discard_buffers ()
 	OLDCPUSET=$(taskset -p $$)
 
 	diag "Test local UST snapshots with small discard buffers"
-	taskset -p 0x1 $$ 1>/dev/null 2>&1 	# CPU 0 only
+	taskset -c "$(get_any_available_cpu)" -p $$ 1>/dev/null 2>&1
 	create_lttng_session_no_output $SESSION_NAME
 	enable_mmap_small_discard_ust_channel $SESSION_NAME $CHANNEL_NAME
 	enable_ust_lttng_event_ok $SESSION_NAME $EVENT_NAME $CHANNEL_NAME
@@ -257,7 +257,7 @@ function test_ust_local_snapshot_small_overwrite_buffers ()
 	OLDCPUSET=$(taskset -p $$)
 
 	diag "Test local UST snapshots with small overwrite buffers"
-	taskset -p 0x1 $$ 1>/dev/null 2>&1 	# CPU 0 only
+	taskset -p "$(get_any_available_cpu)" $$ 1>/dev/null 2>&1
 	create_lttng_session_no_output $SESSION_NAME
 	enable_mmap_small_overwrite_ust_channel $SESSION_NAME $CHANNEL_NAME
 	enable_ust_lttng_event_ok $SESSION_NAME $EVENT_NAME $CHANNEL_NAME
@@ -304,7 +304,7 @@ function test_ust_local_snapshot_max_size ()
 	local snapshot_max_size
 	local channel_max_size_per_cpu
 
-	IFS=" " read -r -a cpus_list <<< "$(get_exposed_cpus_list)"
+	IFS=" " read -r -a cpus_list <<< "$(get_online_cpus)"
 
 	possible_cpus=$(get_possible_cpus_count)
 	subbuf_size=$(getconf PAGE_SIZE)
diff --git a/tests/regression/tools/tracefile-limits/test_tracefile_count b/tests/regression/tools/tracefile-limits/test_tracefile_count
index 37d15905a..9d6c2d9a8 100755
--- a/tests/regression/tools/tracefile-limits/test_tracefile_count
+++ b/tests/regression/tools/tracefile-limits/test_tracefile_count
@@ -22,30 +22,15 @@ TRACEFILE_SIZE=$PAGE_SIZE
 
 source "$TESTDIR"/utils/utils.sh
 
-NUM_CPUS=$(conf_proc_count)
-
 if [ ! -x "$TESTAPP_BIN" ]; then
 	BAIL_OUT "No UST events binary detected."
 fi
 
 function pick_random_cpuid ()
 {
-        local cpuid=0
-
-        # On Linux pick a random available cpuid from sysfs
-        if [ -d "/sys/devices/system/cpu" ]; then
-                local cpuids=()
-
-                for i in /sys/devices/system/cpu/cpu[0-9]*; do
-                        cpuids+=("${i#/sys/devices/system/cpu/cpu}")
-                done
-
-                cpuid=${cpuids[ $RANDOM % ${#cpuids[@]} ]}
-        else
-                cpuid=$((RANDOM % NUM_CPUS))
-        fi
-
-        echo $cpuid
+	local cpuids
+	read -r -a cpuids <<< "$(get_online_cpus)"
+	echo "${cpuids[ $RANDOM % ${#cpuids[@]} ]}"
 }
 
 function enable_lttng_channel_count_limit ()
diff --git a/tests/utils/utils.sh b/tests/utils/utils.sh
index 89624a421..004e90948 100644
--- a/tests/utils/utils.sh
+++ b/tests/utils/utils.sh
@@ -130,6 +130,27 @@ function randstring()
 	echo
 }
 
+# Return a space-separated string of online CPU IDs, based on
+# /sys/devices/system/cpu/online, or from 0 to nproc - 1 otherwise.
+function get_online_cpus()
+{
+	local cpus=()
+	local range_re
+	if [ -f /sys/devices/system/cpu/online ]; then
+		range_re='([0-9]+)-([0-9]+)'
+		while read -r range ; do
+			if [[ "${range}" =~ ${range_re} ]] ; then
+				mapfile -t -O "${#cpus[*]}" cpus <<< $(seq "${BASH_REMATCH[1]}" "${BASH_REMATCH[2]}")
+			else
+				cpus+=("${range}")
+			fi
+		done < <(tr ',' $'\n' < /sys/devices/system/cpu/online)
+	else
+		read -r -a cpus <<< $(seq 0 $(( $(conf_proc_count) - 1 )) )
+	fi
+	echo "${cpus[*]}"
+}
+
 # Helpers for get_possible_cpus.
 function get_possible_cpus_count_from_sysfs_possible_mask()
 {
@@ -202,8 +223,8 @@ function get_exposed_cpus_list()
 # value, e.g. that it could be 0.
 function get_any_available_cpu()
 {
-	for cpu in /sys/devices/system/cpu/cpu[0-9]*; do
-		echo "${cpu#/sys/devices/system/cpu/cpu}"
+	for cpu in $(get_online_cpus); do
+		echo "${cpu}"
 		break;
 	done
 }
-- 
2.34.1