tests: spinlock: Increase thread execution to prevent quick exit

finikorg · carlescufi · commit a3959af4f5e2 · 2023-09-28T15:46:07.000+02:00
On Intel boards (like intel_ehl_crb and intel_rpl_s_crb) for the
trylock_test some part is executed very fast and since there is no
synchronization, there might be situation when there is no
trylock_failures. Increasing time spend in this part fixes the issue.

Signed-off-by: Andrei Emeltchenko &lt;andrei.emeltchenko@intel.com&gt;
diff --git a/tests/kernel/spinlock/src/main.c b/tests/kernel/spinlock/src/main.c
@@ -97,8 +97,9 @@ void bounce_once(int id, bool trylock)
 	 */
 	bounce_owner = id;
 
-	for (i = 0; i < 100; i++) {
+	for (i = 0; i < 5; i++) {
 		zassert_true(bounce_owner == id, "Locked data changed");
+		k_busy_wait(1);
 	}
 
 	/* Release the lock */

Original file line number	Diff line number	Diff line change
`@@ -97,8 +97,9 @@ void bounce_once(int id, bool trylock)`
`97`	`97`	`*/`
`98`	`98`	`bounce_owner = id;`
`99`	`99`
`100`		`- for (i = 0; i < 100; i++) {`
	`100`	`+ for (i = 0; i < 5; i++) {`
`101`	`101`	`zassert_true(bounce_owner == id, "Locked data changed");`
	`102`	`+ k_busy_wait(1);`
`102`	`103`	`}`
`103`	`104`
`104`	`105`	`/* Release the lock */`