fix: support STM32WL family

It happens on STM32WL that case 4 fails.
And it seems to be linked to LiuLayland theoritical bound value,
which is based on ideal system.
It doesn't take into account things like RTOS task swap, ...
So take more margin for case 4 by decreasing the cpu load of one task,
so that global CPU moves away from theoretical bound.

Signed-off-by: Alexandre Bourdiol <alexandre.bourdiol@st.com>

Author: ABOSTM

examples/frLiuLayland/frLiuLayland.ino

@@ -31,7 +31,7 @@ struct task_t {
 task_t tasks1[] = {{10, 5, 2}, {15, 6, 1}};
 task_t tasks2[] = {{10, 5, 2}, {15, 4, 1}};
 task_t tasks3[] = {{10, 3, 3}, {13, 4, 2}, {17, 4, 1}};
-task_t tasks4[] = {{10, 3, 3}, {13, 4, 2}, {17, 2, 1}};
+task_t tasks4[] = {{10, 3, 3}, {13, 4, 2}, {17, 1, 1}};
 task_t* taskList[] = {tasks1, tasks2, tasks3, tasks4};
 int taskCount[] = {2, 2, 3, 3};
 //------------------------------------------------------------------------------