The QTest `test_irq_pin_multiplexer` makes the assumption that the
reset state of irq line 15 is low, which is false since STM32L4x5 GPIO
was implemented (the reset state of pin GPIOA15 is high because there's
pull-up and it results in the irq line 15 also being high at reset).

It wasn't triggering an error because `test_interrupt` was mistakenly
"resetting" the line low.

This commit corrects these two mistakes by :
- not setting the line low in `test_interrupt`
- using an irq line in `test_irq_pin_multiplexer` which is low at reset

Signed-off-by: Inès Varhol <ines.varhol@telecom-paris.fr>
---
 tests/qtest/stm32l4x5_syscfg-test.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)

diff --git a/tests/qtest/stm32l4x5_syscfg-test.c b/tests/qtest/stm32l4x5_syscfg-test.c
index 506ca08bc2..1cdf8f05c8 100644
--- a/tests/qtest/stm32l4x5_syscfg-test.c
+++ b/tests/qtest/stm32l4x5_syscfg-test.c
@@ -223,7 +223,7 @@ static void test_interrupt(void)
     /* Clean the test */
     syscfg_writel(SYSCFG_EXTICR1, 0x00000000);
     syscfg_set_irq(0, 0);
-    syscfg_set_irq(15, 0);
+    /* irq 15 is high at reset because GPIOA15 is high at reset */
     syscfg_set_irq(17, 0);
 }
 
@@ -237,21 +237,21 @@ static void test_irq_pin_multiplexer(void)
 
     syscfg_set_irq(0, 1);
 
-    /* Check that irq 0 was set and irq 15 wasn't */
+    /* Check that irq 0 was set and irq 2 wasn't */
     g_assert_true(get_irq(0));
-    g_assert_false(get_irq(15));
+    g_assert_false(get_irq(2));
 
     /* Clean the test */
     syscfg_set_irq(0, 0);
 
-    syscfg_set_irq(15, 1);
+    syscfg_set_irq(2, 1);
 
-    /* Check that irq 15 was set and irq 0 wasn't */
-    g_assert_true(get_irq(15));
+    /* Check that irq 2 was set and irq 0 wasn't */
+    g_assert_true(get_irq(2));
     g_assert_false(get_irq(0));
 
     /* Clean the test */
-    syscfg_set_irq(15, 0);
+    syscfg_set_irq(2, 0);
 }
 
 static void test_irq_gpio_multiplexer(void)
-- 
2.43.2

On Sat, 29 Jun 2024 at 11:46, Inès Varhol <ines.varhol@telecom-paris.fr> wrote:
>
> The QTest `test_irq_pin_multiplexer` makes the assumption that the
> reset state of irq line 15 is low, which is false since STM32L4x5 GPIO
> was implemented (the reset state of pin GPIOA15 is high because there's
> pull-up and it results in the irq line 15 also being high at reset).
>
> It wasn't triggering an error because `test_interrupt` was mistakenly
> "resetting" the line low.
>
> This commit corrects these two mistakes by :
> - not setting the line low in `test_interrupt`
> - using an irq line in `test_irq_pin_multiplexer` which is low at reset
>
> Signed-off-by: Inès Varhol <ines.varhol@telecom-paris.fr>
> ---



Applied to target-arm.next, thanks.

-- PMM