Hi Heinrich,
On Sun, 31 Oct 2021 at 23:21, Heinrich Schuchardt xypron.glpk@gmx.de wrote:
Am 1. November 2021 00:46:49 MEZ schrieb Simon Glass sjg@chromium.org:
On Thu, 28 Oct 2021 at 04:16, Heinrich Schuchardt heinrich.schuchardt@canonical.com wrote:
Avoid a build failure for CONFIG_WDT_GPIO=n.
We need this setting to test watchdog based system reset.
watchdog-based
Signed-off-by: Heinrich Schuchardt heinrich.schuchardt@canonical.com
test/dm/wdt.c | 4 ++++ 1 file changed, 4 insertions(+)
Reviewed-by: Simon Glass sjg@chromium.org
diff --git a/test/dm/wdt.c b/test/dm/wdt.c index ee615f0e14..07a8eb0e7a 100644 --- a/test/dm/wdt.c +++ b/test/dm/wdt.c @@ -44,6 +44,7 @@ static int dm_test_wdt_base(struct unit_test_state *uts) } DM_TEST(dm_test_wdt_base, UT_TESTF_SCAN_PDATA | UT_TESTF_SCAN_FDT);
+#ifdef CONFIG_WDT_GPIO static int dm_test_wdt_gpio(struct unit_test_state *uts) { /* @@ -75,6 +76,7 @@ static int dm_test_wdt_gpio(struct unit_test_state *uts) return 0; } DM_TEST(dm_test_wdt_gpio, UT_TESTF_SCAN_FDT); +#endif
static int dm_test_wdt_watchdog_reset(struct unit_test_state *uts) { @@ -86,9 +88,11 @@ static int dm_test_wdt_watchdog_reset(struct unit_test_state *uts) uint reset_count; int val;
+#ifdef CONFIG_WDT_GPIO
The #ifdef is unfortunate but I believe it is needed due to DM_DRIVER_GET. You could get by name perhaps, or just get the first device?
Thanks for reviewing
We could move the GPIO test to a separate file and make the decision in Makefile. This would avoid #ifdef.
OK, but I don't think it is worth it. We can keep the #ifdef
Regards, Simon
Best regards Heinrich
ut_assertok(uclass_get_device_by_driver(UCLASS_WDT, DM_DRIVER_GET(wdt_gpio), &gpio_wdt)); ut_assertnonnull(gpio_wdt);+#endif ut_assertok(uclass_get_device_by_driver(UCLASS_WDT, DM_DRIVER_GET(wdt_sandbox), &sandbox_wdt)); ut_assertnonnull(sandbox_wdt); -- 2.32.0
Regards, Simon