tests/qtest/npcm7xx_watchdog_timer-test.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
The test_prescaler() part in the npcm7xx_watchdog_timer test is quite
repetive, testing all possible combinations of the WTCLK and WTIS
bitfields. Since each test spins up a new instance of QEMU, this is
rather an expensive test, especially on loaded host systems.
For the normal quick test mode, it should be sufficient to test the
corner settings of these fields (i.e. 0 and 3), so we can speed up
this test in the default mode quite a bit.
Signed-off-by: Thomas Huth <thuth@redhat.com>
---
tests/qtest/npcm7xx_watchdog_timer-test.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/tests/qtest/npcm7xx_watchdog_timer-test.c b/tests/qtest/npcm7xx_watchdog_timer-test.c
index 4773a673b2..981b853c99 100644
--- a/tests/qtest/npcm7xx_watchdog_timer-test.c
+++ b/tests/qtest/npcm7xx_watchdog_timer-test.c
@@ -172,9 +172,10 @@ static void test_reset_action(gconstpointer watchdog)
static void test_prescaler(gconstpointer watchdog)
{
const Watchdog *wd = watchdog;
+ int inc = g_test_quick() ? 3 : 1;
- for (int wtclk = 0; wtclk < 4; ++wtclk) {
- for (int wtis = 0; wtis < 4; ++wtis) {
+ for (int wtclk = 0; wtclk < 4; wtclk += inc) {
+ for (int wtis = 0; wtis < 4; wtis += inc) {
QTestState *qts = qtest_init("-machine quanta-gsj");
qtest_irq_intercept_in(qts, "/machine/soc/a9mpcore/gic");
--
2.43.0Thomas Huth <thuth@redhat.com> writes:
> The test_prescaler() part in the npcm7xx_watchdog_timer test is quite
> repetive, testing all possible combinations of the WTCLK and WTIS
> bitfields. Since each test spins up a new instance of QEMU, this is
> rather an expensive test, especially on loaded host systems.
I'm not against the change but I do not my home machine runs these tests
in:
1/1 qemu:qtest+qtest-arm / qtest-arm/npcm7xx_timer-test OK 0.18s 180 subtests passed
so I do wonder how the system load can cause such a dramatic increase
for a comparatively simple test.
> For the normal quick test mode, it should be sufficient to test the
> corner settings of these fields (i.e. 0 and 3), so we can speed up
> this test in the default mode quite a bit.
>
> Signed-off-by: Thomas Huth <thuth@redhat.com>
> ---
> tests/qtest/npcm7xx_watchdog_timer-test.c | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/tests/qtest/npcm7xx_watchdog_timer-test.c b/tests/qtest/npcm7xx_watchdog_timer-test.c
> index 4773a673b2..981b853c99 100644
> --- a/tests/qtest/npcm7xx_watchdog_timer-test.c
> +++ b/tests/qtest/npcm7xx_watchdog_timer-test.c
> @@ -172,9 +172,10 @@ static void test_reset_action(gconstpointer watchdog)
> static void test_prescaler(gconstpointer watchdog)
> {
> const Watchdog *wd = watchdog;
> + int inc = g_test_quick() ? 3 : 1;
>
> - for (int wtclk = 0; wtclk < 4; ++wtclk) {
> - for (int wtis = 0; wtis < 4; ++wtis) {
> + for (int wtclk = 0; wtclk < 4; wtclk += inc) {
> + for (int wtis = 0; wtis < 4; wtis += inc) {
> QTestState *qts = qtest_init("-machine quanta-gsj");
>
> qtest_irq_intercept_in(qts, "/machine/soc/a9mpcore/gic");
--
Alex Bennée
Virtualisation Tech Lead @ Linaro
On 15/01/2024 12.32, Alex Bennée wrote: > Thomas Huth <thuth@redhat.com> writes: > >> The test_prescaler() part in the npcm7xx_watchdog_timer test is quite >> repetive, testing all possible combinations of the WTCLK and WTIS >> bitfields. Since each test spins up a new instance of QEMU, this is >> rather an expensive test, especially on loaded host systems. > > I'm not against the change but I do not my home machine runs these tests > in: > > 1/1 qemu:qtest+qtest-arm / qtest-arm/npcm7xx_timer-test OK 0.18s 180 subtests passed That's a different test, I think. Look for "watchdog" in its name, please. > so I do wonder how the system load can cause such a dramatic increase > for a comparatively simple test. The watchdog test is executing a new QEMU instance for each test in the loop, and that can be quite expensive on a loaded system, I think. Thomas
Thomas Huth <thuth@redhat.com> writes: > On 15/01/2024 12.32, Alex Bennée wrote: >> Thomas Huth <thuth@redhat.com> writes: >> >>> The test_prescaler() part in the npcm7xx_watchdog_timer test is quite >>> repetive, testing all possible combinations of the WTCLK and WTIS >>> bitfields. Since each test spins up a new instance of QEMU, this is >>> rather an expensive test, especially on loaded host systems. >> I'm not against the change but I do not my home machine runs these >> tests >> in: >> 1/1 qemu:qtest+qtest-arm / qtest-arm/npcm7xx_timer-test OK >> 0.18s 180 subtests passed > > That's a different test, I think. Look for "watchdog" in its name, > please. Ahh -ETOOMANYNPCMS... So yes I get a drop in time by half: 1/1 qemu:qtest+qtest-arm / qtest-arm/npcm7xx_watchdog_timer-test OK 10.35s 15 subtests passed to 1/1 qemu:qtest+qtest-arm / qtest-arm/npcm7xx_watchdog_timer-test OK 4.68s 15 subtests passed Reviewed-by: Alex Bennée <alex.bennee@linaro.org> >> so I do wonder how the system load can cause such a dramatic increase >> for a comparatively simple test. > > The watchdog test is executing a new QEMU instance for each test in > the loop, and that can be quite expensive on a loaded system, I think. > > Thomas -- Alex Bennée Virtualisation Tech Lead @ Linaro
On 15/1/24 08:02, Thomas Huth wrote: > The test_prescaler() part in the npcm7xx_watchdog_timer test is quite > repetive, testing all possible combinations of the WTCLK and WTIS > bitfields. Since each test spins up a new instance of QEMU, this is > rather an expensive test, especially on loaded host systems. > For the normal quick test mode, it should be sufficient to test the > corner settings of these fields (i.e. 0 and 3), so we can speed up > this test in the default mode quite a bit. > > Signed-off-by: Thomas Huth <thuth@redhat.com> > --- > tests/qtest/npcm7xx_watchdog_timer-test.c | 5 +++-- > 1 file changed, 3 insertions(+), 2 deletions(-) Good idea. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
© 2016 - 2024 Red Hat, Inc.