diff options
Diffstat (limited to 'tests')
| -rw-r--r-- | tests/ports/rp2/rp2_machine_idle.py | 36 | ||||
| -rw-r--r-- | tests/ports/rp2/rp2_machine_idle.py.exp | 1 |
2 files changed, 37 insertions, 0 deletions
diff --git a/tests/ports/rp2/rp2_machine_idle.py b/tests/ports/rp2/rp2_machine_idle.py new file mode 100644 index 000000000..f9c282847 --- /dev/null +++ b/tests/ports/rp2/rp2_machine_idle.py @@ -0,0 +1,36 @@ +import machine +import time + +# Verify that machine.idle() resumes execution within 0.1 and 1.1ms (should be +# 1ms max but allowing for some overhead). +# +# (A minimum sleep time for machine.idle() isn't specified but in a system like +# this with no active interrupts then we should expect some idle time before +# resuming. If it's consistently resuming immediately then that indicates a bug +# is preventing proper idle.) +# +# This test doesn't contain any rp2-specific code, but rp2 is currently the only +# tickless port - which is what led to the bug this is a regression test for. +# Some other ports (unix, esp32) have idle behaviour that resumes immediately on +# a quiet system, so this test is also less useful for those. +# +# Verification uses the average idle time, as individual iterations will always +# have outliers due to interrupts, scheduler, etc. + +ITERATIONS = 500 +total = 0 + +for _ in range(ITERATIONS): + before = time.ticks_us() + machine.idle() + total += time.ticks_diff(time.ticks_us(), before) + +total /= 1000 # us to ms +average = total / ITERATIONS + +# print(f"Total {total}ms average {average}ms") # uncomment for debug + +if 0.1 < average < 1.1: + print("PASS") +else: + print(f"Total {total}ms average {average}ms, out of spec") diff --git a/tests/ports/rp2/rp2_machine_idle.py.exp b/tests/ports/rp2/rp2_machine_idle.py.exp new file mode 100644 index 000000000..7ef22e9a4 --- /dev/null +++ b/tests/ports/rp2/rp2_machine_idle.py.exp @@ -0,0 +1 @@ +PASS |