[wifi_matfunc] network_WiFi_OverlappingBSSScan fails on (Daisy and Daisy_skate) Marvell boards. |
|||
Issue descriptionThe test fails with the following error message "Significant difference in rtt due to OBSS" https://stainless.corp.google.com/search?view=list&first_date=2018-01-20&last_date=2018-02-16&suite=wifi_matfunc&test=network_WiFi_OverlappingBSSScan&board=daisy&status=WARN&status=FAIL&status=ERROR&status=ALERT&status=RUNNING&status=TEST_NA&status=NOSTATUS&exclude_cts=true&exclude_not_run=false&exclude_non_release=true&exclude_au=true&exclude_acts=true&exclude_retried=true&exclude_non_production=true
,
Nov 7
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/autotest/+/39eca23bb752e778096ff82528d82398d163cf2c commit 39eca23bb752e778096ff82528d82398d163cf2c Author: Brian Norris <briannorris@chromium.org> Date: Wed Nov 07 19:40:57 2018 [autotest] network_WiFi: raise dwell time thresholds network_WiFi_BgscanBackoff and network_WiFi_OverlappingBSSScan often fail excessively with warnings like this: "Significant difference in rtt due to OBSS" It appears that some Wifi drivers have a dwell time closer to 150 ms, so it's trivially easy to be waiting for one channel to finish scanning while traffic is pending, and then see >150 ms latency. Let's push this to 250 ms, to give appropriate margin for a variety of reasonable drivers. This will still catch excessive delays (e.g., if the driver continues to process several more channel scans while data traffic is pending) but avoid complaining about reasonable delays for a single channel scan. Also, it doesn't make a lot of sense to have different threshold parameters for network_WiFi_BgscanBackoff and network_WiFi_OverlappingBSSScan, so make them the same. BUG= chromium:812942 , b:68970426 TEST=run this test on a variety of boards Change-Id: I6cf82625be880ccb829adc57d9e0dd8aef761d6c Signed-off-by: Brian Norris <briannorris@chromium.org> Reviewed-on: https://chromium-review.googlesource.com/1302973 Reviewed-by: Kirtika Ruchandani <kirtika@chromium.org> Reviewed-by: Grant Grundler <grundler@chromium.org> [modify] https://crrev.com/39eca23bb752e778096ff82528d82398d163cf2c/server/site_tests/network_WiFi_BgscanBackoff/network_WiFi_BgscanBackoff.py [modify] https://crrev.com/39eca23bb752e778096ff82528d82398d163cf2c/server/site_tests/network_WiFi_OverlappingBSSScan/network_WiFi_OverlappingBSSScan.py
,
Nov 7
Let's monitor this one. I'm not sure if these Marvell boards reliably stay within the new threshold.
,
Nov 16
Daisy has been solid since then. Overall, the test is fairly stable across platforms. There are some odd failures which could be worth looking into someday, but nothing as bad as this. |
|||
►
Sign in to add a comment |
|||
Comment 1 by harpreet@chromium.org
, Feb 16 2018