commit d56252bf6dea55d97d29e4842977441657f42c9b
author Andres Oportus <andresoportus@google.com> Tue Jun 20 13:25:50 2023 +0000
committer CQ Bot <fuchsia-internal-scoped@luci-project-accounts.iam.gserviceaccount.com> Tue Jun 20 13:25:50 2023 +0000
tree 0a1d35db5e9086519c71ee7df8e0b9ea34613365
parent 1094f6c74f0978ecc9c7900bd15c8e38883c1dca

[audio] Fix flake if virtual audio transfer bytes is non-zero

Before this CL, the virtual audio StreamConfig driver would
report a start time driver_transfer_bytes in the future with
the intent of starting to report position notifications after
the time equivalent to driver_transfer_bytes has passed,
this should not be needed per clarification in
fxrev.dev/871737.

The position reported by virtual audio as part of position
notifications should not start at a negative value, we have
assertions to make sure this is the case. Before this CL
the first position notification was intended to happen after
a time equivalent to driver_transfer_bytes has passed and
hence it would have been position 0+. However, the virtual
audio FIDL API allows for updating the position notification
frequency via the SetNotificationFrequency API call, and
in e2e tests this would change the timestamp used to report
the position notification. In fxrev.dev/858124 we
changed driver_transfer_bytes for e2e tests from 0 to 48
(the default value provided by the driver), this would
change the timestamp used to report the position
notification sometimes to a value before
driver_transfer_bytes had passed, triggering the assertion
mentioned above manifesting in the flake.

By not returning a start time driver_transfer_bytes in the
future and also not setting the start of position
notifications in the future, we both comply with the
clarifications in fxrev.dev/871737 and avoid the flake
condition since a SetNotificationFrequency could not result
in a position with a negative value anymore.

Fixed: 129056
Test: host_x64/knowledge_card_test
(with driver_transfer_bytes=1024 to make the failure deterministic)

Change-Id: I287762b227368a8130f247ca524f83720f90b435
Reviewed-on: https://fuchsia-review.googlesource.com/c/fuchsia/+/871738
Commit-Queue: Andres Oportus <andresoportus@google.com>
Reviewed-by: Cam Lloyd <camlloyd@google.com>
Reviewed-by: Martin Puryear <mpuryear@google.com>