By allowing the max timer backoff duration to be limited, a socket can
fast recover in case of intermittent network issues. Before this CL, the
exponential backoff algorithm could result in very long retry durations
(in the order of minutes), when connection has been lost or been flaky
for a long while.
Note that limiting the maximum backoff duration might require
compensating the maximum retransmission limit to avoid closing the
socket prematurely due to reaching the maximum retransmission limit much
faster than previously.
Bug: webrtc:13129
Change-Id: Ib94030d666433e3fa1a2c8ef69750a1afab8ef94
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/230702
Reviewed-by: Florent Castelli <orphis@webrtc.org>
Commit-Queue: Victor Boivie <boivie@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#34913}
The restart limit for timers can already be limitless, but the
RetransmissionErrorCounter didn't support this. With this change, the
max_retransmissions and max_init_retransmits can be absl::nullopt to
indicate that there should be infinite retries.
Bug: webrtc:13129
Change-Id: Ia6e91cccbc2e1bb77b3fdd7f37436290adc2f483
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/230701
Reviewed-by: Florent Castelli <orphis@webrtc.org>
Commit-Queue: Victor Boivie <boivie@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#34882}
This was caught in an integration test which had stricter assertions
than the FakeTimeout which is used in unit tests, so the first thing was
to add the same assertions to the FakeTimeout.
The issue is that when a Timer triggers, and if it's set to
automatically restart (possibly with an exponential backoff), the
`is_running_` field was set to true while the timer callback was called
to allow the client to know that the timer is in fact running, but the
timer was actually not started until the callback returned. Which made
sense, as the callback can with its return value override the duration,
which should affect the backoff algorithm.
The problem was when a timer was manually started within the callback.
As the Timer itself thought that it was already running, it first would
Stop() the underlying Timeout, then Start(). But calling Stop() on a
timeout that is not running is illegal, which set of assertions.
So the solution is to don't lie; Don't say that a timer is running when
it's not. Make sure that the timer is running when the timer callback is
triggered, which makes it consistent at all times. That may result in
unnecessary timeout invocations (stopping and starting), but that's not
too expensive.
Bug: webrtc:12614
Change-Id: I7b4447ccd88bd43d181e158f0d29b0770c8a3fd6
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/217522
Reviewed-by: Florent Castelli <orphis@webrtc.org>
Commit-Queue: Victor Boivie <boivie@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33926}
Ensuring that timer durations never go beyond a safe maximum duration
and that timer IDs are not re-used.
Bug: webrtc:12614
Change-Id: I227a2e9933da16669dc6ea0a39c570892010ba2c
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/215063
Commit-Queue: Victor Boivie <boivie@webrtc.org>
Reviewed-by: Tommi <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33860}
Timer is a high-level timer (in contrast to the low-level `Timeout`
class). Timers are started and can be stopped or restarted. When a timer
expires, the provided callback will be triggered.
Timers can be configured to do e.g. exponential backoff when they expire
and how many times they should be automatically restarted.
Bug: webrtc:12614
Change-Id: Id5eddd58dd0af62184b10dd1f98e3e886e3f1d50
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/213350
Reviewed-by: Tommi <tommi@webrtc.org>
Commit-Queue: Victor Boivie <boivie@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33666}
