Prevent time comparison overflow in S.Port driver #3536
Conversation
src/main/telemetry/smartport.c
Outdated
| @@ -538,7 +538,11 @@ void handleSmartPortTelemetry(void) | |||
| } | |||
| } | |||
|
|
|||
| if (cmpTimeUs(micros(), lastTelemetryFrameReceivedUs) >= SMARTPORT_MIN_TELEMETRY_RESPONSE_DELAY_US) { | |||
| // XXX: Checking for lastTelemetryFrameReceivedUs == 0 is needed for S.Port, since payload will | |||
| // only be non-null (and hence, lastTelemetryFrameReceivedUs will be updated) when using FPort. | |||
There was a problem hiding this comment.
On closer inspection, this is actually wrong. Payload is pointing to the incoming payload for both SmartPort and FPort:
https://github.com/iNavFlight/inav/pull/3536/files#diff-5d6c27b6b68ea6ef24a6103f1bc2ced4L535
A correct fix would be to change
https://github.com/iNavFlight/inav/pull/3536/files#diff-5d6c27b6b68ea6ef24a6103f1bc2ced4R536
to
if (payload || clearToSend) {
Alternatively (and maybe even a bit cleaner), lastTelemetryFrameReceivedUs should be updated inside smartPortDataReceive().
Also, n.b. that handleSmartPortTelemetry() isn't run for FPort.
There was a problem hiding this comment.
I thought about that, but clearToSend might reset lastTelemetryUs if we get an stray byte due to intererence. I wanted to fix it with minimal changes so it can go into inav 2.0
fiam
force-pushed
the
agh_fix_smartport_time_overflow
branch
from
a9183a5 to
8909396
Compare
|
👍 |
fiam
force-pushed
the
agh_fix_smartport_time_overflow
branch
from
8909396 to
74e5ae8
Compare
Would stop S.Port telemetry after ~36 minutes.
Fixes #3282