[
https://issues.apache.org/jira/browse/MYNEWT-536?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15800009#comment-15800009
]
ASF subversion and git services commented on MYNEWT-536:
--------------------------------------------------------
Commit d1af59d32818a2ec079b9475ab595dcc9c3d86b6 in incubator-mynewt-core's
branch refs/heads/develop from [~wes3]
[ https://git-wip-us.apache.org/repos/asf?p=incubator-mynewt-core.git;h=d1af59d
]
MYNEWT-536: mbuf exhaustion can cause supervision timeout
The fix to this bug involved removing the supervision timers and replacing
it with a variable in the connection state machine that records the last
received pdu time (that was valid). This reduced RAM and code usage but
does slightly change when the supervision timeout will actually occur.
The supervision timeout will now occur before the actual timer due
to the fact that we check the supervision timeout at connection
event end. There is no reason to not end the connection at this point
since there is no possible way to receive a data pdu before the timer
would have fired off. There is also the slight possibility that the supervision
timeout will occur slightly later than expected but this would be for a short
amount of time and be unlikely.
> MBUF exhaustion can cause connection supervision timeout
> --------------------------------------------------------
>
> Key: MYNEWT-536
> URL: https://issues.apache.org/jira/browse/MYNEWT-536
> Project: Mynewt
> Issue Type: Bug
> Components: Nimble
> Affects Versions: v1_0_0_beta1
> Reporter: William San Filippo
> Assignee: William San Filippo
> Fix For: v1_0_0_rel
>
>
> If no mbufs are available for reception it is possible that a connection will
> terminate due to a supervision timeout. In the current code, a mbuf needs to
> be available to hand to the LL task in order for the supervision timeout to
> be reset.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
