[ 
https://issues.apache.org/jira/browse/MYNEWT-83?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15368502#comment-15368502
 ] 

ASF subversion and git services commented on MYNEWT-83:
-------------------------------------------------------

Commit 28e2a8b4efcda9d241720f25749137de37ad71e8 in incubator-mynewt-core's 
branch refs/heads/master from [~wes3]
[ https://git-wip-us.apache.org/repos/asf?p=incubator-mynewt-core.git;h=28e2a8b 
]

MYNEWT-83: Controller should use a fixed buffer for reception

In preparation for the rather large code change that will come
with this ticket, I am moving some things from the connection
receive end ISR processing into either the rx start is or the
Link Layer. The reason is that when we use a fixed buffer for
reception we will have to copy the receive buffer into a mbuf.
This will take extra time so we need to do everything we can do
to remove time from the rx end isr.

The changes are pretty simple: move the check for proper access
address into the rx start isr and we can do that there. We also
moved the anchor point processing into the receive start ISR as
we use the receive start time to calculate the anchor point.


> Controller should use a fixed buffer for reception as opposed to mbufs
> ----------------------------------------------------------------------
>
>                 Key: MYNEWT-83
>                 URL: https://issues.apache.org/jira/browse/MYNEWT-83
>             Project: Mynewt
>          Issue Type: Improvement
>          Components: Nimble
>    Affects Versions: v0_8_0_beta1
>            Reporter: William San Filippo
>            Assignee: William San Filippo
>            Priority: Minor
>             Fix For: v0_10_0
>
>
> The current controller code allocates mbufs for reception. During the 
> unplugfest it was noted that is was fairly easy to overload the 
> host/controller with packets. A probably better way to implement this is to 
> have a global receive buffer and we copy out into a mbuf if needed. Note that 
> the transmit portion of the controller already uses a global buffer for 
> transmission.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to