[
https://issues.apache.org/jira/browse/MYNEWT-486?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
William San Filippo resolved MYNEWT-486.
----------------------------------------
Resolution: Won't Fix
I am going to resolve this issue for now. While it is possible to address this
issue, we are not going to do it until we see a compelling need. The changes
needed in the controller are not extensive: ble_ll_conn_enqueue_pkt() would
need to take a uint16_t length and the ble header 'offset' field would need to
be made a uint16_t. For controller only blehci, the mbufs would also have to be
sized to accommodate the increased length; not sure how this would impact the
combined host-controller.
The performance improvement with having larger ACL sizes is minimal anyway.
> ACL data packet sizes of 256 or greater will not work.
> ------------------------------------------------------
>
> Key: MYNEWT-486
> URL: https://issues.apache.org/jira/browse/MYNEWT-486
> 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_beta1
>
>
> It appears that there is a bug in the controller. If the ACL data packet
> length is set to greater than 255, the controller will probably fail to send
> the packet. The issue is that the controller uses a uint8_t internally at
> some point and is truncating the packet length.
> There is a work-around for now: setting the BLE_LL_ACL_PKT_SIZE syscfg value
> to 255 or less. This can be done in a target or APP if desired. We will do
> some more testing on this to verify that this is an issue and that the
> work-around is functional. The code will also be fixed in the beta 1 release.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
