You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@mynewt.apache.org by GitBox <gi...@apache.org> on 2022/06/27 05:03:18 UTC
[GitHub] [mynewt-nimble] sada45 opened a new issue, #1291: The connection subrating needs CONN_SUBRATING_HOST feature
sada45 opened a new issue, #1291:
URL: https://github.com/apache/mynewt-nimble/issues/1291
Hi!
I have tried the connection subrating in the Spec 5.3 and may found an issue.
Before the central sends the `SUBRATE_IND`, it will check wether the peripheral has the `BLE_LL_FEAT_CONN_SUBRATING_HOST`. The corresponding code is in `ble_ll.c` Line 4131. However, the peripheral will not initiate this feature, so the `ble_ll_conn_subrate_req_hci()` always return `-ENOTSUP`.
A quick fix is modify the function `ble_ll_init()` at `ble_ll.c` Line 1965 to:
```C
#if MYNEWT_VAL(BLE_LL_CFG_FEAT_LL_ENHANCED_CONN_UPDATE)
features |= BLE_LL_FEAT_CONN_SUBRATING;
features |= BLE_LL_FEAT_CONN_SUBRATING_HOST;
#endif
```
By the way, it seems that there is no connection subrating API in the GAP layer, is this normal?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@mynewt.apache.org.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [mynewt-nimble] sada45 commented on issue #1291: The connection subrating needs CONN_SUBRATING_HOST feature
Posted by GitBox <gi...@apache.org>.
sada45 commented on issue #1291:
URL: https://github.com/apache/mynewt-nimble/issues/1291#issuecomment-1182959619
I am happy that helps!
I take a look at the fix. I think there is one more bug.
When checking the continuation number, it should ignore the connection events before the last base event.
An example is shown in Fig 4.42, the event `n+11` is skipped.
Therefore, my fix is shown as follow (I rename the `last_pdu_event` to `last_data_event`):
```C
if ((connsm->cont_num > 0) &&
(connsm->event_cntr + 1 != connsm->subrate_base_event +
connsm->subrate_factor) &&
(connsm->last_data_event >= connsm->subrate_base_event) &&
(connsm->event_cntr - connsm->last_data_event < connsm->cont_num)) {
next_is_subrated = 0;
}
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@mynewt.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [mynewt-nimble] sada45 commented on issue #1291: The connection subrating needs CONN_SUBRATING_HOST feature
Posted by GitBox <gi...@apache.org>.
sada45 commented on issue #1291:
URL: https://github.com/apache/mynewt-nimble/issues/1291#issuecomment-1182782695
Besides, currentlly, the `connsm->last_pdu_event` will be updated once received a PDU no matter it is empty or not.
However, this value should be only be updated if there is data transmitted in current connection event.
The spec give the defination of continuation event in `Vol.6 PartB 4.5.1`:
_A continuation event is a connection event where, in at least one of the previous connContinuationNumber connection events (ignoring any before the last subrated connection event), at least one packet was transmitted or validly received containing a Link Layer PDU **with a non-zero Length field**._
The figure 4.41 in `Vol.6 PartB 4.5` gives an example for this:
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@mynewt.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [mynewt-nimble] sada45 commented on issue #1291: The connection subrating needs CONN_SUBRATING_HOST feature
Posted by GitBox <gi...@apache.org>.
sada45 commented on issue #1291:
URL: https://github.com/apache/mynewt-nimble/issues/1291#issuecomment-1183854288
Yes, that sounds reasonable to me.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@mynewt.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [mynewt-nimble] sada45 commented on issue #1291: The connection subrating needs CONN_SUBRATING_HOST feature
Posted by GitBox <gi...@apache.org>.
sada45 commented on issue #1291:
URL: https://github.com/apache/mynewt-nimble/issues/1291#issuecomment-1179503065
Thanks for your reply! I'll just use the connection subrating with link-layer APIs for now.
Here I found another question.
If the BLE works in the noisy RF environment, the keep-alive packet may be lost, and the link-layer chooses to retransmit the empty in the next connection event.
However, when there is a data packet enqueued, the link-layer still choose to retransmit the empty packet, and the data packets will be stalled.
Is this policy defined in the Bluetooth spec. or necessary?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@mynewt.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [mynewt-nimble] andrzej-kaczmarek commented on issue #1291: The connection subrating needs CONN_SUBRATING_HOST feature
Posted by GitBox <gi...@apache.org>.
andrzej-kaczmarek commented on issue #1291:
URL: https://github.com/apache/mynewt-nimble/issues/1291#issuecomment-1183559883
Checking SN in `ble_ll_conn_rx_isr_end` does only mean that previously txd PDU was acked, but the spec does not say anything about acking, it just states that a valid non-empty PDU has to be txd/rxd:
> A continuation event is a connection event where, in at least one of the previous connContinuationNumber connection events (ignoring any before the last subrated connection event), **at least one packet was transmitted or validly received containing a Link Layer PDU with a non-zero Length field**.
So `has_nonempty_pdu` is set in two cases:
- for tx, when `ble_phy_tx` for non-empty PDU was successful, meaning it will be txd
- for rx, in `ble_ll_conn_rx_data_pdu` when rxd PDU was determined to be valid and is non-empty
There's one small catch for tx: in theory it's possible that radio will be disabled prior/during tx so in fact PDU will not be transmissted, but at the moment we do not handle this properly; also it likely won't be even noticed irl.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@mynewt.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [mynewt-nimble] andrzej-kaczmarek commented on issue #1291: The connection subrating needs CONN_SUBRATING_HOST feature
Posted by GitBox <gi...@apache.org>.
andrzej-kaczmarek commented on issue #1291:
URL: https://github.com/apache/mynewt-nimble/issues/1291#issuecomment-1166976101
Host supported features can be set by `HCI LE Set Host Feature`. NimBLE host does not have support for connection subrating so it does not set that bit.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@mynewt.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [mynewt-nimble] sada45 commented on issue #1291: The connection subrating needs CONN_SUBRATING_HOST feature
Posted by GitBox <gi...@apache.org>.
sada45 commented on issue #1291:
URL: https://github.com/apache/mynewt-nimble/issues/1291#issuecomment-1181833269
I test the connection subrating but find the `Continuation Number` does not work.
I think I find the reason. In the function `ble_ll_conn_next_event()` in `ble_ll_conn.c`, Line 2275. The central and peripheral decide wether they should turn on radio on the next connection or the next subrate base event.
Accroding to the Bluetooth spec. If the next connection event is not a base event and the connection event number since we last received a PDU is less than `Continuation Number`, nodes will turn on their radio in the next connection even, rather than the next base event.
So, to my understanding, the correct code should be:
```C
if ((connsm->cont_num > 0) &&
/* (connsm->event_cntr + 1 == connsm->subrate_base_event + */
(connsm->event_cntr + 1 != connsm->subrate_base_event +
connsm->subrate_factor) &&
(connsm->event_cntr - connsm->last_pdu_event < connsm->cont_num)) {
next_is_subrated = 0;
}
```
Looking forward to your reply :-) @andrzej-kaczmarek
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@mynewt.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [mynewt-nimble] andrzej-kaczmarek commented on issue #1291: The connection subrating needs CONN_SUBRATING_HOST feature
Posted by GitBox <gi...@apache.org>.
andrzej-kaczmarek commented on issue #1291:
URL: https://github.com/apache/mynewt-nimble/issues/1291#issuecomment-1182943669
@sada45 good catch, thanks. I created a fix for both https://github.com/apache/mynewt-nimble/pull/1297.
I still need to validate those more carefully since apparently software we use to run qualification test cases do not implement those test cases properly and they pass in all cases...
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@mynewt.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [mynewt-nimble] andrzej-kaczmarek commented on issue #1291: The connection subrating needs CONN_SUBRATING_HOST feature
Posted by GitBox <gi...@apache.org>.
andrzej-kaczmarek commented on issue #1291:
URL: https://github.com/apache/mynewt-nimble/issues/1291#issuecomment-1183325249
@sada45 I reworked continuation events handling a bit more, it should be easier to understand now and I believe it should now be handled properly. let me know how it works for you.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@mynewt.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [mynewt-nimble] andrzej-kaczmarek commented on issue #1291: The connection subrating needs CONN_SUBRATING_HOST feature
Posted by GitBox <gi...@apache.org>.
andrzej-kaczmarek commented on issue #1291:
URL: https://github.com/apache/mynewt-nimble/issues/1291#issuecomment-1183554896
The spec does not say anything about acking, it just states that a valid non-empty PDU has to be txd/rxd:
`has_nonempty_pdu` is updated in `ble_ll_conn_rx_data_pdu` when PDU was
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@mynewt.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [mynewt-nimble] sada45 commented on issue #1291: The connection subrating needs CONN_SUBRATING_HOST feature
Posted by GitBox <gi...@apache.org>.
sada45 commented on issue #1291:
URL: https://github.com/apache/mynewt-nimble/issues/1291#issuecomment-1183349582
@andrzej-kaczmarek I have some doubts that wether the `has_nonempty_pdu` should be set before or after the packet is acked.
For me, I update the `last_data_pdu` in `ble_ll_conn_rx_isr_end`, after the code checked the CRC and SN, which means this data PDU is successfully transmitted and acked.
However, the Spec. has not given a clear explanation about that.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@mynewt.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org