[GitHub] [mynewt-nimble] KKopyscinski commented on issue #586: l2cap coc: Add timer guard on the receiver side

2021-06-09 Thread GitBox


KKopyscinski commented on issue #586:
URL: https://github.com/apache/mynewt-nimble/issues/586#issuecomment-857532339


   The problem with missing frames is solved: both host and controller are 
working fine. Blestress was just configured not well enough: setting 
`BLE_ACL_BUF_SIZE: MYNEWT_VAL_BLE_L2CAP_COC_MPS+4` solves the issue - test is 
not stuck, all data is transferred properly. Basically, we need to use ACL data 
packets that are able to contain MPS and HCI data header. For default 
BLE_L2CAP_COC_MPS size (MSYS_1_BLOCK_SIZE-8 = 286) default BLE_ACL_BUF_SIZE of 
255 is just to small


-- 
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.

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org




[GitHub] [mynewt-nimble] KKopyscinski commented on issue #586: l2cap coc: Add timer guard on the receiver side

2020-08-05 Thread GitBox


KKopyscinski commented on issue #586:
URL: https://github.com/apache/mynewt-nimble/issues/586#issuecomment-669183989


   Ok, so disconnecting channel and returning error is sufficient - as it was. 
About first part, I wonder if it's actually  possible to  lose frames - 
shouldn't that been taken care of by link layer? So packets should be 
retransmitted if they haven't arrieved.



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.

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org




[GitHub] [mynewt-nimble] KKopyscinski commented on issue #586: l2cap coc: Add timer guard on the receiver side

2020-08-05 Thread GitBox


KKopyscinski commented on issue #586:
URL: https://github.com/apache/mynewt-nimble/issues/586#issuecomment-669144577


   I periodically tested my solution with blestress - for larger amounts of 
data, frames were indeed missing, and triggered event was returning used 
credits.
   
   I think at least second part, described by @michal-narajowski, is fine - we 
are disconnecting, but before we disconnect, we return credits. This part 
doesn't use timer.



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.

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org




[GitHub] [mynewt-nimble] KKopyscinski commented on issue #586: l2cap coc: Add timer guard on the receiver side

2020-08-03 Thread GitBox


KKopyscinski commented on issue #586:
URL: https://github.com/apache/mynewt-nimble/issues/586#issuecomment-667840316


   Both cases should be fixed by this patch: 
https://github.com/apache/mynewt-nimble/pull/851



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.

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org