[GitHub] haukepetersen opened a new issue #301: L2CAP: static size of BLE_L2CAP_COC_MTU inefficient?!

[GitBox <gi...@apache.org>]

haukepetersen opened a new issue #301: L2CAP: static size of BLE_L2CAP_COC_MTU inefficient?!
URL: https://github.com/apache/mynewt-nimble/issues/301
 
 
   Lately, I was playing with some stress-tests for sending data over raw l2cap connection oriented channels, and I had a lot of trouble with overflowing buffers. It took me a while to understand NimBLE's internal buffering for l2cap, but I think I have it figured out...
   
   However, when looking more closely into the code, I found that `BLE_L2CAP_COC_MTU` in `ble_l2cap_coc_priv.h` is statically set to `100`. If I understand it correctly, this value is only used for the computation of the initial credits for a connection (in `ble_l2cap_coc_chan_alloc()`). These credits then define the maximum number of 'mbuf's, that are allocated in a row from the `msys` pool, when `ble_l2cap_coc_continue_tx()` is called.
   
   Example:
   - open a l2cap channel with a MTU of 5000 byte -> the channel will initial have 50 credits `(5000 + (BLE_L2CAP_COC_MTU - 1)) / BLE_L2CAP_COC_MTU`
   - now I want to send a chunk 5000 byte through the channel:
     - put the 5000 byte into an application specific `mbuf` pool
     - hand the data to NimBLE using `ble_l2cap_send()`
     - `ble_l2cap_coc_continue_tx()` is called. This function will allocate 50 `mbuf`s from the `msys` pool, putting `BLE_L2CAP_COC_MTU -> 100` byte of user data into each `mbuf`
   
   -> on RIOT, `MYNEWT_VAL_MSYS_1_BLOCK_SIZE` is per default set to `292`. This leads to a lot of wasted memory, as each of the 50 `mbuf`s allocated above uses only 100 of these 292 byes, leaving 50 * ~190 byte -> ~9.5kb of RAM unused
   
   So is there any reason I missed for setting `BLE_L2CAP_COC_MTU` statically to `100`? I would suggest to set it simply to the value of `MYNEWT_VAL_MSYS_1_BLOCK_SIZE`, possibly minus a little overhead. Would that make sense?

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services