Hi all,
I'm experiencing an issue by using the GATT_WriteLongCharValue () function to send 244 bytes from a central device to
a peripheral device. The steps performed by the central device are the following:
1) GAPCentralRole_EstablishLink () => event GAP_LINK_ESTABLISHED_EVENT received with no errors
2) GATT_DiscPrimaryServiceByUUID () => event ATT_FIND_BY_TYPE_VALUE_RSP received with no errors
3) GATT_DiscAllChars () => event ATT_READ_BY_TYPE_RSP received with no errors
4) GATT_WriteCharValue() to send one byte to the peripheral device => event ATT_WRITE_RSP received with no errors
5) GATT_WriteLongCharValue () to send the first chunk of data to the peripheral device => no events received by the BLE stack,
communication interrupted with the peripheral device (no more packets exchange visible with the sniffer)
I've tried to replace the GATT_WriteLongCharValue () with ATT_WriteLongCharValue () and I'm able to see with the sniffer the
packets "ATT PREPARE WRITE REQ" from central to peripheral and "ATT PREPARE WRITE RESP" from peripheral to central
but no indications have been raised by the BLE stack.
Is there a bug in the BLE stack ? Have somebody experienced that issue ? How could I think about a workaround ?
Best Regards
/Alessandro
Hi, Alessandro
Too bad I have only passing experience with WriteLongCharValue, but as it looks, possible chokepoints are
1) (off-chance) for some reason, WriteLongCharValue simply fails to start (returns error status) - but you'd probably had it caught if it happened.
2) A bug in profile description that prevents Central from going on with exchange.
Does it happen when pocking at peripheral via BTool, too?
BR,
Oleg
Hi Oleg,
I haven't tried with BTool yet. The call to GATT_WriteLongCharValue () always returns SUCCESS but it causes the BLE stack
to hang up as I wrote in my post. I suspect a bug in the BLE stack since another guy posted a similar issue in the forum
without anyone from TI who replied to him. I'm going to divide my 244 bytes packet in order to be able to use GATT_WriteCharValue ()
which seems to work well.
Thank you for your reply.
I see, then (1) is out of question.
I've tried calling GATT_WriteLongCharValue and GATT_ReadLongCharValue counterpart from BTool. They both worked ok when communicating with my prototype board.
However, those reads/write were done from/to attribute which is only 16 bytes long and has offset processing code deliberately ripped out from GATT profile. So they are of course not representative of most interesting part of LongCharValue procedure (splitting/reassembling data array).