Custom TWIS DFU transport in SDK15 on S132v6: Hash verification fails in TWIS - passes in BLE.

RE: Custom TWIS DFU transport in SDK15 on S132v6: Hash verification fails in TWIS - passes in BLE.

tian bnb — Fri, 08 Mar 2024 09:21:14 GMT

Hello, could you please send the final debugging case to me for reference?

RE: Custom TWIS DFU transport in SDK15 on S132v6: Hash verification fails in TWIS - passes in BLE.

lesher — Mon, 30 Apr 2018 22:07:55 GMT

That was it. Now to clean things up (react to TWI rx done in the controller instead of delaying).

RE: Custom TWIS DFU transport in SDK15 on S132v6: Hash verification fails in TWIS - passes in BLE.

lesher — Mon, 30 Apr 2018 19:41:37 GMT

Looks like it was a mistake to stop using a memory pool for the transferred data (while still keeping queued flash writes). I'm using an MTU of 129 (128 data Bytes plus a command). It appears the 6th MTU is being programmed into the beginning of app flash. Now I understand that the CRC response (both to PRN-expired and to the direct command) applies to the data transferred, not to the flash. Will attempt to bring back the balloc pool, now.

RE: Custom TWIS DFU transport in SDK15 on S132v6: Hash verification fails in TWIS - passes in BLE.

lesher — Sat, 28 Apr 2018 00:23:10 GMT

Here is the source (copied from other locations, so it won't build in place).

It is a first-pass, and is not very clean - but it produces the results above.devzone.nordicsemi.com/.../nrf_5F00_dfu_5F00_twis.zip