SDK Migration - nrf_log back to experimental in SDK14.2

RE: SDK Migration - nrf_log back to experimental in SDK14.2

MartinBL — Wed, 17 Jan 2018 14:22:15 GMT

Strange, don't know how I missed that one.

Anyway, I found internal reports for both of the bugs. They are different. The one fixed in Sdk 12.3.0 relates to an issue where peer manager under certain circumstances decides to fill up the entire flash with garbage data. The other one relates to a case where pds_peer_data_store() corrupts FDS page offsets.

RE: SDK Migration - nrf_log back to experimental in SDK14.2

Roman — Tue, 16 Jan 2018 06:53:23 GMT

nRF5 SDK v12.3.0: Release Date: Week 19, 2017 - Bugfixes - Drivers and libraries - "Fixed a bug in Peer Manager which made it possible for Peer Manager to corrupt the flash memory under certain circumstances."

RE: SDK Migration - nrf_log back to experimental in SDK14.2

MartinBL — Mon, 15 Jan 2018 15:56:20 GMT

I'm unable to find a mention of any FDS vs Peer manager bug in the SDK 12.3 release notes. Can you point me to where it is mentioned?

I found the internal report of the bug fix in SDK 13.0. It lists SDK 13.0.0-1.alpha and SDK 12.2. as affected versions, so it is plausible that it is the same bug.

RE: SDK Migration - nrf_log back to experimental in SDK14.2

Roman — Tue, 09 Jan 2018 09:18:55 GMT

Hi Martin,

Thanks for your answer. I think, there is no reason to go to 14.x unless I need one of the new features. So far, I only have to decide whether to go to 12.3 or 13.1.

SDK v12.3.0 Release notes say that there has been «Fixed a bug in Peer Manager which made it possible for Peer Manager to corrupt the flash memory under certain circumstances». SDK v13.0.0 Release notes say that there has been « Fixed a bug where Peer Manager corrupted FDS data in flash». Is this a different description of the same bug? Or is this bug not relevant when using “fstorage” instead of “FDS”?

Hmm, I have realized now that there is an other aspect as well: According the release notes of Softdevice S132 v4.0.2, I don’t see a reason for an update. But according nRF52832 Errata Rev 1, SoftDevice S132 v4.0.2 includes workarounds for two anomalies ([102], [106]).

RE: SDK Migration - nrf_log back to experimental in SDK14.2

Austin — Tue, 09 Jan 2018 07:43:44 GMT

Also see devzone.nordicsemi.com/.../

RE: SDK Migration - nrf_log back to experimental in SDK14.2

MartinBL — Tue, 09 Jan 2018 07:25:48 GMT

Hi,

Between SDK 13 and 14 the logger was significantly refactored and extended. It had some stability issues, but most of them are fixed in 14.2. It is also not properly documented yet. Especially with regards to some of the new modules that it is using, like ring buffers and memory handling. That is why we moved it back into experimental.