maemo.org Bugzilla – Bug 4874
Publically document CAL config entries in /dev/mtd1 config area
Last modified: 2011-02-13 13:38:06 UTC
You need to log in before you can comment on or make changes to this bug.
As Nokia has reasons to keep libcal (+ cal-tool + wlan-cal + bt-cal) closed, I wrote opensource alternative*. It is already capable of readonly access, adding write mode is just a question of time. However, in order to use CAL data properly, it would be good to have documentation on entries format. For example, currently, open-wlan-cal sets same limits regardless of uk/europe tablet because I don't know how to determine country. Entry header format already known (except for flags = 4), only entries data structure is required. Having things documented will allow better user experience in post-diablo n8x0 life. * http://git.slonopotamus.org/?p=opendsme
Updating version info as libcal is still in nokia-binaries in Fremantle.
This is an enhancement request.
So, in practice this request means that libcal should be open source. The reasons not to offer the information requested and not having the component open are roughly the same. Question: how badly needed is this? I'm trying to prioritize the components that the community wants to see open.
(In reply to comment #3) > So, in practice this request means that libcal should be open source. The > reasons not to offer the information requested and not having the component > open are roughly the same. > > Question: how badly needed is this? I'm trying to prioritize the components > that the community wants to see open. > I do NOT request libcal source since I rewrote alternative to it. I only request documentation of conf blocks data in /dev/mtd1. It's required in order to produce free (as in speech) initfs. Having requested documentation will reduce risks of bricking/damaging/improper use of user devices.
Weighing in on this one: I personally think we should open source libcal. The potential Security/Liability issues with people editing CAL already exists in the fact it is perfectly possible to maim your tablet / edit calibration area etc with libcal-dev from nokia-binaries if you so wanted. It would solve a source closed dependancy of DSME, libcal. Due to the existence of BME data in CAL requiring proper read and write from BME and the fact that charging ability should be possible in alternative OS'es, I would advise that we either open source libcal or document it. I vote for the first - due to this problem I'm facing with BME on N900: * In N8x0 we could just chroot into the initfs to get BME and charging ability. We can't any longer and hence libcal just became a lot more relevant. * BME speaks to DSME. DSME depends on libcal. libcal is closed and hence I can't compile against it in the Mer build system. * BME requires CAL read/write ability in order to store PMM data (see /dev/mtd1 dump). I personally don't want to trust a 3rd party library to store and retrieve this. * If libcal was open sourced, we would be able to build our own DSME, drop in a binary BME and have safe, proven charging and CAL read/writing ability. My initial impression of libcal code is that it looks clean, well-written and presentable and does not contain anything specific to Nokia devices. It is an algorithm on top of a MTD/NAND device as Marat's open replacement also has found it to be. libcal is on the top of my list of things that would be useful to open source and I'd like to work the open sourcing process on this item.
Some news, libcal is getting open sourced as part of the MeeGo on N900 effort.
(In reply to comment #6) > Some news, libcal is getting open sourced as part of the MeeGo on N900 effort. Also, DSME's dependency on libcal is being removed.
Let me know when they're doing the same for sysinfod and libsysinfo. That's where all the good stuff in CAL got moved to (well, it still remains in CAL) for the N900.
(In reply to comment #6) > Some news, libcal is getting open sourced as part of the MeeGo on N900 effort. Ping! Is that still the plan? It's present in MeeGo 1.1, but in the non-oss repo.
Quim: ping - comment 9?
I have asked Sakari Poussa and Mikko Ylinen (MeeGo architects at Nokia), CCing Carsten (maintainer of MeeGo images for N900).
There discussion about providing an equivalent functionality to libcal in MeeGo is being discussed publicly at http://lists.meego.com/pipermail/meego-architecture/2010-November/000016.html & following posts. You are encouraged to get involved in that discussion. Sakari and Mikko have been active in it. The chances of progress are more likely in that MeeGo architecture discussion than here.
Quim: So should this get closed as WONTFIX for Maemo or what is the status?
If we couldn't get any resolution as for yet, now I'm not expecting any release on this for Maemo 5. Any discussion and work needs to happen in the MeeGo context, as said in my previous comment. Therefore yes, resolving as WONTFIX.
FYI, in case it's of use to anyone: the OpenWrt project now has another free implementation in <https://dev.openwrt.org/browser/packages/utils/calvaria>.