Searching informations (was Re: dsme, Re: Would like to know the status of (was RE: Corporate ownership of open source projects [LWN]))
Thats a great site for searching the archives BTW, thanks! Allow me another plug for tabletSearch: http://www.google.com/coop/cse?cx=014783630360138230012%3A5hwd9yj0uno This is a Google Custom Search Engine (CSE) that I put up and maintain. It allows specific searches for Maemo/NIT-related information; I find it extremely handy whenever I look for anything in this domain. As an example of your needs, just search for dsme and look at the results :-) In addition, I also made a plugin for the Internet Search Applet on the Hildon Home screen. You can get it here: http://tomch.com/wp/?page_id=63 Best regards -Tom ___ maemo-developers mailing list maemo-developers@maemo.org https://lists.maemo.org/mailman/listinfo/maemo-developers
Re: dsme, Re: Would like to know the status of (was RE: Corporate ownership of open source projects [LWN])
Hello! On Wed, May 7, 2008 at 3:45 PM, Frantisek Dufka [EMAIL PROTECTED] wrote: nick loeve wrote: Yep, I have been reading the archives and have compiled a shortlist of info based on past discussions (mainly that you have started! ). I guess I might start documenting what is known already and try and get it all in one place. Yes, please do if you can. https://maemo.org/community/wiki/poweranddevicemanagement/ At the moment it is just a link dump and i intend to summarize some of the info in that page. If you anyone has more useful links/info just dump them on the page or send them in this thread. Cheers I am guilty of not doing it myself over years. I'm sure people on this list can fill some gaps. Apart from examining dsme, its modules and dsmetest in initfs (and its runtime behaviour using strace), check also powerlaunch http://powerlaunch.garage.maemo.org/ It is replacement of mce (also closed). mce talks to dsme too and is done in similar modular way so it can help with some understanding. Powerlaunch imitates some mce-dsme communication, see dsme.c,h at https://garage.maemo.org/plugins/scmsvn/viewcvs.php/trunk/powered/?root=powerlaunch As for bme and figuring out how battery charging works, one could strace it and see what tahvo/retu register it uses, some bits are mentioned in kernel headers (not sure about name now, something like tahvo_retu_user.h) and there is also this info http://marc.info/?l=linux-omapm=120752529631862w=2 Frantisek -- Nick Loeve ___ maemo-developers mailing list maemo-developers@maemo.org https://lists.maemo.org/mailman/listinfo/maemo-developers
Would like to know the status of (was RE: Corporate ownership of open source projects [LWN])
Hi Dave, Let's use this list as a template, then. Can people identify very specific things (and please, assume I don't know what the acronyms mean, and don't know the history behind the issues ;) and I'll get started on a wiki page? Okay, I'll get started then: 1. DSP SBC task - In the roadmap there used to be (what I interpreted as) mention of implementing a DSP task to handle encoding audio for A2DP use (where A2DP stands for Advanced Audio Distribution Protocol and is a Bluetooth protocol for sending data to Bluetooth stereo headsets [1]). SBC stands for subband codec, which is the default encoding method required by A2DP hardware. At the moment this encoding must be carried out by the ARM in software, but if the task could be offloaded to the DSP it would free up the ARM to do other things (such as decoding video better). This is really just a yes/no, are Nokia doing anything as if they are, I won't bother. 2. PowerVR MBX driver [2] and OpenGL implementation. We have PowerVR MBX chips sat in our N8x0 machines, doing, afaict, not a lot. It would be nice if it worked (think OpenGL loveliness in games, and I would hope some frivolous but eye catching snazzy desktop features - prettier UI = more publicity = more sales; perhaps...). There are Linux 2.6.x kernel drivers available, though not for the right kernel version and for the OMAP2430 rather than OMAP2420. It would be nice to know if Nokia are pursuing this, and if not, what the limitation is - if it's hardware could we know, as otherwise we might try fiddling with the existing driver and see if we get anywhere (though that will be a long road). 3. IVA - this is part of the OMAP2420 and stands for Imaging and Video Accelerator chip. Its abilities are described at the end of this document [3], in the specifications. It looks like it could do some useful acceleration. There is however no real information available (from Ti or elsewhere) about what it is, and how to program it, though there are some hints, e.g. [4]. The DSP gateway (which is the bridge software between the ARM and the DSP) has been expanded to allow it to interface with the IVA, but the patches say that there's no driver for the IVA itself [5]. Is this being pursued (as there's little we can do about it as we have no information) and if not, why not? 4. AGPS - Assisted GPS [6]. We have an AGPS chip in the N8x0, it doesn't use the assistance feature. Is this being worked on? This is something we as the community could do some work on, if we had some sort of API to allow us to inject the updated almanac/ephemeris data (leave how to work these out as part of the challenge ;)). Obviously if Nokia come out with an all encompassing system (e.g. giving us a way of talking to supl.nokia.com to provide us with the almanac/ephemeris data that I presume it provides for mobile phones), then that would be good (though probably less fun for hacking purposes ;)) And last but not least, all my questions should be read with pretty please added to them rather than as demands, and thanks (to Dave and Nokia) for taking this on. Cheers, Simon [1] http://en.wikipedia.org/wiki/A2DP#Advanced_Audio_Distribution_Profile_.28A2D P.29 [2] http://www.imgtec.com/factsheets/powervr/POWERVR_MBX_IP_Core_Family_%5B1.0%5 D.pdf [3] http://focus.ti.com/pdfs/wtbu/TI_omap2420.pdf [4] http://osdir.com/ml/linux.ports.arm.omap/2006-08/msg00177.html [5] http://osdir.com/ml/ports.arm.omap/2006-09/msg00067.html [6] http://en.wikipedia.org/wiki/AGPS ___ maemo-developers mailing list maemo-developers@maemo.org https://lists.maemo.org/mailman/listinfo/maemo-developers
Re: Would like to know the status of (was RE: Corporate ownership of open source projects [LWN])
On May 7, 2008, at 14:18, Simon Pickering wrote: 1. DSP SBC task - In the roadmap there used to be (what I interpreted as) mention of implementing a DSP task to handle encoding audio for A2DP use (where A2DP stands for Advanced Audio Distribution Protocol and is a Bluetooth protocol for sending data to Bluetooth stereo headsets [1]). SBC stands for subband codec, which is the default encoding method required by A2DP hardware. At the moment this encoding must be carried out by the ARM in software, but if the task could be offloaded to the DSP it would free up the ARM to do other things (such as decoding video better). This is really just a yes/no, are Nokia doing anything as if they are, I won't bother. AFAIK no. However, you might want to take a look at the recent optimizations that have been made to the SBC codec that comes as part of BlueZ. Several of the optimizations specifically target arm so the value of pushing the encoding to the DSP side is less than it was e.g. one year ago (however it might still have enough value for being worth doing if you have the time and interest). Johan ___ maemo-developers mailing list maemo-developers@maemo.org https://lists.maemo.org/mailman/listinfo/maemo-developers
Re: Would like to know the status of (was RE: Corporate ownership of open source projects [LWN])
Hi! On Wed, May 7, 2008 at 1:18 PM, Simon Pickering [EMAIL PROTECTED] wrote: Hi Dave, Let's use this list as a template, then. Can people identify very specific things (and please, assume I don't know what the acronyms mean, and don't know the history behind the issues ;) and I'll get started on a wiki page? Okay, I'll get started then: Here is my list (which only outlines subjects not touched below): 1. Binary only wifi modules Is there any work being done to make this open completely (definitively)? Or is it possible for Nokia to supply it in object form or at least updated as the kernel API changes over time? 2. Is it possible to at least get some basic documentation of what bme and dsme are doing (beyond obvious basics)? That way if the code from Nokia cannot be opened we have a starting point for an alternative implementation. 3. (my own personal interest) is it possible to document the serial pins for the n8x0 devices? Thats if they differ from the 770 which is documented (kind of, its hidden on a wiki page somewhere). Thats my questions... Im looking forward to the irc meeting and also talking to Nokia/maemo people at linuxtag. By the way, thank you Nokians for being as open as you have and creating a great device. I really see that there is room for compromise from both sides in the recent discussions. Cheers 1. DSP SBC task - In the roadmap there used to be (what I interpreted as) mention of implementing a DSP task to handle encoding audio for A2DP use (where A2DP stands for Advanced Audio Distribution Protocol and is a Bluetooth protocol for sending data to Bluetooth stereo headsets [1]). SBC stands for subband codec, which is the default encoding method required by A2DP hardware. At the moment this encoding must be carried out by the ARM in software, but if the task could be offloaded to the DSP it would free up the ARM to do other things (such as decoding video better). This is really just a yes/no, are Nokia doing anything as if they are, I won't bother. 2. PowerVR MBX driver [2] and OpenGL implementation. We have PowerVR MBX chips sat in our N8x0 machines, doing, afaict, not a lot. It would be nice if it worked (think OpenGL loveliness in games, and I would hope some frivolous but eye catching snazzy desktop features - prettier UI = more publicity = more sales; perhaps...). There are Linux 2.6.x kernel drivers available, though not for the right kernel version and for the OMAP2430 rather than OMAP2420. It would be nice to know if Nokia are pursuing this, and if not, what the limitation is - if it's hardware could we know, as otherwise we might try fiddling with the existing driver and see if we get anywhere (though that will be a long road). 3. IVA - this is part of the OMAP2420 and stands for Imaging and Video Accelerator chip. Its abilities are described at the end of this document [3], in the specifications. It looks like it could do some useful acceleration. There is however no real information available (from Ti or elsewhere) about what it is, and how to program it, though there are some hints, e.g. [4]. The DSP gateway (which is the bridge software between the ARM and the DSP) has been expanded to allow it to interface with the IVA, but the patches say that there's no driver for the IVA itself [5]. Is this being pursued (as there's little we can do about it as we have no information) and if not, why not? 4. AGPS - Assisted GPS [6]. We have an AGPS chip in the N8x0, it doesn't use the assistance feature. Is this being worked on? This is something we as the community could do some work on, if we had some sort of API to allow us to inject the updated almanac/ephemeris data (leave how to work these out as part of the challenge ;)). Obviously if Nokia come out with an all encompassing system (e.g. giving us a way of talking to supl.nokia.com to provide us with the almanac/ephemeris data that I presume it provides for mobile phones), then that would be good (though probably less fun for hacking purposes ;)) And last but not least, all my questions should be read with pretty please added to them rather than as demands, and thanks (to Dave and Nokia) for taking this on. Cheers, Simon [1] http://en.wikipedia.org/wiki/A2DP#Advanced_Audio_Distribution_Profile_.28A2D P.29 [2] http://www.imgtec.com/factsheets/powervr/POWERVR_MBX_IP_Core_Family_%5B1.0%5 D.pdf [3] http://focus.ti.com/pdfs/wtbu/TI_omap2420.pdf [4] http://osdir.com/ml/linux.ports.arm.omap/2006-08/msg00177.html [5] http://osdir.com/ml/ports.arm.omap/2006-09/msg00067.html [6] http://en.wikipedia.org/wiki/AGPS ___ maemo-developers mailing list maemo-developers@maemo.org https://lists.maemo.org/mailman/listinfo/maemo-developers -- Nick Loeve ___ maemo-developers mailing list maemo-developers@maemo.org
Re: Would like to know the status of (was RE: Corporate ownership of open source projects [LWN])
Hello, On Wed, 2008-05-07 at 12:18 +0100, ext Simon Pickering wrote: ... 4. AGPS - Assisted GPS [6]. We have an AGPS chip in the N8x0, it doesn't use the assistance feature. Is this being worked on? This is something we as the community could do some work on, if we had some sort of API to allow us to inject the updated almanac/ephemeris data (leave how to work these out as part of the challenge ;)). Obviously if Nokia come out with an all encompassing system (e.g. giving us a way of talking to supl.nokia.com to provide us with the almanac/ephemeris data that I presume it provides for mobile phones), then that would be good (though probably less fun for hacking purposes ;)) I saw some Nokia-internal A-GPS SW for N810. Can't say anything about the SW maturity, nor release dates, though. Somebody in Nokia seems to be working on it. BR, Kimmo And last but not least, all my questions should be read with pretty please added to them rather than as demands, and thanks (to Dave and Nokia) for taking this on. Cheers, Simon [1] http://en.wikipedia.org/wiki/A2DP#Advanced_Audio_Distribution_Profile_.28A2D P.29 [2] http://www.imgtec.com/factsheets/powervr/POWERVR_MBX_IP_Core_Family_%5B1.0%5 D.pdf [3] http://focus.ti.com/pdfs/wtbu/TI_omap2420.pdf [4] http://osdir.com/ml/linux.ports.arm.omap/2006-08/msg00177.html [5] http://osdir.com/ml/ports.arm.omap/2006-09/msg00067.html [6] http://en.wikipedia.org/wiki/AGPS ___ maemo-developers mailing list maemo-developers@maemo.org https://lists.maemo.org/mailman/listinfo/maemo-developers ___ maemo-developers mailing list maemo-developers@maemo.org https://lists.maemo.org/mailman/listinfo/maemo-developers
Re: dsme, Re: Would like to know the status of (was RE: Corporate ownership of open source projects [LWN])
Hi! On Wed, May 7, 2008 at 2:43 PM, Frantisek Dufka [EMAIL PROTECTED] wrote: nick loeve wrote: 2. Is it possible to at least get some basic documentation of what bme and dsme are doing (beyond obvious basics)? There is a lot of info. Search list archives for dsme (or bme) http://www.gossamer-threads.com/lists/maemo/ If you still have some specific question not answered, ask again, maybe someone knows, maybe not. I'm not sure what 'obvious basics' means to you. Yep, I have been reading the archives and have compiled a shortlist of info based on past discussions (mainly that you have started! ). I guess I might start documenting what is known already and try and get it all in one place. Thats a great site for searching the archives BTW, thanks! -- Nick Loeve ___ maemo-developers mailing list maemo-developers@maemo.org https://lists.maemo.org/mailman/listinfo/maemo-developers
Re: dsme, Re: Would like to know the status of (was RE: Corporate ownership of open source projects [LWN])
nick loeve wrote: Yep, I have been reading the archives and have compiled a shortlist of info based on past discussions (mainly that you have started! ). I guess I might start documenting what is known already and try and get it all in one place. Yes, please do if you can. I am guilty of not doing it myself over years. I'm sure people on this list can fill some gaps. Apart from examining dsme, its modules and dsmetest in initfs (and its runtime behaviour using strace), check also powerlaunch http://powerlaunch.garage.maemo.org/ It is replacement of mce (also closed). mce talks to dsme too and is done in similar modular way so it can help with some understanding. Powerlaunch imitates some mce-dsme communication, see dsme.c,h at https://garage.maemo.org/plugins/scmsvn/viewcvs.php/trunk/powered/?root=powerlaunch As for bme and figuring out how battery charging works, one could strace it and see what tahvo/retu register it uses, some bits are mentioned in kernel headers (not sure about name now, something like tahvo_retu_user.h) and there is also this info http://marc.info/?l=linux-omapm=120752529631862w=2 Frantisek ___ maemo-developers mailing list maemo-developers@maemo.org https://lists.maemo.org/mailman/listinfo/maemo-developers
Re: dsme, Re: Would like to know the status of (was RE: Corporate ownership of open source projects [LWN])
Frantisek Dufka [EMAIL PROTECTED] writes: As for bme and figuring out how battery charging works, one could strace it and see what tahvo/retu register it uses, I did this once for the 770 (by modding the kernel, same basic results). There are some undocumented RETU and TAHVO ports that control the charge circuit. They're undocumented because setting them wrong causes the battery to overheat and possibly explode. Read at your own risk! retu interrupt mask 0x0100 is a 5 second timer used for charge timing. the ADCs are documented in that link you mentioned. retu register 9 has bits that control the charge process but I never decoded them fully. reg 15 and 16 are also involved in controlling the regulator. reg 20 mask 0x1000 is set when the charger is plugged in tahvo register 4 is the charge control frob. Values from 0 (disable charging) to 255 (maximum charge rate) can be written. Must coordinate with ADC values or the battery goes boom! tahvo reg 8 controls charging but I don't know exactly what the bits do (except that mask 0x0001 global-enables the charger). Same for reg 12. tahvo reg 13 is a signed 16 bit ADC that tells you the current flowing into the battery (positive when charging, negative when running). I don't know the scale. The overall charge process is one of setting the charge current to try to get a specific voltage (i.e. constant current (max) at first, followed by variable current (mostly reducing) to get constant voltage) and periodically shutting down the charger to read the idle voltage/current. When you need to set the charge to zero current to get the right voltage, you're done. This is a standard lithium charge cycle. I'm guessing the periodic shutdown helps it deduce charge time ETA. ___ maemo-developers mailing list maemo-developers@maemo.org https://lists.maemo.org/mailman/listinfo/maemo-developers
