IRC log of #maemo-ssu for Saturday, 2013-09-14

*** nox- has joined #maemo-ssu00:00
*** dos11 has joined #maemo-ssu00:11
*** dos1 has quit IRC00:14
*** tg has quit IRC00:15
*** Vlad_on_the_road has quit IRC00:18
*** discopig has quit IRC00:23
*** tg has joined #maemo-ssu00:24
*** discopig has joined #maemo-ssu00:34
*** amizraa has quit IRC00:54
*** amizraa has joined #maemo-ssu00:56
*** M4rtinK has joined #maemo-ssu01:09
*** LauRoman has quit IRC01:17
*** NIN101 has quit IRC01:57
*** sunny_s has quit IRC02:04
*** Pali has quit IRC02:20
*** Martix has quit IRC02:26
*** dos11 is now known as dos103:05
*** M4rtinK has quit IRC03:10
*** dos1 has quit IRC03:30
*** nox- has quit IRC03:47
*** jonwil has joined #maemo-ssu04:06
*** LaoLang_cool has joined #maemo-ssu05:00
*** amiconn has quit IRC05:04
*** amiconn_ has joined #maemo-ssu05:04
*** amiconn_ is now known as amiconn05:04
*** LaoLang_cool has joined #maemo-ssu06:06
LaoLang_coolanyone use fetchmail on maemo?06:06
LaoLang_coolI'm looking for maemo's version of fetchmail06:07
jonwilI dont think fetchmail has been ported to maemo06:11
LaoLang_cooljonwil: oh...06:12
LaoLang_coolthanks, I think maybe someone has compiled it for maemo06:13
*** LaoLang_cool has quit IRC06:19
*** jonwil has quit IRC08:03
*** dhbiker has quit IRC08:20
*** dhbiker has joined #maemo-ssu08:34
*** oldtopman has quit IRC09:26
*** LauRoman has joined #maemo-ssu09:44
*** Pali has joined #maemo-ssu09:44
*** arcean has joined #maemo-ssu09:45
*** arcean has quit IRC10:40
*** nedko has quit IRC10:47
*** arcean has joined #maemo-ssu10:48
*** nedko has joined #maemo-ssu10:48
*** LauRoman has quit IRC10:50
*** arcean has quit IRC10:53
*** arcean has joined #maemo-ssu10:54
sixwheeledbeastPut a quick stub on the wiki at wmo/porting to navigate between the porting pages easier. Also added some categories.10:56
*** Vlad_on_the_road has joined #maemo-ssu10:58
*** amiconn has quit IRC11:12
*** NIN101 has joined #maemo-ssu11:27
*** amizraa4 has joined #maemo-ssu11:33
*** amizraa has quit IRC11:37
*** Vlad_on_the_road has quit IRC11:58
*** Vlad_on_the_road has joined #maemo-ssu11:58
*** Vlad_on_the_road has joined #maemo-ssu12:00
*** LaoLang_cool has joined #maemo-ssu12:09
*** jonwil has joined #maemo-ssu12:17
*** amiconn has joined #maemo-ssu12:17
*** amiconn has joined #maemo-ssu12:26
sixwheeledbeastjonwil: <<(08:56:55) sixwheeledbeast: Put a quick stub on the wiki at wmo/porting to navigate between the porting pages easier. Also added some categories.>>12:28
jonwilok, great :)12:28
jonwilI think the # of packages that will need stuff done to them for the Neo900 is not as big as I thought it would be12:30
jonwilThe hardest part will be the cellular modem stuff12:31
freemangordonjonwil: great job :)12:43
*** Vlad_on_the_road has quit IRC12:47
*** Vlad_on_the_road has joined #maemo-ssu12:54
*** arcean_ has joined #maemo-ssu13:00
*** arcean has quit IRC13:02
sixwheeledbeastAny objections to new layout can rollback if so?13:06
jonwillooks good to me13:10
sixwheeledbeast:) it a bit easier to read IMO13:11
*** bsdmaniak has joined #maemo-ssu13:15
*** Martix has joined #maemo-ssu13:24
Palifreemangordon: look at my patches for ke-recv:
PaliI reverted more commits and implemented multipartition support again and reused your code13:27
*** janemba has left #maemo-ssu13:40
*** bsdmaniak has quit IRC13:49
*** bsdmaniak has joined #maemo-ssu13:55
*** DrCode has quit IRC14:04
*** piscodig has joined #maemo-ssu14:05
*** discopig has quit IRC14:06
*** thedead1440 has quit IRC14:07
*** thedead1440_ has joined #maemo-ssu14:07
sixwheeledbeastPali: I have been told KP fixes the issue where stock maemo bootup causes uSD card to locate at mmc0. Is this correct?14:11
Paliorder of mmc devices in maemo and also in upstream kernel is random14:11
Palireal name is later configured by udev14:12
Palikernel-power has patch which allocate mmc0 only for internal eMMC14:12
Paliso when SD card is initializing it will always get num 114:13
sixwheeledbeastSo what would be the "correct" or "best" way to depend on this?14:13
Palibut just to note, this is not issue in kernel... kernel assign first available number for first device which ask for it14:14
Paliany other policy operations is up to userspace --> udev14:14
Palisixwheeledbeast: so which problem do you have?14:15
Paliudev which setting correct mmc number is starting before maemo gui14:15
sixwheeledbeastflopswap uses fixed locations for swap spaces 1p2 and 1p3 due to the way it flip-flops in the script. an issue obviously appears without kp installed.14:17
Paliand when is flopswap started? before udev?14:18
Palithen it should care about numbers!14:18
*** DrCode has joined #maemo-ssu14:18
Paliand what is flopswap doing?14:19
sixwheeledbeaststarted? it not that clever.14:19
sixwheeledbeastscripts that swapon/swapoff partitions. the only thing that started is an option to add upstart script to put swap on 1p2 at start.14:21
keriojust use LABEL=foo14:21
kerioit's a solved problem14:21
*** dos1 has joined #maemo-ssu14:22
Paliuse blkd for locating swap partitions14:22
kerioor, you know, use the LABEL directive14:22
Palicssu-devel generating swaps in /etc/fstab from blkid output14:22
Paliand it can be configured to add also swaps from SD card paritions14:22
Palihardcoding any paths is BAD14:23
Palieven maemo ke-recv can use MyDocs if eMMC is mmc114:24
sixwheeledbeastok i see I will have to look at doing that then. I understand hardcoding it not ideal but I 'm still learning14:24
Palisixwheeledbeast: you should use only swap partitions specified in /etc/fstab14:25
Paliswapon/swapoff use them14:25
FatPhilDocScrutinizer05: my irc window closed, but I found the meeting proceedings on the website.14:26
sixwheeledbeastso i need to modify scripts to get locations from /etc/fstab14:27
FatPhilif you want someone local to finland, or a natve english speaker, I can be that person (even though I go back to Estonia soon, it's not a long trip)14:29
DocScrutinizer05((<jonwil> ... # of packages that will need stuff done...not as big. The hardest part will be the cellular modem stuff)) I'm really happy to see you are 100% in line with my own estimation regarding this :-) \o/14:37
DocScrutinizer05FatPhil: thanks! much appreciated.14:38
*** piscodig is now known as bromide14:38
*** Vlad_on_the_road has quit IRC14:43
DocScrutinizer05FatPhil: I think we hardly could find anybody better qualified to draft or review such a letter to Nokia, knowing their "mentality", with perfect English, and even a background that helps with some internals that mere mortals are for sure not aware of. If you're willing to help council (+HiFo) to write such a letter/mail to Nokia asking/offering about managing sourcecode of maemo for them, we might actually have a small chance that14:49
DocScrutinizer05this will fly14:49
*** arcean_ has quit IRC14:50
DocScrutinizer05If otoh you say it's futile effort, we for sure should reconsider that idea14:51
jonwilDocScrutinuizer05: The GPS will be the other hard part :)14:54
jonwilAnd a few smaller parts like MCE14:54
DocScrutinizer05yep, exactly14:54
DocScrutinizer05GPS is low prio, MCE is a MUST HAVE14:55
DocScrutinizer05IOW GPS we can make all FOSS apps work, even when switching to completely different API14:56
DocScrutinizer05(though that's not really the idea behind FP)14:57
jonwilbtw it is my view that we should set the goal of doing things such that we dont need to touch telepathy-ring, rtcom-call-ui or rtcom-messaging-ui at all.14:59
DocScrutinizer05but we can jerry-build sth that kinda works, even with a thin compatibility layer above gypsy14:59
DocScrutinizer05yes, absolutely15:00
FatPhilhaving stuff written down on paper is the best way of knowing what the territory really is. If they reply with contradictions, that leaves room for further approaches.15:01
DocScrutinizer05for modem we could start development on N900 with UMTS-USB-modem attached via H-E-N15:02
jonwilfor GPS the way to go now that I look at it is to figure out what* does and implement that on top of the the new GPS hardware15:03
DocScrutinizer05...which will not give us proper audio most likely, but we can test everything else - like call setup, answer, hold, end. Data/GPRS15:03
DocScrutinizer05jonwil: please consider using FSO wherever possible15:04
jonwilif FSO has a GPS stack, adding support to it to spit out* seems like the way to go15:05
DocScrutinizer05jonwil: we have some great support/knowhow regarding FSO in openmoko community ;-)15:05
DocScrutinizer05FSO is using gypsy for GPS15:05
dos1but gypsy API is only implemented in python implementation (frameworkd)15:06
DocScrutinizer05why that?15:06
jonwilbtw if by some miracle we CAN get more source code out of Nokia but cant get everything for some reason, I have a wishlist of about 10 packages where having source would be of use and another 10-15 where having docs/info would be of use :)15:06
DocScrutinizer05(not that it matters too much)15:06
*** arcean_ has joined #maemo-ssu15:06
dos1in cornucopia, there's no dedicated GPS API implemented by FSO15:07
dos1and there's some gpsd integration in fsotdld15:07
dos1DocScrutinizer05: well, I'm not sure. I remember that mickeyl had some reason behind it, but I don't remember what it was15:07
DocScrutinizer05have you "seen" him during the last weeks?15:08
DocScrutinizer05his help would be an *invaluable* benefit to FPTF15:09
dos1I guess direct e-mail would be the only way to contact him right now15:09
DocScrutinizer05will you do or shall I...?15:09
dos1basing on latest post on his blog15:09
DocScrutinizer05we at very least should let him know about Neo900, FPTF, and our plans to base that on FSO where possible15:10
dos1DocScrutinizer05: maybe you, I already wasted too much time not learning to exams soon :x15:11
DocScrutinizer05any help from his side highly appreciated15:11
DocScrutinizer05k, another point on my ToDo list15:11
DocScrutinizer05~seen mickeyl15:11
infobotmickeyl <~mickey@openmoko/coreteam/mickey> was last seen on IRC in channel #openmoko-cdevel, 373d 21h 27m 33s ago, saying: 'hi, btw.'.15:11
DocScrutinizer05[2013-09-14 14:12:04] [Notice] -NickServ- Last addr  : ~mickey@digitaleclipse.org15:12
DocScrutinizer05[2013-09-14 14:12:04] [Notice] -NickServ- Last seen  : Jun 26 20:35:58 2013 (11 weeks, 2 days, 15:36:06 ago)15:12
*** Vlad_on_the_road has joined #maemo-ssu15:12
*** sixwheeledbeast has left #maemo-ssu15:52
freemangordonPali: so, now the partition exported through mass storage in case there is swap is /dev/mmcblk1k1, always?15:59
Palinow yes16:00
freemangordonwhat if this is the swap partition?16:00
Paliany non deterministic behaviour should be removed now16:00
Palithen exporting fail16:01
Palithis needs to be fixed...16:01
freemangordonPali: I still think we should export first vfat. And that this is not hard to be implemented in a deterministic way16:02
freemangordonalso we should mount the first vfat under /media/mmc116:02
Paliyou do not know if HAL reported all partitions when you enabled mass storage mode16:02
freemangordonso? we have them in /dev/16:03
Palibut you can expect that user will enable mass storage mode when HAL report all partitions16:03
Paliyes, we have them in /dev/ but not in internal ke-recv structures16:03
freemangordonwe can enumerate /dev/mmcblk1pN, get the partition fs and make the decision16:04
freemangordonwe don;t need HAL, after all we unmount at that moment16:04
freemangordonor better said "going to unmount"16:05
Paliand what happens when you reformat some ext partition to fat which was before first fat?16:05
freemangordonPali: it will become /media/mmc116:05
Palisame problem as before, tracker will se new path and database will be broken16:05
Paliand if there is bug in ke-recv/HAL which enumerate partitions and inserting them into ke-recv structures, users will have this problem...16:06
Palisee report on TMO where user reported this problem16:06
freemangordonPali: I think this is a better solution. Imagine what will happen if you have p1 and p2 as swap. I know it is pretty unusual, but still16:07
Palike-recv has (or had) bug in code which adding/removing partitions in structures which cause that his partition swap between mmc1 and mmc1p116:07
freemangordonlets fix it then (the bug)16:07
freemangordonwhat you're going to export is p1 and p2 are swap?16:08
infobotfreemangordon meant: what you're going to export if p1 and p2 are swap?16:08
Palifixing this bug means: 1. removing ke-recv or 2. rewirting it16:08
Palior 3. make sure that any future bugs which depends on non determinisic behaviour will not be triggered16:09
freemangordonI think this is what I propose :)16:09
Palimeans that ke-recv must be deterministic in way how to choose partition for /media/mmc1 and partition for exporting via USB16:10
Paliand this means that this partition should be hardcoded16:10
Paliwe can add new option to some config file and user can change first partition to any16:10
jonwilso bored :(16:10
freemangordonPali: if you want once a vfat partition is mounted as /media/mmc1 the same to always be mounted there, I think we can just store either volume id or some unique id and use it next time16:11
Palithis will be broken after user format partition16:12
Paliboth windows and mkfs genereting new volume uuid16:12
freemangordonbut we don;t care if that happens, as it is not the *same* partition anymore16:12
Palirather use partition number and not uuid16:12
Paliand we can use partition number from some gconf key16:12
freemangordonshould work too16:13
freemangordongconf is what I had on my mind16:13
Paliby default it could be first partition16:13
freemangordon1. a new uSD is inserted16:13
freemangordon2. enumerate /dev/mmcblk1pN to find the first vfat16:14
freemangordon3. store uSD id and N in some gconf key16:14
freemangordon4. use that from now on to mount under /media/mmc1 and to export (if there is a swap)16:14
freemangordonthere should be a change in p.1. if uSD had changed16:15
infobotfreemangordon meant: there should be a verification in p.1. if uSD had changed16:15
PaliuSD id is problematic and enumerating partitions in ke-recv is not easy, because ke-recv working in async mode - this is reason why using hal/dbus16:16
freemangordonalso a check in p4 if partition type has change, if yes goto 116:16
freemangordonPali: we can use libblkid or directly go to shell16:16
PaliI do not want to patch ke-recv to mix sync/async calls more --> we will have other problems...16:17
freemangordonPali: all this can be done in a shell script16:17
Palinow it is problematic with mount commands, and we can be happy that ke-revc working...16:17
freemangordon(deterministic first vfat detection)16:18
Palipartitions are mounted by ke-recv immetiately when they appear16:18
Paliand partitions are reported in random order by HAL16:18
Palithis is problem16:19
PaliHAL working in async mode16:19
Paliand for all above to work we need sync function for enumeration16:19
freemangordonso? we can call get_first_vfat() everytime ke-recv tries to mount a partition and if get_first_vfat() == current+partition, mount it under /media/mmc116:19
Paliand there could be problems, that we miss some HAL info about partitions...16:20
Paliand if you have some non vfat partitions before vfat and you format non vfat to vfat, then first vfat will change16:21
Paliand this cause tracker problems16:21
freemangordonI know, but that will happen *after* we did a first time mount, so we'll have it stored in gconf key16:22
Paliand I do not trust any uSD ids...16:22
Palireally one simple gconf key which will be by default 116:23
freemangordonPali: we should take care of uSD changed16:23
Paliand are you sure that some ids reported by hal/kernel will work?16:24
freemangordonthey should work16:24
PaliI bet that there will be problems...16:24
*** arcean__ has joined #maemo-ssu16:24
freemangordonmmc.cid = '1b534d30303030301079af83c700cb63'  (string)16:25
freemangordonmmc.csd = '400e00325b590001d47a7f800a404079'  (string)16:25
freemangordonwhatever is that :)16:25
freemangordonoh,  mmc.serial = 2041545671  (0x79af83c7)  (int)16:25
freemangordonPali: iiuc mmc.cid is what we want to use16:27
*** DrCode has quit IRC16:27
*** arcean_ has quit IRC16:27
freemangordonPali: or some part of it, maybe manfid, date and serial16:28
freemangordonwe can just keep card specific info in gconf tree, I bet noone will put 100 uSD cards in his n90016:30
*** DrCode has joined #maemo-ssu16:30
freemangordonso: 1. a card is put16:32
freemangordon2. ke-recv receives "new partition added" and calls get_first_vfaf(), gfv() checks if the card is already in gconf, if it is, it verifies if the "first_vfat" from gconf is still vfat. if it is, it returns it. if not, it enumerates /dev/mmcblk1pN to find the first vfat16:34
freemangordonif the card is not in gconf, gfv() enumerates16:35
*** dhbiker has quit IRC16:35
freemangordon3. ke-recv checks if the current_partition is the one returned from gfv(), if ye, it updates gconf and mounts it under /media/mmc116:35
freemangordonPali: I think ^^^ will work just fine16:36
PaliI will implement something to my new ke-recv git tree16:36
freemangordonPali: like the above?16:37
Paliwill see if there are some problems...16:37
Palibut I will try something as above16:37
freemangordonPali: ooh, and if there is *no* vfat, I don;t think anything should be mounted under /media/mmc116:38
freemangordonPali: ok, nice :)16:38
*** dhbiker has joined #maemo-ssu16:38
freemangordonPali: just to make it clear - I insist on that "vfat" because there are lots of windows users. most of them are clueless enough to don;t know what to do if we hardcode /media/mmc1->/dev/mmcblk1p1 and they already have uSD partitioned in some incompatible way16:40
Paliif they have partition in other way, they probably know what to do...16:41
freemangordoneven if they know, thay'll have to reformat the card16:41
freemangordonimagine a 6gGB card full with data ;)16:42
freemangordon64GB that is16:42
freemangordonone have to copy that date to their desktop, format the card with a partition layout that makes us happy and copy the data back16:43
freemangordonI don;t think this is fare :)16:43
freemangordondata even, not date16:44
freemangordonPali: btw is that kernel camera drivers guy helpful?16:45
Palifreemangordon: last week I got email from him, and some months ago he sent me that board data for front webcam16:46
freemangordonlets hope he'll answer me :)16:46
freemangordonPali: one more thing and I'll stop pestering you - which is the next thing in kernel you think I should try to fix, while wating for sakari to answer?16:48
freemangordonmodem drivers?16:48
freemangordon(I remember those working in 3.5)16:49
Paliyes, look at modem drivers why not working with maemo516:49
Paliand why rtcom-call-ui crashing16:49
freemangordonoh, bwt hulda tries to use some /sys/kernel/low/high_watermark, WTF is this?16:49
freemangordonhmm, goping to as google16:50
Palispecial nokia sysfs16:50
Palipatch hulda to not fail if that sysfs file not exists16:50
freemangordonI don;t think it fails, it just spits some warning in syslog. but I'll look in the code anyway16:50
ShadowJKit's lowmem stuff16:51
ShadowJKto get advance warning of running out of memory, before things slow down16:51
PaliI already patches rcS-late to not fail if that file does not exists16:51
Paliand this was reason why maemo not booted on 3.x kernels16:51
ShadowJKiirc google also used this stuff in android at some point16:52
freemangordonmaybe we should forward port it16:53
freemangordonI don;t think it is that complicated16:54
freemangordonwill look at it when everything else works :D16:54
*** arcean__ has quit IRC17:10
*** mkaindl has left #maemo-ssu17:11
*** thedead1440_ is now known as thedead144017:15
*** mkaindl has joined #maemo-ssu17:23
* jonwil wishes there was something he could usefully do :(17:25
DocScrutinizer05jonwil: why don't you help on reviewing and sanitizing ke-recv specifications?17:45
jonwilI dont know the first thing about ke-recv17:46
*** oldtopman has joined #maemo-ssu17:46
DocScrutinizer05when you read backscroll you'll notice it has a whole lot of very intriguing implications17:46
DocScrutinizer05even with tracker *COUGH*17:46
jonwilbesides, pouring over source code to ke-recv sounds like the sort of thing that would make me MORE bored17:51
*** sunny_s has joined #maemo-ssu17:59
*** oldtopman has quit IRC18:09
*** sunny_s has quit IRC18:12
*** sunny_s has joined #maemo-ssu18:12
jonwilbtw doc, I suspect the FPTF is gonna need someone who knows PulseAudio and probably someone who knows telepathy18:12
*** oldtopman has joined #maemo-ssu18:19
*** oldtopman has joined #maemo-ssu18:19
*** DrCode has quit IRC18:22
nedkoi'm trying to setup a cross-compilation environment for n900. is still scratchbox the proper/best way?18:26
Siceloyes sir :P18:27
*** tg has quit IRC18:29
nedkohow good is to use the maemo sdk virtual image?18:32
*** DrCode has joined #maemo-ssu18:32
nedkoi guess i could install a virtual debian system and then install scratchbox18:34
freemangordonnedko: VMWare image is what i use18:35
nedkofreemangordon: i'm a bit worried that the version is from 2009 but the latest sdk is from 2010...18:36
freemangordonit is ok, just fix your DEB_BUILD_OPTIONS18:37
nedkofreemangordon: is there an online resource on how to do this?18:38
freemangordondo what?18:38
nedkofix your DEB_BUILD_OPTIONS18:39
freemangordonexport DEB_BUILD_OPTIONS=$OPTIONS, the default on contains thumb18:39
nedkofreemangordon: do you use image from or from ?18:41
freemangordonI downloaded it 2-3 ears ago, so I don;t know what I use :)18:42
nedkohmm, maybe they are the same18:42
freemangordonmost probably18:42
nedkomy main problem is that i use gentoo and virtualbox :]18:44
jonwilif you use Gentoo, why not just install Scratchbox for N900 development?18:45
jonwilI have a Gentoo box and I did exactly that18:45
nedkojonwil: how to install scratchbox for n900 development on gentoo?18:47
nedkoi see there is dev-embedded/scratchbox18:48
nedkoit has no use flags18:48
nedkoand i see no packages matching "n900" search pattern18:48
nedkoor maybe i should use scratchbox2?18:49
jonwilI downloaded the SDK from this page and installed it that way (including Scratchbox)18:51
jonwili.e. I installed the Scratchbox Nokia suggested and ignored anything from the Gentoo repos18:51
nedkojonwil: you used the two scripts and they worked ok on gentoo?18:55
jonwilNot sure which script I used18:55
jonwilbut yes I used the Nokia installer and it worked on Gentoo18:55
jonwilmy Gentoo box is not 100% functional right now otherwise I would check more18:56
nedkook, thanks18:56
*** jonwil has quit IRC19:03
* Sicelo is going to join the scratchbox banddwagon soon.. 19:05
Sicelohow big is the vmware image btw?19:05
*** dhbiker has quit IRC19:06
*** dhbiker has joined #maemo-ssu19:08
nedkothe one i'm downloading is 1457450621 bytes19:12
nedkothey provide no digital signatures for the files :(19:13
nedkothe only way to check are md5 checksums19:13
*** arcean__ has joined #maemo-ssu19:15
*** bsdmaniak has quit IRC19:19
*** arcean__ is now known as arcean19:19
*** tg has joined #maemo-ssu19:27
DocScrutinizer05nedko: huh?19:28
DocScrutinizer05what's wrong with md5?19:29
nedkoDocScrutinizer05: it doesnt protect you from compromised servers19:31
nedkoman in the middle attacks too19:31
DocScrutinizer05when our server gets compromised, they as well can steal our private signing key19:32
DocScrutinizer05MITM attacks you hardly can counteract at all19:32
nedkoDocScrutinizer05: you put your private keys on a server?19:33
nedkothe one for SSL certificate (for https) is ok, obviously19:34
nedkoi actually downloaded the file from a host in buglaria19:34
nedkoi bet it is a local mirror of nokia19:34
nedkoor probably a CDN is a better term here19:36
DocScrutinizer05nedko: when those are private keys, what do you think where we should store them?19:36
nedkodepends on what you use them for, no?19:37
DocScrutinizer05in MrFlop's bedside locker?19:37
nedkothe two scripts for installing the SDK require root19:38
DocScrutinizer05so what? I can't help that19:38
nedkodon't you think it is a good idea to have them at least somewhat verified before executing code as root19:38
nedkoor it is expected that people should use a dedicated machine for maemo development19:39
DocScrutinizer05don't you think, since they are scripts anyway, that it should be *YOU* to verify them, on sourcecode level?19:39
DocScrutinizer05or would you trust *me* to do that for you?19:40
nedkoDocScrutinizer05: yup, this is the good side. they are scripts19:40
DocScrutinizer05I can signe those files with my private key, no problem19:40
DocScrutinizer05but that doesn't mean a thing19:41
nedkoif i trust you...19:41
nedkoand my trust on you may be different that trust on someone controlling the internet access in my country19:41
freemangordonDocScrutinizer05: what is GAZOO and what is RAPUYAMA?19:41
DocScrutinizer05honestly, I don't see how techstaff can help19:42
DocScrutinizer05freemangordon: RAPU = BB5 baseband, GAZOO = companion chip with power, maybe RF amp19:42
nedkoDocScrutinizer05: you are from nokia techstaff?19:43
DocScrutinizer05nedko: if you maybe haven't noticed yet: is a community driven project since start of this year19:44
DocScrutinizer05no NOKIA techstaff around at all19:44
nedkoDocScrutinizer05: yes, i noticed19:45
DocScrutinizer05so yes, i'm from techstaff, and NO I'm NOt NOKIA19:45
nedkoi just complained that i have almost no mean to verify nokia maemo sdk vmware image19:45
DocScrutinizer05verify against what? what is your damn reference to verify it against?19:46
nedkoDocScrutinizer05: in a corporate world, i'd expect the MD5SUM file to be available via https19:46
DocScrutinizer05for all I could tell it may contain malware since 5 years19:47
nedkowith proper certificate19:47
nedkossl certificate19:47
nedkoif it was open source, i'd expect someone from upstream to sign the releases with a private key stored on his own machine, not a server19:48
DocScrutinizer05we CANNOT get a proper SSL cert since domain (DNS) still owned by NOKIA19:48
nedkothis is what i do when i act as a upstream19:48
DocScrutinizer05and i'm not your dang upstream19:48
nedkoDocScrutinizer05: i'm not telling that it is your fault19:48
DocScrutinizer05YOU are upsteam!19:48
* nedko shuts up, there is no point19:50
DocScrutinizer05there's zilch I could do for you19:50
DocScrutinizer05I can quote the md5sum out of band in this IRC channel, if that helps19:51
DocScrutinizer05directly off the commandline on server19:52
nedkobtw, it is i'm talking about19:52
DocScrutinizer05prolly not really19:52
*** Vlad_on_the_road has quit IRC19:52
nedkothe MD5SUM file is available from
DocScrutinizer05which is a CDS to
freemangordonFatPhil: ping19:53
infobot , or http://tabletsdev.maemo.org19:53
DocScrutinizer05you're free to compare all 4 information sources to verify that you indeed got that opaque blob that Nokia donated to community. Nobody will give any warranty for its content though, except that it been unchanged since years19:55
nedkonow this is useful, thank you19:56
nedko sdk link points to nokia site...19:57
DocScrutinizer05then edit the URL19:57
DocScrutinizer05it's stage for
DocScrutinizer05of course the links point to
DocScrutinizer05CDS, remember?19:58
nedkoi admit i dont really know what CDS is19:58
DocScrutinizer05Content Distribution System19:59
DocScrutinizer05a server farm of proxies basically19:59
DocScrutinizer05 are slaves, is master aka stage19:59
nedkoso how i could at least check that the MD5SUM i downloaded from the bulgarian server is same as the one you use?19:59
nedkos/use/will get if you try to download it/20:00
infobotnedko meant: so how i could at least check that the MD5SUM i downloaded from the bulgarian server is same as the one you will get if you try to download it?20:00
DocScrutinizer05get it from http://tabletsdev.maemo.org20:00
DocScrutinizer05the URL on is identical to that on
DocScrutinizer05just edit it, to access the master20:01
nedkoedit what?20:01
nedkothe url?20:01
freemangordonurl in your browser20:01
freemangordonor wget or whatever you use20:01
freemangordonthean download the sdk20:02
freemangordonBTW I am nor sure #maemo-ssu is the channel to discuss SDK stuff on20:02
nedkoi got redirect to
nedkofreemangordon: what is a better channel for this?20:02
freemangordonthen replace that with  tabletsdev.maemo.org20:02
freemangordontry on #maemo20:02
nedkoclick accept and you get to
nedkofreemangordon: thank you20:03
* nedko shuts up20:03
infobotfrom memory, skeiron is the semi-official backup and emergency standin for all internet borne maemo resources:, or see: ~tabletsdev20:03
*** arcean has quit IRC20:14
*** Vlad_on_the_road has joined #maemo-ssu20:16
freemangordonDocScrutinizer05: hmm, need a little help with that cmt shit. by looking at the schematics, gpio 72 seems to be input to omap (APE_RST_RQ). But I have gpio 72 as .cmt_rst_ind_gpio = 72 in board data. this makes no sense20:26
DocScrutinizer05both are RST at least :-)20:27
DocScrutinizer05dunno if cmt resets APE or APE resets cmt ;-D20:27
freemangordonwtf is APE?20:27
DocScrutinizer05or both? :-o20:27
DocScrutinizer05App Proc Env20:28
DocScrutinizer05== linux20:28
DocScrutinizer05== SoC OMAP343020:28
freemangordonhmm, but this is input to omap (gpio 72)20:28
freemangordonyeah, got it20:28
freemangordonso, modem resets omap?20:28
DocScrutinizer05possibly. One of the WD20:29
DocScrutinizer05makes a lot of sense to me20:29
freemangordonhmm, yeah20:29
DocScrutinizer05either it's one of our 2 or 3 generic WD, or a special one that only kicks in during 2orphaned" phonecalls20:30
DocScrutinizer05you wouldn't want the modem to continue a call ad infinitum while APE locked up, eh?20:30
freemangordonhmm, but this is just a gpio20:31
DocScrutinizer05check OMAP34xx TRM aka spruf98 for exact meaning/semantics/function of APE_RST_RQ20:31
DocScrutinizer05*each* GPIO has at least a 2nd function, often a 3rd and 4th20:32
DocScrutinizer05you define that via IO-mux or whatever it's called20:32
freemangordonok, this is fine then20:33
DocScrutinizer05actually it's rather the other way round: each function pin can also get used as GPIO when you don't need the function20:33
DocScrutinizer05except a very few ones like VDD, GND ;-) and maybe CLK etc20:34
DocScrutinizer05and most of GPIO can also create an IRQ20:35
DocScrutinizer05well, several20:35
DocScrutinizer05dunno if "most of"20:35
DocScrutinizer05page 16425820:36
freemangordonhmm, iirc every gpio can act as irq too20:36
DocScrutinizer05maybe reading service manual L3_4 helps!20:36
DocScrutinizer05they have _some_ nice high level function descriptions in there20:37
DocScrutinizer05strongly recommended read20:37
*** dos1 has quit IRC20:41
DocScrutinizer05N900_RX-51_SM_L3_4.pdf  page 16520:42
DocScrutinizer05p180 "Cellular RF Block Diagram"20:43
*** Vlad_on_the_road has quit IRC20:44
DocScrutinizer05actually not that enlightening20:44
freemangordona black box20:45
freemangordonhmm, cawake gpio control seems fine20:48
*** dos1 has joined #maemo-ssu20:50
DocScrutinizer05freemangordon: ok, forget about L3_4, it has nothing interesting. Just ask me instead ;-)20:51
freemangordonok :)20:52
DocScrutinizer05watch the funny stuff around BSI they done to RAPU20:53
DocScrutinizer05(p 11)20:54
*** dos1 has quit IRC20:54
DocScrutinizer05you evidently can signal special modes (TEST[!!] Battery) directly to BB5 RAPU, via BSI20:55
DocScrutinizer05shutting down BME won't change any of that20:56
PaliDocScrutinizer05: it is possible to (re)charge RTC battery in n900?20:56
*** Vlad_on_the_road has joined #maemo-ssu20:56
freemangordonhmm. maybe replacing the battery with a diagnostic tool is the way used to test bb520:56
DocScrutinizer05Pali: it recharges automatically20:59
DocScrutinizer05freemangordon: exactly20:59
PaliDocScrutinizer05: without BME?20:59
freemangordonPali: yes20:59
freemangordonnolo sets the charging21:00
DocScrutinizer05built-in function of TWL403021:00
Paliok, so then my rtc battery is emtpy or dead...21:00
DocScrutinizer05needs one or two bits set in one of the twl4030 registers21:00
freemangordonas is mine and all the others :D21:00
infobotsomebody said bupbat was use the capacitive type, LiIon are breaking during 12 months, or, or, or
freemangordonDocScrutinizer05: yeap, charge current and enable21:01
Palialready reading21:05
*** dos1 has joined #maemo-ssu21:06
DocScrutinizer05ooh, it was already in infobot's factoid21:10
* DocScrutinizer05 feels silly for searching 10min for an URL that's been in front of his nose21:11
DocScrutinizer05Pali: you noticed ~zapman ? :-)21:12
DocScrutinizer05zapman zaps fapman21:12
DocScrutinizer05~zapman usb21:14
DocScrutinizer05OOOOH! Mobile HotSpot V0.3.4 is from Eero/Rambo!21:15
freemangordonDocScrutinizer05: hmm, can;t find gpio 157 on the schematic :(21:16
freemangordonthat one is defined as CMT_BSI21:16
freemangordonin board data21:16
DocScrutinizer05better ddat21:17
*** nox- has joined #maemo-ssu21:17
DocScrutinizer05err nope21:17
DocScrutinizer05where from you got GPIO_157?21:17
freemangordon#define RX51_GPIO_CMT_BSI15721:18
freemangordon#define RX51_GPIO_CMT_BSI 15721:18
freemangordonDocScrutinizer05: I am verifying gpio defs in board data for validity21:18
DocScrutinizer05it's probably some AD_IN4 or whatever21:18
freemangordonomg21:19 "cmt_bsi",21:19
freemangordon.gpio= RX51_GPIO_CMT_BSI,21:19
freemangordon.flags= OMAP_GPIO_SWITCH_FLAG_OUTPUT,21:19
freemangordon.type= OMAP_GPIO_SWITCH_TYPE_ACTIVITY,21:19
freemangordon.debounce_rising= RX51_GPIO_DEBOUNCE_TIMEOUT,21:19
freemangordon.debounce_falling= RX51_GPIO_DEBOUNCE_TIMEOUT,21:19
DocScrutinizer05you need to: get spruf98, find GPIO157, check for 2nd function, search for $SECONDFUNCTION in schematics21:19
DocScrutinizer05or get the pin coords (H21 or sth) and search for those21:20
*** LauRoman has joined #maemo-ssu21:21
DocScrutinizer05try ADCIN421:22
DocScrutinizer05ooh wait21:22
DocScrutinizer05hmm, no clue. would actually like to hear which pin number you find for GPIO_15721:24
freemangordonI find nothing but that define in board data21:25
DocScrutinizer05you need to check SPRUF9821:25
DocScrutinizer05search GPIO 15721:25
DocScrutinizer05read the pin ID (AF21)21:26
infobotDocScrutinizer05 meant: read the pin number (AF21)21:26
freemangordonhmm, lost it, need to find it again21:27
DocScrutinizer05since it seems the pin number is the only label that *every* SoC pin got, in schematics21:27
freemangordonsee on page 77921:28
freemangordonin trm21:28
DocScrutinizer05I don't have spruf98 opened21:28
freemangordonTable 7-4. Core Control Module Pad Configuration Register Fields (continued)21:28
DocScrutinizer05that one21:28
freemangordonCONTROL_PADCONF_MCBSP1_CLKR[31:16] 0x4800 218C mcbsp1_fsr adpllv2d_dither cam_global_ gpio_157 safe_mode21:28
freemangordonin node4 this is gpio 15721:29
DocScrutinizer05and other modes seem mcbsp1_fsr adpllv2d_dither21:30
DocScrutinizer05i'd rather like to learn the pin number ;-)21:31
freemangordonme too :)21:31
DocScrutinizer05MCBSP1_CLKR  ?21:31
DocScrutinizer05makes no sense21:32
freemangordonno, this is gpio 15621:32
freemangordonmcbsp1_fsr is gpio 15721:32
freemangordonyou mean the define?21:32
freemangordonI know, that is why I asked :)21:33
freemangordonhmm, going to check in meego kernel21:33
DocScrutinizer05there's mcbsp1_fsx21:34
DocScrutinizer05but no mcbsp1_fsr21:34
DocScrutinizer05and there's MCBSP1_CLKR[21:34
DocScrutinizer05which is pin Y2121:35
DocScrutinizer05search SPRUF98 for Y2121:35
DocScrutinizer05you should find a GPIO definition next to it21:36
freemangordonbtw what I have is not spruf98, but omap3430 public trm21:36
DocScrutinizer05search in same table for GPIO_15721:36
DocScrutinizer05what is written at bottom of that TRM?21:37
DocScrutinizer05each page?21:37
freemangordonSWPU223M–July 2007–Revised February 201121:37
DocScrutinizer05SWPU223? DUH!21:37
DocScrutinizer05right, prolly SPRUF has no pin numbers at all21:39
DocScrutinizer05no SWPU here21:41
freemangordonDocScrutinizer05: see
freemangordongpio157 is put in safe mode21:42
DocScrutinizer05can't say this means much to me21:45
freemangordonI guess safe mode means "floating"21:45
DocScrutinizer05at *some* time *some* GPIO_157 is set to a certain mode, maybe21:45
freemangordonoh, well21:45
DocScrutinizer05probably float/high-Z, yes21:46
DocScrutinizer05somewhere they define what "safe_mode" actually means21:47
DocScrutinizer05I didn't bother so far to check21:47
freemangordonme neither21:47
*** lansiir has joined #maemo-ssu21:48
DocScrutinizer05anyway possible functions of this pin are:  mcbsp1_fsr     adpllv2d_dithering_en1     cam_global_reset       and gpio15721:50
DocScrutinizer05neither of those shows up in N900 schematics21:51
*** oldtopman has quit IRC21:52
DocScrutinizer05so your only last chance is to find out about the pin number for the package used in N900, and search for that. You likely won't find that either, which means the pin is unused21:52
DocScrutinizer05is GPIO_157 used *anywhere* in maemo kernel?21:52
DocScrutinizer05I mean: *used*21:52
DocScrutinizer05like in: read it, write to it21:53
DocScrutinizer05define a IRQ handler for it21:53
freemangordonwill check if maemo sscd uses that gpio21:53
DocScrutinizer05cmt_bsi sounds pretty odd21:54
freemangordonhmm, "ssc_run.c:304 ssc_hw_write(".../cmt_bsi/state", "inactive")"21:55
*** lansiir has quit IRC21:55
freemangordonlemme check where is that wired on 2.6.2821:55
DocScrutinizer05I ponted you to what cmt does with respect to BSI. I don't see much of any cmt_bsi signal from or to APE21:55
freemangordonme neither, but sscd seems to use that21:55
freemangordonsee ^^^21:56
DocScrutinizer05the line you quoted above sounds more like "not used at all"21:56
freemangordonthis is syslog when "/usr/sbin/sscd -w -d 3"21:56
DocScrutinizer05what the heck is sscd?21:57
freemangordonthe one that talks to the modem21:57
freemangordonsee /sys/devices/platform/gpio-switch/cmt_bsi21:58
DocScrutinizer05IroN900:~# cat /sys/devices/platform/gpio-switch/cmt_bsi/state21:59
DocScrutinizer05meh, you beaten me to it21:59
DocScrutinizer05ssc_run.c:304 ssc_hw_write(".../cmt_bsi/state", "inactive")   looks like it does just that: create a sysnode and fill it with "inactive"22:01
freemangordonsee /sys/kernel/debug/omap_gpio22:01
DocScrutinizer05to see what?22:01
freemangordongpios :)22:02
freemangordonthis is not userspace creatde22:02
DocScrutinizer05GPIO 157 (cmt_bsi             ): out lo22:02
freemangordonand our schematic is missing that22:02
DocScrutinizer05I wonder if our schamitcs are missing that or it's just cruft in kernel22:03
DocScrutinizer05schematics even22:03
freemangordonoh, I think I have some clue while the modem does not work22:04
freemangordonjust a sec22:04
freemangordonDocScrutinizer05: Pali: ^^^22:06
Sicelowhy modem does"mt work? you mean in 3.x?22:06
freemangordonpio-72  (cmt_rst_ind         ) in  lo22:06
freemangordongpio-72  (ape_rst_rq          ) in  hi irq-232 edge-falling wakeup22:06
DocScrutinizer05freemangordon: yep, seems no IRQ registered on 72 and 151 in 3.1022:08
freemangordonalso, in lo/in hi22:08
DocScrutinizer05so modem tries to "talk2 but APE is deaf22:09
DocScrutinizer05of course you need IRQ handlers registered for those IO22:09
DocScrutinizer05init fsckdup22:10
freemangordonPali: any idea what that hi/lo means? is that the current state, or acteve state?22:10
DocScrutinizer05hmm I think current state22:10
Palifreemangordon: I have no idea22:12
DocScrutinizer05clearly current state22:12
DocScrutinizer05watch --differences=cumulative cat /sys/kernel/debug/omap_gpio22:12
DocScrutinizer05then engage lockswitch22:12
freemangordonok, got it22:13
freemangordonok, lets see why cmt does not register interuupts22:13
freemangordonor rather ssi driver22:13
DocScrutinizer05ssi driver BS?22:13
DocScrutinizer05board file BS?22:13
freemangordonno idea, have to find it :D22:14
freemangordonhmm, most probably irqs are just not reported :(22:17
DocScrutinizer05anyway, have to start my RL weekend. finally22:17
*** Pali has quit IRC22:19
DocScrutinizer05wtf acmelite reset22:24
DocScrutinizer05watch -n 0,5 --differences=cumulative cat /sys/kernel/debug/omap_gpio ;<--awesome22:25
*** Pali has joined #maemo-ssu22:29
DocScrutinizer05freemangordon: anyway cmt_bsi aka 157 constantly out_low here22:32
DocScrutinizer05stingray reset???22:36
freemangordonback camera22:36
freemangordonhmm, no, modem uses that, whatever it is22:36
*** arcean has joined #maemo-ssu22:46
*** arcean has quit IRC23:04
*** Vlad_on_the_road has quit IRC23:18
*** Vlad_on_the_road has joined #maemo-ssu23:32

Generated by 2.15.1 by Marius Gedminas - find it at!