**** BEGIN LOGGING AT Mon Dec 12 02:59:56 2011 Dec 12 06:34:30 pespin: elsa is the login manager in PROTO Dec 12 06:34:50 pespin: edm if you want Dec 12 07:34:03 moin Dec 12 07:34:40 moin JaMa Dec 12 07:38:15 mrmoku: fsosystemd work fine, thanks for fixing that! Dec 12 08:00:16 JaMa: yeah, n900 boots a lot faster now :P Dec 12 08:00:23 without that firmware delay Dec 12 08:00:38 mickeyl: awake? Dec 12 08:01:04 utilities.c:701:2: error: too many arguments to function 'backtrace_symbols' Dec 12 08:01:51 which comes from: Dec 12 08:01:51 string[] symbols = Linux.backtrace_symbols( buffer, size ); Dec 12 08:02:18 and my vapi has: Dec 12 08:02:19 public string[] backtrace_symbols (void* buffer, int size); Dec 12 08:03:15 _tmp7_ = backtrace_symbols (_tmp4_, _tmp5_, &_tmp6_); Dec 12 08:03:20 is what vala does out of it Dec 12 08:04:33 hmm... it is trying to pass the length of the symbols array in _tmp6 Dec 12 08:04:36 but why? Dec 12 08:17:55 guess I have to wait for some vala guru to wake up Dec 12 08:18:48 duh Dec 12 08:31:03 vala was too old Dec 12 08:34:34 heh Dec 12 08:46:36 Dec 10 11:34:14 (if not GarthPS, dos1 or HeinervdmOff change their mind spontaneously :) <- I will come, if everything goes fine on Thr and Fri at Work :) Dec 12 08:48:07 HeinervdmWork: great :-) Dec 12 08:48:23 HeinervdmWork: IIRC you didn't want a bed, right? Dec 12 08:50:57 mrmoku: no, i will bring a sleeping bag Dec 12 08:51:09 i think there are not enough bed's for everyone, right? Dec 12 08:51:53 HeinervdmWork: well... depends Dec 12 08:52:17 HeinervdmWork: it looks like where an odd number... as the rooms are double-rooms there is one bed left Dec 12 08:52:23 depends on antrik though Dec 12 08:52:59 in any case I think there should be a bed left for whoever wants one Dec 12 08:53:56 some lucky girl on street too? :) Dec 12 08:55:04 :) Dec 12 08:57:46 Btw. who can create user accounts for the fso wiki? it seems that i don't have an account Dec 12 08:58:08 HeinervdmWork: morphis and mickeyl Dec 12 08:58:13 ok Dec 12 08:58:52 if there is time left at the weekend, i would like to add fsopimd ;) Dec 12 09:02:04 cool :) Dec 12 10:25:32 antrik: news from the bank? Dec 12 10:26:04 HeinervdmWork: regarding odd and even... that changed... we've got two double rooms and one triple room Dec 12 10:26:32 JaMa: bad luck for the girl ;) Dec 12 10:30:25 mrmoku: nope :-( Dec 12 10:40:51 antrik, HeinervdmWork: I have to send LH the final bedlist... Dec 12 10:40:54 whom do I put in? Dec 12 10:41:45 lucky girl .. Dec 12 10:41:49 :P Dec 12 10:42:07 I can share room if nobody volunteers.. :) Dec 12 10:42:19 * mrmoku doubts antrik or HeinervdmWork want to be JaMa's lucky girl ;) Dec 12 10:42:51 hehe Dec 12 10:43:14 the bigger problem is I have to distribute us in rooms... and nobody responded to my mail for preferences Dec 12 10:44:59 GarthPS: what about you? already decided to come? :) Dec 12 10:46:00 moin! Dec 12 10:46:31 mrmoku: well I think I will not comme becose it is to expensive Dec 12 10:46:45 ok Dec 12 10:46:51 around 300€ so.. just to come Dec 12 10:48:28 mrmoku: I would love to.. Dec 12 10:51:43 mrmoku: what's cost of LH? just to decide how much € I should exchange Dec 12 10:54:29 JaMa: 15 * 2 (room) + 5 * 2 (breakfast) + 5 * 3 (seminar room) Dec 12 10:54:41 makes 55 € I think Dec 12 10:55:02 plus some money for lunch + dinner we will organize somehow somehwere Dec 12 10:55:25 ahh Dec 12 10:55:31 + 5€ final cleaning Dec 12 10:55:36 so it's 60 € :-)= Dec 12 10:56:34 ok, thanks Dec 12 11:00:22 * DocScrutinizer51 wonders whether to hate or appreciate ClearCase Dec 12 11:00:52 DocScrutinizer51: daywork making you learn new revision control stuff? Dec 12 11:01:04 yeah Dec 12 11:01:10 MEH Dec 12 11:01:33 my head feels like filled with wool Dec 12 11:03:27 might be the weather though. or the flu Dec 12 11:04:03 DocScrutinizer51: are you free for beer in Nurnberg on Sunday? Dec 12 11:04:07 or simply Monday ;-) Dec 12 11:04:15 sure Dec 12 11:04:27 JaMa: visiting N? Dec 12 11:04:51 DocScrutinizer51: my bus is leaving on Monday morning (1:45 am) so after mrmoku leaving me in N I have time for few beers :) Dec 12 11:05:18 I see :-) Dec 12 11:05:37 let's chat about details later Dec 12 11:05:47 ok :) Dec 12 11:23:21 <[Rui]> hi Dec 12 11:23:26 <[Rui]> go where? Dec 12 11:23:42 <[Rui]> mrmoku, go where? Dec 12 11:25:23 Essen http://wiki.freesmartphone.org/index.php/FSOSHRCON_2011 Dec 12 11:26:22 [Rui]: yup Essen Dec 12 11:28:00 <[Rui]> bua Dec 12 11:28:27 <[Rui]> travelling abroad for funny events is limited to fosdem :| Dec 12 13:07:10 mickey_office: ping Dec 12 13:27:37 mrmoku: pöng Dec 12 13:30:19 mickey_office: yo, I'm into netlink with a problem :) Dec 12 13:30:29 they removed the env-style dumping Dec 12 13:31:06 and now I'm not sure what the best way is to get the attributes in a general way Dec 12 13:31:16 arg, bummer Dec 12 13:31:28 i tried to avoid the work getting these in a general way Dec 12 13:31:31 I *think* that's not possible and must be done per type Dec 12 13:32:29 I can get the list of attributes contained in a message with attr_list Dec 12 13:32:37 but no idea how to get back to the attribute then Dec 12 13:33:00 from something like this: Dec 12 13:33:01 mtu,txqlen,qdisc,map,address,broadcast,flags,name,ifindex,family,arptype,stats,change,operstate,linkmode,num_vf Dec 12 13:33:19 iirc there are functions to grab that Dec 12 13:33:28 so we need a massive case statement Dec 12 13:33:34 and return a string Dec 12 13:33:37 hmm Dec 12 13:33:38 well Dec 12 13:33:42 other idea Dec 12 13:33:44 hmm... but the attributes themselves don't contain the name Dec 12 13:33:49 can you grab the code from the old library? Dec 12 13:33:53 :D Dec 12 13:34:07 http://paste.pocoo.org/show/519689/ Dec 12 13:34:17 mickey_office: hehe, and patch it back in? Dec 12 13:34:35 mrmoku: either that or, if it only uses public API, put it into our library as C code Dec 12 13:34:59 mickey_office: what it does is casting the generic obj to the appropriate type Dec 12 13:35:01 like link Dec 12 13:36:08 if (link->ce_mask & LINK_ATTR_IFINDEX) nl_dump(p, "index %u ", link->l_index); Dec 12 13:37:43 mickey_office: there's not that many types Dec 12 13:37:53 so maybe the cleanest thing would be to add them to the vapi Dec 12 13:38:01 hmm, perhaps Dec 12 13:40:27 and then have a helper in fsobasics which takes the object and converts it to the wanted HashTable of properties Dec 12 13:40:42 which it does by casting to the correct type and manually filling it for that type :/ Dec 12 13:40:58 yep Dec 12 13:41:22 at least my main goal is achieved :) Dec 12 13:41:24 i tried to save time doing that when i discovered that the dump() function did exactly what I needed Dec 12 13:41:35 yeah, was quite elegant Dec 12 13:41:57 I can build libfsobasics against libnl-3 now Dec 12 13:42:02 cool Dec 12 13:42:04 the netlink stuff won't work Dec 12 13:42:13 but I'm only interested in working on fsoaudiod for now Dec 12 13:42:29 and libnl was my build blocker :-) Dec 12 13:42:36 good Dec 12 13:43:07 ok, now shower and client :( Dec 12 13:43:13 bbl Dec 12 13:44:19 cu Dec 12 14:41:52 VQuickSilver, hi Dec 12 14:42:14 VQuickSilver,can you show me a patch of your armv6-novfp thing? Dec 12 14:48:35 GNUtoo, ok, give me a second Dec 12 14:49:19 VQuickSilver, I also should point you to the old patch Dec 12 14:49:53 http://paste.pocoo.org/show/519737/ Dec 12 14:50:36 also I added this file meta/conf/machine/include/tune-arm1136ej-s.inc Dec 12 14:51:52 that was mine: http://www.pastie.org/3005401 (wrong way) Dec 12 14:52:10 I changed armv6 to armv6-novfp and added armv6-vfp Dec 12 14:52:41 which would be good in a perfect world but people already started using armv6 (that has vfp) Dec 12 14:53:12 ok Dec 12 14:53:45 I added armv6-novfp and armv6t-novfp for thumb, that should be enough AFAIK Dec 12 14:53:56 ok I'll look now Dec 12 14:54:14 If I have to change something or I did something wrong just tell me Dec 12 14:54:46 seem ok Dec 12 14:55:21 at least AVAILTUNES seem ok Dec 12 14:55:24 let me look the rest Dec 12 14:55:29 ok Dec 12 14:55:43 however the tunes themselves don't seem ok Dec 12 14:56:03 since you got my patch take inspiration from it Dec 12 14:57:08 my concens are that you add armv6 in novfp tunes files Dec 12 14:57:17 the probelm is that armv6 already has vfp Dec 12 14:57:32 however since armv5 was done correctly Dec 12 14:57:47 I guess armv5 == the novfp variant Dec 12 14:59:04 hmmm it's kind of complicated Dec 12 14:59:14 maybe try it out Dec 12 14:59:16 and see if it works Dec 12 14:59:45 GNUtoo, what kind of images are available for shr? Dec 12 14:59:52 all the oe images Dec 12 15:00:11 including shr-image and shr-lite-image and aurora-image Dec 12 15:00:32 I just did bb shr-lite-image Dec 12 15:00:38 ok fine Dec 12 15:00:40 it's building now Dec 12 15:00:44 because in one hand it's long to build Dec 12 15:00:53 but in the other hand it has Xorg etc... Dec 12 15:01:05 maybe I should rather do the tune file? Dec 12 15:01:18 because I've an htcdream that is already somewhat supported Dec 12 15:01:22 but that lacks the tune file Dec 12 15:01:26 GNUtoo, ok Dec 12 15:01:47 also I did meta-geeksphone on meta-smartphone Dec 12 15:01:52 btw what's your sign-off-by? Dec 12 15:01:58 hmmm Dec 12 15:02:04 why not putting in meta-htc? Dec 12 15:02:16 ask JaMa what's best between both Dec 12 15:02:54 GNUtoo, uhm, I saw meta-nokia, meta-htc and just created one for geeksphone Dec 12 15:03:00 ok Dec 12 15:03:09 I don't know what's best but JaMa may know Dec 12 15:03:16 since I'll take your patch and improve it.... Dec 12 15:03:23 I need your sign-off by on it Dec 12 15:03:28 just do: Dec 12 15:03:31 got commit --amend -s Dec 12 15:03:42 or just get me that line Dec 12 15:03:50 GNUtoo, sign-off? I'm a newbie to git Dec 12 15:03:51 ok Dec 12 15:04:00 it's something like that: Dec 12 15:04:27 Signed-off-by: Denis 'GNUtoo' Carikli Dec 12 15:05:00 it's just to give you credit Dec 12 15:05:22 ok ok Dec 12 15:06:54 Signed-off-by: Victor 'vquicksilver' Enriquez Dec 12 15:07:07 something like that I guess Dec 12 15:07:50 ok thanks a lot Dec 12 15:10:34 GNUtoo, so what I need to do now? create a recipe for the kernel? Dec 12 15:11:16 yes and wait for me Dec 12 15:11:27 create all you can Dec 12 15:11:34 while waiting for the vfp thing Dec 12 15:11:39 GNUtoo, ok Dec 12 15:12:04 GNUtoo, so now I'm building with vfp? Dec 12 15:12:16 or whats wrong with the file? Dec 12 15:12:22 I've no idea, it's kind of complicated Dec 12 15:12:32 the arch are wrongs Dec 12 15:12:43 armv6 -> armv6-novfp Dec 12 15:13:52 http://paste.pocoo.org/show/519747/ http://paste.pocoo.org/show/519748/ Dec 12 15:14:04 I did this two files too, for the machine definition Dec 12 15:14:10 still now finished of course Dec 12 15:14:14 not* Dec 12 15:15:06 where can I see, what a MACHINE_FEATURE means? Dec 12 15:17:03 stuff that a machine has Dec 12 15:17:05 like wifi Dec 12 15:17:06 or vfat Dec 12 15:17:09 or similar stuff Dec 12 15:17:19 or alsa, usb host etc... Dec 12 15:18:49 GNUtoo, yes, but how can I see what packages each feature adds? Dec 12 15:19:08 it's in the tasks Dec 12 15:19:15 for wifi it adds wpa_supplicant for example? or it doesn't work that way? Dec 12 15:19:23 yes it works that way Dec 12 15:19:29 wireless-tools etc... Dec 12 15:20:02 ok Dec 12 15:21:06 I don't know if I'm doing the building right, how many recipes does shr-lite-image builds? Dec 12 15:21:42 a lot Dec 12 15:21:52 it's building gstreamer and a lot of things Dec 12 15:21:54 yes Dec 12 15:21:57 yes Dec 12 15:22:05 it's not very lite then Dec 12 15:22:18 it's lite in size Dec 12 15:22:25 oh, ok Dec 12 15:22:31 there is lighter but with less chances of working tough Dec 12 15:22:38 like micro-base-image Dec 12 15:22:45 that one is really light Dec 12 15:22:59 GNUtoo, ok, then I'm going to stop it, and build that instead Dec 12 15:23:05 ok Dec 12 15:23:09 or stop and wait for me Dec 12 15:23:28 GNUtoo, ok Dec 12 15:23:39 but building don't hurt...so Dec 12 15:23:57 anyway of making bitbake more verbose? Dec 12 15:24:04 -DDD Dec 12 15:24:13 like emerge in gentoo for example, to see if the CFLAGS are ok for example Dec 12 15:24:15 ok Dec 12 15:24:18 is geeksphone built by HTC? Dec 12 15:24:52 JaMa, no, it's from qualcomm msm7225 Dec 12 15:25:58 JaMa, it's built by foxconn AFAIK, the use the same hardware that commtiva z1 and other russian clone that I don't remember the name now Dec 12 15:26:56 and also they have two models, now geeksphone one, and geeksphone zero, maybe someone wants to work on zero in the future Dec 12 15:27:16 so maybe it's better to have a folder for them? Dec 12 15:30:21 JaMa, is that ok: http://www.pastie.org/private/34jpagln7mhrgymsql2bqq Dec 12 15:30:24 or is that better: Dec 12 15:30:42 http://paste.pocoo.org/show/519737/ Dec 12 15:30:47 uhm GNUtoo I get this error ERROR: Nothing PROVIDES 'micro-base-image' Dec 12 15:30:47 ERROR: Command execution failed: Traceback (most recent call last): Dec 12 15:30:50 I'm a bit lost in the tunes files Dec 12 15:31:00 VQuickSilver, you need to include meta-micro I guess Dec 12 15:31:08 ok Dec 12 15:31:26 http://www.openembedded.org/wiki/LayerIndex Dec 12 15:32:13 alternatively build a fso2-image Dec 12 15:32:19 or something like that Dec 12 15:32:21 it's cleaner Dec 12 15:32:29 since meta-micro will do some stuff you don't want Dec 12 15:32:31 ok Dec 12 15:32:33 (some bbappend) Dec 12 15:33:00 fso2-console-image Dec 12 15:33:02 to be exact Dec 12 15:35:38 GNUtoo, when I have a suitable userspace to boot it on my phone, I will need and android update.zip that writes boot.img changing the root= parameter to sdcard partition having the userspace files Dec 12 15:36:03 GNUtoo: PACKAGE_EXTRA_ARCHS_tune-armv6t-novfp seems missing armv6t Dec 12 15:36:14 ok Dec 12 15:37:20 thanks a lot Dec 12 15:37:34 VQuickSilver, no need Dec 12 15:37:48 VQuickSilver, your phone has a microsd card slot Dec 12 15:37:57 GNUtoo: please submit it to oe-core yourself as reply to that previous version Dec 12 15:38:03 so just flash the boot.img containing the kenrel and root=/dev/mmcblk0p1 Dec 12 15:38:09 JaMa, ok Dec 12 15:38:15 JaMa, I'll try it before Dec 12 15:38:19 GNUtoo, ok Dec 12 15:38:23 sure Dec 12 15:38:39 JaMa, I've also added a header on 1 of the 2 patch of FBreader Dec 12 15:38:57 I've also tried to upgrade fbreader without success(compilation failure with C++ and curl wrapper) Dec 12 15:48:08 VQuickSilver, could you post your arm1136ejs tune file? Dec 12 15:48:26 GNUtoo, yup Dec 12 15:49:43 http://paste.pocoo.org/show/519759/ Dec 12 15:50:57 GNUtoo, I just copied it from other similar file and changed the name and the TUNE_FEATURES Dec 12 15:51:38 the original was this tune-arm1136jf-s.inc Dec 12 15:52:32 ok Dec 12 15:53:26 ah that's why you have a - Dec 12 15:54:13 ahh I start to understand Dec 12 15:54:20 you were right and I were wrong Dec 12 15:54:31 my modified version of the armv6 is not good Dec 12 15:54:41 because vfp is a supplementary flag Dec 12 15:55:04 GNUtoo, yeah, like thumb Dec 12 16:00:04 so change http://paste.pocoo.org/show/519759/ and restart building Dec 12 16:00:08 TUNE_FEATURES_tune-arm1136ejs = "$-{TUNE_FEATURES_tune-armv6t-novfp} arm1136ejs" Dec 12 16:00:11 change it in Dec 12 16:00:16 TUNE_FEATURES_tune-arm1136ejs = "${TUNE_FEATURES_tune-armv6t-novfp} arm1136ejs" Dec 12 16:01:35 ok Dec 12 16:01:43 also I've a small issue Dec 12 16:01:45 I'll fix Dec 12 16:02:54 TUNE_FEATURES_tune-armv6t ?= "armv6t vfp thumb" Dec 12 16:02:58 I guess it need the t Dec 12 16:03:09 ah no sorry Dec 12 16:03:13 wrong line Dec 12 16:03:24 GNUtoo, I need to make clean or something like that? Dec 12 16:03:36 no idea Dec 12 16:03:58 the t is for thumb Dec 12 16:04:00 I think Dec 12 16:04:56 yes it is Dec 12 16:05:06 you got it nearly working Dec 12 16:05:08 I didn't Dec 12 16:05:24 TUNE_FEATURES = "$-{TUNE_FEATURES_tune-armv6t-novfp} arm1136ejs" Dec 12 16:05:24 TARGET_FPU = "soft" Dec 12 16:05:36 just remove the - and it should work for you Dec 12 16:06:26 ok, I didn't notice the - until you told me about it Dec 12 16:07:46 GNUtoo, done, I need to add the TARGET_FPU line too? Dec 12 16:09:06 no Dec 12 16:09:32 the lines I've pasted is what you pasted in private Dec 12 16:10:12 they come from there Dec 12 16:11:01 GNUtoo, ok ok Dec 12 16:11:22 GNUtoo, then bitbake fso2-console-image and cross fingers Dec 12 16:11:43 yes Dec 12 16:12:01 or rather Dec 12 16:12:06 rm your tmpdir Dec 12 16:12:10 and restart from scratch Dec 12 16:12:14 and cross fingers Dec 12 16:12:15 ok Dec 12 16:15:36 SHR: 03Martin.Jansa 07meta-smartphone * r6ab80148f703 10/meta-shr/recipes-navigation/orrery/ (orrery/orrery.png orrery/use.GdkPixbuf.patch orrery_2.7.bb): orrery: import from OE-classic Dec 12 16:15:37 SHR: 03Martin.Jansa 07meta-smartphone * r1850995cb1a0 10/meta-shr/recipes-extended/gnuplot/ (6 files in 3 dirs): gnuplot: import from OE-classic, upgrade to 4.4.4 Dec 12 16:15:37 SHR: 03Martin.Jansa 07meta-smartphone * ra56fe1751668 10/meta-shr/recipes-sato/claws-mail/ (9 files in 3 dirs): claws-mail: import from OE-classic Dec 12 16:15:38 SHR: 03Martin.Jansa 07meta-smartphone * r3412f3836f99 10/meta-shr/recipes-navigation/omgps/ (5 files in 2 dirs): omgps: import from OE-classic Dec 12 16:15:39 SHR: 03Martin.Jansa 07meta-smartphone * rf87eec697ef8 10/meta-shr/recipes-support/libetpan/ (3 files in 2 dirs): libetpan: import from OE-classic Dec 12 16:15:39 SHR: 03Martin.Jansa 07meta-smartphone * r776038f0febf 10/meta-shr/recipes-support/links/ (7 files in 2 dirs): links(-x11): import from OE-classic Dec 12 16:15:39 SHR: 03Martin.Jansa 07meta-smartphone * r4d5df6b3964c 10/meta-shr/recipes-support/gpm/ (4 files in 2 dirs): gpm: import from OE-classic Dec 12 16:15:39 SHR: 03Martin.Jansa 07meta-smartphone * r5e19c1ff9840 10/meta-shr/recipes-shr/3rdparty/podboy_svn.bb: podboy: fix gst-plugins names to match oe-core Dec 12 16:15:39 SHR: 03Martin.Jansa 07meta-smartphone * r74ea1d5ade06 10/meta-shr/recipes-shr/3rdparty/podboy_svn.bb: podboy: remove unused dir to fix QA issue Dec 12 16:15:50 SHR: 03Martin.Jansa 07meta-smartphone * r35a2482631e7 10/meta-shr/recipes-shr/3rdparty/podboy_svn.bb: podboy: upgrade to 1.7.3 Dec 12 16:15:50 SHR: 03Martin.Jansa 07meta-smartphone * r4c1554f84998 10/meta-shr/recipes-shr/tasks/task-shr-feed.bb: task-shr-feed: add packages from #1436 and #1569 Dec 12 16:16:10 wow claw mails!!!! Dec 12 16:16:45 GNUtoo, I need to make git pull or something before I start building? Dec 12 16:18:11 cause I get Error, the PACKAGE_ARCHS variable does not contain TUNE_PKGARCH (armv6t). Dec 12 16:20:11 ah me too Dec 12 16:20:14 then let me try to fix it Dec 12 16:21:12 GNUtoo, ok Dec 12 16:21:39 JaMa, any idea on how to clean a machine totally(where I've only 1 machine for the same arch(armv6)) Dec 12 16:21:52 ? Dec 12 16:22:50 you can try something like this http://paste.pocoo.org/show/519788/ if I understand you right Dec 12 16:24:12 JaMa, for instance I change the arch of the htcdream too often, will it be a problem? Dec 12 16:25:17 ok Dec 12 16:26:50 ah no that's for changing machine at build Dec 12 16:26:58 not for changing the machine architecture Dec 12 16:27:06 for instance the machine was armv6-vfp Dec 12 16:27:13 and becomes armv6-novfp Dec 12 16:27:21 that would contaminate staging etc...right? Dec 12 16:28:24 TUNE_FEATURES = "armv6 thumb arm1136js" Dec 12 16:28:24 TARGET_FPU = "soft" Dec 12 16:28:30 sort of.. as staging is machine specific now Dec 12 16:28:39 you should rebuild from scratch that machine Dec 12 16:28:43 so I rm htcdream staging and restart Dec 12 16:29:11 and stamps (see the script) Dec 12 16:29:16 ok Dec 12 16:30:10 VQuickSilver, Dec 12 16:30:13 -PACKAGE_EXTRA_ARCHS_tune-armv6t-novfp ?= "${PACKAGE_EXTRA_ARCHS_tune-armv5e} armv6" Dec 12 16:30:13 +PACKAGE_EXTRA_ARCHS_tune-armv6t-novfp ?= "${PACKAGE_EXTRA_ARCHS_tune-armv5e} armv6 armv6t" Dec 12 16:30:32 ah wait a second Dec 12 16:30:37 I had sanity disabled Dec 12 16:30:43 (for getting bitbake -e to work) Dec 12 16:44:21 GNUtoo, ok Dec 12 17:05:23 hmmm Dec 12 17:05:27 where is FREESMARTPHONE_GIT defined? Dec 12 17:05:31 I don't dinf it Dec 12 17:05:32 ls Dec 12 17:05:33 oops Dec 12 17:06:26 found? Dec 12 17:07:35 oe-core/meta/conf/bitbake.conf Dec 12 17:07:43 ah ok Dec 12 17:07:52 thanks Dec 12 17:44:15 hi, is there someone that speaks german and has a gta04? Dec 12 17:49:30 SHR: 03Martin.Jansa 07shr-chroot * rdd8c1188802b 10/ (572 files in 59 dirs): system upgrade Dec 12 18:55:15 JaMa: now I'm starting to be free for a chat, though maybe connectivity issues in tube will hit Dec 12 19:26:20 dos1, hi Dec 12 19:43:12 heyho Dec 12 19:43:14 GNUtoo: ping Dec 12 19:43:51 hi Dec 12 19:45:05 morphis, pong Dec 12 19:45:23 GNUtoo: I am really wondering how waking up on incoming call works on htcdream Dec 12 19:45:54 as it can't work for the nexus s with just theses two patches Dec 12 19:46:09 the 2 patches are specific to the htcdream Dec 12 19:46:13 ok Dec 12 19:46:15 did you do what I said? Dec 12 19:46:22 cat /proc/wakelocks Dec 12 19:46:29 look there, there is no SMD_RPC Dec 12 19:46:35 yes Dec 12 19:46:39 but thats not the problem Dec 12 19:46:40 but there are modem_ things Dec 12 19:46:46 ah ok Dec 12 19:46:52 it never arrives there Dec 12 19:46:57 think about this: Dec 12 19:47:05 you put your device in suspend Dec 12 19:47:09 did you add printks to the kenrel? Dec 12 19:47:15 with disconnecting usb and don't touch the device Dec 12 19:47:20 ok Dec 12 19:47:22 so it will power down Dec 12 19:47:28 so real suspend Dec 12 19:47:42 yes Dec 12 19:48:16 now a phone call arrives Dec 12 19:48:24 ok Dec 12 19:48:28 the wake lock for the modem is locked by the modem driver Dec 12 19:48:36 or is it the modem itself? Dec 12 19:48:49 no let me explain how it works on normal phones Dec 12 19:48:52 like htcdream Dec 12 19:48:57 s/normal/other Dec 12 19:49:04 on htcdream you suspend Dec 12 19:49:11 you disconnect the USB Dec 12 19:49:20 then it really suspend, asyncronously Dec 12 19:49:30 when nothing access any wakelocks Dec 12 19:49:37 wakelocks prevent real suspend Dec 12 19:49:42 then you call it Dec 12 19:49:52 and the modem is on a separate CPU Dec 12 19:50:04 which will wake up the CPU Dec 12 19:50:08 trough an interupt line Dec 12 19:50:11 via a GPIO, right? Dec 12 19:50:15 then the cpu is awake Dec 12 19:50:16 no idea Dec 12 19:50:18 ok Dec 12 19:50:26 but once awake the cpu has to make a decision Dec 12 19:50:32 will I completely wake up? Dec 12 19:50:39 or will I go back to sleep Dec 12 19:50:39 ? Dec 12 19:50:46 that is done by userspace Dec 12 19:50:57 trough the fsousaged plugin Dec 12 19:51:05 yes Dec 12 19:51:10 it look at the resume reason(was it a phone call?) Dec 12 19:51:19 and if it's a phone call it resume totally Dec 12 19:51:23 and prevent suspend Dec 12 19:51:26 not trough wakelocks Dec 12 19:51:36 but trough echo on > /sys/power/state (full power) Dec 12 19:52:03 so now, can I see your kernel debug code? Dec 12 19:52:32 because I guess it works with wakelocks somehow Dec 12 19:52:52 but it can't work with the lowlevel_android plugin Dec 12 19:53:00 as it only watches for the power key and nothing else Dec 12 19:53:15 no Dec 12 19:53:21 it watches the resume reason too Dec 12 19:53:33 from /proc/ or /sys Dec 12 19:53:36 I don't remember Dec 12 19:53:37 really? Dec 12 19:53:43 yes, read the code Dec 12 19:53:50 maybe there are some bugs in the code Dec 12 19:53:54 the core to do this was there Dec 12 19:53:55 but that's the intention Dec 12 19:53:59 s/core/code/ Dec 12 19:54:00 yes Dec 12 19:54:01 morphis meant: the code to do this was there Dec 12 19:54:09 nomally it should be there Dec 12 19:54:15 there may have some bugs tough Dec 12 19:54:19 but should the device wakeup if there's only some signal strength coming in? Dec 12 19:54:38 no Dec 12 19:54:55 because the modem won't send the wakeup for some signal strength Dec 12 19:54:59 but thats the case if we wakeup the device with watching the modem_rx wakelock Dec 12 19:55:02 *wakeup signal Dec 12 19:55:09 ok Dec 12 19:55:16 hmmm Dec 12 19:55:39 we should only wakeup on calls or sms Dec 12 19:55:54 else if there is nothing better Dec 12 19:55:58 we're awake already but can check if we should wake up completely Dec 12 19:56:00 we have to implement the following Dec 12 19:56:06 the userspace checks if it's a call Dec 12 19:56:17 and goes back to sleep if it's just a signal strength Dec 12 19:56:21 yes Dec 12 19:56:30 I already started with some code in fsogsmd to do this Dec 12 19:56:45 but I have the investigate a bit more before making the final decision Dec 12 19:56:50 yes Dec 12 19:56:54 2 questions: Dec 12 19:56:59 1)how replicant does this Dec 12 19:57:13 2)how android does this with its proprietary things Dec 12 19:57:30 freesmartphone.org: 03morphis 07cornucopia * rc6f746e6d191 10/fsousaged/src/plugins/lowlevel_android/plugin.vala: Dec 12 19:57:30 freesmartphone.org: fsodeviced: lowlevel_android: cleanup code a bit and comment unused parts to avoid compiler warnings Dec 12 19:57:30 freesmartphone.org: Signed-off-by: Simon Busch Dec 12 19:57:32 freesmartphone.org: 03morphis 07cornucopia * r5eb4d38db067 10/fsogsmd/src/plugins/modem_samsung/channel.vala: Dec 12 19:57:32 freesmartphone.org: fsogsmd: modem_samsung: defer suspend as long as we have pending requests with no responses Dec 12 19:57:32 freesmartphone.org: Signed-off-by: Simon Busch Dec 12 19:57:34 freesmartphone.org: 03morphis 07cornucopia * re66ac380fd72 10/libfsotransport/fsotransport/commandqueue.vala: Dec 12 19:57:34 freesmartphone.org: libfsotransport: commandqueue: add simple method to get known about the current state of the queue Dec 12 19:57:34 freesmartphone.org: Signed-off-by: Simon Busch Dec 12 19:57:34 freesmartphone.org: 03morphis 07cornucopia * r79f2ce3eac64 10/fsogsmd/src/plugins/modem_nokia_isi/plugin.vala: Merge branch 'master' of ssh://git.freesmartphone.org/cornucopia Dec 12 19:57:36 freesmartphone.org: 03morphis 07cornucopia * r0977cc90fa40 10/fsousaged/ (2 files in 2 dirs): Dec 12 19:57:36 freesmartphone.org: fsousaged: lowlevel_android: make resume reason for modem a config option Dec 12 19:57:36 freesmartphone.org: Not on every device the resume reason for modem activity is named SMD_RPCCALL. This is Dec 12 19:57:36 freesmartphone.org: mostly the case on MSM devices but for example not for the Nexus S from Samsung. Dec 12 19:57:37 freesmartphone.org: Signed-off-by: Simon Busch Dec 12 19:57:37 freesmartphone.org: 03morphis 07cornucopia * r8d65a9bd49b4 10/fsousaged/src/plugins/lowlevel_android/plugin.vala: Dec 12 19:57:38 freesmartphone.org: fsousaged: lowlevel_android: refactor screen resume handling into own method Dec 12 19:57:45 I don't know Dec 12 19:58:00 I searched a little bit in the android sources but didn't found anything valuable Dec 12 19:58:10 ok Dec 12 19:58:34 freesmartphone.org: Signed-off-by: Simon Busch Dec 12 19:59:17 maybe it's not in the android sources? Dec 12 19:59:26 maybe it could be somewhere else like in the kernel? Dec 12 19:59:43 there I didn't found anything too Dec 12 19:59:56 ok Dec 12 20:00:13 do you have more details on your steps? Dec 12 20:00:29 more precisely the debuggiing patches and result for getting what wake the phone Dec 12 20:00:32 like printk Dec 12 20:00:33 I did some grep'ing through the android sources Dec 12 20:00:48 I enabled some debug output but I don't saved something Dec 12 20:00:54 for instance in the 2 patches mickeyl did, did you insert some printks? Dec 12 20:01:01 I did that while I was in the train Dec 12 20:01:04 yes Dec 12 20:01:04 ok Dec 12 20:01:11 what was the result of it? Dec 12 20:02:21 first it didn't received at mickeyl's code Dec 12 20:02:31 so I started to work around it Dec 12 20:02:34 the 2 kenrel patches Dec 12 20:02:38 ok Dec 12 20:02:39 I never got some resume reason Dec 12 20:02:41 yes Dec 12 20:02:44 ah ok Dec 12 20:03:11 so you enabled wakelocks debugging right? Dec 12 20:03:13 I tried with replacing SMD_RPCCALL but it didn't worked too Dec 12 20:03:16 yes Dec 12 20:03:24 the printk's Dec 12 20:03:45 with /sys/module/wakelock/parameters/debug_mask Dec 12 20:03:49 right Dec 12 20:04:29 I tried with putting the device into suspend Dec 12 20:04:34 and connecting the usb cable Dec 12 20:04:45 so it's not fully resumed when I am on it with ssh Dec 12 20:05:09 what about replicant? Dec 12 20:05:20 does it work? Dec 12 20:05:22 it seem so Dec 12 20:05:31 yes Dec 12 20:05:47 ok so we have a starting point Dec 12 20:05:54 I searched through the strings of the samsung-ril too Dec 12 20:06:03 (I mean the original samsung ril) Dec 12 20:06:11 ok Dec 12 20:06:48 I have some initial code here: http://pastie.org/3006889 Dec 12 20:06:55 for fsogsmd Dec 12 20:07:15 as we should have more stuff in userland than in kernel Dec 12 20:07:28 so we should deliver only the plain resume reason Dec 12 20:07:56 let's look how replicant does it Dec 12 20:07:59 which we can check but we need to find out if we can resume completely on our own when we're in awake state Dec 12 20:07:59 ok Dec 12 20:08:11 it should be somewhere in the ril source, right? Dec 12 20:08:26 is hardware/libhardware_legacy/power/power.c still used? Dec 12 20:08:41 morphis, for resuming it's dead simple Dec 12 20:08:49 echo on > /sys/power/save Dec 12 20:08:54 oops Dec 12 20:08:59 echo on > /sys/power/state Dec 12 20:09:13 see hardware/libhardware_legacy/power/power.c Dec 12 20:10:14 yes Dec 12 20:10:19 I saw that Dec 12 20:10:25 ok Dec 12 20:10:28 but where is the decision to call that? Dec 12 20:10:31 when a call arrives? Dec 12 20:10:37 no idea Dec 12 20:10:41 echo on > ... is the last resume step Dec 12 20:10:42 maybe in the java layer? Dec 12 20:10:46 to wake up completely Dec 12 20:10:47 hm Dec 12 20:10:55 hard to find I think Dec 12 20:11:00 rilj maybe Dec 12 20:11:43 rilj? Dec 12 20:11:59 yes the java part of the ril Dec 12 20:12:24 where is it? Dec 12 20:13:08 I think somewhere in frameworks/base/ Dec 12 20:13:46 what's the mock ril? Dec 12 20:15:18 else there is also ./android/telephony in frameworks/base/ Dec 12 20:16:26 static final String LOG_TAG = "RILJ"; Dec 12 20:16:49 frameworks/base/telephony/java/com/android/internal/telephony/RIL.java Dec 12 20:17:30 theses phone stuff is like mission impossible where you have to assemble a team but need one android person, one modem specialist, one kernel specialist etc.... Dec 12 20:19:06 set_screen_state should be the highest lock if I remember well Dec 12 20:19:48 look in rilj: Dec 12 20:19:55 if (intent.getAction().equals(Intent.ACTION_SCREEN_ON)) { Dec 12 20:19:55 sendScreenState(true); Dec 12 20:26:07 morphis, fsogsmd fails to compile Dec 12 20:26:15 snd.h: no such file or directory Dec 12 20:27:16 http://developer.android.com/reference/android/content/Intent.html Dec 12 20:27:32 for more bacground on rilj part Dec 12 20:29:10 yo Dec 12 20:29:17 GNUtoo: snd.h? Dec 12 20:29:33 GNUtoo: too old libsamsung-ipc? Dec 12 20:29:43 I use autorev Dec 12 20:30:58 http://pastie.org/private/xuwxsakysljpavfold9qa Dec 12 20:33:59 I'll cleansstate libsamsung-ipc and fsogsmd Dec 12 20:35:23 hm Dec 12 20:35:55 GNUtoo: maybe AUTOREV is the problem ... Dec 12 20:36:58 ah ok Dec 12 20:37:48 com/android/internal/telephony/SMSDispatcher.java -> grep for acquire Dec 12 20:39:23 morphis, found it Dec 12 20:39:29 /** This is probably an MT call that we first saw in a CLCC response */ Dec 12 20:39:48 and similar messages Dec 12 20:40:03 where? in SMSDispatcher.java? Dec 12 20:40:25 ah maybe I was wrong Dec 12 20:40:30 never mind I'll look better Dec 12 20:40:40 com/android/internal/telephony/gsm/GsmConnection.java Dec 12 20:40:49 in frameworks/base/telephony/java Dec 12 20:40:52 hmmm Dec 12 20:43:43 anyway if I remember well the sms only rings and doesn't wake up the screen Dec 12 20:43:55 so I hope I find for the calls Dec 12 20:44:55 SMS is something we just need save and then we can sleep again Dec 12 20:44:58 so no real resume Dec 12 20:45:24 but I will check first for the call scenario Dec 12 20:45:26 we need to resume Dec 12 20:45:30 else it never rings Dec 12 20:45:32 btw. we have to create the ALSA state files too Dec 12 20:45:34 yes Dec 12 20:45:34 when you have an sms Dec 12 20:45:35 I know Dec 12 20:45:40 I will implement this Dec 12 20:45:42 like on htcdream.... Dec 12 20:45:48 where you misses all sms because of that Dec 12 20:45:51 we have the org.freesmartphone.Usage.Resume method Dec 12 20:46:35 I will implement it that way that if we're calling the Resume method the suspend method finishes and the system is alive Dec 12 20:46:59 maybe just ring and go back to sleep Dec 12 20:47:37 LOL - I just had a glance (lengthy glance though) at out old friend's http://www.gizmoforyou.net site. Dang, they have modular *projects*, and merged *projects*, and interesting *projects*, and dunno what else *projects* and *devices*. I wasn't able to find a google hit on the site for the term "phone" Dec 12 20:48:00 yes they have some expensive phones Dec 12 20:48:02 s/out/our/ Dec 12 20:48:02 DocScrutinizer meant: LOL - I just had a glance (lengthy glance though) at our old friend's http://www.gizmoforyou.net site. Dang, they have modular *projects*, and merged *projects*, and interesting *projects*, and dunno what else *projects* and *devices*. I wasn't able to fi... Dec 12 20:48:19 GNUtoo: hm Dec 12 20:48:30 GNUtoo: when a call comes in the system should wake up Dec 12 20:48:31 uhuh, and what'S the name of the *project* that's supposed to be a phone? Dec 12 20:49:23 DocScrutinizer, let me find it Dec 12 20:49:25 SHR? ;) Dec 12 20:49:28 really - I rarely ever seen a site with worse advertising Dec 12 20:49:54 GizmoForYou FLOW G1.55 Dec 12 20:51:52 be back soon, phone ringing Dec 12 20:54:46 http://www.gizmoforyou.net/site/en/gizmos.html doesn't have the term phone, though it's better than the pages http://www.gizmoforyou.net/site/en/projects.html or http://www.gizmoforyou.net/site/en/projects.html?section=projects&workspace=59&task=display_details&id=59&35c8424a7e4f9b57c8050790bf5f368a=1 which completely omit all info what the *projects* are meant to DO Dec 12 20:56:18 probably I could read the ShellFish schematics or documentation to eventually find out if it has a GSM radio Dec 12 20:58:46 back Dec 12 21:00:31 so I am off for today Dec 12 21:00:32 gn8 Dec 12 21:42:11 DocScrutinizer: and now I'm back too :) Dec 12 21:42:16 for a while.. **** ENDING LOGGING AT Tue Dec 13 01:20:13 2011 **** BEGIN LOGGING AT Tue Dec 13 01:22:14 2011 **** ENDING LOGGING AT Tue Dec 13 01:30:04 2011 **** BEGIN LOGGING AT Tue Dec 13 01:32:14 2011 **** ENDING LOGGING AT Tue Dec 13 02:59:57 2011