**** BEGIN LOGGING AT Fri Sep 07 02:59:59 2012 Sep 07 06:35:13 woosh fsogsmd/libfsotransport/lbgsm0710x do not handle voice message enabled phone Sep 07 06:35:53 what is a voice message enabled phone? Sep 07 06:49:39 yes (and sorry for the fuss ... it might ends up at another issue being the culprit -> same freerunner, same sim card -> the sms are seen as disabled then fsogsmd tells me there are no sim card Sep 07 06:51:24 I have debug logs from before and after which helps but does not avoid mistakenly blaming the worng sequence . ie CPVWI=1 returns CME error in both case which looks fine now that I looked at it a second time Sep 07 07:04:42 pabs3: looks like the feature (AT %CPVWI) that enable one to get the phone to retrieve the voice messages locally Sep 07 07:05:24 found out in Cphs4_2.ww6.doc Sep 07 07:06:09 this is not part of gsm specs but a defacto standard made by an european (british at first) consortium Sep 07 07:07:41 interesting Sep 07 07:07:55 thankfully I did not paid the 445$ for the MWI/VMWI specs from telico as they are useless for this matter (and the cphs doc was availble online without NDA or else and free as it seems Sep 07 07:09:45 indeed interesting and there is way more than voice messages in this "spec". But one thing at a time (I will focus on getting the sms and sim card detection back on trail on this freerunner . Sep 07 07:36:37 might be 1d41feb3b65152a91e15656c6d58910de2d95f5f on cornucopia at fault Sep 07 09:25:53 k, continuing yesterdays adventure Sep 07 09:26:10 i'v got the older sim, and sms retrieval does work Sep 07 09:26:19 so indeed it was an issue with 3g sim Sep 07 09:27:41 i'll go and buy other carrier's sim, that will also be 3g, and test that one too. it'll be sad sad day if my openmoko cant handle the only sim that i can buy here :| Sep 07 09:59:31 ggoZ: do you have the moko11 firmware for the gsm modem? Sep 07 09:59:52 see http://wiki.openmoko.org/wiki/Flashing_the_GSM_Firmware Sep 07 10:00:44 "Moko11 should fix #666 (compatibility with some 3G sim cards), and also has a new command AT+CSIM. " Sep 07 10:00:49 probably not, will look into it Sep 07 11:11:54 "gdouble _tmp7_ = log10(31.0);" <- gcc segfault on freerunner latest shr compiling that line Sep 07 11:13:37 ggoZ: I found out that /etc/freesmartphone/conf/default/fsogsmd.conf is not the file to change . /etc/freesmartphone/conf/GTA02/fsogsmd.conf is (set every log to DEBUG and you will get plenty of details Sep 07 11:15:35 http://git.freesmartphone.org/?p=cornucopia.git;a=commit;h=1d41feb3b65152a91e15656c6d58910de2d95f5f is the commit I am trying to revert Sep 07 11:16:41 I can reproduce the beginning of your issue (afterwards I get a no sim card error ... I have the log from May that worked with that exact sim card and the same freerunner Sep 07 11:17:06 hm Sep 07 11:17:16 so I compare and revert what could be at fault (if only gcc was not segfaulting at me while compiling fsogsmd Sep 07 11:19:07 in parallel I rebuild fsogsmd in shr-chroot though it also takes way more time Sep 07 11:38:27 I am afraid that one of the patch from openembedded breaks constant 31.0 as not a double Sep 07 11:50:16 cross compiler on debian does not suffer this issue, debian gcc neither ... even trimmed down the test futher : "#include \n#include \n int main () {double tmp7 = log10(31.0); return 0;}" gcc -S input_file.c -lm Sep 07 11:50:56 well debian cross for arm and debian x86_64 gcc Sep 07 11:51:23 did not tested native build on the freerunner with a debian image yet Sep 07 13:05:23 GNUtoo, hello , I have made some progress Sep 07 13:06:14 after boot into linux, I modprobe -i fbcon then the screen can show log and never stop as modprobe never finish Sep 07 13:06:54 ah? Sep 07 13:06:56 dmesg? Sep 07 13:07:36 yo GNUtoo ! Sep 07 13:07:41 hi Sep 07 13:08:08 how are you? Sep 07 13:08:15 fine Sep 07 13:08:17 and you? Sep 07 13:09:45 GNUtoo, noting in dmesg Sep 07 13:09:59 maybe it fails in __init Sep 07 13:10:47 http://blog.csdn.net/bonderwu/article/details/5805961 Sep 07 13:10:47 it said:" it will load driver by reading vendor from MDDI Bridge, in debug, bootloader can read correct vendor, so bootloader do the initial for LCD, but kernel can't, using JTAG, MODULE INIT in kernel can't read correct vender, so driver is not loaded, if we don't do MDDI Bridge inital in bootloader, MODULE INIT can do well Sep 07 13:10:48 GNUtoo, but screen really can't stop print log, and lsmod show it is loaded Sep 07 13:11:24 good. feeling really well these days. just returned from my first real open source conference, which means trying to get back into OSS Sep 07 13:11:27 what is printed on screen? Sep 07 13:11:34 wow Sep 07 13:11:37 and another, it is not important, but it can explain why msm_fb_type is unknown Sep 07 13:11:38 what conference? Sep 07 13:11:45 smartdevcon.eu, in poland Sep 07 13:11:55 ok Sep 07 13:12:15 did anyone make any advancements wrt. n900? Sep 07 13:12:30 i think i can donate mine, if someone wants. not really using it anymore Sep 07 13:12:30 https://paste.debian.net/188420/ Sep 07 13:12:36 same as for a bunch of more hardware, btw. Sep 07 13:13:03 will try to make a list Sep 07 13:13:27 what it print to screen is hard to read, screen looks like scan, even it print a new line, it scan from up to down Sep 07 13:14:13 mickey_office, with n900 yes and no Sep 07 13:14:48 basically there are progress in a branch for the forwarder Sep 07 13:14:48 but not integrated in mainline Sep 07 13:14:50 mainline cornucopia I mean Sep 07 13:15:04 ok Sep 07 13:15:11 everyone seem interested in having n900 work, no one works on it Sep 07 13:15:16 heh Sep 07 13:15:17 pity Sep 07 13:15:28 indeed so I unblocked the n900 forwarder Sep 07 13:15:36 giving a clear direction on what needed to be done Sep 07 13:15:39 cool Sep 07 13:15:46 but still, no contributions Sep 07 13:17:30 morphis: how applicable is the work you did to, say, a galaxy nexus? Sep 07 13:18:36 mickey_office, we already support galaxy nexus Sep 07 13:18:48 GNUtoo, for the screen thing(the kernel source) can I modify it to this https://paste.debian.net/188423/ Sep 07 13:18:58 GNUtoo: ah, cool Sep 07 13:19:05 you see i've been somewhat out-of-the-loop ;) Sep 07 13:19:43 mickey_office, please see PM Sep 07 13:19:50 I'll put you in the looop Sep 07 13:20:36 oki Sep 07 13:28:36 bbl, will be here more often in the future now ;) Sep 07 13:28:41 weekend now though Sep 07 13:37:32 GNUtoo, then any idea about the first thing? __init failed should show in dmesg , doesn't it? Sep 07 13:37:57 not if __init doesn't print anything Sep 07 13:38:26 I'll look later Sep 07 13:40:45 GNUtoo, thank you, then what shall I do at present? Sep 07 13:42:34 add printk in the module __init Sep 07 13:42:40 of fbcon Sep 07 13:42:52 also you need CONFIG_VT Sep 07 13:43:03 and in CONFIG_VT add console binding and unbinding Sep 07 13:46:20 GNUtoo, yes, I will, but about printk in the module __init meaning I shall modify kernel source ? Sep 07 13:49:23 yes Sep 07 13:49:34 let me look 2s Sep 07 13:51:07 add some printk here: Sep 07 13:51:09 static int __init fb_console_init(void) Sep 07 13:51:13 inside that function Sep 07 13:51:42 I see, the function is in the fbcon.c ? Sep 07 13:52:58 but in my source, it is static int __init fb_console_setup(char *this_opt) Sep 07 13:53:43 yes Sep 07 13:54:01 then printk there Sep 07 13:54:05 but doubble check Sep 07 13:54:09 *doubble check before Sep 07 13:55:51 yes, I will, but actually, I don't know much about this, I don't know what I shall printk, but I will try Sep 07 13:57:49 printk(KERN_ERR," HERE I am %:%i\n",,__FILE__,__LINE__); in all if sentence Sep 07 14:05:08 wait... Sep 07 14:05:09 ,, Sep 07 14:05:27 you have ,, in what you pasted Sep 07 14:05:43 and % instead of %s Sep 07 14:07:56 GNUtoo, sorry, the line is too long, I just done it I find a pastebin Sep 07 14:08:03 ok Sep 07 14:08:09 pastie.org Sep 07 14:09:21 http://pastie.org/4680143 Sep 07 14:10:13 sorry, I need modify it Sep 07 14:10:42 http://pastie.org/4680146 Sep 07 14:26:24 also be sure it's the init function Sep 07 14:26:26 like that Sep 07 14:27:16 module_init(fb_console_init); Sep 07 14:32:14 GNUtoo, ? Sep 07 14:32:31 back Sep 07 14:32:59 but what is meaing of "also be sure it's the init function" ? Sep 07 14:33:24 module_init(fb_console_init); Sep 07 14:34:09 yes, but what does it mean? I didn't modify the correct function? Sep 07 14:34:23 just double check that there is a line like that: Sep 07 14:34:27 module_init(fb_console_init); Sep 07 14:34:36 and that inside module_init it's your function Sep 07 14:34:43 the one you're modifying Sep 07 14:35:12 yes, there is a such line Sep 07 14:36:10 ok Sep 07 14:39:37 do I need to modify the "static int __init fb_console_init(void)" function? Sep 07 14:45:32 GNUtoo, ? Sep 07 14:45:49 no idea Sep 07 14:45:56 please pastebin your current source Sep 07 14:48:25 sorry, it is more than 64KB, let's use my site Sep 07 14:51:15 https://sumomo.r.gd/fbcon.c Sep 07 14:54:34 ok Sep 07 14:55:24 you did some mistakes Sep 07 14:55:47 1) move the file and restore the previous one Sep 07 14:55:54 2) hook that function instead: Sep 07 14:55:55 fb_console_init Sep 07 14:56:10 else there are some stuff in kernel hacking for debugging Sep 07 14:56:15 like function tracer etc... Sep 07 14:56:23 or maybe it was not integrated yet Sep 07 14:56:28 I must look Sep 07 14:56:55 at least there were patches from tim bird (sic) Sep 07 14:57:35 I see, I will move all hook in fb_console_setup Sep 07 14:57:37 (he's famous for encouraging the violation of the GPL) Sep 07 14:57:49 and do that instead: Sep 07 14:58:15 printk(KERN_ERR,"HERE I am %s:%i\n",__FILE__,__LINE__); Sep 07 14:58:19 with one , Sep 07 14:58:21 not two Sep 07 14:58:24 at the beginning Sep 07 14:59:11 (so you don't propagate your typo which may be long to fix if you don't master sed well) Sep 07 14:59:32 ? just put "printk(KERN_ERR,"HERE I am %s:%i\n",__FILE__,__LINE__);" in the fb_console_init ? Sep 07 14:59:45 yes add some Sep 07 15:00:01 put one Sep 07 15:00:03 compile Sep 07 15:00:06 if it works put more Sep 07 15:00:23 and then run it Sep 07 15:06:19 sorry a network problem, GNUtoo how about the new https://sumomo.r.gd/fbcon.c ? Sep 07 15:18:26 let me look Sep 07 15:18:57 sorry, the network is too bad, if the source is correct, please tell me, I will go to sleep soon Sep 07 15:21:28 ayaka, I cannot fetch it.... Sep 07 15:21:34 Connecting to sumomo.r.gd|112.111.22.246|:443... Sep 07 15:21:40 and it's blocked there Sep 07 15:21:59 GNUtoo, no, wait the out of day of dns, it is 3 minutes Sep 07 15:22:13 the new ip is 112.111.22.196 Sep 07 15:23:20 ok Sep 07 15:23:35 I try the http instead of https(certificate issues with the IP) Sep 07 15:24:04 http doesn't seem to work Sep 07 15:24:36 and in https it says: WARNING: cannot verify 112.111.22.196's certificate, issued by `/C=CN/ST=FuJian/L=FuZhou/O=SUMOMO Computer Assciate/OU=localCA/CN=ayaka/emailAddress=ayaka@mail.sumomo.pri':, and at the end: Sep 07 15:24:42 2012-09-07 17:23:02 ERROR 404: Not Found. Sep 07 15:24:45 which is the real error Sep 07 15:24:49 let me try with a browser Sep 07 15:25:17 hmmm Sep 07 15:25:18 of course it can't, there are forum.sumomo.rg, and mail on it Sep 07 15:25:23 ok Sep 07 15:25:29 so vhost stuff? Sep 07 15:25:39 it is multi domains server Sep 07 15:25:44 ah ok Sep 07 15:25:47 so vhost then Sep 07 15:25:55 http 1.1 stuff Sep 07 15:26:06 just wait 3 minutes Sep 07 15:26:15 in plain english all that means that there are multiples websites on the same apache server Sep 07 15:26:22 using the same IP Sep 07 15:26:28 s/IP/IP address/ Sep 07 15:26:30 GNUtoo meant: using the same IP address Sep 07 15:26:37 ok Sep 07 15:26:59 I see, I think it has been out of time Sep 07 15:27:27 I shall change ttl to 1 minutes next time Sep 07 15:27:35 ok I have it Sep 07 15:28:20 ok build and run it Sep 07 15:28:23 also add that: Sep 07 15:28:30 printk(KERN_ERR,"HERE I am %s:%i\n",__FILE__,__LINE__); Sep 07 15:28:32 right here: Sep 07 15:28:47 before and after console_lock() Sep 07 15:28:55 and after fb_register_client Sep 07 15:28:56 etc... Sep 07 15:29:06 not only in the errors (IS_ERR) Sep 07 15:30:15 I see I will Sep 07 15:30:20 ok nice Sep 07 15:31:50 but I have to report tomorrow, and I will go to university in 2 days, maybe I can get JTAG there, I may lose connection that time because I am forced to add army Sep 07 15:33:04 ok Sep 07 15:35:29 s/add/join/ Sep 07 15:35:29 ayaka meant: but I have to report tomorrow, and I will go to university in 2 days, maybe I can get JTAG there, I may lose connection that time because I am forced to join army Sep 07 15:35:48 good bye Sep 07 16:01:13 JaMa|Off, hi Sep 07 16:01:30 wasn't the intone+enjoy bug supposed to be fixed? Sep 07 16:01:38 because there was someone who sent patches right? Sep 07 18:11:13 except I did not send the patch to the tracker :/ Sep 07 18:11:33 he s gone ... will send asap Sep 07 18:25:35 is there a mean to reinstall without --force-reinstall (for libc6 --force-reinstall only breaks the system) Sep 07 18:25:46 ie it removes then fails to do anything Sep 07 18:25:50 morphis: ping, any progress building cornucopia building on jenkins ? Sep 07 18:26:27 I have a segfault building cornucopia fsogsmd on freerunner gta02 , is this related ? Sep 07 18:27:22 I nailed it down to a single line and made a test case ... now I am investigating so if the issue is known and already worked on it might save me time :) Sep 07 18:27:23 prahal2: no, its abould building standalone cornucopia on x86 Sep 07 18:27:37 ok Sep 07 18:28:30 prahal2: are you building ON or FOR freerunner ? Sep 07 18:28:46 ON Sep 07 18:29:07 for I bet will work (I am at 1100 on 2300 tasks) Sep 07 18:29:34 because the issue will likely be with a patch applied to gcc Sep 07 18:30:09 prahal2: may be but i am somehow outdated... Sep 07 18:30:16 ie I tried in the shr-chroot and it worked cross compiling Sep 07 18:30:29 prahal2: this i expected too Sep 07 18:31:19 prahal2: but compiling for x86 leads to error: http://norman-schleicher.de/jenkins/job/cornucopia-nightly/14/consoleFull Sep 07 18:31:26 this I was not confident (isn't the same patchset applied to /OE/shr-core/tmp-eglibc/sysroots/x86_64-linux/usr/bin/armv4t-oe-linux-gnueabi/arm-oe-linux-gnueabi-gcc ? Sep 07 18:31:38 than at gta02 runtime gcc ? Sep 07 18:33:31 prahal2: this i cannot tell you so JaMa is one of the experts or you should ask in #oe Sep 07 18:33:49 should ask in #oe Sep 07 18:34:03 ok thanks Sep 07 18:35:48 prahal2: but sometimes its important where the compiler is running gcc is a very fragile software in that point Sep 07 18:36:47 prahal2: so compiling on x86 and x64 host may have different results Sep 07 18:37:53 you miss a gir files Sep 07 18:38:13 prahal2: what are gir files ? Sep 07 18:38:22 error: Package `gisicomm-1.0' not found in specified Vala API directories or GObject-Introspection GIR directories Sep 07 18:38:40 they are introspection data for bindings (including vala) Sep 07 18:38:48 python , perl , etc Sep 07 18:39:49 hum sorry you just need libgisi-dev Sep 07 18:40:17 prahal2: ok the build process is incomplete it depends on stuff which is created during build and is not installed yet Sep 07 18:40:22 gir or vala api ... in this case it is a vala api not a gobject introspection file that was missing (and it is in above package Sep 07 18:41:16 yes there is a bug (installing the package by hand is just a workaround Sep 07 18:41:43 prahal2: but this is what i wont to do Sep 07 18:41:59 probably the dependency is not written at all Sep 07 18:42:54 prahal2: i think morphis: will continue to fix the build process... so installing any fso packages is not necessary Sep 07 18:44:01 in meta-smartphone/meta-fso/recipes-freesmartphone/cornucopia/fsogsmd.inc you can try adding "libgisi-dev" to DEPENDS . If it works then you got the issue and the fix Sep 07 18:44:54 welll wait this are runtime depends .... I will have to look a bit more to check if this will apply to build deps Sep 07 18:49:43 prahal2: this is not a bitbake build Sep 07 18:50:17 prahal2: its a git clone and pure make build on x64 Sep 07 18:51:45 prahal2: building using bitbake works Sep 07 18:53:11 oh pure make build cannot guess which package holds the -dev data Sep 07 18:53:28 I don't know if there is a m4 to check for vala api files Sep 07 18:54:01 prahal2: so dont care about my problem.. morphis: will fix that. Sep 07 18:54:03 indeed bitbake as a side effect when it build the libgisi get the libgisi-dev as colateral damage Sep 07 18:54:26 thus the issue is hidden even if no explicit depends is told to bitbake Sep 07 18:54:33 ok Sep 07 18:55:43 prahal2: thank you trying helping me :-) Sep 07 19:05:40 should I send the patch for enjoy only to e17 project or is there a component in shr tracker about packages shipped I should also send to ? Sep 07 19:12:41 nvm about the "how to reinstall w/o --force-reinstall" I found out : opkg download ; opkg install ./.ipk Sep 07 19:13:21 hum no sorry ... then my libgcc1 was outdated Sep 07 19:14:50 wait yes it works :) I had an old download in the directory thus been confused Sep 07 21:09:17 GNUtoo: should I send the patch for enjoy to shr tracker ? Sep 07 21:09:25 or only to enlightenment one Sep 07 21:11:43 prahal2, hi Sep 07 21:11:53 enlightenment I guess Sep 07 21:11:57 and then we backport? Sep 07 21:12:03 so send to both Sep 07 21:12:24 yes but shr under component "all" ? Sep 07 21:12:33 I guess so Sep 07 21:12:46 is intone fixed now btw? Sep 07 21:13:29 no Sep 07 21:13:43 ok Sep 07 21:13:51 so you tried to fix enjoy first? Sep 07 21:14:01 I am currently looking at breakage of my 3g card support fsogsmd Sep 07 21:14:06 yes Sep 07 21:14:09 ok Sep 07 21:14:37 I think what we can do is that you send the patch to both enlightenent and SHR Sep 07 21:14:55 and we tell in the patch header that it's pending or something like that Sep 07 21:15:02 (for SHR/oe) Sep 07 21:20:22 ok .http://shr-project.org/trac/ticket/2058 is the shr ticket Sep 07 21:35:15 thanks a lot Sep 07 21:35:25 I'll try to apply it as soon as possible Sep 07 21:38:05 I'll build enjoy right now and try to apply the patch Sep 07 21:38:12 I hope I succeed before going to bed Sep 07 21:38:18 because else tomorow I can't Sep 07 22:04:10 gah I failed I must go to sleep and it's still compiling.... Sep 07 23:44:39 SHR: 03dos 07opimd-utils * re57a0c9624c2 10/ (5 files in 2 dirs): fixes for new python-elementary **** ENDING LOGGING AT Sat Sep 08 02:59:58 2012