IRC log of #maemo-ssu for Saturday, 2013-10-12

*** xes has joined #maemo-ssu00:07
*** _rd has joined #maemo-ssu00:16
*** _rd has quit IRC00:21
*** _rd has joined #maemo-ssu00:22
*** LauRoman has quit IRC00:29
*** _rd has quit IRC00:33
*** _rd has joined #maemo-ssu00:34
*** rd_ has joined #maemo-ssu00:44
*** _rd has quit IRC00:44
*** rd_ has quit IRC00:49
*** WielkiTost has joined #maemo-ssu00:51
*** WielkiTost is now known as dos100:51
*** dos11 has quit IRC00:51
*** Pali has quit IRC01:02
*** dhbiker has quit IRC01:04
*** freemangordon has quit IRC01:08
*** freemangordon has joined #maemo-ssu01:14
*** nox- has joined #maemo-ssu01:24
*** NIN101 has quit IRC01:34
*** kolp has quit IRC01:38
*** joshgillies has joined #maemo-ssu01:55
*** Martix has quit IRC02:23
*** joshgillies has quit IRC02:42
*** tom has quit IRC02:43
*** wmarone_ has joined #maemo-ssu02:52
*** wmarone has quit IRC02:52
*** dafox has quit IRC03:02
*** joshgillies has joined #maemo-ssu03:21
*** jonwil has joined #maemo-ssu03:22
*** jonwil_ has joined #maemo-ssu03:27
*** jonwil_ has quit IRC03:28
*** BCMM has quit IRC03:28
*** jonwil has quit IRC03:29
*** jonwil has joined #maemo-ssu03:31
*** xes has quit IRC03:45
*** Vlad_on_the_road has quit IRC03:47
*** joshgillies has quit IRC03:51
*** dos1 has quit IRC04:40
*** rlinfati has joined #maemo-ssu04:46
rlinfati a copy of the nolo, kernel, and modules for crypto on N900 http://las.menoscero.com/virtu/virtu-datos/Programas/device/maemo-fremantle/flash/RX-51_2009SE_crypto.tar.gz04:47
*** rlinfati has quit IRC04:48
DocScrutinizer05o.O .oO(???)05:07
jonwilLooks risky if it needs to flash a new xloader05:09
DocScrutinizer05>>3. pug usb cable and "IMMIDIATELY" press power button.<< W*T*F?!?!!05:10
DocScrutinizer05well, "risky"05:12
DocScrutinizer05there's always coldflash05:12
DocScrutinizer05but this new xloader might give us some valuable other options05:13
*** discopig has quit IRC05:14
*** nox- has quit IRC05:15
DocScrutinizer05since a) it must be signed, so we have two signed xloaders now, and b) it seems to enable stuff that formerly not been enabled - yet to check what else is possible now except AES, and c) this new xloader (and Nolo?) might come with other goodies aka easter eggs05:15
jonwilThe real question is where these files came from and who published them (and whether it was done with approval or not)05:18
*** amizraa has quit IRC05:21
jonwilstill interesting for those who are game enough to fiddle with it :)05:21
DocScrutinizer05jonwil: all that is a non-question as far as I'm concerned05:24
DocScrutinizer05xloader is *tiny* and easy to diff to original05:24
DocScrutinizer0514.5kB05:25
DocScrutinizer05freemangordon and pali will think xmas came early this year05:25
DocScrutinizer05there's also a flasher.rover in there05:26
DocScrutinizer05kds@kds:~/download$ flasher.rover -i05:26
DocScrutinizer05flasher v2.8.2 (Jan  8 2010)05:26
jonwilnot sure what that flasher is though05:27
DocScrutinizer05flasher-3.5 basically05:27
*** tom has joined #maemo-ssu05:27
DocScrutinizer05probably identical, though who knows for sure, before we checked it05:28
jonwilhmmm, I cant find flasher-3.5 right now05:38
jonwiltablets-dev.nokia.com isnt working and I cant find a mirror of the Linux version05:39
*** amiconn has quit IRC05:39
*** amiconn_ has joined #maemo-ssu05:39
*** amiconn_ is now known as amiconn05:39
jonwilok, found it05:41
jonwiland no flasher.rover is not the same as flasher-3.505:42
jonwilflasher.rover is smaller in size than flasher-3.505:43
jonwiland it appears to be a newer version (2.8.2 vs 2.5.2)05:45
jonwilin any case I will allow those more knowledgeable than me on such things fiddle with that interesting looking tarball05:47
jonwilbtw it looks like it supports sha1/md5 hardware as well as AES hardware05:48
DocScrutinizer05anyway, here's a screenshot I took when managing maemo.cloud-7.de/maemo5/fiasco&co/    http://wstaw.org/m/2013/10/12/plasma-desktopoU3743.png06:00
DocScrutinizer05a pity that this tarball doesn't include the sources for xloader and NOLO/2nd and no signing tool and keyfile for xloader signing06:02
*** amizraa has joined #maemo-ssu06:11
jonwilanyhow, that tarball is a good find :)06:36
jonwiljust like some of the other "good finds" that have popped up lately (like the fact that http://mirror.thecust.net/harmattan-dev.nokia.com/pool/harmattan/nokia-binaries/41667a5bd857be02f487c2ce806fbf85/nokia-binaries/p/pulseaudio-nokia/ contains debug packages for all those pulseaudio blobs)06:37
DocScrutinizer05do we have a mirror of that stuff already?06:59
jonwilthat site looks like a mirror done by someone of harmattan-dev.nokia.com07:00
DocScrutinizer05yeah, and anybody's guess when it vanishes07:00
jonwilfull mirror by the looks of it07:00
jonwilsomeone should mirror it then :P07:01
DocScrutinizer05brkn: around?07:01
* jonwil must remember to find out what the difference between gst-plugins-good, gst-plugins-bad and gst-plugins-ugly is07:05
*** jonwil has quit IRC07:07
*** FlameReaper has joined #maemo-ssu07:49
*** discopig has joined #maemo-ssu08:14
*** joshgillies has joined #maemo-ssu08:46
*** M13 has joined #maemo-ssu08:46
*** rd_ has joined #maemo-ssu08:47
*** LauRoman has joined #maemo-ssu08:56
*** kolp has joined #maemo-ssu09:16
*** joshgillies has quit IRC09:33
*** joshgillies has joined #maemo-ssu09:38
*** FlameReaper has quit IRC09:54
*** rd_ has quit IRC10:03
*** DrCode has quit IRC10:17
*** NIN101 has joined #maemo-ssu10:23
*** rd_ has joined #maemo-ssu10:27
*** DrCode has joined #maemo-ssu10:27
*** FlameReaper has joined #maemo-ssu10:27
*** Pali has joined #maemo-ssu10:58
*** rd_ has quit IRC10:59
*** Vlad_on_the_road has joined #maemo-ssu11:03
*** Vlad_on_the_road has quit IRC11:09
*** joshgillies has quit IRC11:22
*** Vlad_on_the_road has joined #maemo-ssu11:24
*** joshgillies has joined #maemo-ssu11:27
*** rd_ has joined #maemo-ssu11:48
*** Martix has joined #maemo-ssu11:52
Palidocscrutinizer05, jonwil: that tarball with xloader, nolo and crypto modules was sent to maemo ML years ago11:56
Paliand people who worked on it are real and well known11:56
PaliI sent link to that email thread here in irc more times11:56
Paliand I already wrote that I have that tarball on hdd which is disconnected...11:57
Paliso if nobody is able to find it I can provide it (after I connect that hdd somewhere...)11:58
Paliand if you want more info about that flasher.rover please read doc/ folder in 0xFFFF11:59
Palibasicly that flasher has full support for MkII flashing protocol11:59
Paliflasher-3.5 support flashing only eMMC via MkII protocol, that flasher.rover support flashing any image type (except rootfs)12:00
Palijonwil: gst-plugins-good/bad/ugly are different packages and all contains different modules12:01
Paliit is like stable/testing/devel12:02
*** rd_ has quit IRC12:02
Palifreemangordon, jonwil: can you do some diff between pr1.3 xloader and that new crypto xloader?12:05
*** NIN102 has joined #maemo-ssu12:05
Paliand also between secondary nolo?12:05
freemangordonPali: trying it right now12:05
Paliok12:06
freemangordonbut I forgot how to switch between ARM <-> THUMB mode in IDA :(12:06
Paliand look at strings flasher.rover12:07
freemangordonwhy?12:07
Palithere is something "writing certificate"12:07
Paliwhat it could means?12:07
Paliwhere and how can flasher.rover write certificate (and which)?12:08
*** NIN101 has quit IRC12:08
freemangordonI guess we can load a pubkey, as long as it is signed with TI private ;)12:08
Palipubkey of what?12:08
Palipublic key for omap boot rom cannot be changed...12:08
freemangordonpubkey that is used to verify PPA signiture :)12:09
Paliand there is no other signing process...12:09
freemangordonxloader is signed12:09
Paliyes, but signature is verified by omap boot rom and this key cannot be changed by design12:09
Paliharmattan has signed also second part which can be maybe changed...12:10
*** NIN102 is now known as NIN10112:11
Paliwho is going to flash that xloader? :-)12:12
Palior maybe only secondary nolo image? (which is not signed)12:12
freemangordonPali: are you sure xloader signiture is verified against TI key? I doubt12:12
PaliI do not understand why secondary image must be changed too...12:12
Palifreemangordon: yes it is verified against key hardcoded in omap/arm processor12:13
freemangordonwhere did you get that from12:13
freemangordonas usually this is not how signing works12:13
Palithere was page about xloader structure12:13
Paliand there it was written12:13
freemangordonPali: xloader has RSA key in it12:14
freemangordonso I guess there is a certificate chain rather a single key12:14
Palitry to look for xloader strucutr12:15
Palilink should be in log12:15
*** rd_ has joined #maemo-ssu12:26
*** jonwil has joined #maemo-ssu12:40
*** rd_ has quit IRC12:54
*** rd_ has joined #maemo-ssu13:01
*** Vlad_on_the_road has quit IRC13:05
*** dafox has joined #maemo-ssu13:20
*** LauRoman has quit IRC13:20
*** Martix has quit IRC13:20
*** Martix has joined #maemo-ssu13:22
*** FReaper has joined #maemo-ssu13:27
*** FlameReaper has quit IRC13:29
*** Pali has quit IRC13:32
*** FlameReaper has joined #maemo-ssu13:54
*** FReaper has quit IRC13:56
*** Pali has joined #maemo-ssu14:05
*** psycho_oreos has quit IRC14:06
*** psycho_oreos has joined #maemo-ssu14:15
jonwil~seen pali14:22
infobotpali is currently on #maemo (17m 39s) #maemo-ssu (17m 39s), last said: 'or ask google for support'.14:22
jonwil~seen docscrutinizer0514:23
infobotdocscrutinizer05 is currently on #maemo (3d 14h 17m 59s) #n9 (3d 14h 17m 59s) #harmattan (3d 14h 17m 59s) #meego (3d 14h 17m 59s) #openmoko (3d 14h 17m 59s) #maemo-ssu (3d 14h 17m 59s) #infobot (3d 14h 17m 59s) #openmoko-cdevel (3d 14h 17m 59s). Has said a total of 613 messages. Is idling for 7h 57s, last said: 'seems he didn't like my advice'.14:23
Palifreemangordon: now I looked at gst omx plugin and it provide support for h264 and wmv14:25
freemangordonPali: so?14:25
Paliand dsp gst plugin does not support wmv14:25
freemangordonit does14:25
freemangordonnot stock one ofc14:25
Palihm, I looked at stock only14:26
freemangordonlook at the one on 720p thread14:26
freemangordonit provides everything OMX one provides14:26
Palistock omx: decoder-video/x-wmv, wmvversion=(int)3;14:26
Paliok14:26
Palifreemangordon: and where is hardcoded priority of omx and dsp?14:27
Paliin plugin? or in gst library?14:27
freemangordonin the plugin14:27
Palior somewhere else?14:27
Paliand cannot we increase priority of dsp?14:28
freemangordonwe can, but why?14:28
Palione package less which needs to be modified14:28
Pali(removed)14:28
freemangordonok, but omx just will stay in rootfs14:29
freemangordonand iirc it already have max priority14:29
Palisize is 40k14:29
freemangordonwe can just do "replaces" iiuc14:29
Paliok if we cannot increase priority because it has max, then create dummy transtitonal package14:30
freemangordonyep14:30
PaliI can do packaging when laptop is back14:30
Palibut what to do with hd codecs libs?14:31
Paliwe did not get any definite answer from ti/nokia14:31
freemangordonI first have to finish gstdsp to support both stock and hd codecs14:31
freemangordonif we're not going to replace codecs14:31
Paliand harmattan-dev will be down one day14:31
Paliso we cannot depends on that server14:32
freemangordon:nod:14:32
PaliI'm for pushing that libs to r.m.o14:32
Palinobody (we/cssu/nokia/ti) wrote any objections14:33
freemangordonPali: I am fine with that, but we should get a green light from merlin199114:33
Paliok14:33
Palimerlin1991: ping14:33
Palifreemangordon: so then pusing dsp and oc patches to cssu-kernel?14:34
freemangordonPali: do you remember why we decide to not push them?14:34
Palifreemangordon: can you give me link to cssu kernel wiki maemo page?14:35
Palioc because of powertop14:35
Paliand dsp because needs oc14:35
freemangordonhttp://wiki.maemo.org/CSSU_kernel_assess14:35
freemangordonoh, yeah, powertop14:35
freemangordonI guess this reason is no longer valid14:35
freemangordonFatPhil: ping14:35
* freemangordon needs to find something for lunch14:36
Palifreemangordon: powertop not working only if kernel is overclocked?14:36
freemangordonno14:36
Palior it not working also with default nokia settings?14:36
freemangordonit has number of freqs hardcoded14:36
freemangordonso having more then that breaks powertop14:36
freemangordons/freqs/OPPS14:37
Palifatphil: did you write letter about powertop license/source code to nokia?14:50
*** dafox has quit IRC14:55
Palipowertop sources could be usefull for fixing that bug in powertop14:57
*** Mihanizat0r has joined #maemo-ssu15:02
Palifreemangordon: and there is open source theora dsp library... if you have time can you try to create gst-dsp plugin for it?15:05
*** M13 has quit IRC15:06
*** dos1 has joined #maemo-ssu15:10
freemangordonPali: for theora we need newev gstreamer15:20
freemangordon*newer15:20
freemangordonand newer gstreamer needs newer libc15:20
freemangordonwhich needs newer kernel (i guess)15:20
*** rd_ has quit IRC15:21
Paliwhy we need new gstreamer?15:26
PaliI do not know how is gst-dsp working, but why cannot be added here also theora support which will use theora dsp codec?15:27
*** rd_ has joined #maemo-ssu15:47
* jonwil still cant find anything to work on...16:01
*** joshgillies has quit IRC16:04
*** lenoch has joined #maemo-ssu16:11
*** FlameReaper has quit IRC16:23
*** Pali has quit IRC16:38
*** Pali has joined #maemo-ssu16:48
Palijonwil: app-detect?16:49
jonwiltried it, I can figure out in broad terms what its doing but its doing something wierd so I cant do more about it16:52
jonwilas in clone it16:52
jonwilwhat do you want to know about it?16:54
jonwil?17:00
Paliwhat is that app-detect doing :-)17:02
Paliand something other to RE: systemui actdead plugin17:03
Palibecause it using libbme17:03
Palijonwil: that plugin is small17:03
jonwilsystemui actdead plugin isn't using libbme17:04
Palithen some other systemui plugin is linked with libbmeipc.so.017:04
PaliI thought it was actdead17:04
jonwilits using hald-addon-bme17:04
jonwilchecking some things now17:05
Palino no there is plugin which depends on libbmeipc0 package17:05
jonwilchecking now who actually links to libbmeipc.so.017:05
Paliok17:05
*** FlameReaper has joined #maemo-ssu17:09
Palithere is some problem with tablet-browser-daemon/eal/microb17:09
freemangordonPali: what is that problem thi https://gitorious.org/linux-n900/linux-n900/commit/1f98ccf8d0ceb3a7f1fe1c0695f7bfd8c39c9492 is fixing?17:09
freemangordon*this17:09
Palieverytime when I edit maemo wiki page it show me this message in syslog:17:09
Palirowser[2757]: GLIB WARNING ** default - Message sending from "ui" to socket "eal" failed, len:-1 error: Resource temporarily unavailable (11)17:09
Palisorry browser is freezed17:10
Palicannot open link now17:10
jonwilthings that use libbmeipc are17:13
jonwil /usr/bin/testserver17:13
jonwil /usr/lib/dsme/libthermalobject_surface.so17:14
jonwilvarious things in /usr/lib/testserver17:15
jonwilhald-addon-bme17:15
jonwiland the pulseaudio-nokia things17:15
jonwilthats all that I can find that uses libbmeipc17:16
Paliok17:17
Palijonwil: other thing which needs to be done is determinate why rtcom-call-ui crashing on 3.12 kernel17:19
Paliand create fix for that (kernel or userspace)17:19
jonwilI dont know the first thing about kernel and I know almost zilch about rtcom-call-ui17:20
Palijonwil: also what is needed: check who using /proc/component_version and /proc/bootreason17:21
jonwilThat I can find17:21
Paliand fix all these program to not fail if that files not exists17:21
jonwilchecking now for who is using that17:21
Palijonwil: please check programs from cssu-devel17:22
Palibecause I already fixed something17:22
jonwilI am just going to make list of all binaries on N900 that use those /proc entries for you17:22
jonwilfixing it is another matter :)17:22
jonwilok, files that use it:17:22
jonwil10hildon_welcome (script)17:23
jonwilrcS17:23
jonwilalso script17:24
jonwilsgx (script)17:24
jonwilsoftupd.rcS (script)17:24
jonwilfb-progress.sh17:24
jonwilalso script17:24
jonwilgetbootstate17:25
jonwilhald17:25
jonwilham-notifier-status-menu-item.so17:26
jonwilham-updates-status-menu-item.so17:26
jonwilusr/lib/apt/methods/http17:26
jonwilusr/lib/apt/methods/https17:26
jonwilmce/modules/libaccelerometer.so17:27
jonwilmce/modules/libalarm.so17:27
jonwilmce/modules/libaudiorouting.so17:27
jonwilmce/modules/libbattery.so17:27
jonwilmce/modules/libcallstate.so17:27
jonwilmce/modules/libcamera.so17:27
jonwilmce/modules/libdisplay.so17:28
jonwilmce/modules/libfilter-brightness-als.so17:28
jonwilmce/modules/libfilter-brightness-simple.so17:28
jonwilmce/modules/libhomekey.so17:28
jonwilmce/modules/libinactivity.so17:28
jonwilmce/modules/libkeypad.so17:28
jonwilmce/modules/libled.so17:28
jonwilusr/bin/testserver17:29
jonwilsbin/preinit17:29
jonwilsbin/mce17:29
jonwilmicrob-engine/libxul.so17:29
jonwilmce/modules/libvibrator.so17:29
jonwilusr/lib/libsysinfo.so17:29
jonwilusr/lib/libossoproductinfo.so17:29
jonwilso basically thats it thats using those 2 /proc entries17:31
*** Mihanizat0r has quit IRC17:31
jonwilonly things that arent foss or scripts are mce bits, libsysinfo and libossoproductinfo IIRC17:31
jonwilgetbootstate is now foss I believe17:32
jonwiloh and testserver isnt foss17:32
jonwildoes that help?17:32
Paliand what is using sysinfo?17:33
Palitestserver can be removed, it is not used17:33
jonwilsysinfo or ossoproducinfo?17:33
Palisysinfo (it is daemon + library)17:34
PaliI would like to know what we need to fix/replace/re to remove that sw17:34
jonwilossoproductinfo is used by camera-ui.launch, eapd, hildon-input-method-configurator.bin, /usr/lib/libconbtui/libaudio.so, libconnui.so, liblocation_applet.so, osso-connectivity-connui-conndlgs.launch, osso-product-info and sscd17:37
jonwilas for sysinfo, checking that now17:37
jonwilwl1251-cal (not my replacement though)17:38
Paliah that productinfo list is too big... bad...17:39
jonwilgen-obex-capability.sh17:41
jonwilbluetooth-sysinfo (not my replacement)17:41
jonwilK30ke-recv17:42
jonwilke-recv17:42
jonwilS30ke-recv17:43
jonwillibossoproductinfo.so17:44
jonwilsoftupd17:45
jonwilsscd17:45
freemangordonPali: I sent a mail for omapfb CMA failures, but don;t see it appearing in lkml :(17:46
freemangordonyou're in the cc17:46
jonwildsme-thermal (script)17:47
jonwilcontrolpanel.launch17:49
jonwilmediaplayer.launch17:49
jonwilosso_startup_wizzard.launch17:49
Palifreemangordon: lkml.org does not update website immediately, need wait17:49
freemangordonoh, ok :)17:50
jonwilfmtxd17:50
jonwilthats everything using sysinfo17:50
Palifreemangordon: now I updated cssu kernel table. Two things: 1. enable OC or not 2. patch ARM_6066_1_Fix_BUG_scheduling_while_atomic_swapper_0_0x00000002.diff17:50
freemangordonPali: well, the reason for not enabling OC was powertop, that problem no longer stands17:51
freemangordonI don;t know that patch17:51
freemangordonnut if it is enabled in KP, I see no reason why not enable it in kcssu17:51
freemangordon*but17:52
jonwilok, anything else I can check for while I am doing checks?17:52
Palifreemangordon: and 3. usb host mode is disabled, so in defconfig need to disable all usb drivers because are useless17:52
freemangordon:nod:17:52
Paliand that should be all17:53
freemangordongreat17:53
freemangordonut you have no PC to build it :D17:53
freemangordon*but17:53
Palifor kp53 I have two patches17:53
* freemangordon needs some cigarettes17:53
jonwilthose things will kill you :P17:53
freemangordonnaah, my liver will die faster :D17:54
jonwilor if not kill you, put you in hospital for massive surgeries :P17:54
Sicelo+117:54
Sicelo:P17:54
Palione from pavelm which adding optional better support for headset button17:54
Palibutton press is reported to system immediately (as new different keycode) but sometimes can be send also when nothing was pressed17:55
Paliand second patch for adding new usbids to some usb network driver17:55
Palisecond is of cource now irrelevant for kcssu17:56
Paliand I will test first (maybe with new omp patches) how it work...17:56
Paliand decide if that can go to kcssu or not17:56
Palibutton keypress is reported as KEY_PRG1 (I think this is not used anywhere) so no maemo application should be broken17:57
Palibecause for listening for that is needed HAL17:57
Paliand tell it to listen for key_prg117:58
Palijonwil: can you integrte your wl1251-cal and bt-cal to udev?17:59
jonwilI dont know how to program udev18:00
jonwilNever used it18:00
PaliI mean: creating udev rules which call them after wifi/bluetooth appear in system18:00
jonwilI dont know udev rules either18:00
Paliah :-(18:00
Palijonwil: other task: kernel driver can load nvs wl1251 file from /lib/firmware/... location18:01
Palican you create program which generate that nvs file?18:02
Paliyou have already program which sending that nvs file via netlink to kernel18:02
Pali(nvs file is generated/read from cal)18:02
freemangordonPali: what is the problem with cal?18:03
Palifreemangordon: kernel cannot read cal18:03
jonwilWe have libcal clone already18:04
Paliand patches which can add that support will be rejected18:04
jonwiloh ok18:04
jonwilyeah makes sense18:04
jonwilbtw the wl1251-cal and bluetooth-sysinfo clones I made put the startup script in /etc/event.d18:04
Paliproblem is that proper way is to send nvs data to kernel driver is via /lib/firmware file18:04
Paliand it is already implemented18:04
Palijonwil: but that your script will not work with upstream kernel18:05
freemangordonPali: is your idea to read nvs from cal and to store it in /lib/firmware before wifi driver is loaded?18:05
Paliyes18:05
freemangordonok18:05
freemangordonmakes sense18:05
Paliand do that only once (when file not exists)18:05
freemangordonsure18:05
freemangordonPali: what is store in nvs file? MAC address?18:06
jonwilcalibration data IIRC18:06
freemangordonok18:06
jonwilit gets changed slightly depending on regulatory domain18:06
Palinvs file has calibration data without MAC address18:06
freemangordonbut there is nvs in upstream fw iirc18:06
freemangordonor this is device dependent?18:06
Palipermanent mac address *cannot* be configueed for wl1251 now18:06
Palinvs is device dependent18:07
freemangordonwhaat?18:07
Palisome "general" is in fw tree18:07
Palifreemangordon: there are two mac address18:07
Pali1. permatent18:07
freemangordonoh, I see18:07
Pali2. configured by user18:07
jonwilWiFi (and bluetooth) MAC addresses come from CAL18:07
jonwilfeel free to adapt my wl1251-cal clone to spit out nvs data into a file (and use udev) if that's what you need18:07
Palikernel driver report to userspace both addresses18:08
jonwilmy dev box is having issues right now due to a botched emerge18:08
freemangordonok, so maemo configures amc18:08
jonwilso I cant write the program for you18:08
jonwilonly reverse engineer stuff18:08
Paliand wl1251 generating some random permanent18:08
freemangordon*MAC18:08
Paliand that reports to user18:08
freemangordonhmm, wait, I have MAC filtering on my router18:08
jonwilanything else you need me to do?18:09
jonwilor want me to look at?18:09
freemangordonso if MAC is random, my devel device won;t be able to connect18:09
Paliof course jonwil/nokia program set user configuable address correctly18:09
freemangordonok18:09
Paliand kernel driver using always that user configurable18:09
Palibut you (as computer user) can ask kernel if driver has its native permanent address or changed18:10
jonwildifference between my program and nokia program is that my program doesnt use sysinfo to get mac addresses from CAL and it uses value from CAL for country code rather than asking cellmo for country code18:10
PaliI will ask on lkml what to do with wl1251 driver - how to transfer permanent address to kernel18:10
PaliI'm for creating sysnode (as in bluetooth) for that18:11
Paliso userspace (root user) can send permanent address to kernel driver18:11
Paliwl1271 kernel driver has mac address at begining of nvs data18:11
Palibut because wl1251 nvs data are different and structure is already defined, we cannot add somewhere (at begining or at end) needed mac address18:12
jonwilso yeah feel free to use anything from my bluetooth-sysinfo or wl1251-cal replacements if you want :)18:13
Paliso I think sysnode is good idea18:13
Palijonwil: your bluetooth sysinfo program only read mac address and send it to kernel via sysfs?18:13
Palinothing more?18:13
jonwilyes thats all it does18:14
jonwilgrabs it from CAL and feeds to sysfs18:15
*** freemangordon has quit IRC18:15
*** freemangordon_ has joined #maemo-ssu18:15
Paliok, then this program should be called by udev when bluetooth device appear18:15
jonwilok, well like I said I know nothing about udev18:15
Paliok18:15
jonwilbtw is that crypto tarball from earlier (the one you said was posted before) any actual use to us?18:17
jonwili.e. does it acutally let us get at the hardware crypto bits we want to get at18:18
freemangordon_i'd say no18:18
jonwilor does the fact that it needs new xloader and new nolo make it useless?18:18
freemangordon_unless we find a way to enable crypto HW without having to replace NOLO18:19
freemangordon_yep18:19
freemangordon_maybe we ahould ask the author wht exactly needs to be done18:19
jonwilok, so someone needs to figure out the binary layout of xloader and nolo and how to load both into IDA and then someone needs to find out whats different between the old and new xloader/nolo18:19
*** freemangordon_ is now known as reemangordon18:19
*** reemangordon is now known as freemangordon18:20
jonwilI bet it does something that cant be done from kernel, only from lower level (xloader or nolo)18:20
freemangordonjonwil: I have those loaded in IDA18:20
freemangordonwith the correct offsets18:20
jonwilstock or the ones from that tarball?18:20
jonwilbut yeah talking to author of that tarball is best18:20
freemangordonjonwil: it is not about the level (low/high), xloader/part of secondary run in so-called secure mode18:21
jonwilyes thats what I thought18:21
jonwilsecure mode18:21
freemangordonjonwil: both stock and the ones from the tarrball18:21
freemangordonbut we should at least know which is the binary to look in18:21
jonwilsame as how on modern PCs there are things only the UEFI bios bits can do and that OS kernels cant do18:21
jonwilso yeah maybe we should talk to the author and find out where the magic sauce is18:23
jonwilwhich part (xloader or nolo or both)18:23
jonwilstuff that low (close to the hardware) is outta my league18:23
jonwilI like nice sane elf executables calling functions with nice names and prototypes (preferably functions that I have header files for :)18:24
jonwilheck, the closest I get to kernel stuff is when I need a kernel upgrade on my Gentoo box :)18:26
freemangordonjonwil: both18:27
jonwilok, anything else I can do?18:28
*** M4rtinK has joined #maemo-ssu18:29
freemangordonjonwil: no idea18:29
jonwilok18:29
freemangordonjonwil: oh, ...18:29
jonwiloh what?18:29
freemangordonyou may want to look at VKB portrait mode18:29
*** M4rtinK has quit IRC18:29
jonwilcant do that either, my dev box is down (as I said)18:30
freemangordonok18:30
jonwilso I can reverse engineer stuff but not write code right now :)18:30
*** M4rtinK has joined #maemo-ssu18:30
freemangordonI see18:30
jonwilnot that I know enough about GTK to do vkb portrait mode even if my dev box wasn't down18:30
freemangordonjonwil: it is not about GTK afaik18:30
freemangordonit needs testing and layouts18:30
freemangordonnicolai promised to do it, but it is obviously he has lost interest18:31
jonwilI have no interest whatsoever in portrait vkb, my interest is in system bits, bits that will benefit Neo900 or etc18:32
jonwilhence things like wl1251-cal18:32
jonwiland bluetooth-sysinfo18:32
Palijonwil: re hildon notify library18:32
freemangordonok, ok :)18:32
freemangordonre sscd :P18:32
jonwilhildon notify lib is which one?18:32
freemangordonhmm, yes18:32
Palithat hildon library has bugs18:32
freemangordonwait, I have IDA project, can give it to you18:33
jonwilwhich library is it?18:33
Paliannoying bugs18:33
freemangordonlibhildonnotify?18:33
freemangordonwait a minute18:33
Palifreemangordon: which library it is?18:33
Paliade also need to modify that library for clock18:33
freemangordonhildon-plugins-notify-sv18:34
freemangordonmaybe18:34
jonwilalready tried hildon-plugins-notify-sv and given up, its too hard18:34
Paliah :-(18:34
freemangordonjonwil: it is not *THAT* hard18:34
jonwilwell you do it then :P18:35
jonwilI have tried and failed18:35
freemangordonand you'll work on the kernel? :P18:35
jonwilnope :P18:35
freemangordon:)18:35
freemangordonVKB is more complicated than it18:35
jonwilas for sscd, I plan to reverse engineer the external interfaces to it (dbus calls etc) but not the binary itself18:36
freemangordonmakes sense18:36
jonwilexternal interfaces are needed since Neo900 will need some daemon that takes same dbus interfaces and matches them to Neo900 cellmo18:36
jonwilsame thing with cellular services daemon and bits18:36
Paliand that hildon library is needed RE because has bugs18:37
freemangordon:nod:18:37
Paliade need it for clock-ui18:37
freemangordonPali: I'll take a moth or so to RE it, I'd rather spend such a time on the kernel now18:38
jonwilif only I had a spare couple thousand dollars for latest hex-rays ARM, then certain things that are currently hard would become easier18:38
freemangordon*It'll18:38
jonwilbut I am unemployed with no spare money :P18:38
freemangordonjonwil: this lib has x86 binary18:38
jonwilyes I know it does18:38
Palifreemangordon: yes18:39
jonwilbut other things that dont have x86 binaries would become easier to RE18:39
jonwillike csd bits18:39
freemangordonjonwil: sure. BTW did I send you those 10 lines of code I REed?18:39
jonwilfrom what?18:39
freemangordonhildon-plugins-notify-sv18:39
jonwilyeah I think you did18:40
freemangordonthe whole lib is like that18:40
freemangordonjust GObject interfaces, which are easy to RE18:40
freemangordon(once yoiu grok them)18:40
jonwilI never found gobject easy to RE18:40
freemangordonbut it is, because you have all the structures defined18:41
jonwilbut thats probably because I know basically nothing about glib and programming it18:41
freemangordonyou just import the headers from SDK into IDA18:41
freemangordonand hexrays is more than happy18:41
jonwilyeah that I know18:41
freemangordonyou simply need to figure out the private structure (if any)18:42
freemangordonmost of the objects don;t have one, in hildon-plugins-notify-sv18:42
freemangordondammit, I can;t find the project :(18:43
jonwilarm hexrays would help me with so many things, csd bits, icd bits, connectivity UI bits, pulseaudio stuff and more :P18:43
freemangordonstill not leaked?18:43
jonwilnope, surprisingly it hasn't18:43
freemangordonwhat the hell is with the russians?!? :D18:44
jonwilprobably because the entities that actually need ARM HexRays AND can afford it are so few18:44
jonwilmuch less than need x86 hexrays18:44
jonwilplus I suspect IDA people are very careful who they sell ARM Hexrays to18:45
freemangordonhmm, on the other hand - there is not much I can do on the kernel side while waiting for Pali's laptop18:45
Palifreemangordon: maybe fixing snd rx51?18:46
Paliforward porting changes from kp...18:46
freemangordonPali: I want KP driver forward-ported and upstreamed18:46
freemangordonyeah18:46
freemangordonhmm, ok18:47
Palifreemangordon: or look why that rtcom-call-ui crashing18:47
Paliif you have working strace...18:47
Palior use ltrace/latrace18:47
freemangordonPali: no, audio is something I want to have working18:47
Palimaybe rtcom-call-ui crashing because of audio...18:48
jonwilre rtcom-call-ui crash, maybe getting a dump file of the crash and finding out exactly where its crashing might help?18:48
freemangordonPali: yep, I suspect the same18:48
freemangordonjonwil: it SEGFAULTs18:48
freemangordonhmm, or aborts, can;t remember18:48
Paliin dump will be nothing usefulll...18:48
freemangordonyep18:49
jonwilok18:49
Paliwe do not have debug symbols18:49
freemangordon^^^18:49
jonwilyeah good point18:49
Paliso to know that it crash at instruction mov r0 ... is useless18:49
jonwildo we know which library it crashes in? (or main rtcom-call-ui binary itself)? That might help narrow things down a bit18:50
Palimy idea is that it checking some sysfs file which not exists and then it crash18:50
freemangordonI think the same too18:51
Palijonwil: or oher task? document what is stored in cal mtd partition :-)18:51
Paliand now I do not have any other idea what to do....18:52
jonwildont know anything about CAL other than what wl1251-cal and bluetooth-sysinfo do :P18:52
jonwilanyhow, its late, zzz time :)18:52
jonwilcya later18:52
freemangordonjonwil: bye18:52
*** jonwil has quit IRC18:52
freemangordonPali: oh, we need nokia-av driver too20:08
*** rd_ has quit IRC20:09
*** FlameReaper has quit IRC20:10
merlin1991Pali: pong20:11
*** mnzaki has joined #maemo-ssu20:19
*** rd_ has joined #maemo-ssu20:37
*** NIN101 has quit IRC20:51
*** rd_ has quit IRC21:04
*** xes has joined #maemo-ssu21:06
*** rd_ has joined #maemo-ssu21:06
*** rd_ has quit IRC21:19
*** mnzaki has quit IRC21:21
Palimerlin1991: look at discussion about harmattan dsp ti codecs for cssu/rmo21:26
freemangordonPali: please think about how to integrate nokia-av driver21:29
freemangordonPali: lets have a discussion/brainstorm tomorrow21:29
Palifreemangordon: I thought that nokia-av was integrated into some other kernel driver21:30
Palimaybe snd?21:31
Paliso this was reason why I deleted nokia-av from elinux n900 page21:31
freemangordoncould you find it?21:32
Palibut maybe I did not look correctly21:32
PaliI do not remember21:32
freemangordonanyway, I don;t have time now, lets leave it for tomorrow21:32
*** NIN101 has joined #maemo-ssu21:32
Paliok21:33
*** rd_ has joined #maemo-ssu22:32
*** tom has quit IRC22:35
Palisuggection for camera-ui in cssu: add support for front camera23:11
*** nox- has joined #maemo-ssu23:19
merlin1991I'm all for that :D23:19
Palimerlin1991: so you agree to push harmattan dsp codecs (closed nokia/ti) to cssu/rmo?23:23
*** xes has quit IRC23:28
merlin1991what's the (c) on those23:32
*** xes has joined #maemo-ssu23:32
*** _rd_ has joined #maemo-ssu23:32
*** rd_ has quit IRC23:33
Palimerlin1991: nokia or ti - do not know now23:35
*** lenoch has quit IRC23:44
*** LauRoman has joined #maemo-ssu23:46

Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!