IRC log of #maemo-ssu for Thursday, 2012-07-19

*** Effilry has quit IRC00:02
*** Effilry has joined #maemo-ssu00:02
*** Effilry is now known as FireFly00:02
lufWhat movie? Dexter is still running :D00:11
*** mase76 has joined #maemo-ssu00:11
merlin1991luf: .mkv file on my pc ;)00:12
merlin1991and dexter is a series not a movie ;)00:12
lufWho cares? :D00:12
lufShit how to easily create some async callback?00:20
lufpbap in obexd is piece of ...00:23
lufIt doesn't count that backend should be synchronous :(00:24
DocScrutinizer06GSM signal meter fucked up00:25
DocScrutinizer06shows 2.5 all the time00:25
DocScrutinizer06either it got stuck at 2.5, OR there actually is constant data traffic via GSM now, despite the network connection via WLAN is active00:26
*** DocScrutinizer06 is now known as DocScrutinizer0500:30
*** mase76 has quit IRC00:36
*** Guest92502 has quit IRC00:36
*** _rd has quit IRC00:36
lufOk. I patched pbap.c to support async backends. I hope I don't broke something else :(00:39
luf:D I get list of incoming, outgoing and missed calls :)00:42
lufThe async versus sync took me 2 hours of my life.00:43
DocScrutinizer05BUG: CSSU-T4.1 showing GSM:2.5 even when no EDGE00:51
DocScrutinizer05and UNplugging fastcharger pops up "Ass Rage Mode?" "PC suite mode?"00:54
DocScrutinizer05time for a second boot I'd say00:54
*** jon_y has joined #maemo-ssu00:57
DocScrutinizer05hmm, I'm starting to feel unsure about whether this 2.5G thing was working "correctly" in previous CSSU00:57
Woody146192.5G Is edge...00:58
DocScrutinizer05anyway now it's stuck at 2.5, while I'm quite sure there's been a time where it switched back to 2G when no data transfer (quite a natural think as you can't tell if GSM is EDGE or mere 2G unless you actually transmit data - at least that's what I seem to know)00:58
DocScrutinizer05Woody14619: yes, so what?00:59
Woody14619The only time I've had it drop down below 2.5 was when I was in an area that didn't support it.  (Or was on the edge of service.)00:59
DocScrutinizer05that seems odd01:00
DocScrutinizer05it's been different in former times01:00
jon_yhey guys, what is the preferred way to cross compile n900 kernels/apps, the nokia ubuntu VM?01:00
Woody14619I'm confused... Are you saying it's staying in 2.5, or it's dropping to 2.0 when idle?  I've always had it do the former, not the later.01:00
DocScrutinizer05now it's stuck at 2.501:01
DocScrutinizer05the nokia flavour01:01
Woody14619I'm on CSSU-T4.0 right now (haven't updated to 4.1)01:01
merlin1991just do it, only good stuff comes with it :D01:01
Woody14619And get the same behavior... Stays on 2.5, even when not connected.01:01
DocScrutinizer05Woody14619: you can find a lot of "why is this thing switching to 2G?" probably even tickets in bugzilla01:02
Woody14619merlin1991: I will, just haven't had the time to scratch my ass, yet alone deal with an update. ;)01:02
DocScrutinizer05update went flawlessly (haven't checked my SIP yet ;-P)01:03
Woody14619May be a regional issue?  FWIW: I was running on AT&T in the US for over a year, with no option of 3G, so...01:03
DocScrutinizer05SIP ok01:03
jon_yDocScrutinizer05: does scratchbox2 require chroot?01:03
Woody14619It may be it does something different in 3G mode?  I don't tick that on often unless I need it (I'm now on T-MO)01:03
DocScrutinizer05it IS a chroot01:04
jon_yah ok, means no cygwin for me01:04
DocScrutinizer05a mega monster chroot01:04
jon_ywell, looks like I'll be setting up Debian :)01:04
* jon_y is unfortunate to own a n900 with HW revision 220401:05
jon_yuboot seldom work on it01:05
jon_yit just do a reboot loop01:05
DocScrutinizer05or rather "I never heard that before"01:05
jon_yPali's uboot01:05
jon_ythe new 2012.4-201:06
jon_ythe Nokia PR1.3 uboot works though01:06
jon_yloading a kernel with fbcon says it couldn't mount ubi0:root or some such01:06
DocScrutinizer05there's no hw diff between any board revisions that could explain / cause that01:06
jon_ythis scratchbox?
merlin1991jon_y: no01:07
jon_yyeah, I read Pali's thread about requesting atags01:07
merlin1991n900 dev happens on good old sb101:07
jon_ymerlin1991: hmm ok01:08
jon_ymerlin1991: this one?
* merlin1991 doesn't find the bloody wiki entry01:08
jon_yDocScrutinizer05: the thread ->
jon_ywell, I read all 38 pages :|01:09
DocScrutinizer05jon_y: well, what can I say?01:09
jon_ymerlin1991: thanks01:10
jon_yDocScrutinizer05: hopefully pali is still around to help on 220401:10
DocScrutinizer05there's no reason why uBoot *must* fail on any particular hw rev. atags might differ as they describe the storage and that might have changed slightly in abstract descriotion, NOT in function01:10
jon_yanything I can do to debug it?01:11
merlin1991DocScrutinizer05: afaik palis uboot strips the amount of atags, maybe something vital is going missing there?01:11
DocScrutinizer05quite possible01:11
DocScrutinizer05some storage might get reported as one chunk, while another flash chip might report two chunks of half the size01:12
DocScrutinizer05same for RAM01:12
DocScrutinizer05jon_y: ask pali when he's around01:13
DocScrutinizer05~seen pali01:13
infobotpali <~pali@unaffiliated/pali> was last seen on IRC in channel #maemo-ssu, 1d 2h 10m 5s ago, saying: 'only'.01:14
DocScrutinizer05~seen pali_01:14
infobotpali_ <> was last seen on IRC in channel #maemo, 321d 22h 37m 50s ago, saying: 'usbtty - cdc_acm working fine, but when I booted maemo kernel, dsme start rebooting and I got always kernel panic (when device entered reboot state)'.01:14
jon_yhopefully I can poke about uboot around next week01:14
jon_yhmm, speaking of reboot panics01:14
jon_ythe nokia pr1.3 uboot also gets stuck if I power on too quickly after reboot01:15
*** Woody14619a has joined #maemo-ssu01:40
Woody14619ajon_y: had the same issue with one of my two N900s. (though I don't know the hardware rev of my devices)01:41
Woody14619aeven had them identical software wise, kernel flash + backup-menu restore... One would boot, the other loop twice and shutdown.01:42
*** Woody14619 has quit IRC01:43
*** Woody14619a has quit IRC01:43
*** Woody14619a has joined #maemo-ssu01:43
*** Woody14619a is now known as Woody1461901:44
*** Woody14619 has quit IRC01:46
*** Woody14619 has joined #maemo-ssu01:51
*** Woody14619 has joined #maemo-ssu01:51
*** nox- has joined #maemo-ssu02:11
*** MohammadAG has quit IRC02:26
*** MohammadAG has joined #maemo-ssu02:26
*** Woody14619 has quit IRC02:29
*** LaoLang_cool has joined #maemo-ssu03:31
lufIf someone want to test first dev preview of PBAP/IrMC with call history (only pullall is working - test pbap-client) so he can. You can find it here:
lufTime to go to bed for me ...03:54
*** luf has quit IRC03:54
*** arcean_ has quit IRC04:05
*** M4rtinK has quit IRC04:07
*** nox- has quit IRC05:05
*** jd has joined #maemo-ssu05:08
*** Jade has quit IRC05:10
*** amiconn_ has joined #maemo-ssu05:52
*** amiconn has quit IRC05:52
*** amiconn_ is now known as amiconn05:52
*** taziff1 has joined #maemo-ssu06:21
*** taziff has quit IRC06:22
*** LaoLang_coo_ has joined #maemo-ssu08:48
*** LaoLang_cool has quit IRC08:51
*** Pali has joined #maemo-ssu09:13
*** guly has joined #maemo-ssu09:16
*** freemangordon has joined #maemo-ssu09:20
kerioDocScrutinizer05: (re: 2G vs 2.5) i've *always* seen 2.5, even with no data transfer, in edge-enabled zones, and 2G otherwise, in every phone i've used09:24
DocScrutinizer05maybe I ignored that change since PR1.209:25
keriomy old 6120c did that too09:26, what a fucking awesome phone09:26
DocScrutinizer05a pity since I found it quite helpful to have an idea when data gets transfered09:26
kerioumts/hsdpa does that09:26
kerio3.5 only when stuff is being transferred09:26
DocScrutinizer05or that, I'm getting alzheimer09:27
kerio*pat pat*09:27
Palihi, I have in my DEB_BUILD_OPTIONS thumb string in default Maemo 5 SDK image09:34
Paliwho set it?09:35
PaliI cannot find application/script in scratchbox which set thumb...09:35
DocScrutinizer05anyway my cbsms operator widget works quite well - doesn't replace "o2 -de" but simply extends it by "35631134" for chan 221 Gauss Wegener coords09:35
keriofreemangordon hacked the nokia servers to enable thumb by default in the sdk image09:35
kerioso eventually more people will release thumb binaries09:35
Paliso this was reason why my HAM still crashed... (on non thumb kernel)...09:36
Paliand I never installed harm into scratchbox09:36
Paliscratchbox is running in ubuntu maemo 5 sdk image09:36
Paliand this image is running in qemu09:37
Paliqemu has disabled network, so scratchbox cannot connect to nokia server :-)09:37
DocScrutinizer05strange, and scary09:38
kerioPali: no, i mean09:39
keriothe sdk image you downloaded was changed by fmg09:39
kerioor something09:39
Palikeriom no, I downloaded it years ago :-)09:39
keriofmg is a clever planner09:39
PaliI found it now: /targets/FREMANTLE_ARMEL.environment09:41
DocScrutinizer05Pali: so what makes that differ on your (and MohammadAG's?) installation to what most other people have?09:46
DocScrutinizer05for mag we assumed he fecked it by installing harm SB09:46
PaliDocScrutinizer05, I downloaded hdd image of maemo 5 sdk and started in under qemu09:46
Paliand now I see that this image is thumb enabled09:47
DocScrutinizer05aaaah, while usually you install it09:47
DocScrutinizer05so some funny guy uploaded a thumb-enabled image09:48
DocScrutinizer05please give me full path so I can check here09:48
Palican somebody else check if lucid desktop sdk image from really contains thumb string in DEB_BUILD_OPTIONS?09:49
DocScrutinizer05Pali: I meant my local SB09:50
DocScrutinizer05where do I find the incriminated file?09:50
DocScrutinizer05in /sacratchbox/09:51
Palifull path is /targets/FREMANTLE_ARMEL.environment not /scratchbox/targets...09:51
DocScrutinizer05inside SB I suppose09:51
DocScrutinizer05I've not started SB09:52
Palilogin to SB09:52
DocScrutinizer05wouldn't even know how09:52
DocScrutinizer05I prefer to give locate a shot09:52
DocScrutinizer05jr@halley:~> locate /targets/FREMANTLE_ARMEL.environment09:53
DocScrutinizer05jr@halley:~> cat /scratchbox/users/jr/targets/FREMANTLE_ARMEL.environment09:54
DocScrutinizer05export DEB_BUILD_OPTIONS=maemo-launcher,thumb,vfp09:54
Palithumb enabled too :D09:54
Palithis is real nokia problem!09:55
Paliwhy nobody found this problem earlier?09:55
DocScrutinizer05pali king for this week!09:55
DocScrutinizer05I'm pretty amazed we got any working hostmode kernel09:57
PaliDocScrutinizer, problem is only with apps which check that string in debian/rules09:57
DocScrutinizer05I for one never built any such app09:57
PaliDEB_BUILS_OPTIONS is some hint for debian apps: you can enable this options...09:58
DocScrutinizer05I only built kernel, and brainless *#-ena09:58
DocScrutinizer05anyway, that's clearly a topic for council09:58
DocScrutinizer05to get that brainfsck fixed09:59
DocScrutinizer05and MAKE IT PUBLIC09:59
Palibut must pass correct flags to gcc10:01
Pali./configure --enable-maemo-launcher10:01
Palior: ./configure CFLAGS=-mthumb10:01
Palidebian/rules must adjust these flags, debhelper is not doing it automatically10:01
Paliwrite email to community mailinglist10:01
kerioDocScrutinizer05: we could fix it by pushing an updated kernel with the thumb fix! :D10:01
Palisome maemo devs which using scratchbox read mailinglist10:01
DocScrutinizer05meh, don't pester me with that stuff10:02
DocScrutinizer05kerio: ^^^10:02
keriok :c10:02
DocScrutinizer05Pali: I would prefer *you* write that mail, since a) you found it so the credit is yours, and b) you know about it, while I got NFC10:04
PaliDocScrutinizer05, ok10:04
DocScrutinizer05add a line...10:04
DocScrutinizer05sed -i "s/export DEB_BUILD_OPTIONS=maemo-launcher,thumb,vfp/export DEB_BUILD_OPTIONS=maemo-launcher,vfp" /scratchbox/users/jr/targets/FREMANTLE_ARMEL.environment10:05
DocScrutinizer05sed -i "s/export DEB_BUILD_OPTIONS=maemo-launcher,thumb,vfp/export DEB_BUILD_OPTIONS=maemo-launcher,vfp/" /scratchbox/users/jr/targets/FREMANTLE_ARMEL.environment10:06
DocScrutinizer05sed -i "s/export DEB_BUILD_OPTIONS=maemo-launcher,thumb,vfp/export DEB_BUILD_OPTIONS=maemo-launcher,vfp/" $(find /scatchbox -name FREMANTLE_ARMEL.environment)10:07
DocScrutinizer05sed -i "s/,thumb,/,/" $(find /scatchbox -name FREMANTLE_ARMEL.environment)10:09
DocScrutinizer05sed -i "s/thumb//;s/,,/,/" $(find /scatchbox -name FREMANTLE_ARMEL.environment)10:12
DocScrutinizer05isn't it funny how many iterations it takes my sleepy brain to get such a friggin simple regex right? :-S10:15
*** andre__ has joined #maemo-ssu11:16
*** andre__ has joined #maemo-ssu11:16
*** jd is now known as Jade11:26
*** andre__ has quit IRC11:28
*** andre__ has joined #maemo-ssu11:28
*** andre__ has joined #maemo-ssu11:28
*** Jade has quit IRC11:43
*** Jade has joined #maemo-ssu11:43
*** Jade has quit IRC11:43
*** Jade has joined #maemo-ssu11:43
jon_yPali: hey, are you still needing test subjects for HW revsion 2204?12:14
Palijon_y, do you mean atags?12:15
jon_yu-boot 2012.4-2 still fails to boot on it12:15
jon_ywell, anything I can do to help12:15
jon_yfb kernel says it can't mount ubi0:rootfs12:15
Palijon_y, I only need atags if are different from that which was posted12:15
jon_yiirc they were the same, but I can recheck later12:16
Palijon_y, if atags are same, then it is not problem in atags12:16
jon_yhmm ok, any other pointers?12:16
Palijon_y, ok recheck if are same as reported in that thread12:16
DocScrutinizer05number of bad blocks too high is a valid reason for uboot not fitting into partition12:17
DocScrutinizer05unrelated to hw rev though12:17
Paliif stock kernel can mount rootfs and stock kernel started from u-boot cannot then problem should be in partition offset12:17
jon_ythe flasher skips bad blocks automatically?12:17
PaliDocScrutinizer05, flasher refuse to flash code if is bigger then part size12:18
DocScrutinizer05ahh ok12:18
DocScrutinizer05jon_y: it should12:18
Palibut this can be problem, that if more blocks are bad then it try to continue flashing on next good block12:19
Palimaybe bug in flasher12:19
PaliI think nokia did not tested what happen on too menu bad blocks...12:19
DocScrutinizer05probably not12:19
jon_yatag dump
*** arcean has joined #maemo-ssu12:21
Paliwtf: dpkg --compare-versions 3.90.7-11.1-thumb3 '>' 3.90.7-11.2 return zero12:22
Paliwhy is 3.90.7-11.1-thumb3 high version as 3.90.7-11.2?12:22
PaliDocScrutinizer05, ^^12:22
DocScrutinizer05test "<"12:23
DocScrutinizer05and "="?12:23
Palino I pushed to merlin repo new version of modest (3.90.7-11.2) but apt-get installed old (3.90.7-11.1-thumb3)12:25
Paliand when I forced apt to install 11.2 it told me that it will downgrade modest12:25
jon_yPali: they're exactly the same12:25
jon_ywell, diff'ed with another 2204 post12:25
Palijon_y, then it is not problems with atags12:25
Palionly difference in some dumps is that some provides two MEM tags and some only one12:26
jon_ybtw, PR1.3 stock uboot works, somehow12:26
DocScrutinizer05what I suspected12:26
jon_yiirc it was the same uboot as KP-4612:26
Palibut that which provides only one is connected first and second12:26
Paliand if mem tags are connected then kernel it parse as one12:27
jon_yDocScrutinizer05: ?12:27
DocScrutinizer05jon_y: MEM tags12:27
Palipr1.3 uboot is 2 years old uboot version12:27
Paliarm code was more times cleaned and some rewritten12:28
jon_yyeah, it was lacking all the cool new features, like a proper menu12:28
Paliso problem can be in uboot code not in board nokia rx51 code12:28
jon_ywhich is why I clean flashed my phone recently :)12:28
jon_yDocScrutinizer05: what do mem tags do?12:29
jon_yI'm assuming atags store the cmdline12:29
* jon_y n00b12:29
Paliatags store more then cmdline12:29
Palisee your dump :-)12:29
jon_yok, details about the nand layout?12:30
Paliif cmdline atag is not specfied , default hardcoded into kernel zImage is used12:30
DocScrutinizer05jon_y: yep12:30
*** Guest92502 has joined #maemo-ssu12:31
*** _rd has joined #maemo-ssu12:31
jon_yok, so, what does the memtag do?12:31
DocScrutinizer05cat /proc/mtd12:31
jon_yah, partitions, sort of12:31
*** M4rtinK has joined #maemo-ssu12:31
jon_yPali: do you also use scratchbox to build your kernels?12:32
Palimemtag defines RAM12:32
Palijon_y, no12:32
Palibut kernel-power can be build in scratchbox too12:32
jon_yok, I have been itching to stick fbcon in it since I first used it :)12:33
jon_yanything I can do to help with the mem tags?12:34
DocScrutinizer05atag mem  is about ram, partition is about... you guessed it ;-)12:35
DocScrutinizer05some ram reports as two banks, some as one12:36
DocScrutinizer05so atags *might* have two atag mem12:37
jon_y2204 /proc/mtd ->
jon_yanything different about it?12:37
jon_yso, not a mem tag issue?12:38
DocScrutinizer05proc/mtd is about partitions12:38
DocScrutinizer05       0396 - 001c:4f0b PARTITION:    bootloader       : size=00020000 offset=00000000 mask=00000003  ff  in your paste12:39
jon_yah ok, I got it backwards then12:39
DocScrutinizer05should always fit, as otherwise nothing works12:40
jon_yok, what should I do next to debug?12:40
DocScrutinizer050020 - 0004:54410002 ATAG MEM size=10000000 start=8000000012:40
DocScrutinizer05is what we are discussing12:41
DocScrutinizer05good question12:41
Pali0032 - 0004:54410002 ATAG MEM size=08000000 start=8000000012:41
Pali0048 - 0004:54410002 ATAG MEM size=08000000 start=8800000012:41
Palithis is passed by u-boot12:41
jon_yI have: 0020 - 0004:54410002 ATAG MEM size=10000000 start=8000000012:42
jon_yno 8800000012:42
Palisize=08000000 start=80000000 and size=08000000 start=88000000 are connected12:42
DocScrutinizer05two halfes12:42
Paliso it is same as start=80000000 size=08000000+0800000012:43
jon_yyeah, same size, but I don't have it split12:43
jon_yit wouldn't matter?12:43
Palikernel convert it to one memory definition if are connected12:43
Paliso it is same12:43
DocScrutinizer05but uboot?12:43
Palimem tag is no problem12:44
PaliDocScrutinizer05, uboot alocating memory in first or second mem12:44
Palibut if are connected then there is no problem12:45
Paliyou can use full memory12:45
DocScrutinizer05well, jon_y's mem tag is one anyway12:45
DocScrutinizer05so hardly the root cause of uboot problem12:46
PaliI suggest you to flash u-boot and load fia flasher-3.5 NIN rescue OS12:46
Palirescue os can mount rootfs12:46
Paliand kernel is based on 2.6.37+12:46
Paliso maybe there is kernel bug in onenand/ubi code in 2.6.28...12:47
Palifirst need to know if corrupted (after flashing uboot) is rootfs or not!12:47
jon_yit won't wipe my phone data?12:48
jon_yhmm, google won't tell me where to get NIN rescue OS12:48
DocScrutinizer05~listkeys rescue12:49
infobotFactoid search of 'rescue' by key (7): htcrescue ;; cmd: rescue (.*?) ;; superrescue ;; rescue disk ;; rescuecd ;; rescue disks ;; reiserfs-rescue-disk.12:49
jon_y ?12:49
jon_yok, its loaded not flashed, cool12:49
DocScrutinizer05~rescue-os is
infobotokay, DocScrutinizer0512:50
Paliadd also this info file:
jon_yok, will try rescue OS in about half an hour's time, I need to eat12:51
jon_yhmm download gives 50312:52
jon_y*http 50312:52
DocScrutinizer05~rescue-os is also
infobotokay, DocScrutinizer0512:53
jon_yDocScrutinizer05: looks like the download moved to
jon_ystatic ip :(12:53
jon_ythe .cx site leads to a dead 503 server12:54
DocScrutinizer05 load just fine here12:54
jon_yhmm ok, probably my end then12:55
jon_y0.5.3 right?12:55
DocScrutinizer05as does
jon_yI get redirected to with the cx link12:56
jon_y503s there12:56
Paliyes, NIN posted static ip address here on irc more time12:56
jon_ywell, bbl for real :)12:56
DocScrutinizer05~rescue-os is also
infobotokay, DocScrutinizer0513:00
infobotextra, extra, read all about it, rescue-os is, or, or
* DocScrutinizer05 ponders to mirror the whole bunch on
*** Guest92502 has quit IRC13:15
*** _rd has quit IRC13:15
*** LaoLang_coo_ has quit IRC13:18
*** _rd has joined #maemo-ssu13:18
*** Guest92502 has joined #maemo-ssu13:18
*** LaoLang_coo_ has joined #maemo-ssu13:18
DocScrutinizer05meh, I'm too dizy to do a decent wget :-S13:20
*** Pali has quit IRC13:23
*** _rd has quit IRC13:27
*** Guest92502 has quit IRC13:27
*** LaoLang_coo_ has quit IRC13:29
keriohm, the wiki page is not clear13:46
keriodo you *need* the tarball on the µsd or can you go with initrd from usb only?13:46
*** jonwil has joined #maemo-ssu13:55
*** jonwil_ has joined #maemo-ssu13:57
jon_ykerio: one or the other iirc14:00
*** jonwil has quit IRC14:00
*** jonwil_ is now known as jonwil14:00
jon_ydang pali left14:05
jon_yDocScrutinizer05: ping14:05
jon_yhow do I check the FS?14:05
jon_yDocScrutinizer05: looks like nand ubi0:rootfs mounted correctly14:17
jon_yI assume it means it was a bug in 2.6.28?14:18
*** Pali has joined #maemo-ssu14:28
jon_yPali: hey14:32
jon_yafter flashing uboot, NIN rescue can still mount ubi0:rootfs14:32
Palijon_y ok, then rootfs is not corrupted14:33
Palican you try to boot stock zImage via flasher?14:34
jon_yflash or load?14:34
jon_yalready tried load with fbcon, it got further, but rebooted anyway14:34
jon_ythe screen was dim, so I couldn't read the output14:35
jon_yPali: any more ideas?14:38
Palijust load, do not flash14:38
jon_yyeah, it still fails14:38
Palialso try to load zImage with framebuffer14:39
jon_yok, will try it back to back14:39
Paliand write full framebuffer output (picture can be good too)14:39
jon_ydim screens are impossible to read14:39
jon_ythe backlight was off14:39
jon_yPali: ok, it booted up, somehow14:44
jon_yflashed with uboot, loaded with fbcon kernel14:44
Paliopen xterm14:44
jon_yit is not a reliable way to start, fails nearly all the time14:45
Palirun: dmeg > MyDocs/dmesg14:45
Paliand upload dmesg to pastebin14:45
jon_ywill do it over usbnet ssh then14:45
jon_yusbnet somehow failed to connect on windows14:50
Palijon_y, thanks14:50
PaliDocScrutinizer05, can you look at this dmesg output too if there is something bad? ^^^^14:51
Pali[  131.246978] UBI: number of bad PEBs:         114:52
Pali[  131.247222] UBI: max/mean erase counter: 139/4914:52
jon_yis that bad? :(14:52
PaliDo not know, but nothing abnormal is in log14:53
PaliI do not know how bad is bad PEB14:53
Paliask Doc about it14:53
jon_yok, hopefully it will not stop me from using uboot14:53
Palijon_y, can you try if stock zImage work when you load it from flasher (without flashing?)14:54
jon_yPali: yes, this is without flashing14:58
jon_yit loaded and started miraculously14:58
*** lizardo has joined #maemo-ssu14:58
jon_ymaybe it was because I disconnected the usb cable after flasher load was done14:58
jon_yPali: those "isp_mod: Unknown symbol iommu_restore_ctx"...14:59
Paliignore isp_mod messages15:00
PaliI have them too15:00
jon_yare they because /lib/module/current is pointing to the pk modules?15:00
jon_yah ok15:00
Palimaybe yes15:00
jon_ywell, its supposed to be the stock nokia kernel15:00
jon_y*kp modules15:00
Palinow you have booted framebuffer compiled kernel? yes?15:01
Palitry now to boot stock kernel15:01
jon_ywithout flashing?15:01
merlin1991ffs every single application on harm I looked at is closed source15:03
* merlin1991 wants to hack not look at binaries :/15:03
Palimerlin1991, I pushed all my updates to cssu-devel15:05
Palinew options in XTerm are in Fonts settings dialog15:06
merlin1991okay, I'll try to see what happens if I add -devel now to a clean device (to see if the new ham works :D)15:07
jon_yPali: nope, not working15:07
jon_yPali: fb kernel did say error 19 or some such when mounting ubi0:rootfs before quickly powering off15:08
jon_yand also uknown block(0,0)15:08
Palijon_y, can you try to use camera (or other n900 :-)) to take photo of that error?15:09
jon_yok, will try15:10
merlin1991hm if we ever start a harm cssu we'll have to RE a bunch of ui things since EVERY SINGLE STOCK APPLICATION IS CLOSED15:11
* merlin1991 punches a wall15:11
merlin1991even the friggin home screen is a closed package15:11
merlin1991interestingly the settings ui is opensource though15:13
Palimerlin1991, forget about harm cssu15:17
Palibetter fixing maemo15:17
Palilist of maemo closed apps are not too big:
jonwilI do wonder which maemo package would be best to replace next15:18
DocScrutinizer05Pali: [  131.246978] UBI: number of bad PEBs:         1  -- looks ok'ish, one bad block aiui15:18
PaliDocScrutinizer05, ok15:19
Palijonwil, bme of cource15:19
DocScrutinizer0510, or 50 probably big problem. One though, normal15:19
Palibme is in yellow color on that page :-)15:20
Palijonwil, but something like keyboard plugins (that input method) will be usefull15:20
jon_yPali: are you OK with dcc send?15:21
jon_yits 2MB15:21
jon_ywell, bad camera is bad15:21
Palijon_y, if I can read it it is ok15:21
jonwildid you end up emailing the Nokia browser UI guy like you said the other day?15:21
Palijonwil, forgot about it15:22
PaliI will write now15:22
jon_yPali: sent, waiting for accept15:23
jon_yPali: got it yet?15:24
*** LaoLang_cool has joined #maemo-ssu15:25
jon_yPali: second send attempt15:25
Palijon_y, I did not got anything15:26
Paliupload it to some picture server15:26
jon_yPali: I can transcribe it, crappy upload speed stalled it15:30
PaliDocScrutinizer05, look ^^15:36
PaliUBIFS error (pid 1): ubifs_get_sb: cannot open "ubi0:rootfs", error -1915:36
jon_yin the successful load, it had "UBIFS: recovery needed" and "UBIFS: recovery completed" instead15:37
DocScrutinizer05find "UBIFS error (pid 1): ubifs_get_sb: cannot open "ubi0:rootfs", error -19"  in kernel UBI driver15:37
jon_yI don't have the kernel source :(15:38
DocScrutinizer05in some cases block0 is special and needs to be OK15:38
Palioneline source code ^^^15:39
DocScrutinizer05grep string: cannot open15:39
DocScrutinizer05rather "cannot open \"%s\", error %d" or sth like that15:40
*** ThreeM has quit IRC15:41
jon_yabout line 200415:41
jon_yDocScrutinizer05: anything?15:44
DocScrutinizer05Pali needs to check, I got no time now15:45
DocScrutinizer05find out what makes it return -1915:45
jon_y if (name[0] != 'u' || name[1] != 'b' || name[2] != 'i')15:46
jon_ythat check failed15:46
jon_yor it does not match /* ubiX:NAME method */ somehow15:47
DocScrutinizer05I guess it's  return ERR_PTR(-EINVAL); though15:47
DocScrutinizer05which doesn't tell us much15:47
jon_ychecking EINVAL value15:47
jon_y#define EINVAL 2215:48
DocScrutinizer05yeah, hmmm15:48
Pali-19 is what error?15:48
jon_yprobably ubi_open_volume15:48
Paliin errno.h15:48
Pali#defineENODEV19/* No such device */15:49
jon_ymisreading the ubifs superblock?15:50
jon_y%-ENODEV if there is no such UBI device :|15:50
jon_ylooks like it goes to ubi_get_device failed?15:54
Paliseems that this should not fail
Paliso problem is here:
*** mase76 has joined #maemo-ssu15:56
Paliif ubi_open_volume fail (no matter what is error code) it always return -enodev15:57
DocScrutinizer0595  * If a static volume is being opened for the first time since boot, it will be15:57
DocScrutinizer0596  * checked by this function, which means it will be fully read and the CRC15:57
DocScrutinizer0597  * checksum of each logical eraseblock will be checked.15:57
jon_yhmm, so how is it working sometime only?15:58
DocScrutinizer05wacky bits in NAND?16:03
* DocScrutinizer05 frowns16:03
DocScrutinizer05not a good story16:04
DocScrutinizer05what's been the question? why new uboot doesn't work, while "stock kernel" uboot does?16:05
Palino, why stock kernel is not working if nwe u-boot is flashed?16:06
Paliand working again when new u-boot is replaced?16:06
jon_yalso, Pali's power kernel works, if loaded without flashing uboot16:07
jon_yjust tested, loaded and working16:08
DocScrutinizer05basically it's kernel failing to ubi-mount mtd3 in one case, while works in another kernel16:08
jon_yDocScrutinizer05: interesting, when loaded with power50, there is no bad PEB message16:09
DocScrutinizer05now that's a strong indication towards root cause16:10
jon_ycare to clue me in? :)16:10
* DocScrutinizer05 suspects a one-off probelm16:10
jon_yoff by one array/ptr?16:11
DocScrutinizer05sth like that16:11
DocScrutinizer05possibly by transfering the atag info16:11
jon_yuboot issue?16:11
Palikerne-power has patched this file kernel-power-2.6.28/fs/ubifs/file.c16:12
DocScrutinizer05pali should have a decent starting point now16:12
Palipatch is *official* linus' version from 2.6.28 to
jon_ystrage, how are the other HW revision working?16:12
Palino other ubi patches are in kernel-power16:13
DocScrutinizer05jon_y: that's hearsay16:13
DocScrutinizer05no verified relation to hw rev16:13
jon_yhmm ok, there should have been reports on uboot-2012.4-2 failing all over16:14
DocScrutinizer05I suspect something like that patch under certain conditions doesn't handle bad PEB correctly16:15
DocScrutinizer05or the flasher doesn't handle it correctly16:15
DocScrutinizer05and it never shown up without the patch16:16
DocScrutinizer05or whatever16:16
jon_ythe stock kernel is 2.6.28 baseline?16:17
Sc0rpiusI read that mail about having "thumb" in the FREMANTLE_ARMEL.environment file16:17
Sc0rpiusit seems I have always had it16:17
Sc0rpiusand NONE of my programs have ever crashed on me...16:18
Sc0rpius(I mean the ones I develop)16:18
Sc0rpiusso I'm not sure if I want to change it16:18
DocScrutinizer05believe me you want16:18
Sc0rpiusthat will make the executable files larger .(16:18
Sc0rpiusor bigger actually16:19
DocScrutinizer05IF it does then you definitely want16:19
DocScrutinizer05odds are it won't though, as your projects didn't use that file and the settings in there16:19
DocScrutinizer05[2012-07-19 08:58:14] <Pali> DEB_BUILS_OPTIONS is some hint for debian apps: you can enable this options...16:20
Sc0rpiusI thought every project compiled in ARM used the setting16:20
merlin1991Sc0rpius: no16:21
DocScrutinizer05[2012-07-19 08:57:32] <Pali> DocScrutinizer, problem is only with apps which check that string in debian/rules16:21
Sc0rpiusthat means everything you build using dpkg-buildpackage16:21
merlin1991the rules file has to explecitely check for that and adjust the gcc flags accordingly16:21
merlin1991only a few maemo packages actually do hat16:21
merlin1991modest does for example16:21
Sc0rpiusso even if I remove the thumb option my projects won't be affected16:22
DocScrutinizer05that's why modest sucked a lot in CSSU for quite some time16:22
merlin1991Sc0rpius: yes16:22
Sc0rpiusI don't think modest does, but I'm pretty sure tinymail does16:22
*** Pali has quit IRC16:22
Sc0rpiusand since modest runs over tinymail...16:22
Sc0rpiusyeah I remember the crashes, and trying to debug them16:22
*** Pali has joined #maemo-ssu16:22
merlin19918 phones on my desk, I think I have to rethink what I'm doing here16:24
keriomerlin1991: yeah, they should be at least 1016:24
merlin1991kerio:  :D16:24
kerioalso goddammit Pali, where's KP51 D:16:25
*** mase76 has quit IRC16:25
Palikerio, ask autobuilder :D16:25
keriogoddammit autobuilder, where's KP51 D:16:25
PaliI did not have time for investigation...16:25
keriohm, he's not here16:25
Paliyes, autobuilder is on
Palihe (or she?) is not on #maemo16:26
merlin1991kerio: the good news though is, that I only payed for 4 of them :D16:26
jonwilwhat phones are they?16:26
merlin19914 times n900 3 times n9 1 n95016:27
kerioyeah, i see your problems16:27
kerioyou only have 50% of good phones16:27
merlin1991exactly :)16:27
kerioand you don't have nearly enough of those16:27
jonwilat least you aren't being blasphemous and owning devices with fruit or robots on the back :P16:28
merlin1991Pali: did you push the new enabler to cssu-devel aswell?16:28
merlin1991hm the cssu-devel repo isn't signed yet, so I gues ham is going to complain?16:29
*** LaoLang_cool has quit IRC16:29
Paliyes, this could be problem16:29
Paliwhen you start signing repos, write me public key16:29
PaliI will update enabler16:29
merlin1991doing this right away16:30
merlin1991can't be that the repo suddenly breaks ham :)16:30
Palithis new enabler now using proper signing ham method16:30
kerioenabler for what?16:30
Paliand dpkg-divert on apt-worked will be removed16:30
keriooh, the cssu enabler16:30
Palikerio, enaber for cssu16:30
keriowhat does that even *do*?16:30
Paliit will install cssu repo into HAM16:31
kerioother than magic?16:31
Paliand should also gpg keys16:31
Paliand install cssu metapackage16:31
keriocan't you just do that manually?16:31
merlin1991kerio by default ham ignores packages from a "wrong" domain, so the enabler works around this16:31
keriowrong domain? what?16:31
merlin1991it's a bit how aegis on harm blocks installation of system packages that are not signed by nokia, only less restrictive16:32
Paliincept on fremantle is to open xterm :D16:32
Paliwhich is preinstalled16:32
kerioinception on fremantle is to open xterm from xterm16:33
Palikerio, no open xterm from menu :D16:33
* merlin1991 reads reprepro docs16:33
jon_yPali: btw, can you give a quick description of the n900 boot process? :)16:33
merlin1991nolo something kernel done16:34
merlin1991ah yeah upstart in userspace :D16:34
jon_ythere's another I think called xloader?16:34
merlin1991that would be the something part ;)16:34
Paliomap bootrom --> x-loader --> nolo --> kernel --> /sbin/preinit --> /sbin/upstart ...16:34
keriowhat can be changed?16:35
Palionyl x-loader is signed16:35
kerioi see16:35
kerioand well, the bootrom is a rom16:35
jon_ywhere kernel can be uboot I assume16:35
merlin1991jon_y: yes16:35
kerionolo is flashable, right?16:35
Paliyes, nolo is flashable16:35
Palibut u-boot is too big to fit into nolo partition16:35
merlin1991omfg rekonq is bs "the key sequence ctrl+f is ambiguous, use configure...."16:36
jon_yprobably needed to do things like init the DRAM16:36
jon_yhow does the other payload "multiboot" differ from uboot?16:36
Paliand I was not able to run u-boot spl instead nolo in qemu-system-arm16:36
Pali(u-boot spl is something which init hw, dram, ... something like x-loader)16:37
jon_yI hear it flashes the kernel on everyboot, but maybe its red herring16:37
Paliyes, true16:37
Palimultiboot is started from /sbin/preinit16:37
Paliand check which kernel is running16:37
Palialso show some menu16:37
Paliif you choose other kernel, it flash it and reboot device16:37
jon_yI see, so if I used it to boot a kernel that does not run /sbin/preinit, the thing breaks?16:38
jon_ymultiboot is bad for my nand :(16:39
jon_yI love the n900 so much, there is no replacement for it :(16:40
jon_ymaybe the n950, but nobody can get it now16:41
merlin1991forget the n95016:41
merlin1991system is too closed to be fun16:41
keriowait, multiboot flashes the kernel on every boot? :o16:42
infobotsomebody said multiboot was
merlin1991dang no it's fucking evil text?16:42
jon_yI still remember MS saying Nokia WinMo will be an "Open" platform16:45
jon_ynope, not prying the N900 from my hands16:46
kerioand windows is a "good" OS16:46
jon_yI'll consider it if I can put plain Debian on them16:46
merlin1991dang where's my remove all the bs I don't need apt get line16:48
keriomerlin1991: apt-get autoremove16:50
keriooh you mean from a newly-flashed n900?16:50
merlin1991those are not the droids you are looking for ;)16:50
kerioapt-get remove .*ovi.* cherry tutorial-home-applet16:51
Paliseems, that harm flasher can backup CAL partition:
Paliwill be this flasher work for fremantle?16:51
jon_yI tried too, it was looking the APE images16:52
jon_yno idea what those are16:52
jon_yAPE algorithm or something16:52
merlin1991apt-get remove amazon-installer foreca-installer facebook-installer ap-installer dtg-installer tutorial-home-applet16:52
keriooh right, the installers16:52
keriomerlin1991: and ovi16:52
jon_ycommand saved :)16:52
keriodon't remove .*ovi.* though, that'll get provider and provisioning16:53
*** DocScrutinizer has quit IRC16:55
*** DocScrutinizer has joined #maemo-ssu16:55
*** DocScrutinizer05 has quit IRC16:55
*** DocScrutinizer06 has joined #maemo-ssu16:55
jon_yhmm, nolo is responsible for the flashing?16:57
jon_yPali: feature request, option to power down from uboot :)16:58
kerioPali: what advantages does uboot give to the users?16:59
jon_ykerio: it allows you to boot custom kernels without any flashing?17:00
kerioapart from that?17:00
merlin1991kernel can be on the microsd ;)17:01
merlin1991and if you use palis uboot I think it even recognises if the keyboard is open and boots flashed / microsd kernel accordingly17:01
jon_yalso, pick and choose your kernels!17:02
jon_ysadly, u-boot+payload isn't working yet on my phone17:03
keriojon_y: have you tried uboot-only?17:03
jon_ykerio: I mean uboot has issues loading the kernels from disk too17:05
jon_yask DocScrutinizer06 for the details17:06
jon_yI was just helping out on debugging it17:06
jon_yapparently it fails to read the ubifs if uboot is used17:06
jon_yit will fail *sometimes* when using the flasher -l loader17:07
Palikerio, power down from uboot --> ask u-boot developers17:07
Palithis needs to be done in arm code17:07
jon_ybut should pass if flasher -f17:07
jon_yPali: ok17:07
merlin1991Pali: I just noticed that cssu-devel isn't on 4.1 yet17:08
merlin1991I'll call the update17:08
keriomerlin1991: what's in cssu-d?17:12
keriois it an officially-recognized cssu flavour?17:12
merlin1991stuff for when you're feeling like living on the edge :D17:12
kerioyeah but right now, what's there that's not on cssu-t?17:12
merlin1991bit of ham hackery, and other stuff, pali knows best :)17:13
RaimuSlightly newer, untested builds of apps, mostly17:13
RaimuOr rather, basic program components for maemo.17:13
kerio:D i like all of those things!17:14
jon_ylike phonet?17:14
jon_yyeah, Pali's patches17:15
keriomeh, cba to install it17:16
infobotextra, extra, read all about it, cba is Can't Be Arsed.17:17
merlin1991Pali: as far as I can see you didn't upload the update cssu-enabler17:24
Palimerlin1991, ok I will reupload it17:25
merlin1991hm Pali it is there17:25
merlin1991I mean the files17:25
Paliok, so is cssu-enabler in devel or not?17:26
merlin1991it's only in the incoming dir17:26
merlin1991slightly odd17:26
Palimerlin1991, can you fix it?17:26
merlin1991did you upload the changes file last17:27
merlin1991uh oh17:28
merlin1991No distribution found for 'community-ssu-enabler_0.7_armel.changes'17:28
merlin1991ah yeah you have Distribution "fremantle" in the changes file again17:29
merlin1991should be unstable17:29
PaliI deleted builded dsc & debs17:29
Pali(I deleted from my hdd)17:29
* merlin1991 wonders why 50% of cssu packages have unstable and the others have fremantle17:30
* merlin1991 edits the repo config accordingly17:30
merlin1991Pali it's imported17:32
Palimerlin1991, ok I used scheme from debian/changelog in each package17:33
Palino idea17:33
merlin1991well the repo now accepts unstable and fremantle so we're fine :)17:33
* merlin1991 heads back to getting things signed17:33
*** M4rtinK has quit IRC18:01
merlin1991freemangordon: ping18:07
freemangordonmerlin1991: pong18:20
merlin1991freemangordon: I saw that one package for microb didn't get imported properly18:21
freemangordonmerlin1991: hmm, which one?18:21
freemangordondeb? tgz?18:22
merlin1991diff.gz and tar.gz18:22
freemangordonlemme check18:23
merlin1991hm actually they are there, but there are some leftovers in the incoming dir18:24
freemangordonmerlin1991: source code is in repo,so it might be some failed attempt. it took me several tries18:24
merlin1991yeah there are 3 failed imports in the log18:24
Paliplease test new operator cbs widget update in cssu-devel if installation of custom operator widget working...18:25
PaliDocScrutinizer05, see ^18:26
Palijonwil, also see ^^18:26
freemangordonPali: could you push the latest patches from KP51 to kernel-cssu?18:28
keriofreemangordon: why?18:29
keriokernel-cssu is just a rebranding of kp5118:29
keriothat's stupid18:29
*** Guest92502 has joined #maemo-ssu18:29
*** _rd has joined #maemo-ssu18:29
freemangordonkerio: after all it seems you did not read tho OP on thumb thread :(18:29
kerioprobably not18:29
kerioso yeah18:30
keriokernel-cssu is just a rebranding of kp50 plus the thumb errata18:30
keriomerging the kp51 patches it'll become just a rebranding of kp5118:30
freemangordonkerio: CSSU CANNOT depend on stuff in external repos, write that down somewhere18:32
kerioso... host it?18:32
kerioas kernel-power, i mean18:32
keriopackages that are the same but conflict with each other are awful18:33
freemangordonthe final goal is to have kernel-cssu, not kenel-power18:33
keriopackages that are the same and don't conflict with each other are even worse18:33
keriofreemangordon: once you have it, depend on it in some way18:34
freemangordonkerio: ^^^18:34
kerioperhaps place it at a higher priority? i don't know if that can be done on the repo side18:34
freemangordonkerio: soon or later KP will be obsoleted in favor of kernel-cssu18:34
keriono u :c18:34
freemangordon(most probably)18:35
kerioi thought kernel-cssu was supposed to be omap1 + thumb18:35
keriowith no extra stuff18:35
freemangordonand I cannot risk to put untested SW in CSSU-repo18:35
kerioso DocScrutinizer06 would be able to sleep at night18:35
freemangordonno, kernel-cssu is still to be defined18:35
kerioby the way, have you made any progresses in the omap1 ABI-compatibility?18:37
freemangordonkerio: by the way I was on holiday since monday18:37
kerioyeah, so you had free time :D18:37
kerioi've been on holiday for two weeks now18:38
kerioprobably will be for another two months or so18:38
keriodoes that make you feel better? :318:38
freemangordonyeah, now I feel much better18:39
freemangordonkerio: now you have two months of free time, mind to check that omap1 ABI-compatibility?18:40
Raimu*the sound of crickets*18:50
*** NIN101 has joined #maemo-ssu18:51
*** jonwil_ has joined #maemo-ssu18:56
*** jonwil has quit IRC18:59
*** jonwil_ is now known as jonwil18:59
*** Guest92502 has quit IRC19:03
*** _rd has quit IRC19:03
*** Pali has quit IRC19:27
*** Pali has joined #maemo-ssu19:29
merlin1991DocScrutinizer you'll be happy about this: Pali made a patch to community-ssu enabler that gets rid of the hack we use for apt-worker :)19:35
PaliHere is merge request:
*** jonwil_ has joined #maemo-ssu20:00
*** jonwil has quit IRC20:01
*** jonwil_ is now known as jonwil20:01
merlin1991DocScrutinizer51: got a min?20:31
DocScrutinizer51two even20:31
DocScrutinizer51cig break20:31
merlin1991dpkg-query -W -f \${Version} busybox     gives me the version just fine20:31
merlin1991bug `dpkg-query -W -f \${Version} busybox` does an endless loop of busyboxbusyboxbusybox.....20:31
merlin1991hm $(dpkg-query -W -f \${Version} busybox) works though20:33
merlin1991FUCKING MESSYBOX20:33
DocScrutinizer51considered triple-escape?20:34
DocScrutinizer51oh my20:34
DocScrutinizer51cig over, sorry20:35
DocScrutinizer51bbl, now+1h20:35
merlin1991thanks anway :)20:35
*** arcean has quit IRC20:42
*** Guest92502 has joined #maemo-ssu20:42
*** _rd has joined #maemo-ssu20:42
*** arcean has joined #maemo-ssu20:42
DocScrutinizer51ok, this might take a while (android SDK installer, comment on android site: "make sure you have 90GB free for several builds")20:54
*** Pali has quit IRC20:55
merlin1991and I complain about the size of my /scratchbox dir xD20:55
merlin1991it's 20 gigs in total which is 3 times maemo target and 2 times harmattan target20:55
merlin1991and a huge amount of cssu / whatever in source form20:56
*** M4rtinK has joined #maemo-ssu20:59
DocScrutinizer51hmm, for android it's like 1.5 1.6 2.1 2.2 2.3.3 3.0 3.1 3.2 4.0 4.0?e, and all that for several platforms21:03
DocScrutinizer51well, 190GB free on office /home ;-P21:04
DocScrutinizer51I should run! XP21:05
*** Guest92502 has quit IRC21:06
*** _rd has quit IRC21:06
*** luf has joined #maemo-ssu21:12
*** Woody14619 has joined #maemo-ssu21:13
*** Woody14619 has joined #maemo-ssu21:13
merlin1991pali is a genious21:30
DocScrutinizer51pali is the greatest21:41
merlin1991yes we have a working uninstaller for cssu now :)21:42
DocScrutinizer51uninstalls what?21:42
merlin1991everything back to stock21:43
merlin1991and it works21:44
merlin1991(I flashed 2 times whilst debugging it :D)21:44
DocScrutinizer51I'm believing in that after a diff ls_-!R_before.log ls_-lR_after.log21:45
DocScrutinizer51gawd, life sucks21:49
*** mase76 has joined #maemo-ssu21:50
*** mase76 has quit IRC22:03
*** mase76 has joined #maemo-ssu22:05
keriofreemangordon: i neither have the knowledge nor the will22:41
*** DocScrutinizer06 is now known as DocScrutinizer0522:49
*** Guest92502 has joined #maemo-ssu22:53
*** _rd has joined #maemo-ssu22:53
*** NIN101 has quit IRC23:11
*** mase76 has quit IRC23:18
lufmerlin1991: I don't see new commits in git://
*** jonwil has quit IRC23:23
*** _ade_ has joined #maemo-ssu23:25
merlin1991luf: I've been working on the cssu enabler today with pali, will do curl tomorrow probably23:47
*** Guest92502 has quit IRC23:51
*** _rd has quit IRC23:51
lufmerlin1991: ok. I think you put it there yesterday.23:54
lufSo my mistake.23:54

Generated by 2.15.1 by Marius Gedminas - find it at!