**** BEGIN LOGGING AT Wed May 02 02:59:58 2012 May 02 03:48:11 i had a dream once, where the USB port broke on the back of my midi controller May 02 03:48:22 and there was just this giant shot of the usb port May 02 03:48:48 the midi controller was exaggeratedly macroscopic in this dream May 02 03:49:44 and i think the gist of this was that the whole midi controller hinged on this funny little plastic port that was supposedly "universal", but was fragile and could only interface with one shape May 02 03:50:44 Unless they are micro-usb ports, they're easy enough to re-solder May 02 05:06:32 Great May 02 05:07:00 I finally decide that yes, I need to buy a backup n900, and then realize my ebay coupon code expired yesterday...oh well. May 02 05:08:44 On the bright side, having just spent that money will make it easier for a while to resist the temptation to decide I also need an element of the set {Archos 5, Dell Streak 5 AWS, Galaxy Note}. May 02 07:52:04 How do you get/read email on maemo? May 02 08:14:27 hello May 02 08:14:32 ius there anyone left here? May 02 08:33:58 vi___: I think so May 02 08:35:00 woot! May 02 08:35:48 what is the name of the program that lets you change modest folder subscriptions? May 02 08:46:30 sry I have nfc May 02 08:46:49 what near field communication has to do with it? May 02 08:47:00 :D May 02 08:47:17 nfc is no fucking clue, the fancy new stuff you're talking about has nothing todo with it ;) May 02 08:48:07 i know May 02 09:03:36 nfi, not nfc i think merlin1991 meant May 02 09:08:03 Hi, any mail client suggestion? May 02 09:08:32 I think no fetchmail + procmail + mutt for maemo May 02 09:39:53 LaoLang_cool: there is mutt for maemo. May 02 09:55:28 Sicelo: Yes, looking for fetchmail/getmail :) May 02 09:56:00 for offline use? May 02 09:56:27 Sicelo: yes May 02 09:59:58 oh, mutt: Depends: slang1a-utf8 (> 1.4.9dbs-4) but it is not installable... May 02 10:21:12 Can't update with deb http://repository.maemo.org/ fremantle/sdk free non-free May 02 10:23:13 oh, notice a space between maemo.org/ and fremantle, will try again May 02 10:25:10 good, works now May 02 13:13:42 if i were to sell the top half of my (dead) second n900 as a spare part May 02 13:13:54 how much would it sell for do you think May 02 13:29:22 50 bucks maybe May 02 13:29:51 probably less May 02 13:31:03 what happened to the main body? May 02 13:31:12 eaten by snakes? May 02 13:31:27 i needed the pcb May 02 13:32:47 Why? May 02 13:33:25 look, I'd probably pay 50 bucks for a complete defect device, with a good (US) qwerty keymat, an intact case, maybe domesheet, LCD halve (for sure with some minor scratches) etc May 02 13:34:45 i have the whole device except for the pcb May 02 13:34:54 its gota dutch keymat May 02 13:35:00 eeeew May 02 13:35:01 which is pretty standard May 02 13:35:11 4 cursor keys? May 02 13:35:13 yes May 02 13:35:21 Why did you want the PCB? May 02 13:35:21 fine, how much for all? May 02 13:35:22 idk if its special or international May 02 13:35:30 alt-c = euro sign May 02 13:35:50 == standard (US) May 02 13:35:55 ok :) May 02 13:36:15 the screen itself is not scratched May 02 13:36:19 only the metal edge May 02 13:36:43 without any mainboard (domesheet, SIM holder, whatnot else, camera...) I'd pay 25 EUR May 02 13:37:25 how much do you think itll go for on ebay? May 02 13:37:37 I doubt it will go at all May 02 13:37:47 * SpeedEvil ponders. May 02 13:38:18 look, for ~80€ I get used *working* N900 May 02 13:38:49 for ~120€ refurbished perfect condition ones May 02 13:40:05 also iirc for ~80 USD I can get *new* spare parts for a complete display halve May 02 13:42:01 I bet you could do something cool with the top half and a RPi May 02 13:42:10 meh May 02 13:42:38 wouldn't buy one for it, but if I had one lying about May 02 13:42:39 i reseachred the lcd, couldnt find much info May 02 13:42:52 there's none May 02 13:43:32 surely a fairly generic resistive screen running at 800x480 May 02 13:43:44 haha May 02 13:44:20 I thought the custom X was the clever part May 02 13:44:29 nope May 02 13:44:38 the display controller in the omap is May 02 13:45:01 the display controller is in LCD May 02 13:45:13 ah - interesting May 02 13:45:15 w/e you call it May 02 13:45:22 the thing driving the lcd interface May 02 13:45:38 i dont even know what interface its using May 02 13:46:47 that sounds quite 'special' - explains why xssh never worked well May 02 13:46:51 http://wiki.maemo.org/N900_LCD_panel -- is all i have May 02 13:47:10 the X server they are using doesnt even directly talk to the LCD May 02 13:47:20 it uses the GPU May 02 13:48:46 ls -l /sys//class/backlight/acx565akm -- none of that is in OMAP display controller May 02 13:48:57 no May 02 13:49:05 thats the I2C part of it May 02 13:49:08 but the actual graphics May 02 13:49:18 i've looked into the acx565akm driver source May 02 13:49:31 15 SCLK May 02 13:49:32 16 RESET May 02 13:49:32 17 MOSI May 02 13:49:32 18 MISO May 02 13:49:37 is control May 02 13:49:40 and video is May 02 13:49:40 4 Data0- May 02 13:49:40 5 Data0+ May 02 13:49:40 6 Clock- May 02 13:49:40 7 Clock+ May 02 13:49:40 8 Data1- May 02 13:49:41 9 Data1+ May 02 13:49:47 WHAT? May 02 13:50:01 the pins on the lcd connector May 02 13:50:12 again, so what? May 02 13:50:28 they are driven from some muxed non GPIO pins May 02 13:50:59 cabc_mode is in 565akm chip which is *real* LCD-controller May 02 13:51:11 i looked those pins up and they can be any of the display buses the OMAP supports May 02 13:51:25 again, so what? May 02 13:51:32 It's the controller only of the backlight and stuff - it doesn't do actual display data May 02 13:51:48 SpeedEvil: I honestly dount that May 02 13:52:02 still, you cant use the control bus (i2c) to do graphics May 02 13:52:09 cabc-mode is automatic from data May 02 13:52:10 even if it was on the same chip May 02 13:52:15 hm May 02 13:52:16 m May 02 13:52:24 I should in principle look up the datsheet. May 02 13:52:34 i tried that May 02 13:52:38 couldnt find it May 02 13:52:38 But today, I need to clear the lawn of stuff so it doesn't go wierd and yellow in bits. May 02 13:52:40 :/ May 02 13:52:46 * SpeedEvil stabs unavailable datasheets. May 02 13:53:31 look, guys, I know the datasheets of waaay simpler OM FreeRunner display May 02 13:54:00 and even this jkdunno chip has gamma correction and whatnot else May 02 13:54:16 so? May 02 13:54:52 if anybody wants to use the n900 lcd for another purpose they need to be able to interface with its video port May 02 13:54:55 so N900 display isn't a simple plain vanilla LCD you just need to hook on to LVM May 02 13:55:02 LVD* May 02 13:55:45 which is some weird differential signalling 2 data line 1 clock line port May 02 13:56:00 which = n900 display port May 02 13:56:00 afk May 02 14:39:11 so any clue on what the RX pin of the serial port is May 02 14:39:16 TX has been identified May 02 14:39:20 but RX??? May 02 14:45:59 Which serial port? May 02 14:47:00 N900 May 02 14:48:01 http://www.cpkb.org/wiki/File:Nokia_n900_pinout.jpg May 02 14:48:38 thats MBUS May 02 14:48:42 i mean rs232 May 02 14:48:57 there is no rs232, kinda obviously ;) May 02 14:49:04 there is May 02 14:49:30 rs232 is the electrical interface with +12/-12v levels... May 02 14:49:54 According to jacekowskis measurements J2000 is serial console, pin 5 is confirmed to be TX (115200 8-N-1 ~2.6 V, connected to 1st UART in omap - /dev/ttyS0 in Linux). Since J2000 is group of 5 pads one could suppose the serial console uses HW flow control. May 02 14:50:04 okay 3.3v serial May 02 14:50:46 http://www.cpkb.org/wiki/Nokia_N900_pinout May 02 14:50:51 http://wiki.maemo.org/N900_Hardware_Hacking#Debug_ports May 02 14:50:55 "According to jacekowskis measurements J2000 is serial console, pin 5 is confirmed to be TX (115200 8-N-1 ~2.6 V, connected to 1st UART in omap - /dev/ttyS0 in Linux). Since J2000 is group of 5 pads one could suppose the serial console uses HW flow control." May 02 14:51:29 yea thats what i cp'd May 02 14:51:40 but no one has done any research since then? May 02 14:51:45 and NO, this is neither "MBUS" nor is it 3.3V which will definitely fry your SoC May 02 14:52:31 actually i'd be willing to do tests to locate RxD, if it really hasn't been done May 02 14:52:53 how would you test it May 02 14:53:24 all you really need to do is: sort out pins that are definitively not inputs, and then short the others with TxD until you get a loopback May 02 14:53:41 what if you hit a gnd ? May 02 14:53:47 you can sort out those easily May 02 14:54:13 also, uart IO pins are usually relatively forgiving May 02 14:54:39 but i also just said i'd take the risk of killing mine if it really hasn't been done May 02 14:57:34 scratch my last post, for the mbus part May 02 14:58:46 r00t|home: (forgiving) not this UART May 02 14:59:35 from the service manual: "15) While removing the engine board, be careful not to damage the USB connector." May 02 14:59:40 there's no risk of killing anything by conecting any TP to any other of that "jack" May 02 14:59:49 that's ridiculous... even nokia's internal docs mention the usb port is fragile... May 02 15:00:27 DocScrutinizer: so why don't you know which is RxD (yet)? May 02 15:00:51 no, that's a hint for manually disabled to please notice that USB port is latched in the case hole May 02 15:01:32 "Be aware of the sharp point of the dental tool - be careful not to injure yourself!" May 02 15:01:39 ok, that seems to be a valid explanation May 02 15:01:56 r00t|home: because I'm still waiting for you paying the invoice for finding and documenting TX May 02 15:02:20 DocScrutinizer: i'm sorry, i didn't receive the invoice, send again plz May 02 15:12:11 btw the TxD , how do they count May 02 15:12:18 is that bottom left pin of j2000 May 02 15:15:15 txd is _trivial_ to find, if you have a 'scope, i'd re-check that anyway May 02 15:15:30 just send data and find the pin it's being sent on May 02 15:15:30 well i cant rd mode my n900 May 02 15:15:32 no working usb May 02 15:15:47 "rd mode"? May 02 15:16:04 r&d mode to enable serial console May 02 15:16:42 i'd assume the serial port is available as ttyS0 in linux userspace? May 02 15:17:46 um... cat /dev/ttyS0 -> "io error" May 02 15:18:08 with default kernel its bound to console iirc May 02 15:18:59 /proc/cmdline has "console=ttyMTD,log" for me May 02 15:20:44 weird May 02 15:20:51 mine is not assembled atm May 02 15:21:01 so i cant verify May 02 15:24:29 dmesg appears to say the four ttyS ports are from the 16550 driver... that doesn't seem right May 02 15:25:08 16550 clones are common May 02 15:25:25 i have the rx51 board driver source open May 02 15:25:31 let me look it up May 02 15:25:57 static struct omap_uart_config rx51_uart_config = { May 02 15:25:58 .enabled_uarts = ((1 << 0) | (1 << 1) | (1 << 2)), May 02 15:25:58 }; May 02 15:26:23 dat lijken er maar 3 May 02 15:26:28 mayb 4 just isnt enabled May 02 15:27:13 oops May 02 15:27:19 that looks like only 3 May 02 15:27:30 is what i said May 02 15:27:36 i see /dev/ttyS[1-4] on my device, no idea if they are static or were created by udev - all of them just return io error anyway May 02 15:27:50 hmm May 02 15:27:55 try a boot shell May 02 15:28:02 before anything runs May 02 15:28:11 brb dinner May 02 15:33:51 http://www.slashgear.com/nokia-sues-htc-rim-and-viewsonic-for-patent-infringement-02225635/ May 02 15:34:33 it begins May 02 15:34:48 company going downhill? try to make moneys by sueing May 02 15:34:49 I wonder how much of that is because of "we're the legal department, you can't fire us because we're the moneymakers" :-) May 02 16:01:43 root, you still workijng on serial May 02 16:01:44 ? May 02 16:01:50 try killall getty May 02 16:04:31 peterbjornx: /dev/ttyGS[0-2] , see http://ibot.rikers.org/%23maemo/20101201.html.gz May 02 16:06:29 ah May 02 16:06:31 but May 02 16:06:38 ttyS0 is the debug port on n900 May 02 16:06:48 its in use by getty ;) May 02 16:07:16 where is the definite list of maemo council candidates? May 02 16:07:21 afaik it shoudl have been out yesterday May 02 16:09:59 damn phone won't boot without the third battery contact connected... ;) May 02 16:10:22 never needed 3rd contact on mine May 02 16:11:20 it shows the nokia logo for a while and then shuts off May 02 16:11:56 also, i have entries for getty on ttyS0 in etc/inittab, but commented out, none running May 02 16:14:21 upstart ;) May 02 16:14:32 none running, i said May 02 16:14:34 event.d May 02 16:14:41 inittab is ignored May 02 16:15:02 ok, my OTHER n900 boots with externally hooked up battery (only two pins), weird May 02 16:15:12 still, no getty running May 02 16:15:38 mayb different nolo rev May 02 16:16:57 the second one also doesnt have the green text on the nokia logo screen May 02 16:17:18 thats r&dmode May 02 16:17:21 the green text May 02 16:17:28 ah May 02 16:17:46 min had that enabled sice i got it, bought used May 02 16:18:35 ok, device a, no green text, with external battery shuts down after showing the nokia logo May 02 16:19:00 device b, green text, boots with external battery, but is not stuck at the 5-dots screen... May 02 16:19:51 but backupmenu works... May 02 16:20:22 external battery? May 02 16:20:52 ShadowJK_: battery hooked up with wires, so i can get to the debug port with the device running May 02 16:21:01 ah May 02 16:21:30 i used the battery holder from a battery charger and soldered wires with micro-clips to it May 02 16:27:40 ok, phone running, shell open, scope ready May 02 16:31:54 nice May 02 16:35:46 is it just me, or is the usb/serial console (via backupmenu) rather flaky? it drops characters all the time, or hangs completely... May 02 16:36:08 idk May 02 16:36:12 but did you get rx? May 02 16:37:09 still need to find the right serial port device, then tx, THEN rx May 02 16:41:03 any suggestions welcome May 02 16:44:18 ttyS0 should be it May 02 16:44:29 does maemo boot in qmeu? May 02 16:46:04 ~ # yes >/dev/ttyS0 May 02 16:46:04 yes: Input/output error May 02 16:47:26 ps -A May 02 16:47:52 ok, the ttyGS devices are the n900-end of the ttyACM device it seems May 02 16:48:50 ~ # ls -l /proc/*/fd/* 2>/dev/null | grep ttyS May 02 16:48:50 ~ # May 02 16:48:51 # is, like, pound, or hash, or sometimes the number sign. a tic-tac-toe board, or octothorpe May 02 16:49:08 there IS NO GETTY running, i can paste the process list in a query if you insist May 02 16:49:44 ok May 02 16:50:15 320 root 0 SW< [mmcqd] May 02 16:50:15 363 root 1660 S < /sbin/udevd --daemon May 02 16:50:15 581 root 0 SW< [vibra] May 02 16:50:15 685 root 1392 S /usr/sbin/watchdogs May 02 16:50:15 783 root 0 SW< [bluetooth] May 02 16:50:16 792 root 0 SW< [nokia-av] May 02 16:50:18 809 root 0 SW< [mboxd/0] May 02 16:50:20 835 root 0 SW< [wl12xx] May 02 16:50:22 1180 root 2836 S ash /usr/share/backupmenu/BackupMenu.item May 02 16:50:24 1802 root 2840 S -sh May 02 16:50:26 1923 root 2704 S sleep 15 May 02 16:50:28 1926 root 1392 S /usr/sbin/evkey -u -t 1000 /dev/input/event1 May 02 16:50:30 1927 root 2840 R ps auxwww May 02 16:51:22 weird May 02 16:51:37 ls May 02 16:51:38 what are the minor,major ids of your ttySes May 02 16:52:10 crw-rw---- 1 root 20 4, 64 Jan 1 1970 /dev/ttyS0 May 02 16:52:10 crw-rw---- 1 root 20 4, 65 Jan 1 1970 /dev/ttyS1 May 02 16:52:24 crw-rw---- 1 root 20 250, 0 Jan 1 1970 /dev/ttyGS0 May 02 16:52:24 crw-rw---- 1 root 20 250, 1 Jan 1 1970 /dev/ttyGS1 May 02 16:56:13 any new suggestions? else i'll boot my phone back into normal mode for now... May 02 16:56:35 does it tell what io error? May 02 16:57:18 i don't have strace installed May 02 16:57:47 try May 02 16:57:55 minicom May 02 16:57:57 or getty May 02 16:58:04 maybe its not initialized yet May 02 17:01:06 'getty 115200 /dev/ttyS0' runs, but i can't verify if it does anything... 'yes >/dev/ttyS0' still goes io error with getty running and after killing it May 02 17:01:30 let me google it May 02 17:11:35 weird May 02 17:11:36 peterbjornx: does "omap-serial" make any sense? http://omappedia.org/wiki/Kernel_UART May 02 17:13:02 the service manual indicates that there are three uarts, one on the test pads, one unused, and one connected to "BT/FM" May 02 17:13:54 still that shouldnt matter for usermode access May 02 17:14:09 well, omap-serial is not present May 02 17:14:14 let me reassemble my n900 and look for a way to spam the serial port May 02 17:14:21 it can do serial output May 02 17:14:24 from console May 02 17:14:31 so why not like this? May 02 17:14:40 waitt a minute May 02 17:14:48 define "from console" May 02 17:14:52 they have to be initialized May 02 17:14:55 kernel console May 02 17:15:02 in research and development mode May 02 17:15:16 console=ttyS0 is also passed May 02 17:16:09 yeah, i also saw that... May 02 17:16:09 ~ # cat /proc/cmdline May 02 17:16:10 init=/sbin/preinit ubi.mtd=rootfs root=ubi0:rootfs rootfstype=ubifs rootflags=bulk_read,no_chk_data_crc rw console=ttyMTD,log console=tty0 snd-soc-rx51.hp_lim=42 snd-soc-tlv320aic3x.hp_dac_lim=6 May 02 17:16:45 do you have a phone in r&d mode? May 02 17:18:00 the one i am using right now May 02 17:18:29 then enable serial port using flasher-3.5 May 02 17:18:50 link? May 02 17:19:11 connect it in update mode as if you were to flash it May 02 17:19:37 i never flashed it... can't be done from the device itself? May 02 17:20:15 then how did u set r&d mode? May 02 17:20:30 do this when its connected: flasher-3.5 --set-rd-flags=serial-console May 02 17:21:15 as i said before, it seems i bought the phone with r&d already enabled May 02 17:21:21 okay May 02 17:21:28 are you on windows or linux May 02 17:22:28 linux May 02 17:22:56 Meanwhile: Treasure Hunter Says He Has Located Bin Laden's Body May 02 17:23:02 OMG OMG OMG May 02 17:23:15 "treasure"... May 02 17:23:21 download flasher May 02 17:23:35 http://tablets-dev.nokia.com/maemo-dev-env-downloads.php?f=maemo_flasher-3.5_2.5.2.2.tar.gz ? May 02 17:26:48 yes May 02 17:27:38 # ./flasher-3.5 --set-rd-flags=serial-console May 02 17:27:38 flasher v2.5.2 (Oct 21 2009) May 02 17:27:38 USB device found found at bus 001, device address 118. May 02 17:27:38 Found device RX-51, hardware revision 2101 May 02 17:27:38 NOLO version 1.4.14 May 02 17:27:39 Version of 'sw-release': RX-51_2009SE_21.2011.38-1_PR_MR0 May 02 17:27:41 # May 02 17:28:27 reboot device and it should be printing console to serial May 02 17:28:28 :) May 02 17:34:05 i still have the problem that it doesn't boot up in normal mode with the external battery setup May 02 17:34:28 even bootloader dumps info :) May 02 17:34:44 else, just connect 3rd pin May 02 17:34:57 3rd? May 02 17:35:09 battery May 02 17:35:37 i have no battery holder with 3 pins, would have to build something first May 02 17:36:09 well, you need R&D enabled anyway, to make debug UART powered up May 02 17:36:35 as we said, that had been enabled all along May 02 17:36:52 logging to UART during boot is happening always, when R&D and console enabled via flasher May 02 17:37:21 in R&D mode it won't check for BSI May 02 17:37:33 so I don't get your problem May 02 17:37:59 it gets stuck at the animated five dots May 02 17:38:01 don't ask me May 02 17:38:15 that doesnt matter May 02 17:38:36 its output shitloads of data over debug before that happens May 02 17:39:31 indeed May 02 17:40:26 http://wiki.maemo.org/N900_Hardware_Hacking/serial_dump May 02 17:41:53 r00t|home, are you still here? May 02 17:43:11 why is no one reverse engineering nolo May 02 17:43:34 its a pity to be stuck with such a low level "blackbox"component May 02 17:43:41 because it's too demanding May 02 17:44:04 ack for the pity May 02 17:44:07 probing debug port is worse than watching porn - no hands free for typing... May 02 17:44:18 i think i have something that looks like txd now May 02 17:44:56 i get some transmissions there while the nokia logo is displayed May 02 17:45:01 BEWARE of max voltage on those pins! May 02 17:45:19 what's the voltage on my multimeter probe? May 02 17:45:32 feeding 3V3 to it already seems to fry things May 02 17:45:40 DocScrutinizer: also, there is a 5 volts output right next to those pins... May 02 17:45:59 hmm indeed May 02 17:46:19 input actually May 02 17:46:22 its VBUS iirc May 02 17:46:31 the v+ pin from usb May 02 17:46:32 yup May 02 17:46:50 yeah, it measures 5 volts only with usb connected May 02 17:47:06 or with booston running ;-D May 02 17:47:30 peterbjornx: well, txd... i'll rather not try to describe it, but mark it on a picture May 02 17:47:43 next we need a reliable way to read input for finding rxd...? May 02 17:47:53 anyway, voltage of that UART likely being 2V8 May 02 17:48:07 it looks like it was even less, on my scope May 02 17:48:28 I'll recheck May 02 17:48:36 cat /dev/ttyS0 if that were working May 02 17:48:45 wait May 02 17:48:57 do you have a levelshifter? May 02 17:49:16 not for <3.3v May 02 17:49:22 ok May 02 17:49:38 ttyS0 should work with rd mode on May 02 17:49:43 its used by nokia too May 02 17:49:44 the peaks of the signal i am seeing for what is likelty rxd are only 200mV... May 02 17:49:49 *likely May 02 17:50:19 maybe i'm using the wrong ground... (the shield inside the battery compartment right now...) May 02 17:50:23 how did you determine it? May 02 17:50:27 WRONNG May 02 17:50:29 scope May 02 17:50:33 thats a bad gnd May 02 17:50:49 it's a batter gnd than the battery negative terminal May 02 17:51:14 and i'm not going to hold two probes into the debug port... May 02 17:51:21 According to my measurements with DMM (Digital Multimeter) the following pads are ground (counting from top-left and always going left-to-right in case of group of pads): first of J2000, last of J2060, first of J5602 and J3115. May 02 17:51:23 might be even 1.8V May 02 17:52:16 i guess i'll cut out the shield metal between the ports, and then solder wires... May 02 17:52:45 oufff May 02 17:53:02 battery minus is not GND! May 02 17:53:03 ? May 02 17:53:37 DocScrutinizer: it's connected to the shield at least... and to a bunch of pins on the debug port May 02 17:53:58 with low enough impedance for my multimeter to consider them connected, at least May 02 17:54:05 no, it isn't May 02 17:54:28 the impedance is at very least sth like 20mOhm May 02 17:54:40 USB has proper GND May 02 17:54:44 im using the debug pins for gnd May 02 17:54:46 battery hasn't May 02 17:54:46 DocScrutinizer: 200mOhm is below what a multimeter can reasonably measure May 02 17:55:02 yes but there might be other electrical properties May 02 17:55:04 DocScrutinizer: and won't make a difference for measuring the uart signal May 02 17:55:07 capacitance May 02 17:55:25 you are dealing with relatively high frequencies May 02 17:55:37 peterbjornx: a uart isn't high frequency May 02 17:55:38 r00t|home: again wrong. if you also have battery connected across same impedance May 02 17:56:16 use the goddamn USB shield! May 02 17:56:20 it's true GND May 02 17:57:04 http://www.theregister.co.uk/2012/05/02/nokia_patent_suits/ May 02 17:57:11 DocScrutinizer: i'll try just to show you it doesn't make a difference... May 02 17:57:12 r00t , relatively high May 02 17:57:35 115200 baud May 02 17:57:45 r00t|home: you won't show anything to a EE with 40 years of experience May 02 17:57:50 thats over 100khz May 02 17:57:57 so it might make a difference May 02 17:57:58 i wonder how long until i mix up the battery clips and blow up my phone... May 02 17:58:33 might be a nice guide May 02 17:58:40 protection diode mod May 02 17:58:52 a led over the pins in inverse direction May 02 17:59:13 and something to prevent overvoltage May 02 17:59:30 peterbjornx: so the led blows first, and then, 2ms later, the phone? ;) May 02 17:59:46 that fast? May 02 18:00:11 and a diode between b+ and actual b+ May 02 18:00:13 dunoo... try yourself May 02 18:00:26 ive blown up a n900 once and as they are getting more rare by the day it might be a good idea to prevent killing more May 02 18:00:31 can't put a diode there, too much voltage drop, and breaks charging May 02 18:01:06 use your best tool! your *brain* May 02 18:01:19 true May 02 18:01:33 but the idea is good isnt it May 02 18:02:29 the idea would work if your diode would clamp to <~1V @ >2A May 02 18:02:51 then battery protection would cut out when you reverse polarity May 02 18:02:52 bigass diode... May 02 18:06:34 i'd guess this is TxD... just the low voltage is weird... http://www.constancy.org/~r00t/IMG_9243.JPG May 02 18:07:41 project steps: a) find testpoint TX (by monitoring http://wiki.maemo.org/N900_Hardware_Hacking/serial_dump). b) find /dev/ttyX that also sends to same TX TP. c) now *read* from that tty while feeding some signal of appropriate level to other TP under investigation for RX May 02 18:07:49 try the one underneath it in the picture May 02 18:08:11 DocScrutinizer: sure May 02 18:09:35 _OR_: you find the RX pin of SoC, tell me which one it is, and I check with DMM on this one: http://maemo.cloud-7.de/Gallery-N900-exploded/platine_01.jpeg May 02 18:09:57 peterbjornx: right, the one below has a good signal! May 02 18:10:24 peterbjornx: i was going by the "number 5" and testing the pins that might be #5 May 02 18:10:46 i was too May 02 18:10:55 DocScrutinizer: no pintout for the SOC? May 02 18:10:59 but right above it says pin "1" isnt part of the group May 02 18:11:18 the 35 one has a datasheet online May 02 18:11:19 sure, somewhere at UM from TI May 02 18:11:33 peterbjornx: really stupid way to number pins anyway... May 02 18:12:39 guys, we haven't invented that numbering May 02 18:13:01 no, the author of the wiki page did? May 02 18:13:05 it's from Nokia_N900_RX-51_Schematics.pdf May 02 18:13:09 ewww May 02 18:13:22 r00t|home: that'd be basically me, partially May 02 18:14:54 anyway... here's the actual TxD now: http://www.constancy.org/~r00t/IMG_9246.JPG May 02 18:15:40 found a pinout May 02 18:15:40 which is 2.8V May 02 18:15:45 ti3530 datasheet May 02 18:15:51 most pins match :) May 02 18:17:09 tx: e26 rx: f28 May 02 18:17:42 wait May 02 18:17:53 nvm May 02 18:18:01 those pinouts are valid May 02 18:18:12 http://www.ti.com/lit/ds/sprs507f/sprs507f.pdf -- for a different soc May 02 18:18:17 but pincompatible :) May 02 18:19:27 refresh http://wiki.maemo.org/N900_Hardware_Hacking#Debug_ports May 02 18:23:19 refresh http://wiki.maemo.org/N900_Hardware_Hacking#Debug_ports :-D May 02 18:24:08 Ok, Nokia is now officially in the business of patent trolling. May 02 18:24:16 Logical conclusion of events, yes May 02 18:24:19 DocScrutinizer: better upload the image to the wiki... May 02 18:24:36 I'd like to IF I knew how ;-D May 02 18:24:42 heh May 02 18:24:46 Why build your own products, or even try selling Microsoft's? May 02 18:29:13 ok, this also works now, from the backupmenu shell, apparanly because of enabling the serial console: ~ # yes >/dev/ttyS0 May 02 18:29:54 nice May 02 18:30:06 so, try cat /dev/ttyS0 May 02 18:30:15 patience, patience, ... :P May 02 18:30:17 with the yes running in bg :_ May 02 18:30:19 :) May 02 18:31:46 rxd is pin 4, by whatever weird logic May 02 18:34:09 so confirmed ? May 02 18:34:16 y May 02 18:34:16 y May 02 18:34:16 [ 516.770996] SysRq : HELP : loglevel0-8 reB[ 51y May 02 18:34:19 then what is pin 2 May 02 18:34:46 pin1 = gnd ,2 = ? ,3 = VBUS, 4 = RX, 5 = TX May 02 18:34:57 * DocScrutinizer **HATES** auto-capitalization of all page names incl URLs May 02 18:36:50 no, no, counted the wrong way May 02 18:37:02 :o May 02 18:37:15 picture please? May 02 18:37:21 no :P May 02 18:37:24 so no confusion May 02 18:37:31 or edit wiki Ascii art May 02 18:38:26 this is serial loopback: http://www.constancy.org/~r00t/IMG_9251.JPG May 02 18:38:52 is that my imei on there? ... May 02 18:39:34 looks like it May 02 18:45:46 i created an account on garage as instructed, but it doesn't seem to work on the wiki... May 02 18:51:12 bro, this sucks... (wiki upload of photos, etc) May 02 18:51:19 anyway May 02 18:51:23 refresh http://wiki.maemo.org/N900_Hardware_Hacking#Debug_ports :-D May 02 18:51:49 add the rx May 02 18:51:58 o.O May 02 18:52:00 ?? May 02 18:52:13 good thing the probe is coverting the imei on that one ;) May 02 18:52:18 this is serial loopback: http://www.constancy.org/~r00t/IMG_9251.JPG May 02 18:52:25 this is serial loopback: http://www.constancy.org/~r00t/IMG_9251.JPG May 02 18:52:31 wait, let me edit that one first May 02 18:53:24 YES pretty please! edit it to 1/3 of pixel size, to make it more like 250k raher than 2M May 02 18:53:42 those were just 1:1 off my digicam... i never inteneded to post them anyway May 02 18:59:18 DocScrutinizer: http://www.constancy.org/~r00t/n900_uart_loopback.jpg May 02 18:59:31 working on TxD May 02 18:59:56 (note: when inlining the picture in the wiki, you can have the wiki scale it down - RTFM or tell me why i can't login) May 02 19:03:13 meh, nemo is slow as hell on n900, it takes 15 seconds to launch the calculator. WTF? May 02 19:03:45 n900 is short onmemory May 02 19:03:53 run it on some other hw May 02 19:04:11 nope, it uses only 38 mb of swap May 02 19:04:53 right now free shows 27MB free ram, and 34MB of swap May 02 19:05:33 on fremantle my swap is usually around 130MB May 02 19:05:40 it is not the ram May 02 19:05:51 it *might* be my sd card May 02 19:06:29 which is very slow on random writes, but still 12MB/s on sequental May 02 19:08:05 DocScrutinizer: http://www.constancy.org/~r00t/n900_uart_txd.jpg May 02 19:10:22 \o/ r00t|home May 02 19:11:14 r00t|home: you can't even upload files >150k, and that's all sane May 02 19:11:44 DocScrutinizer: i'd prefer to upload a large-enough-to-reuse version, and inline a scaled down view... May 02 19:11:54 but ok, i guess whoever runs that site doesn't agreei May 02 19:11:56 *agree May 02 19:12:04 i'll make smaller versions then May 02 19:12:26 still, tell me why i can't login and upload myself... are the wiki accounts synced by a nightly job or something? May 02 19:14:27 but it's impossible to include both the scope screen AND have the probe location clearly visible in 150k... May 02 19:15:07 DocScrutinizer: http://www.constancy.org/~r00t/n900_uart_loopback_small.jpg May 02 19:18:52 DocScrutinizer: http://www.constancy.org/~r00t/n900_uart_txd_small.jpg May 02 19:21:17 r00t|home: I dunno May 02 19:22:27 I basically finished editing that page, I think it maybe doesn't need the additional small pictures above, or at least I'd not embed them but rather link to them May 02 19:22:37 too.. many... quits/joins May 02 19:22:51 Sazpaimon, you good now? :P May 02 19:22:56 none here ;-P May 02 19:23:03 STFU /ignore-r :P May 02 19:23:24 * DocScrutinizer checks chanlog May 02 19:24:13 shall I temporarily ban him? May 02 19:24:17 DocScrutinizer: do you have a picture of the other side of the bare n900 pcb? it'd make a nice base for a pinout picture to replace the ascii-art May 02 19:24:26 freenode keeps killing my connection when I join too many channels May 02 19:24:35 kinda lame May 02 19:24:37 http://maemo.cloud-7.de/Gallery-N900-exploded/ May 02 19:24:49 N900 terrorism at its best May 02 19:25:23 ah, the one you linked was the right side anyway May 02 19:25:36 and this is exploded http://www.phoneworkshop.com/wp-content/uploads/N900-Exploded-View.png :P May 02 19:25:45 I'd suggest to keep the ASCII art May 02 19:26:45 i maybe wouldn't remove it, but it's not very useful for finding the points on the device :P May 02 19:27:43 r00t|home, if it's any help to you http://mohammadag.xceleo.org/n900-disassembly May 02 19:28:04 hate dust? sue Israel May 02 19:28:11 heh May 02 19:28:20 r00t|home: how hard is finding the TP on device, given we already have that one picture of yours? May 02 19:28:32 not anymore May 02 19:28:44 but it'd be nice to have a clean picture with all (known) pins labeled May 02 19:28:48 TP? May 02 19:29:02 TestPoints May 02 19:29:21 refresh http://wiki.maemo.org/N900_Hardware_Hacking#Debug_ports May 02 19:29:39 ah May 02 19:29:51 about dust May 02 19:29:58 how do I clean a PCB without damaging stuff? May 02 19:30:07 dishwasher! May 02 19:30:18 pressurized air for dust May 02 19:30:34 wouldn't that break the USB port? :P May 02 19:30:47 r00t|home, I've already ovened my iPhone, not using another kitchen utility :P May 02 19:30:53 maybe a brush (not too hard, toothbrush about right) May 02 19:30:54 lol May 02 19:30:55 it was painful, but it fixed the problem May 02 19:31:04 190C for 2 mins May 02 19:31:17 over fan was off May 02 19:31:20 painful to you, or the phone? May 02 19:31:23 me May 02 19:31:37 I was in the shower and a phone call came in, I answered it and apparently water got into the earpiece May 02 19:31:56 the volume decreased a LOT and the sound was distorted May 02 19:32:05 I figured a hairdryer would cause more problems May 02 19:32:25 RST38h, NFC what is going on, but fennec starts almost as fast as the calculator (16 vs 12 secs) May 02 19:32:46 calculator takes 12s? May 02 19:32:52 yep May 02 19:32:55 the fuck May 02 19:33:13 yay, got my diy usb connector + rs232 working May 02 19:33:36 peterbjornx: that was quick May 02 19:33:37 and in the meanwhile meego-qml-launcher uses 100% CPU May 02 19:33:38 no protection circuitry but it works May 02 19:33:57 freemangordon, this isn't Maemo 5 then? May 02 19:34:04 so long to the lack of RAM story :( May 02 19:34:17 nome May 02 19:34:21 *nemo May 02 19:35:30 MohammadAG, of course it is not fremantle, if it takes 12 seconds for the calculator to start, I will fire Pali :D May 02 19:36:26 take 3 seconds in fremantle, and it is Qt too, something is terribly wrong in nemo, meego, or whatever is the name now May 02 19:36:39 Oh - not the stock calc May 02 19:36:43 I was about to say May 02 19:36:47 no, the CSSU one May 02 19:37:01 gtk one should be faster May 02 19:37:39 peterbjornx: what are you using the uart port for? May 02 19:38:07 (sth terribly wrong in mer/meego/nemo) well, so nothing new really May 02 19:38:37 going to use it for debugging other phones May 02 19:38:55 o.O?? May 02 19:39:14 and i can connect it to a microcontroller now May 02 19:39:22 peterbjornx: heh... so no need for a 2.8V level shifter? ;) May 02 19:39:27 yea May 02 19:39:50 for uC id need a shifter May 02 19:40:01 for a uC, i'd find a bluetooth interface more handy May 02 19:40:16 DocScrutinizer: but, but... it is the same API. Ok Qml on top of Qt, but even with it. cutetube-qml takes 4 seconds to launch in freantle May 02 19:40:21 *fremantle May 02 19:40:58 qml :-x May 02 19:41:24 cutetube-qml is qml too (as the name imposes) May 02 19:42:12 for anything I find a USB dongle way more handy May 02 19:43:30 debug UART on N900 is mad useful for bootloader and lowlevel kernel hacking May 02 19:43:58 freemangordon, now time sociality May 02 19:44:43 about 4-5 seconds May 02 19:45:11 sociality can't be slower than cutetube-qml May 02 19:45:23 it is May 02 19:45:52 something's messed up May 02 19:46:04 or DBus is slowing it down May 02 19:46:06 sorry when I glare and frown at Nokia's NIH approach to duplicate android's JAVA-only app development May 02 19:46:11 well, ok, it is a bit unfair, as my primary device is OC to 805, while the ine using nemo runs at stock, but still May 02 19:47:06 it takes about 16 seconds easy debian chroot to show me lxde desktop, an that time includes a button press May 02 19:47:20 and 12 seconds for calculator. May 02 19:47:23 in nemo May 02 19:47:46 and honestly before I drool at qml coming to maemo5, I rather hope for alien-dalvik getting ported to our platform May 02 19:48:05 DocScrutinizer, qml is ok in fremantle May 02 19:48:45 it's abraindead concept, like dalvik May 02 19:49:13 well, I am not arguing about the cooncept, but about the speed May 02 19:49:22 maybe May 02 21:11:12 Oh. May 02 21:12:41 Great. May 02 21:12:55 Not flashing N900 today then... May 02 21:13:29 Error 349 (net::ERR_RESPONSE_HEADERS_MULTIPLE_CONTENT_DISPOSITION): Multiple Content-Disposition headers received. This is disallowed to protect against HTTP response splitting attacks. May 02 21:13:59 Thanks Nokia. \o/ May 02 21:27:52 http://www.google.de/search?q=ERR_RESPONSE_HEADERS_MULTIPLE_CONTENT_DISPOSITION May 02 21:27:59 thanks chromium May 02 21:36:32 Zucca: use a decent browser ;-P May 02 21:36:43 WTF? May 02 21:37:18 Zucca: fix your away msg! it's absolutely deprecated May 02 21:45:19 Zucca: and while you're at it, maybe your IRC client is as weird as your browser: " [CTCP] CTCP-VERSION-Antwort von Zucca empfangen: What? CLIENTINFO for available commands.." May 02 21:54:31 Zucca: you might want to tell chrome devels on #chromium-support (esp the author of the post) that >> Comment 5 by willchan@chromium.org, Nov 14, 2011 Marking as WontFix because it is working as intended. You should fix the header field in your server.<< is *very* bad style, arrogant, and won't fly anyway. Mr Willchan nor any other Chrome devel will force the internet at large to fix their headers, nor will chrome users appreciate this May 02 21:54:32 policy in chrome. May 02 21:55:31 Zucca: fix your away msg! it's absolutely deprecated May 02 21:55:48 the message did not appear in the channel... a notice is appropriate i'd say May 02 21:56:02 nope it's not May 02 21:56:12 well, maybe May 02 21:56:31 no mater what, the CTCP answer is BS anyway May 02 21:58:53 r00t|home: a [notice} might be appropriate in /query, but not for highlighting him in channel May 02 22:04:04 i'd say, when you try to talk to him, it's not completely inapproriate to send you a response May 02 22:04:47 r00t|home: basically virtually all automatic response to any channel action is deprecated May 02 22:05:25 no matter whether it's a notice, a replay (msg to chan) a /query, whatever May 02 22:05:34 reply* May 02 22:06:56 that's btw the reason why bots aren't allowed per se, but only when approved by chanop or owner May 02 22:09:13 the only action where a client is supposed to auto-answer is CTCP, and that is fsckdup too for Zucca's client May 02 22:18:56 r00t|home: btw it might be just due to +C chanmode that this notice didn't show up in channel May 02 22:23:37 Zucca: test May 02 22:24:02 >> :Zucca!zucca@dsl-hkibrasgw4-fe54df00-80.dhcp.inet.fi NOTICE r00t|home :Sorry. I'm away. Since: 2012/05/03 0:32:07 Reason: "Away - Sleep" May 02 22:26:18 not sent to channel May 02 22:31:29 yeah, thanks to the design and purpose of +C you don't see any notice sent to channel, so you have no way to tell if this client has sent a channel wide notice or not May 02 22:35:17 DocScrutinizer: i tried on another channel... in the irc protocol message you can see who was the target of the notice May 02 22:35:50 *sigh* May 02 22:36:08 you won't see anything like that in *your* protocol May 02 22:37:19 really, I don't wanna waste lots of time in investigating _how_ terribly broken this IRC client actually is May 02 22:39:35 the fact it is sending me a [notice] on me highlighting the user in chan, _plus_ the user showing in my chan user list as "Zucca (CTCP USERINFO might tell more.)" where the (value) should be realname rather than a stupid info about non.standard CTCP interface... That's just enought to ask Zucca to goddamn fix that May 02 22:46:22 * DocScrutinizer glares at "trumee (Got ZNC?)", "venusaur ()", "speedevil (purple)", PeterWolf(purple)", ghjstdfdgf(hdgsfdgdhg), and of course at all *(unknown) and *(user) May 02 22:53:50 the special punchline in all this is: Zucca complained about a stupid intolerant reaction of his chrome browser to a not-exactly-standard HTML header sent by Nokia server, while his own IRC client obviously doesn't comply to (de facto?) CTCP standard May 02 23:04:21 DocScrutinizer: you have to admint though that ctcp isn't exactly part of the irc standard, it's somwhat of an extension May 02 23:05:37 I see that of 280 users, exactly one has such a bogus "realname" msg May 02 23:06:29 and you can think 700 rather than 280, as I've never seen this on any of the other channels May 02 23:07:36 it being an extension doesn't mean you are supposed to do with it whatever you think is hip May 02 23:08:39 I.E. implement your very own proprietary idea of CTCP May 02 23:10:30 [2012-05-03 01:10:14] [CTCP] Sende CTCP-VERSION-Anfrage an merlin1991. May 02 23:10:31 [2012-05-03 01:10:14] [CTCP] CTCP-VERSION-Antwort von merlin1991 empfangen: Quassel IRC v0.6.3 (dist-de4dd62) (built on Oct 19 2011 21:33:07) -- http://www.quassel-irc.org. May 02 23:11:20 looks fine to me May 02 23:11:32 [2012-05-03 01:11:14] [CTCP] Sende CTCP-VERSION-Anfrage an Zucca. May 02 23:11:34 [2012-05-03 01:11:14] [Abwesend] Zucca ist abwesend: Away - Sleep May 02 23:11:35 [2012-05-03 01:11:15] [CTCP] CTCP-VERSION-Antwort von Zucca empfangen: What? CLIENTINFO for available commands.. May 02 23:12:11 looks like an asshat irc client developer's idea to me May 02 23:13:00 hm I'd like to see the ctcp rfc May 02 23:13:06 I think CLIENTINFO is part of it May 02 23:13:12 (and my client doesn't support it) May 02 23:14:06 whatever the rfc says, I'm damn sure there's no "What? CLIENTINFO for available commands.." May 02 23:14:19 :D May 02 23:15:13 and definitely no "CTCP USERINFO might tell more" May 02 23:19:17 merlin1991: your client is free to ignore CTCP queries at your discretion. It's NOT supposed to answer BS to well known defacto standard CTCP queries May 02 23:22:40 [2012-05-03 01:22:02] [CTCP] CTCP-CLIENTVERSION-Antwort von Zucca empfangen: irssi ver. 0.8.15 on zelan.zapto.org. May 02 23:22:53 0.8.15 yeah suuuure May 02 23:23:53 well 0.8.15 is out already May 02 23:24:36 merlin1991: sounded like a hoax to me, 08-15 May 02 23:25:09 that's a german thing ;) May 02 23:25:51 anyway it seems to me there are more irssi users, and none of them is answering with "CTCP USERINFO might tell more" to my client's realname inquiry May 02 23:26:30 I wonder if that behaviour is a new clusterfuck introduced with 0.8.15 May 02 23:27:24 quite in line with new clusterfuck "Error 349" in chrome, which Zucca suffers from as well May 02 23:28:07 seems he got a truly poettering distro installed May 02 23:28:18 totally ot, but do you know a pop3 to imap "bridge" ? May 02 23:28:30 I'm tired of May 02 23:28:32 err May 02 23:28:36 umm, webservice? May 02 23:28:47 I'd prefer to run my own on my server :) May 02 23:29:03 so I was wondering if you have a tool name at hand :) May 02 23:29:21 hmm, my server answers both imap and pop3, but I admit it doesn't work nicely May 02 23:30:03 my idea is to have my server poll several pop3 accounts over the net from myself and offer the accumulated mails over imap May 02 23:30:24 that's what I'd like to have ultimately .) May 02 23:30:37 the one of both locks maildir et al on my one client checking mail, and that causes the other one fail during the next 5 or 10 minutes, til that lock times out May 02 23:31:50 aah. well, nope. Nothing at hand right away. I'd think you can configure sendmail or any of the other MTA's to do exactly that May 02 23:33:34 http://conversations.nokia.com/2012/04/01/nokia-ice-the-new-super-cool-phone/ - I hadn't seen :) May 02 23:40:03 merlin1991: apropos cyrus May 02 23:40:26 hm? May 02 23:43:08 merlin1991: man cyrus-master May 02 23:43:30 I'd guess you'll find what you're looking for there May 02 23:45:25 duh the homepage suggests it is scaleable to large enterprises aka insanely hard to configure :D May 02 23:46:53 hehe, sounds like my "sysadmin" who worked for a internet provider and now does admin at BA-systemhaus May 03 00:39:31 This is getting moderately old. I think I'll go start dinner before I take my next crack at breaking this reboot loop. May 03 00:40:47 In other news, how the heck did the design factor of the BT8010 not catch on better. May 03 01:04:14 Kilroo: you lost me on the last one May 03 01:18:16 Jabra BT8010 headset. I laundered mine by accident, and looking for a replacement brought me to the realization that it's out of production and nobody seems to be selling anything similar. May 03 01:18:30 :/ May 03 01:20:06 And still in a reboot loop, darn it May 03 01:21:45 http://www.ebay.co.uk/itm/NEW-JABRA-STONE-BLUETOOTH-STEREO-HEADSET-BT8040-BT8010-/320546938595?pt=PDA_Accessories&hash=item4aa2161ee3 May 03 01:31:25 cool headset indeed May 03 01:32:00 Kilroo: what did cause the loop? May 03 01:36:01 When I get rich May 03 01:36:08 I will reverse engineer N900 May 03 01:36:15 And come out with a new N900S May 03 01:36:26 It would be excatly the same as N900 outside May 03 01:36:35 But on the inside it would be more awesome May 03 01:36:45 and based on sid so it is always recent May 03 01:47:34 DocScrutinizer: The first time, trying to get u-boot/bootmenu + NITDroid working. The second time, flashing back to stock and restoring my backupmenu backup from before the first attempt. May 03 01:48:36 umm May 03 01:49:51 when was the BN-backup done? before or after installing u-boot and/or nitdroid? May 03 01:49:58 Before. May 03 01:50:21 that's *rather* starnge then May 03 01:50:50 Well, one thing I read did make me a little worried that the problem might be because it was also before I resized my partitions...but... May 03 01:51:03 duh May 03 01:51:38 Oh, that would do it? May 03 01:52:01 you reflashed both or only rootfs? May 03 01:52:19 Well, first I tried only kernel. Then kernel and rootfs. May 03 01:52:32 hmmm May 03 01:52:49 how exactly did you resize May 03 01:53:04 gparted from my desktop. May 03 01:53:12 I mean, what was the resulting change? May 03 01:53:25 Oh. 21 and 8. May 03 01:53:39 no change in fs type, or sequence? May 03 01:53:42 No. May 03 01:53:48 should work May 03 01:54:38 you are using recent BM? May 03 01:54:52 I.E. not the deprecated v-1 May 03 01:55:12 as of ~1 year ago May 03 01:59:50 well, if I should guess, I'd suspect your kernel modules don't match the kernel, after restoring BM-backup. /lib/modules/current -> ./2.6.28-omap1 May 03 02:00:21 you might check your BM .tgz content for that May 03 02:01:39 but maybe your problem is with bootmenu in /etc May 03 02:02:48 you might take care to move away any file /bootmenu.sh after restore and before reboot May 03 02:04:16 I.E. don't restore /lib/modules/* and /bootmenu.sh on a fresh reflash, and there shouldn't probably be any bootloop May 03 02:06:14 DocScrutinizer: dox! May 03 02:10:28 no dox here, no caz either May 03 02:12:35 * DocScrutinizer idly wonders how to repartition a standard N900 OS to 8G /home via PC's gparted May 03 02:12:50 wouldn't even get exported to PC May 03 02:13:24 you'd need raw eMMC exported to repartition it May 03 02:13:52 which is a bit, err, hard to do May 03 02:14:10 ...Er, what? May 03 02:14:25 unless you already use backupmenu or similar May 03 02:14:55 Right. I installed backupmenu, made backups, used mass storage mode to allow gparted to repartition May 03 02:14:59 I think BM can export /dev/mmcblk0 May 03 02:15:31 while usually you export /dev/mmcblk0p1 via mass storage May 03 02:16:17 Then I tried to effect the correct combination of http://forum.nitdroid.com/index.php?topic=73.0 and http://talk.maemo.org/showthread.php?t=81613 May 03 02:16:47 yeah, multiboot May 03 02:16:51 :-S May 03 02:16:54 and failed horribly. May 03 02:17:01 ~multiboot May 03 02:17:01 multiboot is probably http://maemo.org/packages/view/multiboot/ May 03 02:17:03 Fuck multiboot, I refuse to go back to it. May 03 02:17:14 ...Well, so far. May 03 02:17:14 ~maemo-multiboot May 03 02:17:15 rumour has it, maemo-multiboot is deprecated, and a horrible hack. PROBLEMS WITH NITDROID/MULTIBOOT? reflash rootfs&kernel aka COMBINED May 03 02:18:43 I don't get it why your system bootloops when you restore a 'clean' maemo5 backup on a clean fresh maemo5 May 03 02:18:59 Neither do I, yet. May 03 02:19:19 so I still suspect May 03 02:19:21 The idea of blaming multiboot is somewhat appealing. May 03 02:19:47 you either did backup while already powerkernel or uboot or multiboot installed May 03 02:20:42 Okay, see, now, if you could point me to where I should have found it clearly indicated that would screw it up, I'd greatly appreciate it. May 03 02:20:47 or you did some special manual tweaks to your initscripts et al, to accomodate your new partitioning May 03 02:21:00 Because that was not one of the many CAUTIONS I found before I started. May 03 02:23:08 sorry, the nitdroid guys seem reluctant to mention their multiboot abomination frequently breaks stuff, esp in combination with non-standard kernels (where uBoot counts as kernel too) May 03 02:24:38 Well, in this case multiboot had ridden along with powerkernel 49, not nitdroid, but the fact remains that if there was anything that mentioned that added complications to using backupmenu I missed it. May 03 02:24:57 as a matter of fact, I've not yet heard of a single user who claimed he usually uses nitdroid and only rarely boots to maemo May 03 02:25:10 nonono May 03 02:25:24 multiboot is completely unrelated to powerkernel May 03 02:25:41 powerkernel comes with uBoot, not multiboot May 03 02:26:29 In that case I suppose it's possible that I got confused at some point, but I assure you, I only had multiboot installed in order to use powerkernel May 03 02:27:11 and BM is just fine, it's basing on bootmenu which is still a 3rd class of "tool", though aiui multiboot also messes with bootmenu May 03 02:27:25 And I could just about swear that the last time I was setting it up I was irritated that I couldn't find any clear instructions on how to get kernel-power to play nicely with u-boot instead of multiboot, but regardless, it was there. May 03 02:28:38 look, multiboot is really literally reflashing each time you change the kernel you want to boot up May 03 02:28:44 I know. May 03 02:29:11 it of course can handle powerkernel as well May 03 02:29:33 As evidenced by the fact that it was. May 03 02:29:42 but this for sure isn't the prefered/suggested method, at least not by powerkernel devels May 03 02:30:25 now I bet your kenrel modules of backup didn't match stock kernel May 03 02:30:51 so after restoring those, you ran into bootloop May 03 02:31:02 As I noticed when I looked up the latest on it the other day, after having not messed with it other than to allow automatic updates since (I think) 47 or possibly 46 May 03 02:31:45 But now that I know that's the problem, I can most likely fix it. I have to get back on PK before I restore the backup. May 03 02:31:49 kernel update will most surely cause bootloop with multiboot May 03 02:32:21 :nod: May 03 02:32:26 The question is, do I have to get back specifically on 49, or will 50 work. Or am I totally screwed. May 03 02:32:59 you for sure have to use same kernel version May 03 02:33:05 50 won't work May 03 02:33:32 Ok. That's what I thought, but since getting 49 reinstalled is at least slightly more work, I figured I'd check. May 03 02:33:52 I guess just flashing KP49 will stop your bootloop May 03 02:35:15 though, OTOH you still have a multiboot infested system then, and probably need to reflash anyway to get a clean restart May 03 02:36:15 so I'd rather suggest flashing stock, then restore *without* /lib/modules and without /etc/*boot* May 03 02:36:33 O...kay...if you're right about that, that's even more retarded than I thought multiboot was... May 03 02:36:41 and without /bootmenu.sh May 03 02:37:35 I was under the impression that if you multibooted to the stock kernel and then uninstalled multiboot, it would in fact actually be gone. May 03 02:38:13 I never touched multiboot, so I can't tell for sure May 03 02:39:20 I guess I'll find out. May 03 02:39:25 I however know that no user ever succeeded to recover from multiboot caused boot problems the 'easy' way. Always involved reflash May 03 02:40:20 I was about to contradict that, but technically I guess all the boot problems from which I recovered easily using multiboot weren't exactly caused by multiboot. May 03 02:40:57 :-D May 03 02:41:06 They were caused by those meditating gurus. May 03 02:41:29 good luck anyway May 03 02:41:34 * DocScrutinizer is off now May 03 02:41:37 Much appreciated May 03 02:41:39 later May 03 02:41:42 yw **** ENDING LOGGING AT Thu May 03 03:00:01 2012