**** BEGIN LOGGING AT Fri May 09 03:00:01 2014
May 09 05:39:03 <ds_shadof> is there any way to send files via skype from n900?
May 09 08:08:00 <wizbit> anybody know how to use sfdisk, its complicated compared to fdisk
May 09 08:21:07 <wizbit> start= 56631360, size=  4194304  hmmm i can convert gb into blocks, but how can i work out the start position?!
May 09 08:23:07 <wizbit> ok worked it out
**** BEGIN LOGGING AT Fri May 09 08:32:12 2014
May 09 11:03:06 <wizbit> sfdisk is complicated.
May 09 11:03:08 <wizbit> arghghghghghgh
May 09 11:21:31 <DocScrutinizer05> IroN900:~# apt-cache policy cfdisk
May 09 11:21:32 <DocScrutinizer05> cfdisk:
May 09 11:21:34 <DocScrutinizer05>   Installed: 2.12r-19
May 09 11:21:35 <DocScrutinizer05>   Candidate: 2.12r-19
May 09 11:21:37 <DocScrutinizer05>   Version table:
May 09 11:21:38 <DocScrutinizer05>  *** 2.12r-19 0
May 09 11:21:40 <DocScrutinizer05>         500 http://extras-devel.merlin1991.at fremantle/free Packages
May 09 11:26:38 <wizbit> DocScrutinizer05: excellent, i hope cfdisk is easier
May 09 11:37:44 * Maxdamantus likes parted.
May 09 11:38:03 <Maxdamantus> parted from Debian seems to have problems on the N900 though.
May 09 11:38:51 <Maxdamantus> root@orcus:/# parted 
May 09 11:38:51 <Maxdamantus> *** glibc detected *** parted: munmap_chunk(): invalid pointer: 0x4019f548 ***
May 09 11:39:20 <Maxdamantus> iirc, it runs under valgrind, but random characters are inserted to the command line.
May 09 11:39:56 <Maxdamantus> (parted)  (i-search)`': ¸(parted) ú¸[1@| ! þÿJþÿÿ þÿÿª
May 09 11:41:15 <Maxdamantus> have to kill it from another shell.
May 09 11:41:51 <Maxdamantus> Randomly generates commands and tries to evaluate them, but they're (hopefully) always invalid, so it prints the help text.
May 09 11:48:58 <wizbit> cfdisk doesnt recognize : cfdisk /dev/mmcblk1
May 09 11:49:06 <wizbit> FATAL ERROR: Bad primary partition 0: Partition ends in the final partial cylinder
May 09 12:02:12 <DocScrutinizer05> http://wstaw.org/m/2014/05/09/plasma-desktopx31889.png
May 09 12:03:31 <DocScrutinizer05> -z: Start with a zero partition table, instead of reading the pt from disk;
May 09 12:04:09 <wizbit> ok done it with -z option!
May 09 12:04:10 <DocScrutinizer05> beware to not mix up mmcblk1 with mmcblk0
May 09 12:04:35 <wizbit> DocScrutinizer05: swap is on /dev/mmcblk1p1
May 09 12:04:43 <wizbit> and extra vfat space is on /dev/mmcblk1p2
May 09 12:05:16 <wizbit> maybe swap should be on /dev/mmcblk1p2 
May 09 12:05:27 <DocScrutinizer05> system swaps names of mmcblk1 and mmcblk0 during boot
May 09 12:05:31 <DocScrutinizer05> sometimes
May 09 12:07:19 <DocScrutinizer05> but yeah, usually mmcblk0 is internal eMMC which you MUST NOT touch 
May 09 12:07:21 <DocScrutinizer05> ~2119
May 09 12:07:22 <infobot> The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED",  "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119.
May 09 12:07:30 <DocScrutinizer05> this time for real
May 09 12:07:41 <DocScrutinizer05> well
May 09 12:07:51 <DocScrutinizer05> but yeah, usually mmcblk0 is internal eMMC which you SHOULD NOT touch 
May 09 12:08:30 <wizbit> here we go
May 09 12:08:31 <wizbit> swapon.debian -p 0 /dev/mmcblk1p2
May 09 12:10:03 <wizbit> DocScrutinizer05: do i need to turn off the existing swap 
May 09 12:10:10 <wizbit> before i swapon the partition
May 09 12:10:24 <wizbit> im not touching eMMC
May 09 12:12:39 <DocScrutinizer05> no, you shouldn't turn of swap before turning on alternative swap. You are supposed to turn off "old" swap *after* new swap got enabled
May 09 12:15:40 <wizbit> ok
May 09 12:15:41 <wizbit> /dev/mmcblk1p2                          partition       1953112 0       -1
May 09 12:15:47 <wizbit> its worked
May 09 12:15:51 <wizbit> i wonder if this will take a reboot
May 09 12:16:33 <wizbit> or should they both be active with different priorities
May 09 12:18:50 <wizbit> hmm
May 09 12:18:58 <wizbit> after reboot it rememebers built in swap
May 09 12:18:59 <wizbit> /dev/mmcblk0p3                          partition       786424  0       -1
May 09 12:19:35 <wizbit> http://wiki.maemo.org/Startup_script  ?
May 09 12:21:57 <wizbit> i will put this in startup script
May 09 12:22:00 <wizbit> swapoff /dev/mmcblk0p3 && swapon /dev/mmcblk1p2
May 09 12:27:15 <DocScrutinizer05> WRONG
May 09 12:27:19 <DocScrutinizer05> see above
May 09 12:29:00 <DocScrutinizer05> what is your system supposed to do after swapoff? I tell you what it's going to do: it suspends all preloaded apps to free up RAM so system can still kinda work after swapoff
May 09 12:29:26 <DocScrutinizer05> this is a very pathological state for your system
May 09 12:30:20 <DocScrutinizer05> and will take quite a while (up to a few minutes) to complete, and only after it completed swapoff will return so swapon will add available virtual memory to system again
May 09 12:31:11 <DocScrutinizer05> kudos to hildon desktop for handling all this, other systems quite possibly would simply blow chunks
May 09 12:32:08 <DocScrutinizer05> swapon /dev/mmcblk1p2 && swapoff /dev/mmcblk0p3
May 09 12:32:14 <DocScrutinizer05> THIS makes perfect sense
May 09 12:32:21 <wizbit> eek my phone wont boot now :/
May 09 12:32:35 <DocScrutinizer05> I'm not surprised
May 09 12:32:40 <wizbit> shit :(
May 09 12:34:35 <wizbit> got it to boot phew
May 09 12:35:23 <wizbit> ok turn on the other swap BEFORE i turn off the other
May 09 12:36:15 <wizbit> DocScrutinizer05: i have created this startup-script inside /etc/event.d/
May 09 12:36:16 <wizbit> http://dpaste.com/123RR90/
May 09 12:36:22 <wizbit> does that look ok?
May 09 12:39:48 <wizbit> i added this to the script: echo 30 > /proc/sys/vm/swappiness
May 09 12:42:31 <DocScrutinizer05> wizbit: this script looks nonsensical since it's not linked to any other event during bootup
May 09 12:44:55 <wizbit> /etc/event.d/rcS-late
May 09 12:45:08 <DocScrutinizer05> I got no clue how upstart *really* works in detail, but if I had to geuss, i'd say such script gets started before or in parallel with any other non-conditional script, *very* early during boot, when neither mounting of original swap nor renaming of mmcblk1 vs 0 has already happened (possibly)
May 09 12:45:13 <wizbit> i changed the swap -a line to swapon /etc/event.d/rcS-late
May 09 12:45:23 <wizbit> swapon /dev/mmcblk1p2
May 09 12:45:24 <wizbit> i mean!
May 09 12:45:37 <DocScrutinizer05> that sounds *much* beter
May 09 12:45:54 <wizbit> it worked!!!
May 09 12:45:55 <wizbit> /dev/mmcblk1p2                          partition       1953112 0       -1
May 09 12:47:00 <wizbit> as long as i reboot the phone every now and again i dont have to worry about swap fragmentation
May 09 12:51:37 <DocScrutinizer05> meh. A "swapon /dev/mmcblk0p3 && swapoff /dev/mmcblk1p2 && swapon /dev/mmcblk1p2 && swapoff /dev/mmcblk0p3" every now and then will do as well
May 09 12:52:27 <wizbit> aye ok!
May 09 12:52:44 <DocScrutinizer05> or simply use any of the several reswap tools
May 09 12:52:47 <wizbit> could create a mini, swap-refresh.sh script
May 09 12:53:09 <DocScrutinizer05> that's what ereswap etc does
May 09 12:53:24 <wizbit> aye ok
May 09 12:53:35 <wizbit> i found another cool feature
May 09 12:53:51 <wizbit> my alarm clock went off on the n900 this morningm then it said what would you like to turn the phone on
May 09 12:53:54 <wizbit> how cool
May 09 12:53:59 <wizbit> it must of been in sleep mode
May 09 12:54:20 <wizbit> am i right in thinking that the alarm clock still works even in sleep mode?
May 09 12:54:32 <DocScrutinizer05> it even works in off mode
May 09 12:54:36 <wizbit> jeeeeze
May 09 12:54:43 <DocScrutinizer05> it's a true cmos clock alarm
May 09 12:54:43 <wizbit> but how? its off
May 09 12:54:51 <wizbit> cooool
May 09 12:56:11 <DocScrutinizer05> you could create a tool that powers up N900 once per month, does something, and goes back to powered down state after 5 minutes. This could run for a year or longer, depending on self discharge of battery
May 09 12:57:19 <DocScrutinizer05> alas a broken bupbat will cause cmos clock to forget any alarm that been set up, when you swap battery
May 09 12:57:33 <DocScrutinizer05> ~bupbat
May 09 12:57:33 <infobot> well, bupbat is use the capacitive type, LiIon are breaking during 12 months, or http://www.digikey.de/product-detail/de/PAS414HR-VG1/587-2157-1-ND/1959153, or https://hbe-shop.de/Art-2112777-TAIYO-YUDEN-PAS414HR-VA5R-KONDENSATOR006F-33V-STAKED-COIN, or http://talk.maemo.org/showthread.php?t=90864
May 09 12:58:14 <DocScrutinizer05> cya
May 09 12:58:17 <DocScrutinizer05> o/
May 09 15:37:16 <DocScrutinizer05> wizbit: why didn't you simply change the fstab?
May 09 15:46:33 <DocScrutinizer05> nevermind, I thought this nonsense got fixed meanwhile in CSSU
May 09 15:59:35 <DocScrutinizer05> I honestly never got the rationale behind doing this messy fstab "auto-config" on every boot, instead of doing it once in postinstall script of a better cleaned-up bootscripts package, and then only doing it MANUALLY whenever somebody changed partition MANUALLY
May 09 16:00:22 <DocScrutinizer05> and I really thought I conviced Pali and the rest of CSSU team, a maybe one year back
May 09 16:13:19 <kerio> DocScrutinizer05: you can disable fstab autogeneration
May 09 16:13:39 <kerio> it's enabled by default because of reasons
May 09 16:13:57 <DocScrutinizer05> I'm not interested in raisins
May 09 16:14:00 <DocScrutinizer05> ;-P
May 09 16:14:08 <kerio> but they're hysterical!
May 09 16:14:19 <DocScrutinizer05> me too!
May 09 16:14:45 <DocScrutinizer05> which reason?
May 09 16:18:11 * DocScrutinizer05 glares at /usr/lib/genfstab.awk
May 09 16:18:58 <DocScrutinizer05> braindead
May 09 16:20:04 <DocScrutinizer05> and /etc/event.d/rcS-late isn't any better tbh
May 09 16:20:44 <kerio> yeah, rcS-late is a fucking mess
May 09 16:21:18 <DocScrutinizer05> and it lacks any proper signaling via indicator LED to user, before "fsck.ext3 -y $HOME_DEV >> /var/lib/fsck_ext3_home.log 2>&1" - which by itself is also quite braindead, at least for logging to /var/lib if nothing else
May 09 16:22:49 <kerio> i think you said that already
May 09 16:24:01 <DocScrutinizer05> I wonder how hard it can be to add a few "echo $pattern >/sys/class/i2c-adapter/i2c-2/2-0032/engine1_mode" etc
May 09 16:28:09 * DocScrutinizer05 wonders if /dev/console would write to screen unconditionally, even during boot. Particularly s§>> /var/lib/fsck_ext3_home.log§tee -a  /var/lib/fsck_ext3_home.log >/dev/console§
May 09 16:30:02 <DocScrutinizer05> does your /var/lib/fsck_ext3_home.log hold *any* reasonable info?
May 09 16:32:41 <DocScrutinizer05> and who the heck ever deletes that file when it grows infinitely until rootfs explodes?
May 09 16:34:13 <DocScrutinizer05> nota bene >>, not >
May 09 16:34:51 <DocScrutinizer05> into /var/lib/ - ARRRRGHHH!
May 09 16:43:38 <DocScrutinizer05> LOL! epic!  http://privatepaste.com/e9462d7b32
May 09 16:48:09 <DocScrutinizer05> sorry WUT?  CSSU_VERSION="CSSU_VERSION=$(dpkg-query -W -f \${Version} mp-fremantle-community-pr 2>/dev/null || echo '<unknown>')"
May 09 16:48:48 <DocScrutinizer05> recursion: see <recursion>
May 09 16:52:16 <DocScrutinizer05> anyway obviously doesn't work, neither for similarly obscure KERNEL_VERSION="KERNEL_VERSION=$(uname -a 2>/dev/null || echo '<unknown>')"
May 09 16:54:38 <silviof> uuhhh
May 09 16:55:04 <silviof> from where is the shell script excert?
May 09 16:55:24 <DocScrutinizer05> oooh
May 09 16:55:26 <silviof> s/excert/excerpt/
May 09 16:55:27 <infobot> silviof meant: from where is the shell script excerpt?
May 09 16:55:43 <DocScrutinizer05> echo $OSSO_VERSION; echo $CSSU_VERSION; echo $KERNEL_VERSION ) > /sys/kernel/debug/panic_info_buff
May 09 16:56:24 <DocScrutinizer05> silviof: /etc/event.d/rcS-late
May 09 16:56:30 <DocScrutinizer05> CSSU
May 09 16:58:17 <DocScrutinizer05> gals and guys, I can tell you PR1.2 /etc/event.d/rcS-late looked QUITE different
May 09 16:58:32 <silviof> :-)
May 09 16:59:52 <DocScrutinizer05> well, actually RX-51_2009SE_3.2010.02-8_PR_MR0
May 09 17:11:31 <DocScrutinizer05> damn! vibrator output is bitbang-PWMed
May 09 17:12:18 <DocScrutinizer05> which is kinda nice and funny to listen and feel CPU clock frequency, but otherwise gives you pretty weird user experience
May 09 17:13:01 <DocScrutinizer05> real fun: echo 100 >/sys/class/leds/twl4030\:vibrator/brightness
May 09 17:13:21 <DocScrutinizer05> do that via wlan ssh on a locked&idle device
May 09 17:14:26 <DocScrutinizer05> NB that any vibra-kicj from MCE for touch event or keypress resets vibra to what would be a "echo 0 >/sys/class/leds/twl4030\:vibrator/brightness"
May 09 17:14:35 <DocScrutinizer05> kick*
May 09 17:16:00 <DocScrutinizer05> so all the weirdness you hear and feel when doing the above via ssh from remote is actually just system itself aka CPU going to clock down to zero and leaving output stale at either full power or zero
May 09 17:16:31 <DocScrutinizer05> while as long as CPU runs it does proper PWM and vibra runs at 40%
May 09 17:17:17 <DocScrutinizer05> thus you see vibra doing a pseudo-random pattern between idle, 40% and 100% operation
May 09 17:19:07 <DocScrutinizer05> I bet TI chip engineers didn't think about system wide zero-clock when designing the TWL4030 vibra output
May 09 17:20:24 <ShadowJK> I've experienced random prolonged haptic vibrate during extreme system load :)
May 09 17:21:27 <DocScrutinizer05> yeah, when MCE gets delayed then that's expected
May 09 17:22:01 <DocScrutinizer05> however PWM(!) stalling is sth I find a bit ... puzzling
May 09 17:22:23 <DocScrutinizer05> I even think PWM is done by a dedicated timer inside TWL4030, no?
May 09 17:23:29 <DocScrutinizer05> for this timer to go stale, I think the system wide clock-request signal shutting down the main oscillator for CPU (et al) clock is the only explanation
May 09 17:23:47 <DocScrutinizer05> (if there's actually such timer)
May 09 17:25:45 <bencoh> "vibrator/brightness" uuuh :]
May 09 17:25:50 <DocScrutinizer05> hehe
May 09 17:26:02 <DocScrutinizer05> yeah, for the kernel dudes all output is a LED
May 09 17:26:07 <bencoh> :)
May 09 17:26:13 <bencoh> LEDs everywhere
May 09 17:26:28 <DocScrutinizer05> they obviously have poor imaginative skills
May 09 17:26:52 <DocScrutinizer05> s./skills/power/
May 09 17:28:00 <DocScrutinizer05> the problem is: they use one driver for IO, and that driver been written and named with a LED in mind. Pathetic
May 09 17:30:56 <DocScrutinizer05> when you design a generic driver meant to get used as template for everything, you rather use a range of 0 .. 100.00% instead of 0 .. 255, and you write that to a sysnode named "power" rather than "brightness"
May 09 17:31:12 <bencoh> 0 .. 255 is fine with me :]
May 09 17:31:36 <bencoh> but "brightness" really is fun
May 09 17:31:40 <DocScrutinizer05> maybe with you, but not with me when the PWM hardware can do 16bit and not only 8bit
May 09 17:32:10 <DocScrutinizer05> generic range is 0% .. 100%
May 09 17:32:27 <DocScrutinizer05> not some fux0red 0..255
May 09 17:32:46 <DocScrutinizer05> of course you leave out the "%" in sysnode value
May 09 17:34:05 <bencoh> hmm true pwm can be in 16
May 09 17:34:57 <DocScrutinizer05> only valid alternatives where 0.0000 .. 1.00000000  and the other few commonly used metrics 0/00 and PPM
May 09 17:39:09 <DocScrutinizer05> bencoh: PWM also can be in 6, 10, 12, 15,...
May 09 17:40:06 <DocScrutinizer05> and for software-bitbanged PWM it can be any arbitrary resolution up to the theoretical limits of system timing precision
May 09 18:08:31 <DocScrutinizer05> cam somebody help me out with the name of the sysnode to control the kbd debug LEDs?
May 09 18:32:22 <ShadowJK> DocScrutinizer05; I wonder if there's lots of interrupt disable in mmc code, and I wonder if cpu can go sleep in middle of ongoing vibra event
May 09 18:32:39 <DocScrutinizer05> echo active > /sys/devices/platform/gpio-switch/sleep_ind/state
May 09 18:33:04 <DocScrutinizer05> will answer the 2nd question at least
May 09 18:34:53 <DocScrutinizer05> btw why mmc code?
May 09 18:39:30 <DocScrutinizer05> do: open kbd slider; (either via ssh from PC, or after a leading "sleep 10;) enter "echo active > /sys/devices/platform/gpio-switch/sleep_ind/state; echo 100 >/sys/class/leds/twl4030\:vibrator/brightness"; (if on device, hit <enter>) lock device with lock slider switch; (on ssh, hit <enter>)
May 09 18:39:42 <Kabouik> Hey guys, my (elder !) father just purchased a Samsung Rex 60. I don't know that phone, but it seems like a terrible choice at first look. Do you confirm? Is an Asha the best choice in that price range (50€) for someone who will not use internet or any smartphone feature? (Asked the same question on #Harmattan)
May 09 18:40:16 <DocScrutinizer05> no idea. Are you sure this is the right channel?
May 09 18:41:26 <Kabouik> Well it's kind of off-topic, but I think it's relevant to ask to people using the same phones as me, and generally acknowledging the pros good old phones with nice UIs
May 09 18:41:37 <Kabouik> of*
May 09 18:42:23 <DocScrutinizer05> hmm, maybe only few of us actually use anything else besides those good old phones
May 09 18:42:59 <DocScrutinizer05> I for one neither know Asha nor Rex60
May 09 18:44:02 <Kabouik> Ok, I just thought I'd ask because some among us TMO-users are really good at trying many phones. I don't know many IRC communities about phones so I just tried here
May 09 18:45:44 <DocScrutinizer05> fair enough
May 09 20:47:31 <DocScrutinizer05> about debug LEDs: left side is sy_sclkreq, right is nsleep
May 09 20:59:18 * DocScrutinizer05 wonders what exactly nSLEEP means
May 09 21:00:06 <DocScrutinizer05> sys_clkreq means something in the system needs proper high frequency clock
May 09 21:00:48 <DocScrutinizer05> the puzzling bit: nSLEEP is longer "active" than SYS_CLKREQ
May 09 21:02:01 <DocScrutinizer05> seems it's time to open the monster TRM again, eh?
May 09 21:02:42 * DocScrutinizer05 glares at the free memory on his machine and decides that it's not wise to try and open a ~4000p pdf
May 09 21:17:50 <DocScrutinizer05> ooh, nSLEEP[12] of twl4030 controls the regulators between max power and "economy mode"
May 09 21:22:32 <ShadowJK> Someone (a noob)  in #electronics has been building an smps the last 6 months, then discovered the twl4030 randomly, and concluded it makes every other smps obsolete
May 09 21:34:52 <DocScrutinizer05> haha
May 09 21:35:36 <DocScrutinizer05> I just read spruf98f 4.11.4.1.2 Device Off-Mode Transition Using Only the SYS_OFF_MODE Signal
May 09 21:35:53 <DocScrutinizer05> which is about nSLEEP1
May 09 21:36:33 <DocScrutinizer05> and I really don't get it why it seems like sysclock gets shut down while nSLEEP is still active
May 09 21:36:54 <DocScrutinizer05> this seems to be a nonsensical pathological system state
May 09 21:37:30 <DocScrutinizer05> or the schematics are incorrect and sys_clkreq is right side of kbd and nSLEEP is left side
May 09 21:38:07 <DocScrutinizer05> but hof the F* I'd check that without scope and other overkill in instrumentation
May 09 21:38:12 <DocScrutinizer05> how*
May 09 22:06:48 * DocScrutinizer05 also wonders of device off mode is actually the zeroclock used to reduce power consumption in N900
May 09 22:08:05 <DocScrutinizer05> >>The CORE power domain can be switched from ON to OFF or retention state.<< retention state sounds more like it
**** ENDING LOGGING AT Sat May 10 02:59:59 2014