**** BEGIN LOGGING AT Sun Jan 31 02:59:58 2016
Jan 31 04:46:09 <Hurrian> wow, so apparently... that Twitter update on Harmattan that breaks mp-harmattan-001-pr causes pkgmgrd to remove all the packages that depend on it
Jan 31 04:46:44 <Hurrian> jeeeeeeeeeeeeeeesus
Jan 31 11:04:10 <Pali> Luke-Jr: hi! about kdepim-noakonadi... if you have idea how to fix this bug https://bugs.kde.org/show_bug.cgi?id=166653 I would be happy...
Jan 31 11:04:12 <povbot> 04Bug 166653: was not found.
Jan 31 11:59:21 <freemangordon> Pali: hi!
Jan 31 11:59:30 <Pali> freemangordon: hi!
Jan 31 11:59:36 <freemangordon> could you rebase to 4.5-rc1 :)
Jan 31 11:59:51 <Pali> ok
Jan 31 12:22:57 <freemangordon> Pali: any idea how to track fs corruption?
Jan 31 12:23:04 <freemangordon> rootfs that is
**** BEGIN LOGGING AT Sun Jan 31 12:28:53 2016
Jan 31 12:42:30 <KotCzarny> darn splits, fmg: tracing io block access?
Jan 31 12:58:57 <freemangordon> KotCzarny: trace it where?
Jan 31 12:59:22 <freemangordon> also, the corruption happens when a reboot/shutdown is attempted
Jan 31 12:59:40 <KotCzarny> not on boot up?
Jan 31 12:59:52 <freemangordon> no
Jan 31 12:59:57 <KotCzarny> hmm
Jan 31 13:00:10 <freemangordon> everything works perfectly, until you try a reboot
Jan 31 13:00:28 <freemangordon> on next boot rootfs is unmountable
Jan 31 13:00:30 <KotCzarny> and it only happens with mainline?
Jan 31 13:01:31 <KotCzarny> is corruption random or just overwritten beginning of the nand?
Jan 31 13:08:52 <freemangordon> KotCzarny: mainline only
Jan 31 13:09:18 <freemangordon> no idea if it is random on on the beginning of nand, whatever "begining of nand" is
Jan 31 13:09:30 <KotCzarny> sector 0
Jan 31 13:09:38 <freemangordon> root is on mtd5, so it is hardly a beginning :)
Jan 31 13:09:48 <KotCzarny> hmm
Jan 31 13:11:10 <KotCzarny> did you try to add mount -o remount,ro /; sleep 10; before final reboot line ?
Jan 31 13:13:59 <freemangordon> no
Jan 31 13:14:28 <freemangordon> do you know where the final reboot line in maemo is?
Jan 31 13:14:52 <KotCzarny> nope, but i can look for it
Jan 31 13:15:19 <freemangordon> KotCzarny: you think something is not flushed?
Jan 31 13:15:47 <KotCzarny> hrm /etc/init.d/minireboot
Jan 31 13:15:54 <ceene2> KotCzarny: reboot should sync and umount everything on its own, shouldn't it?
Jan 31 13:16:01 <KotCzarny> it already does remount
Jan 31 13:16:30 <KotCzarny> ceene, what if nand controler is in the middle of the write when device reboots?
Jan 31 13:17:19 <ceene2> sync, remount or umount shouldn't exit until after the controller behind it has finished, be it nand, sata or whatever
Jan 31 13:18:02 <KotCzarny> 'shouldnt', but it happens
Jan 31 13:19:12 <ceene2> but i'll say that sync and mount are more than probably bug free
Jan 31 13:19:32 <KotCzarny> yes, thats why i suggested sleep
Jan 31 13:19:34 <ceene2> so then, we could think that it's the nand controller the one returning too soon from a sync()
Jan 31 13:19:48 <KotCzarny> so it has a chance to flush internal queues if it has them
Jan 31 13:19:54 <DocScrutinizer05> ~moo
Jan 31 13:19:54 <infobot> mooooooo! I am cow, hear me moo, I weigh twice as much as you. I'm a cow, eating grass, methane gas comes out my ass. I'm a cow, you are too; join us all! type apt-get moo.
Jan 31 13:20:08 <ceene2> at least it's worth trying it
Jan 31 13:20:17 <KotCzarny> fmg: try adding sleep before reboot in /etc/init.d/minireboot
Jan 31 13:20:44 <ceene2> i can't test it today, my dev n900 is at home and i've lost access to my building machine
Jan 31 13:20:55 <ceene2> thus the abscense of 'ceene' :)
Jan 31 13:20:56 <KotCzarny> also actually checking if remount worked wouldnt hurt either
Jan 31 13:21:20 <freemangordon> hmm, seems kernel partition got corrupted too
Jan 31 13:21:30 <DocScrutinizer05> o.O
Jan 31 13:21:34 <ceene2> kernel partition shouldn't be read only?
Jan 31 13:21:39 <freemangordon> it is RO
Jan 31 13:21:44 <KotCzarny> then its random
Jan 31 13:21:44 <freemangordon> afaik
Jan 31 13:21:48 <DocScrutinizer05> well no
Jan 31 13:21:58 <freemangordon> well, it is not mounted at all
Jan 31 13:21:58 <DocScrutinizer05> none of mtd? is ro per se
Jan 31 13:22:04 <DocScrutinizer05> yep
Jan 31 13:22:20 <freemangordon> so this is not really a FS corruption
Jan 31 13:22:30 <DocScrutinizer05> no, this sounds terrible
Jan 31 13:22:31 <KotCzarny> still, try that sleep
Jan 31 13:22:47 <KotCzarny> unless corruption happens even before reboot
Jan 31 13:22:51 <DocScrutinizer05> though I miss the OP, thanks to /ignore
Jan 31 13:22:58 <KotCzarny> but is just not triggered before reboot
Jan 31 13:23:03 <freemangordon> KotCzarny: this happens if I boot rescueos with upstream kernel as well
Jan 31 13:23:24 <KotCzarny> upstream == mainline ?
Jan 31 13:23:30 <freemangordon> yes
Jan 31 13:23:56 <freemangordon> anyway, now I prepared cramfs with a couple of mtd tools to try to debug further
Jan 31 13:24:12 <KotCzarny> even without mounting anything in rescueos or after mounting rootfs?
Jan 31 13:24:13 <DocScrutinizer05> freemangordon: what's the story?
Jan 31 13:24:38 <freemangordon> mainline kernel corrupts maemo footfs on reboot/shutdown
Jan 31 13:24:43 <freemangordon> *rootfs
Jan 31 13:24:46 <DocScrutinizer05> hihihi
Jan 31 13:24:50 <DocScrutinizer05> nice typo
Jan 31 13:24:54 <KotCzarny> fmg, more than rootfs apparently
Jan 31 13:24:56 <freemangordon> :)
Jan 31 13:25:03 <freemangordon> KotCzarny: yeah
Jan 31 13:25:14 <freemangordon> but this is visible immediately ;)
Jan 31 13:25:26 <DocScrutinizer05> mainline kernel maybe shuts down too fast?
Jan 31 13:25:37 <DocScrutinizer05> a page erase tages ages
Jan 31 13:25:42 <freemangordon> actually it hangs because of a bug in g_nokia
Jan 31 13:25:43 <DocScrutinizer05> takes*
Jan 31 13:26:03 <freemangordon> but even if it succeed to reboot, it makes no difference
Jan 31 13:26:29 <KotCzarny> fmg, then its happening even before starting to reboot
Jan 31 13:26:43 <KotCzarny> ie. even after mounting anything and doing some writes
Jan 31 13:26:54 <DocScrutinizer05> I guess OMAP TRM has something to say about correct system shutdown
Jan 31 13:26:59 <KotCzarny> you can try writing some patterns
Jan 31 13:27:13 <KotCzarny> then seeing if they are spotted in corrupted places
Jan 31 13:27:25 <KotCzarny> maybe you will see some regularity
Jan 31 13:27:32 <freemangordon> KotCzarny: I don't have tools to trace what happens on shutdown
Jan 31 13:28:00 <DocScrutinizer05> I also guess there's a reason why maemo proper takes some 10 to 30s to *actually* shut down. change your livelight pattern to see
Jan 31 13:28:01 <KotCzarny> fmg, not on shutdown, just start rescueos, write 1G of abcd repeated
Jan 31 13:28:22 <KotCzarny> then see if and where it starts showing (ie. in kernel partition and at which offsets)
Jan 31 13:28:41 <DocScrutinizer05> freemangordon: you still don't have serial console attached?
Jan 31 13:28:57 <freemangordon> DocScrutinizer05: no, I don;t have devel board
Jan 31 13:29:02 <freemangordon> this is a production device
Jan 31 13:29:08 <DocScrutinizer05> hmm
Jan 31 13:29:19 <KotCzarny> if what you say is true, then regular writes corrupt random parts of nand
Jan 31 13:29:20 <DocScrutinizer05> N900 I assume?
Jan 31 13:29:47 <freemangordon> and I don;t have the right tools to play with connectors under the battery
Jan 31 13:29:50 <freemangordon> n900, yes
Jan 31 13:30:18 <DocScrutinizer05> how about getting a devel N900 and solder some wires to the UART3?
Jan 31 13:30:20 <freemangordon> KotCzarny: or some buffer overflow which rewrites onenand
Jan 31 13:30:52 <freemangordon> DocScrutinizer05: I'd rather pester tmlind who has devel n900 on his rack
Jan 31 13:30:53 <KotCzarny> fmg, add printk to every nand write with debug info?
Jan 31 13:31:03 <freemangordon> KotCzarny: and read it how?
Jan 31 13:31:13 <DocScrutinizer05> freemangordon: why?
Jan 31 13:31:21 <freemangordon> why what?
Jan 31 13:31:25 <Pali> freemangordon, now fixing compile errors
Jan 31 13:31:31 <Pali> looke like there were more api changes
Jan 31 13:31:32 <KotCzarny> if you start writing with dd directly to partition, it should be increasing numbers
Jan 31 13:31:40 <DocScrutinizer05> what would you do with a package with a modded N900 on your door step?
Jan 31 13:31:51 <KotCzarny> if there will be anything with odd offsets, then you can spot the bug
Jan 31 13:32:00 <KotCzarny> also, network log ?
Jan 31 13:32:17 <freemangordon> DocScrutinizer05: I guess I'll start searching for a serial cable :)
Jan 31 13:32:41 <DocScrutinizer05> dang, it might even come with a serial<->USB adapter
Jan 31 13:33:06 <KotCzarny> also, wasnt there a debug usb gadget?
Jan 31 13:33:40 <freemangordon> hmm, a revert of a recent patch seems to have fixed that, lets do the full boot and check
Jan 31 13:41:18 <DocScrutinizer05> freemangordon: http://paste.opensuse.org/84665277
Jan 31 13:42:23 <DocScrutinizer05> IroN900:~# file /sbin/reboot
Jan 31 13:42:25 <DocScrutinizer05> /sbin/reboot: symbolic link to `/bin/busybox'
Jan 31 13:45:07 <DocScrutinizer05> http://paste.opensuse.org/94794124
Jan 31 13:46:59 <Pali> git rebase in automode moved omap_dispc_set_plane_ba0 function (needed for pvr) from omap2 dss fbdev driver to omapdrm
Jan 31 13:46:59 <DocScrutinizer05> funny:  >> umount -r /;   /bin/mount -n -o remount,ro /  <<
Jan 31 13:46:59 <Pali> :D
Jan 31 13:47:12 <Pali> maybe we should move pvr driver to omapdrm too
Jan 31 13:47:17 <Pali> freemangordon ^^^
Jan 31 13:48:39 <DocScrutinizer05> moin Pali
Jan 31 13:48:44 <Pali> hi!
Jan 31 14:04:37 <DocScrutinizer05> OT: Trump, Oh My Gosh! I wonder if USA is a political system with just 2, just one, or zero parties
Jan 31 14:05:34 <KotCzarny> usa is ruled by .il, anything in the media is just for the show
Jan 31 14:12:56 <Pali> freemangordon: v4.5-rc1-n900 is in linux-n900 tree
Jan 31 14:13:17 <Pali> after patching and fixing API changes I tested it in qemu and it is working here...
Jan 31 14:23:51 <Luke-Jr> Pali: no idea, sorry. I don't use local folders at all
Jan 31 14:24:20 <Pali> Luke-Jr: that is also for dimap or imap folders
Jan 31 14:24:43 <Pali> if you create search view
Jan 31 14:37:55 <DocScrutinizer05> Pali: you're still suffering this indexing mess? 
Jan 31 14:38:11 <Pali> DocScrutinizer05: in kmail?
Jan 31 14:38:14 <DocScrutinizer05> yup
Jan 31 14:38:21 <Pali> yes, on every startup
Jan 31 14:38:25 <DocScrutinizer05> dang
Jan 31 14:38:30 <Pali> if there is stored search view
Jan 31 14:38:35 <DocScrutinizer05> ooh
Jan 31 14:38:40 <Pali> if I delete search view, everything is OK
Jan 31 14:38:47 <DocScrutinizer05> makes sense, somehow
Jan 31 14:39:12 <Pali> $ rm ~/.kde/share/apps/kmail/search/*
Jan 31 14:39:14 <DocScrutinizer05> the search in kmail never was really usable
Jan 31 14:39:33 <Pali> this will delete all search views ^^^
Jan 31 14:39:40 <DocScrutinizer05> it was about felt factor 100 slower than a simple grep
Jan 31 14:39:53 <Pali> need to be called when kmail is not running...
Jan 31 14:40:22 <DocScrutinizer05> alas I'm still doomed from akonadi
Jan 31 14:41:27 <bencoh> what about mutt or claws?
Jan 31 14:41:56 <DocScrutinizer05> not like akonadi wouldn't have similar problems, but at least I can (and have to) suspend email indexing on every boot and every akonadi/kontact restart
Jan 31 14:44:54 <DocScrutinizer05> virtuoso eating gigabytes of RAM
Jan 31 14:53:49 <KotCzarny> bencoh, having own webmail is nice too (especially on dedicated machine)
Jan 31 14:56:59 <DocScrutinizer05> bencoh: I know there are lots of nice 'apps' like mutt or thunderbird, there are even better desktops than KDE. My problem is: I'm not willing to migrate, I got used to that stuff and tailored it to fit during 15 years
Jan 31 14:57:35 <DocScrutinizer05> I'm definitely not one of the users who install a pristine system every 6 months
Jan 31 14:58:34 <DocScrutinizer05> heck, my windows deco looks like 1990 ;-)
Jan 31 15:00:49 <bencoh> DocScrutinizer05: not telling you to install from scratch, just suggesting a solution that might fit your needs better ;)
Jan 31 15:00:52 <bencoh> but okay
Jan 31 15:04:02 <DocScrutinizer05> bencoh: I pondered it several times, but http://wstaw.org/m/2016/01/31/plasma-desktopsO3616.png among others
Jan 31 15:06:55 <DocScrutinizer05> not even starting to think about migrating a 500k mails
Jan 31 15:08:03 <DocScrutinizer05> each time I tried to do something like this, it ended in a 4 weeks disaster where I missed or lost important email stuff
Jan 31 15:10:25 <ceene2> why would you want to migrate half a million emails?
Jan 31 15:10:37 <ceene2> can't you just archive them and forget about them?
Jan 31 15:11:02 <KotCzarny> ceene, ever heard about hoarders?
Jan 31 15:11:15 <ceene2> :)
Jan 31 15:11:22 <ceene2> once i lost a hard drive due to malfunctioning
Jan 31 15:11:32 <ceene2> and i started to panic because i had lost hundreds of gigs
Jan 31 15:11:43 <ceene2> later on, i tried to remember what was there and i couldn't
Jan 31 15:11:50 <KotCzarny> ;)
Jan 31 15:11:56 <ceene2> so i decided that it wasn't that important in the end
Jan 31 15:12:04 <ceene2> i just keep a backup now of family photos and little else
Jan 31 15:12:24 <KotCzarny> i have different problem, i have many notes scattered over multiple drives/machines
Jan 31 15:13:48 <DocScrutinizer05> well, everything is in those emails, and archiving them seems a flawed approach since all metainfo like threads etc tends to get lost
Jan 31 15:14:36 <DocScrutinizer05> I however admit that kmail isn't up to handling my mails in the way it's supposed to be, anymore
Jan 31 15:14:57 <ceene2> yeah, but those threads aren't eternal
Jan 31 15:15:08 <ceene2> in a couple weeks, you won't have any of the old threads active anymore
Jan 31 15:15:14 <DocScrutinizer05> more often than not I have a hard time finding mails only a few weeks old, thanks to completely broken search function in kmail
Jan 31 15:16:48 <DocScrutinizer05> ceene2: when I want to check what a subcontractor answered to an inquiry I sent a year ago, I need threads when I only know and thus find my original inquiry but not exactly the answers
Jan 31 15:17:55 <DocScrutinizer05> browsing though a maildir with random nonsensical filenames with grep doesn't really help 
Jan 31 15:18:03 <ceene2> maybe you can set a previous archiving date
Jan 31 15:18:13 <ceene2> and keep last 365 days or so of email
Jan 31 15:18:19 <ceene2> but i'm sure you don't need anything from 2004 :)
Jan 31 15:19:54 <DocScrutinizer05> I think my active mail is from last 3 years, the rest already moved to a "archived" resp "old stuff" folder, however still in kmail. I exculded those dirs from friggin akonadi indexing, which somehow works most of the time
Jan 31 15:20:25 <Sicelo> akonadi is something like tracker on gnome?
Jan 31 15:20:32 <DocScrutinizer05> yes
Jan 31 15:21:00 <DocScrutinizer05> even worse: akonadi actually is a MTA backend
Jan 31 15:21:57 <DocScrutinizer05> the fucking idiotic thing gets started on KDE startup and starts polling my email accounts even when kmail/kontact isn't started at all >:-(
Jan 31 15:22:56 <DocScrutinizer05> it gets started just because. everything in KDE using akonadi now
Jan 31 15:23:33 <DocScrutinizer05> akonadi must have been invented by poettering
Jan 31 15:23:38 <DocScrutinizer05> ;-)
Jan 31 15:25:53 <DocScrutinizer05> http://wstaw.org/m/2016/01/31/plasma-desktopHb3616.png nepomuk virtuoso akonadi all one big bullshit
Jan 31 15:26:37 <DocScrutinizer05> NB you must not disable email indexing when you want even minimal search functionality in kmail
Jan 31 15:26:54 <ceene2> maybe you want to disable semantic desktop?
Jan 31 15:26:58 <ceene2> i don't even know what it does
Jan 31 15:27:03 <KotCzarny> ^^^ that's why i disable all such crap
Jan 31 15:27:09 <KotCzarny> and autoindexers
Jan 31 15:27:12 <ceene2> yeah, it all sounds just like trackerd :)
Jan 31 15:27:18 <DocScrutinizer05> when I do this, nothing works anymore
Jan 31 15:27:38 <ceene2> several years ago I also used kmail
Jan 31 15:27:44 <ceene2> but didn't have so much email, of course
Jan 31 15:27:46 <KotCzarny> i mostly use webmail
Jan 31 15:27:51 <ceene2> in the end... i just use gmail :/
Jan 31 15:27:54 <DocScrutinizer05> disabling sematic desktop implies disabling email indexing for example
Jan 31 15:28:16 <ceene2> it has its problems, but it also has the best searching capabilities, and labels are very nice
Jan 31 15:28:43 <KotCzarny> ceene, also, they have much better hardware
Jan 31 15:28:45 <KotCzarny> ;)
Jan 31 15:29:07 <DocScrutinizer05> no way. If I sell my ass to google, I don't do it for any less than google glass extreme
Jan 31 15:29:33 <DocScrutinizer05> and even then I wouldn't disclose my email to google
Jan 31 15:29:45 <ceene2> i can understand that
Jan 31 15:30:16 <DocScrutinizer05> I even replaced the standard search from google to https://ixquick.com
Jan 31 15:30:36 <ceene2> i'd be happy to have a gmail-like system on a server of mine 
Jan 31 15:30:46 <ceene2> but it doesn't exist, that i know of
Jan 31 15:31:21 <DocScrutinizer05> prolly will change from ixquick to https://startpage.com/
Jan 31 15:33:13 <sixwheeledbeast^> ddg.gg here
Jan 31 15:33:33 <KotCzarny> file:///usr/local/share/index.html here
Jan 31 15:33:39 <DocScrutinizer05> try startpage, it's great
Jan 31 15:34:57 <DocScrutinizer05> note the "proxy" option which allows to even visit result sites anonymously
Jan 31 15:35:40 <sixwheeledbeast^> the thing is startpage uses google services, ddg don't they are yandex bing etc
Jan 31 15:35:54 <DocScrutinizer05> sure, you "trust" the startpage company, but I do trust them more than I trust in google's "do no evil" which is only blabla
Jan 31 15:36:25 <sixwheeledbeast^> :nod:
Jan 31 15:36:39 <DocScrutinizer05> yes, startpage explicitly uses google, but that's the argument to use startpage instead of google
Jan 31 15:37:40 <DocScrutinizer05> if I want search results that are _not_ biased by google's point of view, I explicitly use a different search engine
Jan 31 15:37:53 <sixwheeledbeast^> I would prefer not to support Google at all, Youtube is prolly the only Google service I can think of I use.
Jan 31 15:38:07 <DocScrutinizer05> if you want "without google", use https://ixquick.com
Jan 31 15:38:17 <KotCzarny> swb: actually 'uses bing' is anti-cool imo ;)
Jan 31 15:39:14 <DocScrutinizer05> note how they are almost identical
Jan 31 15:39:32 <sixwheeledbeast^> anti-cool? meh! 
Jan 31 15:39:36 <DocScrutinizer05> the search results however differ
Jan 31 15:39:55 <DocScrutinizer05> hmm? anti-cool?
Jan 31 15:42:01 <DocScrutinizer05> https://startpage.com/eng/protect-privacy.html?hmb=1
Jan 31 15:46:14 <DocScrutinizer05> startpage makes money from google :-) I think this is a very reasonable business model
Jan 31 15:50:59 <freemangordon> Pali: great, thanks
Jan 31 16:02:23 <KotCzarny> wth
Jan 31 16:15:23 <Pali> freemangordon: we will need LD_PRELOAD library for sscd to use new /sys/class/gpio instead gpio_switch
Jan 31 16:15:42 <Pali> because nokia-modem.pm=0 will be probably removed :-(
Jan 31 16:47:43 <drathir> ddg rulez...
Jan 31 16:49:23 <drathir> have proxy too but by default anonymize search trough google is available throught tor too if needed...
Jan 31 17:57:46 <sunshavi> Hola
Jan 31 19:57:08 <freemangordon> Pali: yes, I know
Jan 31 19:58:17 <freemangordon> Pali: TBH it makes sense, no matter how much me and you dislike it :)
Jan 31 19:58:55 <Pali> I do not see reason for removing that support, now after long time from kernel
Jan 31 19:59:12 <freemangordon> Pali: the other option is to introduce an API to gpiolib that creates symlinks to exported gpios 
Jan 31 20:00:06 <freemangordon> Pali: because, by default you need patched kernel to export those gpios
Jan 31 20:00:23 <freemangordon> also, we can always unexport those, afaik
Jan 31 20:01:21 <freemangordon> Pali: also, I guess I can RE sscd
Jan 31 20:01:31 <Pali> sscd is big
Jan 31 20:01:43 <freemangordon> is it?
Jan 31 20:01:59 <Pali> LD_PRELOAD for sscd will be easier
Jan 31 20:02:04 <freemangordon> sure
Jan 31 20:02:30 <freemangordon> sscd is 100k
Jan 31 20:02:37 <freemangordon> yes, big one
Jan 31 20:03:38 <freemangordon> Pali: BTW, that onenand corruption happens if you mount in rescueos as well
Jan 31 20:04:14 <freemangordon> I wonder if it will happen in qemu too
Jan 31 22:25:30 <DocScrutinizer05> freemangordon: RE sscd sounds mad useful however
Jan 31 22:26:19 <DocScrutinizer05> it might be huge, but I think complexity should be rather low
Jan 31 22:29:28 <DocScrutinizer05> most stuff is sort of "chat (8)"
Jan 31 22:30:45 <DocScrutinizer05> or maybe that's even a layer above? And sscd simply implements the modem power-up and establishing of the logical channels via SSI?
Jan 31 22:39:51 <jonwil> A quick look shows that SSCD is making calls to a whole bunch of functions in libisi
Jan 31 22:40:15 <jonwil> so its obviously sending and receiving isi packets
Jan 31 22:44:44 <jonwil> and a lot of em too
**** ENDING LOGGING AT Mon Feb 01 02:59:58 2016