IRC log of #maemo-ssu for Saturday, 2013-09-21

*** _rd has quit IRC00:08
*** nox- has joined #maemo-ssu00:11
sixwheeledbeast~seen Pali00:17
infobotpali is currently on #maemo (12h 32m 47s) #harmattan (12h 32m 47s) #maemo-ssu (12h 32m 47s). Has said a total of 49 messages. Is idling for 2h 6m 52s, last said: 'DocScrutinizer05 ^^^^'.00:17
PaliI'm here00:17
sixwheeledbeasthi00:24
sixwheeledbeastso I have been fixing flopswap to use blkid and mount-ops-overwrite. works great on cssu-devel but I am not sure how to support older versions that don't support new mount-ops.00:28
DocScrutinizer05haha00:34
DocScrutinizer05clearly fell for the convenience trap00:35
sixwheeledbeastmhm00:48
sixwheeledbeastwhen I say "support", I mean how to tell flopswap to do "something else" because mount-ops-overwrite is not supported on this device.00:57
*** Vlad_on_the_road has quit IRC01:00
sixwheeledbeast"something else" being fallback to "$swap0" (device swap in blkid) at boot.01:01
*** NIN101 has quit IRC01:05
*** xes_ has joined #maemo-ssu01:23
*** xes has quit IRC01:25
DocScrutinizer05err, cehck for presence of some unique file?01:29
DocScrutinizer05if CSSU doesn't support a way to detect presence of the feature, I'd honestly open a ticket against that01:32
*** Martix_ has joined #maemo-ssu01:32
*** xes_ has quit IRC01:35
*** dhbiker has quit IRC01:35
*** xes has joined #maemo-ssu01:35
*** dhbiker has joined #maemo-ssu01:36
*** xes has quit IRC01:36
*** Pali has quit IRC01:40
DocScrutinizer05I mean, what worth is a feature when you don't provide a method to check for its presence01:44
DocScrutinizer05hostmode kernel exports a feature "hostmode"01:44
DocScrutinizer05I honestly hope thumb enabled kernel provides a feature "thumbified"01:45
*** Martix_ has quit IRC02:25
*** kolp has quit IRC02:42
*** n900-dk has quit IRC02:47
FatPhilpali: (and yes, I know you're not online right now) - I vaguely remember some failure-to-wake-up bug to do with the crypto accelleration, but that was probably in a n950 context02:52
FatPhiland I presume the slowness was due to things being chopped into too many too-small chunks, so that the overhead communicating with the security subsystem was too high. THat's just a guess.02:55
freemangordonDocScrutinizer05: for sure it does, kernel-errata-xxx-workaround02:56
DocScrutinizer05:-)02:56
FatPhilI think they got it working on the n9 eventually. (I seem to remember the patches went physically through my hands, and I'm pretty sure I remember who wrote them, and I sit at a desk next to him for 2 more weeks)02:56
DocScrutinizer05failure-to-wake-up? sounds like the secret SiErr in IMAP402:57
DocScrutinizer05OMAP even02:57
*** n900-dk has joined #maemo-ssu02:58
freemangordonguys, any idea what to check for if I get wake-up events on gpois, but no irqs?02:58
freemangordons/irqs/interrupts02:58
DocScrutinizer05what means "wake-up event"?02:59
freemangordon"WAKEUPEVENT: Wake-up event status for the pin."02:59
DocScrutinizer05no idea, can't associate it to a CPU status or event03:00
*** dos1 has quit IRC03:03
*** kolp has joined #maemo-ssu03:03
DocScrutinizer05when the GPIO is set up to create a IRQ, when the .... oooh WAKEUP, you mean that special type of IRQ that wakes up CPU from sleep.03:05
DocScrutinizer05that's kinda tricky I heard03:06
freemangordon:nod:03:06
FatPhilneeds to be in a wake-up domain03:06
freemangordonFatPhil: I guess it is if I got wakeupvenet bit set03:06
freemangordonwakeupevent even03:06
freemangordonhowever, the problem is that I don;t reeive ISP and SSI interrupts03:07
DocScrutinizer05since when CPU and particularly IRQ controller are powered down, a wake event will cause the power and clock management to ramp up those function blocks, but the IRQ controller not necessarily notices and fires an IRQ in parallel to the wake event03:07
freemangordonI checked the clocks in IDLEST status reg and they seem to be running03:07
DocScrutinizer05iirc03:07
FatPhiledge-triggered, or level-triggered?03:08
freemangordonhmm, not sure what the driver sets, but the same driver (SSI for example) works with 3.503:08
DocScrutinizer05I heard colleagues discussing exactly this problem for a HSI(?) interface, and I seem to recall they had to take special care in a wake-handler to make sure the IRQ handler gets triggered03:09
freemangordonDocScrutinizer05: there is some HW bug, but Iput the workaround in place03:10
DocScrutinizer05the IRQ controller doesn't do the edge detection by static means like gates etc, it actually "polls" the GPIOs to detect a state change03:10
DocScrutinizer05this doesn't work when the IRQ-controller is sleeping03:10
FatPhilwhich sounds broken, but a level-triggered interrupt might be a possible work-around?03:10
freemangordonI don;t believe IRQ controller is sleeping while ISP is awake03:10
DocScrutinizer05yes, level triggered is always the better choice03:11
DocScrutinizer05when you got a handler that's actually servicing the IRQ source03:11
freemangordonwait, wait, ISP and SSI irq are internal , not GPIO IRQs03:12
freemangordonwell, ISP for sure03:12
DocScrutinizer05doesn't make a difference, I guess03:12
DocScrutinizer05it's about the irq controller and its status03:12
DocScrutinizer05all I know for sure is: a wake event doesn't always cause an IRQ, even when the source is configured for both03:13
freemangordonwell, but it should cause at least some times :)03:14
DocScrutinizer05yes, when the triggering condition is met during SoC _not_ in suspend/sleep mode03:15
DocScrutinizer05IOW when no wake event happens03:15
DocScrutinizer05then IRQ happens03:15
freemangordonmhm. But I see NO interrupts. actually sensors (switches etc) gpio interrupts are raised03:16
DocScrutinizer05this MUST be documented somewhere in the intarwebs03:16
freemangordonbut no interrupts from ISP and SSI03:16
DocScrutinizer05it seems a ARM idiosyncrasy03:16
freemangordonhmm, is it possible that twl wakeup scripts to miss those power domains?03:17
DocScrutinizer05err can't parse that03:17
freemangordonanyway, I need to have some sleep :)03:18
freemangordonnight03:18
DocScrutinizer05n803:18
DocScrutinizer05can WAKE config even be set to edge-triggered at all?03:20
DocScrutinizer05I'd think for IRQ that are meant to WAKE, level trigger is mandatory03:21
*** jonwil has joined #maemo-ssu03:33
DocScrutinizer05hmmmmm   >>There are two ways of doing it currently: with or without DT.<<    https://www.kernel.org/doc/Documentation/gpio.txt03:33
*** kolp has quit IRC03:43
DocScrutinizer05WOW, >>only *edge*-triggered GPIOs are capable of generating  a module-level wakeup to the PRCM (c.f. 34xx NDA TRM section 25.5.3.2.)<<03:47
DocScrutinizer05> >In order to ensure that devices using level-triggered GPIOs a sinterrupts can also cause wakeups (e.g. from idle), this patch enable sedge-triggering for wakeup-enabled, level-triggered GPIOs when a GPI Obank is runtime-suspended (which also happens during idle.)<<03:48
DocScrutinizer05http://lists.infradead.org/pipermail/linux-arm-kernel/2012-March/087973.html03:49
DocScrutinizer05dunno if that helps to find the problem03:58
*** mkaindl has joined #maemo-ssu04:13
*** M4rtinK has quit IRC04:24
*** gregoa has quit IRC04:48
*** gregoa has joined #maemo-ssu04:49
*** nox- has quit IRC04:51
*** cybertheque has quit IRC05:52
*** amiconn has quit IRC05:54
*** amiconn_ has joined #maemo-ssu05:54
*** amiconn_ is now known as amiconn05:54
*** DrCode has joined #maemo-ssu07:27
*** oooaaaooo has joined #maemo-ssu07:31
oooaaaooohi all so im looking at http://wiki.maemo.org/User:Joerg_rw/tools#On_Device07:32
oooaaaoooand am wondering if i need the miniwrapper thing07:32
*** oooaaaooo has quit IRC08:04
*** dhbiker has quit IRC08:17
*** aap has quit IRC08:52
*** aap has joined #maemo-ssu09:10
*** aap has quit IRC09:19
*** sunny_s has joined #maemo-ssu09:53
sixwheeledbeast<<DocScrutinizer05: err, cehck for presence of some unique file?>> exactly, but "what". I have been comparing /etc/ on two devices and I can't see a unique file. only changes in files (mount-ops, rcS-late, fstab), and not something I think I could positively grep. unless I am missing something. I can only think of dpkg-query, but I don't know what package.10:13
*** _rd has joined #maemo-ssu10:15
*** Vlad_on_the_road has joined #maemo-ssu10:28
*** NIN101 has joined #maemo-ssu10:51
*** aap has joined #maemo-ssu10:52
*** Pali has joined #maemo-ssu10:52
*** kolp has joined #maemo-ssu10:58
*** aap has quit IRC11:06
*** aap has joined #maemo-ssu11:28
*** _rd has quit IRC11:42
*** _rd has joined #maemo-ssu11:42
sixwheeledbeastah it seems to be "upstart 0.3.8-68+0cssu5"11:44
sixwheeledbeastdpkg-query -W -f='${Version}\n' upstart12:01
*** Martix_ has joined #maemo-ssu12:03
*** sunny_s has quit IRC12:06
*** dafox has joined #maemo-ssu12:22
Paliyou can ask dpkg-query for upstart version and then use dpkg --compare-versions to check if upstart version is at least version which you need12:37
* sixwheeledbeast goes searching for dpkg --compare-versions, thanks12:42
*** oooaaaooo has joined #maemo-ssu12:53
oooaaaoooDocScrutinizer05: im trying to install less but apt-get says "Couldn't find package less" do i have to enable the extras repo to be able to follow http://wiki.maemo.org/User:Joerg_rw/tools12:54
DocScrutinizer05yes13:03
oooaaaooothnx13:03
DocScrutinizer05~pkg13:03
infobotfrom memory, pkg is http://maemo.org/packages/13:03
*** Vlad_on_the_road has quit IRC13:03
*** Vlad_on_the_road has joined #maemo-ssu13:06
*** Martix_ has quit IRC13:08
*** Vlad_on_the_road has quit IRC13:10
oooaaaoooDocScrutinizer05: how do i install find on the N900? aptitude says "Package: findutils13:13
oooaaaoooState: not a real package"13:13
*** M13 has joined #maemo-ssu13:16
*** Vlad_on_the_road has joined #maemo-ssu13:17
DocScrutinizer05findutils-gnu13:22
DocScrutinizer05nfc about aptitude13:23
DocScrutinizer05http://maemo.org/packages/view/findutils-gnu/13:23
DocScrutinizer05^^^ tells you it's in Fremantle Extras-devel free armel13:24
DocScrutinizer05aka "extras-devel catalog"13:24
*** _rd has quit IRC13:29
*** dafox has quit IRC13:34
*** dafox has joined #maemo-ssu13:35
oooaaaoooDocScrutinizer05: thnx13:35
oooaaaoooDocScrutinizer05: in your wiki entry, what is this all about: "# execute or source this to fix the missing end-times in eventslogger db" ?13:37
*** _rd has joined #maemo-ssu13:40
oooaaaoooDocScrutinizer05: ok so i now understand what that case switch at the beginning of your wiki wasa all about!13:41
oooaaaoooNokia-N900:~# apt-get autoremove13:41
oooaaaoooNONONO!! Don't do that!13:41
oooaaaooololz13:41
*** M4rtinK has joined #maemo-ssu13:41
jonwilpali: ping13:42
Palijonwil: pong13:43
oooaaaoooso if i get a "Ouch!  Got SIGSEGV, dying..13:46
oooaaaoooSegmentation fault" that usually means that my memory allocation isnt enough to run the program?13:46
DocScrutinizer05oooaaaooo: it's about what it says: fixing the timestamps in eventsdb, so tools like extended call logger can show duration of a call13:47
FatPhilwhat's the right channel for n950 questions?13:47
DocScrutinizer05ooabtw all this is rather OT in this channel13:47
DocScrutinizer05oooaaaooo: ^^^13:48
oooaaaoooDocScrutinizer05: ok13:48
DocScrutinizer05please take it to #maemo13:48
oooaaaoooDocScrutinizer05: ok13:48
*** bsdmaniak has joined #maemo-ssu14:16
*** xes has joined #maemo-ssu14:33
*** jonwil has quit IRC14:36
*** jonwil has joined #maemo-ssu14:37
*** M13 has quit IRC14:40
*** DrCode has quit IRC14:52
*** bsdmaniak has quit IRC14:56
*** DrCode has joined #maemo-ssu14:57
*** dos1 has joined #maemo-ssu15:02
*** bsdmaniak has joined #maemo-ssu15:04
*** BCMM has joined #maemo-ssu15:07
*** BCMM has quit IRC15:13
*** BCMM has joined #maemo-ssu15:14
*** BCMM has quit IRC15:14
*** DrCode has quit IRC15:36
*** Vlad_on_the_road has quit IRC15:39
*** DrCode has joined #maemo-ssu15:39
*** oooaaaooo has left #maemo-ssu16:26
*** BCMM has joined #maemo-ssu16:48
*** dos1 has quit IRC17:29
*** dos1 has joined #maemo-ssu17:29
*** _rd has quit IRC17:36
*** oldtopman has joined #maemo-ssu17:42
*** dafox has quit IRC18:14
*** bsdmaniak has quit IRC18:18
jonwilmerlin1991: ping18:43
jonwil~seen merlin199118:44
infobotmerlin1991 is currently on #maemo (10d 16h 4m 33s) #harmattan (10d 16h 4m 33s) #maemo-ssu (10d 16h 4m 33s). Has said a total of 5 messages. Is idling for 9d 22h 19m 39s, last said: '#neverdothatatall'.18:44
*** arcean has joined #maemo-ssu20:51
Palimerlin1991, freemangordon: because of gitorious problems, what about create mirrors on github?20:59
*** jonwil has quit IRC21:09
*** _rd has joined #maemo-ssu21:11
*** nox- has joined #maemo-ssu21:19
*** arcean has quit IRC21:23
*** arcean has joined #maemo-ssu21:24
*** _rd has quit IRC21:27
*** MohammadAG has quit IRC21:28
*** MohammadAG has joined #maemo-ssu21:29
*** mkaindl has left #maemo-ssu21:35
merlin1991jonwil pong21:37
merlin1991in case of reading backlog ;)21:37
*** mkaindl has joined #maemo-ssu21:47
*** _rd has joined #maemo-ssu22:01
Palimerlin1991: now when I should have permission, I will create cssu gitorious repositories...22:09
Palialso I have links for jonwils repo files22:10
merlin1991Pali: sure go ahead :)22:13
*** arcean has quit IRC22:15
*** arcean has joined #maemo-ssu22:15
*** mkaindl has left #maemo-ssu22:26
*** arcean has quit IRC22:50
*** Pali has quit IRC23:13
*** mkaindl has joined #maemo-ssu23:21
*** kolp has quit IRC23:39
*** NIN101 has quit IRC23:42
DocScrutinizer05FatPhil: ooops I completely missed your question - right channel for N950 Q is #harmattan23:46
*** _rd has quit IRC23:47
*** _rd has joined #maemo-ssu23:47
*** LauRoman has quit IRC23:50
*** Martix has joined #maemo-ssu23:52

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