FlatPhil | I spend most of my time in a shell, my keymapping is about as awesome as I could hope for. | 00:03 |
---|---|---|
*** FlatPhil has quit IRC | 00:06 | |
DocScrutinizer05 | then I dunno how to help. I don't see how irssi would use a keymap different to the shell's one | 00:09 |
DocScrutinizer05 | maybe due to me never having used irssi | 00:09 |
DocScrutinizer05 | please everybody involved in FPTF post or query me your full name, I'll eventually edit #2 of http://talk.maemo.org/showpost.php?p=1372710&postcount=2 to list who's contributing | 00:11 |
FatPhil | nope, irssi was fine, it was "communi" that was ignoring the keymap | 00:15 |
sixwheeledbeast | re keymappings >> http://wiki.maemo.org/Remapping_keyboard/user_vis_awesome_kbdmapping | 00:16 |
dos1 | DocScrutinizer05: Sebastian Krzyszkowiak, commited to FSO integration | 00:16 |
*** amiconn has quit IRC | 00:38 | |
*** amiconn has joined #maemo-ssu | 00:40 | |
kerio | ~vi-kbd | 00:53 |
infobot | vi-kbd is, like, http://wiki.maemo.org/Remapping_keyboard/user_vis_awesome_kbdmapping | 00:53 |
kerio | and yeah, qt will still happily ignore your actual keymap | 00:53 |
kerio | which is a giant pile of WTF really | 00:53 |
kerio | it's what happens when a UI toolkit wants to reinvent everything | 00:54 |
*** M4rtinK has joined #maemo-ssu | 01:13 | |
*** NIN101 has quit IRC | 01:28 | |
*** Pali has quit IRC | 01:31 | |
*** lansiir has joined #maemo-ssu | 01:50 | |
*** oldtopman has quit IRC | 01:52 | |
*** discopig has quit IRC | 02:07 | |
*** discopig has joined #maemo-ssu | 02:07 | |
*** dafox has quit IRC | 02:29 | |
*** dos1 has quit IRC | 02:42 | |
*** LauRoman has quit IRC | 02:58 | |
*** sunny_s has quit IRC | 03:25 | |
*** M4rtinK has quit IRC | 04:07 | |
*** nox- has quit IRC | 04:48 | |
*** amiconn has quit IRC | 05:11 | |
*** amiconn_ has joined #maemo-ssu | 05:11 | |
*** amiconn_ is now known as amiconn | 05:11 | |
*** jonwil has joined #maemo-ssu | 05:20 | |
*** iDont has joined #maemo-ssu | 05:22 | |
jonwil | hi | 05:24 |
*** iDont has quit IRC | 05:34 | |
*** LaoLang_cool has joined #maemo-ssu | 06:08 | |
*** LaoLang_cool has quit IRC | 06:12 | |
*** LaoLang_cool has joined #maemo-ssu | 06:12 | |
*** LaoLang_cool has quit IRC | 06:23 | |
*** lansiir has quit IRC | 08:10 | |
*** FlatPhil has joined #maemo-ssu | 08:36 | |
freemangordon | merlin1991: (and others) what to do with this: https://admin.fedoraproject.org/updates/FEDORA-2013-15995/telepathy-gabble-0.18.1-1.fc19 ? | 09:41 |
*** LauRoman has joined #maemo-ssu | 09:46 | |
freemangordon | merlin1991: please pull http://maemo.gitorious.org/maemo-af/freemangordons-ke-recv-extra/source/24307e38817e801f91cbd2165c3c08d618b0edab: in CSSU | 09:48 |
*** Pali has joined #maemo-ssu | 10:06 | |
FlatPhil | is there a version of "ar" in one of the repos? | 10:13 |
freemangordon | FatPhil: isn't that part of the toolchain? | 10:18 |
jon_y | is part of binutils | 10:18 |
freemangordon | FatPhil: http://repository.maemo.org/pool/fremantle/free/b/binutils/ | 10:19 |
freemangordon | FlatPhil: but be warned - if you install gcc/binutils without making sure it will install on /opt, you'll have your / left with no free space | 10:20 |
freemangordon | Pali: did you try ke-recv-extras? | 10:23 |
Pali | installed on kp52 and now working | 10:23 |
jon_y | alternatively extract it yourself | 10:23 |
jon_y | oh wait | 10:23 |
jon_y | deb archive no ar around | 10:23 |
Pali | did not see any problems yet | 10:24 |
freemangordon | Pali: ke-recv has problems on 3.x kernels, who's going to fix that? | 10:24 |
freemangordon | Pali: also, I think we should make that more "organized" :) | 10:24 |
Pali | now I'm going to prepare upstream kernel patches for board chargering code | 10:25 |
FlatPhil | would worked around the problem, so don't need it | 10:25 |
freemangordon | FlatPhil: how's microb on cssu-thumb? | 10:26 |
FlatPhil | no more braaindead than it ever was | 10:26 |
freemangordon | can't parse | 10:26 |
FlatPhil | on the whole, my first impressions of cssu thumb are positive | 10:27 |
FlatPhil | well, microb sucks. | 10:27 |
freemangordon | well, I disagree, microb is the best mobile browser so far. Did you try the browser on Harmattan? I guess you did. | 10:28 |
Pali | remove adblock plugin | 10:28 |
FlatPhil | as does basically any program that has things like uttons | 10:28 |
freemangordon | Pali: this is freshly flashed device iiuc | 10:28 |
Pali | ok | 10:28 |
FlatPhil | almost evverything about harmatton sucks, even the shell sucks | 10:29 |
freemangordon | :nod: :( | 10:29 |
freemangordon | FlatPhil: I have 2 n950s sitting in the drawer, I gave one of them to my GF, she used it for couple of months, and she gave up. now she is happy n900 user :D | 10:31 |
FlatPhil | xchat sucks... gonna fid a working irc client that doesn't waste battery when it should be doing nothing... | 10:31 |
jon_y | n900 was better than n950? | 10:31 |
FlatPhil | i have an n950 made fron the spare parts of 3 or 4 broken ones | 10:31 |
FlatPhil | my g/f has an offiial developer n950 | 10:32 |
freemangordon | jon_y: n900 *IS* better than N9/50 | 10:32 |
jon_y | :| | 10:32 |
freemangordon | or even goes circles around it IMO | 10:32 |
*** FlatPhil has quit IRC | 10:32 | |
jon_y | I wish I knew anybody using an n9, or n900 or n950 | 10:32 |
jon_y | they cannot comprehend the openness | 10:33 |
freemangordon | I know 6 people (including me) | 10:33 |
freemangordon | all of them use n900 | 10:33 |
*** Fat_Phil has joined #maemo-ssu | 10:34 | |
jon_y | everyone around uses an iphone or samsung | 10:34 |
freemangordon | jon_y: and know about 2 more (again n900) | 10:34 |
jon_y | and then there is that one guy with blackberry | 10:34 |
Fat_Phil | some people seem very happy with then. The n900 fans are a minority. | 10:34 |
jon_y | whats wrong with the n950? | 10:35 |
jon_y | I would cont myself as an n900 fanatic | 10:35 |
Fat_Phil | I don't view the n900 as a phone, I view it as a computer. | 10:35 |
freemangordon | jon_y: everything. from HW to SW to capacitive screen etc | 10:35 |
jon_y | computer with gsm capability | 10:35 |
freemangordon | FatPhil: sure, but you can happily use it as phone as well | 10:36 |
jon_y | freemangordon: it wasn't as responsive as the n900? | 10:36 |
kerio | Fat_Phil: how does xchat waste battery? | 10:36 |
Fat_Phil | anyone saying positive things about the n900s resistive screen makes me happy | 10:36 |
jon_y | Fat_Phil: try irssi? it runs in a console | 10:36 |
jon_y | you could ssh to somewhere to use it | 10:37 |
Fat_Phil | kerio: it doesn't. I just got in a mode where I was typing white on white. | 10:37 |
jon_y | iirc ssh doesn't do much idle traffic | 10:37 |
freemangordon | jon_y: n900 on cssu-thumb OC to 805 and swap on uSD is 3-4 and even more times faster for most of the things. HAM excluded ofc :D | 10:37 |
jon_y | whoa, I should move my swap | 10:37 |
jon_y | I have a class 10 SD | 10:37 |
freemangordon | don;t forget to tweak your VM settings | 10:37 |
jon_y | any guides? | 10:38 |
freemangordon | or better - just use swappolube | 10:38 |
jon_y | ~swappolube | 10:38 |
freemangordon | I never used it, but afaik it sets VM to a way that is good for most stuff | 10:38 |
Fat_Phil | jon_y still keeping irssi as a possibility. I think the sterminals themselves are the wakeup cause. Need powertop to verify. | 10:38 |
freemangordon | ~ping | 10:39 |
infobot | ~pong | 10:39 |
Fat_Phil | s/sterm/xterm | 10:39 |
freemangordon | ~swappolube | 10:39 |
jon_y | infobot: swappolube is http://wiki.maemo.org/Swappolube | 10:39 |
infobot | jon_y: okay | 10:39 |
freemangordon | ~swappolube | 10:39 |
infobot | it has been said that swappolube is http://wiki.maemo.org/Swappolube | 10:39 |
freemangordon | :) | 10:39 |
freemangordon | jon_y: make sure you set page-cluster to 0 and swappiness to 30 | 10:40 |
jon_y | freemangordon: btw, just manual swapon/swapoff? | 10:40 |
jon_y | if I want to move my swap to SD | 10:40 |
Fat_Phil | is there something lke the desktop command execution widget that will run one of two different arbitrary commands in order to toggle between 2 states? | 10:41 |
freemangordon | I put a script in /etc/event.d as I am lazy guy, do you want it? | 10:41 |
Fat_Phil | shit, I need a boat | 10:41 |
jon_y | sure | 10:41 |
freemangordon | jon_y: http://pastebin.com/2z8xi0pm | 10:42 |
freemangordon | jon_y: if you use that script, you don't need swappolube | 10:42 |
jon_y | thanks | 10:42 |
Fat_Phil | of course I can just write a stateful script for DCEW to run, but it tracking the state would be good. | 10:43 |
*** sunny_s has joined #maemo-ssu | 10:43 | |
Fat_Phil | doh, that's what my output is, as text. who needs colours | 10:43 |
freemangordon | jon_y: also, for some reason swapoff command doesn't work, I guess it needs sleep 5 before it, but CBA to check what is going on | 10:43 |
jon_y | ok, assuming /dev/mmcblk1p3 is swap | 10:44 |
jon_y | I need to check my SD for space | 10:44 |
freemangordon | yeah | 10:44 |
*** kerio has quit IRC | 10:44 | |
jon_y | also, is there an unmount event for this? | 10:44 |
jon_y | eg when I want to remove back cover | 10:44 |
freemangordon | jon_y: unmount of what? | 10:44 |
freemangordon | ooh, no, don;t remove backcover if you have swap on uSD :D | 10:45 |
jon_y | the n900 dismounts the card when back cover is ope right? | 10:45 |
sixwheeledbeast | jon_y: You "could" try my FlopSwap. if you have stock filesystem, it's aimed for "users" | 10:45 |
jon_y | oh ok | 10:45 |
jon_y | actually that makes sense | 10:45 |
jon_y | ~FlopSwap | 10:45 |
sixwheeledbeast | http://wiki.maemo.org/Flopswap | 10:45 |
jon_y | infobot: FlopSwap is http://wiki.maemo.org/Flopswap | 10:45 |
infobot | okay, jon_y | 10:45 |
sixwheeledbeast | It basically 4 scripts called from a Qt GUI, but the script are designed to flip/flop between two uSD swaps | 10:47 |
*** scoobertron has quit IRC | 10:51 | |
freemangordon | Pali: BTW lkml still moderates my emails, any clue? | 10:53 |
Pali | lkml has no moderators | 10:53 |
freemangordon | shall I change to gmail? | 10:53 |
jon_y | ask lkml admins? | 10:53 |
*** Fat_Phil has quit IRC | 10:54 | |
freemangordon | hmm, it is not lkml, but linux-arm-kernel-owner@lists.infradead.org | 11:00 |
freemangordon | merlin1991: ping | 11:01 |
*** Vlad_on_the_road has joined #maemo-ssu | 11:02 | |
Pali | yes, linux-arm is moderated | 11:08 |
freemangordon | Pali: are you going to fix ke-recv? (for 3.x kernels that is) | 11:10 |
freemangordon | I want to see what is the problem with camera | 11:10 |
Pali | not now, preparing patches | 11:10 |
freemangordon | ooh, sure not now :) | 11:10 |
freemangordon | I ask in principal | 11:10 |
Pali | I belive I find time for it | 11:10 |
freemangordon | ok | 11:10 |
freemangordon | Pali: I'd prefer you to do it, as you know the code better | 11:11 |
Pali | ok | 11:11 |
Pali | freemangordon, maybe you should start working with front webcam | 11:11 |
Pali | because for it there is already kernel driver in upstream | 11:11 |
freemangordon | Pali: any particular reason? | 11:11 |
freemangordon | aah, ok | 11:11 |
Pali | I got some board data (are in v3.10-n900 branch) from camera driver developer | 11:12 |
freemangordon | Pali: BTW I see some errors from camera driver(on boot), going to check what happens | 11:13 |
Pali | freemangordon, to activate camera on 3.x kernels you need to set correct configuration | 11:14 |
Pali | you need for it program mediactl | 11:14 |
Pali | here is paste of args which I used: http://pastebin.com/8GD7RceQ | 11:14 |
freemangordon | "et8ek8 3-003e: invalid register list et8ek8-0002.bin, no POWERON mode found" | 11:15 |
freemangordon | Pali: our firmware seems wrong/dated | 11:15 |
Pali | "vs6555 binner" is front webcam and "et8ek8 3-003e" is back camera | 11:15 |
Pali | freemangordon: yes for et8ek8 you need new firmware | 11:15 |
Pali | vs6555 (from webcam) not using any firmware anymore | 11:15 |
Pali | compatible firmware for et8ek8 is here: https://gitorious.org/omap3camera/camera-firmware | 11:16 |
freemangordon | Pali: I'd rather fix the driver to workr with the old firmware | 11:16 |
Pali | not possible | 11:16 |
Pali | due to changes in kernel structrues | 11:16 |
freemangordon | as: "init: omap3camd0 main process (1527) terminated with status 1" | 11:16 |
Pali | it is not "firmware", but C kernel structure | 11:17 |
freemangordon | Pali: I understand that, but my point is that most probably oma3camd is using it too | 11:17 |
Pali | and struct something {} was changed somewhere after 2.6.28, so it is not possible | 11:17 |
freemangordon | if we can't use omapcam3d, we can throw the whole thing in the trash | 11:18 |
Pali | so you cannot use omap3camd without fixing and recompiling | 11:18 |
freemangordon | all Noka image processing and such is there | 11:18 |
freemangordon | *Nokia | 11:18 |
freemangordon | omap3camd is closed source | 11:18 |
Pali | someting was moded to kernel drivers | 11:18 |
freemangordon | yes, and I am going to find what and rever/make it compatible | 11:18 |
freemangordon | already did similar to DSP driver ;) | 11:19 |
freemangordon | Pali: if we plan to use the same binaries, we'll have to learn to live with such hacks I guess :) | 11:19 |
Pali | but that et8ek8 driver in 3.x tree is not API/ABI compatible with 2.6.28, because full omap video/camera api was changed | 11:19 |
freemangordon | oh | 11:20 |
Pali | reason why you need mediactl | 11:20 |
Pali | forgot about modifing 2.6.28 omap code for 3.x | 11:20 |
Pali | it is very very different | 11:20 |
Pali | rather try to fix upstream driver (or board data) for front webcam | 11:21 |
freemangordon | Pali: I'd rather modify 3.x code to be compatible with omap3camd. if possible | 11:21 |
freemangordon | both cameras use omap3camd | 11:21 |
Pali | et8ek8 is broken also because of kernel changes... | 11:21 |
freemangordon | Pali: so, what we do then? | 11:22 |
Pali | and it is forward-ported from 2.6.37 meego kernel | 11:22 |
Pali | [10:21:04] <Pali> rather try to fix upstream driver (or board data) for front webcam | 11:22 |
Pali | and then look at et8ek8 | 11:22 |
freemangordon | but what about backcam? | 11:22 |
freemangordon | aah, ok :) | 11:22 |
Pali | front webcam should be easier | 11:22 |
freemangordon | ok | 11:22 |
Pali | because of existence of driver | 11:23 |
freemangordon | ok, got it | 11:23 |
*** scoobertron has joined #maemo-ssu | 11:33 | |
*** discopig has quit IRC | 11:36 | |
*** discopig has joined #maemo-ssu | 11:36 | |
*** discopig has quit IRC | 11:41 | |
*** discopig has joined #maemo-ssu | 11:42 | |
freemangordon | Pali: iiuc, the problem is the change from EINVAL to ENOTTY in v4l2 code. I wonder which is the best way to tacle that. Change the code in v4l2 or make a LD_PRELOAD ioctl wrapper. I guess the former is better | 11:47 |
freemangordon | sorry, the latter | 11:47 |
freemangordon | i sbetter | 11:47 |
Pali | maybe you can use mplayer for testing camera | 11:48 |
Pali | and killing omap3camd for front webcam | 11:48 |
freemangordon | Pali: why do you hate omap3camd so much? :D | 11:49 |
Pali | because upstream driver for front webcam is totally different as that in 2.6.28 | 11:49 |
Pali | for front webcam it will not work | 11:49 |
freemangordon | iiuc omapcam3d use v4l2 API | 11:49 |
Pali | but some private omap ioctls | 11:52 |
Pali | and changing some memory via /sys | 11:52 |
Pali | this really can break driver to work | 11:53 |
Pali | freemangordon, patches sent | 11:55 |
freemangordon | Pali: great | 11:56 |
freemangordon | Pali: however, I am lazy, so I will try to go via the easy route :P | 11:57 |
freemangordon | (use omapcam3d) | 11:57 |
Pali | from http://elinux.org/N900 missing: hsi/ssi stuf (sre working on it), thumb2 (patches sent), bq2415x board data (now I sent), front webcam board data (you are woking on it), board data for ADP 1653 Flash Torch (patche sent, waiting for responce) | 11:59 |
freemangordon | Pali: I am working on front webcam board data? Am I? I was under the impression it is there. Isn't it? | 12:00 |
Pali | and then totally missing: back camera, AD5820 Autofocus, BCM2048 (Bluetooth and FM/RDS receiver) and OMAP3 ROM Random Number Generator support (will resent patches after thumb2 errata will be accepted) | 12:01 |
Pali | freemangordon: yes there is, but not working :-) | 12:01 |
freemangordon | oh, ok, so first I should fix the kernel I guess :D | 12:02 |
Pali | yes :D | 12:02 |
Pali | ^^^^ this is what missing in 3.11 kernel | 12:02 |
freemangordon | I thought vs6555 driver is loaded and works | 12:02 |
*** discopig has quit IRC | 12:03 | |
*** discopig has joined #maemo-ssu | 12:03 | |
*** FlatPhil has joined #maemo-ssu | 12:11 | |
FlatPhil | yay - just wrote my first scipt for DCEW, which totally ROCKS! http://fatphil.org/linux/maemo/DCEW/longblue.sh | 12:12 |
FlatPhil | How did I live without DCEW?!?!? | 12:12 |
FlatPhil | oh bollocks, sync issues, broken version uploaded | 12:14 |
FlatPhil | but you get the idea | 12:14 |
FlatPhil | fixed | 12:17 |
*** kerio has joined #maemo-ssu | 12:19 | |
*** LauRoman has quit IRC | 12:22 | |
Pali | FlatPhil: this is possible to configure via ukeyboard package (it replace text insert menu in settings application) | 12:23 |
*** FlatPhil has quit IRC | 12:26 | |
*** kerio has quit IRC | 12:28 | |
*** kerio has joined #maemo-ssu | 12:29 | |
*** kerio has quit IRC | 12:32 | |
*** kerio has joined #maemo-ssu | 12:32 | |
*** NIN101 has joined #maemo-ssu | 12:38 | |
*** NIN101 has quit IRC | 13:02 | |
*** NIN101 has joined #maemo-ssu | 13:03 | |
*** dafox has joined #maemo-ssu | 13:13 | |
jonwil | I think we need a Neo900 wiki (or some wiki pages) | 13:16 |
*** NIN101 has quit IRC | 13:19 | |
*** NIN101 has joined #maemo-ssu | 13:19 | |
*** M4rtinK has joined #maemo-ssu | 13:20 | |
*** NIN101 has quit IRC | 13:23 | |
*** dafox has quit IRC | 13:45 | |
*** dafox has joined #maemo-ssu | 13:53 | |
FatPhil | Pali: I got the idea from a package that was a 20 line python script, but I'm not putting python on this device, so just lifted the important line from the source package. | 13:59 |
FatPhil | Pali: what mailing list(s) are your kernel patches sent to? I'll do some code review. I can probably call it "upstream work' | 14:00 |
FatPhil | and get ${EMPLOYER} to pay me to do it! | 14:00 |
*** NIN101 has joined #maemo-ssu | 14:09 | |
Pali | FatPhil: every patch I CCing to linux-kernel@vger.kernel.org so it appear on https://lkml.org/ archive | 14:18 |
Pali | and mailinglists and email addresses where I sending patches comes from script ./scripts/get_maintainer.pl | 14:19 |
*** dos1 has joined #maemo-ssu | 14:25 | |
FatPhil | Pali: thanks. I was just wondering if you sent them to some arch-specific lists too. I'll hunt them out on LKML though. | 14:43 |
Pali | I sent them also to linux-omap | 14:45 |
FatPhil | Pali: issue with battery on cssu-thumb. I was somewhere below 55% one minute, and suddenly I was at 0%, recieving the "pew" tones. | 14:45 |
Pali | FatPhil: are you using BME or my kernel drivers? | 14:46 |
freemangordon | FatPhil: how old is that battery? | 14:46 |
FatPhil | freshly flashed cssu-thunm, almost no tweaking | 14:46 |
FatPhil | Oooooooold battery | 14:46 |
freemangordon | any other questions? :P | 14:47 |
FatPhil | IT's back at 36% again though | 14:47 |
FatPhil | after an hour of me just leaving it alone | 14:47 |
Pali | now you see why I shred BME binaries :D | 14:47 |
FatPhil | I've never seen it on my daily device with the stock nokia kernel | 14:47 |
freemangordon | FatPhil: yes, because the internal resistance makes voltage to drop under heavy load | 14:47 |
freemangordon | FatPhil: buy a new battery | 14:48 |
FatPhil | well, I wouldn't say it was under heavy load, I was just chatting over wifi, and editing scripts | 14:49 |
FatPhil | screen draining more than anything else put together probably. | 14:49 |
FatPhil | I'll try it with different batteries though, and see if there's a correlation. | 14:49 |
freemangordon | FatPhil: trust me, it is the battery | 14:50 |
FatPhil | I am 99% sure you're right. | 14:50 |
*** _rd has joined #maemo-ssu | 15:25 | |
freemangordon | Pali: WTF am I supposed to do with that ?!? http://pastebin.com/9uu4kmvR | 15:25 |
Pali | I sent mediactl commands | 15:25 |
Pali | on previous pastebin | 15:25 |
freemangordon | Pali: BTW how did you manage to compile media-ctl, I copied the missing headers. | 15:26 |
Pali | do not remember | 15:26 |
Pali | will try build dir | 15:26 |
freemangordon | Pali: it is not important, don;t waste your time | 15:27 |
Pali | freemangordon: git clone git://git.ideasonboard.org/media-ctl.git and dpkg-buildpackage -b | 15:27 |
Pali | there is debian directory | 15:28 |
freemangordon | Pali: won't work in SB :) | 15:28 |
Pali | I compiled it in madde | 15:28 |
freemangordon | anyway, not important | 15:28 |
Pali | ah, now I see :D what I done | 15:28 |
Pali | manual compilation :D:D | 15:28 |
Pali | I commented #include "config.h" and compiled each source file manually with gcc -c | 15:29 |
Pali | and then linked all together into one executable | 15:29 |
freemangordon | hehe :D | 15:29 |
freemangordon | Nokia-N900:~# ./media-ctl -l '"vs6555 binner":1 -> "OMAP3 ISP CCP2":0 [1]"' | 15:29 |
freemangordon | Unable to parse link: Invalid argument (22) | 15:29 |
Pali | btw, you have some other version | 15:29 |
Pali | my binary has name mediactl without dash | 15:29 |
Pali | checkout to 67b66c052e75645439e230fae81d969c834046ac and compile again | 15:30 |
Pali | (from above git repo) | 15:30 |
freemangordon | ok | 15:30 |
Pali | this is what I used | 15:30 |
freemangordon | Pali: no difference | 15:32 |
*** kerio has quit IRC | 15:32 | |
*** kerio has joined #maemo-ssu | 15:32 | |
freemangordon | hmm, wait, I didn't make clean | 15:34 |
freemangordon | Pali: hmm, no, still no joy, I guess you tried that on a different kernel | 15:36 |
Pali | freemangordon: what you need is to connect "entity 19: vs6555 binner (2 pads, 2 links)" to "pad1: Source "OMAP3 ISP CCP2":0 []" | 15:36 |
freemangordon | ok | 15:37 |
Pali | and above error is "Unable to parse link:" | 15:37 |
Pali | so sounds like incorrect arguments | 15:37 |
freemangordon | yep | 15:37 |
Pali | kernel is OK, bcause you see this configuration | 15:37 |
freemangordon | Pali: but, but, isn't that already connected? | 15:39 |
freemangordon | entity 19: vs6555 binner (2 pads, 2 links) | 15:39 |
freemangordon | type V4L2 subdev subtype Unknown flags 0 | 15:39 |
freemangordon | device node name /dev/v4l-subdev11 | 15:39 |
freemangordon | pad0: Sink | 15:39 |
freemangordon | [fmt:SGRBG10/756x596 | 15:39 |
freemangordon | compose.bounds:(0,0)/756x596 | 15:39 |
Pali | no | 15:39 |
freemangordon | compose:(0,0)/756x596] | 15:39 |
Pali | last line is -> "OMAP3 ISP CCP2":0 [] | 15:39 |
freemangordon | <- "vs6555 pixel array":0 [ENABLED,IMMUTABLE] | 15:39 |
freemangordon | pad1: Source | 15:39 |
freemangordon | [fmt:SGRBG10/756x596 | 15:39 |
freemangordon | crop.bounds:(0,0)/756x596 | 15:40 |
freemangordon | crop:(0,0)/756x596] | 15:40 |
freemangordon | -> "OMAP3 ISP CCP2":0 [] | 15:40 |
Pali | and should be [ENABLED] | 15:40 |
Pali | or somethign | 15:40 |
freemangordon | oh | 15:40 |
Pali | [] means disabled | 15:40 |
Pali | and you can see "- entity 1: OMAP3 ISP CCP2 (2 pads, 3 links)" | 15:40 |
Pali | that here is sink "<- "vs6555 binner":1 []" disabled | 15:40 |
Pali | see this: http://shr-project.org/trac/wiki/Devices/NokiaN900/InstallGuide | 15:42 |
Pali | there are commands for et8ek8 | 15:42 |
freemangordon | I am playing with the front camere BTW | 15:42 |
Pali | in section "New camera ISP (2.6.37 meego kernel)" | 15:42 |
Pali | yes, but here are example commands how to use that tool | 15:43 |
freemangordon | Pali: Setting up link 19:1 -> 1:0 [1] | 15:51 |
freemangordon | there is an extra double quote in your commands, at the end | 15:52 |
Pali | ah... | 15:52 |
Pali | also there is second command for "vs6555 binner" for selecting format | 15:53 |
Pali | ok, I edited my local file :-) thanks for catching up parse error | 15:54 |
freemangordon | I guess this should not be needed | 15:54 |
freemangordon | Pali: hmm, I guess someone needs to enable /dev/video1 too | 15:56 |
Pali | freemangordon: from that graph I see that you need to enable /dev/video2 | 16:00 |
Pali | freemangordon: entity 18 --> entity 19 --> entity 1 --> entity 5 --> entity 6 | 16:02 |
Pali | and then output from webcam should be in "entity 6" /dev/video2 | 16:02 |
Pali | freemangordon: try this command: mediactl -l '"vs6555 binner":1 -> "OMAP3 ISP CCP2":0 [1], "OMAP3 ISP CCP2":1 -> "OMAP3 ISP CCDC":0 [1], "OMAP3 ISP CCDC":2 -> "OMAP3 ISP preview":0 [1], "OMAP3 ISP preview":1 -> "OMAP3 ISP resizer":0 [1], "OMAP3 ISP resizer":1 -> "OMAP3 ISP resizer output":0 [1]' | 16:04 |
freemangordon | Setting up link 19:1 -> 1:0 [1] | 16:05 |
freemangordon | Setting up link 1:1 -> 5:0 [1] | 16:05 |
freemangordon | Setting up link 5:2 -> 7:0 [1] | 16:05 |
freemangordon | Setting up link 7:1 -> 10:0 [1] | 16:05 |
freemangordon | Setting up link 10:1 -> 12:0 [1] | 16:05 |
Pali | working now? | 16:06 |
freemangordon | no, still green | 16:07 |
Pali | maybe it is needed to configure resizer | 16:07 |
freemangordon | will try with gstreamer | 16:08 |
Pali | freemangordon: can you please post output of media-ctl -p now? | 16:08 |
Pali | and also look into dmesg | 16:08 |
freemangordon | dmesg is scary :D | 16:08 |
Pali | kernel is compiled with debug info and all printing to dmesg | 16:08 |
freemangordon | Pali: lemme try gstreamer first | 16:08 |
freemangordon | Pali: no | 16:08 |
Pali | ok, but upload output of media-ctl -p | 16:09 |
freemangordon | ok | 16:10 |
freemangordon | Pali: http://pastebin.com/sNE332PY | 16:10 |
Pali | freemangordon, are you trying to play from /dev/video6 ? | 16:13 |
freemangordon | yes | 16:13 |
Pali | ok | 16:13 |
kerio | what are you guys doing? | 16:14 |
freemangordon | Pali: mplayer -tv driver=v4l2:width=320:height=240:outfmt=uyvy:device=/dev/video6 -vo xv -vf screenshot tv:// | 16:14 |
Pali | and what about disable resizer and doing output direclly via "OMAP3 ISP CCDC output"? | 16:15 |
freemangordon | Pali: with that command something is moving on the top of the green output | 16:15 |
Pali | freemangordon, "width=320:height=240" this is maybe bad | 16:15 |
dos1 | kerio: they're trying to get camera running on new kernel | 16:15 |
freemangordon | Pali: if it is bigger, mplayer fails to setup the overlay | 16:15 |
freemangordon | I guess kernel config specifies too low vram | 16:15 |
freemangordon | kerio: front cam that is | 16:16 |
Pali | freemangordon, try to congure resizer for lower resolution | 16:16 |
freemangordon | ok | 16:16 |
Pali | freemangordon, try: mediactl -V '"vs6555 binner":1 [SGRBG10/756x596 (0,0)/756x596 (0,0)/756x596], "OMAP3 ISP CCP2":1 [SGRBG10/756x596], "OMAP3 ISP CCDC":2 [SGRBG10/756x596], "OMAP3 ISP preview":1 [UYVY/756x596], "OMAP3 ISP resizer":1 [UYVY/320x240]' | 16:22 |
freemangordon | pali it is lowercase -v | 16:22 |
Pali | ok | 16:22 |
freemangordon | Pali: ^^^ does nothing | 16:23 |
Pali | can you look at -p if something was changed? | 16:24 |
freemangordon | looked, nothing is chaged | 16:24 |
freemangordon | hmm, wat | 16:24 |
freemangordon | *wait | 16:24 |
Pali | bad, we need to change "4096x4096" to "756x596" | 16:24 |
freemangordon | wait | 16:24 |
*** DrCode has quit IRC | 16:26 | |
freemangordon | Pali: this doesn't look ok to me: "<- "OMAP3 ISP resizer input":0 []" | 16:26 |
freemangordon | only the preview is enabled | 16:27 |
freemangordon | "<- "OMAP3 ISP preview":1 [ENABLED]" | 16:27 |
Pali | it is ok | 16:27 |
Pali | look at preview | 16:27 |
Pali | preview has resizer enabled | 16:28 |
Pali | this is one big chain | 16:28 |
Pali | maybe you can try to enable "OMAP3 ISP preview output" on /dev/video4 what happen | 16:29 |
freemangordon | too big for me even :D | 16:29 |
freemangordon | ok | 16:29 |
Pali | or enable directly "OMAP3 ISP CCDC output" on /dev/video2 | 16:29 |
freemangordon | lemme try to figure out how to do that | 16:29 |
Pali | first think of entify chain and then enable it | 16:30 |
Pali | each entity has sink and source | 16:30 |
Pali | and what you need is to connect "entity 18: vs6555 pixel array" to some entify which has /dev/video* output | 16:31 |
Pali | so connect "entity 18" source to some entify with /dev/video* sink | 16:31 |
Pali | this is classic DFS problem on DAG :D | 16:33 |
Pali | and command '"vs6555 binner":1 -> "OMAP3 ISP CCP2":0 [1]' enabling connection from "vs6555 binner" pad1 to "OMAP3 ISP CCP2" pad0 | 16:35 |
*** DrCode has joined #maemo-ssu | 16:38 | |
*** _rd has quit IRC | 16:44 | |
freemangordon | Pali: I understand the connection principle, what I don;t know is what should be connected to what | 16:44 |
Pali | freemangordon: you need to create path between entity 18 and some /dev/video* entity | 16:46 |
Pali | so you can see output in mplayer | 16:46 |
freemangordon | yes, but all I see is green :D | 16:46 |
Pali | and did you tried another path? | 16:47 |
freemangordon | everything else I tried just fails | 16:47 |
Pali | another output entity and setting other resolutions at entities? | 16:47 |
Pali | maybe this userspace part is OK and problem is somewhere in kernel | 16:48 |
Pali | smiapp driver has enabled all debug messages | 16:48 |
Pali | and you can see that I reverted and added some patches to smiapp driver | 16:49 |
*** arcean has joined #maemo-ssu | 16:55 | |
Pali | freemangordon: this is what I found in my mailbox about front webcam: http://pastebin.com/aSPXeCA0 | 16:56 |
freemangordon | Pali: hmm, I *think* I connected everything ok up to previewer | 16:57 |
freemangordon | but it is still green, maybe the format is not ok | 16:57 |
Pali | try to first GRBG10 format (without preview and resizer) | 16:58 |
Pali | and tell it to mplayer | 16:58 |
Pali | freemangordon: can you try to revert this git commit 7b2e1277598e4187c9be3e61fd9b0f0423f97986 and recompile kernel again? | 17:00 |
Pali | maybe reverting this commit helps | 17:00 |
Pali | (I forgot to revert it when I was reverting 6d1aa02f10497b138e01ebe6eafabd6071729334) | 17:01 |
freemangordon | Pali: ok | 17:03 |
*** LaoLang_cool has joined #maemo-ssu | 17:08 | |
*** _rd has joined #maemo-ssu | 17:09 | |
*** jon_y_ has joined #maemo-ssu | 17:21 | |
*** jon_y has quit IRC | 17:23 | |
*** jon_y_ is now known as jon_y | 17:23 | |
*** discopig has quit IRC | 17:37 | |
*** dhbiker has joined #maemo-ssu | 17:40 | |
Pali | freemangordon: now I'm looking at ke-recv mount point paths and your commits | 17:46 |
Pali | this way how you changed code is path to hell | 17:47 |
Pali | freemangordon: I will try to fix mount point problem in ke-recv properly | 17:47 |
Pali | freemangordon: also I want to export ext4 (and other fs too) partitions when swap is active on sd card | 17:48 |
Pali | going to fix it too | 17:48 |
Pali | ah :-( for this I need more usb luns, so fix is not possible | 17:51 |
freemangordon | Pali: :nod: | 17:57 |
freemangordon | we don;t have an option but to export only the first vfat | 17:57 |
Pali | what is needed is to patch function add_volume to assign correct mount point to volume structure | 17:57 |
Pali | like in function add_usb_volume | 17:58 |
freemangordon | Pali: what exactly you don;t like about my patch? | 17:58 |
Pali | and then each mount/unmount function will use vol->mount_point and no other hacks are needed | 17:58 |
freemangordon | hmm, okj | 17:58 |
freemangordon | Pali: btw reverted that kernel patch, makes no difference | 17:58 |
Pali | I do not like your patch "Unmount the correct mountpoint for the first vfat partition" | 17:58 |
freemangordon | ah, I see | 17:59 |
freemangordon | well, ok | 17:59 |
freemangordon | :) | 17:59 |
Pali | and after this will be fixed, then I can much simplify your another patch "Export the first vfat partition only through mass storage..." | 17:59 |
freemangordon | Pali: ok | 17:59 |
Pali | because it will be always mounted on /media/mmc1 | 18:00 |
Pali | no need for hacks... | 18:00 |
freemangordon | Pali: http://pastebin.com/XCFVshvi | 18:00 |
freemangordon | iiuc everything is setup fine for /dev/video4 (previewer) | 18:01 |
Pali | no, this is incorrect "entity 5: OMAP3 ISP CCDC (3 pads, 8 links)" "pad0: Sink" "[fmt:SGRBG10/4096x4096]" | 18:03 |
Pali | fix resolution | 18:03 |
Pali | freemangordon: and is /dev/video4 working? | 18:03 |
freemangordon | ok | 18:03 |
freemangordon | no, still green, but I see driver active in dmesg | 18:04 |
freemangordon | so it somehow works | 18:04 |
freemangordon | lets try to set the resolution | 18:05 |
Pali | and before driver not reporting nothing to dmesg? | 18:05 |
freemangordon | with /dev/video6 and your connections? | 18:05 |
freemangordon | it was | 18:05 |
freemangordon | sec | 18:05 |
Pali | is omap3camd running? | 18:06 |
Pali | maybe it doing something with /dev/video0/1 | 18:06 |
freemangordon | no | 18:06 |
Pali | ok | 18:06 |
freemangordon | no, it is stopped | 18:06 |
freemangordon | ./media-ctl -v -V '"OMAP3 ISP CCDC":1 [SGRBG10/756x596]' | 18:06 |
freemangordon | Setting up format SGRBG10 756x596 on pad OMAP3 ISP CCDC/1 | 18:06 |
freemangordon | Format set: SGRBG10 4096x4096 | 18:06 |
freemangordon | :D | 18:06 |
Pali | ah | 18:07 |
freemangordon | Pali: http://pastebin.com/st0phiUm | 18:07 |
freemangordon | looks ok | 18:08 |
freemangordon | to me | 18:08 |
Pali | looks ok | 18:09 |
Pali | did you specified correct resolution and format to mplayer? | 18:10 |
Pali | try also gstreamer | 18:10 |
Pali | and maybe camera-ui (with and without omap3camd) | 18:10 |
freemangordon | mplayer -tv driver=v4l2:width=738:height=588:outfmt=uyvy:device=/dev/video4 -vo x11 tv:// -v | 18:10 |
Pali | ok | 18:10 |
Pali | freemangordon: also try without omap previewer (which doing SGRBG10 to UYVY conversion) | 18:13 |
Pali | and also try to change format to some another, maybe green color will change to something other... | 18:14 |
Pali | and maybe post dmesg output | 18:14 |
freemangordon | Pali: seems ISP is working like it should, we just don;t have input from the camera :( | 18:23 |
Pali | so camera is disabled? | 18:24 |
freemangordon | maybe | 18:24 |
freemangordon | no matter what I do, the only think I see is green | 18:24 |
Pali | did you tried gstreamer? | 18:25 |
freemangordon | yes, it is even worse, some ioctl code is not supportd (query controls or something) | 18:25 |
freemangordon | Pali: WTF is n00000013 [label="{{<port0> 0} | vs6555 binner\n/dev/v4l-subdev11 | {<port1> 1}}", shape=Mrecord, style=filled, fillcolor=green] | 18:25 |
freemangordon | fillcolor=green | 18:26 |
Pali | where?? | 18:26 |
freemangordon | ./media-ctl --print-dot | 18:26 |
Pali | no idea | 18:26 |
Pali | freemangordon: maybe I have one needed patch | 18:31 |
Pali | will post it asap on pastebin | 18:31 |
freemangordon | ok | 18:31 |
Pali | freemangordon: here is patch: http://pastebin.com/SaVNHrPU | 18:37 |
Pali | it reverting to old code which using board functions for enabling front camera instead passing gpio value to driver | 18:38 |
freemangordon | Pali: --print-dot prints .dot file, which is a graphical graph representation, we can use that to verify if the connections are ok | 18:38 |
freemangordon | ok, going to apply that | 18:38 |
Pali | ok | 18:39 |
Pali | freemangordon, and when you boot new kernel, please give me dmesg output after hildon-desktop startup (e.g before using camera) | 18:40 |
freemangordon | Pali: cannot apply that patch, should I modify the board file by hand? | 18:43 |
Pali | freemangordon: maybe you have problems with TABs | 18:44 |
Pali | edit by hand | 18:44 |
freemangordon | ok | 18:44 |
Pali | you see what was changed, so... | 18:44 |
*** _rd has quit IRC | 18:47 | |
freemangordon | Pali: http://pastebin.com/BgT6CtDk | 18:57 |
Pali | ok | 18:57 |
freemangordon | Pali: hmm, seems dmesg has overfilled | 18:57 |
Pali | "[ 3.614624] No soundcards found." | 18:58 |
freemangordon | do you want the syslog from the last boot? | 18:58 |
freemangordon | yeah | 18:58 |
Pali | sometimes it write this message | 18:58 |
Pali | [ 21.247680] smiapp 2-0010: wrote 0x1 to offset 0x103 error -121 | 18:58 |
Pali | [ 21.263214] smiapp 2-0010: software reset failed | 18:58 |
freemangordon | yep, I was just looking at that | 18:59 |
Pali | [ 21.283813] omap3isp omap3isp: isp_register_subdev_group: Unable to register subdev smiapp | 18:59 |
Pali | so drivers not working... | 18:59 |
Pali | try to shutdown phone | 18:59 |
freemangordon | ok | 18:59 |
Pali | remove battery, insert it | 18:59 |
Pali | and start again | 18:59 |
freemangordon | ok :D | 18:59 |
freemangordon | so microsofty :D:D:D | 18:59 |
*** _rd has joined #maemo-ssu | 19:00 | |
*** TheOneLaw has joined #maemo-ssu | 19:01 | |
freemangordon | Pali: the same | 19:01 |
Pali | boot 2.6.28 kernel and do shutdown from it | 19:01 |
freemangordon | hmm | 19:02 |
freemangordon | ok | 19:02 |
Pali | maybe in uboot or in upstream kernel is missing something magic which initialize hw | 19:02 |
Pali | sometimes uboot cannot read internal eMMC | 19:02 |
Pali | or do reboot from uboot | 19:03 |
freemangordon | Pali: well, either there is no power or clock, as -121 is timeout iirc | 19:07 |
Pali | freemangordon, is not your battery empty? | 19:08 |
freemangordon | it is fully loaded | 19:08 |
freemangordon | just not calibrated | 19:08 |
Pali | and do you see every device in media-ctl? | 19:09 |
Pali | can you check if that error -121 was there when you had green output? | 19:09 |
Pali | (in syslog) | 19:09 |
freemangordon | Pali: i think your patch broke it | 19:09 |
Pali | so try to revert patch | 19:10 |
*** janemba has joined #maemo-ssu | 19:10 | |
Pali | and maybe try to comment all calls to function "omap_ctrl_writel" in camera board file | 19:10 |
janemba | hi | 19:11 |
janemba | I have an issue with the n900 I set up my own mail server and it works fine. So I set my server to the server and I get no message. So I'd like to debug it but I don't know where to find the log for the mail client. Do you know where to look ? | 19:11 |
freemangordon | Pali: hmm, ok. you think our board file is screwing it? | 19:12 |
Pali | I do not know if calling that writel is needed | 19:12 |
freemangordon | that one: omap_ctrl_writel(OMAP343X_CSIB_RESET | OMAP343X_CSIB_PWRDNZ |... ? | 19:13 |
freemangordon | Pali: who wrote gpio code for smia? | 19:14 |
Pali | that patch? | 19:15 |
janemba | I used imaps | 19:15 |
freemangordon | Pali: no, what is in your git tree | 19:15 |
freemangordon | Pali: for example rx51_acmelite_init | 19:15 |
Pali | all functions comes from older 2.6.37 meego kernel | 19:16 |
freemangordon | hmm | 19:16 |
Pali | and some parts I changed to work with new 3.x kernel | 19:16 |
Pali | that patch also revert code from 2.6 kernel | 19:16 |
freemangordon | ok, but it seems ISP code takes care of all that iiuc | 19:16 |
freemangordon | Pali: why gpio_free(ACMELITE_RESET_GPIO); ??? | 19:19 |
Pali | freemangordon: if you specify that SMIAPP_NO_XSHUTDOWN, then ISP will not care bout gpio | 19:20 |
Pali | and in rx51_acmelite_set_xshutdown there is also call to omap_ctrl_writel | 19:20 |
Pali | so I do not know if is needed or not | 19:20 |
Pali | and ISP calling xshutdown before/after set_xclk function | 19:20 |
Pali | so I added that code to that board function | 19:21 |
freemangordon | Pali: what I mean - this exists with .xshutdown= ACMELITE_RESET_GPIO, | 19:21 |
freemangordon | so ISP to take care | 19:21 |
Pali | but does not calling omap_ctrl_writel | 19:21 |
freemangordon | oh, I see | 19:22 |
freemangordon | Pali: my point is - stingray_init() requests gpio and doesn't free it acmelite_init() rewuests gpio and frees it | 19:24 |
freemangordon | one of those fuctions doesn't behave correctly | 19:24 |
Pali | yes, this is correct, only camera board code xan use STINGRAY_RESET_GPIO | 19:26 |
Pali | and now when my patch disabled using ACMELITE_RESET_GPIO in ISP, camera board is only user of ACMELITE_RESET_GPIO | 19:26 |
Pali | so gpio_free(ACMELITE_RESET_GPIO); in rx51_acmelite_init must be removed | 19:27 |
freemangordon | Pali: so, iiuc, your patch ^^^ to work, we should not free ACMELITE_RESET_GPIO | 19:27 |
freemangordon | yep | 19:27 |
freemangordon | :) | 19:27 |
freemangordon | ok, going to reapply your patch and remove that free | 19:27 |
Pali | rx51_camera_set_xshutdown must have access to ACMELITE_RESET_GPIO | 19:27 |
*** TheOneLaw has quit IRC | 19:28 | |
freemangordon | Pali: exactly ;) | 19:40 |
freemangordon | Pali: it seems to work no, lemme try to get some video | 19:40 |
freemangordon | *now | 19:40 |
Pali | can you paste dmesg output? | 19:41 |
*** sunny_s has quit IRC | 19:48 | |
freemangordon | Pali: after rmmod/modprobe omap3_isp http://pastebin.com/abA1yqpp | 19:50 |
Pali | no error here | 19:51 |
freemangordon | yep, but this was the same like before your patch | 19:51 |
Pali | :-( | 19:52 |
freemangordon | btw don;t we need .strobe_setup too? | 19:52 |
Pali | no idea | 19:53 |
freemangordon | see https://gitorious.org/omap4-v4l2-camera/omap4-v4l2-camera/source/9e2afcfe99cf52c70e993321f9a348fac1be724d:arch/arm/mach-omap2/board-rm680-camera.c | 19:53 |
Pali | maybe you can try to contact Sakari Ailus who wore that driver | 19:53 |
Pali | board data I got from him | 19:53 |
Pali | there is nothing in sec_camera_platform_data | 19:55 |
freemangordon | Pali: where, in rx51 board ? | 19:55 |
Pali | no in that your rm680 link | 19:55 |
Pali | strobe_setup is only in main camera | 19:56 |
freemangordon | aah, yes | 19:56 |
Pali | try to contact Sakari Ailus via email sakari dot ailus at iki dot fi | 19:56 |
freemangordon | hmm, I want to play with that a bit more before asking questions :) | 19:57 |
Pali | and maybe you can try to remove all commits which reverting ISP/SMIA changes | 19:57 |
Pali | I reverted some commits because it caused that kernel driver failing... | 19:58 |
Pali | but maybe this is reason why it not working... | 19:58 |
freemangordon | Pali: which commits? | 20:05 |
Pali | every with subject revert and touching ISP or SMIA parts | 20:05 |
*** jonwil has quit IRC | 20:27 | |
*** arcean has quit IRC | 20:27 | |
*** discopig has joined #maemo-ssu | 20:30 | |
*** _rd has quit IRC | 21:11 | |
*** Vlad_on_the_road has quit IRC | 21:19 | |
*** Vlad_on_the_road has joined #maemo-ssu | 21:34 | |
*** Vlad_on_the_road has quit IRC | 21:40 | |
*** dafox has quit IRC | 21:51 | |
*** Vlad_on_the_road has joined #maemo-ssu | 21:54 | |
*** arcean has joined #maemo-ssu | 22:14 | |
*** arcean has quit IRC | 22:20 | |
Pali | nice, ke-recv does not initliaze stack variable to NULL and then checking is not NULL and using string... | 22:22 |
Pali | who wrote that piece of shit? | 22:23 |
Pali | if ke-recv not working, it is not obvious bug, but bug by design | 22:23 |
DocScrutinizer05 | :-P | 22:24 |
Pali | DocScrutinizer05: I recomend you to stop ke-recv, do not know if NSA not using all these bugs to overwrite stack and execute code... | 22:26 |
NIN101 | frankly. | 22:27 |
NIN101 | somebody checked what openssl version we have? Is it vulnerable? What about the openssh version in the repo? What about openvpn? And tons of other stuff. | 22:27 |
Pali | :D | 22:28 |
Pali | ke-recv is really shit sw | 22:28 |
DocScrutinizer05 | go ahead, check it! it can use a good audit | 22:28 |
Pali | we already found in ke-recv this C code: va_list arg = NULL; | 22:28 |
DocScrutinizer05 | mhm | 22:29 |
Pali | now there is: char *dev; if (bla) dev = someting; if (dev) do_something(dev); | 22:29 |
Pali | looks like nokia developers does not know when they mustnt assing NULL and when they must | 22:30 |
DocScrutinizer05 | hehe | 22:30 |
Pali | I was surpirsed that ke-recv was able to run more then hour without crash | 22:31 |
DocScrutinizer05 | I'd say it clearly proves that lint and similar tools been forbidden at Nokia | 22:31 |
Pali | and more surprised when GCC in scratchbox (4.2) do not rejected compile va_list arg = NULL; | 22:32 |
Pali | 4.7 reject to compile it | 22:32 |
Pali | DocScrutinizer05: maybe this is reason why couple of maemo code is closed? | 22:32 |
Pali | they do not want to show how bad is what they created? | 22:33 |
DocScrutinizer05 | not an unusual motivation to keep code closed | 22:37 |
Pali | ok, ke-recv and HAL are totally non deterministic | 22:44 |
Pali | and sometimes they reporting partitions on SD card in random order | 22:45 |
Pali | freemangordon: because of that using "first" VFAT partition is not safe | 22:45 |
Pali | because we do not know if HAL do not report another VFAT partition | 22:46 |
Pali | and also we do not know if HAL reported mmcblk0p1 as first or not | 22:46 |
*** LauRoman has joined #maemo-ssu | 22:47 | |
Pali | so I'd rather mount device /dev/mmcblk1p1 (really first partition on SD card) to /media/mmc1 and do not check filesystem type | 22:47 |
freemangordon | Pali: we don't care if the "first" vfat partition is the one with the lowest number | 22:47 |
freemangordon | Pali: no, you can't, there are lots of windoze users, and guess what will happen when windoze see EXTn partition | 22:48 |
Pali | on TMO somebody reported that there is real problem with this | 22:48 |
Pali | somebody who have on SD card non VFAT partition should care about it | 22:48 |
freemangordon | Pali: I doubt someone will create uSD with more than one vfat | 22:48 |
freemangordon | and even if there is more than one, we don;t car, as we'll export whatever is reported as "firs" | 22:49 |
freemangordon | *care | 22:49 |
Pali | if somebody is able to create non vfat on SD card, then also is able to set correct write permission | 22:49 |
freemangordon | WTF has to do write permissions with the partitions order? | 22:50 |
Pali | write permission is needed (as you wrote) for broken maemo applications which writing to /media/mmc1/ | 22:50 |
Pali | and crashing if write failed | 22:50 |
freemangordon | Pali: also, imagine someone who has created extX as p1 and vfat as p2, what those ahould do after your "patch"? | 22:51 |
freemangordon | reformat their uSD? | 22:51 |
Pali | if they do not have activated SWAP on SD card, they will see exported via USB same what now | 22:51 |
freemangordon | yes, it is all about the case with swap on uSD | 22:51 |
freemangordon | so, extN on p1, swap on p2 and vfat on p3 | 22:52 |
freemangordon | you'll export p1, ain't? | 22:52 |
Pali | yes | 22:52 |
freemangordon | but DCIM is on p3 | 22:52 |
Pali | why? | 22:52 |
freemangordon | because we mount the first vfat as /media/mmc1 | 22:53 |
Pali | this is not true, due to HAL randomness | 22:53 |
freemangordon | and /media/mmc1 is what EVERYONE use as extarnal card | 22:53 |
freemangordon | *external | 22:53 |
Pali | and I want to fix these HAL & ke-recv bugs | 22:53 |
freemangordon | Pali: sure, but forcing the users to reformat their uSD cards by our linking is a nogo IMO | 22:54 |
Pali | I'm thinking what to do... | 22:55 |
Pali | current situation and implementation in ke-recv is buggy | 22:55 |
freemangordon | Pali: :nod: | 22:55 |
Pali | and we have users which reported bug that sometimes ke-recv mount partition as mmc1p1 and sometimes as mmc1 | 22:56 |
Pali | and this is because of non deterministic behaviour of search for "first" vfat partition | 22:56 |
freemangordon | Pali: I think my patches fix that | 22:56 |
Pali | of course not | 22:56 |
freemangordon | Pali: hmm, maybe we should store what was mounted the last time that uSD was in | 22:57 |
Pali | your patches using first vfat partition from mmc->volumes list | 22:57 |
freemangordon | yep | 22:57 |
Pali | which is first vfat partition reported by HAL | 22:57 |
Pali | not first partition in SD card | 22:57 |
freemangordon | Pali: I got what the problem is | 22:57 |
freemangordon | BTW you meant "not first VFAT partition in SD card"? | 22:58 |
freemangordon | Pali: kets just store the volume id for that uSD somewhere and mount the same volume id everytime | 22:58 |
freemangordon | with some fallback in case uSD was externally formatted | 22:59 |
Pali | problem is that: partition order can be totally random in ke-recv | 22:59 |
freemangordon | I know it is hacky, but can't think of a better solution now | 22:59 |
Pali | and we cannot use something like that: use first partition in this list which has fstype vfat | 22:59 |
Pali | and because of broken apps we should mount something to /media/mmc1 | 23:00 |
freemangordon | Pali: sorry, gtg, will think a bit about that tomorrow. I'd recommend you to not code anything, maybe we can get some help from the others on the channel | 23:00 |
freemangordon | or ask google, whatever :D | 23:00 |
Pali | and there is problem on TMO that sometimes partition is mounted as mmc1 and sometimes as mmc1p1 | 23:01 |
Pali | freemangordon: I'd rather mount partition with num "1" which is first in MBR table | 23:01 |
freemangordon | Pali: we can use blkid or similar, to find the first vfat | 23:02 |
Pali | and this will be always reported as "1" (until somebody does not change MBR table), this does not depends on random HAL order | 23:02 |
Pali | freemangordon: and if there is no VFAT? | 23:02 |
freemangordon | Pali - we want vfat mounted under /media/mmc1 | 23:03 |
freemangordon | well, this is another story | 23:03 |
Pali | this is real situation | 23:03 |
freemangordon | I know, but is different | 23:03 |
freemangordon | oh, we just don;t mount anything under /media/mmc1 | 23:03 |
Pali | why? | 23:04 |
freemangordon | anyway, gtg, sorry | 23:04 |
Pali | ok | 23:04 |
*** DrCode has quit IRC | 23:13 | |
*** dafox has joined #maemo-ssu | 23:20 | |
Pali | fuck, now after gitorious change I'm not able to login | 23:25 |
Pali | Error message is only "OpenID verification failed" | 23:25 |
Pali | freemangordon: new ke-recv code is here: https://gitorious.org/community-ssu/pali-ke-recv/commits/master | 23:29 |
*** DrCode has joined #maemo-ssu | 23:30 | |
*** nox- has joined #maemo-ssu | 23:31 | |
*** Vlad_on_the_road has quit IRC | 23:35 | |
*** _rd has joined #maemo-ssu | 23:36 | |
*** dhbiker has quit IRC | 23:36 | |
*** dhbiker has joined #maemo-ssu | 23:42 | |
*** _rd has quit IRC | 23:54 | |
*** xes has joined #maemo-ssu | 23:58 |
Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!