| Maxdamantus | Yes, but you can at least make sure the filesystem is in some previously intended state. | 00:00 |
|---|---|---|
| Maxdamantus | using either CoW or full data journalling. | 00:00 |
| Maxdamantus | the latter means you write everything twice. | 00:00 |
| KotCzarny | or just minimize writes | 00:00 |
| DocScrutinizer05 | aha, writing twice and mimimize writing serve same purpose? hardly. I prefer the latter | 00:01 |
| Maxdamantus | Databases like sqlite3 are written such that they can operate under that assumption. | 00:01 |
| Maxdamantus | If you're removing a non-journalled drive with a sqlite3 database on it in the middle of a write, there's no guarantee your database isn't corrupt. | 00:02 |
| KotCzarny | removing medium during writes is BAD IDEA | 00:02 |
| KotCzarny | no matter what | 00:03 |
| Maxdamantus | if it's fully journalled or CoW, you do get that guarantee: the database might have lost the last few seconds, but it won't be corrupt. | 00:03 |
| Maxdamantus | Of course, but it hapeens. | 00:03 |
| Maxdamantus | So it should be handled. | 00:03 |
| DocScrutinizer05 | journalling does NOT help writing that very data to the medium. It only helps not even *trying* to write aborted transactions to the medium | 00:03 |
| DocScrutinizer05 | from an app's perspective that photo doesn't get stored nevertheless | 00:04 |
| DocScrutinizer05 | plus journaling not only doubles but at least triples writing to media | 00:05 |
| Maxdamantus | If you're doing full journalling (data included) it basically means double. | 00:06 |
| Maxdamantus | There's of course extra metadata around the journalling itself, but that's not going to be as large as writing the data twice. | 00:06 |
| DocScrutinizer05 | aha, and who's deleting the jounal after commit? | 00:06 |
| kerio | yea but how can i avoid not unplugging the uSD if i remove the back cover? | 00:06 |
| Maxdamantus | kerio: modify the kernel module. | 00:06 |
| DocScrutinizer05 | I'm not talking about volume, i'm talking about IO ops | 00:06 |
| kerio | Maxdamantus: ayy lmao | 00:07 |
| KotCzarny | um | 00:07 |
| KotCzarny | or keeping sd journal on emmc | 00:07 |
| KotCzarny | :) | 00:07 |
| DocScrutinizer05 | KotCzarny: that might be a nice idea | 00:07 |
| Maxdamantus | If it's a full journal, that's going to halve write speeds. | 00:07 |
| DocScrutinizer05 | I already pondered that approach | 00:07 |
| KotCzarny | i wouldnt bother personally tho | 00:08 |
| KotCzarny | user is the weakest link | 00:08 |
| Maxdamantus | It wouldn't really be any better than journalling on the same SD card, since the write speed on the N900 doesn't seem to be bottlenecked by the SD card. | 00:08 |
| Maxdamantus | but by the controller. | 00:08 |
| Maxdamantus | which the eMMC is also on. | 00:08 |
| DocScrutinizer05 | add an electric lock to the battery cover ;-D | 00:09 |
| KotCzarny | electroshocking lock? | 00:09 |
| KotCzarny | :) | 00:09 |
| DocScrutinizer05 | only unlocks when uSD is properly unmounted | 00:09 |
| Maxdamantus | https://talk.maemo.org/showpost.php?p=1457369&postcount=4 | 00:09 |
| KotCzarny | such locks would be a pita | 00:09 |
| KotCzarny | last time when some app locked my cdtray i was .. unhappy | 00:10 |
| DocScrutinizer05 | cya l8r | 00:10 |
| Maxdamantus | (that bottleneck comment is assuming you have an SD card that is meant to be able to write at more than 8 MiB/s or so, can't remember the actual number) | 00:12 |
| * Maxdamantus tries. | 00:12 | |
| *** kerio has quit IRC | 00:14 | |
| *** kerio has joined #maemo | 00:15 | |
| Maxdamantus | seems to be about 6 MiB/s | 00:19 |
| KotCzarny | what is iops@512b for internal flash? | 00:19 |
| KotCzarny | and for emmc? | 00:20 |
| Maxdamantus | I don't want to try that on the flash. | 00:20 |
| KotCzarny | test can be short, 10-20s | 00:20 |
| KotCzarny | there are sd cards capable of ~100 iops | 00:21 |
| KotCzarny | most are in 0.1-10 range tho | 00:21 |
| Maxdamantus | Okay, not 6 MiB/s .. can do 13 MiB/s | 00:24 |
| Maxdamantus | and yeah, that's shared across the SD card an eMMC. | 00:26 |
| Maxdamantus | oh, no, nvm .. I'm writing 11 MiB/s to the SD card now and 9 MiB/s to the eMMC. | 00:26 |
| KotCzarny | try reading to one and writing to other | 00:28 |
| Maxdamantus | reading from eMMC at 14 MiB/s, writing to SD card at 10 MiB/s | 00:30 |
| KotCzarny | so they use independent controllers in chip | 00:31 |
| KotCzarny | arm chips usually have 2-4 mmc controllers | 00:32 |
| Maxdamantus | Maybe it would be nice to use btrfs with its "raid0" mode. | 00:33 |
| KotCzarny | but most of the time only 1 is connected to some slot | 00:33 |
| Maxdamantus | actually, the things I said were inverted: when I said SD card before I meant eMMC and vice versa. | 00:36 |
| *** lobito has quit IRC | 00:44 | |
| *** lobito has joined #maemo | 00:45 | |
| *** kerio has quit IRC | 00:52 | |
| *** kerio has joined #maemo | 00:53 | |
| *** florian has quit IRC | 01:25 | |
| ZetaR | To answer the question about ext4 stability: "On 11 October 2008, the patches that mark ext4 as stable code were merged in the Linux 2.6.28 source code repositories, denoting the end of the development phase and recommending ext4 adoption." (Wikipedia) | 01:34 |
| stryngs | DocScrutinizer05: I've never used osso-backup for those purposes. | 02:11 |
| DocScrutinizer05 | you should give it a try | 02:11 |
| DocScrutinizer05 | (ext4) we got stable ext4 support on maemo | 02:12 |
| DocScrutinizer05 | not for rootfs though | 02:12 |
| stryngs | afk for a bit, bbib | 02:17 |
| Maxdamantus | Why not rootfs? | 02:20 |
| Maxdamantus | Just in terms of it not being there by default? | 02:21 |
| *** vectis has quit IRC | 02:28 | |
| *** LauRoman has quit IRC | 02:51 | |
| *** LauRoman has joined #maemo | 02:51 | |
| ZetaR | stryngs: I have determined the optimal ext4 parameters for the N900's eMMC geometry: "-b 4096 -E stride=2,stripe-width=256". This will make ext4 respect the write block size of 8192, and stripe the eMMC in sections respecting the erase group size of 256 write blocks. | 02:54 |
| *** xes_ is now known as xes | 03:03 | |
| *** M4rtinK has quit IRC | 03:13 | |
| * Maxdamantus is now using ext4 on his root. | 03:17 | |
| Maxdamantus | Though the boot process seems very unreliable. | 03:17 |
| Maxdamantus | Booting is too complicated .. hard to debug issues. | 03:18 |
| * Maxdamantus wonders if there's an option to make it start a shell instead of rebooting. | 03:19 | |
| *** futpib has quit IRC | 03:24 | |
| ZetaR | I wouldn't use ext4 for root. I would probably use ext3 for / and ext4 for /usr and then link other things to /usr | 03:47 |
| ZetaR | Or rather rootfs. | 03:47 |
| Maxdamantus | Why not? | 03:50 |
| ZetaR | ext4 isn't as widely supported for booting. Not sure what is supported on the N900 besides rootfs (if anything). | 03:52 |
| Maxdamantus | I haven't actually enabled any ext4 features yet, just using the ext4 module. | 03:52 |
| Maxdamantus | N900 normally uses ubifs on NAND for root. | 03:52 |
| ZetaR | AFAICT, mounting as ext4 improves performance even if it was created as an ext2 or ext3 partition. | 03:53 |
| Maxdamantus | and ext2 iirc for /opt | 03:53 |
| Maxdamantus | er, ext2 for /home, which includes /opt | 03:54 |
| Maxdamantus | By "widely supported for booting" are you talking about non-N900 distributions, because of lack of ext4 in an initramfs or something? | 03:55 |
| ZetaR | It was sort of a vague general statement made with little confidence in my knowledge of the N900-specific booting process. | 03:56 |
| Maxdamantus | The normal kernel has ubifs and the relevant MTD support built-in, so the mounts that itself and executes /sbin/preinit which is some shell script that checks some stuff then execs the upstart init system. | 03:58 |
| Maxdamantus | So the easiest way to use something else as root is probably to just replace /sbin/preinit with something that loads the modules required (omap_hsmmc, ext{3,4}, ..), mount the desired root filesystem then pivot_root/chroot and exec /sbin/preinit from there. | 04:01 |
| ZetaR | ubifs is a pretty good choice for the NAND, since it is unmanaged. | 04:01 |
| Maxdamantus | Indeed. | 04:01 |
| ZetaR | Come to think of it, since it is unmanaged that would mean that ext2/3/4 would be a pretty bad choice. | 04:01 |
| *** Pali has quit IRC | 04:01 | |
| Maxdamantus | You would need to manage it somehow to use ext2/3/4 | 04:02 |
| ZetaR | It isn't designed for it. | 04:02 |
| Maxdamantus | but iirc the simple things in Linux are unsuitable for that, since they just make all the blocks visible as a single large block device. | 04:02 |
| Maxdamantus | Theoretically, you could have something more intelligent to provide that block device. That's pretty much what the controllers in SSDs and SD cards do. | 04:03 |
| ZetaR | Well, performance on / probably isn't as critical as performance on /usr and /home | 04:03 |
| ZetaR | So fiddling with filesystems won't get you much. | 04:04 |
| Maxdamantus | /usr is normally part of / iirc | 04:04 |
| Maxdamantus | It's /opt that's on a different filesystem. | 04:04 |
| ZetaR | Yeah, I know, but I am thinking of the deoptified layout. | 04:04 |
| *** beford has joined #maemo | 04:04 | |
| Maxdamantus | so packagers are encouraged to put their data in /opt. | 04:04 |
| ZetaR | i.e. what stryngs is working on. | 04:05 |
| Maxdamantus | I don't think it's very useful having a separate /usr | 04:05 |
| Maxdamantus | but I think the NAND is likely to perform better than the eMMC. | 04:05 |
| ZetaR | Why? Get rid of opt and just mount /usr as the eMMC. | 04:05 |
| Maxdamantus | because otherwise why would they have it? | 04:06 |
| ZetaR | That is what was being discussed earlier anyway. | 04:06 |
| Maxdamantus | is it just meant to be like a large initramfs? | 04:06 |
| ZetaR | I don't follow. | 04:06 |
| Maxdamantus | Why wouldn't they put everything on the eMMC in the first place? | 04:07 |
| Maxdamantus | and avoid putting an extra tiny NAND chip in? | 04:07 |
| ZetaR | Honestly, I am not sure. | 04:08 |
| ZetaR | I guess you mentioned that it might perform better, but that doesn't seem like a great reason to increase the complexity. | 04:08 |
| Maxdamantus | It probably does. | 04:09 |
| Maxdamantus | Especially considering everything to do with booting is on it. | 04:09 |
| ZetaR | I guess it could help ensure some sort of minimum required boot time, set by whatever design groups Nokia had. | 04:10 |
| Maxdamantus | It could also have to do with filesystem reliability. | 04:11 |
| ZetaR | So what do you think of stryngs' proposal then? | 04:11 |
| Maxdamantus | I have a feeling ubifs is probably more reliable than ext2 across unexpected shutdowns. | 04:11 |
| Maxdamantus | it's probably a log-based filesystem, since it's meant for MTDs. | 04:11 |
| ZetaR | Reliability for the root filesystem is not as important if you don't typically modify it. | 04:13 |
| *** Humpelstilzchen has quit IRC | 04:14 | |
| ZetaR | Ideally you would have a full separation between user-installed applications and system applications so that you could mount the root filesystem as read-only during normal use. | 04:15 |
| Maxdamantus | I don't really like that distinction. | 04:15 |
| Maxdamantus | I like the idea of just having one filesystem and not having to worry about allocating certain sizes to things. | 04:16 |
| Maxdamantus | especially on something as restricted as a phone. | 04:16 |
| Maxdamantus | Filesystem Size Used Avail Use% Mounted on | 04:16 |
| Maxdamantus | /dev/sdc2 11T 7.6T 3.3T 70% / | 04:16 |
| ZetaR | Having / as a rarely modified partition would allow for diagnostic booting even if the other filesystems get screwed up. | 04:17 |
| ZetaR | That is the main reason to have it IMO. | 04:17 |
| Maxdamantus | That's a reason to have another system to boot into. | 04:19 |
| Maxdamantus | possibly just an initramfs. | 04:19 |
| ZetaR | Possibly. So what would you do with the NAND if it was easy to shuffle everything around? | 04:20 |
| *** Humpelstilzchen has joined #maemo | 04:20 | |
| Maxdamantus | If I had greater control over kernel versions, I'd just have a single btrfs filesystem over the eMMC and the SD card. | 04:20 |
| Maxdamantus | (except for a boot partition and swap) | 04:20 |
| *** LauRoman|Alt has joined #maemo | 04:21 | |
| Maxdamantus | depending on its abilities, possibly swap, but dunno if that's a good idea. | 04:21 |
| Maxdamantus | I don't really know much about MTD, so don't know what should be done with it. | 04:21 |
| Maxdamantus | so atm it's basically just a writable initramfs. | 04:22 |
| ZetaR | Well, you need some sort of management for it so that you don't destroy it. | 04:23 |
| ZetaR | Either that, or just don't write to it often. | 04:23 |
| Maxdamantus | It's still using ubifs. | 04:23 |
| Maxdamantus | You need some sort of "management", or you can't mount it. | 04:24 |
| ZetaR | Why would you combine eMMC and SD? Your system would fail if you removed the SD card. | 04:24 |
| Maxdamantus | it's an MTD, not a block device. | 04:24 |
| *** LauRoman has quit IRC | 04:24 | |
| Maxdamantus | Why would I remove the SD card? | 04:24 |
| ZetaR | Or it broke. | 04:25 |
| ZetaR | It just seems more frail. | 04:25 |
| Maxdamantus | I have backups already. | 04:25 |
| ZetaR | Personally I would use the SD for booting other OSes. | 04:25 |
| ZetaR | Or storing files to transfer to other systems. | 04:25 |
| Maxdamantus | You can already store files in a filesystem. | 04:26 |
| Maxdamantus | That's what it's for. | 04:26 |
| ZetaR | But you cant pull the eMMC out and put it in another computer. | 04:26 |
| ZetaR | Not practically anyway. | 04:26 |
| Maxdamantus | I wouldn't want to be constantly pulling the SD card out anyway. | 04:27 |
| Maxdamantus | They seem physically fragile. | 04:27 |
| ZetaR | That is actually one of the few things I don't like about the N900. I would prefer a slot on the exterior for the SD card, and with a more robust mechanism. | 04:27 |
| ZetaR | i.e. a push-push insertion/ejection slot. | 04:28 |
| Maxdamantus | I'd find that very inconvenient. | 04:28 |
| *** ZetaR has left #maemo | 04:28 | |
| Maxdamantus | You should be unmounting it every time you do it. | 04:29 |
| *** ZetaR has joined #maemo | 04:29 | |
| ZetaR | Oops. Accidentally closed the chat. | 04:29 |
| Maxdamantus | and to do that you need to think about what processes you have running depending on it. | 04:29 |
| ZetaR | Why would you find it inconvenient? | 04:29 |
| *** trumee has quit IRC | 04:29 | |
| Maxdamantus | 13:29:06 < Maxdamantus> You should be unmounting it every time you do it. | 04:29 |
| Maxdamantus | eg, I have mpd running with files open from the SD card. | 04:30 |
| Maxdamantus | so it wouldn't let my unmount it cleanly without stopping mpd. | 04:30 |
| ZetaR | It wouldn't be less convenient, because you could just choose not to remove it. | 04:30 |
| Maxdamantus | I just transfer things over the network, either WiFi or USB (g_ether). | 04:31 |
| Maxdamantus | if I choose not to remove it, there's still an annoyingly convenient way to remove it. | 04:31 |
| ZetaR | What about having a safe-eject button? | 04:31 |
| Maxdamantus | I don't see the point if I don't intend to eject it. | 04:32 |
| Maxdamantus | I'm also currently using a partition on the SD card as swap. | 04:32 |
| ZetaR | You might not, but I think most users would find it much more useful. | 04:32 |
| Maxdamantus | Yeah, I'm saying what would be more convenient for me, not others. | 04:33 |
| Maxdamantus | Though tbh, it does seem silly having removable devices nowadays. | 04:33 |
| Maxdamantus | Networking already facilitates transferring data. | 04:34 |
| ZetaR | So you just aren't a fan of the sneakernet. :P | 04:34 |
| Maxdamantus | Not really. | 04:34 |
| Maxdamantus | People usually still use it because they're not running ssh on their computers. | 04:35 |
| Maxdamantus | my phone and all my computers already all run ssh. | 04:35 |
| Maxdamantus | I do remember when floppy disks were the most convenient form of transport for me though. | 04:36 |
| ZetaR | Floppy drives are pretty cool. Very simple electrically. | 04:38 |
| Maxdamantus | I was still relying on a floppy drive until a couple of weeks ago. | 04:38 |
| *** eMHa_ has joined #maemo | 04:39 | |
| ZetaR | Have you seen the strange things people do with floppy drives and microcontrollers? Here is an example: https://www.youtube.com/watch?v=cM_sAxrAu7Q | 04:39 |
| Maxdamantus | Yes. | 04:39 |
| ZetaR | :) | 04:40 |
| Maxdamantus | That view count is 88 views off 2^19 | 04:41 |
| *** eMHa has quit IRC | 04:42 | |
| *** trumee has joined #maemo | 04:44 | |
| Maxdamantus | Another thing I'd quite like to have which seems slightly relevant to these discussions is .. GRUB. | 04:55 |
| Maxdamantus | but I couldn't get that to work last time I tried it. | 04:56 |
| Maxdamantus | (GRUB2 supposedly has the ability to build to load from uboot and use its terminal etc) | 04:57 |
| ZetaR | Yes, more boot options would definitely be nice. I'm not sure what stryngs has set up now, but he said it supports some diagnostics and OS selection IIRC. | 04:57 |
| *** trumee has quit IRC | 05:00 | |
| Maxdamantus | atm I just have this: https://gist.github.com/Maxdamantus/949e81b445ffaa068449 | 05:00 |
| Maxdamantus | (you can run that Python script (it just prints stuff to standard output) and laugh at it) | 05:01 |
| *** trumee has joined #maemo | 05:04 | |
| *** trumee has quit IRC | 05:20 | |
| stryngs | ZetaR: Interesting thought on the ext4. tbh, I've not considered filesystem type, for this unification. I believe it's a good thought to add though. | 05:21 |
| *** trumee has joined #maemo | 05:23 | |
| stryngs | So, I've almost got this post written up. I believe it's one of the core arguments to be made for unification. I'll post when it's done. | 05:23 |
| Maxdamantus | Hmm .. looking at /usr/sbin/osso-mmc-umount.sh it looks like it normally does a lazy unmount of ~/MyDocs, presumably before letting g_file_storage see it. | 05:24 |
| Maxdamantus | That is basically *COMPLETELY AND UTTERLY* wrong | 05:24 |
| Maxdamantus | lazy unmount just means it's removed from the filesystem tree. | 05:24 |
| Maxdamantus | if another process is still in a directory on it or has a filehandle open, it can continue writing to the filesystem. | 05:24 |
| Maxdamantus | ie, if you rely on the UMS feature in Maemo, you can easily have the same filesystem mounted writable on multiple devices (computer and phone) | 05:26 |
| * Maxdamantus doesn't understand why Linux doesn't seem to have a way to actually force unmount a particular filesystem. | 05:27 | |
| Maxdamantus | seems like the only way to do it is using sysrq to force unmount (or remount readonly) all filesystems | 05:28 |
| Maxdamantus | .. or you could just reexport the block device as a file in FUSE or an nbd device and make sure only one thing has it writable there. | 05:28 |
| Maxdamantus | unless it does something like that, it should simply fail if it can't unmount non-lazily. | 05:30 |
| *** sparetire_ has quit IRC | 05:34 | |
| *** trumee has quit IRC | 05:41 | |
| *** trumee has joined #maemo | 05:44 | |
| stryngs | Alright. I have laid out the argument for bindmounts. | 05:47 |
| stryngs | If anyone has any "Move /usr to eMMC" knowloedge, PLEASE PLEASE PLEASE, respond accordingly. | 05:47 |
| stryngs | This post for the thread, is actually one of the more "core" issues towards unification. | 05:48 |
| stryngs | DocScrutinizer05: ^^, you mentioned "movement of /usr". If ya got it written down would love your insight. | 05:48 |
| * stryngs goes back to packaging | 05:49 | |
| Maxdamantus | Just put everything on eMMC. | 05:51 |
| stryngs | HOW | 05:51 |
| stryngs | HOW | 05:51 |
| stryngs | HOW | 05:51 |
| stryngs | Do ya have it written down!?! | 05:51 |
| stryngs | If so, please describe | 05:51 |
| Maxdamantus | No. | 05:51 |
| stryngs | Maxdamantus: Then it's a pointless point. | 05:52 |
| Maxdamantus | It's probably not something I'd encourage to people who can't figure it out :\ | 05:52 |
| stryngs | Maxdamantus: if you can provide instructions, then we can definately use it. | 05:52 |
| stryngs | Maxdamantus: That's not the point! Let's GIVE BACK to this community. Let us, unify. | 05:52 |
| stryngs | Maxdamantus: None of us wrote the original Linux kernel, but we all use it. | 05:52 |
| stryngs | Maxdamantus: I doubt, you or I, could have figured out how to do so. | 05:52 |
| stryngs | Maxdamantus: yet, we all encourage people to use linux dont we? =) | 05:53 |
| Maxdamantus | stryngs: the easiest way is to create a different script to replace /sbin/preinit that loads the necessary kernel modules and mounts the appropriate filesystem then chroots into it and execs the original preinit from there. | 05:53 |
| stryngs | Maxdamantus: Very cool, could you write something up? | 05:53 |
| stryngs | Maxdamantus: I'd test on my own device. | 05:53 |
| Maxdamantus | Not sure. I might've had to modify some other things in /etc to stop it complaining. | 05:54 |
| stryngs | Maxdamantus: if you "start" something, I'm willing to experiment on my own box. | 05:54 |
| Maxdamantus | But if you're prepared to go through figuring it out, I can show my current init script. | 05:54 |
| stryngs | Maxdamantus: I would do it myself, but as you see in my post, I'm not knowledgable enough in that lowlevel to do so. | 05:54 |
| stryngs | Maxdamantus: When you say "my current init" script, you refer to something you use on a live n900? | 05:55 |
| Maxdamantus | Yes. | 05:56 |
| Maxdamantus | https://gist.github.com/Maxdamantus/c931687be8fdee4cb60c | 05:56 |
| stryngs | Word. | 05:56 |
| * stryngs checks it out | 05:56 | |
| stryngs | Maxdamantus: Line 14 is your failsafe? | 05:57 |
| stryngs | Maxdamantus: You're doing an fsck. does that significantly increase boottime? tbh, do most systems fsck before boot? Again, I know 'nothing' about the 'boot' process on this level. | 05:58 |
| Maxdamantus | That just sets up networking over USB so I can do things over ssh rather than trying to type on the keyboard. | 05:58 |
| stryngs | Maxdamantus: I see | 05:59 |
| stryngs | if false; then | 05:59 |
| stryngs | So it looks more like a failsafe than anything. Is it always false? | 05:59 |
| stryngs | It's the wording of the logic I guess | 05:59 |
| *** trumee has quit IRC | 05:59 | |
| Maxdamantus | If I want to actually use it I'll just modify it. | 05:59 |
| Maxdamantus | with the fsck, the reason I did that is because the mount after not unmounting properly would itself take a long time (probably about as long as the fsck) | 06:00 |
| Maxdamantus | but I think it stops sending ticks to the watchdog during mount, so it would usually reboot during that. | 06:00 |
| stryngs | Maxdamantus: Okay. So without breakign this script down, line by line. In essence, you have modified an n900 you have, and you currently use that script, in place of /sbin/preinit; and in doing so, your whole filesystem is on the eMMC versus the horrible and current stock nokia way? | 06:00 |
| Maxdamantus | stryngs: yes. | 06:01 |
| Maxdamantus | It's the phone I normally carry around with me. | 06:01 |
| stryngs | Maxdamantus: Okay, so with that thought; how did you get the egg before the chicken? | 06:01 |
| stryngs | Maxdamantus: i.e. What steps did you take place to copy over the files to eMMC? | 06:01 |
| *** trumee has joined #maemo | 06:01 | |
| stryngs | Maxdamantus: if I threw that on my n900 right now; it wouldnt boot afaik | 06:01 |
| stryngs | Maxdamantus: because I haven't done your prep steps | 06:02 |
| Maxdamantus | Yes, you should already have a way to boot into it and modify things. | 06:02 |
| stryngs | Maxdamantus: You had to, afaik, at some point, copied over files. | 06:02 |
| Maxdamantus | The easiest way to do that would be using some "recovery OS" someone made somewhere, but that requires a computer when you want to use it. | 06:02 |
| stryngs | Maxdamantus: Right, but your method, would take only a 'one-time' setup with a computer, no? | 06:03 |
| Maxdamantus | I also have a 3.14 kernel I can boot with a decent console map built in, so I can modify things from that too. | 06:03 |
| stryngs | Maxdamantus: I say this, because no matter what, 'flashing' takes a computer as well. | 06:03 |
| Maxdamantus | Yes. | 06:03 |
| stryngs | Maxdamantus: Yer rocking 3x on the n900? | 06:03 |
| Maxdamantus | Not usually. I just have that as my rescue kernel. | 06:03 |
| stryngs | Maxdamantus: Holy crap dude. | 06:03 |
| stryngs | Maxdamantus: Why is your knowledge not public!? | 06:04 |
| Maxdamantus | It is. | 06:04 |
| stryngs | Maxdamantus: Where? | 06:04 |
| Maxdamantus | You can build Linux 4 from Linus' repository and boot it on the N900. | 06:04 |
| Maxdamantus | make defconfig_rx51 | 06:04 |
| stryngs | Maxdamantus: nonononono, I mean on maemo.org, the spot "most folk" go to for "knowledge". Not the hackers who seek elsewhere... | 06:04 |
| stryngs | Maxdamantus: This is what I'm getting at with my proposal. We take kickass ideas, like yours, VOTE on the best course of action, and then move towards it. | 06:05 |
| stryngs | Maxdamantus: It would seem, that you have abundant knowledge that would truly help this community. | 06:05 |
| stryngs | Maxdamantus: Let me ask this. Would you be able to describe from a fresh flash, how to move towards your current setup? | 06:06 |
| stryngs | Maxdamantus: And I don't necessarily mean here on chat, feel free to post on the thread. | 06:06 |
| stryngs | Maxdamantus: If you have the knowledge, and can take sometime to jot it down. it would really help out =) | 06:07 |
| * stryngs blinks like Bambi. | 06:07 | |
| Maxdamantus | No. It's all ad-hoc and relies on the person doing it understanding how things work. | 06:07 |
| stryngs | Maxdamantus: I can script for that and help the enduser. | 06:07 |
| stryngs | Maxdamantus: if this, then, else, else else else, etc... | 06:07 |
| stryngs | Maxdamantus: I just require a little bit of help is all =) | 06:07 |
| stryngs | Maxdamantus: But if nobody is willing to try.... =/ | 06:07 |
| Maxdamantus | But you can't really describe every scenario. | 06:08 |
| Maxdamantus | What if someone's system randomly stops booting? | 06:08 |
| stryngs | Maxdamantus: Then they did something to cause it. | 06:08 |
| stryngs | Maxdamantus: Things don't just happen. | 06:08 |
| Maxdamantus | If the person already understands the extra stuff he/she has added, he/she could hack around to try to fix it. | 06:08 |
| Maxdamantus | That's pretty much the state my system is in. | 06:09 |
| stryngs | Maxdamantus: A a normal "unified maemo" user, won't just do those random things. And if they do, then they are prolly a 'power-user' of sorts. | 06:09 |
| stryngs | Maxdamantus: Well, it's late here, and I want to continue repackaging a couple things I've got. Please... Take some time and think about what I've asked. If you find it in your heart to help out in anyway, it would really be appreciated; and not just by me. | 06:10 |
| * stryngs tips his hat towards Maxdamantus | 06:10 | |
| *** obsed has joined #maemo | 06:12 | |
| *** Luke-Jr has quit IRC | 06:27 | |
| *** Luke-Jr has joined #maemo | 06:30 | |
| ZetaR | Maxdamantus: Wait, you have linux 3.x/4.x working on Maemo? I thought that a lot of stuff would need to be recompiled for that. It would be a great upgrade if we could roll that into stryngs work. We could use btrfs then too. | 06:31 |
| stryngs | ^^ | 06:32 |
| Maxdamantus | No, not on Maemo. | 06:32 |
| ZetaR | Oh, I misunderstood then. | 06:32 |
| stryngs | Maxdamantus: Can you use the phone on your n900? make textmessages, etc..? | 06:32 |
| Maxdamantus | You can obviously run the busybox etc programs on the rootfs, which is sufficient as a rescue system. | 06:32 |
| Maxdamantus | Though I think the aim of the separate n900 branch is kind of to be able to run Maemo pretty much as-is. | 06:33 |
| Maxdamantus | but I haven't really tried that much. | 06:33 |
| ZetaR | Right. I wonder if we could get the latest in the 2.6.x branch though. | 06:34 |
| Maxdamantus | stryngs: I normally use my phone as normal, using a 2.6.28 kernel. | 06:34 |
| Maxdamantus | Though I've been able to send/receive text messages using ofono on a 3.14 mainstream kernel. | 06:34 |
| Maxdamantus | (in a normal Debian system) | 06:35 |
| Maxdamantus | It was just using the example scripts that came with ofonod, didn't have a fancy UI. | 06:35 |
| Maxdamantus | Supposedly you should be able to use empathy or something, but that seemed too complicated to set up. | 06:36 |
| ZetaR | I would leave major kernel upgrades to the people trying to build Maemo from scratch. The unification could contribute to that sort of complete rebuild too. | 06:36 |
| DocScrutinizer05 | I see a lot of enthusiasm and excitement for leete new kernels and swapping around files between NAND, eMMC and possibly even uSD. I'm happy with the way my phone works and I really don't care which kernel runs on it and where the files get stored | 06:36 |
| Maxdamantus | I care which kernel runs mostly because I want to use btrfs. | 06:36 |
| ZetaR | Same. | 06:36 |
| Maxdamantus | It'd also be nice to run an up-to-date Debian system. | 06:37 |
| Maxdamantus | I currently just have a separate root for that that I can chroot into. | 06:38 |
| Maxdamantus | I notice I was able to run Firefox and Chromium from that with the Maemo graphics drivers, but that seemed to cause random reboots. | 06:39 |
| stryngs | So. quick poll, who ssh's into their n900? | 06:47 |
| * Maxdamantus does. | 06:47 | |
| ZetaR | I do not atm. | 06:48 |
| Maxdamantus | quick control poll, who's reading this but not going to respond to the other poll? | 06:48 |
| stryngs | Maxdamantus: How do you ssh in? usb or wifi? | 06:48 |
| stryngs | This is going somewhere, I promise | 06:48 |
| stryngs | ZetaR: Why don't ya? | 06:49 |
| Maxdamantus | Usually WiFi. If I want to transfer stuff faster I'll plug it in using USB but I'll still initiate the transfer over WiFi. | 06:49 |
| stryngs | Maxdamantus: Interesting | 06:50 |
| ZetaR | stryngs: I haven't found that I need it that much. | 06:50 |
| stryngs | Maxdamantus: Do you play with monitor mode on your n900? | 06:50 |
| ZetaR | But I am not doing the sort of stuff that Maxdamantus is. | 06:50 |
| Maxdamantus | Probably not, since I don't know what monitor mode is. | 06:50 |
| stryngs | Maxdamantus: Packet injection, via wifi | 06:50 |
| stryngs | Anyways, curious of ya'lls take on this: http://termbin.com/psc5 | 06:51 |
| Maxdamantus | Ah, no. | 06:51 |
| ZetaR | stryngs: If you are thinking of having sshd by default, IMO it should never be on by default. Though I guess it could be a default-off installer option. | 06:51 |
| stryngs | ZetaR: not at all. | 06:52 |
| ZetaR | Okay. | 06:52 |
| stryngs | ZetaR: It's a tunnel program I wrote. | 06:52 |
| stryngs | ZetaR: I tired of having to push keys and type, so i made it icon clickable | 06:52 |
| stryngs | ZetaR: These are some of the things I'm trying to push with unification. | 06:53 |
| stryngs | ZetaR: Ease of Use.. | 06:53 |
| Maxdamantus | If you are thinking of having sshd by default, you should provide a `passwd` tool that doesn't just use crypt(3) | 06:53 |
| stryngs | Maxdamantus: Way ahead of ya, and no, no sshd by default, but i do use crypt -m to change crypt(3) | 06:54 |
| Maxdamantus | Try sshing in and using just the first 8 characters. | 06:54 |
| Maxdamantus | (of your password) | 06:54 |
| DocScrutinizer05 | PasswordAuthentication no | 07:06 |
| Maxdamantus | Yeah, you could do that too. | 07:06 |
| DocScrutinizer05 | you can't use other encryption since some programs don't know any other | 07:10 |
| Maxdamantus | md5 works. | 07:11 |
| DocScrutinizer05 | we tried to fix that issue like 4 years ago and gave up | 07:11 |
| stryngs | chpasswd -m | 07:11 |
| stryngs | It works... | 07:11 |
| DocScrutinizer05 | aha, and which programs do work with such password? sudo? su? | 07:12 |
| stryngs | well...... | 07:12 |
| stryngs | I dunno | 07:12 |
| stryngs | I use sudser | 07:12 |
| * stryngs tries | 07:12 | |
| DocScrutinizer05 | tzz | 07:12 |
| Maxdamantus | I haven't used either where they ask for a password. | 07:12 |
| stryngs | foo | 07:13 |
| stryngs | Nokia-N900:/root/drop/pwnies$ su | 07:13 |
| stryngs | su: must be suid to work properly | 07:13 |
| stryngs | hmm | 07:13 |
| Maxdamantus | I just type `root`, which I think does something like `sudo gainroot`, which doesn't require a password for user. | 07:13 |
| DocScrutinizer05 | failwhale | 07:13 |
| DocScrutinizer05 | mine DOES | 07:13 |
| *** protem has quit IRC | 07:13 | |
| DocScrutinizer05 | ROOT password, not that idiotic user passowrd asking for getting root | 07:14 |
| Maxdamantus | Just for gainroot, or for everything? | 07:15 |
| DocScrutinizer05 | for su | 07:15 |
| Maxdamantus | I think Maemo uses sudo for a bunch of things. | 07:15 |
| Maxdamantus | Ah, so not sudo. | 07:15 |
| DocScrutinizer05 | sudo itself never needs a password, it's +s | 07:16 |
| DocScrutinizer05 | that's the purpose of sudo | 07:16 |
| Maxdamantus | sudo asks for a password if it's configured to. | 07:16 |
| DocScrutinizer05 | aha! | 07:16 |
| DocScrutinizer05 | guess what I did | 07:16 |
| DocScrutinizer05 | ~jrtools | 07:16 |
| infobot | jrtools is probably http://wiki.maemo.org/User:Joerg_rw/tools | 07:16 |
| DocScrutinizer05 | # fix NOPASSWD for root / sudo gainroot | 07:18 |
| Maxdamantus | Right, so that's just for gainroot. | 07:18 |
| Maxdamantus | if there's an exploit in any of the other things still not requiring passwords, then .. | 07:19 |
| Maxdamantus | user ALL = NOPASSWD: /usr/bin/vim * | 07:19 |
| Maxdamantus | modify any file, vim also lets you execute commands easily. | 07:20 |
| Maxdamantus | sudo vim | 07:20 |
| Maxdamantus | :!bash | 07:20 |
| DocScrutinizer05 | there are exploits | 07:21 |
| DocScrutinizer05 | but sudo vim is not allowed afaik | 07:22 |
| Maxdamantus | Well, I didn't add that. | 07:22 |
| Maxdamantus | Something else must've, and if you have vim installed I suspect you can do the same thing still. | 07:23 |
| Maxdamantus | unless you've explicitly removed that. | 07:23 |
| stryngs | DocScrutinizer05: Hmm...... You've motivated me to create a new .deb =) | 07:23 |
| DocScrutinizer05 | http://paste.opensuse.org/38121768 | 07:24 |
| stryngs | DocScrutinizer05: sudo-on-off_1.0_armel.deb | 07:24 |
| DocScrutinizer05 | no vim | 07:24 |
| Maxdamantus | DocScrutinizer05: look at /etc/sudoers | 07:24 |
| DocScrutinizer05 | and no vim in any of the other files in sudoers.d | 07:24 |
| Maxdamantus | Seems to be part of `filebox.sudoers` | 07:25 |
| * Maxdamantus thinks he knows what that is. | 07:25 | |
| DocScrutinizer05 | WTF??? | 07:26 |
| DocScrutinizer05 | yes | 07:26 |
| Maxdamantus | Yeah, that's an alternative file browser. | 07:26 |
| Maxdamantus | it also sucks about as much as the osso one. | 07:26 |
| stryngs | DocScrutinizer05: Question for you. using chpasswd -m. You're saying this breaks things? | 07:26 |
| * Maxdamantus will just uninstall it. | 07:26 | |
| DocScrutinizer05 | it did when we checked. At least we checked if we could use any other password hashes in etc/passwd and it turned out some binaries had no support for others | 07:27 |
| DocScrutinizer05 | prolly messybox | 07:27 |
| DocScrutinizer05 | or even sudo itself | 07:28 |
| DocScrutinizer05 | dunno | 07:28 |
| stryngs | DocScrutinizer05: Interesting. So with your current method of invoking a password whenever something wants sudo. You're stuck with the 8 char issue, yes? | 07:28 |
| Maxdamantus | sudo with a password works for me with md5. | 07:28 |
| DocScrutinizer05 | it's basically a nice-to-fix but no real problem | 07:28 |
| DocScrutinizer05 | after all what's the attack vector? | 07:29 |
| DocScrutinizer05 | somebody hacking the root password on a local xterm? | 07:29 |
| DocScrutinizer05 | there are huger flaws than simple 8char limit in password | 07:30 |
| Maxdamantus | If you want to use sshd. | 07:30 |
| Maxdamantus | (without authorized_keys) | 07:30 |
| DocScrutinizer05 | who want that? | 07:30 |
| DocScrutinizer05 | :-D | 07:31 |
| DocScrutinizer05 | ultra inconvenient and known to be not safe | 07:31 |
| stryngs | DocScrutinizer05: You've motivated me to create a new .deb =) | 07:31 |
| stryngs | DocScrutinizer05: It will unlock or lock the sudo part of the n900 =) | 07:31 |
| DocScrutinizer05 | oh my god!! | 07:31 |
| stryngs | DocScrutinizer05: i.e. Most of my icons use root, and so my system is unlocked by default. But now, with a click of the button, I can require a password for usage =),, and with a click of the button, turn it off. | 07:32 |
| stryngs | DocScrutinizer05: Remember, I play with pentest stuff, so having to type a password for each icon, is lame. | 07:32 |
| stryngs | DocScrutinizer05: Point being, good idea =) | 07:32 |
| DocScrutinizer05 | BWAHAHA you're kidding me, right? | 07:32 |
| stryngs | DocScrutinizer05: Nope. | 07:32 |
| stryngs | DocScrutinizer05: I hack with my n900 | 07:32 |
| DocScrutinizer05 | great, I thought I'll eventually get around to lock down HAM for good, so nobody can uninstall and reinstall rootsh. Now it seems I only need to install that great new deb pkg via HAM to make my efforts to lock down sudo gainroot futile | 07:34 |
| DocScrutinizer05 | on the click of an icon | 07:35 |
| Maxdamantus | or filebox. | 07:35 |
| stryngs | DocScrutinizer05: bwhahaha =) | 07:35 |
| stryngs | DocScrutinizer05: It's still a good idea for my purposes =) | 07:35 |
| stryngs | DocScrutinizer05: All someone has to do, to trash my system, is click my root shell icon and p00f, they got root on da box. | 07:36 |
| stryngs | DocScrutinizer05: Not like "the man" is waiting to do so, but if i travel out of the country, it's a possibility | 07:36 |
| stryngs | DocScrutinizer05: Just never thought to lock it down, tbh. | 07:36 |
| DocScrutinizer05 | so yeah, pretty please make a package out of that so everybody can fuck theri device on the click of an icon | 07:36 |
| stryngs | DocScrutinizer05: I don't think it would fuck the device, not if done right. I've got to play with it, but I think it would work. Whether or not md5 is an issue, is a seperate discussion. | 07:38 |
| stryngs | i.e. | 07:38 |
| stryngs | http://termbin.com/ivzy | 07:40 |
| stryngs | From your work | 07:40 |
| Maxdamantus | Interesting. Just realised VDQ actually stays across reboots. | 07:40 |
| stryngs | DocScrutinizer05: From your work, make a simple script, that ALWAYS boots to safe mode. Now, when a user wants to relax permissions for the current power cycle, they simply click an icon, this will relax permissions. Upon powerdown, permissions go back into effect. | 07:41 |
| stryngs | DocScrutinizer05: It'll be made clearer when I make it, but I'm gunna make it =) | 07:41 |
| DocScrutinizer05 | you should stop ircing while drincing | 07:41 |
| DocScrutinizer05 | drinking even | 07:41 |
| stryngs | It's hard to explain what I'm talking about DocScrutinizer05 =) | 07:42 |
| stryngs | DocScrutinizer05: Just know you motivated me to make something, useful. | 07:42 |
| stryngs | DocScrutinizer05: And leave my Bud Ice out of it! =) | 07:42 |
| * stryngs tosses back a 20oz. | 07:42 | |
| Maxdamantus | Hmm .. booting does seem a lot quicker using the ext4 module instead of ext3, though I haven't been timing anything. | 07:42 |
| Maxdamantus | (ie, just mounting with `-t ext4`, not enabling any ext4 features) | 07:43 |
| DocScrutinizer05 | "do you want to use a password protected root account [1] or a fuckedup open root account[2]? Both have same permissions" | 07:43 |
| DocScrutinizer05 | "please note we'll ask you same question again for first login to root after next reboot" | 07:44 |
| stryngs | DocScrutinizer05: chmod 4755 /bin/busybox | 07:45 |
| stryngs | DocScrutinizer05: Problem solved | 07:45 |
| DocScrutinizer05 | stop trolling please! | 07:46 |
| stryngs | DocScrutinizer05: ? I thought I was helping? I was diagnosing my use of chpasswd -m. You said it would break things, and it did. I googled my error, and it was a setuid one, as root i did 4755 to busybox, and it allowed me to su via a non privileged user. Isn't this a good thing? | 07:47 |
| Maxdamantus | I suspect you're meant to just copy busybox as chpasswd then set +s on that copy if you want to use it from other users. | 07:50 |
| Maxdamantus | but dunno. | 07:50 |
| DocScrutinizer05 | NO YOU MUST NOT SET SUID ON MESSYBOX!!1!11!! | 07:52 |
| stryngs | DocScrutinizer05: Why? | 07:52 |
| DocScrutinizer05 | for the same reason you don't set SUID on /*bin/* | 07:54 |
| DocScrutinizer05 | because of `ls -l /bin/* /sbin/* /usr/bin/* /usr/sbin/*|grep -- '-> busybox'` | 07:57 |
| Maxdamantus | Hm, chpasswd for me comes from the passwd package, not busybox. | 07:58 |
| DocScrutinizer05 | right, I never even heard of that command# | 07:58 |
| Maxdamantus | but busybox does also have its own chpasswd built in. | 07:59 |
| Maxdamantus | There's just no symlink for it. | 07:59 |
| Maxdamantus | Maybe stryngs generated all the symlinks. | 07:59 |
| DocScrutinizer05 | "built in" is exactly the point | 08:00 |
| DocScrutinizer05 | no matter what 'command' you execute, it all runds SUID busybox | 08:00 |
| stryngs | Alright, perhaps I am a little bit tipsy. Now we're delving into security, and rather than provide troll content as DocScrutinizer05 has suggested, I'm going to leave the convo for the meantime. I want to keep moving this project forward, not have it stall on the account of my silliness =) | 08:00 |
| DocScrutinizer05 | runs* | 08:00 |
| stryngs | I'll pick up at this point tomorrow =), and tbh, Maxdamantus there are TONS of symlinks on my n900; its a bastard child from PwnPhone v2. | 08:01 |
| * stryngs waves | 08:01 | |
| * DocScrutinizer05 waves too | 08:02 | |
| DocScrutinizer05 | busybox-symlinks-procps - dummy package to trick the CSSU metapackage \o/ | 08:04 |
| Maxdamantus | DocScrutinizer05: just wondering, is there meant to be something to replace bq27200.sh when you're using the bq27x00_battery module? | 08:05 |
| DocScrutinizer05 | err, no not really | 08:05 |
| DocScrutinizer05 | except the module itself and the /sys/*/*/*/* files it creates | 08:05 |
| Maxdamantus | I just added `eval "$(sed 's/^/registers[$((/; s/=/))]=/' </sys/class/power_supply/bq27200-0/registers)"; ./i2cget (){ echo "${registers[$(($4))]}"; }` to it. | 08:05 |
| Maxdamantus | Though the /sys files don't seem to provide as much information, unless you know how to interpret the registers, which that script does. | 08:07 |
| DocScrutinizer05 | if I had my way in that midule, it had a sys/*/*/*/dump | 08:07 |
| Maxdamantus | it has sys/*/*/*/registers, if that's what you mean. | 08:07 |
| DocScrutinizer05 | dunno, maybe | 08:08 |
| Maxdamantus | which is what my modification uses. It just returns values from that instead of using i2c directly. | 08:08 |
| DocScrutinizer05 | http://maemo.cloud-7.de/maemo5/usr/local/sbin/bq27k-detail2 | 08:09 |
| DocScrutinizer05 | dmppath=`find /sys/devices/ -path '*/hdq/*' -name dump` | 08:09 |
| DocScrutinizer05 | dmpcmd="cat ${1:-${dmppath}}| tr -s '\n ' ' '" | 08:10 |
| DocScrutinizer05 | else | 08:11 |
| DocScrutinizer05 | dmpcmd="i2cdump -y 2 0x55 b|tail -n +2|cut -d ' ' -f 2-17" | 08:11 |
| DocScrutinizer05 | n8 | 08:12 |
| DocScrutinizer05 | what a 'new maemo' *really* could use was for example a more recent version of sudo, which knows about genuine sudoers.d | 08:19 |
| DocScrutinizer05 | so we could get rid of that unspeakable update-sudoers and the crap in /etc/sudoers file, and iirc it needs a newer version as well for the password-query-client function (or whatever been the name for that little binary that gets called by sudo when a password query is needed) | 08:21 |
| DocScrutinizer05 | then we could not only secure HAM with root password, but actually start with s/NOPASSWD/PASSWD/g and see what doesn't make sense and only annoys and revert those to NOPASSWD | 08:27 |
| *** hubutm20 has joined #maemo | 08:28 | |
| * Maxdamantus doesn't really want to run maemo. | 08:30 | |
| * Maxdamantus just needs a suitable WM and appropriate telephony software. | 08:30 | |
| Maxdamantus | can probably write the former in xmonad. | 08:30 |
| DocScrutinizer05 | actually we could add a comment like #legit-nopasswd to some files in /etc/sudoers.d/* and then run a bootjob like sed -i 's/NOPASSWD/PASSWD/' $(grep -L '#legit-nopasswd') | 08:32 |
| * Maxdamantus wonders if there are any application container projects. | 08:47 | |
| Maxdamantus | Where the sorts of applications you install on phones are executed under different filesystem namespaces as different users (or in different user namespaces), etc | 08:48 |
| * Maxdamantus wonders what levels of isolation applications are undern in Android and iOS. | 08:53 | |
| * Maxdamantus suspects not much in Android. | 08:53 | |
| *** _rd has joined #maemo | 09:05 | |
| Wizzup | use a proper rbac | 09:07 |
| Wizzup | none of that containenr stuff | 09:07 |
| Maxdamantus | That's potentially what I mean by containers. | 09:21 |
| Maxdamantus | POSIX already has containment, eg, between uids. | 09:22 |
| KotCzarny | sshd on device is a must for any bigger job/hacking | 09:29 |
| KotCzarny | but yeah, having no root passwd is asking for troubles | 09:30 |
| KotCzarny | Maxdamantus: do you use journal for that ext3 ? | 09:30 |
| Maxdamantus | KotCzarny: yes, the default journalling mode. | 09:37 |
| KotCzarny | i just cant wait for my secondary n900 to start hacking | 09:38 |
| KotCzarny | also, i think things should be kept as simple as possible, last time i tried overlayfs it was slowing everything noticably (it was on a 200mhz 486 tho) | 09:42 |
| KotCzarny | in my idea nand would be used for boot kernels, boot menu, rescuefs etc | 09:43 |
| KotCzarny | just dont know how much to use as / and how much for the /home/user/MyDocs | 09:44 |
| Maxdamantus | Why do you need ~/MyDocs? | 09:44 |
| KotCzarny | can usb module export only one partition instead of whole device? | 09:44 |
| KotCzarny | Maxdamantus: for above, usb connectivity | 09:45 |
| Maxdamantus | The reason it's separate normally is so it can be mounted over USB, but that's a silly idea and it looks like it's implemented incorrectly. | 09:45 |
| Maxdamantus | You can just create a file on the filesystem and export that as a UMS device. | 09:45 |
| KotCzarny | ums device? | 09:46 |
| KotCzarny | ~ums | 09:46 |
| KotCzarny | or maybe we can just run samba and export it via usb networking | 09:46 |
| KotCzarny | without all that mount/umount tricks | 09:47 |
| Maxdamantus | USB mass storage | 09:47 |
| Maxdamantus | Yes, that works too. | 09:47 |
| Maxdamantus | or sshfs, if you're running sshd. | 09:47 |
| Maxdamantus | (less configuration that way) | 09:47 |
| KotCzarny | ssh eats cpu cycles | 09:47 |
| KotCzarny | but could be workable | 09:47 |
| Wizzup | does it? | 09:47 |
| Wizzup | isn't it just select() | 09:47 |
| KotCzarny | yes, encryption | 09:47 |
| Wizzup | ah, sure... | 09:48 |
| KotCzarny | and windows doesnt have sshfs | 09:48 |
| KotCzarny | which is a pity | 09:48 |
| Wizzup | it does | 09:48 |
| Maxdamantus | I think you can use filezilla. | 09:48 |
| KotCzarny | wizzup, link? | 09:48 |
| Wizzup | using dokan | 09:48 |
| Wizzup | dokandev | 09:48 |
| KotCzarny | nice | 09:49 |
| KotCzarny | last time i tried (few years ago) there wasnt one | 09:49 |
| Maxdamantus | dokan translates other filesystems to smb/cifs, iirc | 09:49 |
| Maxdamantus | (which is supported by Windows of course) | 09:49 |
| KotCzarny | nice | 09:50 |
| KotCzarny | maybe we can export little ums device that would have some autorun.bat :> | 09:50 |
| KotCzarny | that will be created dnamically and enabled usb networking on windoze | 09:51 |
| KotCzarny | and opened explorer window | 09:51 |
| * Maxdamantus just uses it for bootable OS images. | 09:51 | |
| KotCzarny | that way all gotchas about unclean fs would reduce only to unexpected shutdowns | 09:52 |
| KotCzarny | and journal could be disabled | 09:52 |
| Maxdamantus | The journal is there for unexpected shutdowns. | 09:54 |
| KotCzarny | does enabling usb networking on n900 and connecting to windows work out of the box or requires some configuration/user interaction? | 10:04 |
| Maxdamantus | It'd probably work without configuration if you run a DHCP server on the N900. | 10:05 |
| KotCzarny | cool | 10:06 |
| KotCzarny | very nice | 10:06 |
| Wizzup | KotCzarny: I found that some windows versions do not have usbnet support | 10:09 |
| Wizzup | And if you're without internet, you also cannot fetch it | 10:10 |
| Wizzup | This was windows 7 | 10:10 |
| KotCzarny | thats why i also though about autorun.bat | 10:10 |
| KotCzarny | and providing the driver just in case | 10:10 |
| KotCzarny | *thought | 10:10 |
| * Maxdamantus wonders when things like USB are going to be replaced by ethernet. | 10:26 | |
| KotCzarny | never? | 10:26 |
| KotCzarny | usb is owned by intel | 10:26 |
| KotCzarny | and they are making moneys out of it | 10:27 |
| KotCzarny | so few things would have to happen, intel going out of bussiness, ethernet starting providing power for attached devices, connectors/protocol going mobile friendly etc | 10:28 |
| KotCzarny | i think wifi was an attempt in what you wanted | 10:29 |
| Maxdamantus | It just seems weird having these special connections for "peripherals" and "drives" or something. | 10:30 |
| Maxdamantus | When all you're doing is networking devices. | 10:30 |
| DocScrutinizer05 | USB<->ethernet adapter for N900 | 10:32 |
| KotCzarny | old as world rule, if something is unclear for you, it's about moneys | 10:32 |
| * Maxdamantus also wonders when people are going to stop using money. | 10:33 | |
| KotCzarny | we are already on the way | 10:33 |
| DocScrutinizer05 | is today world money day? | 10:33 |
| KotCzarny | watch few infomercials on youtube about "fiat money" | 10:33 |
| DocScrutinizer05 | seems everybody but me does, eh? | 10:34 |
| DocScrutinizer05 | they talk about money in several channels | 10:34 |
| KotCzarny | doc, thats because my secondary n900 would only arrive next week | 10:34 |
| KotCzarny | maxd: usb is also about charging devices | 10:35 |
| DocScrutinizer05 | >>das ist prinzipiell gut, wenn jemand sagt: Geld interessiert mich nicht<< | 10:35 |
| KotCzarny | that's its second most important function today | 10:35 |
| DocScrutinizer05 | >>wir nähern uns einer idealen gesellschaft an.<< | 10:35 |
| KotCzarny | doc: if you have enough moneys for other tasks, yes | 10:36 |
| kerio | yeah let's turn a perfectly serviceable master-slave protocol into some weird bidirectional networking crap | 10:36 |
| Maxdamantus | KotCzarny: ethernet can potentially do that as well. | 10:36 |
| *** xorly has joined #maemo | 10:36 | |
| KotCzarny | maxd: ethernet wasnt designed with that in mind | 10:36 |
| kerio | i'm sure nothing will go wrong with that | 10:36 |
| KotCzarny | it was about as you say 'networking devices' | 10:36 |
| KotCzarny | not about one cable fits them all | 10:36 |
| Maxdamantus | until a few days ago I had my USB peripherals attached over ethernet. | 10:37 |
| Maxdamantus | well, ethernet cable, at least. | 10:37 |
| KotCzarny | if you start adding features to it, you will enter incompatibility hell | 10:37 |
| DocScrutinizer05 | meh USB-C hot shit ;-P | 10:37 |
| KotCzarny | and sure, you can use 2 pairs for your other tasks, but. | 10:37 |
| KotCzarny | doc, usb is licensed :P | 10:38 |
| DocScrutinizer05 | s/hot// | 10:38 |
| infobot | DocScrutinizer05 meant: meh USB-C shit ;-P | 10:38 |
| kerio | DocScrutinizer05: how do you feel about the new macbook? | 10:38 |
| DocScrutinizer05 | feel? what's that? | 10:38 |
| KotCzarny | he asks about your moneys | 10:38 |
| KotCzarny | probably wants to sell you something | 10:39 |
| DocScrutinizer05 | tbh I have no clue what's the "new macbook" | 10:39 |
| kerio | it's a netbook that costs 1.2k | 10:39 |
| DocScrutinizer05 | so it's a tad hard to have feelings for it | 10:39 |
| Maxdamantus | (I soldered an ethernet cable to a USB header plug and some USB ports, as an alternative to using a hub or multiple extension cables) | 10:39 |
| kerio | it has one usb C port, and one headphone jack | 10:39 |
| kerio | and that's it | 10:39 |
| Maxdamantus | but now I'd like to use that other internal USB port. | 10:40 |
| KotCzarny | maxd: o.O | 10:40 |
| Maxdamantus | er, s/port/header/ | 10:40 |
| kerio | not even a specialized charging port | 10:40 |
| kerio | the charger is usb C as well | 10:40 |
| DocScrutinizer05 | less holes \o/ | 10:40 |
| DocScrutinizer05 | does it also have less keys? | 10:40 |
| KotCzarny | i dont mind holes | 10:40 |
| KotCzarny | holes are nice to put things in | 10:41 |
| DocScrutinizer05 | yes ::-) | 10:41 |
| KotCzarny | and it feels good usually | 10:41 |
| kerio | hm, no | 10:41 |
| kerio | but the keyboard has some supposedly better keys | 10:41 |
| KotCzarny | so, in a way, they are sterilizing it | 10:41 |
| KotCzarny | thinkpad keys > everything | 10:41 |
| DocScrutinizer05 | it's just this was the common industry paradigm even 5 years ago already. Heck 7 years | 10:41 |
| KotCzarny | real ibm thinkpad keys, not chinovo ones | 10:41 |
| kerio | the new macbook is probably as tall as a thinkpad key | 10:41 |
| DocScrutinizer05 | less keys less holes | 10:41 |
| kerio | :> | 10:42 |
| kerio | yea in a way it's similar to an embedded device | 10:42 |
| KotCzarny | i wonder when they split macbook in a similar way as ipod went | 10:42 |
| kerio | phones are really popular, right | 10:42 |
| KotCzarny | one with screen and one without | 10:42 |
| kerio | ROFL | 10:42 |
| KotCzarny | macbook shuffle | 10:43 |
| KotCzarny | you turn it on, it does random things | 10:43 |
| DocScrutinizer05 | stay patient - can't take long. Tehy already started their own competirot product to Oculus Rift | 10:43 |
| KotCzarny | user is happy because it looks cool | 10:43 |
| DocScrutinizer05 | competitor | 10:44 |
| KotCzarny | competiriot | 10:44 |
| kerio | is the ipod even a thing anymore | 10:44 |
| KotCzarny | never had one, apart from the 5$ find on some garage sale | 10:44 |
| kerio | well, apparently they keep updating the ipod touch in lockstep with the iphone | 10:45 |
| KotCzarny | hmm, i wonder how dpkg handles situation when /opt is a symlink/bind into /usr | 10:48 |
| KotCzarny | and there are files in /opt at similar place as same name symlinks in /usr | 10:49 |
| *** TabTwo has joined #maemo | 10:58 | |
| *** _rd has quit IRC | 11:13 | |
| *** Kabouik_ has joined #maemo | 11:17 | |
| *** vectis has joined #maemo | 11:17 | |
| *** Kabouik has quit IRC | 11:19 | |
| *** _rd has joined #maemo | 11:20 | |
| *** M4rtinK has joined #maemo | 11:29 | |
| *** _rd has quit IRC | 11:34 | |
| KotCzarny | hmm, nice thing about partitioning would be having stock/cssu/cssu-t all on one device for testing | 11:38 |
| KotCzarny | (partitioning and moving to one / without nand) | 11:38 |
| KotCzarny | or even multiple sd cards with different flavors | 11:39 |
| *** _rd has joined #maemo | 11:41 | |
| KotCzarny | hmm, anyone had a problem that ussd codes stopped working even if there is cellular connectivity? (sms and calls work) | 11:47 |
| KotCzarny | it says 'cellular network not available' | 11:47 |
| *** `0660 has quit IRC | 11:48 | |
| *** `0660 has joined #maemo | 11:48 | |
| *** Pali has joined #maemo | 11:50 | |
| *** futpib has joined #maemo | 11:59 | |
| *** LauRoman|Alt has quit IRC | 11:59 | |
| *** florian has joined #maemo | 12:02 | |
| *** xorly has quit IRC | 12:05 | |
| KotCzarny | sending the code through pnatd works tho, hum | 12:11 |
| KotCzarny | and i didnt update/install anything recently, why is dialer acting up out of sudden? | 12:11 |
| kerio | did you install starhash enabler | 12:12 |
| KotCzarny | nope | 12:13 |
| KotCzarny | but it was working before | 12:13 |
| KotCzarny | thats why i didnt bother with starhash enabler | 12:13 |
| DocScrutinizer05 | KotCzarny: check out *#ena, it's a binary-less packahe :-) | 12:15 |
| KotCzarny | um? | 12:15 |
| DocScrutinizer05 | pretty simple to grok what postinstall does and do same thing manually, resp check if it's needed | 12:15 |
| KotCzarny | http://pastebin.com/raw.php?i=d7VcZaV6 | 12:16 |
| DocScrutinizer05 | starhash enabler is a glorified shellscript with a button to launch | 12:16 |
| DocScrutinizer05 | once, at install time | 12:16 |
| KotCzarny | doc, but it was working few days ago | 12:17 |
| KotCzarny | just stopped out of sudden | 12:17 |
| KotCzarny | and the message is baffling 'cellular network is not available' which is clearly a lie | 12:17 |
| KotCzarny | at worst i can just write something in pygtk, but it was working not so long time ago | 12:18 |
| DocScrutinizer05 | http://repository.maemo.org/extras/pool/fremantle/free/source/s/starhash-enabler/starhash-enabler_0.2.tar.gz | 12:19 |
| DocScrutinizer05 | http://paste.opensuse.org/1328588 | 12:19 |
| DocScrutinizer05 | echo -e "[supplementary]\nssc=1" >> /home/user/.osso/call-ui.ini | 12:19 |
| KotCzarny | done, restarted rt-com ui, nothing changed | 12:20 |
| KotCzarny | same error | 12:20 |
| KotCzarny | could it be my network provider did something funky? | 12:21 |
| *** MikaT has quit IRC | 12:21 | |
| KotCzarny | at+cusd=1,"*100#",1 didnt work, at+cusd=1,"*100#",15 did (in pnatd) | 12:21 |
| DocScrutinizer05 | 15 sounds familiar | 12:22 |
| *** Oksanaa has joined #maemo | 12:23 | |
| DocScrutinizer05 | isn't it like 1 means first message, 15 means answer to a previous message? | 12:23 |
| DocScrutinizer05 | I've seen USSD fail when I didn't answer a "multiple choice" request. It usually 'fixes itself' after some 5 min of timeout, or after a boot the latest | 12:24 |
| KotCzarny | booted just yesterday when device went empty in 5s after enabling edv1=1 | 12:25 |
| DocScrutinizer05 | killall -9 rtcom-call-ui | 12:25 |
| KotCzarny | did that too | 12:25 |
| DocScrutinizer05 | the -9 is since otherwise it reverts any edits you did to /home/user/.osso/call-ui.ini | 12:26 |
| KotCzarny | yes, but i did that | 12:26 |
| DocScrutinizer05 | recheck /home/user/.osso/call-ui.ini | 12:26 |
| *** arossdotme has joined #maemo | 12:26 | |
| KotCzarny | and it wasnt the problem dialer not accepting *100# but erroring with 'cellular network not available' | 12:26 |
| DocScrutinizer05 | yes, I noticed that | 12:27 |
| KotCzarny | also, using at+cusd=1,"*100#" results in error too | 12:27 |
| DocScrutinizer05 | err, wait, 15 might be encoding which is *always* 15 | 12:28 |
| KotCzarny | http://mybroadband.co.za/vb/showthread.php/167974-What-is-a-syntax-of-AT-CUSD | 12:29 |
| DocScrutinizer05 | AT+CUSD=<ussdMode>[, <ussdReq>[, <ussdDCS>]] | 12:29 |
| KotCzarny | its clearly optional, but default is failing | 12:30 |
| DocScrutinizer05 | <ussdDCS> 3GPP TS 23.038 [37] Cell Broadcast Data Coding Scheme (default 15). ï€ | 12:31 |
| KotCzarny | anyway, that seems only related to pnatd, while dialer having different problem | 12:31 |
| *** xorly has joined #maemo | 12:32 | |
| DocScrutinizer05 | dialer needs nssc=1 | 12:34 |
| KotCzarny | it worked before | 12:34 |
| KotCzarny | just few days ago | 12:34 |
| DocScrutinizer05 | and it needs some plugin in the telephony stack | 12:34 |
| DocScrutinizer05 | /usr/sbin/csd -m -p call -p gprs -p info -p net -p sim -p simpb -p sms -p ss #*maybe*, the -p ss | 12:37 |
| KotCzarny | im missing non-castrated ps | 12:38 |
| DocScrutinizer05 | unrelated but funny: hald-addon-mmc: listening on /sys/class/mmc_host/mmc0/cover_switch | 12:38 |
| *** `0660 has quit IRC | 12:39 | |
| *** `0660 has joined #maemo | 12:39 | |
| KotCzarny | from /proc/cmdline: /usr/sbin/csd.-m.-p.call.-p.gprs.-p.info.-p.net.-p.sim.-p.simpb.-p.sms.-p.ss. | 12:39 |
| KotCzarny | and obviously ussd is available because it works via pnatd | 12:40 |
| *** florian has quit IRC | 12:42 | |
| DocScrutinizer05 | no idea. Try call redirection from settings. Try *#006#. Try *#67# | 12:52 |
| kerio | what's 67 for | 12:52 |
| DocScrutinizer05 | some redirect | 12:52 |
| DocScrutinizer05 | can't recall which one | 12:52 |
| DocScrutinizer05 | iirc | 12:53 |
| DocScrutinizer05 | 67:when busy (other call); | 12:53 |
| DocScrutinizer05 | *#21# all calls | 12:54 |
| DocScrutinizer05 | 004: eq: 61+62+67; | 12:54 |
| *** xorly has quit IRC | 12:54 | |
| DocScrutinizer05 | **61*<number>**<time-in-seconds,5,10,..,30># : set up redirect to <number> after <seconds> ringing | 12:54 |
| Maxdamantus | Mm .. turns out my SD card has pretty much always been mounted as ext2 -_- | 12:55 |
| kerio | LOL | 12:55 |
| Maxdamantus | stupid maemo. | 12:55 |
| Maxdamantus | I think it's now resulting in crashes. | 12:55 |
| Maxdamantus | Tried to run an fsck on it and it spent about 40 minutes fixing things before segfaulting. | 12:56 |
| DocScrutinizer05 | normal | 12:56 |
| DocScrutinizer05 | fsck had a bug, and generally it's extremely RAM greedy | 12:57 |
| DocScrutinizer05 | which means it gives the device a nice swap hell | 12:59 |
| DocScrutinizer05 | and may run literally for days | 12:59 |
| *** bef0rd has joined #maemo | 12:59 | |
| *** beford has quit IRC | 13:02 | |
| Maxdamantus | Mm, I think it's causing the rest of the system to misbehave too though. | 13:02 |
| * Maxdamantus has multiple places where it tries to mount it :\ | 13:03 | |
| *** bef0rd has quit IRC | 13:04 | |
| DocScrutinizer05 | 1h, 2 reasons, why I don't want a leete new but a rugged mature system for my phone | 13:05 |
| Maxdamantus | Linux 4.1 is maturer than Linux 2.6.28.10 | 13:05 |
| DocScrutinizer05 | mhm | 13:06 |
| bencoh | "The details of the referendum can be found at: http://talk.maemo.org/showthread.php?t=[#####]" | 13:06 |
| Maxdamantus | It probably has bug fixes for ext4. | 13:06 |
| bencoh | looks like something's wrong with the link | 13:06 |
| *** beford has joined #maemo | 13:06 | |
| *** `0660 has quit IRC | 13:06 | |
| DocScrutinizer05 | seems like quite a bit is wrong with the whole election | 13:06 |
| *** `0660 has joined #maemo | 13:06 | |
| DocScrutinizer05 | http://wstaw.org/m/2015/09/06/plasma-desktopVV1947.png | 13:07 |
| bencoh | hmmm | 13:08 |
| *** Oksanaa has quit IRC | 13:08 | |
| DocScrutinizer05 | I'm not surprised | 13:10 |
| KotCzarny | :) | 13:12 |
| KotCzarny | things falling apart | 13:12 |
| DocScrutinizer05 | nah, things not used the way they were meant to get used | 13:13 |
| KotCzarny | doc, as for your 'leeter os' im running stock fremantle | 13:14 |
| KotCzarny | didnt hack it because its my main phone | 13:14 |
| DocScrutinizer05 | that's pretty odd then | 13:14 |
| KotCzarny | but as i said, it was working few days ago | 13:15 |
| KotCzarny | so i consider network changing in a way it made n900 baffled | 13:15 |
| KotCzarny | well, not n900, dialer mostly | 13:15 |
| KotCzarny | where does it get information if cellular network is available? | 13:15 |
| KotCzarny | or in other words, does it do ussd directly or calls some other app | 13:16 |
| KotCzarny | btw. *#67# fails with the same error | 13:19 |
| KotCzarny | so it somehow fails to acknowledget that cellular is available for all ussd codes | 13:20 |
| KotCzarny | yet, calls/sms work as expected | 13:20 |
| KotCzarny | *#06# works, but i guess its handled differently from other codes | 13:20 |
| *** Kabouik has joined #maemo | 13:23 | |
| *** Kabouik_ has quit IRC | 13:25 | |
| *** Oksanaa has joined #maemo | 13:25 | |
| DocScrutinizer05 | try a different network (aka SIM) | 13:26 |
| DocScrutinizer05 | I mean, ebay PP payment fucked up, why shouldn't your carrier's USSD handling be fucked up | 13:28 |
| DocScrutinizer05 | obviously they set a wrong default encoding | 13:28 |
| DocScrutinizer05 | you could try sending a SMS to yourself | 13:29 |
| DocScrutinizer05 | or send a SMS to some carrier's services, often they have that stuff to query your prepaid bakance etc | 13:30 |
| *** merlin1991 has quit IRC | 13:39 | |
| *** MikaT has joined #maemo | 13:39 | |
| *** merlin1991 has joined #maemo | 13:39 | |
| Ras_Older | Good morning everyone! I could order a coffee to bed... *yawns* | 13:41 |
| *** obsed has quit IRC | 13:45 | |
| *** wnd has joined #maemo | 13:45 | |
| *** sparetire_ has joined #maemo | 13:58 | |
| *** RedW has quit IRC | 14:02 | |
| *** Oksanaa has quit IRC | 14:03 | |
| *** RedW has joined #maemo | 14:04 | |
| *** Oksanaa has joined #maemo | 14:06 | |
| *** Oksanaa has joined #maemo | 14:06 | |
| *** _rd has quit IRC | 14:34 | |
| *** _rd has joined #maemo | 14:39 | |
| *** Cor-Ai has quit IRC | 14:50 | |
| *** Cor-Ai has joined #maemo | 14:51 | |
| *** mavhc has quit IRC | 14:54 | |
| *** mavhc has joined #maemo | 14:57 | |
| Pali | hi! is there any guru-guru for XS language? | 15:09 |
| KotCzarny | xs? | 15:43 |
| *** _rd has quit IRC | 15:45 | |
| *** kerio has quit IRC | 15:48 | |
| *** fuz_ has quit IRC | 15:49 | |
| *** robink has quit IRC | 15:50 | |
| *** kerio has joined #maemo | 15:51 | |
| *** _rd has joined #maemo | 15:53 | |
| *** Oksanaa has quit IRC | 16:04 | |
| *** fuz_ has joined #maemo | 16:05 | |
| *** trumee has quit IRC | 16:07 | |
| *** _rd has quit IRC | 16:07 | |
| *** trumee has joined #maemo | 16:10 | |
| *** Oksanaa has joined #maemo | 16:12 | |
| *** KotCzarny has quit IRC | 16:13 | |
| *** _rd has joined #maemo | 16:17 | |
| *** KotCzarny has joined #maemo | 16:23 | |
| Pali | KotCzarny: yes XS | 16:26 |
| KotCzarny | never heard of it | 16:27 |
| Pali | that language in which you can write perl functions which can call native C functions | 16:27 |
| Pali | uses xsubpp compiler (XS to C) | 16:27 |
| KotCzarny | so, is there any wiki admin online? | 16:31 |
| KotCzarny | <a style="background-image: url(/var/lib/mediawiki/skins/common/images/maemo-logo.png);" href="/Main_Page" title="Visit the main page [alt-shift-z]" accesskey="z"></a> | 16:31 |
| KotCzarny | someone messed path a bit | 16:31 |
| KotCzarny | tough it might be a job for anyone with file access to wiki server | 16:32 |
| * Sicelo is by the way not in favor of the "image" concept being discussed here in the last few days. | 16:52 | |
| *** ssvb has quit IRC | 16:56 | |
| *** Oksanaa has quit IRC | 16:57 | |
| *** _rd has quit IRC | 17:06 | |
| *** trumee has quit IRC | 17:30 | |
| *** trumee has joined #maemo | 17:34 | |
| *** Tekk_ has quit IRC | 17:38 | |
| *** ssvb has joined #maemo | 17:56 | |
| *** xorly has joined #maemo | 17:57 | |
| *** LjL-Cat is now known as LjL | 18:05 | |
| *** trumee has quit IRC | 18:07 | |
| *** trumee has joined #maemo | 18:10 | |
| vectis | hmm, my N810 is stuck at the 'Nokia hands' screen, but I can ssh into it. Is it fixable? what to look for? | 18:12 |
| *** Venusaur has quit IRC | 18:13 | |
| sixwheeledbeast | Well I'd guess flashing would fix it. | 18:25 |
| sixwheeledbeast | I assume you don't wish to do that | 18:25 |
| *** Venusaur has joined #maemo | 18:28 | |
| kerio | vectis: first thing | 18:28 |
| kerio | do you have a way to charge the battery? | 18:28 |
| kerio | other than the N810 itself | 18:28 |
| vectis | I have another N810, so yes. | 18:29 |
| vectis | it's been up and working for days, so it's probably something i've built/installed recently | 18:31 |
| KotCzarny | hmm | 18:32 |
| Humpelstilzchen | or a filesystem corruption | 18:33 |
| KotCzarny | if you can ssh, you can check what is the latest pid | 18:33 |
| KotCzarny | then you will see how far did it go | 18:33 |
| KotCzarny | also, dmesg can provide a hint | 18:33 |
| *** beford has quit IRC | 18:36 | |
| vectis | ok, it's not the 'hands' rather the 'nokia' screen. dmesg gives the following: http://pastebin.com/3dKH3br1 | 18:46 |
| KotCzarny | paste process list maybe | 18:46 |
| vectis | process list = top? http://pastebin.com/xyNdZaWX | 18:49 |
| KotCzarny | ps axw | 18:50 |
| KotCzarny | but it looks like things have started | 18:50 |
| vectis | http://pastebin.com/9Z81nAfa | 18:51 |
| vectis | asui is running after I press the power button | 18:51 |
| KotCzarny | were you modifying startup scripts? | 18:52 |
| vectis | not to my knowledge, but it's been up for a week+ | 18:54 |
| KotCzarny | if you are feeling adventurous you may try killing things | 18:54 |
| KotCzarny | start from 1509 and decrease pids | 18:55 |
| KotCzarny | you may lose network connection, but apparently something is blocking | 18:55 |
| KotCzarny | you may also try installing some syslog and reboot | 18:55 |
| vectis | could you explain further? I can always rewrite the image file, but I'd like to find out the cause | 18:57 |
| KotCzarny | nothing special in your pastes | 18:57 |
| KotCzarny | looks like window manager and friends are started | 18:57 |
| KotCzarny | so, i would start poking in the dark: kill 1424 (you might be disconnected, so reboot might be required later) | 18:58 |
| *** kerio has quit IRC | 19:01 | |
| *** kerio has joined #maemo | 19:01 | |
| vectis | sudo kill 1424 still the same | 19:02 |
| KotCzarny | kill 1481 1479 | 19:03 |
| vectis | no change | 19:04 |
| KotCzarny | ok, lets go more radical: kill 1002 (should kill and restart window manager) | 19:05 |
| *** darkschneider has quit IRC | 19:07 | |
| *** darkschneider has joined #maemo | 19:08 | |
| vectis | still the same but I've lost asui now | 19:08 |
| KotCzarny | ok, even more radical, lets kill Xomap: kill 956 | 19:08 |
| Sicelo | out of curiosity, what's asui? | 19:09 |
| vectis | http://asui.garage.maemo.org/ | 19:11 |
| vectis | still no change :( | 19:11 |
| KotCzarny | ps ax and check if things restarted on their own | 19:12 |
| vectis | http://pastebin.com/gcmwb2Km | 19:13 |
| Sicelo | i think i'll buy an N810 if i ever come across one :) | 19:13 |
| KotCzarny | Xomap didnt restart | 19:14 |
| vectis | lol, just don't press the buttons in the sequence that I did :D | 19:14 |
| KotCzarny | hmm? | 19:14 |
| vectis | should I give up? | 19:18 |
| KotCzarny | dont remember command to restart xomap | 19:19 |
| vectis | lol, ok | 19:19 |
| KotCzarny | check scripts in /etc/init.rd | 19:19 |
| KotCzarny | then run one with restart as a param | 19:19 |
| KotCzarny | i mean /etc/init.d | 19:19 |
| KotCzarny | asui looks nice | 19:22 |
| *** _rd has joined #maemo | 19:26 | |
| Sicelo | [N900] .. i was just testing a few programs on N900 which exhibit extreme slowness, e.g. opening a large pdf file with lots of pictures. i always assumed that our slowness is due to low ram & having to swap a lot, but htop seems to indicate we just have a slow processor for those things. is my observation just placebo effect? | 19:28 |
| KotCzarny | iostat -k 1 | 19:28 |
| KotCzarny | or vmstat | 19:28 |
| KotCzarny | you will see | 19:28 |
| Humpelstilzchen | Sicelo: tried swappolube? | 19:28 |
| KotCzarny | remember that random write access to flash is extremely slow | 19:29 |
| KotCzarny | so even 10iops is plenty | 19:29 |
| vectis | hmm im getting command not found for 'start' and 'restart' | 19:30 |
| KotCzarny | use: ./somescript restart | 19:30 |
| vectis | /etc/init.d/x-server: line 50: dsmetool: not found | 19:31 |
| KotCzarny | hum | 19:33 |
| Sicelo | so we have slow chips? Humpelstilzchen i don't use swappolube, but occasionally do enable swap on uSD | 19:33 |
| vectis | hmm, I missed out sudo command. Xomap is restarted. http://pastebin.com/srC6y1UQ | 19:49 |
| KotCzarny | but nothing changed on screen? | 19:50 |
| vectis | no, all I have is a white screen | 19:53 |
| KotCzarny | no hands? | 19:53 |
| vectis | no. | 19:53 |
| KotCzarny | restart wm too | 19:53 |
| KotCzarny | should be in the startup scripts too | 19:53 |
| vectis | ok, thanks for your help | 19:53 |
| KotCzarny | its not final | 19:54 |
| KotCzarny | you still have to figure out what made it stop | 19:54 |
| KotCzarny | as you had x and wm running | 19:54 |
| KotCzarny | just not updating screen | 19:54 |
| vectis | I ran /etc/init.d/hildon-desktop script but no change | 19:58 |
| KotCzarny | with restart as option? | 19:58 |
| KotCzarny | restart matchbox (its the wm) | 19:58 |
| *** Cor-Ai has quit IRC | 19:59 | |
| *** Cor-Ai has joined #maemo | 20:02 | |
| *** robink_ has joined #maemo | 20:10 | |
| *** _rd has quit IRC | 20:13 | |
| vectis | I'm out of my depth here, I don't know how to restart matchbox :( | 20:14 |
| KotCzarny | what are the script names? | 20:15 |
| KotCzarny | (sorry, my n800 is sleeping quite far away) | 20:15 |
| Sicelo | restart didn't help? | 20:16 |
| vectis | http://pastebin.com/7xWwh3aF | 20:16 |
| KotCzarny | hum | 20:17 |
| KotCzarny | grep matchbox * ? | 20:17 |
| vectis | nothing | 20:18 |
| stryngs | Sicelo: Out of curiosity, why not in favor of the "unification"? | 20:19 |
| KotCzarny | vectis, hildon-desktop restart ? | 20:20 |
| vectis | yep, tried that already. restart is not implimented, only start. | 20:23 |
| vectis | I tried start as well | 20:23 |
| vectis | if I stop it, will just reboot? | 20:27 |
| KotCzarny | dont know what watchdog monitors | 20:28 |
| vectis | I think I should have mentioned that I have r&d mode enabled | 20:28 |
| vectis | hmm, stop and start still leaves me with a white screen | 20:30 |
| KotCzarny | r&d shouldnt stop things from showing up | 20:30 |
| KotCzarny | can you run xterm or some other x app? | 20:30 |
| vectis | osso-xterm returns no errors nor any change to the screen. | 20:33 |
| KotCzarny | hmm | 20:33 |
| KotCzarny | i assume you've rebooted few times once you got that white screen for the first time? | 20:34 |
| vectis | yer | 20:34 |
| KotCzarny | is there x log in /var/log ? | 20:35 |
| Sicelo | stryngs: i still don't understand what fragmentation is in this context. that's just the beginning though. | 20:36 |
| vectis | no its empty. should I just install syslog? | 20:36 |
| KotCzarny | vectis, nah, take note of X params, then kill it, and run from terminal maybe | 20:36 |
| Sicelo | stryngs: so i don't yet really know which problem unifications is attemptiong to solve .. XY? | 20:37 |
| vectis | KotCzarny: I don't understand :( | 20:38 |
| *** _rd has joined #maemo | 20:38 | |
| KotCzarny | vectis, right now its started and forks into background | 20:39 |
| KotCzarny | maybe it will print some debug to terminal | 20:39 |
| KotCzarny | /usr/bin/Xomap -mouse tslib -nozap -dpi 96 -wr -nolisten tcp | 20:39 |
| KotCzarny | dont know what -wr does | 20:39 |
| KotCzarny | but kill Xomap then start it directly from terminal (without script) | 20:41 |
| Sicelo | stryngs: i'm not saying it should not be done or abandoned .. | 20:42 |
| vectis | KotCzarny: I ran sudo killall Xomap then ran the command above. apart from some errors regarding keycodes, no change. | 20:44 |
| KotCzarny | vectis, somehow your xomap doesnt render things | 20:45 |
| *** Tekk_ has joined #maemo | 20:47 | |
| vectis | I guess it must be some software that I've compiled and installed. Trouble is that I've done quite a lot without a reboot. | 20:49 |
| KotCzarny | try reinstalling xomap package | 20:50 |
| KotCzarny | or xorg or xorg-server | 20:52 |
| KotCzarny | dpkg -l|grep -i xomap | 20:52 |
| KotCzarny | or xorg | 20:52 |
| KotCzarny | or just dpkg -S /usr/sbin/Xomap | 20:52 |
| KotCzarny | will give you name of the package containing it | 20:52 |
| *** M4rtinK has quit IRC | 20:52 | |
| *** protem has joined #maemo | 20:57 | |
| vectis | I re-installed libxsp0 and xserver-xomap then rebooted. No change though :( | 21:01 |
| KotCzarny | hmm, powerdown, take out battery, wait 20s, restart? | 21:02 |
| vectis | lol, I guess I should just overwrite the microsd with a copy of the image I made and start again. Just wish I could figure out how I broke it. | 21:05 |
| KotCzarny | reboot more often after serious system changes :) | 21:06 |
| vectis | hmm, I didn't think anything I was doing was that serious. | 21:07 |
| *** M4rtinK has joined #maemo | 21:21 | |
| *** _rd has quit IRC | 21:29 | |
| KotCzarny | hmm | 21:40 |
| KotCzarny | council people are english illiterate? | 21:40 |
| KotCzarny | or just drunk? | 21:40 |
| Sicelo | haha, why? | 21:40 |
| *** protem has quit IRC | 21:41 | |
| KotCzarny | Dear kotczarny, | 21:41 |
| KotCzarny | The Council apologises for the inconvenience caused by the fact that there were some glitches in the election machie, making this already the fourth attempt at setting up the Referendum. | 21:41 |
| KotCzarny | We are not spamming your inbox in purpose, and hope that this time it will work allright. | 21:41 |
| Sicelo | authentic? | 21:42 |
| KotCzarny | just got this email | 21:42 |
| KotCzarny | from 127.0.0.1 (EHLO mail.maemo.org) (213.128.137.23) | 21:44 |
| KotCzarny | root <root@www1.maemo.org> | 21:44 |
| KotCzarny | hehe | 21:44 |
| KotCzarny | using root on internet connected machine to send emails? why not | 21:45 |
| Sicelo | the email still gets dispatched ;) | 21:51 |
| KotCzarny | doesnt make them too pro though ;) | 21:53 |
| Sicelo | let's hope it's just a quirky alias | 21:57 |
| Sicelo | and not actual root account | 21:57 |
| KotCzarny | yeah, sure | 21:57 |
| KotCzarny | not to mention that sending that email just made them subscribed to bazilion of spam sites | 21:58 |
| KotCzarny | :) | 21:58 |
| KotCzarny | lets just hope root doesnt get any mail ;) | 21:58 |
| *** lxp has quit IRC | 22:07 | |
| *** lxp has joined #maemo | 22:07 | |
| *** zGrr has joined #maemo | 22:13 | |
| *** lxp has quit IRC | 22:20 | |
| *** lxp has joined #maemo | 22:24 | |
| *** ced117 has quit IRC | 22:25 | |
| *** xes_ has joined #maemo | 22:29 | |
| KotCzarny | xes_: can you modify files on wiki.maemo.org? | 22:30 |
| *** M4rtinK has quit IRC | 22:31 | |
| *** xes has quit IRC | 22:32 | |
| *** _rd has joined #maemo | 22:40 | |
| *** LauRoman has joined #maemo | 22:57 | |
| *** florian has joined #maemo | 23:03 | |
| *** SpeedEvil has quit IRC | 23:13 | |
| *** SpeedEvil has joined #maemo | 23:14 | |
| *** NIN101 has quit IRC | 23:20 | |
| *** NIN101 has joined #maemo | 23:20 | |
| vectis | KotCzarny: I've traced the problem to libcurl - must remember not to overwrite system files so late at night :) | 23:26 |
| *** aloril has quit IRC | 23:38 | |
| *** Oksanaa has joined #maemo | 23:39 | |
| *** aloril has joined #maemo | 23:39 | |
| KotCzarny | vectis, what? | 23:58 |
| KotCzarny | i knew maemo is very fragile, but libcurl? | 23:58 |
| KotCzarny | how does it blocked Xomap rendering? | 23:58 |
Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!
