**** BEGIN LOGGING AT Mon Feb 25 02:59:57 2013 Feb 25 06:48:03 hi there.got a new n900,try to repartition the internal mmc,but get only 7Gb of space Feb 25 06:48:25 how did you repartition? Feb 25 06:48:33 fdisk Feb 25 06:49:11 start 1 end +25000M Feb 25 06:49:13 you did what to the existing partitions? Feb 25 06:50:22 deleted them and made new partitions,than reflashed.my opt partition uses 6 Gb,it is ok,but can't get larger fat space than 7Gb Feb 25 06:50:58 also will use my ext card for swap Feb 25 06:52:03 i'm a power user,very strange,the fat partition is type win95 fat Lba Feb 25 06:53:53 why my fat is only 7Gb? Feb 25 06:56:43 is anyone able to help me?it is very strange,looks like my fat size is limited,but don't know why Feb 25 06:57:53 fat16 or fat32? Feb 25 06:58:47 hmm... i dared not to muck with that despite it would be nice to have more space in /opt... but there's plenty as it is... Feb 25 06:59:16 fat32 Feb 25 06:59:39 i know,that fat16 size is limited Feb 25 07:04:36 what are you using for logical sector size? Feb 25 07:05:43 don't know Feb 25 07:08:20 try mkfs'ing with -S 32768 or -s 128 and see if that works... it's just a guess though, I thought mkfs should figure this out on its own Feb 25 07:11:18 i use failsafe,there is only fdisk Feb 25 07:38:07 DocScrutinizer05: you suck at predictions, dude Feb 25 07:38:17 they're supposed to be a lot vaguer than that Feb 25 07:38:44 Testing Feb 25 07:39:03 Wooooa m on irc Feb 25 07:39:09 wooooooooah Feb 25 07:39:11 welcome Feb 25 07:40:07 Yeah Feb 25 07:40:09 DocScrutinizer05: "and the end will come for the lost winds when the demon from the east will reveal himself" Feb 25 07:40:16 or something Feb 25 07:41:29 problemsolved Feb 25 07:41:39 thank you Feb 25 07:41:47 cheers Feb 25 07:44:54 Chkng Feb 25 07:45:14 checking what? Feb 25 08:46:09 can you guys link me to how to on soldering microusb unit on n900 lots of pics may be even a video Feb 25 08:46:22 ~usbfix Feb 25 08:46:23 [usbfix] http://talk.maemo.org/showthread.php?t=75920 - and **NEVER** use epoxy (unless you want to seal your device for underwater) Feb 25 09:09:16 [00:23] or a guy sounding like him but with a nick of "Piotr" Feb 25 09:09:25 so much about "moral value" :/ Feb 25 09:10:15 If I were you, with your current "position" and "respect", I would be little more carefull with acussing people of using "virtual" identifies to write bullshit Feb 25 09:10:58 especially, that, no matter of your personal antagonism, you must admit, that I was never afraid of writing unpopular things, no matter of ban or whatever harassment risk Feb 25 09:11:26 so please, don't offend me, by implying cowardise of hiding behind some idiot's nick. Thanks in advance. Feb 25 09:15:50 anyone using bme replacement and kp52, experience strange hardware wakeups in powertop? Feb 25 09:16:22 define strange Feb 25 09:16:24 hardware things that should sleep wake up for some time (5-8%) Feb 25 09:16:32 lemme dig up pastebin Feb 25 09:16:44 brb, need to shut down connections for a while Feb 25 09:18:07 C4 | 92.6% | 3090.1ms | Feb 25 09:24:33 hi, long time since I asked this here: any news? Feb 25 09:24:38 any new devices? Feb 25 09:25:15 second question: I got a lenovo laptop, I want to use a gps program on the laptop Feb 25 09:25:37 which option makes more sense: use the N9 gps and internet connection Feb 25 09:25:55 or buy a cellular modem +gps at 200$ ? Feb 25 09:26:13 I mean most of you must tried this - right? Feb 25 09:27:42 here is powertop output Feb 25 09:27:46 http://sebsauvage.net/paste/?38da9ceccfdd5407#oVyh3iVpkQkuALzAtCuEkGFPIVFjrok0ItFy9UwTOyM= Feb 25 09:27:53 ^ kerio Feb 25 09:28:13 well, what's wrong? Feb 25 09:28:16 looks good Feb 25 09:28:35 nope, it's with everything disabled Feb 25 09:28:48 before, ic2_omap wasn't generating more than ~250 wakeups Feb 25 09:28:51 now it's twice as much Feb 25 09:29:04 also, C1 C2 and C3 was quiet Feb 25 09:29:10 only C4 and very little C0 Feb 25 09:29:20 \_o_/ Feb 25 09:29:30 only C4 and very little C0 Feb 25 09:29:33 which troubles me most Feb 25 09:29:36 well, hald-addon-bme is listed, but with only one activity Feb 25 09:29:55 core, neon and mpu are awake for 14% of time Feb 25 09:30:21 yeah, that's kinda bad Feb 25 09:30:22 per for 5% Feb 25 09:30:35 I'm 100% sure before kp52 it was all OK Feb 25 09:30:41 may be nit related, though Feb 25 09:30:47 nit? Feb 25 09:31:27 i, too, have 7.3% of C3 Feb 25 09:31:40 and 10% of core/neon/mpu Feb 25 10:09:17 when I try to install columbus I get: Feb 25 10:09:45 that's not a lot Feb 25 10:09:57 faild to fetch http://repository.maemo.org/extras... Feb 25 10:10:24 is repository.maemo.org up? Feb 25 10:12:46 Failed to fetch http://repository.maemo.org/extras-devel/pool/fremantle/free/libr/librxsupport/librxsupport_0.2.0-1_armel.deb Hash Sum mismatch Feb 25 10:12:53 Failed to fetch http://repository.maemo.org/extras-devel/pool/fremantle/free/libc/libcolumbus/libcolumbus_0.9.3-1_armel.deb Hash Sum mismatch Feb 25 10:13:17 why are the Hashs bad? Feb 25 10:13:59 repo problem Feb 25 10:14:05 use extras-devel.merlin1991.at Feb 25 10:14:30 so I need to edit the sources file Feb 25 10:14:34 ok, lets see Feb 25 10:16:13 deb http://extras-devel.merlin1991.at/extras-devel fremantle free non-free Feb 25 10:16:14 ? Feb 25 10:16:54 ok, it updates well Feb 25 10:17:18 W: GPG error: https://downloads.maemo.nokia.com ./ Release: The following signatures were invalid: KEYEXPIRED 1349249546 KEYEXPIRED 1349249546 KEYEXPIRED 1349249546 Feb 25 10:17:23 ignore that Feb 25 10:17:51 do you use the n9? Feb 25 10:18:20 did anyone managed to add google calender sync to the n900? Feb 25 10:18:27 calendar Feb 25 10:18:42 no, still hash mismatch Feb 25 10:18:49 min Feb 25 10:19:02 why is it using http://repository.maemo.org/extras-testing/pool/fremantle/free/libr/librxsupport/librxsupport_0.1.0-4_armel.deb Feb 25 10:19:44 cityLights, do dpkg -i /var/cache/apt/archives/partial/*.FAILED Feb 25 10:20:08 (check if the path is right) Feb 25 10:20:17 I just typed it out of my head Feb 25 10:20:29 teotwaki: does it sync periodical to google calender? Feb 25 10:20:48 Pali: 0xFFFF seems to flash even with -s Feb 25 10:20:49 W: Failed to fetch http://extras-devel.merlin1991.at/extras-devel/dists/fremantle/non-free/binary-armel/Packages 404 Not Found [IP: 188.40.39.11 80] Feb 25 10:20:52 it's "calendar" Feb 25 10:20:55 not "calender" Feb 25 10:21:42 qwazix: rm /var/cache/apt/archives/partial/* Feb 25 10:22:16 ? Feb 25 10:22:18 ccxCZ, I will look at it Feb 25 10:22:47 cityLights, the hash sum failed is due to repo index file error, the packages are actually fine Feb 25 10:23:12 so apt downloads them in /var/cache/apt/archives/partial with a suffix .FAILED Feb 25 10:23:25 I fixed the source line and I am waiting for the update now Feb 25 10:23:31 you can install them from there with dpkg Feb 25 10:23:44 right Feb 25 10:23:55 the source line is only for extras-devel, there are packages with sum mismatch on testing too Feb 25 10:24:28 so if you still get the error message the solution I just outlined will do Feb 25 10:24:37 right Feb 25 10:25:39 now its ok Feb 25 10:25:44 cool Feb 25 10:26:02 Pali: http://bpaste.net/show/79595/ Feb 25 10:26:05 I aim to use the n900 gps with the laptop Feb 25 10:26:14 I am looking to do the same with the n9 Feb 25 10:26:42 ccxCZ, do you have last version from git? Feb 25 10:26:49 cityLights: your url is wrong Feb 25 10:27:01 there's no such thing as http://extras-devel.merlin1991.at/extras-devel/ Feb 25 10:27:01 git version from before the weekend Feb 25 10:27:06 ok installed fine now Feb 25 10:27:35 kerio: right, I saw and fixed it , and now its ok Feb 25 10:27:44 ccxCz, do you have also full backtrace? (gdb command bt full)? Feb 25 10:29:05 Pali: here you go http://bpaste.net/show/79596/ Feb 25 10:29:29 thanks, I will look at it Feb 25 10:32:54 hrm, now it bootloops and bootmenu is gone. luckily rescueos still works Feb 25 10:34:07 ~rescueos Feb 25 10:34:07 extra, extra, read all about it, rescue-os is http://206.253.166.96/N900/rescueOS/ Feb 25 10:34:48 good, according to the docs rescueos does charge the battery too Feb 25 10:36:33 ccxCZ, now I quickly fixed that NULL pointer and simulate param Feb 25 10:36:48 patches are in git Feb 25 10:39:46 updated, compiled, segv :P Feb 25 10:41:04 http://bpaste.net/show/79599/ Feb 25 10:43:28 -O0 version, remade clean http://bpaste.net/show/79602/ Feb 25 10:44:05 ach :-( glibc detected: double free or corruption this will be harder Feb 25 10:44:42 ccxCZ, gdb is not usefull for memory corruptions... do you have valgrind? Feb 25 10:45:11 yup, but no real experience with it Feb 25 10:45:12 Hi guyz Feb 25 10:45:22 hello Feb 25 10:45:30 now when simulate flag working, I can try it too Feb 25 10:46:12 Finally irc Feb 25 10:46:19 Yippe Feb 25 10:48:13 0 Feb 25 10:53:21 kerio, it doesn't matter if it's a lot or not, there definitelly wasn't such things before Feb 25 10:53:32 ping Pali Feb 25 10:54:14 Pali, could you look at this: Feb 25 10:54:15 http://sebsauvage.net/paste/?38da9ceccfdd5407#oVyh3iVpkQkuALzAtCuEkGFPIVFjrok0ItFy9UwTOyM= Feb 25 10:54:58 since installing kp52 and bme replacement bits, there is strange hardware activity, that I'm 100% sure wasn't the case before Feb 25 10:55:24 100% idle device, settled down, without any connections Feb 25 10:55:53 core, neon, and mpu are active for 16% of times, when it should be asleep Feb 25 10:56:00 16% of time* Feb 25 10:56:36 CPU is 9% on C3, when it should sleep at C4 almost completely Feb 25 10:57:02 also, 2.7% C2, which is (summed up with C3) strange Feb 25 10:57:22 I suspect those core, neon and mpu are the cause Feb 25 10:57:36 it eats more than 10% of standby time Feb 25 10:57:43 i.e. idle power usage got higher Feb 25 10:58:22 freemangordon, ^, you may be interested in it too (no idea if using thumb can be related? But, before kp52, no such things, even in thumb...)) Feb 25 10:58:50 if I can debug it some more, just tell me what to do. Kerio said he also got similar wakeups, he is on kp52 and bme replacement too Feb 25 10:59:04 emove bme repl Feb 25 10:59:09 remove* Feb 25 10:59:33 could do, but bme replacement bits are listed on powertop, with very OK'ish number of wakeups Feb 25 10:59:56 so I would rather blame kp52, if anything Feb 25 10:59:57 bat applet Feb 25 11:00:07 tried without it already Feb 25 11:00:10 same things Feb 25 11:00:15 :-/ Feb 25 11:00:32 stop processes one after another Feb 25 11:00:40 it may be unrelated to KP at all, it's just that due to repos being kinda down, I havent changed anything else Feb 25 11:00:49 prepare for unadvertised reboots though Feb 25 11:01:18 DocScrutinizer05, ok, but which processes? it's hardware wakeup, not software. Or do you mean *all* processes? Feb 25 11:01:36 unload bq* kmods Feb 25 11:01:56 uh, right, could think about that one myself :P Feb 25 11:02:32 will report result at evening... BTW, it would be nice if other people with kp52 and/or bme replacement could test powertop too Feb 25 11:02:40 appreciated Feb 25 11:03:15 device must be settled down (no recent activity, more than 30 minutes since turning on), with gsm and wifi and whatsnot disabled for some minutes already Feb 25 11:03:40 and locked screen? Feb 25 11:03:43 (reminder for others, that may be interested) Feb 25 11:03:44 sure Feb 25 11:04:14 "powertop" from extras-devel mirror, then powertop from terminal, and quickly lock screen and lay device flat and idle Feb 25 11:04:22 I support this investigation Feb 25 11:04:24 after 11 sec, powertop will start sampling for 30 sec Feb 25 11:04:28 * Estel_ nods Feb 25 11:04:55 it may be totally unrelated, but anyway, maybe someone will get idea what causes this, then. I'm quite allergic to hardware wakeups :P Feb 25 11:05:16 same here ;-) Feb 25 11:05:58 Estel_: have you checked IRQs? Feb 25 11:06:20 erm, checked what? :P Feb 25 11:06:54 ~irq Feb 25 11:06:55 hmm... irq is (Interrupt request) This is a gateway for devices to request time from the processor to do their jobs. For instance, every time you hit a key on your keyboard an interrupt is generated on the keyboard IRQ. This is mainly only a concern for PC users. Feb 25 11:07:28 cat /proc/interrupts Feb 25 11:07:49 see who's firing Feb 25 11:08:15 oook, but should I test it, somehow, when device is idle? Feb 25 11:08:20 * DocScrutinizer05 bets on i2c Feb 25 11:08:21 or just normally? Feb 25 11:08:56 cat /proc/interrupts; sleep 60; cat /proc/interrupts Feb 25 11:09:00 compare Feb 25 11:10:00 ok Feb 25 11:10:10 here is result of "just cat": Feb 25 11:10:12 http://sebsauvage.net/paste/?4db15921728c418c#2nUcVu+OhWOKEmxcabxeshmrwHg2FE3eJiSK1GM4S/o= Feb 25 11:10:55 will do sleep 11; cat /proc/interrupts; sleep 60; cat/proc/interrupts, and lock screen after firing, just like with powertop Feb 25 11:11:12 but that need to wait 'till evening, can't drop connections now Feb 25 11:11:26 273: 6189 GPIO kb_lock Feb 25 11:11:34 12:11 46 Tage 11:58 an, 0 Benutzer, Durchschnittslast: 0,03, 0,02, 0,00 Feb 25 11:11:39 ;-) Feb 25 11:12:23 sorry, yopu lost me on this one? Feb 25 11:12:33 s/yopu/you/ Feb 25 11:12:33 Estel_ meant: sorry, you lost me on this one? Feb 25 11:12:36 sorry Feb 25 11:13:14 my device has an uptime of ~46 days and I had 6189 kb-lock slider actions during that time Feb 25 11:13:52 sounds insane Feb 25 11:14:31 ah :) Feb 25 11:14:46 even if each open and close is an IRQ, this would still be ~60 per day Feb 25 11:14:54 sound OK Feb 25 11:15:02 I think I'm sliding it more :P Feb 25 11:15:16 this is tk-lock switch I think Feb 25 11:15:27 well, might be Feb 25 11:15:54 231: 174 GPIO slide Feb 25 11:16:58 which tells you my device is closed now and seen 87 open/close cycles during last 46 days Feb 25 11:17:43 oh, as for slider on the right side, I use it even more Feb 25 11:17:52 I would bet on ~200 per day Feb 25 11:17:58 strangely enough, it's not even worn out Feb 25 11:18:23 I think there is difference between revisions @ plastic quality of some part - for sure keyboard sux on 2204 Feb 25 11:18:44 may be, that people with worn-out sliders got them worse too Feb 25 11:19:11 quite possible, or they have abrassive puches Feb 25 11:19:19 my 1st keyboard, which is used *very* much, is still matt and got 0 worn out key labels or not even slightest sratched thing Feb 25 11:19:36 OTOH, many others are worn out and "glossy" after few months of usage Feb 25 11:19:36 abrasive pouches Feb 25 11:19:48 * Estel_ nods Feb 25 11:20:12 same here Feb 25 11:20:22 keymat is worse than battery Feb 25 11:20:33 new one every year, at least Feb 25 11:23:10 anyway, for you only difference over a idle time in proc/interrupts is relevant Feb 25 11:24:46 those (I suspect): Feb 25 11:24:49 56: 270217988 INTC i2c_omap Feb 25 11:24:50 57: 29466229 INTC i2c_omap Feb 25 11:24:52 61: 601899 INTC i2c_omap Feb 25 11:25:33 bq* kmods would have impact on iirc 57 Feb 25 11:32:33 [root@blade-a ~]# xm list Feb 25 11:32:35 Name ID Mem VCPUs State Time(s) Feb 25 11:32:36 Domain-0 0 1024 12 r----- 23302.5 Feb 25 11:32:38 dns 12 1024 1 -b---- 4.4 Feb 25 11:32:39 firewall-a 1 512 1 -b---- 579.0 Feb 25 11:32:41 monitor 13 2048 2 -b---- 11.3 Feb 25 11:32:42 [root@blade-b ~]# xm list Feb 25 11:32:44 Name ID Mem VCPUs State Time(s) Feb 25 11:32:45 Domain-0 0 1024 12 r----- 15318.9 Feb 25 11:32:47 db 20 8192 2 -b---- 6243.3 Feb 25 11:32:48 firewall-b 3 512 1 -b---- 312.9 Feb 25 11:32:50 garage 27 8192 2 r----- 569.7 Feb 25 11:32:51 wiki 16 1024 1 -b---- 307.7 Feb 25 11:32:53 www 22 2048 2 -b---- 294.4 Feb 25 11:33:27 neat Feb 25 11:39:20 DocScrutinizer05: isn't blade-a a bit under-used? Feb 25 11:39:43 kerio: this will change Feb 25 11:40:48 http://wiki.maemo.org/Migrating_to_Community-driven_Infrastructure#Plan_for_migration_.2F_Timeline Feb 25 11:42:30 for now everything is underused Feb 25 12:12:56 my N900 is shipping - damn you for making me want it for years! Feb 25 12:13:24 yay Feb 25 12:14:30 you people keep making all this cool stuff for it that my N800 didn't get Feb 25 12:15:17 plus the fact the old N800 is starting to get unreliable: random reboots, a tad crashy, not a lot of RAM (although a class 10 SD card and a helping of swap on it helps) Feb 25 12:15:51 :-) Feb 25 12:19:10 no sim card is going in the new thing though Feb 25 12:19:49 celluar service is a bloody scam, it's way too expensive! I'll stick with my home fibre and my city's (and most restraunts) free wifi Feb 25 12:35:12 if I wanted to reformat rootfs and optfs by hand, are there any special flags to mkfs I should use? Feb 25 12:42:51 you probably don't want to reformat rootfs Feb 25 12:42:58 because nobody understands how to format ubifs Feb 25 12:43:07 optfs... meh, whatever you want, really Feb 25 12:43:27 it's going to be mounted as ext3 with the options listed in /etc/fstab Feb 25 12:44:18 ic, thanks Feb 25 12:44:46 how do the rootfs backup in bootmenu work then? plain dd? Feb 25 12:44:51 does* Feb 25 12:44:59 no Feb 25 12:45:07 never ever ever ever use dd on a mtd Feb 25 12:45:20 backupmenu mounts the filesystem and makes a tarball Feb 25 12:45:30 on restore, it deletes every file on the filesystem and then unpacks the tarball Feb 25 12:45:45 ok Feb 25 12:46:41 optfs is recreated Feb 25 12:46:48 but that's just a mkfs.ext3 by default, anyway Feb 25 12:49:29 ok, so I can use rescueos to just do that by hand Feb 25 12:50:22 * ccxCZ thinks rsync --delete and squashfs would be better for rootfs Feb 25 12:50:35 ccxCZ: in which way? Feb 25 12:51:25 to restore the backups, since I don't have backupmenu anymore Feb 25 12:51:46 rsync --delete only overwrites when there are changes Feb 25 12:52:05 and you'd get cheap lzo compression and random access on the backups Feb 25 12:52:54 hm Feb 25 12:52:59 well i don't see why that wouldn't work Feb 25 13:09:17 kerio, I know how to format ubifs Feb 25 13:09:36 you need to use ubiformat Feb 25 13:09:59 oh really :P Feb 25 13:10:10 ubiformat /dev/mtd5 -s 512 -O 512 -f image.ubi Feb 25 13:10:46 but ubifs must be umounted and ubi deattached from mtd device Feb 25 13:10:59 *de*attached? really? Feb 25 13:11:13 yes, because you are flashing ubi image to mtd Feb 25 13:11:18 oic Feb 25 13:11:21 not ubifs image! Feb 25 13:11:47 our rootfs image in FIASCO is ubi image which contains one volume for rootfs Feb 25 13:12:10 Pali: what about creating a brand new ubifs Feb 25 13:12:21 use mkfs.ubifs Feb 25 13:12:28 i assume it's ubimkvol first though Feb 25 13:12:42 this will create new ubifs image on HDD Feb 25 13:12:45 in file Feb 25 13:13:18 and then use ubinize which will create new ubi image from that ubifs file Feb 25 13:13:37 and then you can flash that new ubi image to n900 Feb 25 13:14:28 ubiformat doing all what is needed to flashing new ubi images (preserve erase counters, ...) Feb 25 13:15:09 kerio, but I did not tried it on real n900, only on desktop with nandsim Feb 25 13:15:14 and it worked Feb 25 13:20:22 ubuntu on n900? Feb 25 13:20:24 heh Feb 25 13:21:06 so is maemo.org running on someones eu isdn line now? ;) Feb 25 13:23:41 Macer: not yet, we have a 9k6 modem for that. Feb 25 13:24:10 heh Feb 25 13:24:28 i'm sure DocScrutinizer05 has it hosted somewhere Feb 25 13:24:39 what? Feb 25 13:24:45 maemo.org Feb 25 13:24:52 o.O Feb 25 13:25:03 ie: hosted on some vhost server Feb 25 13:25:10 Macer: warfare is the sysadmin for *all* maemo infra Feb 25 13:25:24 oh Feb 25 13:25:28 and I'm the manager Feb 25 13:25:55 and right now, the two servers from nokia are in free CoLo at IPHH Feb 25 13:26:09 they're from nokia? Feb 25 13:26:20 Macer: if you wanna know details, check http://wiki.maemo.org/Migrating_to_Community-driven_Infrastructure Feb 25 13:26:22 i thought nokia dropped the servers and only redirected their dns Feb 25 13:26:44 the opensource fad isn't what it used to be :) Feb 25 13:26:45 they gifted us two quite powerful servers Feb 25 13:26:59 one Feb 25 13:27:02 ouldn't authenticate against garage. (DB problem) Feb 25 13:27:10 with 4 blades with 2 cpu sockets each Feb 25 13:27:15 Couldn't authenticate against garage. (DB problem) Feb 25 13:27:22 * Macer looks around Feb 25 13:27:31 Macer: oops Feb 25 13:27:46 which service? Feb 25 13:28:03 http://wiki.maemo.org/Migrating_to_Community-driven_Infrastructure Feb 25 13:28:09 o.O Feb 25 13:28:11 wut Feb 25 13:28:26 if i was the cio... you'd be fired as manager lol Feb 25 13:28:28 FUUUUUCK Feb 25 13:28:39 don't make business cards just yet :) Feb 25 13:28:46 shutup Feb 25 13:29:02 bitch at support@nemein.com Feb 25 13:29:09 hehe Feb 25 13:30:09 warfare: please look into it Feb 25 13:30:17 ^^^^ Feb 25 13:30:38 seems aslan has fekkered db when he tried those snapshots Feb 25 13:31:17 [2013-02-25 13:07:45] I wonder if we can get those snapshots out of the grid somehow Feb 25 13:31:19 [2013-02-25 13:07:54] then again, they are huuuge Feb 25 13:31:20 [2013-02-25 13:13:28] and that actually causes interesting problem now Feb 25 13:31:30 "interesting" Feb 25 13:31:30 interesting indeed lol Feb 25 13:31:49 MEH! Feb 25 13:31:53 *sigh* Feb 25 13:35:54 fixed Feb 25 13:36:20 Macer: ^^^ Feb 25 13:40:29 you don't see the irony of that page being down? :) Feb 25 13:41:24 wow.. why is nemein being so kind? Feb 25 13:42:10 HiFo? Feb 25 13:42:11 kerio, look: http://talk.maemo.org/showpost.php?p=1325043&postcount=12 and http://talk.maemo.org/showpost.php?p=1325044&postcount=13 Feb 25 13:42:39 kerio, I wrote what I know about ubi, ubifs, mounting and flashing Feb 25 13:42:56 i see Feb 25 13:43:15 do you know if there's any "tuneubifs" utility Feb 25 13:43:33 and what should that utility do? Feb 25 13:43:46 there are more ubi* utilities Feb 25 13:44:13 Pali: for instance, change the default compression method Feb 25 13:44:38 there are: ubiattach, ubicrc32, ubidetach, ubiformat, ubimkvol, ubinfo, ubinize, ubirename, ubirmvol, ubirsvol, ubiupdatevol Feb 25 13:45:11 kerio, I do not know if there is support in kernel for other compression method for ubifs Feb 25 13:45:43 compression in ubi most probably optimized for NAND Feb 25 13:45:55 DocScrutinizer05: meh, it's LZO Feb 25 13:45:59 you can also use zlib Feb 25 13:46:02 and no compression Feb 25 13:46:07 it's not like 2hey, let's go for ueber-leete new xyz compression" Feb 25 13:46:12 i kinda wanted to test the latter Feb 25 13:46:16 kerio, you can create new ubifs image with mkfs.ubifs Feb 25 13:46:24 and there is option -x Feb 25 13:46:34 compression type - "lzo", "favor_lzo", "zlib" or "none" (default: "lzo") Feb 25 13:46:57 but there is no utility for modifing ubifs filesystem Feb 25 13:47:35 kerio, so you can unpack rootfs, then create new ubifs file, pack it into ubi image and flash it Feb 25 13:47:42 Macer: only thing i see is a very momentary servie glitch that got fixed by maemo tech staff during <5min after report Feb 25 13:47:57 Pali: ooh, you can also mkfs.ubifs directly on a /dev/ubix_y Feb 25 13:48:06 mkfs.ubifs can create new filesystem in regular file or can call it on /dev/ubi Feb 25 13:48:09 yes Feb 25 13:48:24 but you need to umount ubi dev first Feb 25 13:48:34 so you cannot do it from maemo Feb 25 13:48:44 surely you must have the ubi device attached, to make the ubifs on it Feb 25 13:48:47 Macer: which, given the fact that we're just migratinng the whole shite again, is really really not a tragedy Feb 25 13:49:15 anyway, where would i get the correct parameters for it? Feb 25 13:49:32 other maintainers would simply announce service downtime of two days Feb 25 13:49:34 like the leb size Feb 25 13:50:04 and the minimum I/O unit size Feb 25 13:50:43 ubinfo -a Feb 25 13:50:53 ty Feb 25 13:50:53 and from dmesg Feb 25 13:51:10 Count of bad physical eraseblocks: 1 Feb 25 13:51:12 oh noes D: Feb 25 13:51:28 ubi and ubifs modules in debug mode printinf more in dmesg Feb 25 13:51:35 now going offline, bye Feb 25 13:51:50 Macer: so if I were the IT manager, you would earn a place on my ignore list :-) Feb 25 13:52:44 ticket response time of <300sec, a dream Feb 25 13:53:02 what the hell, why is the logical eraseblock size 126.0KiB Feb 25 13:53:06