**** BEGIN LOGGING AT Wed Jun 25 02:59:58 2014 Jun 25 10:38:02 hm what happened to "Power Search" on maemo.org? Jun 25 10:57:04 cssu is looking for people again: http://talk.maemo.org/showthread.php?p=1430952 Jun 25 10:57:54 How much do you pay? :P Jun 25 11:20:47 Power Search has been broken for a while. It's generally to do with what you are searching for I believe. Jun 25 11:27:43 ~sb Jun 25 11:27:44 hmm... scratchbox is a cross-compiling system that uses binfmt_misc, rpc calls, and an nfs mount to make a cross-build appear to be 100% native, and is found at http://www.scratchbox.org/, hosted by maemo now. Also at http://maemo.merlin1991.at/files/SB Jun 25 11:28:39 sixwheeledbeast: well for me it's just "This page will allow you to search the whole of maemo.org and will allow you to filter the search results according to the different sections. Loading" Jun 25 11:28:46 the search doesn't even show up :D Jun 25 11:30:40 ah yes, that search. That has been broken the same amount of time as the other search bar that doesn't like searching for "n900" Jun 25 14:12:57 sixwheeledbeast: http://originalspareparts.com.ua/phones/nokia/n900#c "screw M1.4 x 2.8" Jun 25 14:14:00 or M1.4x5.0 or RF1.4x4 Jun 25 14:34:33 DocScrutinizer05: thanks Jun 25 14:53:48 sixwheeledbeast: don't fall for http://www.gerry-mobilfunk.de/cgi-bin/shop.pl?seite=1&do=artikel&artikel=9991964&status=00029&suchstring=&bestellung=&hersteller=Nokia&hersteller1=N900&ru=Handy&sk=GM&sa= Jun 25 14:54:00 it's genuine but a rip-off basically Jun 25 14:54:27 you can get 2k of them for the same price Jun 25 14:55:00 DocScrutinizer05: yes I would look at cpc.farnell or rs etc first. Jun 25 14:55:14 check ebay, honestly Jun 25 14:55:34 M1.4x5 is a good and precise search string and product spec Jun 25 14:56:03 or whatever the particular screws you need Jun 25 14:56:21 FR1.4x4.0 Jun 25 14:56:25 *shrug* Jun 25 15:01:57 DocScrutinizer05: Also while your about the friend I need the screws for has had an issue with his screen protector. it has "welded" it's self to the screen (never heard this before). Any tips on removal without damaging anything? Jun 25 15:11:39 slooooowly, no force, slightly increased temperature Jun 25 15:12:07 I'm afraid he tried to remove the better half of digitizer ;-) Jun 25 15:12:55 often a patch of scotchtape in one corner can help lift the protector Jun 25 15:14:25 (screws) http://originalspareparts.com.ua/parts/6300174/screw_m14x50_black_camera-cross/ Jun 25 15:14:58 http://originalspareparts.com.ua/parts/6380558/screw_rf14x4_torx_plus_6ip_black/ Jun 25 15:15:33 sounds more reasonable than 8EUR Jun 25 15:25:04 TMO does not resolve. Jun 25 15:25:04 maemo.org does not resolve Jun 25 15:25:04 oh it was my DNS server :P Jun 25 15:25:05 looking good now! Jun 25 15:27:31 DocScrutinizer05: I have not seen it yet. But I too think it's the digitizer Jun 25 15:30:12 I was thinking if it is the protector is stuck then slowly with maybe a thin guitar pick. Didn't think of temperature however having a warm spell here so maybe leave in the sun for a bit. Jun 25 15:31:04 Are there any good quality digitizer replacement links? Jun 25 15:31:10 use scotchtape! Jun 25 15:31:37 quality of replacements up to particular evaluation Jun 25 15:31:59 most claim "ORIGINAL NOKIA" Jun 25 15:32:15 some prolly are, some may be fake of inferior quality Jun 25 15:33:30 e.g when they use whatever liquid they can find,instead of the silicon oil, for separator Jun 25 15:33:59 prolly a lot of phantom shifts done on tzhose parts Jun 25 15:35:02 maybe even true knock offs / copies Jun 25 15:36:05 real fakes (non-functional) rather improbable Jun 25 15:42:55 Never had to replace one yet, however, my worry would be spending time fitting a poor quality part. Jun 25 16:06:12 quality check: connect resistance-meter to upper and lower plane, check for ope-circuit aka infinite-ohms. then make sure digitizer makes contact when you *slightly* touch it with a feather or the like Jun 25 16:07:10 there are only 4 pins on connector of the digitizer, find two of them who are usually not connected Jun 25 16:07:21 NoContact Jun 25 16:08:14 on slightest touch, the resistance has to change from infinite to some 100..1000Ohms Jun 25 16:08:30 easy Jun 25 16:09:10 more subtly quality issues not detected that way, but you will find out after installing only Jun 25 16:09:31 DocScrutinizer05: wait, the digitizer is supposed to be *that* sensitive? :o Jun 25 16:09:38 yes Jun 25 16:10:37 I can hardly wipe of dust flakes with finger, and I'm not an elephant Jun 25 16:11:15 it's almost impossible to touch a good new N900 digitizer with your finger without already making a touch event Jun 25 16:12:20 for sure a sheet of paper touching digitizer with the corner should be able to create a touch even when done right Jun 25 16:13:11 of course the separator oil might age and get more viscous over the years Jun 25 16:14:31 actually meanwhile my daily N900 needs some 20g orso on a pinpoint touch Jun 25 16:15:24 also depends a bit on whether you touch at edge or on center of digitizer Jun 25 16:15:31 edge is harder Jun 25 16:17:00 still woking with a standard sheet of paper corner Jun 25 16:17:15 though it's a tad hard and needs several tries now Jun 25 16:18:19 fold the paper once and for sure you get any touch you want, easily Jun 25 16:19:28 a standard BT USB dongle is a itsy bit too few weight to make it trigger Jun 25 16:20:43 * DocScrutinizer05 will test witl coins of well known currency later on Jun 25 18:22:23 So......... Jun 25 18:46:11 so? Jun 25 18:46:44 (coins) 2EUR is close to the edge, damn close but not entitely sufficinet Jun 25 18:54:50 probably the 2€ coin is not "pinpointed" enough even on the edge, thus it takes a tad more this way Jun 25 19:35:41 Well, I thought about fuckin' some chickens before, eh? If you want to have a good time and you need some pussy, you can (motions with his hands) cut that chicken's head off, stick yo' dick in the ass of that chicken, and that damn chicken'll go crazy on your ass and go "Caaaaah". Jun 25 19:36:53 yeah ..... right. Jun 25 19:37:20 /clear /clear /clear Jun 25 19:37:52 /clear /opaque /clear /fill Jun 25 19:37:56 =) Jun 25 19:47:15 Snafu777 ._. Jun 25 19:48:32 DocScrutinizer05: so you managed to get xchat behaving properly on switching 3g/wifi connections? Jun 25 19:49:25 RiD: =) Jun 25 19:49:43 hey, any easy way (without reflashing) to check if my FM transmitter is hardware-broken? Jun 25 19:50:10 it doesn't seem to work. Jun 25 19:51:31 a radio Jun 25 19:51:51 :] Jun 25 19:52:59 yeah, that's how i know it doesn't work Jun 25 19:53:07 what i want to be sure of, if its a software mess up or hardware Jun 25 19:53:16 but at the same time, i'm too lazy to reflash Jun 25 19:53:29 (i have multiboot set up and all that :\ - yeah kill me for not using uboot too!) Jun 25 19:54:40 multiboot is actively harmful Jun 25 19:55:21 I know. Uboot has issues with what I want to do, though. Tried it a couple times, never got it to work 100% Jun 25 19:55:35 what do you actually want to do ? Jun 25 19:55:53 android + maemo Jun 25 19:55:55 (that wont work with uboot ?) Jun 25 19:56:06 you really use android on n900 ? Jun 25 19:56:11 never got android hardware accel to work with uboot, or charging Jun 25 19:56:15 for an application or two, yes Jun 25 19:56:35 oh, so people actually use it Jun 25 19:56:38 woah Jun 25 19:56:42 very few I guess :| Jun 25 19:56:55 I still prefer maemo, but really not much choice eh Jun 25 19:57:22 and i don't really want to jump into a more powerful phone, while at the same time more garbage Jun 25 19:57:51 I'd rather handle the n900 "slowness" than jump into android right now. Not many alternatives to android either Jun 25 19:58:20 5" display, open keyboard, bam, useless screen Jun 25 19:58:31 annoys the hell out of me Jun 25 19:58:31 * Snafu777 uses his n900 as a computer in his pocket Jun 25 19:58:45 Hey DocScrutinizer05, maybe you can answer this or anyone else for that matter Jun 25 19:58:52 Snafu777: who doesn't Jun 25 19:59:16 I've got easy debian chroot installed Jun 25 19:59:20 and I'm using debian-m5-v3e.img.ext2 Jun 25 19:59:24 dos1: (xchat) sure Jun 25 19:59:31 I've been apt-get purging crap on it to save space Jun 25 19:59:37 but the darn filesize doesnt ever shrink Jun 25 19:59:39 why? Jun 25 19:59:47 I've purged over 700 mb from the chroot Jun 25 19:59:52 DocScrutinizer05: was that lagometer option enough or you found something else? Jun 25 19:59:54 yet the debian-m5-v3e.img.ext2 filesize is the exact same Jun 25 20:00:20 Snafu777: is that an image? Jun 25 20:00:23 it wont shrink Jun 25 20:00:29 Will it expand then? Jun 25 20:00:35 Because that's why i wanted to remove stuff Jun 25 20:00:40 Snafu777: moderator warning, watch speech Jun 25 20:00:41 so that i could add stuff without expanding too much Jun 25 20:00:45 K DocScrutinizer05 Jun 25 20:00:52 Sorry =/ Twas a joke =) Jun 25 20:01:00 ? Jun 25 20:01:03 And yer referring to my chicken thing right? Jun 25 20:01:07 dos1: lagometer did it Jun 25 20:01:51 bencoh: Why will it not shrink? And can it expand if i add to it? Jun 25 20:02:03 And if not, then where does the stuff I add to it go?? Jun 25 20:02:06 It's mind boggling Jun 25 20:02:19 Snafu777: it is a filesystem image Jun 25 20:02:43 Snafu777: it's like a disk partition - it doesn't shrink when you delete files on its filesystem Jun 25 20:03:01 Does it expand if u add? Jun 25 20:03:05 without tias involved Jun 25 20:03:13 does disk partition do? Jun 25 20:05:45 when it's an image file, your only chance is to create a second smaller img file and mount it like first one and copy everything from first to new smaller img Jun 25 20:06:13 Ok, but can it grow in size? Jun 25 20:06:39 That's my concern Jun 25 20:06:50 Because that was my point for removing stuff i didn't use Jun 25 20:06:58 And, with that thought in mind Jun 25 20:07:10 If i purge 500mb, and then I add back in 400mb worth of stuff, it shouldnt expand, correct? Jun 25 20:07:27 only grows when you tell it to Jun 25 20:07:32 usually Jun 25 20:07:46 Ok...., So it can expand, but not shrink Jun 25 20:07:48 depends on the particular solution you use. Some auto-grow Jun 25 20:08:01 well, you can shrink img, but you have to resize it manually Jun 25 20:08:01 Well, I'm using the stock apt-get easydebianchroot thing Jun 25 20:08:09 some can grow when you use a tool to extend them Jun 25 20:08:36 some cannot change size at all after creation Jun 25 20:08:43 _none_ shrinks Jun 25 20:08:47 I guess what I'm getting at is this: I originally added stuff to it, but from ls -l the size did not change on the image, even after I unmounted it Jun 25 20:08:54 So where did the added stuff go? Jun 25 20:08:59 it's ext3 or 4 afaik Jun 25 20:09:04 I'll just add that if you use thin provisioning it can indeed shrink :) Jun 25 20:09:16 wow Jun 25 20:09:25 ok, you da master :-) Jun 25 20:09:34 or to be more exact the thin provisioned LV returns the blocks back to the pool VG Jun 25 20:09:44 when you delete stuff Jun 25 20:10:10 never recently looked into that stuff, so nevermind what i toldya Jun 25 20:10:17 Hmm Jun 25 20:10:23 So does anyone here use the easy debian chroot? Jun 25 20:10:27 but yeah, not really applicable if you don't use LVM :) Jun 25 20:10:30 DocScrutinizer05: resize2fs should shrink fs image file just fine Jun 25 20:10:32 M4rtinK: that's LVM? Jun 25 20:10:42 Logical Volume Manager Jun 25 20:10:43 that's a ext image file mounted with loop Jun 25 20:10:49 :nod: Jun 25 20:11:09 like mount -o loop debian.ext3 /mnt/debian Jun 25 20:11:18 ext img doesn't sound like LVM Jun 25 20:11:24 sits between physical partitions and filesystem Jun 25 20:11:28 yes, it's completely unrelated to LVM Jun 25 20:12:04 or rather, loopmount doesn't sound like LVM Jun 25 20:12:14 Snafu777: you can use resize2fs on *unmounted* image to adjust its size Jun 25 20:12:54 that's the "you tell it to" variant Jun 25 20:12:58 Alright, image resizing aside, I could honestly care less about shrinking the actual filesize of the image. What I am trying to figure out is if I purge 500mb from the image, can I add 500mb without expanding the size of the image Jun 25 20:13:24 Since the repos for maemo are so crazy, I figured i'd go with debian Jun 25 20:13:26 yeah, I'm just saying that there are some scenarios (not directly related to loop mount) where you get space "back" if you delete stuff on a filesystem Jun 25 20:13:29 The filesystem is "sane" at least Jun 25 20:13:32 Snafu777: well, if you purge 500mb from ext3 partition on your PC, can you add 500mb without expanding your harddrive? :P Jun 25 20:13:33 sure, what gets deleted can get added as well Jun 25 20:13:40 which reminds me ... no one runs lvm on n900 ? Jun 25 20:13:43 Ok Jun 25 20:13:50 So then riddle me this Jun 25 20:14:00 Without purging anything Jun 25 20:14:02 I added to the image Jun 25 20:14:04 unmounted Jun 25 20:14:06 bencoh: don't know about anyone Jun 25 20:14:07 yet the size didnt grow Jun 25 20:14:16 *sigh* Jun 25 20:14:20 and NOTHING else on the entire n900 partitions changed in size Jun 25 20:14:21 I think kp has no lvm support Jun 25 20:14:26 bencoh: on the other hand the HP TouchPad uses LVM :) Jun 25 20:14:27 Snafu777: size is *constant*! Jun 25 20:14:28 So where the heck did the added files go? Jun 25 20:14:35 bencoh: as presumably other WebOS devices did Jun 25 20:14:35 M4rtinK: lvm is love :) Jun 25 20:14:38 your image is like a virtual disk Jun 25 20:14:47 Snafu777: it will not shrink, ever :) Jun 25 20:14:50 bencoh: indeed! :) Jun 25 20:14:52 Snafu777: if you remove 500mb from inside filesystem, the file with it does not shrink (unless you run resize2fs on it) Jun 25 20:14:57 So you are saying when I chroot into it, I should issue df to see how much I have on / Jun 25 20:15:01 yes Jun 25 20:15:15 So it's like a fixed little usb thumbdrive so to speak Jun 25 20:15:19 you don't even need chrooting for that, just mounting it will do Jun 25 20:15:25 True Jun 25 20:15:26 Snafu777: somehow Jun 25 20:15:33 well, kindof Jun 25 20:15:33 but the concept is to chroot so u have an OS within the n900 OS Jun 25 20:15:59 fine Jun 25 20:16:11 alright, i'm decompressing it right now to chroot into it Jun 25 20:16:23 I'll do a df and see what is available Jun 25 20:16:27 I think I understand now Jun 25 20:16:33 chroot, use df -h, see how much of your 42.911GB image space got used Jun 25 20:16:33 THANK you for the patience =) Jun 25 20:17:03 do a ls -l 42.911GB.image to notice it never chages size Jun 25 20:17:28 That's what boggled me DocScrutinizer05 =) But now after listen to ya'll it has clicked i do believe Jun 25 20:17:40 :-) Jun 25 20:17:52 chicken enlightnment Jun 25 20:18:02 lol Jun 25 20:18:10 The whole reasoning behind wanting to know is so that I can use the debian environment for things that just "dont" work in maemo Jun 25 20:18:21 Other people try to run different OS's on the n900 and whatnot Jun 25 20:18:24 screw all that Jun 25 20:18:39 easydeb FTW Jun 25 20:18:43 Yes =) Jun 25 20:18:47 beware that you will be limited by old kernel Jun 25 20:18:52 That's ok by me Jun 25 20:18:55 It's a starting pt Jun 25 20:19:03 glibc in newest debian won't work with maemo kernel Jun 25 20:19:10 Oh Jun 25 20:19:23 So my idea of doing ./configure && make to create a binary on the chroot Jun 25 20:19:25 * DocScrutinizer05 idly ponders qemu on fremantle Jun 25 20:19:26 wont port over to maemo? Jun 25 20:19:33 Because that was the plan Jun 25 20:19:34 with win7 inside ;-P Jun 25 20:19:34 so for easydebian you're limited to wheezy(?) Jun 25 20:19:46 squeeze as far as i can tell Jun 25 20:19:55 possibly Jun 25 20:19:56 DocScrutinizer05: *nope*nope* :– Jun 25 20:19:56 its got sid squeeze and lenny in the sources Jun 25 20:20:02 I killed off all non squeeze ones Jun 25 20:20:09 sid for sure won't work Jun 25 20:20:54 dos1: until we get a brand new mainline kernel with phone functions :° Jun 25 20:21:02 yup Jun 25 20:21:08 what's missing btw ? Jun 25 20:21:20 the phone functions? Jun 25 20:21:25 :-D Jun 25 20:21:44 right, but is it just about code merge or rewriting driver ? Jun 25 20:21:57 plus a number of other things, I heard cam doesn't work, audio also a construction site Jun 25 20:22:06 I thought isi driver was enough for phone functions Jun 25 20:22:15 (at least for kernel space) Jun 25 20:22:22 prolly it is Jun 25 20:22:29 So will my idea for chrooting into debian, doing whatever to create a binary, and copying the binary back to my maemo partition, it wont work? Jun 25 20:22:33 it's also about fremantle closed parts having lots of old-kernel specific cruft hardcoded Jun 25 20:22:41 unless you want audio in your phonecall Jun 25 20:22:44 oh, nice! http://www.outflux.net/blog/archives/2012/02/15/discard-hole-punching-and-trim/ Jun 25 20:22:46 DocScrutinizer05: haha Jun 25 20:22:49 so you have to workaround it with LD_PRELOAD stubs etc. Jun 25 20:22:53 what's missing for audio then ? Jun 25 20:22:58 dos1: eww Jun 25 20:23:43 looks like since kernel 3.2 the loopback backing file _should_ shrink if stuff is deleted from the loppback device Jun 25 20:23:48 you need the old kernel API (/sys /proc /dev) for the blobs (dialer et al) to work Jun 25 20:23:58 as long as the FS supports trim Jun 25 20:24:21 cool Jun 25 20:25:09 Snafu777: won't work usually Jun 25 20:25:18 except for some basic cmdline stuff Jun 25 20:25:37 That sucks =) So much for "hacking" it together Jun 25 20:25:49 M4rtinK: even with non-sparse image files ? Jun 25 20:25:50 boot debian *instead* of maemo then :D Jun 25 20:26:00 dos1: haha Jun 25 20:26:01 a ddate might just work fine ;-) Jun 25 20:26:02 The whole lack of a phone is why i dont Jun 25 20:26:07 cmd: Today is Sweetmorn, the 30th day of Confusion in the YOLD 3180 Jun 25 20:26:30 after spending some time on it you could get working phone functionality on plain debian Jun 25 20:26:38 yeah Jun 25 20:26:42 with FSO for instance Jun 25 20:26:49 But that knowledge is limited to those who blog about it Jun 25 20:26:52 Which is nill.... Jun 25 20:26:52 or ofono (have a look at nemo) Jun 25 20:27:12 I'd love to have my phone have 1 partition and 1 partition only Jun 25 20:27:15 THAT would be superb Jun 25 20:27:16 pita though Jun 25 20:27:19 but then you become maintainer of your own phone distro, which has both upsides and downsides :) Jun 25 20:27:21 None of this optimizing nonsense Jun 25 20:28:01 dos1: I've always wished I could maintain a working openembedded tree for n900 (with phone/sms) ... but .... meh :) Jun 25 20:28:07 *dreaming*dreaming* Jun 25 20:28:33 bencoh: there was SHR working on N900 at some point of time Jun 25 20:28:41 phone included ? Jun 25 20:28:45 sadly, latest images didn't boot for me and I didn't have time to investigate why Jun 25 20:28:47 I think FSO missed this part Jun 25 20:28:58 yes, phone included Jun 25 20:29:01 woot Jun 25 20:29:02 just not so stable yet ;) Jun 25 20:29:21 there is ISI modem plugin in fsogsmd Jun 25 20:29:24 do you remember at which point in time ? Jun 25 20:29:51 oh, a few years ago? Jun 25 20:30:07 bencoh: not sure, probably should be sparse Jun 25 20:30:17 when there was still some development on SHR/FSO@N900 Jun 25 20:30:50 BTW, I would recommend everybody to see what nice features have been coming to LVM & device mapper recently :) Jun 25 20:30:58 I guess http://wiki.freesmartphone.org/index.php/Phonet_Implementation isnt up to date then Jun 25 20:31:02 really nice stuff :) Jun 25 20:31:26 M4rtinK: I cant get nicer than it is .... can it ? :) Jun 25 20:31:47 like LVM cache (bcache/hybrid drive alternative) or the upcoming compression/block order targets for DM Jun 25 20:31:56 it can! :) Jun 25 20:31:59 :) Jun 25 20:32:47 IMHO it is basically becoming a UNIXy modular BTRFS that is based on proved tech so you can actually deploy it for serious stuff :) Jun 25 20:34:13 :) Jun 25 20:55:33 * Sicelo wonders why Android still insists on a vm/runtime configuration Jun 25 20:57:17 it makes sense to drop Dalvik, but didn't expect that it would be replaced by just a more 'fancy' brother, ART Jun 25 21:28:54 anybody tested http://www.amazon.com/SanDisk-Ultra-MicroSDXC-Memory-Adapter/dp/B00IIJ6W4S in N900 ? Jun 25 21:38:32 no Jun 25 21:39:51 not that I know of. I have 64GB Sandisk UHS. I don't see why >3TB will not work as long as it's reformatted from exFAT Jun 25 21:41:35 hmm? Jun 25 21:41:41 >3TB? Jun 25 21:42:46 sixwheeledbeast: you tested speed of your SanDisk 64GB? Jun 25 21:43:01 particularly for random writes? Jun 25 21:43:03 nope. Jun 25 21:43:40 I'd be *very* interested in how it performs, e.g for cache etc Jun 25 21:43:59 I needed space over speed. Also it was the only one available and a very good price at the time. Jun 25 21:45:55 It has all my music and documents on there plus two swap spaces for flopswap. Jun 25 21:46:10 maybe I can come up with a random write testsuite based on standard unix toolbox bits and allowing to test for different write chunk sizes Jun 25 21:46:19 flashbench Jun 25 21:46:26 ShadowJK: :-D Jun 25 21:46:28 Sound like an idea. Jun 25 21:46:59 ShadowJK: you got a binary? :-) Jun 25 21:47:28 https://github.com/bradfa/flashbench Jun 25 21:47:40 hmm, sounds like source Jun 25 21:47:47 no SB here Jun 25 21:48:12 https://wiki.linaro.org/WorkingGroups/KernelArchived/Projects/FlashCardSurvey Jun 25 21:48:22 Yeah I've not run it on N900 itself Jun 25 21:52:51 i just want a ram-based microsd :( Jun 25 22:00:52 you want a uSD with a huuuuge RAM buffer Jun 25 22:01:26 I want an alternative for a cheesy slow 32GB eMMC Jun 25 22:02:55 and I wonder if a UHS interface is compatible to the eMMC 8bit interface, i.e. if the OMAP could handle SDXC/UHS on its MMC1 interface that currently hosts the eMMC Jun 25 22:03:37 and if such 128 (or 64) GB uSD would be faster than the N900 eMMC then Jun 25 22:05:32 since *if* all that pans out, you prolly would wnat a uSd tray rather than a soldered eMMC chip in your phone, eh? Jun 25 22:06:40 I mean, why does N900 have one uSD tray only? A: because eMMC is faster, thanks to 8bit databus Jun 25 22:07:23 if a uSD was as fast (or even faster) than a soldered chip, wouldn't you prefer two uSD trays? Jun 25 22:07:59 even if one of both is inconvenient to reach Jun 25 22:10:51 definitely :) Jun 25 22:11:27 nod: Having multiple replaceable uSD would be better IMO Jun 25 22:15:42 Sorry it's 2TB not 3. This is the maximum that SDXC can go. Jun 25 23:17:09 hmmph, uptime 47days and my IroN900 actually starts cocking up. Huge delays in almost everything and audio gets hiccup Jun 25 23:17:36 killall most-of-ui-stuff ? :) Jun 25 23:17:45 lots of them leak Jun 25 23:17:46 sudo reboot Jun 25 23:17:51 that's mean Jun 25 23:18:07 Mem:239M used:181M buffers:9M cache:44M Jun 25 23:18:13 *shrug* Jun 25 23:18:25 hmm Jun 25 23:18:28 (and swap ?) Jun 25 23:18:40 swap 199/767M Jun 25 23:19:00 usually I kill systemui and things are back to "normal" Jun 25 23:19:04 maybe actually fragmented Jun 25 23:19:19 * DocScrutinizer05 does a manual reswap Jun 25 23:22:27 mean? that's mean: Jun 25 23:22:30 IroN900:~# cat /proc/swaps Jun 25 23:22:32 Filename Type Size Used Priority Jun 25 23:22:33 /dev/mmcblk0p3 partition 786424 204676 -1 Jun 25 23:22:35 /dev/mmcblk1p3 partition 136184 0 -2 Jun 25 23:22:36 IroN900:~# swapoff /dev/mmcblk0p3 Jun 25 23:23:41 let's see how long that takes Jun 25 23:24:10 that...will take a while... Jun 25 23:24:18 indeed ;-) Jun 25 23:24:42 hildon will need to end and kick a lot of preloaded stuff Jun 25 23:24:50 it's kind of counterintuitive indeed, it only has 204MB to copy... Jun 25 23:24:53 ooh, already done Jun 25 23:25:29 what speed microsd? Jun 25 23:25:35 Mem:239M used:210M buffers:3M cache:22M Jun 25 23:25:49 swap: 86/132MB Jun 25 23:26:12 Tasks: 189; been 201 Jun 25 23:26:13 whenever I swapoff my internal swap, it takes forever on my "class 10" (which I think is a lie) microsd... Jun 25 23:26:43 >>One study found that, in this random-access use, some Class 2 cards achieved a write speed of 1.38 MB/s, while all cards tested of Class 6 or greater (and some of lower Classes; lower Class does not necessarily mean better small-file performance), including those from major manufacturers, were over 100 times slower.<< Jun 25 23:27:21 HUNDRED times! Jun 25 23:27:35 wow...very possible Jun 25 23:27:37 over 900 Jun 25 23:27:45 i dont have any more class 2 cards Jun 25 23:28:08 class 4 might work too Jun 25 23:28:16 I had a 8GB class 2 microsd card on my 5230 that I was using as backing store Jun 25 23:28:41 putting it in my n900, the n900 fried it (or at least it seems like it, only used it a few hours and it died) Jun 25 23:29:05 I didn't use it as swap, might just be electrical incompatibility, or the card was on its last legs Jun 25 23:29:50 IroN900:~# swapon /dev/mmcblk0p3 Jun 25 23:29:52 IroN900:~# swapoff /dev/mmcblk1p3 Jun 25 23:29:55 also takes a while Jun 25 23:30:25 wonder what's the block size of the emmc vs the microsd cards Jun 25 23:31:34 wonder why this feels like taking even longer Jun 25 23:32:12 have a feeling that in order to increase the "MB/sec" class, they probably have huge block sizes, sacrificing small block read/modify/write Jun 25 23:32:52 exactly Jun 25 23:33:32 ok, reswap finished. At least it's still alive. When it doesn't behave, it's reboot time Jun 25 23:33:38 I really also wonder about how they do wear levelling on these SD cards, it must not be as good as on full sized ssds Jun 25 23:33:48 a boot is good Jun 25 23:34:35 I just swapon with a higher priority to entice the n900 to swap on microsd, alas, the emmc swapping is clearly faster... Jun 25 23:35:05 of course, it has 8bit data Jun 25 23:35:12 while uSD has only 4 Jun 25 23:35:25 now that you mentioned that there was a study on "class 6"+ being so slow, I figure the RMW routines suck on these cards, and it probably will swap itself to death earlier than on emmc. Jun 25 23:35:31 it oughta be faster Jun 25 23:35:48 (meaning, wearout will occur faster than expected) Jun 25 23:36:39 http://www.ccs.neu.edu/home/pjd/papers/fast10-flash.pdf Jun 25 23:38:03 https://wiki.linaro.org/WorkingGroups/KernelArchived/Projects/FlashCardSurvey Jun 25 23:42:39 Sandisk Ultra Class 4 'SD16G' sounds good Jun 25 23:45:44 ohnoes, now even ssh has lags of 0.5..2s Jun 25 23:45:49 reboot time Jun 25 23:46:23 weird, no idea what's causing this Jun 25 23:48:11 top 2 of htop CPU usage: htop:6%, /usr/bin/hildon-status-menu:0..2% Jun 25 23:48:51 CPU total: 8..10% Jun 25 23:48:56 *cough* Jun 25 23:50:08 http://privatepaste.com/650fa0455d :-P Jun 25 23:51:34 mine just did that too... Jun 25 23:51:48 i hate it when all the widgets disappear Jun 25 23:52:09 well, i love it when they re-appear ;-) Jun 25 23:52:24 without me typing in the SIM PIN Jun 25 23:52:43 yeah true, at least they seem to retain their position when you re-add them... Jun 25 23:52:52 now if it also would restart all the tasks that been running... Jun 25 23:53:02 oooh Jun 25 23:53:11 nah, I had no such problems Jun 25 23:53:34 I simply typed "reboot" Jun 25 23:54:02 which is really convenient, compared to the alternatives Jun 25 23:54:49 hmmmm Jun 25 23:54:51 Mem:239M used:131M buffers:22M cache:83M Jun 25 23:55:12 and of course damn tracker-indexer hogging CPU Jun 25 23:56:05 37 CPU seconds already Jun 25 23:56:30 but it's snappy again Jun 25 23:57:46 "Micro Center 16GB Class 10" Jun 25 23:58:28 * dos1 wonders how long will it take for maemo to fill all of neo900's 1gb ram Jun 25 23:58:43 2weeks at most ? :° Jun 25 23:58:51 2min? Jun 25 23:58:59 2 sec? Jun 25 23:59:40 2 sec seems kinda pessimistic :P Jun 25 23:59:49 should be fast enough to run a program that malloc() everything? :D Jun 25 23:59:55 "fill" is relative Jun 26 00:00:21 we're not counting caches and take normal, n900-like usage into account :P Jun 26 00:00:43 Mem:239M used:129M buffers:4M cache:102M Jun 26 00:01:08 Uptime: 00:10:57 Jun 26 00:01:22 and swap? Jun 26 00:01:29 tracker CPU time 1:22.18 Jun 26 00:02:05 do peoples' n900 software leak memory Jun 26 00:02:06 ? Jun 26 00:02:11 59/767MB Jun 26 00:02:29 some does Jun 26 00:02:41 lockscreen known to do Jun 26 00:02:57 I never use lockscreen, particularly not to unlock Jun 26 00:03:25 the default lockscreen or one from repo? Jun 26 00:03:50 default Jun 26 00:03:57 freemangordon fixed that Jun 26 00:04:02 interesting... oh okay Jun 26 00:04:10 iirc Jun 26 00:04:54 cssu has the fix? how fast did it leak? Jun 26 00:04:56 * DocScrutinizer05 feels like killing trackerd with cat /dev/urandom >`which trackerd` Jun 26 00:05:12 ask freemangordon Jun 26 00:05:47 I guess i've seen worse... Jun 26 00:06:57 the Linux ACPI and virt-manager leaks were noticeable ... Jun 26 00:07:05 both have been fixed though... Jun 26 00:07:09 hm, so lockscreen is responsible for systemui leak ? Jun 26 00:07:26 (not for n900 of course) Jun 26 00:07:44 dunno anything about the lockscreen leak unfortunately Jun 26 00:07:59 bencoh: mainly, but iirc among others Jun 26 00:08:04 :] Jun 26 00:08:14 I suspect everything leaks on this thing Jun 26 00:08:23 nope, actually not Jun 26 00:08:41 a lot of stuff is empirically proven to not leak Jun 26 00:12:20 http://privatepaste.com/d695c34a17 the top 4 suuuuuck Jun 26 00:12:29 top 5 actually Jun 26 00:13:40 modest all but modest Jun 26 00:14:05 6.8 MEM% WTF? Jun 26 00:15:57 * DocScrutinizer05 finally feels fed up with modest and patches the .desktop Jun 26 00:30:13 http://privatepaste.com/ae857f2b97 Jun 26 00:30:20 ok, time for a walk **** ENDING LOGGING AT Thu Jun 26 02:59:58 2014