**** BEGIN LOGGING AT Fri Feb 10 03:00:02 2017
Feb 10 08:05:23 <zGrr> moin :)
Feb 10 12:34:45 <Oksanaa> LEDs work now, for no visible reason. Which package is better, xiph-support or decoders-support? 
Feb 10 12:36:24 <KotCzarny> better for what?
Feb 10 12:38:30 <Oksanaa> For tracking and decoding wide variety of audio and video files. So far, I have decoders-support installed, and flac and other audio gets tracked and played, while video gets tracked, but some of it gives "Media format not supported" errors.
Feb 10 12:39:14 <KotCzarny> xiph obviously only means ogg/vorbis/theora
Feb 10 12:39:41 <KotCzarny> i would bet that decoders-support has it ot relies on it
Feb 10 12:43:25 <Oksanaa> Decoders-support relies on ogg-support, which conflicts with xiph-support. And currently, an ogv with theora encoded video gives error video codec not supported error.
Feb 10 12:44:15 <KotCzarny> because probably ogg-support doesnt have video codec support
Feb 10 12:44:58 <KotCzarny> uninstall ogg-support and xiph-support, install decoders-support and see if it plays more formats?
Feb 10 12:45:43 <Oksanaa> Decoders-support depends on ogg-support and is already installed. Ogg-support and xiph-support have identical dependencies.
Feb 10 12:47:02 <Juesto> hey oksanaa
Feb 10 12:47:10 <Juesto> Strange
Feb 10 13:00:34 <Oksanaa> Installing xiph-support seems only to worsen the situation. 
Feb 10 13:02:59 <Juesto> Is ffmpeg available?
Feb 10 13:04:35 <KotCzarny> old one
Feb 10 13:04:57 <KotCzarny> i just compile my player with latest libav statically, less hassles
Feb 10 13:10:50 <Oksanaa> Installed back ogg-support, still not playing this theora.ogv video. Siberian hitch from Wikimedia Commons.
Feb 10 13:11:02 <Juesto> OGV?
Feb 10 13:11:03 <Juesto> uhm
Feb 10 13:11:08 <Juesto> link of the file?
Feb 10 13:11:14 <KotCzarny> ogg is a container, as in 'avi'
Feb 10 13:11:20 <Juesto> yeah
Feb 10 13:11:25 <KotCzarny> real codec is called vorbis for audio and theora for video
Feb 10 13:11:26 <Juesto> i want to look at the file via media info
Feb 10 13:11:47 <Juesto> oksanaa ^
Feb 10 13:20:54 <Oksanaa> https://upload.wikimedia.org/wikipedia/commons/transcoded/f/f1/Siberian_hitch.webm/Siberian_hitch.webm.480p.ogv
Feb 10 13:26:18 <Juesto> Oksanaa: since it's transcoded, what about MP4 ?
Feb 10 13:27:08 <Juesto> sigh
Feb 10 13:27:10 <Juesto> its impossible
Feb 10 13:27:12 <Oksanaa> Prefer open codecs
Feb 10 13:27:18 <Juesto> webm or theora only
Feb 10 13:27:27 <Juesto> Does maemo supprot webm?
Feb 10 13:27:58 <Oksanaa> Webm stutters. Ogv is better. It did work before.
Feb 10 13:28:31 <Juesto> vp8/vorbis
Feb 10 13:28:36 <Juesto> Ouch
Feb 10 13:28:45 <Juesto> Oksanaa: What changed? what broke?
Feb 10 13:28:50 <Juesto> when it broke?
Feb 10 13:30:46 <Oksanaa> No Idea. And installing gstreamer0.10-theora doesn't help.
Feb 10 13:31:35 <Juesto> Do you remember what was installed when it was working?
Feb 10 13:31:55 <KotCzarny> try to play it from command line via gstreamer
Feb 10 13:32:08 <Juesto> KotCzarny: is custom compiled ffmpeg for maemo good?
Feb 10 13:32:15 <Juesto> as in, how well your copy works
Feb 10 13:32:39 <Juesto> it is stable enough for other people to use it?
Feb 10 13:32:46 <KotCzarny> juesto: since my player is mainliny audio player, i only enable audio codecs
Feb 10 13:32:53 <KotCzarny> *mainly
Feb 10 13:32:59 <Juesto> Ah
Feb 10 13:33:00 <KotCzarny> and it's stable
Feb 10 13:33:02 <Juesto> well and video?
Feb 10 13:33:11 <KotCzarny> for video i just call gstreamer
Feb 10 13:33:35 <KotCzarny> not that i use it for video, but it was an easy code, so it's there
Feb 10 13:33:43 <Juesto> Ah i see
Feb 10 13:34:42 <KotCzarny> not to mention software decoding video on single core a9@600mhz isnt going to show fireworks
Feb 10 13:35:08 <Juesto> Should she try 360p version?
Feb 10 13:36:15 <KotCzarny> someone should write dsp optimized video codec for vp8/9 and theora, but that's hard even for people knowing assembly
Feb 10 13:37:19 <Oksanaa> KotCzarny: gst-launch ffdemux_ogg /home/user/MyDocs/.videos/Siberian_hitch.webm.480p.ogv 
Feb 10 13:37:21 <Oksanaa>  works, or says that it works. No idea how to actually see the video on the screen.
Feb 10 13:38:00 <KotCzarny> oksanaa: you can either check gst cheat sheets or try gst123 (i'm using it in oscp so you can grab it from there or compile your own)
Feb 10 13:38:26 <KotCzarny> you have to add proper chains (in/decode/out)
Feb 10 13:42:54 <Juesto> armel assembly specifically right?
Feb 10 13:43:08 <KotCzarny> nope, dsp code
Feb 10 13:43:13 <Juesto> Hm
Feb 10 13:43:18 <Juesto> ah right
Feb 10 13:43:40 <Juesto> isnt there something like mplayer or totem for maemo?
Feb 10 13:43:47 <Juesto> for sapwood/hildon
Feb 10 13:43:55 <KotCzarny> yes, but it uses software decoding
Feb 10 13:43:59 <KotCzarny> ie. cpu
Feb 10 13:44:09 <Juesto> any way to make it use gstreamer?
Feb 10 13:44:20 <Juesto> as in, is CLI accelerated or not?
Feb 10 13:44:48 <KotCzarny> gstreamer doesnt have many modern hw codecs
Feb 10 13:45:32 * Juesto sighs
Feb 10 13:47:02 <L29Ah> damn n900 doesn't want to boot while being charged from my laptop again
Feb 10 13:47:36 <Juesto> L29Ah: charge a bit then turn it on?
Feb 10 13:47:44 <L29Ah> charging indication stops after a few seconds
Feb 10 13:47:46 <Juesto> does it even charge?
Feb 10 13:47:49 <Juesto> ouch
Feb 10 13:47:51 <L29Ah> how do i tell?
Feb 10 13:48:00 <Juesto> wait..?
Feb 10 13:48:14 <KotCzarny> pulsing amber or green?
Feb 10 13:48:17 <L29Ah> now the charging indication is locked on its maximum brightness
Feb 10 13:48:19 <L29Ah> amber
Feb 10 13:48:26 <KotCzarny> pulsing or constant?
Feb 10 13:48:34 <L29Ah> used to be pulsing, now constant
Feb 10 13:48:42 <KotCzarny> might not be charged at all
Feb 10 13:48:45 <KotCzarny> ~flatbat
Feb 10 13:48:49 <KotCzarny> ~ping
Feb 10 13:48:49 <infobot> ~pong
Feb 10 13:48:53 <Juesto> Rip
Feb 10 13:48:55 <KotCzarny> hrm
Feb 10 13:49:29 <KotCzarny> ~flatbatrecover
Feb 10 13:49:29 <infobot> Remove battery for 1 minute. Insert battery. Plug powered ***NOKIA WALLCHARGER*** to device. Watch steady amber. Let sit and charge. Do NOT try to boot. After 30 min, you got either a) a booted up N900, b) flashing amber which means you can boot, c) steady amber shut off -> start over again with ~flatbatrecover while already searching for a new battery. CAVEAT! Only works when ~rootfs OK (no ~bootloop)
Feb 10 13:50:00 <L29Ah> hooked it into a dumb charger, see same behaviour
Feb 10 13:50:15 <L29Ah> > steady amber
Feb 10 13:50:17 <L29Ah> ok i guess
Feb 10 13:50:22 <Juesto> Tried a proper charger?
Feb 10 13:50:28 <Juesto> ah
Feb 10 13:50:30 <Juesto> hm
Feb 10 13:50:31 <L29Ah> dunno what's a "proper charger"
Feb 10 13:50:41 <KotCzarny> proper charger == original nokia one
Feb 10 13:50:50 <L29Ah> :D
Feb 10 13:50:54 <Juesto> sometimes chargers have more power to charge faster
Feb 10 13:50:55 <Juesto> Lol
Feb 10 13:51:05 <Juesto> voltage*
Feb 10 13:51:08 <Juesto> or whatever
Feb 10 13:51:16 <L29Ah> i doubt n900 supports that
Feb 10 13:51:50 <L29Ah> i can just hook my liion charger wires to my 18650 block but i decided to have a try with usb stuff just in case
Feb 10 13:52:17 <Juesto> be careful
Feb 10 14:07:10 <joga> heh... I had problems with charging the n900 before from a powerbank using a certain cable... now I couldn't get it to charge anymore after many attempts, and I was outside home, so got the cheapest microusb cable I could find nearby and woo, no problems charging. It also feels a lot less loose than the previous cable so maybe it's just mechanical wear on the old cable or whatever
Feb 10 16:21:34 <L29Ah> no it didn't charge
Feb 10 16:21:37 <L29Ah> that bastard
Feb 10 16:21:56 <L29Ah> reporting to be charging but not charging
Feb 10 16:23:24 <doppler> :(
Feb 10 16:35:12 <freemangordon> L29Ah: it is charging, but slowly, only 100 mA
Feb 10 16:36:24 <freemangordon> (assuming it is still connected to your laptop USB port)
Feb 10 16:36:53 <L29Ah> it got +10mV in 2h on a 4A*h battery
Feb 10 16:37:05 <L29Ah> (3.40 to 3.41)
Feb 10 16:37:36 <L29Ah> maybe it didn't even charge and it's the battery capacitance recovered from a current surges of starting up n900
Feb 10 16:37:39 <KotCzarny> maybe that low current isnt able to charge it?
Feb 10 16:37:43 <freemangordon> looks normal for 100 mA charge current to me
Feb 10 16:38:22 <freemangordon> L29Ah: just find 1A wall charger
Feb 10 16:38:23 <L29Ah> 3.40 to 3.41 is like 1%
Feb 10 16:38:44 <freemangordon> how did you measure that voltage?
Feb 10 16:38:46 <L29Ah> i used a wall charger
Feb 10 16:38:55 <L29Ah> probing the battery
Feb 10 16:38:59 <freemangordon> is it 1A or 2.1 A?
Feb 10 16:39:00 <L29Ah> with a multimeter
Feb 10 16:39:19 <L29Ah> 1A
Feb 10 16:39:24 <freemangordon> hmm
Feb 10 16:39:51 <freemangordon> could you check with your multimeter if there is a short between D+ and D- ?
Feb 10 16:40:09 <L29Ah> switched to another cable, now it has charged for real
Feb 10 16:40:13 <L29Ah> 3.55V
Feb 10 16:40:24 <KotCzarny> do you have any charge 'doctor' to confirm any usage?
Feb 10 16:40:28 <KotCzarny> hehe
Feb 10 16:40:31 <KotCzarny> figures
Feb 10 16:40:35 <KotCzarny> mark that other cable as bad
Feb 10 16:40:41 <L29Ah> and the former cable works just fine for data
Feb 10 16:41:00 <L29Ah> and it detects the charger somehow over it
Feb 10 16:41:01 <L29Ah> WTF
Feb 10 16:41:07 <L29Ah> what's so bad about the cable?
Feb 10 16:41:13 <freemangordon> contacts?
Feb 10 16:41:19 <KotCzarny> cheap, dumb?
Feb 10 16:41:20 <freemangordon> might be weared
Feb 10 16:41:37 <L29Ah> so they have like 3V worth of drop?
Feb 10 16:42:15 <freemangordon> I doubt bq will wait for 3V drop before stopping the charging
Feb 10 16:42:24 <L29Ah> what microusb cables should i buy from ebay btw?
Feb 10 16:42:37 <KotCzarny> 24avg or less
Feb 10 16:42:46 <L29Ah> they don't specify this as i see
Feb 10 16:42:53 <KotCzarny> then dont buy those
Feb 10 16:42:55 <freemangordon> L29Ah: you'd better buy charger, not cables
Feb 10 16:43:04 <L29Ah> i don't need a charger
Feb 10 16:43:09 <freemangordon> ok
Feb 10 16:43:12 <L29Ah> i need a cable for emergency charging from my laptop
Feb 10 16:43:24 <KotCzarny> then get good data cable
Feb 10 16:43:30 <KotCzarny> 4 wires
Feb 10 16:43:31 <L29Ah> i have a 4A LiIon charger for regular stuff
Feb 10 16:43:31 <freemangordon> L29Ah: it doesn;t work like that
Feb 10 16:43:50 <freemangordon> the most you got over USB is 500 mA
Feb 10 16:43:57 <freemangordon> BUT
Feb 10 16:44:01 <L29Ah> but i did the test to ensure i can charge it a bit in case i'm in a forest with my laptop and a cable
Feb 10 16:44:08 <KotCzarny> fmg, some laptops are fine with much more than 500mA
Feb 10 16:44:16 <freemangordon> you should have the OS booted, iirc, otherwise you got only 100 mA
Feb 10 16:44:41 <freemangordon> KotCzarny: it is not what port can handle, it is what n900 will draw
Feb 10 16:44:47 <freemangordon> *not about
Feb 10 16:44:57 <KotCzarny> fmg, oh
Feb 10 16:45:05 <L29Ah> can it be hacked so it draws 1A in every circumstance?
Feb 10 16:45:24 <freemangordon> L29Ah: yye, you need to short-circuot D+ and D-
Feb 10 16:45:27 <L29Ah> w/o messing with pcb
Feb 10 16:45:30 <L29Ah> :(
Feb 10 16:45:43 <freemangordon> you can rework the canle
Feb 10 16:45:46 <freemangordon> *cable
Feb 10 16:45:58 <L29Ah> i expected a software solution
Feb 10 16:46:39 <L29Ah> i see that n900 boots the kernel even to start charging-blinking stuff
Feb 10 16:46:39 <freemangordon> not sure, maybe there exists such a solution if you use bme replacement
Feb 10 16:46:45 <freemangordon> yes
Feb 10 16:46:49 <freemangordon> oh, well
Feb 10 16:47:11 <freemangordon> NOLO supports charging as well, but iirc it draws 100 mA only
Feb 10 16:47:20 <freemangordon> Pali_: ^^^ ?
Feb 10 16:47:28 <freemangordon> an;t remember the details
Feb 10 16:47:33 <freemangordon> *can't
Feb 10 16:47:51 <Pali_> DocScrutinizer05 ^
Feb 10 16:47:55 <freemangordon> :)
Feb 10 17:25:41 <KotCzarny> https://wiki.maemo.org/Talk:Useful_Mobile_URLs
Feb 10 17:25:46 <KotCzarny> spam detected
Feb 10 17:26:09 <KotCzarny> anyone can clean it/block the user?
Feb 10 17:26:18 <KotCzarny> xes?
Feb 10 17:27:59 <KotCzarny> https://wiki.maemo.org/Talk:Task:Defining_maemo
Feb 10 17:28:01 <KotCzarny> ugh
Feb 10 17:28:56 <KotCzarny> quite a lot of spam in talk:
Feb 10 17:29:13 <KotCzarny> all users are unregistered
Feb 10 17:29:20 <sixwheeledbeast> KotCzarny: I'll look
Feb 10 17:29:47 <KotCzarny> https://wiki.maemo.org/Talk:MaemoCommunity_eV
Feb 10 17:29:49 <KotCzarny> etc
Feb 10 17:30:10 <KotCzarny> for easy spotting click history or diffs
Feb 10 17:31:05 <KotCzarny> i wonder for how long it went
Feb 10 17:31:43 <KotCzarny> maybe server admin can run some cleanup script (and block posting talk: entries containing links
Feb 10 17:32:24 <xes> https://wiki.maemo.org/index.php?title=Special:RecentChanges&days=60&limit=500
Feb 10 17:32:39 <KotCzarny> xes, yeah, a lot
Feb 10 17:33:27 <xes> i will check the wiki's anti spam filter,maybe it's not exaclty aligned to the tmo one
Feb 10 17:33:51 <KotCzarny> i think its because wiki allows anonymous posting in talk: pages
Feb 10 17:34:06 <xes> (the tmo one has looonger memory for bad guys)
Feb 10 17:37:20 <sixwheeledbeast> Ok fixed and bliacklisted IP's
Feb 10 17:38:12 <xes> sixwheeledbeast: thanks
Feb 10 17:38:29 <sixwheeledbeast> KotCzarny: Any others? I used to have the time to check for spam on a daily basis but unfortunately not any more.
Feb 10 17:39:04 <sixwheeledbeast> Any spam on the wiki ping me I can sort it.
Feb 10 17:39:08 <sixwheeledbeast> xes: np
Feb 10 17:39:46 <KotCzarny> swb: check all anonymous (ip based) adds
Feb 10 17:41:11 <KotCzarny> https://wiki.maemo.org/Talk:Community_SSU
Feb 10 17:44:42 <sixwheeledbeast> KotCzarny: just done that, I can only check back 30 days and I have just blacklists and rolledback those.
Feb 10 17:45:19 <KotCzarny> i cant get history older than 3 dec 2016
Feb 10 17:45:48 <sixwheeledbeast> ok 60 days
Feb 10 17:46:32 <sixwheeledbeast> oh I can manually edit the days in the URL
Feb 10 17:46:44 <KotCzarny> but i would advise someone to run some searches on anonymous adds in talk: pages
Feb 10 17:48:06 <sixwheeledbeast> Yep 3 dec is as far back as I can search
Feb 10 17:49:01 <KotCzarny> btw. wmo is missing favicon
Feb 10 17:49:31 <KotCzarny> and favicons make links nicer in bookmarks ;)
Feb 10 17:50:10 <KotCzarny> http://wiki.maemo.org/favicon.ico
Feb 10 17:50:11 <KotCzarny> ;)
Feb 10 17:50:59 <KotCzarny> but its referenced in wmo pages: <link rel="shortcut icon" href="/favicon.ico" />
Feb 10 17:51:56 <sixwheeledbeast> :nod: I already reported that but NFC what the answer was.
Feb 10 17:52:16 <KotCzarny> just copy it from tmo ?
Feb 10 17:52:31 <KotCzarny> maybe att white T over it to make it distinct
Feb 10 17:52:37 <KotCzarny> *add
Feb 10 18:30:49 <sixwheeledbeast> KotCzarny: should point to href="/skins/monobookmaemo2009/static/maemo_favicon.ico"
Feb 10 20:09:13 <ff_> can n900 use all 90MB/s r/w speed of https://www.sandisk.com/home/memory-cards/microsd-cards/extremeplus-microsd ?
Feb 10 20:09:49 <KotCzarny> nope. your target is a card with highest iops you can get, not sequential speed
Feb 10 20:11:10 <ff_> and the lowest access time?
Feb 10 20:12:01 <KotCzarny> yes, but getting real (4k r/w) iops benchmarks is tricky
Feb 10 20:12:52 <ff_> some example?
Feb 10 20:13:43 <ff_> it's possible make block size bigger then 4k?
Feb 10 20:15:04 <KotCzarny> most noname cards have iops in 0.1-1/s range
Feb 10 20:15:30 <KotCzarny> some samsung evo +/pro manage to get ~100iops
Feb 10 21:26:55 <ff_> you're right http://www.cameramemoryspeed.com/reviews/sd-cards/sandisk-extreme-plus-95-90-mbs-64gb-micro-sdxc-memory-card/ only 3MB/s random write, 700ish iops
Feb 10 21:30:51 <ff_> ssd are designed for much more iops https://news.samsung.com/global/samsung-introduces-worlds-first-universal-flash-storage-ufs-removable-memory-card-line-up-offering-up-to-256-gigabyte-gb-capacity 40000 iops, but that's not helpful for n900
Feb 10 21:35:09 <Juesto> Nice!
Feb 10 22:05:47 <ff_> with 35000 iops it could works like RAM, so, n900 with extremely expensive 32GB RAM/ROM :)
Feb 10 22:05:58 <L29Ah> 23:15:04]<KotCzarny> most noname cards have iops in 0.1-1/s range
Feb 10 22:05:58 <L29Ah> what
Feb 10 22:29:55 <ff_> I couldn't be surprised if Chinese people could produce something like that
Feb 10 22:33:41 <ff_> anyone know speed of microSD data bus in n900?
Feb 10 22:34:07 <ff_> maybe it's not worth the bother with fast microSD
Feb 10 22:36:22 <Maxdamantus> I'm getting around 12 MB/s read.
Feb 10 22:36:56 <Maxdamantus> I have a feeling that the maximum speed does kind of matter.
Feb 10 22:37:18 <ff_> what are specs of sd?
Feb 10 22:37:30 <Maxdamantus> iirc I would only get around 4 MB/s on a lower-end Kingston microSD, which was rated at something like 30 MB/s.
Feb 10 22:38:38 <ff_> speed of read or write?
Feb 10 22:38:46 <Maxdamantus> SanDisk 128GB Ultra® microSDXC™ UHS-I Class 10 card
Feb 10 22:39:52 <Maxdamantus> SD cards don't really have a particular speed.
Feb 10 22:40:10 <Maxdamantus> It depends on things like how the controller has laid out the data.
Feb 10 22:41:00 <ff_> with it you get 12MB/s in n900?
Feb 10 22:41:18 <Maxdamantus> Yes.
Feb 10 22:41:22 <Maxdamantus> 236978176 bytes (237 MB) copied, 20.236 s, 11.7 MB/s
Feb 10 22:41:23 <ff_> and how with write?
Feb 10 22:42:23 <Maxdamantus> write is trickier to test.
Feb 10 22:43:08 <Maxdamantus> can try after remounting with -o sync, I guess.
Feb 10 22:44:20 <ff_> not pressure, I'm just curious
Feb 10 22:45:23 <Maxdamantus> I'm getting 10.6 MB/s writing to ext4 mounted with -o sync using dd with conv=fsync on
Feb 10 22:45:38 <Maxdamantus> (so that's to a filesystem, not just a block device)
Feb 10 22:45:59 <Maxdamantus> though I'm only writing zeroes, dunno if the card does something special with that.
Feb 10 22:46:54 <Maxdamantus> Well, I imagine the card should handle more than 10.6 MB/s anyway .. it'll at least be pushing that 10.6 MB/s in SDIO packets.
Feb 10 22:47:07 <ff_> I don't know also, thanks for real data with n900
Feb 10 22:47:43 <Maxdamantus> The sync options are also going to usually reduce overall throughput
Feb 10 22:48:45 <ff_> it seems that n900 has 10MB/s speed limit on sd bus
Feb 10 22:48:53 <Maxdamantus> These speeds are all sequential btw, not random.
Feb 10 22:50:27 <ff_> random should be lower than this 10MB/s cameramemoryspeed.com/reviews/sd-cards/sandisk-extreme-plus-95-90-mbs-64gb-micro-sdxc-memory-card/
Feb 10 22:51:43 <ff_> 9MB/s for read and 3MB/s for write if that benchmark is with the same class card
Feb 10 22:55:11 <ff_> microSDXC UHS-I U3, I'm not sure for that U3 symbol
Feb 10 23:34:33 <DocScrutinizer05> a pty ff_ left. Was absolutely right it doesn't really make sense to bother, since databus is limited to 4bit buswidth * (iirc) 40MHz clockrate max, so 20MB/s
Feb 10 23:35:08 <DocScrutinizer05> *might* even be just 25MHz
Feb 10 23:35:34 <Oksanaa> Question is, which cards actually work at maximum limit of the databus.
Feb 10 23:38:07 <DocScrutinizer05> all
Feb 10 23:38:10 <DocScrutinizer05> :-)
Feb 10 23:38:36 <Maxdamantus> except for my Kingston one.
Feb 10 23:39:21 <DocScrutinizer05> it's a pretty ancient databus protocol. Recent Ultra card only work since they support legacy protocols, the OMAP3 doesn't support those new speeds
Feb 10 23:39:57 <DocScrutinizer05> UHS-I and UHS-II are not supported by OMAP3
Feb 10 23:40:43 <DocScrutinizer05> for write, the real show stopper is page erase needed to rewrite a block with new data. It can take up to some 100s of MILLIseconds
Feb 10 23:41:12 <DocScrutinizer05> at very least 10ms
Feb 10 23:41:19 <DocScrutinizer05> per page erase
Feb 10 23:42:20 <Oksanaa> Is it possible to have faster databus with OMAP3? And what is page erase, is it like going over the card with eraser to be sure that write to card goes well?
Feb 10 23:43:31 <DocScrutinizer05> yes @ #2
Feb 10 23:43:51 <DocScrutinizer05> no @ #1, it's a physical interface that can't do any faster
Feb 10 23:44:39 <DocScrutinizer05> OMAP has a 4bit and a 8bit storage bus, both with same max clock frequency. In N900 the 8bit is for eMMC
Feb 10 23:44:42 <Oksanaa> Like, would Neo900 be able to have faster databus than N900?
Feb 10 23:44:47 <DocScrutinizer05> no
Feb 10 23:44:51 <DocScrutinizer05> alas not
Feb 10 23:45:36 <DocScrutinizer05> that's also one killer argument why Neo900 must have eMC despite many people would prefer two uSD slots: eMMC supports 8bit
Feb 10 23:46:05 <Oksanaa> About page erase: is it possible to do it beforehand, like use idle time to erase empty blocks on the card so that by the time a write is needed, card already has several blocks ready?
Feb 10 23:46:26 <Maxdamantus> I imagine that's something the card will already do.
Feb 10 23:46:26 <Oksanaa> uSD slot supports only 4bit?
Feb 10 23:46:43 <Maxdamantus> Hosts don't interact with SD cards as flash devices.
Feb 10 23:46:58 <Maxdamantus> That's all stuff that's handled by the computer inside the SD card.
Feb 10 23:47:17 <Oksanaa> Maxdamantus: that would be neat, if it's implemented already.
Feb 10 23:49:23 <DocScrutinizer05> (do it beforehand) that's how all slightly smarter cards work, as long as the OS supports TRUNCATE
Feb 10 23:49:42 <Maxdamantus> TRUNCATE? Do you mean TRIM?
Feb 10 23:50:14 <Maxdamantus> if so, that has nothing no do with SD cards.
Feb 10 23:50:20 <DocScrutinizer05> the card needs to know which blocks can get erased beforehand, even writing all zeroes to a whole page doesn't automatically tell card "this is no data, this is empty storage free to use"
Feb 10 23:50:38 <Maxdamantus> well, that's not what SD cards typically do.
Feb 10 23:51:01 <DocScrutinizer05> uSd has only 4 bit data, yes. count the pins on your uSD card ;-)
Feb 10 23:51:16 <DocScrutinizer05> you need GND, VCC, clock r/w whatnot else
Feb 10 23:51:34 <Maxdamantus> DocScrutinizer05: can you provide a link to some information about this magical "TRUNCATE" thing you're talking about?
Feb 10 23:51:50 <Maxdamantus> or is this another case where you're just saying things about technologies you don't understandL
Feb 10 23:55:21 <DocScrutinizer05> Oksanaa: https://wiki.debian.org/SSDOptimization
Feb 10 23:55:40 <Maxdamantus> We're talking about SD cards, not SSDs.
Feb 10 23:55:57 <Maxdamantus> SD cards don't have TRIM (or "TRUNCATE" if you like making up names for things)
Feb 10 23:56:17 <DocScrutinizer05> basically all SSD rationale also applies to SD cards
Feb 10 23:56:35 <Maxdamantus> SD is a completely different protocol.
Feb 10 23:56:44 <Maxdamantus> and it doesn't have anything like TRIM in it.
Feb 10 23:56:51 <DocScrutinizer05> sorry, DISCARD, not TRUNCATE
Feb 10 23:58:38 <Maxdamantus> apparently it's actually ERASE
Feb 11 00:02:44 <DocScrutinizer05> the best way is to actually never use a maybe 10% of the SD card, by downsizing the partition(s)
Feb 11 00:04:46 <DocScrutinizer05> once the system wrote to all available blocks (and didn't issue a DISCARD or whatever the equivalent name on SD bus, for all the no longer needed blocks, and made sure the non-used blocks concatenate to a whole erase page), the card will significantly drop on write performance
Feb 11 00:05:14 <Oksanaa> So, use up to 90% or so of the card.
Feb 11 00:05:52 <DocScrutinizer05> since the card controller can't fortune tell which pages to erase "preemptively"
Feb 11 00:06:20 <DocScrutinizer05> card always have a few hidden spare pages, but not that many
Feb 11 00:06:59 <DocScrutinizer05> Oksanaa: (90%) I mere guess by me, but sounds good
Feb 11 00:07:26 <Oksanaa> So, if somebody made a card with high enough quantity of hidden spare pages, that would be easier on user.
Feb 11 00:08:16 <DocScrutinizer05> yes
Feb 11 00:08:53 <Maxdamantus> Well, presumably it'll usually have around 73 MB per GB of spare blocks.
Feb 11 00:09:16 <Maxdamantus> if its capacity is given in GB.
Feb 11 00:09:44 <Maxdamantus> since 1 GiB is a bit more than 73 MB more than a GB.
Feb 11 00:10:57 <Maxdamantus> flash usually comes in a power-of-2 number of bits, then the SD card will offer a power-of-2 number of GBs or MBs.
Feb 11 00:11:16 <DocScrutinizer05> maybe interesting (didn't read it completely): https://blogofterje.wordpress.com/2012/01/14/optimizing-fs-on-sd-card/
Feb 11 00:16:08 <DocScrutinizer05> https://haydenjames.io/increase-performance-lifespan-ssds-sd-cards/  here it's called TRIM
Feb 11 00:16:09 <Juesto> Do compactation?
Feb 11 00:16:54 <DocScrutinizer05> though >>Therefore, when you delete a file, your SSD is now able to write data to blocks as if it they were brand new without having to perform the cumbersome deletion process.<< obviously is utter bullshit
Feb 11 00:27:15 <DocScrutinizer05> MEH! I need to avoid looking into chanlog, always an annoyance to read that crap my /ignore luckily usually saves me to read
Feb 11 00:31:24 <DocScrutinizer05> Juesto: one of the major problems about uSD embedded controller is: it doesn't know about ext-fs or anything else than VFAT basically. On ext[234] you have inodes scattered all over the complete storage
Feb 11 00:32:39 <DocScrutinizer05> it's assumed that the SD controller even knows about File allocation Table in VFAT and optimizes using VFAT properties it knows about. Can't do same for any other filesystem
Feb 11 00:36:07 <DocScrutinizer05> Juesto: anyway SSD/SD analysis and optimization is absolutely non-trivial, due to the complex way a flash storage works and the obscure SEKRIT sauce the uSD controller adds on top
Feb 11 00:36:44 <luke-jr> Maxdamantus: SD cards have a virtual block layer on top of the real flash, which adds overhead for remapping bad blocks
Feb 11 00:36:54 <DocScrutinizer05> that's why for really tight control and fast r/w access you want NAND and a filesystem like ubifs
Feb 11 00:37:11 <DocScrutinizer05> oh hi luke-jr :-)
Feb 11 00:37:17 <luke-jr> hi DocScrutinizer05
Feb 11 00:38:39 * DocScrutinizer05 seems to recall there were plans or even initiatives to reflash the uSd controller firmware to allow using the uSD like it was normal un-managed NAND storage
Feb 11 00:39:49 <DocScrutinizer05> obviously such a project would be highly brand&make specific on the particular uSD
Feb 11 00:50:33 <DocScrutinizer05> Oksanaa: re "90%" don't get fooled: see https://www.sandisk.com/home/memory-cards/sd-cards/extremeplus-sd-uhs-i "DISCLOSURES: ... 1MB=1,000,000 bytes. ** 1GB=1,000,000,000 bytes"
Feb 11 00:51:35 <DocScrutinizer05> *plus*:  "Actual user storage less."
Feb 11 00:55:39 <DocScrutinizer05> so a 32GB uSD has a 32E9 bytes, minus a non-marginal fraction for overhead like spare blocks and prolly also flash controller metadata etc
Feb 11 01:19:01 <DocScrutinizer05> AAAH and there's the term I missed: FTL aka Flash Translation Layer (aka "the uSD embedded controller)  http://www.linux-mtd.infradead.org/doc/ubifs.html
Feb 11 01:19:57 <DocScrutinizer05> s/"the uSD embedded controller("the uSD embedded flash controller"/
Feb 11 01:21:01 <DocScrutinizer05> http://www.linux-mtd.infradead.org/doc/ubifs.html#L_raw_vs_ftl
Feb 11 01:21:27 <Maxdamantus> 13:55:39 < DocScrutinizer05> so a 32GB uSD has a 32E9 bytes, minus a non-marginal fraction for overhead like spare blocks and prolly also flash controller metadata etc
Feb 11 01:21:30 <Maxdamantus> No.
Feb 11 01:21:53 <Maxdamantus> It has 32 GiB inside, but provides at least 30 GB that you can write to.
Feb 11 01:23:05 <Maxdamantus> You practically can't get 32 GB in flash chips .. a 32 GB SD card will have 32 GiB of flash inside, but the block device you end up with will be around 32 GB.
Feb 11 01:25:28 <Maxdamantus> Hm, actually, my 128 GB card only has 127865454592 bytes.
Feb 11 01:26:35 <DocScrutinizer05> the trick with optimizing uSD is:http://www.linux-mtd.infradead.org/faq/general.html#L_mtd_vs_hdd you have an interface/protocol for a block device (left side of table, with extensions like TRIM for example), but you try to implement a "usage pattern" aka filesystem and driver that takes care about properties described on right side half of the table
Feb 11 01:27:56 <DocScrutinizer05> you need a priori knowledge (or assumptions) about how the FTL firmware works and what it does in response to what you do on the block device interface
Feb 11 01:33:44 <Maxdamantus> The eMMC on my N900 has 32015122432 bytes though.
Feb 11 01:33:51 <DocScrutinizer05> http://www.linux-mtd.infradead.org/faq/general.html#L_ext2_mtd >>But in many cases using mtdblock is a very bad idea because what it basically does if you change any sector of your mtdblockX device, it reads the whole corresponding eraseblock into the memory, erases the eraseblock, changes the sector in RAM, and writes the whole eraseblock back. This is very straightforward. If you have a power failure when the eraseblock is being erased, 
Feb 11 01:33:52 <DocScrutinizer05> you lose all the block device sectors in it. The flash will likely decay soon because you will wear few eraseblocks out - most probably those ones which contain FAT/bitmap/inode table/etc.<<
Feb 11 01:34:46 <DocScrutinizer05> eraseblock (I called it erase page, sorry) is as large as several dozen or even hundreds of kB
Feb 11 01:35:18 <DocScrutinizer05> and to change a singke bit in it, you need to read modify erase write the whole eraseblock
Feb 11 01:35:41 <Maxdamantus> DocScrutinizer05: yeah, that's why you shouldn't use something like mtdblock, as mentioned above.
Feb 11 01:35:51 <Maxdamantus> DocScrutinizer05: something sensible would keep writing to new blocks.
Feb 11 01:36:06 <Maxdamantus> that's basically the principle behind log-based filesystems.
Feb 11 01:36:33 <Maxdamantus> The FTLs in SD cards probably do something similar.
Feb 11 01:36:44 <DocScrutinizer05> *unless* the bit change is 0->1, assuming all-0s is the result of block-erase (which takes several milliseconds as I already explained)
Feb 11 01:37:47 <Maxdamantus> https://en.wikipedia.org/wiki/List_of_log-structured_file_systems
Feb 11 01:38:15 <Maxdamantus> The filesystems at the bottom are all the standard ones for flash.
Feb 11 01:38:24 <Maxdamantus> including UBIFS, which is used by maemo.
Feb 11 01:38:54 <DocScrutinizer05> so stuff like FTL and ubifs implements very nifty ways to not always read-modify-write for a few bytes that need to be changed, rather it uses journals etc that simply get extended at end, writing to the already/still all-0 region of a erase block
Feb 11 01:38:59 <Maxdamantus> an SD card or MMC has to do the same sort of stuff, but it just provides a block device instead of a filesystem.
Feb 11 01:39:55 <DocScrutinizer05> and TRIM deals with telling FTL that a complete eraseblock is 'clean' again and ready to get erased when there's some spare time in background for doing so
Feb 11 01:43:41 <DocScrutinizer05> (simplified image, since TRIM doesn't really care about erase blocks, it has no notion about them afaik. It's up to the FTL to deduce erase blocks needing a background erase, from what  fs blocks (512) TRIM told the FTL are unused
Feb 11 01:43:46 <DocScrutinizer05> )
**** ENDING LOGGING AT Sat Feb 11 03:00:01 2017