#qi-hardware IRC log for Sunday, 2012-01-01

xiangfuHappy New Year. :D05:18
kristianpaullets gor fot that :)05:19
kristianpauls/gor/go05:19
wolfspraulkristianpaul: thanks for helping pabs308:20
wolfsprauland happy new year!08:20
wolfspraulnice post from Andrew about (advanced) homemade pcbs  http://siliconexposed.blogspot.com/2011/12/multiple-lithography-in-homemade-pcbs.html09:55
xiangfu:D11:57
xiangfu why I add some 'serial_puts' make my u-boot-nand.bin not working? how to debug/find out such root cause?11:58
lars_is there a config option to stub out serial_puts?12:00
lars_or the uart is not properly initalized12:00
lars_clock disabled or incorrect pinmux12:00
xiangfulars_, it output correct. 12:01
xiangfuonly the u-boot-nand.bin not working any more. like failed at load the u-boot.bin to RAM.12:01
xiangfulars_,  http://lists.denx.de/pipermail/u-boot/2012-January/114464.html <--  the patch is here 12:01
xiangfulars_, and nand_spl only several lines12:02
lars_ah, ok. maybe the image is to big now someting like that12:04
xiangfulars_, when I remove the 'extern' lines from nand_spl.c also make u-boot-nand.bin not working. :(12:07
lars_thats rather stange, extern for functions should really matter12:11
lars_have you compared the binaries?12:13
lars_the working and the non-working one?12:13
GorDonFreeManhi12:32
GorDonFreeManrealtime preempted linux kernel's HRT & TSC test result data http://pastebin.com/1QF6bc9T12:32
GorDonFreeManopinions?12:32
xiangfulars_, testing now. 12:51
xiangfuqwebirc3657, you question is about Windows driver. you should ask Ingenic people or NOVO people about jz4770.13:40
xiangfuqwebirc3657, we are support Ben Nanonote which is using jz4720 cpu.13:40
xiangfuqwebirc3657, and most people here is using Linux system :)13:40
qwebirc3657how can i reach novo people?13:41
qwebirc3657i am sorry for the wrong place 13:41
xiangfuqwebirc3657, well. you bought the novo 7 . so there must some email or phone number on the manual.13:41
qwebirc3657my problem is i cannot speak chinese13:41
xiangfuqwebirc3657, oh.13:42
qwebirc3657and ALL information is in chinese13:42
qwebirc3657but with the help of Google translate13:42
xiangfuqwebirc3657, you only have Window system?13:42
qwebirc3657i located and downloaded the firmware upgrade that i need13:42
qwebirc3657yes13:42
qwebirc3657my computer is w7 6413:43
qwebirc3657and i am affraid that i am having trouble installing the drivers because of 64bits system13:43
qwebirc3657or else because of some chinese characters in the folder tree which may be a problem for windows not chinese13:44
qwebirc3657I got this new Novo 7 running honeycomb 3.213:47
qwebirc3657it just crashes everytime i use wifi, or when i try to access some given sttings, or even sometimes it crashes just while booting13:48
qwebirc3657i complained with the seller, butafter a while i understood that thay are really ignorant about tech stuff, they're just sellers13:49
qwebirc3657so i searched and found the ainol site (all in chinese), but with google translate i was able to find that there is a new firmware upgrade13:51
xiangfuqwebirc3657, I just search a little. : you can try this driver: http://www.multiupload.com/9WIR1W1NBA13:51
qwebirc3657i downloaded it13:51
qwebirc3657i am checking13:51
qwebirc3657downloading13:52
qwebirc3657wil this be proper for JZ4770 USB boot device?13:53
xiangfuI think so. 13:54
qwebirc3657thanks13:54
qwebirc3657i'll report back if it worked13:56
kristianpaulwolfspra1l: np, btw you added sdr.osmocom.org rss to our planet?13:57
kristianpaulmorning !13:57
xiangfuqwebirc3657, and here is the new NOVO7 firmware: http://www.ainol.com/plugin.php?identifier=ainol&module=download&action=info&downloadid=7813:58
xiangfuqwebirc3657, it is chinese but you can just try to download and take a look. good luck. :)13:58
qwebirc3657thanks xiangfu13:59
qwebirc3657actually, this is the exact version of the firmware i already downloaded14:00
xiangfuoh14:01
qwebirc3657and the drivers i am having trouble to install came inside this rar14:01
qwebirc3657i even get my system calling to install the driver14:02
qwebirc3657the problem is it allways finishes saying it is not installed14:03
qwebirc3657it might be because of some chinese characters in the folders tree names, which might be not recognized by my system14:04
qwebirc3657or because the drivers are not for x6414:05
qwebirc3657anyway, thank you so much14:05
wolfspra1lkristianpaul: no, not yet [sdr.osmocom] - why are you asking?14:28
kristianpaulnah, just remenbering ;)14:30
wpwrakoh dear. now, how many days of hangover will come from this ...14:55
wolfspra1lkristianpaul: ok, thanks for the reminder!14:56
wolfspra1ltime to add some new great quality feeds to the qi planet14:56
wolfspra1lAndrew's Silicon Exposed blog, OsmoSDR, Moxie, LZX Industries, Video Circuits blog, and Richard Hughes' ColorHug blog :-)14:57
wpwrakGorDo: hmm, quite a lot of measurements seem a bit off. i wonder why this is. e.g., polling 5 ms supposedly taking ~ 8.04 ms.15:06
wolfspra1lwow, I got 50 mails from the qi-planet list now ;-)15:26
wolfspra1l(oh, another 12 :-))15:27
wolfspra1lat least the new blogs definitely made it, let's see whether things calm down15:28
wolfspra1lI think we should try to manage the planet to < 3 to max 5 posts per day so that it can be followed by a human reader and doesn't just become another way to search15:28
DocScrutinizermeh15:28
wolfspra1lso I'm am always favoring feeds that have fewer and higher quality posts15:29
DocScrutinizerhi wolfspra1l, happy new year15:29
wolfspra1lyou too, happy new year115:29
wolfspra1l!15:29
wolfspra1lthis is not a pressing issue on my side, just pops into my mind: now that you are a full-time employee - are you still interested and allowed to take on paid side jobs?15:30
DocScrutinizerdang, my private vhost at hetzer/gismo/central-services acts up, RTT of minutes15:31
DocScrutinizerwolfspra1l: yes, I may, though I must inform my employer about it15:32
DocScrutinizerif they are "on a regular basis"15:32
rjeffriesthis gadget is open and interesting (to me): http://code.google.com/p/micropendous/wiki/Micropendous315:33
wolfspra1lrjeffries: Ron! :-) happy new year!15:33
rjeffriesHappy (&propserous) New Year to you wolfspraul15:33
wolfspra1lwhat do you mean with 'prosperous'?15:33
wolfspra1lI just added a number of great blogs to the Qi planet, maybe something for you in there as well, don't know15:34
DocScrutinizertrue diogenes' answer15:34
rjeffriesDocScruinizer we are all worried you may starve to death before you get your next check. are you ok? still have a pulse?15:34
wolfspra1la few cynics in the channel cannot hurt15:34
rjeffrieswolspraul even MORE prosperous than I assume you are today15:35
wolfspra1lit's pretty obvious that none of us suffer from real starvation15:35
wolfspra1lI love my life in China for teaching me the lessons of real life every day.15:35
DocScrutinizeryes, I think I'm still OK, though now got some minor problems with starting work in ~14h15:35
rjeffriesDoc however did not get his bank transfer.15:35
DocScrutinizerthanks, anyway15:35
rjeffrieshaving a steady job is A Good Thing, no?15:35
wolfspra1lif it makes him happy, yeah sure15:36
wolfspra1lso when you say 'prosperous' you mean USD?15:36
wolfspra1lI had that feeling ;-)15:36
rjeffriesthe currency of ypour chpoice RMB is a fine currency15:36
DocScrutinizerrjeffries: it's annoying especially when you have to PAY for going to work15:36
rjeffriesDoc, yes, that would be ANNOYING15:37
wolfspra1llooked up the definition, indeed it says "successful in material terms"15:37
wolfspra1lso a LOT OF HARDWARE15:37
wolfspra1lopen of course15:37
wolfspra1l:-)15:37
wolfspra1lthanks Ron!15:37
rjeffriesbut also a happy spirit, an active, curious, learning mind full of wonderment and awe15:38
DocScrutinizerFUUUUUU*15:38
wolfspra1l:-)15:39
DocScrutinizerI guess the re-registrations of DocScrutinizer51 are spamming all the channels? (I have join/nick/quit msgs hidden so can't tell for sure)15:39
wolfspra1lDocScrutinizer: hang on a second please, I will check your link in a few15:41
DocScrutinizerZNC bouncer on that vhost, sucks when character roundtrip time in ssh is ~120s15:41
wolfspra1l(sometimes you disappear fast nowadays)15:41
kristianpaulhmm seems planet also fed from previous days, thats it15:41
wolfspra1lyes, don't know why15:41
kristianpauloh cool, http://sdr.osmocom.org/trac/attachment/wiki/WikiStart/osmosdr.jpg15:42
wolfspra1lI changed the 'max posts' setting from 60 to 500, so the html page is really long now15:42
wolfspra1lthat's because I recently realized how great the endless debian planet html page is for searching in the browser15:42
DocScrutinizerindeed15:42
kristianpaulyup15:42
wolfspra1lon the other hand I feel mroe and more the traditional search engines (google, bing) let me down15:42
Action: DocScrutinizer *loves* long pages15:42
wolfspra1lsometimes I just cannot find stuff anymore15:42
wolfspra1lthe links I get are ridiculously bad/off-topic15:43
Action: DocScrutinizer also loves mtr15:45
DocScrutinizerWAY better than traceroute15:45
DocScrutinizerand ping15:45
kristianpauloh, really15:45
kristianpaullets see15:45
kristianpauloh looks automatic15:46
wolfspra1lrjeffries: micropendous looks nice in that they use kicad15:46
wolfspra1ldo they have a blog?15:47
wolfspra1ldo you know any names of key people behind the project?15:47
wolfspra1l(just asking in case you know, I am searching now...)15:48
kristianpaulwow DocScrutinizer indeed15:48
kristianpauli dint knew it, but i had to other command fu before to get this stats15:49
wolfspra1lkristianpaul: what? (sorry you lost me on this)15:51
DocScrutinizerhmmpf, >>7. gagarin.andromedalabs.de                                                       3.9%<< (packet loss)15:51
kristianpaulwolfspra1l: mtr command that DocScrutinizer loves, i dint knew about it15:52
DocScrutinizer4.3%15:53
DocScrutinizer5%15:53
kristianpaulDocScrutinizer: a bit OT for current topics, but are you aware of a gsm/umts/edge/whatever usb dongle than have populate in the PCB trace for direct serial conection?15:57
kristianpaulergh, i hope i explain my self right ;)15:57
DocScrutinizerno, sorry, never heard of any, but that doesn't mean a thing15:58
kristianpaulbecause is not actually same port you got emulated by UBS?15:59
DocScrutinizeranyway I guess all those dongles come with a one-chip solution integrating USB, so *if* there is any serial connection, then it's alternative to USB on same SoC and may or may not be supported by the modem firmware15:59
kristianpauls/UBS&USB15:59
kristianpauli see15:59
DocScrutinizerkristianpaul: query?16:05
kristianpaulhmm?16:06
wolfspra1lrjeffries: micropendous hardware license is cc-by, software seems all open/free indeed16:09
wolfspra1lthat's good :-) no -nd -nd whatever16:09
wolfspra1l-nc16:09
wolfspra1lkicad, good. professional tools are always a good sign :-)16:09
wolfspra1lthe one thing I never understand about such development boards is what is being developed16:10
wolfspra1lwhat's the next step?16:10
wolfspra1lso I am really thrilled by stuff like the ColorHug colorimeter, or Icarus bitcoin mining board (all proper and real open hardware as far as it goes nowadays)16:10
wolfspra1lbecause they zoom in on a particular use case. and whenever you do that you run into *a lot* of problems associated with that use case16:11
wolfspra1lthat's what will make your 'thing' better and better over time16:11
wolfspra1lbut a 'development board' like micropendous?16:11
wolfspra1lhow do they prioritize problems?16:11
wolfspra1lI hope you don't mind but i guess you are not planning to buy one (micropendous) or do anything with it. so you will not find out much about such use-case specific issues either...16:12
wolfspra1lbut I love the license and everything I read about it - a nice find. thanks!16:12
wolfspra1lonly that Ben NanoNote and Milkymist One are years ahead :-)16:13
wolfspra1ltoo bad, I cannot find a micropendous feed to include in the Qi planet...16:15
wolfspra1ladded to the watch list for now, if I can ever find a good feed I'll add it for sure16:21
wolfspra1lhttp://en.qi-hardware.com/wiki/Planet#Blogs_to_watch16:22
wolfspra1lthanks again, really nice find16:22
wolfspra1lI guess if someone wants a starting point for an AVR USB project, they might want to take a Micropendous as their first step...16:22
viricBon any a tothom16:42
viricbtw, my UBI fs too often says 'moving from one PEB to another PEB'16:42
viricslowing down a lot the disk access.16:43
viricmaybe I should get a newer kernel16:43
viricdoes anybody know how to get ubifs statistics?16:45
virickyak: A =>2K< 3>4><. :) I still would like to know the configuration of your utf-8 terminals with small text size16:46
viricbt16:47
viric(btw the keyboard of my nanonote has some keys that I've to press hard to be typed)16:49
DocScrutinizerthat's bad16:49
viricannoying.16:52
rjeffriesI do not understand this machine language: <DocScrutinizer> FUUUUUU*17:00
DocScrutinizerhehehe17:01
DocScrutinizera special CISC opcode set mnemonics, that invokes a WTF-IRQ servicing routine17:02
rjeffrieswolfspraul1 if you are still around, I may buy a micropendous. I also will buy a Raspberry Pi when they ship in volume (mid 2012 my guess). as to what the use case might be:17:06
rjeffriesboth are general purpose computers. Raspebrry Pi is much more powerful, has greta graphivs support, speakes HDMI to any old (modern) monitor or TV.17:07
rjeffriesRapsberry Pi is focused on teaching/enabling a new generation of young people to learn programming. how refreshing.17:07
kristianpaulgood, keep us postead about how you get upstream linux to work on it (pi)17:08
rjeffriesto my mind, Micropendous is simple a relatively easy to use microcontroller with lots of i/o. it's better adapted to so-called physical computing.17:08
rjeffrieskristianpaul you get a working Debian for Raspebrry Pi out of the box. Not sure if that qualifies as "upstream" I have never been clear on that concept. me bad.17:09
rjeffriesdocscrutinizer I suspect but can not prove that FU (U repeated) may be a mnemic for eff yew.17:11
DocScrutinizernot exactly. FU*\* is an alternative for FSCK!*1!*1117:14
DocScrutinizerso the U isn't an opcode modifier but an immanent part of the opcode itself17:16
DocScrutinizerwhile the number of repetitions of the U indicates the severity level of the WTF-IRQ17:17
DocScrutinizerthe trailing '*' is relevant to distinguish it from the FU{1,1} mnemonic17:18
wpwrak(micopendous) interesting ... didn't know USB A-A adapters existed commercially17:28
rjeffrieswhen wpwrak says "interesting> I pay attention. ;) Happy New Year, Werner.17:29
rjeffriesDoc, I suspect you know and may perchance dream in lambda calculus. Just sayin'17:30
wpwrakrjeffries: happy new year ! new year, new hangover ;-)17:37
wpwrakheh, clever. they're doing a ralley in south america, successor of that, paris-darkar thing. and they started today, january 1st, when probably everyone involved is conveniently drunk. seems it took less than an hour for the first fatal accident to happen. some things are just too predictable ...17:46
wolfspra1lrejon_: sounds good! Please keep us posted about your Pi and Micropendous findings. You will learn more about the concept of upstream soon :-)17:47
wolfspra1lnot rejon_, I meant rjeffries 17:47
kyakviric: hi, happy new year to you, too :) what did you want to know? it's all in ben-cyrillic package18:06
virickyak: ah. I'd have to find it18:06
viricwpwrak: dakar, not darkar :)18:08
virichere media considers it quite a popular race18:08
kyakprojects.qi-hardware.com/index.php/p/openwrt-packages/source/tree/master/ben-cyrillic18:09
kyakviric: you can have s look there18:09
viricok!18:09
rohnappy new year to you guys too18:09
viricI've the checkout18:09
roheh happy.. but also nappy.. both good18:09
virickyak: what is the source of all that?18:10
rohrjeffries: i would be really happy if people would stop nagging with vaporware.18:10
rohrasberry is nothing but a 'uboot' by brmc.18:10
virickyak: I mean... you wrote it from scratch?18:10
kyakviric: i took the original keymap and modified it18:11
rohsame as the panda and beagleboard are a project done by ti. just brmc doesnt want that in the big press.18:11
virickyak: you deserve at least an AUTHORS file and a licence18:11
viricor a README telling where it comes from :) otherwise it looks like a file that appeared there magically :)18:11
rohfrom my pov rasberry is NOT open hw and not a open design. so people.. please stop doing PR for a company who doesnt pay you for it.18:12
kyakviric: there is an i18n page at qi-hw wiki with explanations18:12
virickyak: I know how to use your package. I only mean that there should be an explanation of where all that comes from. What author, over what base, etc.18:14
viric:)18:14
kyakit's all git log now ;)18:14
viricI've checked it, and it starts at "adding the files for cyrillic", which is not the best indication of where all that comes from ;)18:15
viricI think your name should go to an AUTHORS file there, or a README, telling who made the files and how. You deserve that glory :)18:16
kyakok, i got your point.. i had a plan to add phonetic keymap as well, so probably i'll add a README, too18:17
viricgreat18:17
viric:)18:17
viricnot for instructions, but for the source and the release conditions of all that18:17
virichm I don't have setfont218:18
kyakit's another package you can find in the same repo18:19
viricah18:19
viricnice! :)18:20
kyakand there is also a kernel patch for setfont2..18:20
viricah really? hm18:20
kyakprojects.qi-hardware.com/index.php/p/openwrt-xburst/source/tree/master/target/linux/xburst/patches-3.0/0030-fbcon-color-fonts.patch18:22
kyakthis one18:22
viricAh you run 3.0?18:22
viricI'll go updating the kernel then18:22
kyakyep18:22
viricbecause the ubifs trouble annoys me18:22
kyakwhich trou le4?18:23
kyaks/trou le4/trouble18:23
viricah18:24
viricwell, I see from time to time, ubifs moving data from PEB to PEB18:24
viricfor example, I started today 'fbterm', and it took 3 or 4 minutes of relocating PEBs18:24
viricI don't know what is that about.18:24
rohwhats a PEB?18:24
viricI imagine it's a nand block18:25
viricat boot ubi says18:25
viric[    9.390000] UBI: number of good PEBs:        407918:25
viric[    9.400000] UBI: number of bad PEBs:         118:25
rohhm. well.. yes it should move stuff around for defragmentation etc... and balanching but it shouldnt take long18:25
viricand I imagine that starting fbterm does not imply writing things to the nand, only reading18:26
viricah, well... there is the fontconfig cache.18:26
viricI see those things quite often18:26
rohheh.. do a strace. maybe it does stupid stuff like touching lots of small files18:26
virichm in fact I don't have any directory with glyphcache files18:27
kyakthis is the first time when you start any application using fontconfig that it takes long to create the cache18:28
viricI know. It wasn't either the first time, nor I can find the cache18:28
kyakjust make sure the cache is not removed upon reboot and you are good18:28
viricwhy it should be? hm18:29
viricgrmbl. I just mplayer of the bunny file with the latest 0.1.5 jz47xx driver and it hanged18:29
kyakyou cab use fc-cache to recreate the cache18:29
viricah, I don't have fontconfig I see18:30
viricmplayer died. grmbl18:30
kyakas i recon, mplayer is also using fontconfig for subs at least :)18:31
viricnot mine :)18:31
viricah, but fbterm yes uses fontconfig18:32
viricoh /var/cache/fontconfig. I thought it would use ~/.fontconfig18:33
viricMPlayer interrupted by signal 4 in module: decode_audio18:34
viricSIGILL18:34
virica bit of gdb...18:36
kyakprobably you can have4 a look at ours config of mplayer, there aare bits related to audio18:38
kyakprojects.qi-hardware.com/index.php/p/openwrt-packages/source/tree/master/mplayer/files/mplayer.conf18:38
viricI'm using the tremor...18:38
viricbut af-adv=2 not18:38
viriccrashed again with that, at a different point18:40
rjeffriesroh to each his own. Raspberry Pi indeed uses BRCM parts. I could care less. The price is aggressive, it has the possibility of opening up the wonderful experience of programming to an entire new generation.18:40
virickyak: do you use 0.1.5 ?18:40
viricjz47xx18:40
viric(I see trunk of packages says 0.1.5)18:40
kyakyep, should be using that if it the latest18:41
viricweird.18:41
rjeffriesNot sure any of that is bad. Yeah, it does not meet the stanrdas of copyleft philosphy. But that doesn't mean it will not poetntailly hange the world18:41
viricit also crashes with vo=null though18:41
virickyak: maybe I need a newer kernel? I run 2.6.3518:42
kyakviric: probably dvdk would be interested.. he is the author of jz47xx18:42
viricok18:43
viricis he coming on irc often?18:43
kyakhm.. it's a god idea to update the kernel anyway18:43
rjeffriesroh while this channel is mainly open hardware, poeple may also be interested in clever linux computers that are very cheap. at least I have the interest. ;) how is your 2012 going so far?18:43
kyaki don't remember that we were officially running 2.6.3518:43
viric:)18:44
viricI know18:44
kyakviric: ocasionally.. but he's re4ading the ML i'm sure18:44
viricok18:44
viricis there anything new in 3.0 or 3.1 for the nanonote?18:45
kyakem.. rtc has some troubles :))18:46
viricperfect :D18:46
viricthen I still don't have a good reason to update18:46
viricah, setfont218:46
viricwhen mplayer crashed, it left my console unusable18:47
virickeyboard not answering18:47
viricany way to recover that?18:47
viric(and what's the sysreq combination for the nanonote?)18:47
viricthe usual alt-sysreq thing on PC18:48
virickyak: I have another video I prepared, and it works great18:49
viricalthough there is some flickering18:49
viricI remember telling dvdk about flickering, but we never got anything clear from that18:50
kyakwhen something kills my console, i just reboot :)18:50
viricuh18:50
viriclars_: do you know any alt-sysreq kind of combination for the nanonote?18:51
viricI'd like the 'remount readonly', or 'kill some processes', or even reboot18:51
kyakyou could try to do this via ssh18:52
viricwell, if the kernel console is set to the serial line (as it was some time ago in your openwrt, iirc), then the sysreq would be only available there.18:52
kyakthere should be a /proc/sysrq or whatever it's called18:53
viricah18:53
viricI never saw anything in /proc related to that18:53
kristianpaulusb gadget is working in 3.1?18:54
lars_viric: i think the problem is that the serial core in the jz4740 does not support the break sequence18:56
lars_or do you mean via the keyboard?19:10
viriclars_: keyboard, yes19:49
viricso, I see that when mplayer crashes, it leaves the console unusable.19:51
viricssh keeps on working fine.19:51
lars_viric: if you think it is a usfull feature we could map KEY_SYSREQ to a key20:10
lars_or key combination20:10
viricaha20:12
lars_fn+f8 or something20:12
viricYes, I'd like that20:12
viricand then, fn+h, fn+b, fn+r, ...20:12
viricisn't it always nice to remount readonly the fs instead of a hard reboot?20:13
lars_yes, i guess20:13
viricDon't you do that with your PC linuxes? :)20:13
viric(Am I alone using sysreq?)20:14
lars_my linux never locks up ;)20:14
viricaaah.then I just have bad luck :)20:14
lars_sometimes it get kernel panics20:15
lars_uptime is usually more than a month20:15
viricah :)20:15
viricwin95 quality :)20:16
lars_reminds me i wanted to upgrade my kernel20:16
lars_to 3.1.x20:16
viricthat sounds full of hope :)20:17
viriclars, maybe you know... when mplayer crashes playing a video using dvdk driver, how can I recover the console?20:17
viricI only have the last picture on screen, and it looks like even not reading the keyb.20:18
lars_no idea20:18
viricok20:18
viricAren't your kernel logs full of this?20:26
viric[ 7040.830000] UBI: scrubbed PEB 2571 (LEB 0:1828), data moved to PEB 173220:26
viric[ 7152.620000] UBI: scrubbed PEB 3112 (LEB 0:2312), data moved to PEB 402620:26
viric....20:26
rohrjeffries: i dont have a problem with that. what i dont like is people promoting something closed and nonfree as open and free.20:27
lars_viric: thats nothing to wory about. ubi is a bit verbose on this20:28
lars_(i think)20:28
viricah20:29
rohi mean.. when i get a completely foss driverstack and maybe some register level documentation about the brmc part i would change my position on it. but for now its advertising fraud to me.20:29
viricwell, today I waited 2 or 3 minutes for fbterm to start20:29
viricand the only think I could see on the screen since typing "fbterm[enter]" was three or four scrubs like those.20:29
lars_i think ubi moves pages if it detects a (correctable) bitflip20:31
lars_so the bitflip also gets corrected on the physical storage20:31
lars_but i'm not quite sure, so i could be wrong20:31
lars_" * This sub-system is also responsible for scrubbing. If a bit-flip is detected20:32
lars_ * in a physical eraseblock, it has to be moved. Technically this is the same20:32
lars_ * as moving it for wear-leveling reasons.20:32
lars_quote from the ubifs source20:32
lars_"20:32
viricah20:34
lars_ok, just read some more and what i said above is correct.20:35
lars_if it detects a correctable (due to ecc) bitflip it will move the block in order to correct the bitflip in the physical storage to reduce the risc of data corruption20:35
viricso, some block reads cause block writes20:36
rjeffriesroh I understand. It will be intresting to see if/when BRCM opens up some of the docs. since Raspberry Pi usesd an old part, thye may well do so.20:36
viricthat should not have slown down the fbterm start though20:36
lars_imo it should not report these in the kernel log, because the kind of storage we use tend to have bitflips from time to time20:36
viricah20:36
viricthank you a lot for the explanation :)20:37
lars_no problem :)20:42
--- Mon Jan 2 201200:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!