#qi-hardware IRC log for Sunday, 2012-01-08

wolfspraulkristianpaul: that's a nice link about hiqsdr.org - thanks!00:34
wolfspraulunfortunately I couldn't find an RSS feed to add to the planet right now, but I will definitely watch this project, from the distance00:34
wolfspraulthey seem to be very serious about high-quality information, bom, etc.00:34
wolfspraulsources are GPL licensed, but overall the licensing and tools is messy, too bad00:35
wolfspraulbut I couldn't find anything like -nc -nd or so, it's all just open and a bit messy00:35
viricso I imagine noone runs 2.6.37 here11:46
viricDoes 3.0 suspend fine?11:46
viricah it just suspended fine11:47
viricmaybe I can't reproduce the crash again11:47
viricI could suspend twice, and then I tried to play a video and crashed11:49
viricwhere is dvdk?11:51
viric1 week...12:33
kyakon vacation, probably :)12:33
viricah, I remember as if rarely seeing him here these last months12:34
viricI've read his code... nice12:34
viricof course, I can't tell why it hangs :)12:34
virickyak: can you try suspend with the usb cable connected?12:36
virichm it somehow works.12:37
viricI think that when the nanonote is 'on' by more than 1h (let's say), then suspend doesn't work anymore12:37
kyaki don't suspend very often, but when i do, it usually works12:38
virichm ok12:39
kyaki remember it crashed several time though12:39
viricyou usually halt?12:39
viricin fact it boots fast. 2.6.37 gives a login prompt in 5 seconds from poweron12:39
virica big improvement over 2.6.36, iirc12:40
virickyak: if the virtual consoles on frame buffer are totally implemented by software in the nanonote, I don't think there should be a memory limit for glyphs12:43
kyakyeah, i usually halt or not switch off at all12:44
kyaki'm not sure where this limit comes from..12:45
viricyou even are not sure of the limit12:45
virickyak: how do you type a ` with your keyb?12:49
viricshift-comma types "12:49
mstevensfuloong progress - I can see the screen now16:25
viricwhat fuloong?16:33
viricmstevens: my fuloong just got a black screen, when I booted for the first time linux 3.216:37
viricmstevens: are you dealing with that16:37
mstevensviric: I have just purchased a new fuloong, and am trying to install debian16:37
mstevensat first I couldn't see the installer16:38
mstevensbut then I switched from DVI to a vga cable, and now I can at least see the installer16:38
viricah, deep debugging16:38
viricmstevens: I know little the fuloong, but I ported NixOS to it..16:39
mstevensviric: someone on #fsf reckoned the DVI out doesn't work right until X starts16:39
mstevensviric: my current problem is that the debian wheezy installer doesn't have the right kernel modules16:40
mstevensno driver for the network card16:40
viricto achieve what?16:40
viricah ok16:40
mstevensMy current test is a combination of the "openrays" kernel the machine ships with, and the debian wheezy initrd16:44
mstevensbah, boot that kernel and it just goes straight into rays16:47
mstevensI think maybe the root device is hardcoded and it's ignoring my initrd16:50
viricthat'd be very strange17:06
viricI used tftp to boot what I wanted17:06
kristianpaulmstevens: i ran debian wheezy (with X patched) and no problem on my yeeloong17:14
mstevenskristianpaul: I can't get the installer to see the fuloong network card17:15
mstevensIt seems to be a rtl 8169, which the netinst images don't seem to have drivers for17:16
kristianpaulwich yeeloong revision is yours?17:17
mstevensI have a fuloong, not a yeeloong17:18
kristianpaulah !17:18
mstevenswhatever they sell if you ordered one 2 weeks ago17:18
kristianpaulnope mine is kinda old..17:18
kristianpauland is yeeloong17:18
mstevensI am attempting to extract the requisite driver from the full kernel package and stuff it into the initrd17:20
viricI'll check mine17:20
viric[    3.200000] r8169 Gigabit Ethernet driver 2.3LK-NAPI loaded17:20
viric[    3.208000] r8169 0000:00:06.0: eth0: RTL8169sc/8110sc at 0x9000000050072000, 00:23:9e:00:0f:c6, XID 98000000 IRQ 3617:20
viricthats in-kernel17:21
mstevensviric: yeah I don't have that17:21
viricdo you want my kernel?17:21
mstevenslet me try this17:24
viricmstevens: I just wrote to loongson-dev, about me failing to boot a linux 3.217:24
viricthe nanonote is still faster than qemu in a home computer, no?17:38
viric(for mips32 code)17:39
mstevensviric: I just made progress by grabbing r8169.ko out of a full kernel package and stuffing it into an initrd17:39
viricmstevens: great :)17:39
viricwhat kernel version do you boot?17:39
mstevensI think 3.1.017:39
mstevensbut I'm not certain17:39
mstevensit's in mid install17:39
viricah ok17:39
viricalt-f2, uname -a ?17:39
viricah nice17:40
viricI should get one of those.17:40
viricI'm measuring the nanonote battery bad.... it says -10% now17:47
mstevensviric: I should update my nanonote17:47
mstevensdebian is installing the base system17:48
viricto run nanonixos. everyone should :)17:48
mstevensfingers crossed this will work at the end of it17:48
mstevensviric: I bought a nanonote when they first came out, played for a while, concluded they didn't do anything useful, and put it in a box17:49
mstevensI should have another go17:50
viricI can't say I use mine much17:50
viricsometimes when my wife drives, I read rss on it17:50
viricmstevens: will you use your fuloong with a screen?17:53
mstevensviric: this is the plan17:53
viricI don't17:54
viricdo you have a fan in it?17:54
mstevensthe goal for me is a desktop for generalised hacking17:54
mstevensviric: it sounds like it!17:54
viricI hate that fan17:54
viricI'd have used them much more, if they hadn't a fan17:54
mstevensviric: it is not as bad as people suggested on irc! but it is noticeable17:54
viricmaybe it gets worse on age17:55
mstevensI am wondering if they made new ones quieter17:55
viricthey should have a big fan, if they wanted a fan at all. big fans are much quieter17:55
mstevensviric: I found buying the fuloong from china it was cheap enough to buy it as an experiment17:55
mstevensif it doesn't work out I won't be totally sad17:55
mstevensand I like to support "free" hardware17:56
viricI received mine as a donation to port nixos to it :)17:56
mstevenseven cheaper!17:56
mstevenswhat is nixos?17:56
mstevensI guess I can google17:56
viricthe only worth GNU/Linux distribution.17:56
mstevenshmm it does not boot, that is not so good18:01
viricwhat does not boot?18:05
mstevensthe freshly installed debian18:07
viricbad debian bad18:07
mstevensI am experimenting. I think it may not be setting up right for pmon18:08
viricI forgot all I knew about pmon18:09
Action: mstevens runs off for a bit18:11
qi-bot[commit] Paul Cercueil: fbcon: prevent possible buffer overflow. (jz-3.1) http://qi-hw.com/p/qi-kernel/d3d327419:02
qi-bot[commit] Paul Cercueil: OpenDingux: initrd: updated mininit to latest version. (jz-3.1) http://qi-hw.com/p/qi-kernel/76a1add19:02
qi-bot[commit] Paul Cercueil: MIPS: JZ4740: SLCD: added a parameter to select the panel. (jz-3.1) http://qi-hw.com/p/qi-kernel/17a69f119:02
qi-bot[commit] Paul Cercueil: OpenDingux: added OpenDingux logo. (jz-3.1) http://qi-hw.com/p/qi-kernel/8845bf519:02
qi-bot[commit] Paul Cercueil: MIPS: JZ4740: SLCD: show linux logo at init. (jz-3.1) http://qi-hw.com/p/qi-kernel/959cebc19:02
qi-bot[commit] Paul Cercueil: fbcon: read parameters as unsigned values. (jz-3.1) http://qi-hw.com/p/qi-kernel/694a33819:02
qi-bot[commit] Paul Cercueil: MIPS: A320: use the new OpenDingux logo. (jz-3.1) http://qi-hw.com/p/qi-kernel/e3487b119:02
qi-bot[commit] Paul Cercueil: fbcon: optimize param read loop with 'continue' (jz-3.1) http://qi-hw.com/p/qi-kernel/ae92f2419:02
qi-bot[commit] Paul Cercueil: fbcon: add parameter to toggle bind on startup. (jz-3.1) http://qi-hw.com/p/qi-kernel/cf5635919:02
qi-bot[commit] Paul Cercueil: MIPS: A320: hide the fb console at startup. (jz-3.1) http://qi-hw.com/p/qi-kernel/1f176f019:02
qi-bot[commit] Paul Cercueil: fbcon: use kstrtouint to respect initial behaviour. (jz-3.2) http://qi-hw.com/p/qi-kernel/61e3c3019:02
qi-bot[commit] Paul Cercueil: fbcon: prevent possible buffer overflow. (jz-3.2) http://qi-hw.com/p/qi-kernel/828e6f619:02
Aylait should have pushed only the two commits for jz-3.2 x_x19:03
mthgit will push every branch that has a remote equivalent19:03
Aylaokay, so what should I do now? ...19:04
mthyou can set the ref head to an earlier commit for jz-3.119:04
mthI don't know the syntax by heart though19:04
larscgit push -f qi jz-3.1~4:jz-3.119:08
larscwhere the 4 is the number of commits you want to go back19:08
Aylaok, I type that?19:08
larscyou type that19:08
larscor you could pass a specific revision instead of the jz-3.1~X19:09
Action: Ayla crosses fingers19:10
larscwhich is  4adea3bca06ca in our case19:10
larscone of your commits is still there19:13
Aylathat one was pushed in november19:13
viricI thought only larsc touched the kernel19:16
Aylathank you for your help, and sorry about that19:16
mstevens /me returns19:17
larscviric: i've been a bit quiet in that regard in the last months19:17
viricah ok19:17
viricso me.19:17
larscAyla and mth are working on the linux port for the dingoo which is a handheld console which uses the same SoC as the Nanonote19:18
mstevenscut out the grub stage, seems promising19:18
larschence we share the same kernel repo19:18
mstevensthere's more hard drive activity, anyway19:20
viricmstevens: you use grub?19:20
viriclarsc: ah ok19:20
mstevensviric: the wheezy installer puts it in as part of the boot process19:21
viricmstevens: ah ok19:21
viricI've never seen grub on the fuloong19:21
mstevensviric: I've told pmon to boot the kernel directly and it seems to be more successful. Although hasn't actually booted yet, so I can't be sure.19:21
viriclarsc: do you know of any suspend problems?19:21
virichm ok19:21
viricSometimes it crashes here19:21
viricand the screen with the console keeps lighted, as fading to white19:21
viricbut it does not reach full white.19:22
larscthats normal if the lcd hasn't received a powerdown command but the SoC isn't sending any pixel data anymore19:22
viricI've seen that three times today, and one time yesterday19:23
viric(on suspend)19:23
viric(on 2.6.37)19:23
larsccan you resume normally in such a case, or has the whole system crashed?19:25
viricin one case I could resume19:25
viricin the other, crashed19:25
viricI mean19:25
viric4 times crashed, one could resume19:25
viricha. right now I suspended (screen black), resumed, and crased.19:26
viricthe cursor blinks, but no key does anything19:26
mstevenshmm, constant drive activity, no boot19:26
viriclarsc:after resume, the console managed to write 'g_ether gadget: high speed config #1: CDC Ethernet (ECM)19:27
viriclarsc: I'm simply running "echo mem > /sys/power/state"19:27
Aylaviric: on Nanonote?19:28
AylaI've been working on the suspend feature19:28
viricAyla: 2.6.37 though19:28
Aylamy changes are on the Qi kernel, but you should upgrade to a more recent kernel19:29
viricI run vanilla + openwrt patches19:29
mstevensseems to be having problems loading the initrd image19:34
viricAyla: so, what's the best to run on the nanonote now?19:35
AylaI don't know, I don't have a nanonote :D19:36
viricumh :)19:38
viricAyla: is there cpuidle for the jz47?19:53
Aylawhat do you mean by 'cpuidle'19:54
virichm linux cpuidle19:56
viricgoing to low power active states 19:57
larscyes. we'll put the cpu into sleep state if it is idle19:57
viricsleep is that of suspend?19:57
viricI see: normal, doze, idle, sleep modes19:57
AylaI believe it's not supported currently19:58
larscthe jz4740 has a special sleep instruction which will halt the cpu until the next interrupt19:58
larscwe use that when the cpu is idle19:58
virichm ok19:59
Aylaah, that's why it consumes much less power when idle19:59
virichm the sleep instruction has different meanings depending on the state of some registers20:01
viricI think larsc means the manual 'idle mode', while suspend would be the manual 'sleep mode'20:03
larscyes. it either suspends the cpu20:03
larscor the whole SoC20:03
viricthere is also the 'doze mode', not used now, right?20:04
viricI don't see what advantadges can bring the doze mode...20:06
Aylasomething great would be a working ondemand CPU governor20:07
viricI think the doze mode is about ignoring some clocks...20:07
viricso doze mode could be used to implement some cpufreq20:09
virickind of.20:09
larscdozemode looks like a extra divider for the cpu clock to me20:10
viriclarsc: that's what I menat20:10
viricisn't that the closest to cpufreq?20:10
viricbtw, how do you test your kernel modifications?20:11
viricdo you use xbboot?20:11
larscflash it and boot, don't know what you want to hear20:11
viricxbboot can "upload file at <path> to <address> then jump to <address>"20:12
viricAyla: and you? Flash and boot?20:13
AylaI boot from USB when I need to test20:14
viricdo you use KGDB or something like that through the serial line?20:17
viricnever tried?20:18
AylaI don't really need KGBD20:19
viricI like to do those kinds of things...20:19
viricAyla: no? hm20:19
viriclarsc: perf for the nanonote does not build, btw20:28
larscany idea why?20:29
viric http://sprunge.us/fNZR20:29
viric^ larsc 20:30
viriclarsc: are you aware the default keyboard in 2.6.37 writes " instead of ` ?20:31
virichm wait. I don't remember if I use the default keyboard.20:31
qi-bot[commit] Wolfgang Spraul: added small bash script to test kicad cmdline patches (master) http://qi-hw.com/p/eda-tools/4d628c321:01
mstevensprogress, i can boot into X21:01
virichm what is the serial port major,minor in the nanonote?21:06
viricwhat I am doing wrong?   'stty -F /dev/ttyS0 raw 57600'21:11
viricI run 'cat ttyS0&'21:11
viricand 'echo hola > ttyS0'21:12
viriccrossing the tx/rx in the serial.21:12
viricI try both S0 and S121:15
viricS1 even blocks on 'echo hola > ttyS1' now...21:17
viricI thought that'd be trivial21:19
viricanyone alive, to tell me the major,minor of the serial port of the backwards of the nanonote?21:23
wolfspraulwpwrak: I'm looking at your bug report saying "pcbnew --plot doesn't pickup global clearance settings"21:27
viricok, I checked that it's ttyS0... the nanonote sends fine, but does not receive.21:28
viricARGH I've a mingetty at ttyS021:29
viricwhich... looks irrelelant21:31
wpwrakwolfspraul: hmm, now if i would remember it ... presumably worked around it, then swiftly reused the precious memory space :)21:35
viricnothing. I fail at the serial port.21:36
viricwpwrak: do you have a serial port connection for the nanonote? :)21:36
viricwpwrak: crossing tx and rx should allow me to get echo, no?21:36
wpwrakviric: echo from where to where ?21:37
wpwrakviric: i use idbg for the serial port21:37
viricI'm running: stty -F /dev/ttyS0 raw 5760021:38
viricthen:   while true; do echo hola; done > /dev/ttyS0 &21:38
viricthen: cat /dev/ttyS0 &21:38
viricand cross the tx and rx pins21:38
viricI get nothing21:38
viricI've checked with a ftdi cable, and the nanonote sends well. But looks like not receiving anything21:39
wpwrakhmm, which set of TX/RX contacts are you using ? the ones near the battery or the ones at the front of the PCB ? the RX near the battery is reassigned to the keyboard, so you'd have to reassign it for UART use21:40
viricat the bottom of the battery21:40
viriceither serial or keyboard?21:41
virichow? an echo to some gpio thing?21:41
wpwrakRX is shared21:41
viricwpwrak: what page of the manual says that? :)21:41
virichow can I 'reassign' it?21:43
wpwrakpage 5 of the schematics ? :) pin 72 (KEYIN8) right below TP7521:43
viric'reassign' means a software operation, no21:44
wpwrakor page 20 of jz4720_ds :)21:44
wpwrak(reassign) yes21:44
viricwhat one?21:44
wpwrakthe reassigning would be page 356 of jz4740_pm, PDFUN21:45
wpwrakyou basically have to write a 1 << 26 to PDFUNC, then you may be able to RX21:46
wpwrak(not sure if there's anything else in the way)21:46
viricCan I do that with 'dd' and /dev/mem? :)21:46
wpwrakan easier approach would be to run a wire to the other set of TX/RX pins21:46
wpwrakwith "poke"21:47
viricI don't have the pm!21:48
viricwhere is that ?21:48
wpwrakfor locating TPs, this map is your friend: http://downloads.qi-hardware.com/people/werner/tmp/tp-map.png21:48
viricI want to go the 'poke' way21:48
wpwraklemme find the complete version ...21:49
viricI have the jz4725pm :)21:49
wpwrakheh, i think i lost that one. but anyway, the registers are probably the same21:52
viricPDFUN is at 0x1001034021:52
viricthat's >32MB21:52
viric>256MiB :)21:52
viric/dev/mem does not let me do that21:52
wpwrakit works with "poke" :)21:53
viricwhat is that poke?21:53
wolfspraulwpwrak: no worries, I have it in my notes, just a heads up21:54
viricok, I stop for today. I can't advance further...21:56
wpwrakcompile with the appropriate CC=...21:58
wpwrakit's quite universal ;-)21:58
viricI'll compilie on the nanonote21:59
viricwpwrak: sure about '1<<26'?22:01
viricpoke will 'peek', if without value?22:01
viricgpio d, 26. ok. 22:02
wpwrakno ... you have to calculate the shift22:03
wpwrakand yes, it will peek if without value22:03
viricI've seen 26 http://en.qi-hardware.com/wiki/Hardware_basics#GPIO_pins22:03
viric1 << 26 looks right22:04
viric# ./poke -32 0x1001034022:05
viric2^26 = 400000022:05
viricso... 0x2e22:05
viric# ./poke -32 10010340 2e003f0022:06
viricusage: ./poke [-8|-16|-32] hex_address [value]22:06
viricah 0x22:06
wpwrakyes, 0x :)22:06
viricI read back, and it says the same as before22:06
viricmaybe in the jz4740 is not in the same place..22:07
wpwrakdid the command run ? or did you only get the usage ?22:07
viricif I write 0x2e003f00, it runs22:07
wpwrakwriting to PDFUN has no effect. you have to write to PDFUNS (set) or PDFUNC (clear)22:08
wpwrakin your case, PDFUNS22:08
viricand with 0x400000022:09
viricWORKS! rx works now22:09
wpwrakwhee ! :)22:09
viricthe wiki page about the serial port could have at least one word about this ;22:10
wpwrakisn't it obvious ? ;-)22:10
kristianpauli did the TP mod for... :-|22:10
wpwrakviric: see, kristianpaul agrees with me ;-)22:10
kristianpaulnope, well argh, this shared wasnt with a keyboard key?22:11
kristianpauli got confused now..22:11
kristianpaulRX shared*22:11
viricI want that for debugging...22:12
viricnot for running serial devices with the nanonote keyboard22:12
kristianpaulahh !22:12
Action: kristianpaul bad memory22:13
viriclots of magic22:14
viricundocummented where it should be.22:14
viricwpwrak: not obvious not obvious :)22:14
kristianpaulYES ;-)22:14
viricwpwrak: why does any wiki document tell anything about soldering the RX pin at all? :)22:15
wpwrakdo they or don't they ?22:16
viricwho they?22:16
wpwrakthe wiki is basically write-only. there's no editorial oversight and no means to organize things for yourself. that's why it's a mess.22:16
virichttp://en.qi-hardware.com/wiki/Serial_console this tells to solder the RX22:16
viricwpwrak: I agree the PM is much better ;)22:16
wpwrak(they = documents)22:17
wpwraknaw, we should have a wiki based on a revision control system. that would make it easier to maintain at least properly organized subtrees.22:17
viricwpwrak: you blame the software? haha :)22:18
viricpeople can write equally messy over a vcs :)22:18
kristianpaulon its own tree/branch i understand?22:19
kristianpauloh well..22:19
viricwpwrak: I had by chance the jz4725pm because that's mentioned in a comment of the dvdk IPU code I was reading today...22:19
viricthere I go with my vcs wiki22:23
viricI wrote that to the qi wiki as well, on 'serial console'22:26
virickristianpaul: what do you do with the nanonote serial port?22:39
viricprogram arduinos like david? :)22:43
viriclong ago22:44
DocScrutinizerwow, that's cute23:17
kristianpaulviric: no23:47
kristianpaulviric: connect to gps receiver and posibly to a jeenode (not yet do)23:47
kristianpaulyou edited the wiki, great !23:51
wolfspraulour under-loved wiki23:51
--- Mon Jan 9 201200:00

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