#milkymist IRC log for Wednesday, 2011-07-27

wpwrakdrums, please !01:02
kristianpaullinux.. :-), i have a osgps running for mips, well why not linux02:08
kristianpauluclinux**02:09
aw_rc3: just captured R157, factory didn't 2 pcs while I measured already 17pcs. ;-) so this makes me remembered that my first 2 / 10 pcs got d2/d3 dimly lit after reflashed. ;-) I'll go back to check them. :)02:41
aw_so far now just measured impedances after 20pcs, I didn't see "short" conditions though, this also makes me felt that I must make sure to get boards fully dried since I put them overnight after cleaned.02:59
rohi have 225buttons of 240 done. need to wait for them to harden finally. also need to make some new ones (lost 2 pieces somewhere and mis-glued some)03:02
rohbut i still have material. will get the rest done later (after sleeping)03:03
aw_roh, sleep firstly then brain get clear. My brains always get stock when work late. ;-)03:07
aw_roh, so 3 * 80pcs = 240pcs, no more spares?03:07
kristianpaul+1 to spares, i also need some ;^)03:09
rohyes. i can do spares. (need to laser replacements from the leftovers anyways) .. how many should i make? like 5 more kits of 3 buttons?04:14
larscI don't think it's correct either, but it works. And there is currently a lot of restruturing done in the upstream memblock with the final goal to get rid of bootmem, so I think it's best to wait until that is done, before cleaning up our code04:15
larscmwalle: ^04:15
rohkristianpaul: do you get a shipment from taiwan? i can put some without the button cap glued on in there for you as well.04:15
rohor you could make them yourself.. its simply a 0.5mm thick disc with 7.9mm diameter as spacer and a 1mm thick disc of 12mm diameter behind it. (so its cap, small and thin disc, big and thicker disc from front to pcb)04:17
rohi use http://www.modulor.de/shop/oxid.php/sid/46adc07dbce552c2c94a88da5571b9e5/cl/alist/cnid/NAH/tpl/-/lang/1 to 'glue'04:18
rohbeware. its evil.04:18
aw_roh, i didn't send any spares for kristianpaul . ;-)04:18
aw_wolfspraul, do you think that how many spares of roh's sets of 3 buttons we need? since surely we haven't known how many boards in the end they will have?04:20
wolfsprauldon't understand04:20
wolfspraulwhat?04:20
wolfspraul:-)04:20
wolfspraulspares, roh, kristianpaul, Adam, Taiwan?04:20
aw_Taiwan here for Adam. ;-)04:21
wolfspraulI ordered 80 full sets of cases, to be shipped to Adam in Taipei. That's step 1 :-) Beyond that I don't know anything.04:21
wolfspraulroh: how are things going?04:22
aw_we mounted 90 pcs for rc3 boards, other accessories stocks here http://en.qi-hardware.com/wiki/Milkymist_One_accessories#available_stocks04:22
rohwolfspraul: got 75*3 buttons done (final qc not yet, but glued)04:24
wolfspraulalmost! cool! 5 to go :-)04:24
rohfound that i am missing 2 buttons, rest follows after sleeping etc04:25
wolfspraulgood, get a rest. Adam will have interesting test results later today...04:25
rohwill laser some more as spares from the leftovers i have. so i will simply do some more so you have spares if some dont want to fit04:25
wolfspraulnice04:25
rohi hope i can complete the rest later/tomorrow (the glued parts should rest 24hours ive read) and if everything goes as planned i should be able to ship the stuff.04:31
rohquestion still is: which carrier ;)04:32
wpwrakavian ? :)04:37
wolfspraulroh: what about the logo?04:43
rohah. yes . that needs to be done too. *nih*04:43
wolfspraulfor shipping, let's see04:43
wolfspraulI guess our options are fedex, dhl/deutsche post, hand-courier by zecke :-)04:44
wolfspraullet's finish the stuff first04:44
wolfspraulactually zecke could take them and we save some money, but I'm a little worried about the VAT return/export thing. He would have to walk to some customs counter at the airport and have a few documents stamped or so. Maybe he even gets the VAT cash back there? I never did this process...04:45
wolfspraulotherwise we just pick deutsche post or fedex or so. no worries.04:46
wolfspraullast time it worked pretty well04:46
rohok05:17
awstill 35pcs need to be measured impedances. phew~05:38
GitHub111[linux-milkymist] larsclausen pushed 1 new commit to master: https://github.com/milkymist/linux-milkymist/commit/ae1dcb3a00d95b87a092689f1b29a05bbfb44be205:46
GitHub111[linux-milkymist/master] milkymist-gpio: Fix copy&paste error - Lars-Peter Clausen05:46
awrc3 impedances measurements done: http://en.qi-hardware.com/wiki/Milkymist_One_run_3_schedule#Test_Results08:04
aw8 / 90 has short conditions: haven't investigated yet. next steps to voltage/current test @ 0.7A limited from lab power supply08:06
wolfspraulaw: interesting08:13
wolfspraulfor example for 0x7B, you have an 'X' but then you corrected R157, right? did it then pass, or still fail?08:14
awwolfspraul, let's move forward firstly, finally PASS. when I marked 'X' means it happened then.08:15
awwolfspraul, 1. R157-4.7K ohm, replaced 10K done. ;-)08:16
wolfspraulaw: but the test still fail after replacing R157?08:17
wolfspraulI noticed 0x7B and 0x87 are both like this08:17
awwolfspraul, NO, Pass after replaced a corrected R157-10K done.08:17
aw0x7B & 0x87 are done. ;-) sorry that I should have noted well. :)08:18
awnotes are revised.08:20
wolfspraulok08:21
wolfspraulaw: let's see how it continues :-)08:21
GitHub117[extras-m1] yizhangsh pushed 1 new commit to master: https://github.com/milkymist/extras-m1/commit/96946ab8a0402630ba9862c2190af5ea69dfc96909:17
GitHub117[extras-m1/master] removed redundant file - Yi Zhang09:17
GitHub167[milkymist] sbourdeauducq pushed 1 new commit to master: https://github.com/milkymist/milkymist/commit/72a4243c9d3ddf5396a2d345ebbbfcf5f840fe9909:57
GitHub167[milkymist/master] demo: cleanup output - Sebastien Bourdeauducq09:57
awvoltage/current results updated but not finished all, check it out.10:35
lekernelwhat's wrong with 31 and 3A reflash?10:56
lekernelthat's a new problem10:56
lekerneland 5510:56
wolfspraulyou mean 0x32 and 0x3A? no they are all old11:04
wolfspraulactually it starts to look much better now, let's see how things continue11:04
wolfspraulmaybe it will be a pretty good run in the end, after all :-)11:04
wolfspraulcertainly the results that came in today, so far, were good11:05
kristianpaulroh: sue11:07
kristianpaulsure*11:07
kristianpauli mean with +1 that having spares is good,11:08
kristianpaulmeant*11:08
kristianpaulwolfspraul: i just tought if roh ship you extra spare parts, adam can ship me later sme screws buttons.. etc.. of course by the cheapest mail i pay the order11:12
kristianpaulwolfspraul: if you agree of course11:12
wolfspraullekernel: I'm trying to figure out what exactly makes up the anti-aliasing circuit we can remove when switching to ADV7181C. Is it L14/15/16 - that part?11:13
kristianpauli dont  need buttons btw :)11:13
wolfspraulkristianpaul: why should I not agree :-) roh can put some into a letter too if it's easy, or Adam can, sure. Just need some spares first somewhere...11:14
kristianpaulroh will be nice11:14
kristianpaulletter*11:15
kristianpauli just need screews and the rubber thing in the botton, too keep air flow okay11:15
wpwrakkristianpaul: can't you find all that in the next DIY market/ferreteria ? :)11:22
kristianpauli'm worried about margins11:22
wpwrakkristianpaul: certainly the "rubber thing" if you mean "feet"11:22
kristianpaulyes11:22
wolfspraulwpwrak: did I identify the anti-alias circuit correctly? I'm having some fun reading schematics but I don't want to be responsible for ripping out life-supporting organs :-)11:23
wolfspraulI think it's L14/15/16? but can the varistors go as well? or was that the surge protection we added? and how about C202-C207 - can they go?11:24
wolfspraul:-)11:24
wpwrakwolfspraul: dunno. this is the first time i even hear of an anti-alias circuit. so far, i think that was some software, maybe hardware ;-)11:24
wolfspraulI was just having fun playing EE11:25
wolfspraulI think I am ready soon, he he11:25
wolfspraulincompetence galore, where is my soldering iron?11:25
wolfspraulAdam's test results so far look good, gives me a second wind11:26
wpwrakah .. aliasing = ghosting maybe ?11:27
wpwraki.e., reflection on the video signal line ?11:27
wpwrakwhat chip was the codec before ?11:28
wolfspraulit's really called antialiasing circuit, see p97 http://www.analog.com/static/imported-files/data_sheets/ADV7181B.pdf11:28
wolfspraulthat is not needed any longer with the ADV7181C (it has been integrated, I guess)11:28
wolfspraulin the m1 schematics, on p12, I think it must be something connected to AIN4-611:29
wolfspraulso I figure - rip out L14-16, and maybe it's the right thing?11:29
wolfspraulthat's the block that is in the way, he he11:30
wpwrakhmm. fig. 44 is off. no input, no output ? :)11:35
wolfspraullook on the next page11:35
wpwrakat least it connects to two power rails ;)11:36
wolfspraulthey say this should be plugged into any incoming line11:36
wpwrakyes, fig. 45 shows the black box as if it had inputs and outputs, then points to fig. 44. that doesn't really help ;-)11:36
wpwrakyou're probably supposed to connect between C93 and R38, and between L10/C102 and R6311:37
wpwraknow, what makes you think you can remove things ? it looks as if you should _add_ quite a lot of filter circuit ?11:38
wpwrak(alias) ah, now i understand the terminology. it's now impedance matching but HF components (from wherever) riding on the input11:39
wpwrakhmm. the whole circuit looks a little fishy. a bead can also remove HF, that's true. what bead are you using there ?11:42
wpwrakthe DC blocking caps seem under-dimensioned. fig. 44 has them 1000 times as large11:42
wolfsprauljust confirmed - we added the varistors V5-V7 as some sort of protection in rc3, so that's definitely not part of it11:43
wpwrakvaristors are just a protection, plus some parasitic capacitance11:43
wolfspraulso it must be L14-16, and maybe C202-C207. my guess.11:43
wpwrakso is there a specific problem with the video at the moment ? or are you just poking around ?11:44
GitHub188[uclibc-lm32] larsclausen pushed 2 new commits to master: https://github.com/milkymist/uclibc-lm32/compare/71624d5...fd3537311:44
GitHub188[uclibc-lm32/master] Fix dup3 - Lars-Peter Clausen11:44
GitHub188[uclibc-lm32/master] lm32: Sync kernel_types.h - Lars-Peter Clausen11:44
wolfspraulno problem. we are switching to ADV7181C which doesn't need the antialiasing filter anymore. I try to figure out what exactly needs to be removed.11:45
wolfspraulthat's for rc411:45
wpwrakaaah !11:46
wolfspraulif I'm even looking at the right place in the schematics, it must be L14-L16. but I may be totally wrong...11:46
wpwrakyes, the circuit looks a LOT more like what ADV7181C wants ;-)11:47
wpwrakand L14-16 do indeed seem redundant. not sure if they actually hurt. they could even prevent some RF creepage. C202 through C207 also seem unnecessary. but maybe make them NC, just in case.11:49
wpwrakwouldn11:49
wolfspraulB works well with what we have today11:50
wpwrak't it be good to connect AIN1 through AIN3 to test points ? in case someone wants try and add something new11:50
wpwrakthe B data sheet suggests that it depends on the nastiness of the source. and manifests itself in noise. maybe you've only tried "nice" sources. or sources that are already noisy. or you just didn't look hard enough for the noise ;-)11:51
wpwrakhave you tried to connect a PC with a clean VGA signal and looked at what the M1 makes of that signal ?11:52
wolfspraulsure, all possible. no test point proliferation... I am just trying to understand what to remove exactly :-)11:52
wpwrak(tp) okay. just struck me as one of those places where a TP could be nice to have. well, it's a component with leads, so you can always improvise later. not like the USB host in the ben ;-)11:53
wpwrakbut yes, it looks as if the beads could probably go. at least half of the 120 pF caps, too. (beads plus caps form a somewhat odd Pi filter. do you know if that filter was put there specifically for the aliasing problem ?12:01
lekernelantialiasing circuits are low-pass filters put before an ADC to block frequencies higher than half the sample rate (as they would cause aliasing, hence the name)12:10
lekernelthere are also optical antialiasing materials to put before CCD sensors, etc.12:11
lekernelin our case it's the "pi" circuit made of the 120pF caps and ferrite beads12:11
lekernelthere's one per channel12:11
wpwraklekernel: so far. so good. now, if the "pi" intended as a replacement for the active filter suggested for the ADV7181B ?12:15
lekernelyes, I copied this from an altera board12:16
wpwrakaah, i see12:16
lekernelthat was fast to make, and I like fast things.12:16
wpwrakso wolfgang would be right and the ADV7181C shouldn't need that pi, agreed ?12:16
wpwrak(fast) ;-)12:16
lekernelyes12:17
lekerneljust DNP the caps and use a 0 ohm resistor instead of the bead12:17
wpwrak(fast) makes me wonder what kind of vehicle you drive :) or will drive once M1 becomes a stunning success :)12:17
lekernelI don't think we should change the PCB layout12:17
wpwrak(DNP/0R) yes, that sounds like the easiest approach12:17
lekernelso it stays compatible with B12:17
wpwrakalso gives you some flexibility if some obscure issues pop up12:18
wolfspraulpcb layout has to be changed for the logic gates, no?12:49
wolfspraulso we can fix L19 and the antialiasing filter removal in the pcb as well, unless it's just not wanted...12:49
wolfspraulso that was correct - L14-L16 and C202-C207 can go?12:50
wolfsprauleither DNP/0R or remove altogether...12:52
lekernellet's test the logic gates thoroughly first12:55
lekernelthey work flawlessly in theory, but the diodes were supposed to as well ...12:55
lekernelbut yes, the PCB layout will need a few changes in all cases12:55
wolfspraullekernel: if the PCB layout changes anyway, do you still want to keep L14-16/C202-C207 or it's ok to remove them entirely?12:58
lekernelit should be ok to remove them entirely12:58
wolfspraulthx12:58
lekernelbut it breaks compatibility with B12:59
wolfspraulsomeone unsoldering the C to revert to B? :-)12:59
wolfspraulbtw our sourcing contact in Shenzhen did indeed confirm that the B is harder to source now, C is cheaper too13:00
wolfspraulC is 1.50 USD cheaper13:01
wolfspraulunless we buy B in larger quantity, then it might be about the same as C :-)13:03
wolfspraulsourcing magic13:04
lekernelwhat is that "unicore32" architecture?13:33
lekernelhttp://www.google.com/search?q=unicore3213:33
lekernelseems to be some research processor by a Chinese lab13:33
lekernel"Based on the Unity-863 SoC, MPRC has built a complete SoC design platform. MPRC is now working at several SoC design projects with the cooperation of several organizations using this platform." ...which ones??13:34
lekernelgoogle gives only qemu and linux patches13:35
lekernelhttp://cadlab.cs.ucla.edu/icsoc/protected-dir/IC-DFN_Agenda_Aug_2007/Xianfeng%20Li-The%20SuperK%20Project%20at%20MPRC.pdf13:37
awrc3 voltage/current measurements before reflash done: total current is 0.53A +/- 0.01A, 0x62 gets 0.6A while my lab power supply sets @ 0.7A.13:39
awnext step to reflash tomorrow13:40
lekernelof course, they get the usual autocrap problems: http://fgsink.blog.163.com/blog/static/1671699702011624111911824/13:41
lekernelaw, some part on 0x62 probably gets hot13:42
awlekernel, ha~i put my fingers to feel, didn't find. ;-)13:43
awlet's temporarily leave 0x62 now, let's go for reflash and use test tool to test others then back to investigate failed boards. :-)13:46
lekernelbtw you are still using 'verify' during flashing?13:48
lekernelthis is painfully slow. I though Xiangfu would have integrated CRC checking into the test program ...13:48
awhi no, only i enable 'verify' while I re-test once I wanted to check flash chip. yes super painfully slow though.13:49
awguessed xiangfu haven't integrated CRC checking into it, but good reminder, I email him.13:52
awnight all. cu13:59
larsclekernel: hm, i accidentally pushed made HEAD point to the lm32 branch instead of master in openwrt-milkymist branch, and apparently you can only change it if you have admin rights for the repo15:24
GitHub129[extras-m1] yizhangsh pushed 1 new commit to master: https://github.com/milkymist/extras-m1/commit/10f3bd0dfe2a63d81129699b469de29f3c3001b715:26
GitHub129[extras-m1/master] changed fonts in use - Yi Zhang15:26
lekernellarsc, better now?15:27
lekerneldoes this openwrt port work btw?15:27
larscyes15:29
larscclone, select milkymist, run make and you get a bootable kernel+rootfs15:30
lekernelyes it's better, or yes it works? :)15:30
larscboth ;)15:30
lekernelwith the master or lm32 branch?15:30
lekernelcool15:30
lekernellet me try :)15:30
lekernelah, and it's got ipv6 ;)15:34
larsckiller feature ;)15:49
larschm, apparently to much of a killer feature15:50
lekernellarsc, http://pastebin.com/x2tkr6kB16:07
larschttps://dev.openwrt.org/ticket/923516:12
lekernelha https://dev.openwrt.org/ticket/923516:13
lekernelyes16:13
GitHub85[extras-m1] yizhangsh pushed 1 new commit to master: https://github.com/milkymist/extras-m1/commit/53f2017f1bd858782f258d7e7e04b4a17dcad32a16:19
GitHub85[extras-m1/master] removed unused font - Yi Zhang16:19
larscso the bug is libgmpxx not linking against libstdc++?16:30
lekernelapparently. the proposed fix worked for me.16:34
lekernelstill compiling now ...16:34
mwallelekernel: larsc: is networking working?16:54
mwallethe linux port still has the old minimac1 driver16:54
mwallelooking at it atm16:54
mwallebbl16:55
lekernelCloning into elf2flt-08d80172231853b2d2a7bf03d94a9d7115c3d348...17:03
lekernelfatal: dumb http transport does not support --depth17:03
larscgit revert 0fbc1903f8d516e03af319419ae90ea1811dbe5517:13
lekernelyup, done. i was just reporting the issue.17:18
lekernelnow the kernel won't compile:     LD      usr/built-in.o17:18
lekernelmake[7]: *** No rule to make target `arch/lm32/kernel/gpio.o', needed by `arch/lm32/kernel/built-in.o'.  Stop.17:18
larschm...17:18
Action: larsc blames mwalle ;)17:19
larsci guess it built here because i already had gpio.o17:19
lekernelthis one is a bit more troublesome *g*17:19
lekerneli'm editing out gpio.o from arch/lm32/kernel/Makefile17:21
lekerneland pray it will link17:21
larscshould work gpio.c was moved to platforms/milkymist17:23
GitHub183[linux-milkymist] larsclausen pushed 1 new commit to master: https://github.com/milkymist/linux-milkymist/commit/2b01591062b9d4ee4bb4c05485f6c8f7b9e73e7e17:23
GitHub183[linux-milkymist/master] lm32: Don't try to build kernel/gpio.o - Lars-Peter Clausen17:23
lekernelso we can play with the LEDs from linux now? :)17:25
larscyes17:25
larsceven via the leds sysfs interface17:25
GitHub189[linux-milkymist] larsclausen pushed 1 new commit to master: https://github.com/milkymist/linux-milkymist/commit/f37bef13afa096897bb49f124a94a0ee9dbec08e17:25
GitHub189[linux-milkymist/master] lm32: milkymist_one.dts: Set bootargs - Lars-Peter Clausen17:25
larscyou'll need the last commit as well, otherwise it won't mount the rootfs17:26
lekernelmh, what is simpleImage.milkymist_one ?17:28
lekernelI just boot that? kernel+initrd combo?17:28
lekernelapparently it's just kernel ...17:29
larscthats just the kernel17:29
lekernelfor some reason it seems it did not build a ramdisk image17:30
lekerneleven though it's selected in menuconfig17:30
lekernells bin/lm32/17:30
lekernelmd5sums  openwrt-lm32-rootfs.tar.gz  packages  simpleImage.milkymist_one17:30
larschm17:30
kristianpaullarsc: (bootable kernel+rootfs) great !!17:31
larsclekernel: just select ext4 under target images17:33
larscand no ramdisk17:33
lekernelthere is no such option17:33
lekernelah, I have to deselect ramdisk before ...17:33
lekernelKernel panic - not syncing: No init found.  Try passing init= option to kernel. See Linux Documentation/init.txt for guidance.17:41
lekernelwith: Kernel command line: init=/sbin/init above17:41
lekernelwtf17:41
larschm17:42
larscrootfs is mounted?17:42
larscroot=/dev/ram0 rootfstype=ext217:42
lekernelyes... maybe that's because of the \n at the end17:42
lekernelof cmdline.txt17:42
lekernelnope. still won't work.17:43
lekernelRAMDISK: Loading 8192KiB [1 disk] into ram disk... done.17:43
lekernelVFS: Mounted root (ext2 filesystem) readonly on device 1:0.17:43
lekerneldevtmpfs: mounted17:43
lekernelFreeing unused kernel mem: 104k freed17:43
lekernelKernel panic - not syncing: No init found.  Try passing init= option to kernel. See Linux Documentation/init.txt for guidance.17:43
lekernelchecking initrd image atm17:43
lekernelwell... there is /sbin/init .. linked to the existing ../bin/busybox17:45
lekernelbut it refuses to work17:45
larsccould you upload your rootfs?17:46
kristianpaullarsc: this linux was tested in wich bitstream SoC version?17:47
GitHub98[extras-m1] yizhangsh pushed 1 new commit to master: https://github.com/milkymist/extras-m1/commit/01b0d4e64262a7d3ac775a7bd2593381043ed13317:48
GitHub98[extras-m1/master] 1) updated hackable-devices contact info; 2) rearranged page for A3 printing - Yi Zhang17:48
larsckristianpaul: i've tested it with a reduced bitstream17:48
lekernellarsc, http://www.milkymist.org/initrd.bin.bz217:49
kristianpaullarsc: reduced from wich SoC version?17:49
lekerneland kernel command line is init=/sbin/init root=/dev/ram0 rootfstype=ext217:50
larsckristianpaul: i've disabled a few peripherals that are not supported by linux anyway17:51
kristianpaulincluding ehernet core? :-)17:53
kristianpaulanyway, i guess it should work..17:55
kristianpauli'll try as well17:55
larschm, lekernel apparently something went wrong with elf2flat on your side17:55
larsclekernel: could you paste your openwrt .config?17:55
larscah...17:56
larscbusybox ignores the default CFLAGS or something like that...17:56
lekernelhttp://pastebin.com/enr7NRXD17:56
larsclekernel: CONFIG_BUSYBOX_CONFIG_EXTRA_CFLAGS="-Wl,-elf2flt"17:56
larscshould make it work17:57
lekernelyes, works :)18:05
lekernelthanks :)18:05
lekerneland congratz :)18:05
larscthanks18:05
lekernelthere's no framebuffer output (even though I have "milkymist_fb e0003000.vgafb: fb0: Milkymist frame buffer at 47200000, size 1536kB" in the log)18:09
mwallelekernel: on real hw?19:57
mwallebtw simpleImage is kernel + dtb19:57
lekernelyes19:57
mwalledtb linked into the kernel, normally, the dtb is an extra blob provided by the board19:57
mwallelekernel: mh i've only tested on qemu19:57
lekernelmaybe it needs a clearing of the reset bit19:58
lekernelthe bios sets it19:58
mwalle/* take hardware out of reset */19:59
mwalleiowrite32be(0, milkymistfb->ctrlbase + VGAFB_CTRL);19:59
mwallemaybe the sync and scan registers are wrong20:00
mwalleall ignored by qemu20:00
mwallelekernel: the control interface should work if hw is in reset, shouldnt it?20:02
lekernelyes20:03
lekernelin fact you should only modify the scan register values when it's in reset20:03
mwallelekernel: could you verify the settings?20:04
mwallebtw does anybody know a way to access the physical memory through /dev/mem with standard unix tools?20:05
lekernelwith dd, maybe?20:06
mwalledidnt work last time i tried it (on ppc) dunni, maybe i've done sth wrong20:06
lekernelthere are known-good register values in https://github.com/milkymist/milkymist/blob/master/software/libhal/vga.c#L26720:08
mwallelekernel: could you upload your working initrd?20:11
lekernelmilkymist.org/openwrt20110727.tar.bz220:13
mwallethx20:14
mwalletypo?20:15
mwalle40420:15
lekernelargh, yes20:17
lekernelshould be ok now20:17
mwallemh, my kernel wont work20:22
mwallelekernel: could you send me your kernel config? :)20:32
GitHub89[linux-milkymist] mwalle pushed 1 new commit to master: https://github.com/milkymist/linux-milkymist/commit/e94ece6b49773e8cf000d6f8cd220695e0f82d6220:33
GitHub89[linux-milkymist/master] lm32: enable flat support by default - Michael Walle20:33
lekernelI used the default config built by openwrt ... which I deleted, sorry20:33
mwalleright after init is started, there a message about uci_load (which i get in the good case too) but no prompt after that20:48
mwallelarsc: do you have any idea?20:49
lekernelI had that message too, but the prompt followefd20:51
lekernel-f20:51
kristianpaulthere is an app in busybox that allow me to transfers files over tty?21:31
kristianpaulat the same time i'm using it..21:31
mwallelekernel: mh vscan and vsync_{start,end} are different21:32
mwallekristianpaul: rx?21:32
kristianpaulah, cool thank one, thks mwalle :)21:34
kristianpaullarsc: can you publish somwhere you're working bitstream with minimac old version?21:44
mwallelarsc: did you test the (old) milkymistfb driver on real hw?21:50
lekernelthe old driver worked for me22:08
mwallemh22:10
mwallei'll flash my mm1 tomorrow :)22:10
mwallelekernel: i guess i have to look at the rc2 schematics for the ac97 rework?22:14
timgHey there I just got a MilkyMist one and the keyboard will not work from a Mac are there specific known working keyboards?22:24
wpwraktimg: you may have to experiment a bit. as a general rule, the simpler the keyboard, the better. USB hub, built-in track pad, and similar things are to be avoided22:30
wpwraktimg: they did a bit of research here: http://en.qi-hardware.com/wiki/Milkymist_One_accessories#keyboard22:30
timgok, I thought I may find that to be the case22:32
rohmwalle: afaik the rework only was removing the codec and placing an new one22:32
rohmaybe one or 2 resistors needed changing. not sure anymore.22:33
wpwrakroh: "only" :) if i remember correctly, it has been performed successfully only once so far22:33
timgany chance anyone got like an apache server on the device to control OSC etc?22:34
rohwpwrak: we did it here in the raumfahrtagentur, so it is, per definition basic kitchen-table engineering ;)22:39
roh C255 47nF/0402, 3D high pass filter. C253 10uF/0805, C254 100nF/0402, CAP2 de-coupling capacitors U1.pin_44, SPDIF_ENABLE U1.pin_48, SPDIF OUT22:39
rohhttp://lists.milkymist.org/pipermail/devel-milkymist.org/2011-March/001328.html22:39
rohthats what i can find about the codec rework. afaik it did just work after swapping. not sure if the cap changes are that important.22:40
timgdoes MM1 come with OSC support? if not is it easy to add?22:47
rohtimg: it does23:02
timgoh cool23:03
timgnow to start work on a new Android App :-)23:03
rohtimg: flickernoise uses https://github.com/milkymist/liboscparse23:05
timggood info23:06
--- Thu Jul 28 201100:00

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