| kristianpaul | antgreen: hello, | 01:16 |
|---|---|---|
| kristianpaul | antgreen: i was wondering (just for intelectual curiosity) where i can find the bsp for moxie in newlib | 01:17 |
| kristianpaul | i just noticed you have lots of libraries in src :-) | 01:17 |
| kristianpaul | xiangfu: seems you are near to have the first nanonote build host for debian :-) | 01:20 |
| wolfspraul | xiangfu: can the Ben detect whether the USB cable is inserted? | 01:31 |
| wolfspraul | and/or whether it is charging, I guess (I think it always tries to draw power from USB when connected) | 01:32 |
| wolfspraul | the reason I am asking is because I find the screen blanking so annoying, so I thought when the Ben is connected to USB, we could disable screen blanking entirely. I think as long as it's connected the screen should stay on. | 01:33 |
| roh | that should be possible, one way or another | 01:34 |
| kristianpaul | how i do recover "stashed" changes so i can commit?.. | 01:37 |
| kristianpaul | oh, dear trying to remenber what did one month ago, and found some verilog bubbles wpwrak may find interesting :-) | 01:38 |
| pabs3 | kristianpaul: git stash pop/apply | 01:42 |
| kristianpaul | pabs3: thanks !, but.. | 01:44 |
| kristianpaul | what is actually pop? | 01:44 |
| kristianpaul | branch? | 01:44 |
| pabs3 | applys the top patch in your stash, then removes it from the stash | 01:44 |
| kristianpaul | ah i undertand is like a cp /tmp/garbage /repo | 01:44 |
| pabs3 | if you run gitk --all you'll see that the stash is just another branch | 01:45 |
| kristianpaul | k | 01:50 |
| kristianpaul | yeah a picture of this is very apreaciated | 01:50 |
| xiangfu | wolfspraul, yes. the /sys/devices/platform/gpio-charger.0/power_supply/USB/online will indicate that if USB cable is inserted. | 02:07 |
| kristianpaul | oh :) | 02:10 |
| wolfspraul | xiangfu: nice | 02:14 |
| wolfspraul | is it possible to disable screen blanking when the USB cable is connected? | 02:15 |
| xiangfu | gmen2x: is kind of easy. we can add some code to gmenu2x detect the usb cable then disable/enable the screen blanking. | 02:16 |
| xiangfu | tty0: it needs a daemon app I guess. :) | 02:16 |
| xiangfu | gmenu2x and tty0/1/2/3 use different blanking method. | 02:17 |
| xiangfu | :) | 02:17 |
| wolfspraul | ok | 02:17 |
| wolfspraul | I was trying the 11-13 image a bit and I'm super excited. great work! | 02:17 |
| wolfspraul | how do you like it? | 02:17 |
| xiangfu | I think it getting better. | 02:18 |
| xiangfu | :-) | 02:18 |
| wolfspraul | yes, definitely | 02:18 |
| xiangfu | I start to exciting when the 08-27 come out. :) | 02:18 |
| xiangfu | this 11-13 is better then 08-27. | 02:19 |
| wolfspraul | is the nanonote support in u-boot upstream now so that we can remove the special u-boot from openwrt? | 02:19 |
| wolfspraul | I'm a little worried that >80% of the 512mb image are full though (in 11-13) | 02:19 |
| xiangfu | when 11-13 come out. all I want is copy more media file to my nanonote. picture/video/music. | 02:19 |
| wolfspraul | I think we already had to remove some sw from the image in the past, and we will run into more trouble there | 02:20 |
| wolfspraul | we need the full 2gb I think | 02:20 |
| wolfspraul | in some way | 02:20 |
| xiangfu | wolfspraul, u-boot. yes. needs some work. since the openwrt uboot-xburst support 3 device, I will try to merge them to new u-boot. or create another uboot-nanonote maybe. not sure. | 02:21 |
| xiangfu | wolfspraul, yes. we removed gcc stuff. | 02:22 |
| wolfspraul | yeah | 02:23 |
| wolfspraul | bad | 02:23 |
| wolfspraul | :-) | 02:23 |
| wolfspraul | I wanted to use irssi, but not there | 02:23 |
| wolfspraul | how do I disable screen blanking in tty0? | 02:25 |
| wolfspraul | another thing that confuses me is our long list of terminal selections | 02:26 |
| wolfspraul | when I'm in a terminal, is there an easy way to iterate over console fonts? | 02:27 |
| wolfspraul | I think there should be some hotkey that just flips through all fonts and redraws the screen | 02:28 |
| xiangfu | (tty0) use setterm -blank 0 | 02:28 |
| wolfspraul | perfect, thanks! | 02:30 |
| xiangfu | wolfspraul, (fonts) yes. since we have triggerhappy. it not hard to write a script file do that. | 02:30 |
| xiangfu | let me try to do that now. :) | 02:30 |
| wolfspraul | opkg list shows 6365 packages!? | 02:32 |
| wolfspraul | are we really building over 6000 packages? | 02:32 |
| wolfspraul | 495 installed | 02:33 |
| wolfspraul | (I installed a few manually, maybe it was around 490 in the 11-13 image) | 02:33 |
| xiangfu | 493 | 02:33 |
| xiangfu | nanonote have totally 2900+ packages in package repo. | 02:34 |
| wolfspraul | do we have a nanonote-docs or so where we can start to write up tutorial/faq/tips/etc ? | 02:35 |
| wolfspraul | I see we have man, no info | 02:36 |
| xiangfu | don't have a nanonote-docs for now. that is the plan. haven't start. :) | 02:36 |
| wolfspraul | ok sure, great | 02:36 |
| wolfspraul | maybe we just use html and one of the many text browsers we have | 02:36 |
| wolfspraul | wow, netsurf is so cool! | 02:37 |
| wolfspraul | :-) | 02:37 |
| wolfspraul | I think I have a new use case for my Ben - I leave it connected to my notebook, and move apps to the Ben | 02:38 |
| wolfspraul | I probably start with chat & email | 02:38 |
| wolfspraul | my emails will get shorter :-) | 02:38 |
| rejon | xiangfu, can you come to that meeting tomorrow at 3 at fisheye? | 03:01 |
| rejon | without you, going to be like me and wolfgang talking in hand gestures | 03:01 |
| wolfspraul | I tried listener, and then mplayer to play the .wav, but it couldn't play because of missing codec... | 03:15 |
| xiangfu | yes. | 03:16 |
| xiangfu | 'aplay' works. | 03:17 |
| xiangfu | still needs some work on mplayer :) | 03:17 |
| wolfspraul | I think now that the base system seems more or less under control (bootloader, kernel, rootfs/console/gui), we have chance to improve big time on the following 3 major features: | 03:17 |
| wolfspraul | 1) backup | 03:17 |
| wolfspraul | 2) encrypt | 03:17 |
| wolfspraul | 3) update | 03:17 |
| wolfspraul | I keep thinking how to achieve it, and the most promising idea I can come up with is to look for existing small utilities/tools, even command line, to help with those. | 03:18 |
| wolfspraul | for update, can we rely on opkg to just keep updating packages? | 03:18 |
| wolfspraul | not sure whether we can update the bootloader or kernel that way, or how this interacts with major openwrt releases | 03:18 |
| xiangfu | rejon, oh | 03:21 |
| xiangfu | update. yes. the only one we should case is kernel. the openwrt kernel package is just a empty package only have some version file. | 03:23 |
| xiangfu | we should add some post-install. flash the uImage to kernel partition. | 03:24 |
| xiangfu | (updating packages). yes. we should try. | 03:24 |
| xiangfu | if the uClibc and gcc update that maybe more trouble? | 03:25 |
| wolfspraul | probably | 03:26 |
| wolfspraul | like you said we need to experiment carefully | 03:26 |
| wolfspraul | and in parallel continue the full image releases and reflash_ben.sh we use today | 03:26 |
| xiangfu | yes. | 03:26 |
| wolfspraul | how do most openwrt users update their devices? (mostly routers I guess) | 03:28 |
| wolfspraul | in my own openwrt usage, it was typically some serious focused effort to get it to a router, say a few hours or days full-time, and then I leave that router alone until the hardware fails x years later :-) | 03:28 |
| wolfspraul | no security updates, nothing | 03:28 |
| wolfspraul | is that the typical openwrt usage model? I hope not :-) | 03:28 |
| xiangfu | openwrt update is like 'upload a image file to router' since mostly router only have 4M or 8M flash. | 03:37 |
| xiangfu | I think this image include kernel and rootfs I guess. (forget the detail) | 03:37 |
| wolfspraul | I'm wondering how people continue after that | 03:39 |
| wolfspraul | like I said - I just leave it like that until the hardware fails | 03:39 |
| wolfspraul | not the most secure approach | 03:39 |
| wolfspraul | I am sure a lot of people at least will install security updates incrementally, via opkg (?) | 03:39 |
| wolfspraul | how about upgrading to the next major openwrt version? maybe everybody just reflashes everything? | 03:40 |
| wolfspraul | we probably should not try something that is not in line with the typical update model of most openwrt users | 03:40 |
| wolfspraul | basically we can just do what is easy to do with openwrt | 03:41 |
| jow_laptop | wolfspraul: there is sysupgrade nowadays which simplifies reflashing | 03:43 |
| jow_laptop | and configs remained largely compatible during the last few years | 03:43 |
| wolfspraul | sysupgrade | 03:44 |
| wolfspraul | sounds good | 03:44 |
| wolfspraul | thanks! | 03:44 |
| wolfspraul | will check out | 03:44 |
| wolfspraul | jow_laptop: how about backup or encryption - any tools you are aware of? | 03:44 |
| jow_laptop | some people use loop-aes afair | 03:44 |
| jow_laptop | as for backup, don't know, what do oyu envision? something networked or more an archivier? | 03:45 |
| wolfspraul | something that protects me in case of nanonote loss or total failure | 03:45 |
| wolfspraul | drop in water, stolen, etc. | 03:45 |
| wolfspraul | so I want to have a backup on my notebook, then restore from that backup | 03:46 |
| jow_laptop | provide an httpd server with a simple backup tar.gz download? | 03:46 |
| jow_laptop | thats what we do for the webui | 03:46 |
| jow_laptop | there is "opkg list-changed-conffiles" which is useful to find user edited package configs | 03:46 |
| jow_laptop | plus home | 03:47 |
| wolfspraul | since we talk about openwrt - how about updates through opkg? | 03:47 |
| wolfspraul | is that possible/supported? | 03:48 |
| jow_laptop | technically doable | 03:48 |
| wolfspraul | even across major releases? | 03:48 |
| wolfspraul | or totally unsupported? | 03:48 |
| jow_laptop | there are two big issues of one doesn't apply for you | 03:48 |
| jow_laptop | one problem is that kernel updates through opkg are not pssoible | 03:48 |
| jow_laptop | the other is space on the squash+jffs2 combo | 03:48 |
| jow_laptop | oh and of curse the uclibc binary compat | 03:48 |
| jow_laptop | but that depends on how you develop | 03:49 |
| jow_laptop | I mena whether you choose to update it or not | 03:49 |
| wolfspraul | why are kernel updates through opkg not possible? | 03:50 |
| jow_laptop | because the kernel image is usaully in a dedicated flash partition | 03:50 |
| jow_laptop | without any fs | 03:50 |
| jow_laptop | which various from model to model | 03:50 |
| jow_laptop | *varies | 03:50 |
| jow_laptop | so there is no generic mechanism | 03:50 |
| wolfspraul | ok makes sense | 03:51 |
| wolfspraul | but we could have a dedicated package or post-install script | 03:51 |
| jow_laptop | if there is a way to write a new kernel from within a running openwrt then this is solvable as well | 03:51 |
| jow_laptop | one can encode the process in a postinstall script then | 03:51 |
| jow_laptop | so the bits and pieces for upgrade are all there | 03:52 |
| wolfspraul | how do most openwrt users upgrade? | 03:52 |
| jow_laptop | sysupgrade | 03:52 |
| jow_laptop | which is basically reflash + config restore | 03:52 |
| jow_laptop | it does not retain installed packages though | 03:53 |
| jow_laptop | most targets are still in the 4MB..8MB flash memory range | 03:54 |
| jow_laptop | there simply is no space for packaging overhead | 03:54 |
| wolfspraul | ok | 03:55 |
| Action: kristianpaul havent update his routers since a year.. | 03:55 | |
| wolfspraul | no regular/automatic security updates? | 03:55 |
| wolfspraul | kristianpaul: hey :-) sounds like my model :-) | 03:56 |
| wolfspraul | I get openwrt to setup the way I want to, then leave it untouched until hardware failure :-) | 03:56 |
| kristianpaul | well not that i wait for that, i regulary load new image as a RC is avaliable | 03:57 |
| jow_laptop | there are package upgrades, but binary releases are not openwrts strong point and we're still figuring it out | 03:57 |
| jow_laptop | organization-wise | 03:57 |
| wolfspraul | ok good | 03:57 |
| wolfspraul | so for now the Ben follows the common/regular openwrt model | 03:57 |
| wolfspraul | that's good | 03:57 |
| jow_laptop | that is basically up to the people building openwrt based distros :) | 03:57 |
| wolfspraul | how is usage of openwrt proper or distros developing actually? | 03:58 |
| wolfspraul | mostly routers? | 03:58 |
| kristianpaul | about backup, at lest consider that most apps will use root home folder, including at least for me some games and nanonote | 03:58 |
| wolfspraul | the number of supported models increasing/decreasing? | 03:58 |
| wolfspraul | I don't follow routers at all I have to admit... | 03:58 |
| jow_laptop | wolfspraul: it is becoming more models | 03:59 |
| kristianpaul | i have loose my maps and game leves after every relfash, but i'm really lazy to backup it up :) | 03:59 |
| kristianpaul | for the rest of precious content i just use a memcard | 03:59 |
| kristianpaul | wolfspraul: you dont need follow, use they work well you ignore :) | 03:59 |
| wolfspraul | kristianpaul: yes, I don't think you are alone. So we need to improve that and make those 3 things dead simple: backup, encryption, update | 03:59 |
| jow_laptop | wolfspraul: you should build a backup facilityi nto reflahs-ben.sh | 04:00 |
| jow_laptop | make it connect the ben and download a tar.gz with changed configs plus user root | 04:00 |
| jow_laptop | then reflash | 04:00 |
| jow_laptop | then connect the ben again and restore it | 04:00 |
| wolfspraul | good idea | 04:00 |
| wolfspraul | do you have stats on the number of openwrt users? | 04:01 |
| jow_laptop | no | 04:01 |
| jow_laptop | I suppose we could evaluate the http server logs | 04:01 |
| jow_laptop | but that is skewed as dd-wrt uses openwrt packages too | 04:01 |
| kristianpaul | or include rootkits in next release ;-) | 04:01 |
| kristianpaul | openwrt is getting really nice those days, the other days i was consider even use like for a swtich OS | 04:02 |
| jow_laptop | http://pastebin.com/CVZQSKQp | 04:03 |
| jow_laptop | this code finds all user changed package configs plus all paths matched by shell globs in /etc/sysupgrade.conf | 04:03 |
| jow_laptop | on my home router this currently looks like this: http://pastebin.com/kfqpQ2tJ | 04:05 |
| jow_laptop | an important thing w.r.t. opkg based upgrades is that package maintainers must be careful | 04:06 |
| jow_laptop | and one needs to establish versioned depends | 04:07 |
| wolfspraul | maybe the idea to enhance reflash_ben.sh is good | 04:07 |
| wolfspraul | as a first small practical step | 04:07 |
| jow_laptop | yes | 04:07 |
| kristianpaul | posible depency issues? | 04:07 |
| jow_laptop | right now one can only depend on "libfoo", not on "libfoo >= v1.0" | 04:07 |
| kristianpaul | at lest that nannonote can upgrade it self (like M1 :D) | 04:07 |
| wolfspraul | but there will be details, for example reflash_ben.sh normally starts right now when the Ben is in a special usb boot mode | 04:08 |
| jow_laptop | this leads to issues like: https://forum.openwrt.org/viewtopic.php?pid=148933 | 04:08 |
| jow_laptop | yeah, I think its will be tricky to establish some kind of communication with the ben | 04:08 |
| jow_laptop | you need to establish some ip connectivity over udp I suppose | 04:09 |
| jow_laptop | erm *over usb | 04:09 |
| jow_laptop | so that in the end the reflash ben.sh is like itunes to the iphone | 04:09 |
| wolfspraul | we have a lot of options I think, but need to find the smallest step that delivers results | 04:09 |
| wolfspraul | oh my god, that sounds scary :-) | 04:10 |
| jow_laptop | yeah I know | 04:10 |
| wolfspraul | I hope people are not inspired by *that* analogy :-) | 04:10 |
| jow_laptop | i just mean the aspect of bundling backups, reflashes, recovery aetc. into one app | 04:10 |
| wolfspraul | you could also say "maybe the Ben will be as easy to use as AOL one day" | 04:10 |
| jow_laptop | we also experiment with web based image builders | 04:11 |
| wolfspraul | go on | 04:12 |
| jow_laptop | http://openwrt.linux-appliance.net/ib/wizard.cgi?target=xburst | 04:12 |
| jow_laptop | this way one can assemble a desired package collection online, provide additional files to embed and let it build on the server | 04:13 |
| wolfspraul | nice | 04:13 |
| jow_laptop | the final image is tailored to the custom needs then | 04:13 |
| jow_laptop | in the upload step one could upload the backup archive for example | 04:13 |
| wolfspraul | but it probably contains patent encumbered software | 04:14 |
| jow_laptop | well that entriely depends on what components are provided in the repo | 04:14 |
| jow_laptop | the link I sent is based on http://downloads.openwrt.org/backfire/10.03.1-rc6/xburst/OpenWrt-ImageBuilder-xburst-for-Linux-i686.tar.bz2 | 04:15 |
| wolfspraul | yeah it's nice | 04:15 |
| wolfspraul | can it build bootable images for the nanonote now? | 04:16 |
| wolfspraul | I'm wondering whether we have more things left to upstream still, or all there already... | 04:16 |
| jow_laptop | tbh I never tried :) | 04:16 |
| jow_laptop | it will probably boot and drop to a busybox prompt | 04:16 |
| jow_laptop | no gmenu etc. | 04:16 |
| wolfspraul | that would be great | 04:19 |
| wolfspraul | jow_laptop: if you feel there is anything we can do better for openwrt upstream, please speak up | 04:19 |
| jow_laptop | can't think of anything | 04:19 |
| jow_laptop | some things will never be applicable mainline | 04:19 |
| wolfspraul | xiangfu worked hard to get nanonote support into u-boot upstream, we can probably cleanup a bit on the openwrt side now | 04:19 |
| jow_laptop | and the other stuff slowly trickles in | 04:20 |
| jow_laptop | like package updates | 04:20 |
| wolfspraul | oh sure, that's fine | 04:20 |
| wolfspraul | yes | 04:20 |
| wolfspraul | I also hope we get a bigger push behind OpenWrt on Milkymist One at some point | 04:20 |
| wolfspraul | Lars has been starting and made good progress, but very few people use that right now (understandably since there are serious limitations right now) | 04:21 |
| wolfspraul | jow_laptop: have you heard of Milkymist? :-) (you probably did, just checking...) | 04:22 |
| jow_laptop | yes, I follow the lists and the channel here | 04:22 |
| wolfspraul | to improve Linux and OpenWrt on it a lot of pieces are missing, such as mmu, lots of work in the Linux kernel, C++ compiler or even more recent gcc compiler support, dynamic linking missing as well I think | 04:23 |
| wolfspraul | let's say 1 year :-) | 04:23 |
| qi-bot | [commit] Werner Almesberger: m1/patches: added soc11_2.diff posted by Sebastien on Nov 23 (master) http://qi-hw.com/p/wernermisc/94fd599 | 08:22 |
| rejon_ | xiangfu, if you want to come tomorrow great | 11:09 |
| rejon_ | but now found others who can help bridge the lang. divide :) | 11:09 |
| rejon_ | wolfspraul and I will meet at fisheye cafe and talk shop | 11:09 |
| rejon_ | hopefully can look at the image stuff on m1 and more | 11:10 |
| rejon_ | http://wiki.creativecommons.org/Content_Registries | 11:13 |
| xiangfu | rejon_, if the JJ want setup the M1 tomorrow. I think I will come to fisheye tomorrow. :) | 11:15 |
| rejon_ | ok cool, its up to you | 11:16 |
| rejon_ | i know its your weekend | 11:16 |
| rejon_ | so don't stress out | 11:16 |
| xiangfu | rejon_, I will update the JJ's m1 to latest stable release. will not include image support. | 11:16 |
| rejon_ | is it still too risky? | 11:16 |
| rejon_ | ;) | 11:16 |
| rejon_ | i'll see wolfgang there at 2 pm, others will arrive at 3 pm | 11:17 |
| rejon_ | like the crowd | 11:17 |
| xiangfu | rejon_, hmm... no. just don't want break normal webupdate. I think ender user follow the stable release is better. | 11:17 |
| rejon_ | totally agree | 11:18 |
| rejon_ | who is JJ? | 11:18 |
| xiangfu | rejon_, sorry. Jing jing. | 11:18 |
| rejon_ | aha | 11:18 |
| rejon_ | cool | 11:18 |
| rejon_ | is she going to come there? | 11:18 |
| xiangfu | I am C developer. :D | 11:18 |
| xiangfu | rejon_, no. I don't think so. but she said will wire the money today, and want setup M1 this weekend. | 11:19 |
| rejon_ | oh great | 11:19 |
| Last message repeated 1 time(s). | 11:19 | |
| xiangfu | rejon_, oh. she didn't CC to you. | 11:19 |
| rejon_ | its ok | 11:19 |
| rejon_ | i'm used to that | 11:19 |
| rejon_ | chinese people don't like to do :CC | 11:19 |
| xiangfu | rejon_, I will forward this email to you. | 11:19 |
| rejon_ | ok cool | 11:19 |
| rejon_ | we have a thxgiving party that evening | 11:20 |
| rejon_ | so she must want to try it there | 11:20 |
| rejon_ | gbraad, we having a meetup at 3 pm | 11:20 |
| rejon_ | sharism | 11:20 |
| rejon_ | at fisheye cafe | 11:20 |
| rejon_ | we will flesh out openwrt + html5 more | 11:20 |
| rejon_ | you should come! | 11:20 |
| rejon_ | bring pasi | 11:20 |
| xiangfu | rejon_, hmm... I think a people working on openwrt router in Beijing, do you want more people :) | 11:21 |
| rejon_ | sure | 11:22 |
| xiangfu | s/I think/I know | 11:22 |
| rejon_ | will be like an openwrt flash mob | 11:22 |
| xiangfu | ok. he buy a Nanonote. then switch to create his own openwrt router. :) | 11:22 |
| rejon_ | want to flesh out how we can get test suite for openwrt + netsurf to get minimal html5 compliance | 11:22 |
| rejon_ | yah, i need to flash mine | 11:22 |
| xiangfu | rejon_, CC you another email. :) | 12:02 |
| rejon_ | great | 12:04 |
| xiangfu | rejon_, updated to latest version and packaged well | 12:21 |
| kristianpaul | morning | 12:46 |
| antoniodariush | xiangfu, hi, the new toolchain now has the wpan support for the atben | 12:47 |
| wolfspraul | good morning! | 12:47 |
| antoniodariush | do I still need zigbee? | 12:48 |
| lekernel | what is that LED power supply? | 13:14 |
| kristianpaul | there are some logs about it, you may found interesting uses a CPLD | 13:20 |
| wpwrak | looks like a hideously complex way of powering a LED :) | 13:22 |
| kristianpaul | fancy indeed, but still interesting | 13:24 |
| kristianpaul | http://en.qi-hardware.com/irclogs/qi-hardware_2011-10-15.log.html#t23:12 | 13:25 |
| kristianpaul | http://en.qi-hardware.com/irclogs/qi-hardware_2011-11-15.log.html#t08:03 | 13:25 |
| wpwrak | i kinda wonder what it does with all those coils. does it connect to mains ? | 13:27 |
| kristianpaul | i wonder how long batteries hold | 13:28 |
| roh | um.. why the fsck. does it need an cpld to drive a led? | 13:30 |
| kristianpaul | i remenber, the person behind is involved the IC design | 13:31 |
| lekernel | that's just a lamp? | 13:36 |
| roh | 13:38 | |
| roh | seems so. a 70W led driver | 13:38 |
| xiangfu | antoniodariush, the new release support atBen. | 13:56 |
| xiangfu | antoniodariush, I think it still not support zigbee. there are some code in kernel but not enable. | 13:57 |
| antoniodariush | yes, I compile all the necessary packages on the toolchain successfully | 13:57 |
| xiangfu | antoniodariush, (zigbee) depends what you want. the new release already support TCP under atBen. | 13:57 |
| antoniodariush | so it should work with just the packages from the toolchain ? | 13:58 |
| xiangfu | antoniodariush, you have both atBen and atUsb? or two nanonote and two atBen? | 13:58 |
| antoniodariush | two nanonotes and two atBen yes | 13:59 |
| xiangfu | antoniodariush, sorry. what packages you mean? the package you compiled? | 13:59 |
| xiangfu | antoniodariush, cool. great. | 13:59 |
| antoniodariush | i mean in the toolchain we need to select all the required drivers and stuff | 14:00 |
| xiangfu | antoniodariush, you mean SDK. yes. we needs select those tools: http://pastebin.com/Lg3sgCbS | 14:01 |
| antoniodariush | let me check my toolchain | 14:02 |
| wolfspraul | yes, led driver/power supply | 14:02 |
| wolfspraul | there's not much else to the project right now than us helping Deng a little with review and feedback | 14:02 |
| wolfspraul | in fact I'm not sure whether/how Deng will continue, that's up to him | 14:03 |
| xiangfu | antoniodariush, sorry. I am a little lose. about toolchain: what do you mean? the SDK under here: http://downloads.qi-hardware.com/software/images/NanoNote/Ben/latest/? or you compile the whole openwrt by yourself. | 14:03 |
| antoniodariush | the second | 14:04 |
| xiangfu | antoniodariush, ok. got it. | 14:05 |
| antoniodariush | xiangfu, I had the atbens working but following werner's instructions http://projects.qi-hardware.com/index.php/p/ben-wpan/source/tree/master/install/INSTALL-Ben | 14:06 |
| antoniodariush | but now that everything is in the toolchain i'm looking forward to simplify few things :) | 14:06 |
| wolfspraul | antoniodariush: if we can help you simplify more, on the software side, please let us know | 14:07 |
| wolfspraul | we'd love to do that | 14:07 |
| antoniodariush | sure | 14:07 |
| antoniodariush | anyway I have these selected already http://pastebin.com/Lg3sgCbS | 14:08 |
| wolfspraul | wpwrak: kristianpaul lekernel roh : since you guys asked about the led supply, a quick background | 14:10 |
| wolfspraul | basically it's a chinese friend who quit his job and is working on this board | 14:11 |
| wolfspraul | I don't fully understand the background of his work or business idea | 14:11 |
| wolfspraul | but he understands the gpl, is very open minded about 'open', etc. | 14:11 |
| wolfspraul | so I offered help in hosting his files, helping a bit with review, cleanup, process, translation to English, etc. | 14:11 |
| wpwrak | that's all very nice. but what on earth is the board doing ? ;-)) | 14:11 |
| wolfspraul | but he has to drive it forward, it's his board and project. if he stops, the project stops and the files stay there for whatever future benefit (or none) | 14:12 |
| wolfspraul | there's a video, no? | 14:12 |
| xiangfu | antoniodariush, if you use the latest openwrt-xborst.git commit. it already included all patch about atBen. | 14:12 |
| wolfspraul | it's a led power supply | 14:12 |
| wolfspraul | I cannot qualify this project | 14:12 |
| xiangfu | antoniodariush, yes. after you select those packages. you have everything you need :) | 14:12 |
| wolfspraul | I only provide the project server, translation, some review, process, etc. | 14:12 |
| wolfspraul | for example it might be a good boom test case for me :-) (though I doubt I get to it) | 14:12 |
| wpwrak | i wonder why the board is so complex. is it driven from mains ? | 14:12 |
| wolfspraul | yes I think so | 14:13 |
| wolfspraul | I know very little, that's one of the goals of pulling it out into the open | 14:13 |
| wolfspraul | we have to see where Deng takes this | 14:13 |
| antoniodariush | xiangfu, excellent, thanks. | 14:15 |
| xiangfu | antoniodariush, keep up update about your two nanonotes. cool. what are you plan on this ? (just want know more about atBen :) | 14:16 |
| xiangfu | I mean keep us(people here) update. :) | 14:16 |
| antoniodariush | I (actually we) always keep you people updated :) | 14:21 |
| antoniodariush | wolfspraul, did you watch the video ? | 14:22 |
| wolfspraul | yes | 14:22 |
| wolfspraul | a little confusing :-) | 14:22 |
| wolfspraul | but it looks cool - I think about all the Bens I sold! | 14:22 |
| antoniodariush | yes we have plenty | 14:26 |
| kristianpaul | wolfspraul: mechanical side for this led lamp still WIP? | 14:33 |
| wolfspraul | he | 14:35 |
| wolfspraul | dream-in-progress | 14:35 |
| wolfspraul | the led array would be a standard one, this project is just the power supply | 14:35 |
| kristianpaul | ok :) | 14:35 |
| wolfspraul | and no mechanical at all, the idea is to maybe look for automotive customers (deng's idea, not mine) | 14:35 |
| kristianpaul | hum i see | 14:36 |
| wolfspraul | led lamps are shooting up everywhere now, I really cannot tell much about the differences etc. | 14:36 |
| wolfspraul | but I read for example that the new Boeing 787 has a spectactular LED lighting system, guess I have to fly with one to see it for myself :-) | 14:36 |
| kristianpaul | hhaha | 14:36 |
| roh | hmmm. http://www.alibaba.com/showroom/70w-led-driver.html | 14:37 |
| xiangfu | antoniodariush, oh. where is the video url? | 14:38 |
| antoniodariush | http://theclashingrocks.org/wiki/doku.php?id=tcr:pics_video | 14:39 |
| antgreen | kristianpaul: newlib bsp for moxie? it depends what you mean by that. libgloss (in the same src repo as newlib) hooks into the gdb simulator and qemu port for IO, etc. | 14:39 |
| roh | uhm. seems there is a chip for that | 14:39 |
| roh | www.micrel.com/_PDF/mic3230.pdf | 14:40 |
| antgreen | kristianpaul: for the gdb simulator, the simulator intercepts software interrupts (swi instruction) to do I/O. | 14:40 |
| kristianpaul | antgreen: sorry, i just wanted to know how you ported implemented stdio for your soc | 14:40 |
| kristianpaul | hum | 14:40 |
| antgreen | for qemu, I model memory mapped IO devices | 14:40 |
| antgreen | libgloss is built for both versions, and you specify one or another with different linker scripts. | 14:41 |
| kristianpaul | ah :), well i need learn more about qemu later | 14:41 |
| antgreen | is there a milkymist qemu port? | 14:42 |
| kristianpaul | yes | 14:42 |
| kristianpaul | i thin is upstream now but be aware of http://milkymist.org/wiki/index.php?title=Using_QEMU | 14:43 |
| kristianpaul | must go, back later, thanks for explanation antgreen :) | 14:44 |
| antgreen | np | 14:44 |
| xiangfu | :) | 14:49 |
| antoniodariush | :-) | 14:49 |
| jivs | :-) | 14:51 |
| DocScrutinizer | wolfspraul: ping | 14:52 |
| DocScrutinizer | wpwrak: ping | 14:52 |
| wolfspraul | yes | 14:52 |
| wpwrak | pong | 14:57 |
| kristianpaul | hello Artyom | 20:06 |
| kristianpaul | btw i'm wokring to make my local changes to upstream | 20:09 |
| kristianpaul | later i will bug you about osgps, wich i notiched the isr is not fully implemented, just polling isnt? | 20:09 |
| kristianpaul | hum, good point | 20:13 |
| Artyom | yeah, I used polling. (I had some troubles with ISR for GPIO in ARM7 and didn't want to spent time on them) | 20:14 |
| kristianpaul | Do you think full osgs can run on that arm7 mcu? | 20:16 |
| kristianpaul | I tried it on my 333Mhz nanonote and it felt not very happy... ah but of course correaltors were gp2021 in software not namuru | 20:17 |
| kristianpaul | may be algortym to get fix point is not very cpu ungry, i havent check it yet.. | 20:18 |
| kristianpaul | btw Artyom , had you seen this http://www.holmea.demon.co.uk/GPS/Main.htm ? | 20:19 |
| Artyom | arm7 should work fine because there is gpl-gps project which runs on ARM7 (about 30 or 50 MHz) | 20:20 |
| kristianpaul | ah yes | 20:20 |
| Artyom | and gpl-gps is based on osgps | 20:20 |
| Artyom | And position solution is not cpu hungry. Usualy position is calculated every second and it's not so often as tracking loops updating | 20:23 |
| Artyom | I've seen this project (Homemade GPS Receiver) but I don't like it. The source code is poor commented and not very clear (like in MM ;) ) | 20:27 |
| kristianpaul | yes i agree | 20:32 |
| kristianpaul | what call me atention is the RF DIY receiver part | 20:32 |
| Artyom | (Sorry, what is DIY stand for?) | 20:34 |
| kristianpaul | do it your self | 20:34 |
| kristianpaul | Artyom: so, you make flterm work your board also? | 20:36 |
| Artyom | Not yet, I had no enough time to experiment with flterm. And there is one potential problem that must be solved before. | 20:37 |
| Artyom | And I also have to prepare some demo-program. The stadard are not for my board | 20:43 |
| kristianpaul | wich kind of demo progam you mean? for osgps mod? | 20:47 |
| Artyom | no, I mean that for testing flterm I could use a program in /software/demo directory. But I can't do it because my board differes from M1. So I have to write my own, or simplify existing ;) | 20:50 |
| kristianpaul | ah yes of course that demo.. (wich i never ran before... and i should :-)) | 20:55 |
| Artyom | :))) (It seems that I do many stupid things that you didn't even try to ;) ) | 20:56 |
| kristianpaul | he no just i met milkymist when demo was not needed, i mean flickernoise wich is the VJ* apps was ready | 20:57 |
| kristianpaul | so i dint get interested,but seems it cames with some debug facilites i should try someday | 20:57 |
| kristianpaul | ready written for rtems | 20:58 |
| Artyom | btw did you test RTEMS on your board? Did you write any additional (pgrogram/task/thread?)? | 20:59 |
| Artyom | I think I've read that M1 uses RTEMS as a base? | 21:00 |
| kristianpaul | yeah well rtems is the better supporte OS in terms of working drivers | 21:00 |
| kristianpaul | and yes i had tried rtems before | 21:00 |
| kristianpaul | i just moved to baremetal meanwhile i get sure all works fine (as i got some hangs in the pst) | 21:01 |
| Artyom | What drivers do you mean? | 21:01 |
| kristianpaul | ethernet (best tcp/ip stack than bios too), memory card (read only support), yaffs for NOR wich is a big plus as the only non-volatile file system with write support | 21:02 |
| kristianpaul | and framebuffer | 21:02 |
| kristianpaul | there is also a fpu but i dont know how usefull could be for us | 21:03 |
| kristianpaul | about programs.. well i tried a bit very basic sstuff, i dont get in to task/threads | 21:03 |
| Artyom | fpu can be usefull for tracking loops and for position calculation (What math functions does it support?) | 21:04 |
| kristianpaul | i need to learn more about that, also rtems have a messae passing system | 21:04 |
| kristianpaul | sorry is pfpu | 21:05 |
| kristianpaul | datasheet here http://milkymist.org/socdoc/pfpu.pdf | 21:05 |
| Artyom | yeah, this part I've already seen... | 21:05 |
| kristianpaul | something usefull in there? | 21:05 |
| kristianpaul | cos sin.... well i dont know all the math in position calculation, i guess is a kind of big matrix? :-) | 21:06 |
| Artyom | Not sure... But I think that new operations can added. It's open source :) | 21:07 |
| kristianpaul | :-) | 21:07 |
| kristianpaul | for tracking loops how usefull can be? | 21:09 |
| kristianpaul | i was actually thinking in a dedicated lm32 core for that maye be to much ? | 21:10 |
| Artyom | may be atan calculation... But many thing must be checked... | 21:10 |
| Artyom | dedicated lm32 core only for tracking? ;) Sounds unusual but interesting ;) | 21:11 |
| kristianpaul | lol | 21:12 |
| kristianpaul | well as i said i was checking some gps baseband chips other days | 21:12 |
| kristianpaul | most of then include an arm procesor and a minimal soc with the correlators and other stuff | 21:12 |
| kristianpaul | but anyway just an idea | 21:12 |
| Artyom | what chips did you check? | 21:13 |
| Artyom | I remember nj1006 (or similar) It contains leon+peripheral+correlator | 21:13 |
| kristianpaul | oh leon :) | 21:13 |
| kristianpaul | just a ATR0625P from atmel | 21:14 |
| Artyom | and gp4021 (or similar): arm+correlator | 21:14 |
| Artyom | Do you leon? Did you test it? | 21:14 |
| kristianpaul | afaik no | 21:14 |
| kristianpaul | i just know it was in space because have a fault tolerance system | 21:15 |
| Artyom | I think Carlos know more about it... ;) | 21:16 |
| kristianpaul | yes he was doing benchmarking lm32 vs leon :D | 21:16 |
| kristianpaul | with linux | 21:16 |
| Artyom | Have you seen any results? | 21:17 |
| kristianpaul | nope, actually long time since i dont see carlos here | 21:17 |
| kristianpaul | i think was univesity asigment, dunno if got in a hurry and... | 21:18 |
| Artyom | btw is it possible to have couple lm32 cores in one FPGA? Did you try to do it? | 21:22 |
| kristianpaul | good questions, lekernel said yes is likely posible | 21:23 |
| Artyom | and how to transmit data between them? | 21:24 |
| Artyom | is it possible to have two cores on one wb-bus? | 21:24 |
| kristianpaul | yes | 21:24 |
| kristianpaul | several masters and use DMA | 21:24 |
| kristianpaul | i guess | 21:25 |
| kristianpaul | (dma part) | 21:25 |
| kristianpaul | or implement a wishbone switch (milkymist used it before) | 21:25 |
| Artyom | Have you ever seen FFT implementation for FPGA (as a separate core)? | 21:26 |
| kristianpaul | urghh no :) | 21:26 |
| kristianpaul | FFT is very ungry in reources i think | 21:27 |
| Artyom | may be, I didn't think about it ;) | 21:27 |
| kristianpaul | and most sutiable for software based processing (Ghz..)? (correct me in that one) | 21:28 |
| kristianpaul | at least in some books about GPS software aprouch imedialy skip serial processing and jump to fft | 21:28 |
| kristianpaul | of course they dont use a fpga likes us :) | 21:28 |
| Artyom | FFT for software? I think it can be parallelized and that's a good task for FPGA. | 21:29 |
| kristianpaul | hum.. perhaps i need read more | 21:29 |
| Artyom | In books they don't use serial approach because it's toooooo slow for software. | 21:29 |
| kristianpaul | :D | 21:30 |
| Artyom | But at the same time it's very simple for hardware :) | 21:30 |
| kristianpaul | but fft is required for acquisition , tracking loops and ... | 21:30 |
| kristianpaul | yeah | 21:30 |
| Artyom | fft is for acquisition only (as I understand) | 21:30 |
| kristianpaul | fft also needs too much routing, oh hell | 21:32 |
| kristianpaul | back in a minute out for a snack | 21:33 |
| Artyom | Kristianpaul, it's time to leave for me... | 21:33 |
| Artyom | too late... | 21:34 |
| kristianpaul | have nice sleep then | 21:34 |
| Artyom | thanks! bye! :) | 21:35 |
| --- Sat Nov 26 2011 | 00:00 | |
Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!