**** BEGIN LOGGING AT Tue Nov 09 02:59:58 2010 Nov 09 08:56:08 elo Nov 09 09:12:55 hrw, can you fix my network driver? Nov 09 09:13:11 actually, I wonder if they have a useable datasheet Nov 09 09:14:11 cwillu_at_work: I am not kernel hacker sorry Nov 09 09:14:20 and? Nov 09 09:15:17 cwillu_at_work: I guess you can google for the datasheet faster than hrw :P Nov 09 09:15:52 hmm Nov 09 09:16:27 I wonder if this stick on the chip is actually intended for heat sinking purposes or something, or if it's just a qa thing Nov 09 09:16:35 given that it completely covers the chip Nov 09 09:16:58 also, who took my small screwdriver? Nov 09 09:18:33 never a prpplague around when you need one :/ Nov 09 09:19:39 * cwillu_at_work gets some rubbing alcohol Nov 09 09:25:47 80 pages Nov 09 09:25:51 this might be useful Nov 09 09:26:34 ;D Nov 09 12:25:31 hi. who could i prod to make a change to the images here? http://cdimage.ubuntu.com/ubuntu-netbook/ports/releases/10.10/release/ Nov 09 12:26:09 i'm not sure if the omap4 ones are meant to be pandaboard-specific (some bits are), but currently it doesn't allow overriding the kernel cmdline arguments Nov 09 12:26:43 which makes it difficult to add a console= argument to get any sign of life Nov 09 12:29:03 bernard_: after writing the image to your sd card you're free to mount the first disk partition and change the boot.scr file Nov 09 12:29:13 yep, i know this Nov 09 12:29:14 with the proper cmdline arguments you need Nov 09 12:29:24 so, what exactly do you want to do? Nov 09 12:29:34 but other people are running into the same issue. and it seems odd that it doesn't do so out of the box Nov 09 12:29:49 i just add console=ttyO2,115200n8 to the cmdline Nov 09 12:30:03 which if it's a pandaboard-specific image is probably fine to do Nov 09 12:30:22 yeah, there was a decision to avoid enabling console by default Nov 09 12:30:35 the omap 4 image is basically targeting panda, so that's ok Nov 09 12:31:14 please take this as a vote to enable console by default :) Nov 09 12:31:32 the kernel can be booted with quiet, so as to not slow down boot, and then a getty put on the terminal Nov 09 12:32:18 yeah, I don't see why we can't at least enable a valid getty for it Nov 09 12:32:49 that would rock Nov 09 12:33:11 a way to append to bootargs without having to edit the SD card would be awesome too! Nov 09 12:33:29 wont happen unless you convince TI to add NAND Nov 09 12:33:35 there is no ways to save them Nov 09 12:33:54 i want to be able to override from u-boot (mainly for dev) Nov 09 12:34:06 e.g. trying to figure out what args to pass to the kernel to get video working how i want Nov 09 12:34:12 iirc sakoman was working in a way to save the env somewhere else than nand Nov 09 12:34:22 if you add the above console= parameter to treh cmdline *before* first boot it will set up a serial console for you automatically Nov 09 12:34:27 don't remember if it was just emmc though Nov 09 12:34:51 yes, we might switch to a raw partition intead of vfat in natty Nov 09 12:35:11 ogra_ac: did you test that, because the all cmdline arguments kernel patch to init was applied later on Nov 09 12:35:14 that way you can treat iut like NAND and have some area to saveenv to Nov 09 12:35:25 and we had to create a different uImage for blaze at least Nov 09 12:35:37 iirc jasper parses /proc/cmdline still Nov 09 12:35:48 so the kernel patch shouldnt be needed Nov 09 12:36:06 hehe, some files parse /proc/cmdline and other don't Nov 09 12:36:40 ogra_ac: but why can't we just enable the getty for uart by default? Nov 09 12:36:40 i will have to check the code Nov 09 12:36:43 I mean, for natty Nov 09 12:36:55 rsalveti, see the jasper-rewrite spec ... Nov 09 12:37:04 we will if upstart allows it Nov 09 12:37:10 it has to happen there Nov 09 12:37:39 yeah, it's hardware specific =\ Nov 09 12:37:52 for x in $(cat /proc/cmdline); do Nov 09 12:37:53 case ${x} in Nov 09 12:37:53 console=tty[A-Z]*) Nov 09 12:38:03 so the kernel patch shouldnt matter Nov 09 12:38:16 but its essential to do it before first boot Nov 09 12:38:18 cool Nov 09 12:38:23 yep Nov 09 12:39:25 i found it a little bit creepy that the first boot modified the system and then rebooted. what does it do? Nov 09 12:40:07 all things the debian-installer would do otherwise Nov 09 12:40:37 bernard_: mainly resize the sd card Nov 09 12:41:40 ah, fair enough Nov 09 12:42:26 it also sets up the board specific stuff Nov 09 12:42:59 it will change a lot in next release, but the resizing will have to stay Nov 09 12:44:03 oh, so an interesting bug report - when I dd'd the image to my SD card, the ext3 partition wasn't where it was supposed to be. Nov 09 12:44:21 running "file - < /dev/sdb2" just reported "data" Nov 09 12:44:43 i ended up running mkfs and copying the files in manually from the image. Nov 09 12:45:23 thats why the resizing bit is there Nov 09 12:45:34 you shouldnever tinker with it after dd Nov 09 12:45:38 the image didn't even boot though, because the partition table was pointing to the wrong place. Nov 09 12:46:21 i'm not sure what it was about the SD card, but i can dig deeper tomorrow morning (in about 9 hours, here) Nov 09 12:47:12 its a 4G or bigger card, right ? Nov 09 12:47:26 4G SD card, yep Nov 09 12:47:39 * ogra_ac only has heard of probs if people didnt stick to the minimal card size Nov 09 12:47:57 dd didn't complain, cmp against the ungzip'd image matched. Nov 09 12:48:21 i didn't dig much deeper at the time, but i can have another look in the morning. Nov 09 12:48:26 oh, you gunzipped Nov 09 12:48:32 rit, you said dd ... Nov 09 12:48:57 just use zcat ;) dont waste your diskspace Nov 09 12:49:08 to write the image, i did 'zcat foo.img | dd of=/dev/sdb'. to do a compare, i had to unzip, and then did 'cmp foo.img /dev/sdb' Nov 09 12:50:47 thats completely worng Nov 09 12:50:57 which install doc did you follow Nov 09 12:51:54 https://wiki.ubuntu.com/ARM/OMAP Nov 09 12:52:57 i was following what i normally do for writing images to media, not the instructions, sorry! Nov 09 12:53:13 is there really a difference between piping to dd of=/dev/sdb and using > /dev/sdb ? Nov 09 12:54:08 no idea, i never tested, but the instructions on the wiki were tested by many people Nov 09 12:54:27 the wiki also says that if the > method doesn't work, to gunzip and dd instead. Nov 09 12:54:35 we usually do quite some QA before releasing something ;) Nov 09 12:55:14 hrm, wo added that Nov 09 12:56:47 maybe there is something crazy about my sd card. i really can't see why they should differ. Nov 09 12:57:04 as i said, i'll reproduce it tomorrow and find out where exactly the image went. Nov 09 12:58:48 will anyone here be awake in 9 hours? ;) Nov 09 12:58:56 berco: sure Nov 09 12:58:58 ops Nov 09 12:59:02 bernard_: sure Nov 09 12:59:59 cool. so i'm not in the craziest timezone! Nov 09 13:01:46 heh 9 hours == business hours Nov 09 14:50:11 ogra: hi! Nov 09 14:50:26 ogra: can you please enable natty on our PPA? Nov 09 14:50:27 hey berco Nov 09 14:50:28 jayabharath, do you have any beyond repair Panda boards? I need something to test size and spacing with. Nov 09 14:50:45 berco, did you ask in #launchpad ? Nov 09 14:50:51 jayabharath, completely dead is perfect if you have any. Nov 09 14:50:59 * ogra_ac doesnt know why that wouldnt happen automatically Nov 09 14:51:06 Yes I think we have a couple sitting in our office... will 1 or 2 board do? Nov 09 14:51:10 ogra_ac: I think I did last week but never got a real reply Nov 09 14:51:29 ogra_ac: maybe everyone was at UDS/ELC... Nov 09 14:51:44 or plumbers :) Nov 09 14:51:48 jayabharath, 2+ is a perfect amount of boards, 1 is better then nothing. :-) Nov 09 14:52:08 ogra_ac: I thought it was you. Shall try again on #launcpad? Nov 09 14:52:23 jayabharath, 2+ allows me to test spacing for the 19" rack case Nov 09 14:52:27 ask for a LOSA (launchpad admin) Nov 09 14:52:30 I have 1 in my office. I do know we burned one more board recently.. so we can give 2 boards to you. How can we get them to you Nov 09 14:52:53 jayabharath, I am more then happy to drive over to the TI offices to pick them up Nov 09 14:53:07 ok great. Nov 09 14:53:54 jayabharath, thanks, that will really help me with the layout of the 19" case and cable assemblies, just let me know when to come get them. Thanks again. Nov 09 14:54:58 I will get my hands on the boards this morning and will email you. You can drop by anytime after 3.30pm today Nov 09 14:55:20 jayabharath, perfect, wow it's coming together nicely Nov 09 14:55:34 nice Nov 09 15:45:35 ogra_ac: told you that they never tested qt neon autodetection on a non neon capable machine Nov 09 15:45:43 it's now an upstream bug Nov 09 15:45:59 even affecting natty Nov 09 15:46:04 yeah Nov 09 15:46:20 good that this also affects meego Nov 09 15:46:22 do you think that will be solved soon with a suitable patch for maverick ? Nov 09 15:46:43 else i'd prefer to do the static build again Nov 09 15:46:47 let's wait until tomorrow's eof Nov 09 15:46:51 *eod Nov 09 15:46:54 like if it takes months it doesnt help much in maverick Nov 09 15:46:59 k Nov 09 15:47:18 if not, at thursday we just push the disabled neon version for maverick Nov 09 15:47:53 well, i'm fine even waiting another week or so Nov 09 15:47:54 the bug is critical, but then I saw that most qt bugs are critical hehe :-) Nov 09 15:47:58 just not months Nov 09 15:48:18 ok then, let's wait at least this week then Nov 09 16:59:07 davidm: ping Nov 09 16:59:29 prpplague, good morning Nov 09 16:59:36 davidm: greetings earthling Nov 09 16:59:43 by the by the AC100 is 19 V Nov 09 17:00:01 davidm: i haven't seen an email come across by inbox, just curious if you are still working on the project? Nov 09 17:00:07 davidm: ahh interesting Nov 09 17:02:44 Yes you might want to have a look at: https://blueprints.launchpad.net/ubuntu/+spec/other-arm-n-public-panda-ppa-build-cluster Nov 09 17:02:50 davidm: let me guess: 19V 2.1A? Nov 09 17:03:00 Where I have the beginnings of the documentatation Nov 09 17:03:34 Nope 1.5x 30W Nov 09 17:03:41 Nope 1.5xA 30W that is Nov 09 17:03:52 Tiny Nov 09 17:04:18 The wiki page has some ASCI art on the relay schematic Nov 09 17:04:23 Very rough Nov 09 17:04:56 I'm coming by TI today to pickup 2 dead panda boards so I can do sizing and such Nov 09 17:05:33 I will then take some perfboard and make "daughter" cards to size Nov 09 17:05:51 and make up my "10" stack for sizing into the 19" rack case Nov 09 17:06:36 davidm: ok, i'd like to show you something when you come by Nov 09 17:07:00 have a nice rest of day Nov 09 17:07:55 hrw|gone: later Nov 09 17:08:59 prpplague, OK I'll mention that to jay when I come in. Nov 09 17:09:28 davidm: i have something that might slightly impact your physical format Nov 09 17:09:39 davidm: but would make things a little easier to work with Nov 09 17:10:04 I'm not sure how to bias the transistor so the relay will latch and not sure what pin to use to poke it for unlatching Nov 09 17:10:08 prpplague, cool Nov 09 17:10:13 easier ++ Nov 09 17:10:15 :D Nov 09 17:10:30 ogra_ac, !!!! :P Nov 09 17:11:00 * prpplague is still tinkering with a netbook interface kit for the pandaboard Nov 09 17:11:16 prpplague, the transistor must allow latching with the panda board hard off (no power applied) Nov 09 17:11:31 davidm: yea looking over your diagram now Nov 09 17:11:58 prpplague, please be gentile remember its ascii art ;-P Nov 09 17:13:06 davidm: np, let me work through the logic and create a schematic over the next few days Nov 09 17:13:54 OK Nov 09 17:14:19 davidm: here is the relay i suggest using - http://www.mouser.com/ProductDetail/NEC/UB2-3NU/?qs=c5n%252bvICFHOa2OX6vjaAytg%3d%3d Nov 09 17:14:39 davidm: low noise, surface mount and is available in several different coil configurations Nov 09 17:15:39 prpplague, the URL shows coil voltage at 3V will 5V damage it? Nov 09 17:15:51 Or is that the min pullin voltage? Nov 09 17:16:28 davidm: it is available in a wide range of coil voltage configurations, that just happens to be the 3V nominal part Nov 09 17:17:17 davidm: http://www.mouser.com/Electromechanical/Relays-I-O-Modules/_/N-5g31?Keyword=UB2&FS=True Nov 09 17:17:47 Cool, then we are good with 5V, perfect Nov 09 17:18:11 YEs, I like that relay, nice Nov 09 18:00:23 ogra_ac: the problem is not only the kernel tree development, it's more the long term maintenance of it Nov 09 18:00:38 I know the linaro folks will integrate most ubuntu sauce and stable fixes Nov 09 18:01:03 so I believe we'll be ok for the release, but don't know then how the long term maintenance will work Nov 09 18:01:20 as long as the config is fine Nov 09 18:01:35 we might need a different config from what linato wants to use Nov 09 18:01:40 (more drivers) Nov 09 18:01:43 that's ok Nov 09 18:01:49 we'll have some kind of linaro-ubuntu tree Nov 09 18:02:14 and we can use the config we want at the package level Nov 09 18:04:51 right, so security and version alignment is our only issues Nov 09 18:07:56 ok, at least for omap 3 I believe we'll be fine Nov 09 18:08:13 right Nov 09 18:08:20 I'll also help on what I can to make sure we have a good working kernel for it Nov 09 18:08:29 i'm more concerned about possible upcoming platforms Nov 09 18:08:32 my concern is if we get the same kind of request for other hardware vendors Nov 09 18:08:34 yup Nov 09 18:08:36 :-) Nov 09 18:08:48 this shouldn't be a common solution Nov 09 18:09:00 omap 3 will be maintained this way because it's important for the community Nov 09 18:09:16 even demoting omap3 to a community image if we cant get security would be fine by me Nov 09 18:09:26 yeah Nov 09 18:09:29 but not for possible other arches Nov 09 18:09:46 if we have commitment we need to support them Nov 09 18:10:20 sure Nov 09 18:35:35 davidm: ok i follow the ascii art but it needs some cleanup and conversion into a proper schematic Nov 09 18:35:40 davidm: i'll start work on that Nov 09 18:36:42 davidm: have you prototyped the circuit yet? Nov 09 18:42:44 prpplague, nope, I've done the latching side many many times, but first time with a transistor to drop it Nov 09 18:43:31 The prior cases I used a mechanical switch to open the relay Nov 09 18:43:40 different use case Nov 09 18:44:18 davidm: ok, i'll do a prototype setup along with the schematic Nov 09 18:44:25 By the by I'll be at TI about 3PM today Nov 09 18:44:32 dandy Nov 09 18:44:33 I'll call Jay for an escort Nov 09 18:44:38 :) Nov 09 18:45:20 Do you have a suggestion on what GPIO pin might be safest to use? Nov 09 18:46:59 davidm: generally speaking they will be about the same, but i'd want to read over the x-load code for blaze/sdp/panda to see if there are any commonalities that might create problems Nov 09 18:48:02 davidm: i'd probably suggest using GPMC_AD13 as an initial item, which is on the secondary expansion header Nov 09 18:48:20 davidm: this leaves the other ones untouch incase you need to rev the board for other features Nov 09 18:49:12 prpplague, OK that works, thanks. I figure you know the board better then I do. Nov 09 18:50:09 davidm: i should be able to get some quick turn boards for the relay section one week from today Nov 09 18:53:48 I have some prototyping boards here I can breadboard if that helps Nov 09 18:54:16 That was what I was going to do anyway Nov 09 19:01:53 davidm: naw, this will be quick turn boards with the correct footprints and everything Nov 09 19:02:12 davidm: basically flush out any issues with the schematic and such Nov 09 19:03:04 Works **** ENDING LOGGING AT Wed Nov 10 02:59:57 2010