IRC log for #openezx on 20070624

00:52.50*** join/#openezx Fry-kun (n=kostya@h-66-166-233-66.snvacaid.covad.net)
01:12.39Fry-kunhello
01:13.01Fry-kunback to check if anyone is alive
01:13.54Fry-kunstill get some qemu-related error, same place in compilation (generating locale)
02:27.00*** join/#openezx AlexZhang (n=alex@219.133.201.184)
02:42.49Fry-kuncheck, check
02:42.51Fry-kunanyone awake?
03:33.23*** part/#openezx Fry-kun (n=kostya@h-66-166-233-66.snvacaid.covad.net)
05:44.06*** join/#openezx wt (n=wt@70-57-62-197.hlrn.qwest.net)
05:52.36*** join/#openezx koen (n=koen@dominion.kabel.utwente.nl)
06:13.26*** join/#openezx wt_ (n=wt@216-160-159-206.hlrn.qwest.net)
06:28.21*** join/#openezx wt__ (n=wt@71-33-200-227.hlrn.qwest.net)
07:03.01*** join/#openezx pleemans (n=peter@d51A5E76A.access.telenet.be)
07:04.02*** join/#openezx ao2 (n=u@2001:1418:117:0:0:0:0:1)
07:38.51*** join/#openezx goxboxlive (n=goxboxli@176.84-48-210.nextgentel.com)
07:39.33*** join/#openezx dion (n=dion@inhex.net)
07:46.51*** join/#openezx chairuou (n=chairuou@unaffiliated/chairuou)
07:56.04*** join/#openezx chairuou_home (n=chairuou@unaffiliated/chairuou)
08:14.10*** join/#openezx chairuou (n=chairuou@unaffiliated/chairuou)
08:29.58*** join/#openezx LuitvD (n=luitvd@beigetower/luitvd)
08:38.19*** join/#openezx chairuou_home (n=chairuou@58.187.127.130)
08:54.47Hopscotchgood morning
10:42.18*** join/#openezx chairuou_home (n=chairuou@58.187.127.130)
11:51.17*** join/#openezx chairuou (n=chairuou@unaffiliated/chairuou)
12:12.11*** join/#openezx chairuou_home (n=chairuou@unaffiliated/chairuou)
12:34.33*** join/#openezx koen (n=koen@dominion.kabel.utwente.nl)
12:59.15*** join/#openezx LoneStar (n=xinzhen@218.81.28.68)
13:45.17*** join/#openezx chairuou_home (n=chairuou@58.187.127.130)
13:45.21*** join/#openezx drath_ (i=vmaster@p5B07E767.dip.t-dialin.net)
13:46.05*** join/#openezx goxboxlive (n=goxboxli@176.84-48-210.nextgentel.com)
14:17.58*** join/#openezx dion (n=dion@inhex.net)
14:52.59*** join/#openezx drWyrM (n=wyrm@201.2.68.83)
14:53.10drWyrMhi
17:17.33*** join/#openezx LoneStar (n=xinzhen@218.81.28.68)
17:43.21*** join/#openezx flo_lap (n=fuchs@f049171058.adsl.alicedsl.de)
17:48.33flo_laphi all
18:31.34*** join/#openezx goxboxlive (n=goxboxli@176.84-48-210.nextgentel.com)
18:46.47*** join/#openezx dion (n=dion@inhex.net)
18:48.37*** join/#openezx koen (n=koen@dominion.kabel.utwente.nl)
20:26.51*** join/#openezx pleemans (n=peter@d51A5E76A.access.telenet.be)
21:11.41*** join/#openezx flo_lap (n=fuchs@f049171058.adsl.alicedsl.de)
21:46.43*** join/#openezx Fry-kun (n=kostya@h-66-166-233-66.snvacaid.covad.net)
21:55.08drWyrMFry-kun: hi.
21:55.30Fry-kunhiya
21:56.48drWyrMhow is OE? still at qemu?
21:57.02drWyrMyou are Konstantin, from the ML, right?
21:57.08Fry-kunyes, I am
21:57.16Fry-kungot email from Simon (you?)
21:57.26Fry-kunseems to have passed the failure point
21:57.50drWyrMyes, i got. Simon suggestion is not 100% right, i think.
21:58.17drWyrMyou will be able to create a rootfs image, but you will have no locale data.
21:58.20Fry-kununfortunately i don't really understand what the fix does
21:58.30Fry-kunhm..
21:58.35Fry-kunthen what's the real fix?
21:59.10drWyrMit disables locale generation, it is generated with qemu assistance, because it should be generated in the same arch.
21:59.57drWyrMi dont know for sure, the last time i generated a complete rootfs with OE was in 2006. :)
22:00.05Fry-kunwhy does it have to be generated from the same arch?
22:00.13Fry-kungcc cross-compiles quite easily
22:00.26Fry-kundoesn't matter what the target code has to be generated, after all
22:00.40drWyrMand i had to install a gcc3.x on my host machine to get a good qemu compilation.
22:01.05Fry-kunbah
22:01.13Fry-kunthat's annoying..
22:01.29drWyrMit has to be generated from the same arch, have no idea why its arch dependent.
22:01.38drWyrMon what phone are you working?
22:02.53Fry-kuna1200
22:03.01drWyrMhumm..
22:03.29drWyrMwell, openezx still doesnt support the transflash on a1200.
22:03.57drWyrMright now, you dont have a use for a rootfs ;)
22:04.14Fry-kuni see..
22:04.40Fry-kuni'm still not sure how it all works, but seeing how a1200 has barely any support, i want to put something in
22:04.47drWyrMif your intention is to contribute to the project, you already has all the tools needed :)
22:04.51drWyrMgood!
22:05.45drWyrMi can help you. Have you already managed to use gcc-cross from OE to compile the OpenEZX kernel?
22:07.51Fry-kunso far all I managed to do is to get OE sources, more-or-less set up the environment, and get past a few errors in "bitbake virtual/kernel"
22:07.59Fry-kunit's finishing up right now
22:09.07Fry-kuni just restarted from last failure a minute or two before you sent a "hi"
22:09.19drWyrMgood. You know that all you need from OE to compile our kernel is gcc-cross, dont you?
22:09.41Fry-kunnope
22:10.02Fry-kunso.. "bitbake virtual/kernel" is too much?
22:10.26drWyrMwell, i think it is easier to compile the kernel out of the OE environment, but using gcc-cross from OE.
22:10.27Fry-kunbecause that's what was on the wiki page..
22:10.43Fry-kunoh..
22:11.00Fry-kunso.. to make it clear in my head
22:11.10Fry-kungcc-cross is different from normal gcc, how?
22:12.18drWyrMyes, it is different from gcc-host. a cross-compiler is meant to compile code for another arch.
22:12.37drWyrMit is the same gcc altough ;)
22:13.04drWyrMthe easy way is:
22:13.16drWyrM1. download kernel 2.6.21
22:13.40drWyrM2. checkout our svn on the patches dir on the root of the kernel tree
22:15.29Fry-kunyou mean, vanilla 2.6.21?
22:15.34drWyrM3. run make -f patches/Makefile.OpenEZX CROSS_COMPILE=<your path to OE>/tmp/cross/bin/arm-angstrom-linux-gnueabi- zImage-a1200
22:15.39Fry-kunand use normal gcc to build it?
22:15.40drWyrMyes, vanilla 2.6.21
22:15.48drWyrMsee 3.
22:16.02Fry-kunhm.. okay
22:16.11Fry-kunoh, so this is after OE finishes compiling?
22:16.20drWyrMno.
22:16.35drWyrMafter OE populates the tmp/cross dir is enough.
22:16.47drWyrMusually gcc-cross is one of the first things built.
22:16.54Fry-kunoh
22:17.13drWyrMyours is probably ok already. :)
22:17.14Fry-kunbut then.. "bitbake virtual/kernel" is overkill after all?
22:17.45drWyrMno, it is a way to compile the kernel using OE, but i dont recommend it for kernel development.
22:17.47Fry-kunoh, just finished compilation
22:18.00Fry-kunoooh
22:18.02Fry-kunokay
22:18.15Fry-kunis there a command to build just the gcc-cross?
22:18.26drWyrMis useful for compiling everything, but is a lot of extra work when you recompile the kernel 50 times a day :)
22:18.31Fry-kuni'll put it up on the wiki once i have a clear picture in my head
22:18.36Fry-kunor try to, anyway
22:18.46Fry-kunok, gotcha
22:19.34drWyrMbb in 20 mins. have to eat something :)
22:19.50Fry-kuni'll try what you suggested, in the meanwhile
22:20.00Fry-kunsince gcc-cross should be ready now
22:20.04Fry-kunwoohoo
22:20.05Fry-kun:)
22:34.30drWyrMworked?
22:41.14Fry-kunno.. didn't start compilation yet -_-
22:41.37Fry-kungot distracted by food
22:42.02Fry-kunoops
22:42.07Fry-kundid i get the wrong ver
22:42.11Fry-kun2.6.21.5
22:42.58Fry-kunthough .5 should not matter so much
22:43.15drWyrMyes, it matters.
22:43.28Fry-kunoh
22:43.41drWyrMit will fail when 'quilt push -a'
22:43.47Fry-kunooh
22:43.55Fry-kundang
22:44.37drWyrMi forgot to tell you above, after you checkout the patches dir, you do a
22:44.49drWyrM2.5. quilt push -a
22:44.50drWyrM:)
22:45.15Fry-kunhttp://www.kernel.org/pub/linux/kernel/v2.6/linux-2.6.21.tar.bz2 <- this one good?
22:45.27drWyrMyes
22:45.45Fry-kunkewl
22:51.16Fry-kunso the general method is
22:51.16Fry-kun1. get OE
22:51.16Fry-kun2. bitbake virtual/kernel to get gcc-cross
22:51.16Fry-kun3. get vanilla kernel
22:51.16Fry-kun4. apply patches
22:51.17Fry-kun5. build vanilla kernel using gcc-cross
22:51.46Fry-kunbetween 1 and 2, do I want to apply patches to OE kernel, too?
22:53.51drWyrM-2.
22:54.18drWyrM+2. i am sure that there is a way to build only gcc-cross. i just dont know how :)
22:54.37Fry-kunright..
22:55.01Fry-kunbut when not tweaking the kernel, the idea is to use OE kernel, instead?
22:55.03drWyrMwhen compiling with OE it applies the patches automatically.
22:55.15drWyrMyes.
22:55.31Fry-kunexcept it doesn't know where i downloaded the patches to
22:55.42drWyrMin fact it ends to be the same kernel, compiling outside OE just makes development faster.
22:56.00Fry-kunright.. gotcha
22:56.56Fry-kunby the way, what does OE offer other than gcc-cross? why do we need it at all, if we can just compile vanilla kernel like this?
22:56.57drWyrMtried to boot_usb it already?
22:57.02Fry-kunnoooo
22:57.05Fry-kunnot that far yet
22:57.07Fry-kun-_-
22:57.28Fry-kunjust now finished extracting the kernel
22:57.31drWyrMwell. kernel is like 2% of the code for a complete new firmware.
22:57.41drWyrMOE takes care of the 98% remaining.
22:57.47Fry-kunah okay
22:58.07Fry-kunsomeone told me yesterday that openmoko would probably take over, since it builds on top of OE
22:58.17Fry-kunso i
22:58.40Fry-kunso i'm inferring that OE is basically like the OS and openmoko is like the phone "application"
22:58.46Fry-kunthe GUI, as it were
22:58.47drWyrMopenmoko is the most likely to be used. but OE can build other kinds of images too.
22:58.49Fry-kunis that right?
22:59.12drWyrMby OE do you mean OpenEmbedded?
22:59.14Fry-kunyes
22:59.19drWyrMor OpenEZX?
22:59.29Fry-kunno, OpenEmbedded
22:59.52Fry-kunit would probably be better to shorthand OpenEZX as OX
23:00.34drWyrMOE is the build system for everything, angstrom is the name of the distribution, OpenEZX is our fork of the kernel. The OS is GNU/Linux.
23:00.48drWyrM:)
23:01.56Fry-kunin my head: only the kernel itself is "linux" - all the programs around it make up a "distribution" which I also refer to as an "OS"
23:02.00drWyrMOpenMoko, GPE, Opie, are GUI environments.
23:02.24drWyrMthats why i said GNU/*
23:02.33Fry-kunlike X/KDE, right?
23:02.50Fry-kunhehe ok
23:03.50Fry-kunso.. (sorry to repeat).. OE is a distro?
23:04.03Fry-kunand OpenEZX is only the kernel
23:04.15Fry-kunor even just a patch set for the kernel
23:04.29drWyrMOE is a build system. It builds distros.
23:04.38drWyrMAngstrom is the distro.
23:04.42Fry-kunaaah
23:04.44Fry-kunokay
23:04.51Fry-kungotcha..
23:05.48drWyrMhave you ever did kernel development?
23:05.53Fry-kunnope
23:06.13drWyrMyou will like it ;)
23:06.16Fry-kunhaha
23:06.18Fry-kunwe'll see
23:06.26Fry-kunquite possibly
23:06.41drWyrMhope that you have enough time to contribute regularly.
23:06.43Fry-kuni never got into it since most of the common hardware I use already works perfectly well
23:07.02Fry-kunyeah, hope so too
23:07.34drWyrMpmarques is already working on the pxafb driver to support a1200.
23:07.45drWyrMso, dont care about the green screen.
23:08.01drWyrMi suggest you to start trying to get the TransFlash to work.
23:08.08Fry-kunpxafb == ?
23:08.18Fry-kunoh, is it the screen?
23:08.25Fry-kun+driver
23:08.30drWyrMi tried it, but i dont have an a1200 to test, so it currently dont work.
23:08.35drWyrMyes, it is the LCD.
23:08.40Fry-kuni see
23:08.43Fry-kunwell, i have one
23:08.57Fry-kunwhich is why i'm interested in writing stuff for it :)
23:09.14Fry-kunoh, back to compilation
23:09.39Fry-kunso i checked out the svn patches
23:10.01Fry-kunis there svn+ssh or anything? or should I just use http?
23:10.18Fry-kunwill probably need something for authentication if i want to check something in...
23:10.25Fry-kunbut i digress..
23:10.44Fry-kunnow i have two directories, linux-2.6.21/ and patches/
23:10.56drWyrMwe will grant you ssh, and svn access as soon as you have something to commit. :)
23:10.57Fry-kunshould patches/ be inside linux?
23:11.22drWyrMyes, on the same level as Documentation/ arch/ drivers/ etc..
23:11.56drWyrMiwbb in 30 mins. cya.
23:12.17Fry-kunk, quilt seems to have worked
23:12.22Fry-kunok
23:14.39Fry-kunuh oh
23:14.56Fry-kunerror in your command line
23:15.17Fry-kuntmp/cross/bin/arm-angstrom-linux-gnueabi- zImage-a1200
23:15.27Fry-kunarm-angstrom-linux-gnueabi- is not a tool :(
23:16.26stefan_schmidtFry-kun: make ARCH=arm CROSS_COMPILE=/path/to/gnueabi zImage-a1200
23:17.02Fry-kunCROSS_COMPILE=/path/to/gnueabi <- what does yours look like?
23:17.40Fry-kunooh, it's just a prefix
23:17.40stefan_schmidtFry-kun: It's just the path to tmp/cross/bin/arm-angstrom-linux-gnueabi-
23:17.47stefan_schmidtFry-kun: exactly
23:17.59Fry-kuni was confused because i thought it'd be a full path :)
23:18.02Fry-kunrunning now
23:18.05Fry-kunseems good so far
23:18.10Fry-kunsome warnings, but that's about it
23:18.11stefan_schmidtfine
23:18.22Fry-kunthanks!
23:18.29stefan_schmidtWarnings are normal. Still much to do.
23:18.33stefan_schmidtnp
23:19.01Fry-kunyeah, i'm sure :)
23:23.13Fry-kunhey, it got built! woohoo!
23:24.28stefan_schmidtFry-kun: Congrats. Now You can try to boot this kernel with boot_usb
23:24.38Fry-kunnow, to try boot_usb
23:24.43stefan_schmidtheh
23:25.37Fry-kunwhoa, it's just one .c file?
23:25.54stefan_schmidtJup, easy code
23:26.08stefan_schmidtJust find the usb device and send message via libusb
23:26.17stefan_schmidtmessages and data
23:26.21Fry-kunneat
23:27.16stefan_schmidtFry-kun: I would like to see it able to doing more stuff, but that is all the EZX bootloader gives us.
23:27.41Fry-kunmore stuff like what?
23:28.12Fry-kunbtw, you don't need to keep writing "Fry-kun:" all the time, it's kind of redundant when nobody else is speaking :)
23:28.30Fry-kunjust a thought ;)
23:29.13stefan_schmidtPersonally I like it as I have a trigger for my nick. Doing a lot stuff besides IRC and like to see if soembody talks with me personally
23:29.40Fry-kunaah ok
23:30.08stefan_schmidt(Bootloader) Loading kernel from SD card, having a real bootloader shell over usbserial, all the neat bootloader stuff. :)
23:30.26Fry-kunaah
23:30.42stefan_schmidtBut at least we are able to upload code and execute it by the bootloader. That'S fine atm
23:30.48Fry-kunheheh
23:30.54Fry-kunok, about to try it
23:31.02Fry-kunwhat's the command
23:31.08Fry-kunjust boot_usb
23:31.16Fry-kunor boot_usb zImage
23:31.17Fry-kunor something
23:31.29stefan_schmidtArgument is the path to the zImage
23:31.36Fry-kunk
23:31.45stefan_schmidtAnd you must be root or change udev accordingly
23:31.56Fry-kunUnable to claim usb interface 1 of device: could not claim interface 0: Operation not permitted
23:31.58Fry-kunoh
23:32.46Fry-kunookay
23:33.00Fry-kunnow i got a jumble of random color pixels on my phone
23:33.05Fry-kunis that it?
23:33.07stefan_schmidtnormal
23:33.25stefan_schmidtAfter some seconds you should see green font kernel messages
23:33.26Fry-kunoh, some are flickering
23:33.42Fry-kunaah okay
23:33.55Fry-kunover 30 sec now..
23:34.03stefan_schmidthmm, that's not normal
23:34.04Fry-kunerr minute
23:34.18Fry-kun:(
23:34.24stefan_schmidtTried a precompiled zImage already?
23:34.28Fry-kunno
23:34.32Fry-kundidn't know there was one
23:34.39stefan_schmidtWould be a good test
23:34.44Fry-kunbah
23:34.47Fry-kunok..
23:34.51stefan_schmidthttp://www.openezx.org/download/zImage-2.6.21-ezx11-a1200
23:34.51Fry-kunphone won't turn off
23:34.56Fry-kununplug battery?
23:35.03stefan_schmidtYou have to remove the battery
23:35.33Fry-kunok whew it's alive
23:36.09stefan_schmidtIt works with the precompiled one?
23:36.45Fry-kunabout to try
23:37.04Fry-kunjumble for 1 second, then black screen
23:37.19Fry-kunstill black
23:37.34stefan_schmidthmm
23:38.16Fry-kunhardware failure??
23:38.21stefan_schmidtGive me a moment to finish another task. 20 min or so
23:38.30stefan_schmidtNo, should be software
23:38.38Fry-kunalright
23:40.14Fry-kuni've suspected before that my phone has a hardware problem because it reboots every once in a while, by itself
23:40.21Fry-kunhope it's not though
23:45.27drWyrMFry-kun: ther kernel opn openezx.org/download is buggy on the backlight driver.
23:45.35drWyrMs/opn/on/
23:46.35drWyrMFry-kun: sometimes, the phone lock-up with a withe screen when booting, happens on other models too. dont know why yet.
23:46.47Fry-kunoh
23:46.49Fry-kuni see...
23:47.01Fry-kunso.. any way of testing it?
23:47.28drWyrMwhat usually solves, is booting it on the original firmware once, and then trying too boot_usb again.
23:47.46drWyrMthe kernel you built is probably better than the version that is on the site.
23:47.50Fry-kunoh, rebooting into flashed firmware
23:47.53Fry-kunthen trying again?
23:47.57Fry-kuncan try
23:48.07Fry-kuner.. i didn't modify it yet
23:48.10drWyrMyes, usualy this fixes the issue.
23:48.11Fry-kunwhy is it better
23:48.51drWyrMare you trying with zImage-a1200 right? do not use vmlinux
23:49.05Fry-kuntrying prebuilt one
23:49.16Fry-kunsame thing - black screen
23:49.32drWyrMit is better because i let a wrong configuration option activated, and ao2 did the release with the buggy code.
23:49.44drWyrMthe prebuilt will always give a black screen
23:49.51Fry-kunaah
23:49.57drWyrMthe backlight driver dont works.
23:50.01Fry-kungotcha
23:50.37Fry-kunshouldn't i be able to see something even w/o the backlight?
23:50.49drWyrMno.
23:51.06Fry-kunHEEEEY!
23:51.08Fry-kunit worked
23:51.12drWyrM:)
23:51.18Fry-kunat least, insofar that i got green text
23:51.20Fry-kunawesome!
23:51.26Fry-kunexcept it kernel-panicked
23:51.52drWyrMit panicked because it is trying to mount the rootfs from the second partition on the card.
23:52.06drWyrMbut we dont support the card yet.. ;)
23:52.13Fry-kunlol ic
23:52.26stefan_schmidtFry-kun: Card support is your job. ;)
23:52.26drWyrMyou can try to upload an mtd-ram image and pass it on the CMDLINE.
23:52.30stefan_schmidtFry-kun: congrats
23:52.34stefan_schmidthi drath
23:52.37stefan_schmidthi drWyrM
23:52.38Fry-kun*yay*
23:52.46drWyrMstefan_schmidt: hi.
23:53.36Fry-kunso.. what's my next step? how do i go about getting the mSD card to work
23:53.50Fry-kuns/mSD/TF
23:54.06drWyrMdownload the 2.4 kernel source from opensource.motorola.com
23:54.09stefan_schmidtTransflash == microSD
23:54.22Fry-kunstefan_schmidt: yeah i know :)
23:54.31drWyrMuse the a1200-mci.patch as reference. You basicaly have to fix this patch.
23:54.42Fry-kunstefan_schmidt: but "mSD" is ambiguous, since it can also mean mini-SD
23:54.48Fry-kunaaah
23:55.02Fry-kunsounds doable
23:55.08stefan_schmidtFry-kun: (mSD) good point
23:56.17stefan_schmidtFry-kun: A good start would be to compare the power on sequenze from 2.4 and 2.6
23:57.14stefan_schmidtFry-kun: Write values into register for power on etc. It's a uninteresting job but should give you an overview over the driver code
23:59.07Fry-kuneek
23:59.10Fry-kunsounds scary
23:59.24drWyrMthe card detect gpio is wrong, this one i am sure, may be a good starting point.
23:59.25Fry-kunprobably not so bad though
23:59.56stefan_schmidtFry-kun: Take your time. This will not happen from alone. Fetch the code, read it, get an overview...

Generated by irclog2html.pl Modified by Tim Riker to work with infobot.