IRC log for #arm-netbook on 20100330

01:05.29*** join/#arm-netbook ibot (ibot@rikers.org)
01:05.29*** topic/#arm-netbook is ARM netbook GNU/Linux hacking - logs: http://ibot.rikers.org/%23arm-netbook
10:34.04mnemocfjp: hi, did you receive your bag from DX yet?
10:34.36mnemocfjp: sorry, it was zecrazytux
10:34.43mnemoczecrazytux: hi, did you receive your bag from DX yet? :-)
10:40.25zecrazytuxhi
10:40.43zecrazytuxI'll go to the post office this afternoon
10:41.03zecrazytuxthey tried to deliver it when I was sleeping, I guess
10:41.09zecrazytux:)
10:43.47mnemoc:D
10:45.25zecrazytuxanyway, shipping takes more time than I thought. A bit less than one month !
10:46.28zecrazytuxfor the first part, as they are splitting it out into three packages, and one is actually for an item that is unavailable
10:57.39mnemocdoh
10:57.45fjpzecrazytux: Did you get your compilation problem sorted out? Did not ring any bells.
11:03.25zecrazytuxfjp: yup, I changed the name of an "ASM loop" I guess (newbie inside)
11:03.57zecrazytuxin mtd/onenand/s3c-onenand.c iirc, there's an inline ASM code
11:04.27zecrazytuxand a loop/target named "read_page", but this symbol is somewhat "already defined"
11:04.34zecrazytuxI changed it to "1"
11:04.44zecrazytuxand it compiled
11:04.57fjpWeird. Not needed here.
11:05.16zecrazytuxmaybe related to the build tools ?
11:05.27fjpNo idea.
11:05.30zecrazytuxasm / linker
11:05.48zecrazytuxI'm using a crosstoolchain based on uclibc and gcc 4.3.4
11:05.57zecrazytuxbuilt with buildroot
11:07.43mnemocto take the chitech a part, what goes first, the screws from the bottom or the kb?
11:10.27fjpDoesn't really matter
11:11.10mnemocany "advise" before voiding my warranty *cough* ? :)
11:12.15fjpI found it all relatively easy.
11:44.17*** join/#arm-netbook lkcl (~lkcl@nat66.mia.three.co.uk)
11:52.57mnemochaving a bad time taking the kb out
11:53.17mnemocand the screws from the bottom seem to not be enough to release the back
11:53.34mnemoc.oO( where is my frelling nail?? )o
11:54.33mnemoc.oO( or however that flat tool like the one using for playing guitar is named in english )o
12:04.19fjpmnemoc: It's been posted to the list early on.
12:05.00fjpYou remove the kbd by lifting it over the 3 clips at the back. Just create a bit more room.
12:05.35fjpThere are 3 more screws below the kbd that need removing before the top will come off
12:06.04fjp(You don't remove the back, but the top with the touchpad.)
12:06.24mnemocic
12:08.51mnemocthat was the spirit of my initial question :)
12:09.24mnemocthis *censored* is thight
12:09.40mnemoctight*
12:09.59fjpThe order in which you do it still doesn't matter: there are no screws holding the keyboard in place.
12:10.41mnemocbut without taking the kb out, you can't take the back out
12:12.29mnemocfinally... kb out
12:14.03mnemocplectrum... funny name
12:16.21mnemocfjp: do i have the take the display out too?
12:17.10mnemocforget that
12:17.12mnemoc:)
12:19.22fjpIf you want to install a webcam: yes ;-)
12:20.34mnemoci'm more interested in a touch layer :)
12:21.49mnemocGREAT! I got both "extras" :D
12:22.05mnemocthe 720p ic and the switch in the sodim
12:22.06mnemocm
12:23.44mnemocit's a pretty cute board indeed
14:31.36mnemocdid yours come with the actual sim card reader or only the hole in the case?
14:43.56fjphas a hole
14:44.38fjphttp://alioth.debian.org/~fjp/log/posts/Debianizing_an_ARM-based_netbook.html
15:01.59zecrazytuxmnemoc: got it
15:02.18mnemoczecrazytux: how is it?
15:03.06zecrazytuxdamn great, regarding the price :)
15:03.36zecrazytuxthe netbook fits perfectly in it
15:03.46mnemocadvisable?
15:04.07mnemocdoes it... protect? :)
15:04.25zecrazytuxmaybe a bit larger than the netbook itself
15:04.48zecrazytuxit does protect as much as a "skin" can :)
15:05.19zecrazytuxdon't expect your notebook to survive to a big jump :p
15:06.27zecrazytuxhalf the width of good quality skins such as my old crumpler one
15:07.16zecrazytuxbut that's something like /15 or /20 the price :p
15:07.40mnemochehe
15:07.53zecrazytuxI'll take pictures if you some
15:07.57zecrazytux+want
15:08.15mnemocsure
15:12.31mnemoci want to buy one bag today, and yours seems to be a good option :)
15:13.30zecrazytuxhow much space does Debian takes on the flash ? I am thinking about running either debian grip (+mainstream then needed)  or a custom rootfs based on busybox, uclibc and a few softwares I may need/want
15:15.47mnemocgrip = squeeze.... requires newer kernel
15:16.56zecrazytuxhmm
15:16.57zecrazytux"Emdebian Grip 1.0 (lenny) released"
15:17.29fjpI have 513MB used for Lenny. Base system + X.Org + bare enlightenment + claws-mail.
15:17.56mnemoczecrazytux: ah, cool :)
15:18.14fjpI've not made any real effort to reduce space. Only removed a few packages and deleted translations of man pages.
15:18.29zecrazytuxhm, great
15:18.52mnemocGrip is busybox based instead of coreutils and it's fat friends
15:19.00mnemocthat has to count for something
15:19.23mnemoczecrazytux: i was sure grip was squeeze-based :)
15:19.26fjpreally doesn't want the extra worry of an external repo...
15:19.38zecrazytuxmnemoc: no
15:19.45zecrazytuxmnemoc: crush is busybox-based
15:19.53mnemoc.oO
15:19.59zecrazytuxmnemoc: grip is based on coreutils as well as Debian
15:20.08mnemocnow i'm puzzled
15:20.19zecrazytux:)
15:20.55zecrazytuxcrush won't be an option after a loooooong time: no armel support, only a few packages are available, etc
15:20.58fjpgrip only really removes unwanted translations and reduces some dependencies.
15:20.59mnemoci had the idea that Crush was a layer upon grip...
15:21.08zecrazytuxnop
15:21.29fjp(or maybe crush; can't keep them separated)
15:22.15zecrazytuxgrip is about 25% smaller than a basic debian installation, but acts the same as Debian
15:22.58zecrazytuxin fact, binaries are the same, only useless files are taken out of the packages
15:23.12zecrazytuxthat's what I understood, indeed
15:23.57zecrazytuxcrush binaries are crosscompiled, completely different packages from mainstream Debian ones
15:25.53mnemocand i wanted to try emdebian in my gdium (mipsel)... damn :(
15:26.15zecrazytuxdunno if mipsel is support
15:26.20zecrazytuxed
15:26.48mnemoccrush "Currently only available for ARM"
15:26.58zecrazytuxwhat about grip ?
15:28.21zecrazytuxI wanted to use buildroot, but it does not support package creation :(... a basic rootfs based on uclibc is easy to get ready, but I would like to be able to add packages easily (even if crosscompilation on another box is needed)
15:28.30mnemoc"arm, armel, i386, amd64, powerpc, mips, mipsel, source."
15:28.39zecrazytuxI won't have much time to have a look at OE
15:28.58mnemoci have used buildroot and OE at work, but i don't feel confortable with any
15:29.08mnemocso i ended up having a custom made build kit
15:29.13zecrazytux:)
15:29.28mnemocbut i really wanted to switch to debian/busybox...
15:35.31mnemocgosh... with all those mails i don't know where to start for making the installation (already moved the switch)
15:38.20mnemocfjp: is it still needed to use qemu?
15:47.39mnemocany idea what the ribbon connector at the left of the keyboard is for?
15:48.14fjpmnemoc: For what?
15:49.40mnemocfjp: to install debian. the wiki reflects that method, and i wanted to know if it's obsolete or i should read it
15:52.49mnemocthere is way too many info in the ML for my stressed brain
16:10.03fjpmnemoc: Just look for my first "Debian Installer" mail.
16:10.35fjpqemu is no longer needed; you can just boot D-I directly
16:10.57mnemocfjp: thanks
16:12.02mnemocstill curious about the 5-wire free connector on the back of the vga controller and the free ribbon connector at the left of the kb's
16:12.31fjpmnemoc: http://lists.phcomp.co.uk/pipermail/arm-netbook/2010-March/000394.html
16:13.32mnemocfjp: i had that exact same mail open in front of me, but still unread. thanks a lot for the pointer and the confirmation! :)
16:15.58mnemocthe amount of mails (and chaotic replies) in the ML make it quite hard to dive in
16:18.53fjpagrees with some of that; I've seen better list behavior
16:20.26mnemocgmail is ugly^3 for long threads
16:20.53fjpWhat's gmail?
16:20.58fjp;-)
16:21.01mnemoc=)
16:47.37zecrazytuxwell, I installed debian base
16:48.22zecrazytuxuse mutt, luke :)
16:57.01zecrazytuxfjp: is your xorg.conf posted somewhere ?
16:57.24fjpNope.
16:57.41zecrazytuxcurently xfbdev does not show up
16:57.42fjpsupposes zecrazytux would like him to
16:57.48zecrazytuxyes, please
17:02.36zecrazytuxdriver "fbdev"
17:02.39zecrazytuxok
17:02.56zecrazytuxhmm, keyboard works
17:03.02zecrazytuxbut that's daaaaammmnnnn slooooow
17:04.43zecrazytuxhm, some keys don't work
17:10.29fjpzecrazytux: sent to list
17:34.13zecrazytuxthx
17:38.35zecrazytuxdon't you find X very slow ?
17:46.41zecrazytuxs/bash/dash and mksh for the user: aaaaaaaaaaaaaah <3 :)
17:56.25zecrazytuxhmm, no driver for wireless :(
17:56.32zecrazytuxI need to flash the kernel, then
18:09.43mnemochow long does it usually take since the G³ to the chinese text of loading the "firmware"?
18:10.28fjp~3 seconds or so?
18:10.53mnemocuhm
18:11.19mnemoci was almost a minute holding :-\
18:11.30mnemocand it's actually still in the G³
18:11.48zecrazytuxsame problem here wanting to flash the kernel
18:12.16zecrazytux3
18:14.04zecrazytuxcrap :(
18:14.25zecrazytuxmaybe the uboot file is needed
18:15.05mnemocany trick to force a power-off pressing something? :(
18:15.26zecrazytuxhold power button for long
18:15.50mnemocover a minute holding it already
18:15.54zecrazytuxoh
18:15.57mnemocand still G³
18:16.01zecrazytuxreset on the back of the netbook
18:16.10mnemocaha! thanks
18:16.15mnemoci forgot that button
18:16.27zecrazytuxdamn, can't flash that fucking kernel
18:16.41zecrazytuxor it takes ages
18:16.49zecrazytuxbut there nothing bout G3 on the screen
18:18.06zecrazytuxhm
18:18.13zecrazytuxred chinese text
18:19.16fjpSo your SD card is not correct.
18:19.47zecrazytuxpk it flashed this time. dunno why but it failed for a while
18:20.23mnemocdamn G³ logo
18:21.01mnemoci can't <reset> either, yuck yuck
18:22.02zecrazytuxWTF ?! I can't boot the kernel :(
18:23.48zecrazytuxI may try to boot a kernel with initrd and dd the kernel to flash :/
18:24.09zecrazytuxooooh
18:24.13zecrazytuxI know
18:24.17zecrazytuxwell
18:24.21zecrazytuxI have an idea
18:24.53zecrazytuxthe chitech kernel haven't conpiled in fonts, and i don't install any dm :p
18:28.35mnemocwonders how long will it take for the battery to get drained while only showing the G³
18:29.22zecrazytuxdid you click quite strongly on the reset button ?
18:29.42fjpYes. You need to feel/hear a click.
18:30.13mnemocthanks! :D
18:30.20mnemocso now it's off
18:30.57mnemoclike when removing the kb, i was been "too gentle"... narf
18:31.22fjpzecrazytux: Yes. That's why I don't really consider the CT kernel an option. On a system with a display I want a console.
18:32.34zecrazytuxfjp: yes, but on a netbook, I *want* a wireless connection
18:32.37zecrazytux:/
18:33.15zecrazytuxisn't there any free driver for that card ?
18:33.44mnemocfb? or xorg?
18:33.55fjpSure there is. Just not in 2.6.24.
18:34.27fjpIt's very simple: we need more people working on the kernel support.
18:35.05mnemoci will
18:36.52zecrazytuxerf
18:37.05zecrazytuxI have no knowledge
18:39.08zecrazytuxhm, ok, rt2x00
18:39.44mnemocthe 3rd edition of love's kernel development guide was rescheduled for june :(
18:40.23mnemoc2nd. edition is great, much easier to digest than "uniderstanding the linux kernel"
18:41.00mnemoc3rd. edition was scheduled for this week... but not anymore
18:42.02zecrazytuxfound in Linux Kernels: from 2.6.24 release still available on 2.6.33 release
18:42.06zecrazytuxhmmm
18:43.32fjpWe need RT2800PCI_SOC
18:45.36fjpWhich was added in .32.
18:45.45mnemoczecrazytux: http://www.amazon.com/Building-Embedded-Linux-Systems-Yaghmour/dp/0596529686 is a good buy too... buying love's linux 2nd while 3rd is to be relased in 2M doesn't make much sense
18:47.36fjp(At least, I think so)
18:49.01zecrazytuxfjp: pci ? it is USB !
18:49.31zecrazytuxanyway I'm building rt2x00
18:49.39zecrazytuxfor our 2.6.24.7
18:49.47zecrazytuxdid you try already ?
18:50.20mnemoczecrazytux: hey, don't get distracted! :) the goal is not to get wifi in .24, the goal is to get the chitech running .33 :)
18:51.27zecrazytuxhuhu
18:51.46zecrazytuxI would like to be able to use my notebook, also :)
18:52.28mnemoc^^
18:53.22fjpzecrazytux: Ah, yes. Anyway, what we need is not available in .24.
18:55.55fjprt2x00: Implement support for rt2800usb (added in 2.6.30)
19:01.28mnemocfjp: D+I should boot fine from a 128MB SD/fat16 card, right?
19:01.37lkclmnemoc: yep
19:02.20mnemoclkcl: thanks
19:02.35lkcli have the GPIO startup sequence for powering on the PCI-e slot.
19:02.56fjpis not 100% sure about fat16
19:03.02fjpTry VFAT
19:03.49fjplkcl: I'm not sure about those mmc changes. I don't see how those could influence the IRQ handling.
19:03.55fjpWill give them a try though.
19:04.26zecrazytuxmine is fat16
19:04.51mnemoczecrazytux: you can't boot either
19:05.02lkclyehh, i'm just beginning at the beginning
19:05.03zecrazytuxI could
19:06.58mnemoczecrazytux: fat16, 1 partition, 3 files from http://arm-netbook.alioth.debian.org/pub/CT-PC89E/debian-installer/ ?
19:07.22zecrazytuxyes
19:08.21zecrazytuxlast time i said i couldnt boot, that was due to the flashed chitech kernel that don't display anything but the g3 logo
19:08.42zecrazytuxit somewhat booted but I did not see anything :)
19:20.41mnemoc:)
19:30.07fjp<PROTECTED>
20:01.49fjplkcl: Just verified that the interrupt really comes in on the IRQ assigned to mmc0.
20:01.55lkclack
20:03.04mnemocany idea why i can get stuck at the G³ when trying to boot the D-I (vfat, 128MB SD, 3 files from your dir)
20:04.17fjpNope. From my experience you should get either a red or blue Chinese test with the firmware boot procedure.
20:04.35mnemocuhm...
20:05.08fjpu-boot may be having problems reading the SD...
20:05.38fjpIt's been working 100% here.
20:07.11mnemocok, i'll try to find another not-SDHC card
20:07.33fjpWhat is the total capacity of the card?
20:08.53mnemoc128MB
20:09.40mnemocand badblocks didn't find any problem on it
20:10.36mnemoci think i have a 4GB SD-only card for the camcorder somewhere
20:24.20mnemocdamn card, it's not even useful to flash a firmware
20:24.28mnemoc2G card worked immediately
20:37.49zecrazytuxfjp: did you try to boot a recent kernel already ?
20:40.06fjpno
20:47.08fjpmay have found the SD card solution...
20:52.44fjpYihaa!
20:52.48fjplkcl: Solved!
20:53.33mnemocuniversal recipe for booting from the SD?
20:54.32fjpNope. The ejecting external SD card fucks with mmc0 problem
20:55.36mnemocreading the thread
21:01.11zecrazytuxhttp://www.handlewithlinux.com/bphone-linux-phone-rotate-screen
21:01.13zecrazytuxwow
21:01.19zecrazytuxexpensive, but amazing
21:08.57lkclfjp: yaaay
21:08.59lkclexcellen
21:10.20fjpNot 100% yet. Seem to have lost mmc0 now. But eject is seen on right device.
21:11.24lkclzecrazytux: yes these things usually are.
21:11.34lkclhehe
21:12.56lkclit's very easy to do expensive.
21:13.13lkclzecrazytux: you sent an email to arm_netbooks a couple days ago
21:13.19lkcli seem to have missed it but meant to reply
21:15.00fjplkcl: Have a look at the mmc related changes here: http://git.debian.org/?p=arm-netbook/arm-netbook.git;a=blobdiff;f=arch/arm/mach-s3c6410/mach-smdk6410.c;h=74dadd1b6f75d776e1c1e3de066d1afb253201a6;hp=ec86f3e21466b85bc2ee3ae0076371ab2404db41;hb=00f370ce61f903ba8697709d58dc459fd65579f5;hpb=9b253055c73de146ed220dc94700f89a1e93c1b0
21:15.04lkclacp
21:15.06lkclack
21:15.09fjpSome of those are the key.
21:15.27zecrazytuxyesterday, iirc, about hackable-devices project to promote the idea ?
21:19.23lkclzecrazytux: yep
21:19.33lkclfjp: you got some keys working as well??
21:19.36lkcl:)
21:19.47fjpwhat keys?
21:20.11lkclhhmid_buttons etc.
21:20.19mnemocthe mappings
21:20.47fjpDamn. Where is mmc0...
21:20.56lkclyes.  ok.  rright.  i see things like +               gpio_set_pin(S3C_GPG6, S3C_GPG6_MMC_CD0);
21:21.01lkclwhich is probably it
21:21.16fjpYep. And the added #if 0 s
21:22.30mnemoc.oO( why in the mach-smdk6410 file? )o
21:24.02fjpThat's not where I'm editing.
21:24.17fjpThat's just what we used as our base
21:24.50mnemoci was sure i had see some pastes from you suing march-ctpc89e
21:25.06mnemocyuck, typos
21:27.43fjplkcl: I can fix mmc1, but lose mmc0...
21:27.46fjphelp!
21:27.49lkclahh :)
21:29.04lkclgimme 1sec... *thinks*... writing a message now, commit with working using mmc0 then branch, then patches to do mmc1-but-no-mmc0 i take a look
21:33.38fjpHmm. Interesting. I now have a state where removing card 1 results in messages that mmc0 is removed with the internal card ID.
21:34.10fjpAh, no. That's what we had all along.
21:35.47lkclfjp: i take a quick look at what ben was referring to, the "permanently inserted" stuff
21:35.52lkclthen i go sleep, it's late for me
21:35.56zecrazytuxethernet is unstable with "our" kernel
21:38.31lkclhmmm, that was CONFIG_MMC_UNSAFE_RESUME, not same thing
21:39.18lkclzecrazytux: the capacitors are not according to spec, it makes the ethernet unreliable, and naijing hacked the dm9000 about, adding extra delays and resets, to cope.
21:39.36zecrazytuxfuckages...
21:39.50lkclsometimes they do initialise or reset _twice_
21:40.03lkclthe first time drains the capacitors but gets things half way there.
21:40.14lkclinside the chip
21:40.25zecrazytuxi've been able to use it for one hour or so but now i can't make it work again
21:40.27lkclthen the second time it works!
21:40.30lkclahh.
21:40.46lkcli saw that happen on chitech as well - it's not just us.
21:41.05lkclexcept you are lucky: for me it was any time the ethernet was disconnected, that was it: reboot time.
21:41.44zecrazytuxgrmbl
21:42.21zecrazytuxhopes the wireless won't be fucked up that way as well
21:43.39fjplkcl: I think you'll have to check what's done in the disassembly.
21:44.01fjpI've tried various combies, but can't find a working one.
21:44.06fjp-> bed
21:44.41lkclfjp: yes.  they have two versions of hsmmc_set_gpio
21:44.51lkclone called s3c_hsmmc0_set_ios
21:44.57lkclanother called s3c_hsmmc1_set_ios
21:44.58lkclwhoops
21:45.41zecrazytux... up 100, down, up 100, down, up 10 !
21:46.25zecrazytuxand no packets...
21:46.31zecrazytux(after a reboot)
21:47.23fjplkcl: Well, at least we have it located now.
21:53.01fjpHmm. That means they must have hacked drivers/mmc/host/s3c-hsmmc.c too...
21:53.23fjpThat is where s3c_hsmmc_set_ios is defined.
22:02.03lkclyes.
22:02.36lkclthe s3c_hsmmc0_set_ios is a complete fuck-up.  direct access to registers (again), bypassing gpio_set_pin etc.
22:03.11lkcl*S3C_GPIO+0xE0 &= ~0xf000000; *S3C_GPIO+0xE0 |= 0x3222222
22:03.14lkclthat sort of thing
22:03.27fjpI think they must somehow force card detection for mmc0
22:03.29lkclbut s3c_hsmmc1_set_ios looks relatively unscathed
22:04.01fjpWhile avoiding linking it to the interrupt.
22:04.38lkclthey're definitely doing S3C_GPG6 -> S3C_GPG6_MMC-CD0
22:04.59lkclso in case 1: they're doing
22:05.35lkclgpio_set_pin(S3C_GPH0, S3C_GPH0_MMC_CLK1);
22:05.55lkclgpio_set_pin(S3C_GPH1, S3C_GPH1_MMC_CMD1);
22:06.11lkclpullups to zero on both
22:06.41lkclgpio_set_pin(S3C_GPG6, S3C_GPG6_MMC_CD0);
22:07.42lkclgpio_pullup(S3C_GPG6, 2);
22:09.57lkclthen... waderfuk?  s3c_gpio_cfgpin and pullup, not set_pin and pullup? beh??
22:10.39lkclok - i leave it, look at it tomorrow.
22:10.43fjpAck
22:33.42lkclnight folks.  3g off for the day.

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