**** BEGIN LOGGING AT Wed Sep 21 02:59:56 2005 Sep 21 03:00:34 RP: see the versionstring in familiar-0.8.3.conf :) Sep 21 03:02:30 RP: I'll poke mickeyl to make weekly snapshots for people to test Sep 21 03:03:36 RP: zecke was working on a dotfile generator for OE Sep 21 03:03:53 RP: so you can view the dependency tree in graphviz Sep 21 03:04:17 I bet it would make a nice big poster Sep 21 03:08:14 koen|slug: Once that happens, a week later you will have a mutlithreaded bitbake of some sort ;-) Sep 21 03:08:44 If I can get the depencency information from it correctly anyway Sep 21 03:15:46 03ph5 07org.oe.dev * r4d90ab5d... 10/packages/bluez/ (3 files in 2 dirs): bluez-utils-dbus: work around some alignment issues in hcid Sep 21 03:19:53 pH5: pull/merge/update before commit and before push/sync Sep 21 03:20:29 pH5: otherwise you generate multiple heads Sep 21 03:23:06 hrw|work: i know, sorry. note to self: do not trust the irc client. Sep 21 03:30:25 RP: thats freaky Sep 21 03:44:17 I don't suppose anyone has an old kernel handy? Sep 21 03:44:22 (on a Z) Sep 21 03:44:50 How old? Sep 21 03:45:08 Noodles: 2.6.13-mmx would do Sep 21 03:45:45 zImage-2.6.13-mm3-spitz-20050917180841.bin ? Sep 21 03:46:04 Noodles: If you don't have it installed, don't worry Sep 21 03:46:23 It's not on my Z; I have 2.6.14-rc1 on it. Sep 21 03:46:37 I do Sep 21 03:46:41 what do you want? Sep 21 03:47:08 lardman|work: I'd just like to know if the hda: hda1 message appears if you insert a CF card, umount it and then mount it again Sep 21 03:48:05 Oh drat, I reflashed the other day, sorry Sep 21 03:48:19 I've got a kernel lying around if anyone wants to flash it though Sep 21 03:48:39 lardman|work: I should have one here somewhere. I'll reflash and experiement... Sep 21 03:49:13 I just need to know if we're looking at any kind of regression :-/ Sep 21 03:50:05 I must admit I've not had any troubles mounting the CF card with 2.6.14-rc1 though... Sep 21 03:51:35 lardman|work: You're not using the latest udev... Sep 21 03:53:50 Hmmm. a spitz updater.sh does not work well on c7x0 :-/ Sep 21 03:55:47 RP: I built this on 19/09/05 @14:44, not sure what that means by way of udev Sep 21 03:58:40 lardman|work: I've made a few changes to udev since then. 2.6.13-mm1 shows the same behaviour... Sep 21 04:01:12 RP: Hmm, but I don't have any troubles anyway, so now I'll have even fewer? ;) Sep 21 04:01:43 lardman|work: As things stand, the recent udev will go into a mount/umount loop ;-) Sep 21 04:02:06 Hi there - I'm just trying my first oe build of a gpe-image and the kernel compile is failing Sep 21 04:02:49 I get these messages in the log: Sep 21 04:02:49 ccache arm-linux-gcc-2.95 -D__KERNEL__ -I/home/johannes/zaurus/build/build/tmp/work/openzaurus-sa-2.4.18-rmk7-pxa3-embedix-r22/linux/include -Wall -Wstrict-prototypes -Wno-trigraphs -Os -mapcs -fno-strict-aliasing -fno-common -fno-common -pipe -mapcs-32 -march=armv4 -mtune=strongarm1100 -mshort-load-bytes -msoft-float -S -o constants.h.tmp.1 getconstants.c Sep 21 04:02:49 arm-linux-gcc-2.95: No such file or directory Sep 21 04:02:55 ~zauruskernels Sep 21 04:02:57 i heard zauruskernels is By default Zaurus use 2.4.18 or 2.4.20 and you need GCC 2.95.3 to compile them - all is described in OE wiki: http://openembedded.org/cgi-bin/moin.cgi/ZaurusKernels There is also work on 2.6: http://openembedded.org/cgi-bin/moin.cgi/ZaurusKernels26 Sep 21 04:03:05 johbrahms: read teh zaruskernels wiki Sep 21 04:03:13 ok Sep 21 04:03:15 thanks Sep 21 04:23:14 hi zecke Sep 21 04:23:35 hail zecke Sep 21 04:24:02 hey Sep 21 04:25:13 Great - it seems to be working now. Thanks again. Sep 21 04:25:53 can uclibc use EABI too? Sep 21 04:29:03 hi zecke Sep 21 04:39:12 zecke: did the tabcompletion work on your assignment? Sep 21 04:57:05 RP: I think I found the udev problem Sep 21 04:59:30 RP: or a possible cause anyway, my Z oops though :-( Sep 21 05:07:55 XorA: I think its a kernel bug Sep 21 05:08:09 XorA: ide_cs shouldn't set the removable flag for CF memory cards Sep 21 05:08:19 XorA: but I'm interested to hear what you found? Sep 21 05:08:41 RP: udev seems to not like it when mount /dev/hda fails Sep 21 05:09:15 RP: but now the whole system is acting differently than before oops :-( Sep 21 05:09:44 RP: why not? it is removable Sep 21 05:10:47 XorA, hrw|work: See http://www.rpsys.net/openzaurus/temp/ide_cs_removable.txt Sep 21 05:11:19 I'm preparing that for LKML. I need to patch and test this fixes it first though Sep 21 05:12:49 ah.. Sep 21 05:12:50 ok Sep 21 05:12:52 lunch Sep 21 05:15:05 03ccsmart 07org.oe.dev * r26c9a4c6... 10/packages/bogofilter/bogofilter_0.96.0.bb: bogofilter: Adding RDEPENDS Sep 21 05:16:08 RP: my theory is bogus, my ooops must have given false results, Ive stuck debugging all through the script and cannot repeat Sep 21 05:16:36 RP: you just get a continuous loop of add /dev/hda1, remove hda1 Sep 21 05:44:09 RP: wont this bug trigger on USB disks plugged into Z's as well? Sep 21 05:51:30 XorA: It depends how their driver works Sep 21 05:54:19 RP: if the scheme youve posted in that email is true thats an impossible to solve loop :-) Sep 21 05:54:48 XorA: If the removable device correctly indicates media_change events, it won't loop Sep 21 05:55:26 RP: ah, its supposed to only rescan partitions when a disk change is detected, not at mount? Sep 21 05:56:02 XorA: Correct Sep 21 05:56:25 XorA: but ide_cs is marked as removeable and with a null media_change function Sep 21 05:56:40 RP: ah I understand now, nasty Sep 21 07:14:25 hi guys Sep 21 07:14:38 hey Bernardo Sep 21 07:15:11 hi koen, how are things? Sep 21 07:15:59 very good! Sep 21 07:16:06 and how are things on your end? Sep 21 07:17:30 ok, I'm now finishing reinstalling my desktop machine, now in pure 64bits... Sep 21 07:17:43 will see if now I can get monotone client to work there Sep 21 07:24:24 RP: there is some duplication between udev.rules and devfs-udev.rules in udev-0.70-r3 Sep 21 07:52:33 pH5: Familiar needs devfs naming apparently. I've just corrected the .bb to make it do what was done previously. If there's a problem, someone with an understanding of what familiar requires will have to fix it... Sep 21 07:54:27 Should EXTRA_CFLAGS add what I give it to the standard CFLAGS if I'm running make with EXTRA_OEMAKE = "-e"? Sep 21 08:01:18 RP: I just noticed that some rules (e.g. input and sound) are in both files Sep 21 08:03:53 pH5: I'm not going to worry about it... Sep 21 08:04:20 The solution might be not to install udev.rules in that case... Sep 21 08:07:58 morning Sep 21 08:10:46 morning chouimat Sep 21 08:33:37 hey kergoth Sep 21 08:34:05 'morning kergoth Sep 21 08:34:08 bleh. overslept. Sep 21 08:34:14 gonna be one of those days, methinks Sep 21 08:34:22 I keep waking up 2-3min before the alarm goes off Sep 21 08:54:45 someone know tool which can generate Changelog in GNU format from monotone repository? Sep 21 08:55:52 I guess your could export the monotone repo to CVS and use cvs2cl Sep 21 08:55:57 s/your/you Sep 21 08:56:48 it is one option Sep 21 08:56:50 hi reenoo|OL Sep 21 09:00:09 hey pb_ Sep 21 09:06:10 current monotone log cleaned from not-so-needed infos has 8800 lines only Sep 21 09:06:53 ok - cu Sep 21 09:11:34 bye guys Sep 21 09:20:32 Anyone here know how to modify the gpe-image.bb file? Sep 21 09:21:37 vi? And more seriously, what do you want to do to it? Sep 21 09:22:44 :) Firstly, I'd like to omit some packages and add some others Sep 21 09:23:08 I'm not sure which bits end up in the image Sep 21 09:23:44 meta-gpe.bb might be a better bet Sep 21 09:24:59 Just remove the bits you don't want, and add in the bits you do Sep 21 09:25:20 OK. How do I know if it will fit? Trial and error? Sep 21 09:25:44 Yes, that's the method Sep 21 09:26:17 So back in gpe-image.bb, what does the export IPKG_INSTALL line control? Sep 21 09:26:53 It tells ipkg what to 'install' in the initrd.bin Sep 21 09:27:12 but as I said, you want to look in meta-gpe.bb, not gpe-image.bb Sep 21 09:27:42 Right. I'll try a few builds... Sep 21 09:28:17 look in that file though, then you'll see the various package names which are included, and then you can just remove them, etc. Sep 21 09:30:06 opie maintainers: the h3600 opie image was >16MB a week ago, has that improved? Sep 21 09:47:39 do I have to (re)move the tmp/rootfs tree between building images? And/or will changes I make in that tree be included in the image? Sep 21 09:49:26 the rootfs dir is rm -rf'd before each image is built Sep 21 09:49:36 so no, they won't be included Sep 21 09:50:08 Thanks! Sep 21 09:50:08 and you don't need to do it yourself either as bitbake does it for you - look in run.do_rootfs (if that's the right name) Sep 21 10:08:06 good morning.... Sep 21 10:09:16 good morning mickeyl Sep 21 10:11:10 mickeyl: any reason why OZ doesn't do weekly public 3.5.4-snaps before release? Sep 21 10:13:58 nothing in particular, no. except the amount of work to make it happen. Sep 21 10:14:27 one script and a night of compiling :) Sep 21 10:16:08 How many would be kept? Sep 21 10:16:16 http://pastebin.com/370091 Sep 21 10:16:27 that's the script I use for familiar Sep 21 10:16:51 koen: hmm, worth a try. Sep 21 10:17:22 and include multimachine.inc in local.conf Sep 21 10:17:34 * raduga would surely snap them up Sep 21 10:18:05 it will build stuff twice since the armv5te machines do some strange stuff Sep 21 10:18:33 *nod* that's expected Sep 21 10:18:52 03mickeyl 07org.oe.dev * rfbf5d9c0... 10/packages/rtaudio/ (librtaudio_3.0.1.bb rtaudio-tests_3.0.1.bb): Sep 21 10:18:52 add librtaudio + tests Sep 21 10:18:52 rtaudio is a highly efficient and portable C++ sound abstraction layer supporting OSS, ALSA, JACK. Sep 21 10:19:02 it will compile glibc twice too Sep 21 10:19:16 since the arch for native packages gets set to armv5te too Sep 21 10:19:26 they will still produce host binaries Sep 21 10:20:00 put twice is a win over 3 to 6 times Sep 21 10:21:25 Why on earth won't ipkg install a package which isn't in a feed (if it can find anything in a feed which is later)? Sep 21 10:22:10 used an absolute path? Sep 21 10:22:24 Yep, I used ./octave-blah.ipk Sep 21 10:22:40 hmm that's not absolute, but it's what i meant. Sep 21 10:22:44 odd then, how Sep 21 10:22:48 is it complaining? Sep 21 10:23:25 I've two versions of octave in a local feed - 2.1.73 & 2.9.3 and I can't make it install version 2.1.73, it always starts downloading 2.9.3; then I kill it Sep 21 10:23:58 this is with my specifying the full filename on the command line Sep 21 10:25:30 It's actually been like this for a long time, which used to be quite nice as I'd give it the name of a package I had saved locally, and it would install something later if it could find it Sep 21 10:30:29 ipkg will always install the latest it has in memory Sep 21 10:30:42 -force-downgrade usually works Sep 21 10:30:50 if not, rm /var/lib/ipkg/* Sep 21 10:31:07 thanks koen Sep 21 10:48:00 Hi, I'm making an opie-image using bitbake for the h1940. When it's assembling the image i get several ipkg: Cannot create directory `/usr/lib/ipkg': Permission denied. And then the build fails. Any ideas why it fails. Sep 21 10:48:04 ? Sep 21 10:48:15 I'm using bitbake as a normal user Sep 21 10:48:28 that's not why the build fails Sep 21 10:48:29 i guess you need to look out for ERROR Sep 21 10:50:44 night Sep 21 10:50:48 where to config the opie-image to build with additional language/locales? Sep 21 10:52:02 no, I just issued a bitbake opie-image, no extra fluff Sep 21 11:00:03 hmm Sep 21 11:00:10 how much memory does bitbake use nowadays? Sep 21 11:01:22 joshua_: not that much, an order of magnitude better than before. I'm building gpe-image with 384M RAM Sep 21 11:01:27 joshua_: and hi! Sep 21 11:02:20 Is it crucial that bitbake makes a whole build in one push or is ok to abort it and resume it later? Sep 21 11:02:34 you can abort and resume later Sep 21 11:02:45 The_Ace: bitbake is pretty smart about stopping and resuming Sep 21 11:02:49 hey mreimer Sep 21 11:02:53 hey koen Sep 21 11:03:05 mreimer: that's 2 orders of magnitude :) Sep 21 11:03:07 Is there any way of purging a bad build or do I have to go into the source folder and delete the packages by hand? Sep 21 11:03:13 koen: even better :-) Sep 21 11:03:15 from 1G -> 30M Sep 21 11:03:44 mreimer: I'm using bitbake with 32mb ram :) Sep 21 11:03:52 (and 700MB swap) Sep 21 11:03:53 koen: serious?! wow Sep 21 11:03:57 koen: on your slug? Sep 21 11:04:03 yes Sep 21 11:06:16 mreimer, hello. Sep 21 11:06:22 netbase question: /etc/network/interfaces has "iface usb0 inet dhcp" and "iface usbf inet static" for 2.4. The problem I'm seeing is, how can the other end be configured for dhcp when the interface doesn't exist until it goes up? Should usb0 be configured statically like usbf, or is there something I'm missing? Sep 21 11:06:23 hey joshua_ Sep 21 11:06:42 hi koen Sep 21 11:06:50 still need to get around to shipping spyro his iPAQ Sep 21 11:07:16 mreimer: ifplugd? Sep 21 11:07:44 zecke_: checking it out. Sep 21 11:08:46 mreimer: usb0 should be static Sep 21 11:08:58 with the same config as usbf Sep 21 11:09:15 koen: ok, that's what I usually do. Shall I put a patch in bugs.oe? Sep 21 11:09:30 sure Sep 21 11:09:31 makes me wonder how many people are using 2.6 Sep 21 11:14:00 * pH5 thinks usb networking should use zeroconf Sep 21 11:14:44 that would be nice Sep 21 11:14:46 pH5: that isnt a good idea. then you'd have no way to know what ip your Z has. Sep 21 11:14:54 zeroconf needs to be coupled with a device discovery mechanism Sep 21 11:15:02 like apple's rendesvouz (dns service discovery) Sep 21 11:15:08 that's where avahi comes is :) Sep 21 11:15:12 in* Sep 21 11:15:18 * mreimer whispers to kergoth, "nuke linear-h2200" Sep 21 11:15:24 ideally we'd have a side-band channel that two user-space daemons could talk over. Sep 21 11:15:32 I'm pretty happy with ipv6 and radvd Sep 21 11:15:41 mreimer, what's the status of usability on h2210? Sep 21 11:15:47 namely, runninig from flash Sep 21 11:16:11 booting from sd my girlfriend is able to play games on it Sep 21 11:16:26 joshua_: I haven't tried running from flash, but it should be possible. I have booted LAB from flash using the stock bootloader, but I haven't tried putting a jffs2 fs on it. Sep 21 11:16:26 I just realized that I probably have some spare compactflash cards if it comes down to that, although I'd prefer to run from flash directly. Sep 21 11:17:14 joshua_: if you're up for it, download http://handhelds.org/~mreimer/2nd-stage-linux-nfs.bin and I'll walk you through flashing that using the rescue bootloader Sep 21 11:17:33 or you can dd that directly to the 2nd-stage partition Sep 21 11:17:52 hey, why not. let me grab my USB cradle and some AC power. Sep 21 11:17:56 joshua_: err, maybe not dd directly, because I think that file has the HTC header on it Sep 21 11:18:09 joshua_: are you home from school early, or does school get out earlier now? Sep 21 11:18:15 school gets out at 1:05 on wednesdays Sep 21 11:18:27 cool Sep 21 11:20:06 this latest gpe-image I built is missing user messagebus. Suggestions? Sep 21 11:20:57 joshua_: if you want to flash using the rescue bootloader, grab a CF card and reader too Sep 21 11:21:08 mreimer: this user should be created in the postinst of the dbus package Sep 21 11:21:24 pH5: ah, so on first boot it must have failed? Sep 21 11:21:51 morning Sep 21 11:22:06 I do not have a CF reader, but let me find a CF card. Sep 21 11:22:41 joshua_: do you know if there's some kind of wince utility that would let you dd a file to CF? Sep 21 11:22:48 mreimer: i think so. Sep 21 11:23:10 worst case, I have to use HaRET and NFS Sep 21 11:23:23 can I just do it all over USB? Sep 21 11:24:12 joshua_: if you can boot linux, we can dd it directly, but I don't know how to have the rescue bootloader pull data from USB (just CF) Sep 21 11:24:47 okay, well, let me see if I have a spare CF that's not being used in monstertruck (the EPIA) Sep 21 11:26:52 hm, that's something I haven't seen in a while Sep 21 11:26:54 my stowaway Sep 21 11:27:01 but where is the CF... Sep 21 11:27:02 is that a keyboard? Sep 21 11:27:08 yah. Sep 21 11:27:11 got it for my h3600. Sep 21 11:27:33 dunno where the h3800+ adaptor is for it. Sep 21 11:28:10 aldot * r11538 busybox/ (5 files in 2 dirs): - add applet nohup(1) Sep 21 11:28:11 cool Sep 21 11:29:11 djeez Sep 21 11:29:25 when will they stop bloating busybox ;) Sep 21 11:29:29 hehe Sep 21 11:30:31 hmm Sep 21 11:30:37 I have a CF card in the EPIA Sep 21 11:30:53 is is possible to do this with no external storage? Sep 21 11:31:04 joshua_: yes, if you boot over nfs Sep 21 11:31:07 okay Sep 21 11:31:12 where do I get the right stuff? Sep 21 11:31:33 alternatively I have an SD card. Sep 21 11:31:46 (no SD reader, though -- the SD reader on my tablet is not supported by Linux) Sep 21 11:32:03 joshua_: you can start by getting linux booting: http://handhelds.org/~mreimer/haret.exe and http://handhelds.org/~mreimer/nfs/ Sep 21 11:32:16 joshua_: I'll build a new nfs kernel Sep 21 11:32:27 alright. let me see if I can figure eout how to get all the data out of WinCE Sep 21 11:32:41 * joshua_ removes mp3s and fires up iPAQ Backup Sep 21 11:34:13 * joshua_ watches iPAQ Backup spin Sep 21 11:34:49 * joshua_ gets bluetooth adapter, since it's probably easier to do that than to set up synce. Sep 21 11:35:34 joshua_: you don't need synce, just usbnet Sep 21 11:35:44 to get haret onto it, I do Sep 21 11:35:48 ah yes Sep 21 11:35:59 and to get the backup stuff off it, I do Sep 21 11:36:06 (unless we now support iPAQ File Store) Sep 21 11:36:14 joshua_: no Sep 21 11:36:32 * joshua_ clicks kbluetooth icon Sep 21 11:36:34 * koen doing a bitbake -k world right now Sep 21 11:36:46 koen: what does that do? Sep 21 11:36:59 build everything and not halt on errors Sep 21 11:37:02 ah Sep 21 11:37:10 although I did rig the build a bit Sep 21 11:37:28 I removed qmake-native to stop any qte stuff from getting in the way Sep 21 11:42:18 one of the hacks I plan to do is see if I can teach bluez to stream oggs Sep 21 11:42:31 joshua_: cool Sep 21 11:42:48 joshua_: would you like jffs2 built into the kernel? I'm building a new zImage for you Sep 21 11:42:57 sure, why not. Sep 21 11:42:59 thanks Sep 21 11:44:15 fsck... I have backup of my blog database only at work ;( Sep 21 11:45:25 koen: you also disabled putty iirc :) Sep 21 11:48:35 man, wince's bluetooth stack is just abysmally unstable Sep 21 11:48:51 joshua_: do you have the bt update from hp installed? Sep 21 11:48:56 joshua_: which one? iirc there are atleast 2 incompatible Sep 21 11:49:34 joshua_: if you're running debian you can just install the synce packages and run the configurer (pretty simple) Sep 21 11:50:24 mreimer, I have not run any updates on it. Sep 21 11:50:35 I didn't plan to run wince long enough ;) Sep 21 11:50:39 :-) Sep 21 11:52:58 hmm, synce-kde is broken Sep 21 11:53:20 synce-pls: Unable to initialize RAPI: An unspecified failure has occurred Sep 21 11:53:21 gah, thanks Sep 21 11:53:37 joshua_: give it three or four seconds to init, then try again Sep 21 11:53:52 Also, you'll need to run dccm Sep 21 11:54:18 yah Sep 21 11:54:21 I got it working once upon a time Sep 21 11:54:22 joshua_: I have these installed: synce-dccm, synce-serial, libsynce0 Sep 21 11:58:29 there we go Sep 21 12:00:53 okay, got the backup off of it. Sep 21 12:02:49 * joshua_ wget ipaq-rootfs Sep 21 12:03:40 joshua_: I just flashed LAB and booted it to a prompt Sep 21 12:03:43 Cool. Sep 21 12:04:15 If I have time tonight I'll build LAB. Right now I'm mostly looking for a usable Linux environment that I can take to school with me. Sep 21 12:04:31 joshua_: ok. I can send you my .config to get you up and running quickly Sep 21 12:04:50 alright. Sep 21 12:05:55 hi joshua_, mreimer Sep 21 12:06:00 hi aquadran Sep 21 12:06:01 hi aquadran Sep 21 12:06:38 hrw: http://www.oesf.org/forums/index.php?showtopic=15129&pid=96584&st=0&#entry96584 Sep 21 12:09:22 cu Sep 21 12:09:43 mreimer, okay, I think I have my NFS set up Sep 21 12:09:49 joshua_: great Sep 21 12:10:04 er, I should Sep 21 12:10:08 where do I plop ipaq_rootfs? Sep 21 12:10:38 joshua_: on the host providing usbnet Sep 21 12:10:48 sure, where thereon? Sep 21 12:11:08 joshua_: and then add a line to /etc/exports; mine is: /usr/home/ipaq_rootfs 192.168.0.202(rw,insecure,sync,no_wdelay,no_root_squash,no_subtree_check) Sep 21 12:11:20 /usr/home, or /home?? Sep 21 12:11:23 s/??/?/ Sep 21 12:11:51 joshua_: that's up to you. you can stick it in your home dir if you want Sep 21 12:12:06 joshua_: just need to make sure the args in startup.txt agree Sep 21 12:12:16 ah, cool Sep 21 12:12:24 mickeyl: morning :) Sep 21 12:13:00 I've confirmed the mount.sh going into the mount/umount loop is a kernel bug btw. Just need to convince LKML of that now... Sep 21 12:13:11 RP: good work! Sep 21 12:13:25 Basically ide-cs is not a removable device... Sep 21 12:13:40 RP: did that change in .13 or .14rc1? Sep 21 12:14:01 mreimer: Its been in every 2.6 kernel. My mount.sh scripts just shows it up Sep 21 12:14:56 RP: so the fix is to mark it as non-removable? sounds simple Sep 21 12:15:16 Yes, except its the core ide code that decides in ide-probe.c Sep 21 12:15:34 RP: do you have a patch? Sep 21 12:16:28 mreimer: http://lkml.org/lkml/2005/9/21/118 Sep 21 12:16:36 koen: just beat me :) Sep 21 12:16:47 There are various ways we can do this... Sep 21 12:16:51 thanks koen-secretary :-) Sep 21 12:17:05 Personally, I'd like to rip all that code out... Sep 21 12:17:25 RP: so you have the latest udev working with MMC too? Sep 21 12:17:39 mreimer: it works beautifully with mmc Sep 21 12:17:58 RP: great! Are those rules/scripts in OE yet? Sep 21 12:18:32 koen: Note familiar's 2.6 kernels will have this problem with all udevs as the script gets installed against all versions of udev... Sep 21 12:18:41 mreimer: Its in OE Sep 21 12:18:48 RP: I'll give it a try then Sep 21 12:19:47 mreimer, cool, usbnet all set up. got a zImage for me? Sep 21 12:19:48 reading through forums lessens my faith in people intelligence Sep 21 12:20:14 how hard is it to seperate a buildsystem (OE) from a distro (OZ) in you mind Sep 21 12:20:28 joshua_: I have a LAB kernel built; let me check my NFS kernel Sep 21 12:20:31 okay. Sep 21 12:20:39 * mreimer hugs 'make O=' Sep 21 12:20:41 thanks. Sep 21 12:20:59 RP: a comment from the pdax people on 2.6: "we don't have smp handhelds, so 2.6 is useless" Sep 21 12:21:10 haha Sep 21 12:21:15 * joshua_ looks at koen Sep 21 12:21:16 * joshua_ explodes Sep 21 12:21:17 their loss Sep 21 12:21:23 :) Sep 21 12:21:29 their ignorance astounds me Sep 21 12:21:37 "and we don't need support from any kernel developers; 2.6 is useless" Sep 21 12:22:04 "we don't need support from outside devs, since they might go away" Sep 21 12:22:22 Ciao all Sep 21 12:22:29 that's why pdax is done by 1 russian dude locked inside a closet with a copy of gcc and ipkg Sep 21 12:22:33 hey Pigi Sep 21 12:22:42 hi koen ! Sep 21 12:22:59 BlueZ with USB is pretty nice, I've found. Sep 21 12:23:07 joshua_: http://handhelds.org/~mreimer/nfs/zImage-20050921 Sep 21 12:23:33 joshua_: indeed it is, I got it working on my slug this afternoon Sep 21 12:23:43 I got a "broken" USB BlueTooth dongle from someone (USB thingus snapped off... a soldered cable fixed it). it's pretty plug'n'play once you get it going Sep 21 12:23:50 kbluetooth is pretty sleek. Sep 21 12:23:57 joshua_: I'll upload modules... Sep 21 12:24:03 mreimer, okay. wgetting Sep 21 12:25:05 from an engineering POV bluetooth is utter crap Sep 21 12:25:11 just like usb Sep 21 12:25:31 oh god, it's awful. Sep 21 12:25:46 joshua_: http://handhelds.org/~mreimer/nfs/modules-20050921.tgz Sep 21 12:25:53 mreimer, wgetting Sep 21 12:26:00 joshua_: bbiab (lunch call) Sep 21 12:26:56 enjoy Sep 21 12:38:16 hmm Sep 21 12:42:04 how do I get Sep 21 12:42:08 er Sep 21 12:42:10 how do I get hotplug to work Sep 21 12:43:43 echo "1" > /proc/sys/kernel/hotplug ? ;) Sep 21 12:44:04 hmm Sep 21 12:44:05 and add lots of scripts in /etc/hotplug Sep 21 12:44:12 * koen suggest making it really hot Sep 21 12:44:13 I can't convince it to hotplug networkingn devices. Sep 21 12:44:17 use a paintstripper Sep 21 12:44:19 I tried the stuff in the wiki... Sep 21 12:44:23 koen, or acetone Sep 21 12:45:10 heh Sep 21 12:45:23 I once cleaned my keyboard with acetone.... Sep 21 12:45:44 is there an option I need to enable in my kernel to get hotplug net? Sep 21 12:48:47 NAiL, joshua_: echo "/sbin/hotplug" > /proc/sys/kernel/hotplug Sep 21 12:49:43 udev disables hotplug at the moment. We'll probably end up reenabling that but it would be nice not to have to Sep 21 12:49:46 hey zecke_ Sep 21 12:49:57 RP, you're quickest than me ;) Sep 21 12:50:01 RP, aha. Sep 21 12:50:26 did that, but no luck.. Sep 21 12:50:26 RP: aha Sep 21 12:50:42 but even when I manually invoke net.agent, nothing.. Sep 21 12:51:40 koen: your post about Angstrom is nice ;) Sep 21 12:52:04 RP, I suspect you should cat as told by RP, and then restart the hotplug system with /etc/init.d/hotplug restart or something like this Sep 21 12:52:07 hrw|tv: I couldn't resist :) Sep 21 12:52:38 Pigi, I'll do that. Sep 21 12:52:54 koen: ;) Sep 21 12:53:41 joshua_, so I suspect you should add some debugging to that script and eventually others in /etc/hotplug to see where it fail Sep 21 12:54:19 yeah. looks like net.ifup isn't. Sep 21 12:55:20 bbl Sep 21 12:56:03 in turn, ifup --allow=hotplug usb0 doens't. Sep 21 12:57:45 there it goes. Sep 21 12:59:09 joshua_: back Sep 21 12:59:22 allow-hotplug usb0 did the trick. Sep 21 12:59:38 joshua_: on which end, host or ipaq? Sep 21 12:59:43 host Sep 21 12:59:56 how long should it take from backlight out to system doing something useful? Sep 21 13:00:05 in debian you just put it in /etc/network/interfaces Sep 21 13:00:08 joshua_: a minute or two Sep 21 13:00:17 joshua_: do you have serial hooked up? Sep 21 13:00:18 I see it txing and rxing like a madpenguin Sep 21 13:00:20 no serial. Sep 21 13:00:27 tablet does not have rs232, nor does this cradle. Sep 21 13:00:45 tx tx rx rx rx rx rx rx rx tx tx tx Sep 21 13:00:58 joshua_: I don't remember what that rootfs does; it might not even start gpe itself Sep 21 13:01:13 heh. Sep 21 13:01:23 joshua_: can you ssh in yet? Sep 21 13:01:31 password is? Sep 21 13:01:39 should be '' (empty) Sep 21 13:01:44 fwiw, I get no frame buffer. Sep 21 13:01:56 joshua_: I'll upload a more recent rootfs Sep 21 13:02:44 root:6ZeEzVhue14Kw:0:0:root:/root:/bin/sh Sep 21 13:02:46 whatever that is Sep 21 13:02:55 joshua_: are you in? Sep 21 13:03:04 I am not in Sep 21 13:03:07 I'm looking at that over nfs. Sep 21 13:04:29 er, duh Sep 21 13:04:30 dug, duh. Sep 21 13:04:31 duh, duh. Sep 21 13:04:38 I was sshing to my tablet, which is obviously not an iPAQ. Sep 21 13:04:41 heh Sep 21 13:04:54 wow, monotone has really sped up recently Sep 21 13:04:56 the iPAQ seems to be refusing connections, in any event. Sep 21 13:05:29 joshua_: I suppose the usb traffic has died down? Sep 21 13:05:37 it has. Sep 21 13:05:43 CosmicPenguin: #define recently Sep 21 13:06:01 like yesterday Sep 21 13:06:04 joshua_: pingable too? Sep 21 13:06:07 pingable Sep 21 13:06:27 CosmicPenguin: nightly build? Sep 21 13:06:30 joshua_: are you ssh'ing from the same host that the usb connection is made to? Sep 21 13:06:35 yes Sep 21 13:06:57 I don't think we have any of the h2200 modules loaded yet, if that means anythinig -- popping it out of the creadle doesn't make the blinking charge light go out. Sep 21 13:07:13 joshua_: does this help: /sbin/iptables -I FORWARD -s 192.168.0.202 -j ACCEPT; /sbin/iptables -I FORWARD -d 192.168.0.202 -j ACCEPT Sep 21 13:07:22 joshua_: wouldn't surprise me, since that's an old rootfs Sep 21 13:07:35 I copied your new modules in. Sep 21 13:07:41 I have masq set up... Sep 21 13:07:45 joshua_: what does /etc/modules have? Sep 21 13:08:09 joshua_: is it refusing connections, or timing out? Sep 21 13:08:23 evdev, h2200_audio,snd-uda1380,snd-pxa-uda1380,snd-pcm-oss,h2200_pcmcia,h2200_buttons Sep 21 13:08:25 refusing connections Sep 21 13:08:43 that's surprising Sep 21 13:10:33 joshua_: tarring up a newer rootfs (one I use to boot mmc) Sep 21 13:10:56 okay. Sep 21 13:10:56 ~curse ljp for fscking datebook Sep 21 13:10:59 May the fleas of a thousand camels infest your most sensitive regions, ljp for fscking datebook ! Sep 21 13:12:44 * joshua_ rm the old ones Sep 21 13:12:51 as an aside... Sep 21 13:13:05 mreimer, does your iPAQ make a high frequency noise when it's turned on? Sep 21 13:13:31 joshua_: not that I've noticed, other than a brief plop when resetting sometimes Sep 21 13:14:23 hm Sep 21 13:14:50 joshua_: ETA 3 mins Sep 21 13:18:41 9651.57Hz. Sep 21 13:19:03 joshua_: try http://handhelds.org/~mreimer/nfs/mmcroot.tgz Sep 21 13:19:07 joshua_: 20M Sep 21 13:19:52 * joshua_ wget Sep 21 13:20:01 that including latest modules? Sep 21 13:20:22 joshua_: no, so you'll want to untar the other modules over this set Sep 21 13:20:35 okay. Sep 21 13:21:38 koen: did you figure out the 2.6.1[23]-hh[23] problem with modules and kernel mismatching? Sep 21 13:22:49 yeah Sep 21 13:22:54 stupidity Sep 21 13:23:04 oh, sorry for bringing it up then :-) Sep 21 13:23:13 the hx4700 has a seperate kernel paritition... Sep 21 13:23:19 ah Sep 21 13:25:15 mreimer, seems to be doing something now Sep 21 13:28:27 hm, taps take considerably more force it seems in X than they do with WinCE. Sep 21 13:28:45 joshua_: is it force, or is it the duration? Sep 21 13:29:02 joshua_: so it's working? Sep 21 13:29:15 coudl be duratio Sep 21 13:29:16 n Sep 21 13:30:22 mreimer, hm, works, but no ssh. Sep 21 13:30:42 can you see if dropbear is running, using the terminal? Sep 21 13:30:56 dropbear - not found Sep 21 13:31:28 joshua_: should be in /usr/sbin/dropbear Sep 21 13:33:11 koen: are you up for committing a few bug fixes? 287 and 337 for starters Sep 21 13:33:11 I ran it, although it seems to not continue to run,. Sep 21 13:33:22 joshua_: /etc/init.d/dropbear start Sep 21 13:33:28 that should do it Sep 21 13:33:33 joshua_: on first start all needed keys are generated Sep 21 13:33:44 the keys should already have been generated Sep 21 13:34:07 premature exit: bad buf_incrwritepos Sep 21 13:34:08 mreimer: thats rootfs created from device? Sep 21 13:34:09 when I do dropbear -F -E Sep 21 13:34:36 what's dropbear actually do? Sep 21 13:34:37 hrw|tv: I don't understand; I built this gpe-image and tweaked it a bit Sep 21 13:34:42 it's an ssh server Sep 21 13:34:51 oh. (sigh!) Sep 21 13:35:00 mreimer: thanks Sep 21 13:35:03 np Sep 21 13:35:26 * raduga thought for some mindless reason that it was an opie gui component Sep 21 13:35:33 or something serving something up for the gui Sep 21 13:35:47 mreimer: dropbear create host keys in initscript (when they was not created before) - I added that code before got r/w in OE Sep 21 13:35:55 mreimer, hmm, buttons seem to not quite work as expected. Sep 21 13:36:05 joshua_: the joypad is squirrelly Sep 21 13:36:28 joshua_: maybe the keys aren't mapped properly? Sep 21 13:36:44 possible. Sep 21 13:39:07 joshua_: is it reading the battery properly? Sep 21 13:39:13 looks to be doing so. Sep 21 13:40:11 joshua_: are you ready to try flashing LAB? Sep 21 13:41:14 ljp: you asked me to not superreview your changes Sep 21 13:41:19 ljp: is that still the case? Sep 21 13:41:25 mreimer, almost. I think I may know why it's bailing. Sep 21 13:41:28 read only file system. Sep 21 13:41:43 joshua_: that'd cause problems :-) Sep 21 13:41:57 though I can't figure out why it's mounted RO. Sep 21 13:42:15 * joshua_ grumbles and goes for the chmod a+rw -R * Sep 21 13:42:15 joshua_: is it exported rw? Sep 21 13:42:24 it is exported rw, I think Sep 21 13:42:49 are you logged in as root? Sep 21 13:43:01 I am. Sep 21 13:43:05 it declares an ro fs, though. Sep 21 13:45:12 mreimer, are you using the kernel server, or the userland server? Sep 21 13:45:33 joshua_: dunno. how do I tell? Sep 21 13:45:41 ps aufx | grep nfsd Sep 21 13:45:45 I think Sep 21 13:45:47 joshua_: I have several nfsd Sep 21 13:45:54 ah, you are using userland Sep 21 13:45:57 joshua_: do you use debian? Sep 21 13:45:57 joshua_: also, rcconf shows nfs-common and nfs-kernel-server getting started Sep 21 13:46:01 koen, I am using debian Sep 21 13:46:07 mreimer, okay, you are using kernel? :confused: Sep 21 13:46:34 joshua_: ok, there was some chatter on the mailinglist about FC messing things up (as usual) Sep 21 13:46:47 heh Sep 21 13:47:22 heh, I restarted nfsd, and it blew up. Sep 21 13:47:26 stale NFS file handle this! *reset* Sep 21 13:51:58 whee, successfully connected Sep 21 13:52:15 joshua_: what was the problem? Sep 21 13:52:50 read onlyness. Sep 21 13:52:59 on host? Sep 21 13:53:04 umask? Sep 21 13:53:07 when untarring Sep 21 13:53:17 when I set up kernel instead of user nfs server, it started working. Sep 21 13:53:24 great Sep 21 13:53:38 joshua_: and it's >10 times faster too Sep 21 13:53:43 joshua_: does armboot accept kernel command line arguments? Sep 21 13:53:48 still hugely slow :P Sep 21 13:53:57 yeah Sep 21 13:53:57 mreimer, it doesn't, but it could be hacked to do so easily. Sep 21 13:54:09 once we get LAB flashed onto this, we'll go through it and see what we can do. Sep 21 13:54:17 joshua_: are you ready to try? Sep 21 13:54:22 go for it! Sep 21 13:54:27 I have JTAG :) Sep 21 13:54:29 :-) Sep 21 13:54:30 well Sep 21 13:54:32 I don't, yet. Sep 21 13:54:33 but I could. Sep 21 13:54:35 yeah Sep 21 13:56:36 joshua_: I'll upload a new shim + LAB... Sep 21 13:56:46 joshua_: in the meantime, modprobe hamcop_nand and see what you get Sep 21 13:57:01 mreimer, wow, the touchscreen is beautifully linear. good work! Sep 21 13:57:06 thanks! Sep 21 13:58:29 joshua_: get http://handhelds.org/~mreimer/2nd-stage-linux-nfs.bin Sep 21 13:58:39 I do see the jitter to which you alluded. I think that's a fact of life with touchscreens, though. Sep 21 13:59:10 wgotten Sep 21 13:59:38 joshua_: you might try editing /usr/share/tslib/ts.conf-h2200 and uncommenting the dejitter_h2200 line Sep 21 14:00:09 anyone know a well written svn phrasebook for cvs users? Sep 21 14:00:14 joshua_: that enables a dejitter tslib module that I cobbled together from aric's hx4700 implementation based on jamey's description Sep 21 14:00:15 something like a single page cheatsheet Sep 21 14:00:28 joshua_: modprobe hamcop_nand Sep 21 14:00:35 modprobed. Sep 21 14:00:46 gb2, most of the commands are the same Sep 21 14:00:57 but take a look at the subversion book. Sep 21 14:01:11 yeah, i know most of it is just s/cvs/svn/ Sep 21 14:01:12 joshua_: you should have four partitions: first-stage, rescue, second-stage, and the rest Sep 21 14:01:15 right Sep 21 14:01:22 just looking for something to hand out to a small group of people Sep 21 14:01:44 joshua_: for jollies, dd if=/dev/mtdblock2 of=2nd-stage.bin Sep 21 14:02:20 * joshua_ modprobe mtdblock Sep 21 14:02:24 joshua_: on mine, MD5 (2nd-stage-wince.bin) = 1c79cd1c533795ef8253f7637aa52143 Sep 21 14:02:35 joshua_: though I'm running ROM 1.1 Sep 21 14:02:47 1548797ea4b1d5eff59daf71c5df2e9d 2nd-stage.bin Sep 21 14:03:04 now DD the other way? Sep 21 14:03:09 joshua_: yes Sep 21 14:03:20 using 2nd-stage-linux-nfs.bin? Sep 21 14:04:20 doing so Sep 21 14:04:30 done Sep 21 14:04:39 joshua_: read it back and check md5 Sep 21 14:04:45 good night Sep 21 14:05:03 mreimer, probably not valid; it didn't fill up the block device. Sep 21 14:05:22 joshua_: ah yes Sep 21 14:05:24 joshua_: cmp? Sep 21 14:05:26 cmp -x Sep 21 14:05:37 cmp not found Sep 21 14:05:59 joshua_: hexdump -C file | head Sep 21 14:06:09 do we have PCMCIA support in linux yeT? Sep 21 14:06:17 joshua_: yes, though it's acting funny with udev Sep 21 14:06:24 -C? what's that? :P Sep 21 14:06:28 * joshua_ shoots busybox Sep 21 14:06:34 joshua_: oops Sep 21 14:06:37 looks good. Sep 21 14:06:42 okay. reset it? Sep 21 14:06:47 ok! but you'll need serial Sep 21 14:06:50 plug into bluefire? Sep 21 14:06:59 sure, let me migrate upstairs. Sep 21 14:07:25 * joshua_ watches kbluetooth declare tha his bluetooth adaptor went away, as is expected (he unplugged it, after all) Sep 21 14:09:32 wow, cool! Sep 21 14:09:37 works? Sep 21 14:09:37 it went to sleep and I successfully woke it up! Sep 21 14:09:45 but of course! ;-) Sep 21 14:09:57 dang, that NEVER happened with my h3600! Sep 21 14:10:39 * joshua_ looks for servo +serial+usb doohicket Sep 21 14:10:41 doohickey Sep 21 14:16:24 well, I am impressed. it wakes up faster than WinCE does. Sep 21 14:21:01 joshua_: is serial working? Sep 21 14:21:18 I'm getting that going now. Sep 21 14:22:44 Whats wrong with bb 1.0.1? Sep 21 14:25:08 ljp: hey Sep 21 14:25:37 CosmicPenguin: no mindcontrol applet Sep 21 14:25:55 hiya Sep 21 14:26:46 mreimer, okay, so I just reset, and I'm getting bricksign. Sep 21 14:26:52 koen: really? I thought I saw an entry in the BB bug tracker for that Sep 21 14:26:59 (blue and yellow LEDs.) Sep 21 14:27:06 joshua_: yikes Sep 21 14:27:22 joshua_: is there anything on the screen? Sep 21 14:27:28 CosmicPenguin: I guess they didn't put it in 1.0.1 Sep 21 14:27:30 nope. Sep 21 14:27:33 what's the magic hand gesture? Sep 21 14:27:42 joshua_: contacts + power + reset Sep 21 14:28:12 gee, how unexpected Sep 21 14:28:22 how unexpected! Sep 21 14:29:46 well I guess I'd better get my JTAG in working order. Sep 21 14:30:06 joshua_: did the rescue bootloader come up when you pressed contacts + power + reset? Sep 21 14:30:10 nope Sep 21 14:30:25 joshua_: that is unexpected. maybe hamcop_nand problems Sep 21 14:30:35 possibly. Sep 21 14:31:55 I'd be happier about this if my JTAG connector hadn't self-fuxed. Sep 21 14:32:35 well, we'll see how much I can repair. Sep 21 14:34:15 three broken pins. the question is how many I need Sep 21 14:38:11 okay, let's see if this will do the trick. Sep 21 14:38:41 03rpurdie 07org.oe.dev * r42b169fa... 10/packages/linux/ (2 files): linux-oz-2.6: Add patch to mark ide-cs devices as none removable. Add some pcmcia ids from hrw. Sep 21 14:38:42 joshua_: are you sure you were pressing the right sequence? contacts is second from the left Sep 21 14:40:34 mreimer: http://www.rpsys.net/openzaurus/patches/ide_not_removable-r0.patch is the patch I'm running with Sep 21 14:40:41 thanks RP Sep 21 14:40:46 Functionally equivalent to the one I gave you Sep 21 14:40:52 before Sep 21 14:44:49 koen: is ewi down? Sep 21 14:46:53 zecke_: I can't seem to be able to ssh into it Sep 21 14:47:04 I guess it didn't like bitbake -k world Sep 21 14:47:36 I'll send a reboot monkey into the lab tomorrow Sep 21 14:48:34 joshua_: I just booted cold -> LAB -> ymodem + mmc rootfs Sep 21 14:48:50 a clear sign to go to bed Sep 21 14:49:02 yeah Sep 21 14:49:02 koen: did you try to build mt on os x? Sep 21 14:49:15 zecke_: no, too lazy for that Sep 21 14:49:26 the static one works great Sep 21 14:52:23 I did forget to update my bd on osx for a while Sep 21 14:52:30 db* Sep 21 14:52:59 anyway Sep 21 14:53:03 'night all Sep 21 14:53:07 night koen Sep 21 14:57:48 joshua_: did you dd 2nd-stage-linux-nfs.bin to mtdblock1 or mtdblock2? Sep 21 15:38:34 'night all Sep 21 15:45:05 mreimer, to mtdblock2. Sep 21 15:45:14 joshua_: ok, that's right Sep 21 15:46:11 mreimer, it would be a lot easier to fix if I was not so incredibly shaky as I am. Sep 21 15:46:18 joshua_: you mean your hands? Sep 21 15:46:21 yeah. Sep 21 15:46:23 I dunno why. Sep 21 15:46:30 joshua_: have you eaten? Sep 21 15:47:02 I did just eat. I've been shaky all day, though. maybe I'll go back and try again after this digests. Sep 21 15:47:19 the other issue is that the the plastic melts before I can get the solder on there, which makes very nasty smoke. Sep 21 15:47:24 joshua_: do you have a cold? maybe that's messing up your equilibrium Sep 21 15:47:36 joshua_: heh, maybe the smoke is making you shaky? Sep 21 15:47:58 well, I just started new medications, so it's possible that that's doing it. I dunno. Sep 21 15:48:34 I think that later tonight I'll just grab some sewing needles, fit them into the side of the thing, hot glue them in such a manner that they make a makeshift connector, and then use that as a TAP. Sep 21 15:48:41 :-) Sep 21 15:48:52 after some homework, preobably. Sep 21 15:48:53 won't be very fun in your pocket though Sep 21 15:48:53 probably. Sep 21 15:49:02 well, not hot glue into the connector. Sep 21 15:49:16 hot glue so that they stay in relative position, and can be inserted and removed from the connector. Sep 21 15:49:26 ah, I see Sep 21 15:50:21 joshua_: I've just uploaded the bootShim stuff Matt Evans wrote that I'm using to make the flashable LAB Sep 21 15:51:34 okay. Sep 21 15:51:50 and hamcop_nand works OK for you? Sep 21 15:52:25 joshua_: I've only done read-only stuff with hamcop_nand Sep 21 15:52:37 what were you using to make a flashable LAB? Sep 21 15:52:38 joshua_: but yes, it seems to work Sep 21 15:52:44 joshua_: the rescue bootloader + CF card Sep 21 15:53:25 joshua_: dd if=2nd-stage-lab.htc of=/dev/sda (my CF), insert CF into h2200, contacts + power + reset, then press Power when prompted on screen, then reset into LAB Sep 21 15:54:04 hm Sep 21 15:54:25 oh, what's the 3fs to get into command prompt (not reflash mode)? Sep 21 15:54:50 joshua_: same, contacts + power + reset with serial or usb connected Sep 21 15:54:58 ah, let me try it with USB connected. Sep 21 15:55:09 joshua_: the rescue bootloader should print some text on the screen Sep 21 15:56:20 ...?! Sep 21 15:56:21 now it boots Sep 21 15:56:25 Hello, LAB! Sep 21 15:56:30 joshua_: sweet! Sep 21 15:56:40 "...?!" Sep 21 15:58:51 joshua_: how's it looking? Sep 21 15:59:00 well, I'm talking to it over serial Sep 21 15:59:25 and apparently it had no problem waking up after I gave it a "suspend" command Sep 21 15:59:31 cool Sep 21 15:59:33 although display didn't come back to life: /usr/home/mreimer/linux/kernel26/drivers/soc/mq11xx_base.c:464 Unknown device ID (4d51:0188) Sep 21 16:00:44 huh Sep 21 16:00:47 one more s/r attempt did it, though Sep 21 16:00:56 haven't seen that before Sep 21 16:01:39 you might want to try a newer 2nd-stage I just uploaded, as it has all the LAB modules built-in (I forgot some important ones like fs) Sep 21 16:01:53 okay. Sep 21 16:02:34 I tweaked LAB slightly to not pass any cmdline args to the kernel, so the cmdline built into the kernel gets used instead Sep 21 16:02:38 okay. Sep 21 16:02:46 boot> copy ymodem: nand:2 Sep 21 16:02:46 Copying [ymodem:] to [nand:2] ... Sep 21 16:02:46 Destination [nand] does not exist. Sep 21 16:02:49 this could be a problem Sep 21 16:02:49 do you want an nfs-root kernel? Sep 21 16:02:59 joshua_: uh yeah, I forgot to compile in that module too Sep 21 16:03:06 * joshua_ armboot ymodem: Sep 21 16:03:13 but the good news is that your rescue bootloader is probably still intact Sep 21 16:03:19 so worst case you can use the CF method Sep 21 16:03:27 well, before I armboot ymodem:, I need a kernel to armboot Sep 21 16:03:44 ok, I'll cook one up for you Sep 21 16:03:48 do you have a LAB kernel for me? Sep 21 16:04:01 (with nothing tacked on the front of it, preferably) Sep 21 16:04:27 http://handhelds.org/~mreimer/zImage-LAB Sep 21 16:04:47 >> LAB autoboot starting in 2 seconds - press a key to abort Sep 21 16:04:47 >> Autoboot aborted Sep 21 16:04:56 I didn't press a key. Bug, or feature? Sep 21 16:05:26 probably a bug Sep 21 16:07:11 worked last time I tried it. Sep 21 16:07:41 * joshua_ watches lab slowly transfer over serial Sep 21 16:09:52 mreimer, do you randomly get ñs popping up? Sep 21 16:10:10 only at first boot Sep 21 16:10:24 that I've noticed Sep 21 16:10:46 but then I don't have needles sticking out of my h2200 like some kind of voodoo doll ;-) Sep 21 16:14:33 heh, I don't have any needles in it at the moment. Sep 21 16:14:58 how do I not have cvs on bluefire??? Sep 21 16:15:02 weird, LAB thinks the 'a' key was pressed and so it thinks autoboot should be aborted. Sep 21 16:15:13 (I recently flattened and reinstalled bluefire) Sep 21 16:15:28 mreimer, perhaps the console code should be modified to eat all the characters that are waiting on startup Sep 21 16:15:36 yeah Sep 21 16:15:39 it did work once Sep 21 16:16:54 shimminy christmas, libjava takes a long time to compile Sep 21 16:17:03 do we really need to have it on in gcc? can't we get our java compiling from elsewhere? Sep 21 16:17:54 * joshua_ sudo rsync -av joshua@192.168.2.200:/usr/local/arm/ /usr/local/arm/ Sep 21 16:18:17 * joshua_ watches das blinkenlights on the hub Sep 21 16:18:33 woo, 156msec latency over the wireless. Sep 21 16:19:03 joshua_: btw, don't pull HEAD, just K2-6-12-hh3 Sep 21 16:19:15 joshua_: the 2.6.13 merge is still in process Sep 21 16:19:15 * joshua_ c-c Sep 21 16:20:59 mreimer, I got a good joke for you. Sep 21 16:21:04 yeah? Sep 21 16:21:07 mreimer, CVS Sep 21 16:21:40 :) Sep 21 16:21:58 :-) I think I get it. CVS is the joke, right? Sep 21 16:22:02 yeah. Sep 21 16:22:06 :-) Sep 21 16:22:08 hehe Sep 21 16:22:14 here's another one Sep 21 16:22:17 GNU hurd Sep 21 16:22:22 ooh ooh, i have one Sep 21 16:22:22 lol Sep 21 16:22:24 sourceforge Sep 21 16:22:32 thats more of a sad joke Sep 21 16:22:37 true :( Sep 21 16:22:41 like a clown with a knife in his forehead Sep 21 16:23:09 that reminds me of that totally non-pc shirt from tshirthell Sep 21 16:23:13 rape is only funny if you're raping a clown Sep 21 16:24:16 mreimer, do you have a .config that I can throw into my linux/kernel26 directory? Sep 21 16:24:28 yep: http://handhelds.org/~mreimer/lab-config.20050921 Sep 21 16:24:36 cool. and any patches? Sep 21 16:25:39 joshua_: just one in lab/modules/armboot.c to comment out the code that sets up ATAG_CMDLINE (so it uses the one in the kernel instead) Sep 21 16:27:54 excellent! so how do I stuff this into flash? Sep 21 16:28:24 joshua_: get http://handhelds.org/~mreimer/bootShim-0.2.tgz Sep 21 16:28:55 joshua_: and copy zImage into that directory, or (better) symlink your zImage into that directory, then make Sep 21 16:30:10 cool. Sep 21 16:30:14 so what do I ymodem? Sep 21 16:30:22 htc_bootshim.bin? Sep 21 16:30:26 htc_bootshim.htc Sep 21 16:30:44 ? Sep 21 16:30:50 yes, the bin is the shim + zImage, the .htc is greatwall_header + shim + zImage + greatwall_trailer Sep 21 16:31:04 use .htc with the rescue bootloader Sep 21 16:31:07 okay. Sep 21 16:31:10 and .bin if you're writing directly to flash Sep 21 16:32:01 * joshua_ erasemtd 2 Sep 21 16:32:05 * joshua_ copy ymodem: nand:2 Sep 21 16:32:20 drumroll Sep 21 16:33:53 sigh Sep 21 16:33:58 ok, well kexec-tools is going to need some hacking Sep 21 16:34:13 das buildsystem isn't as good as it should be Sep 21 16:34:16 * CosmicPenguin goes to drink beer Sep 21 16:35:08 mreimer, we need to hack htcloader to say something other than "Windows start" Sep 21 16:35:18 joshua_: ok :-) Sep 21 16:35:30 but yes, no bombout. Sep 21 16:36:01 joshua_: so erase worked? Sep 21 16:36:12 it did indeed. Sep 21 16:36:16 great! Sep 21 16:36:18 (Hello LAB!) Sep 21 16:36:59 hmm, buttons in LAB support would be nice Sep 21 16:37:19 yes Sep 21 16:39:07 and for standalone use, it would be nice to get autoboot working. I added what I think is the equivalent of tcflush(fd, TCIOFLUSH), but that didn't help. Sep 21 16:39:25 what's tcflush do? Sep 21 16:39:36 it's supposed to flush the tty of input and output Sep 21 16:39:41 aha. Sep 21 16:39:56 I'd just go for the read until select says there's nothing left to read(). Sep 21 16:40:00 ah Sep 21 16:41:17 lab suspend resume is pretty cool. Sep 21 16:41:23 :-) Sep 21 16:41:36 joshua_: so are you getting a fs written? Sep 21 16:42:09 I'm not sure if I remember what the options are that I have to pass to mkfs.jffs2. Sep 21 16:42:10 my rootfs has some extra stuff in it that gpe-image does not, so it's probably too big Sep 21 16:42:32 I can put up a pristine gpe-image if you want, but it doesn't work quite right (udev, etc.) Sep 21 16:43:05 2210 has 32mb of flash, right? Sep 21 16:43:07 yes Sep 21 16:43:21 well, let's see if this rootfs that I already have compresses sufficiently. Sep 21 16:46:07 -rw-r--r-- 1 joshua joshua 26666020 2005-09-21 19:43 root.jffs2 Sep 21 16:46:14 58611 /home/ipaq_rootfs/ Sep 21 16:46:26 cool Sep 21 16:46:46 * joshua_ erasemtd -j 3 Sep 21 16:49:55 okay, erasing after copying a different test image, and starting to trasnfer root.jffs2. Sep 21 16:50:28 * joshua_ to homework Sep 21 16:50:34 see you in 26mb! Sep 21 16:50:38 :-) Sep 21 17:26:44 hey baby Sep 21 17:27:14 hrmm Sep 21 17:45:17 mreimer, jffs2 is feeding me lots of magic bitmask not found. Sep 21 17:45:18 Node at 0x00ce3584 with length 0x00000c15 would run over the end of the erase block Sep 21 17:45:18 Perhaps the file system was created with the wrong erase size? Sep 21 17:45:33 joshua_: erase size should be 16KB, right? Sep 21 17:45:48 or is it an ooblayout issue? Sep 21 17:45:57 -e, --eraseblock=SIZE Use erase block size SIZE (default: 64KiB) Sep 21 17:46:01 " Sep 21 17:46:01 ah Sep 21 17:46:02 er Sep 21 17:46:04 "d'oh!" Sep 21 17:46:33 hamcop_nand.c is using MTD_NANDECC_AUTOPLACE to position the ECC bytes where wince does Sep 21 17:46:46 I would guess that jffs2 would work with that Sep 21 17:46:56 how long did it take to flash? Sep 21 17:47:11 grr Sep 21 17:47:37 stupid date command Sep 21 17:47:45 its not tuesday Sep 21 17:48:04 yesterday it was though Sep 21 17:48:06 mreimer, longer than I'd l;ike.. Sep 21 17:48:19 yeah i wish my ipaq would keep proper time by itself Sep 21 17:48:38 maybe i just need to go back to using the 0.8.1 image Sep 21 17:49:02 the 8.2 seems to have broke everything that worked Sep 21 17:49:28 joshua_: I have to go now, but I'm interested in your progress. Email me if/when you get it up/hobbling? Sep 21 17:50:42 sure. Sep 21 18:26:08 hmm Sep 21 18:26:19 monotone seems to be taking its time ... are the servers up? Sep 21 19:45:48 monotone just sucks Sep 21 19:46:41 lol Sep 21 19:46:48 nah monotone is fine Sep 21 19:47:15 and after the next couple updates the speed issue should be gone Sep 21 19:48:23 hmm Sep 21 19:48:32 my problem is connecting to a server tho Sep 21 19:49:44 hmm Sep 21 19:49:55 dominion works ... hate to use it tho .. Sep 21 19:52:08 gah 112 revs Sep 21 20:38:10 ~seen koen Sep 21 20:38:19 koen was last seen on IRC in channel #oe, 5h 45m 14s ago, saying: ''night all'. Sep 21 20:38:21 he is asleep Sep 21 20:38:28 jhm Sep 21 20:38:32 *hmm...aight Sep 21 20:38:45 anybody else here have h3900 series ipaq? Sep 21 20:38:54 i've a 37 Sep 21 20:39:11 have you built a 2.6 kernel for it? Sep 21 20:39:20 2.6 is likly to not work Sep 21 20:39:27 ya... Sep 21 20:39:29 most of the drivers are bad or dont exist Sep 21 20:39:38 well...koen got one to boot on his h5400... Sep 21 20:39:46 basic kernels will boot Sep 21 20:40:00 the 54 use pxa i belive Sep 21 20:40:04 not sa Sep 21 20:40:04 i think his mail said he got gpe up... Sep 21 20:40:08 hm... Sep 21 20:43:14 i guess it helps it i dont mask out what i am trying to build ... Sep 21 20:44:32 oh well...i'll just stick with the 2.4 series right now... Sep 21 20:44:49 do you know when the proposed switch to 2.6 is supposed to happen? Sep 21 20:44:59 yeah 2.4 def has its own issues as well Sep 21 20:45:09 yep... Sep 21 20:45:16 when the drivers get written so 2-3 yrs Sep 21 20:45:29 unless you start writting some :) Sep 21 20:45:33 haha Sep 21 20:45:43 wow...i didn't think it would take that long...shows what I know... Sep 21 20:45:44 :-D Sep 21 20:46:00 well...i'm workin on the wireless driver right now...but maybe after that :) Sep 21 20:46:04 there are very few people interested in doing drivers for an obsoleate chip Sep 21 20:46:24 the wireless chip? Sep 21 20:46:48 the sa Sep 21 20:46:52 ah... Sep 21 20:47:20 i thought h3900 used the pxa chip? Sep 21 20:47:32 specifically i am unsure Sep 21 20:47:47 i am pretty psoitive most of the 5x do tho Sep 21 20:48:20 sorry for asking so many questions...but...(i am trying to learn as much as possible :) ) Sep 21 20:48:29 what is the difference between the sa and pxa? Sep 21 20:48:46 i actually have a h5555... Sep 21 20:49:07 quite a few things from what i ungerstand Sep 21 20:49:12 understand* Sep 21 20:49:25 you could look at the spec sheets for both for specifics Sep 21 20:49:48 just different technologies for the most part Sep 21 20:50:04 ah... :) Sep 21 20:52:48 now that you ask i am curious if the pxa has the same radiation sheilding the sa had Sep 21 20:53:02 eh? Sep 21 20:53:09 since one of its popular uses was as a lowcost radiationhardened device Sep 21 20:54:04 was one of the manufacturing "features" of it... planned or not i am unaware Sep 21 20:54:22 wow... Sep 21 20:54:32 thats definitely a hell of a 'feature' Sep 21 20:54:33 haha Sep 21 20:54:43 yeah Sep 21 22:09:22 hi all Sep 21 22:18:31 how to make bitbake to build opie to use an alternative locale? I want to try chinese other than the default english. Sep 21 22:20:07 anyone? Sep 21 23:15:16 Hello every1, trying to get oe working on this box and it keeps failing on trying to build gnu-config-native-0.1cvs20050922-r3 and i get this msg "1 out of 7 hunks FAILED -- rejects in file config.guess" while trying to apply this patch -> config-guess-uclibc.patch any idea's?? Sep 21 23:35:49 your patches are outof date with the sources Sep 21 23:36:17 abm_y4k, edit your bb files to use chinese Sep 22 01:03:08 hi ! Sep 22 01:04:39 morning all Sep 22 01:42:02 good morning all Sep 22 02:00:50 mornin' Sep 22 02:01:49 morning Sep 22 02:03:18 abm_y4k: you need to alter opie-image and add opie-i18n-CORRECT there Sep 22 02:03:46 who will have rev 2500 in monotone? Sep 22 02:04:50 which rev it was? Sep 22 02:08:33 according to my db we are at 2331 right now Sep 22 02:08:48 2327 here Sep 22 02:09:05 but I pull only .dev Sep 22 02:25:58 good morning Sep 22 02:27:16 morning all Sep 22 02:49:56 I'm just trying to track down a possible bug in Octave (on the Zaurus), at boot it says that the kernel is compiled with NetWinder Floating Point Emulator V0.97 (double precision), is the fact that it's soft-float a modification to this, or is it this? Sep 22 02:51:23 nwfpe is for hardfloat systems Sep 22 02:51:34 softfloat is toolchain related Sep 22 02:51:59 so the nwfpe is still built into the kernel then? Even though we use softfloat? Sep 22 02:52:20 yep Sep 22 02:52:23 yes Sep 22 02:52:51 Okay. One last question - is softfloat a part of GCC, or is it an add-on? Sep 22 02:52:54 since OE!=OZ Sep 22 02:53:07 koen|hx: okay, okay ;) Sep 22 02:57:17 hello, maybe someone could help me with a build problem Sep 22 02:57:18 ERROR: Cannot satisfy the following dependencies for task-bootstrap: Sep 22 02:57:18 | base-files base-passwd busybox initscripts netbase sysvinit sysvinit-pidof tinylogin modutils-initscripts fuser setserial linux-hotplug update-rc.d libc6 (>= 2.3.5+cvs20050627) libc6 (>= 2.3.5+cvs20050627) apm apmd network-suspend-scripts ipkg-collateral ipkg ipkg-link modutils-depmod module-init-tools-depmod Sep 22 02:57:25 Okay, my understanding is this - there's an arm-softfloat patch which is applied to GCC which then allows us to specify TARGET_FPU=soft? **** ENDING LOGGING AT Thu Sep 22 02:59:56 2005