**** BEGIN LOGGING AT Thu Dec 15 02:59:59 2005 Dec 15 03:05:59 morning all Dec 15 03:06:48 hey RP Dec 15 03:07:22 hey RP Dec 15 03:07:29 hrw|work: I just bumped the PR and altered a couple of fstabs so I doubt I've broken base-files Dec 15 03:07:48 hey RP Dec 15 03:09:04 RP: my fault - bitbake hides info that I run out of disc space Dec 15 03:12:21 hrw|work: I have a typo somewhere in the hostap changes though :) Dec 15 03:12:40 RP: hostap-modules does not RPROVIDE kernel-hostap iirc Dec 15 03:13:13 RP: hostap-utils/hostap-daemon should rather RRECOMMENDS hostap then RDEPEND on them Dec 15 03:13:26 I can compile hostap into kernel... Dec 15 03:14:04 kernel-modules should only be RRECOMMENDED Dec 15 03:14:16 koen: why? Dec 15 03:14:31 because I can compile them into the kernel Dec 15 03:14:41 *boom* task do_rootfs failed Dec 15 03:16:46 koen: I changed a hard dependency on hostap-modules into a depenency on virtual/kernel-hostap which is mow provided by hostap-modules and the oz-2.6 kernel so its a move in the right direction Dec 15 03:17:01 yeah Dec 15 03:17:35 ideally OE parsed the defconfig and adds RPROVIDES for each compiled in module Dec 15 03:18:09 If its compiled in, its not a module, but yes :) Dec 15 03:18:36 You can't work that out from the defconfig though Dec 15 03:18:46 In fact its probably near impossible to work out... Dec 15 03:20:25 <_law_> if i insert my wlan card i always have to say ifup wlan0 this is an pcmcia problem or? Dec 15 03:20:56 _law_: Using pcmciautils or pcmcia-cs? Dec 15 03:21:24 I was tempted to blaim udev Dec 15 03:21:35 but udev doesn't handle networking stuff yet Dec 15 03:21:52 :) Dec 15 03:21:53 <_law_> RP, pcmcia-cs - 3.2.8-r24 Dec 15 03:22:31 _law_: Sounds like pcmcia-cs is at fault if the wlan module gets loaded Dec 15 03:22:47 koen: yeah, generating RPROVIDES based on the kernel .config file would be a good idea. Dec 15 03:23:10 pb_: How do you guess the module names from the .config though? Dec 15 03:27:01 14 17:00 < RP> hrw|work: Try forcing MMC_CLKRT to 6 in pxamci_start_cmd Dec 15 03:27:26 RP: MMC_CLKRT default is 8 - lowering can help? Dec 15 03:27:59 I think some code somewhere will change that? Dec 15 03:28:21 RP: writel(host->clkrt, host->base + MMC_CLKRT); Dec 15 03:28:31 RP: its the only one usage of that Dec 15 03:29:17 and host->clkrt is a constant? Dec 15 03:29:22 8 is an invalid value... Dec 15 03:29:27 It should usually be 0 Dec 15 03:30:18 hrw|work: clkrt is changed in pxamci_set_ios Dec 15 03:30:27 yep Dec 15 03:30:43 Just change that line to writel(6, host->base + MMC_CLKRT); and see if it helps Dec 15 03:31:05 ah. Dec 15 03:31:36 <[lala]> hi Dec 15 03:31:55 <[lala]> i'm trying to update my borzoi kernel to latest rc5 Dec 15 03:32:06 hi [lala] Dec 15 03:32:23 <[lala]> but i noticed that for instance hostap-modules don't get recompiled after switching kernel Dec 15 03:32:29 <[lala]> is this correct behaviour? Dec 15 03:32:52 [lala]: We don't use the external hostal-modules anymore Dec 15 03:33:00 We use the verison internal to the kernel Dec 15 03:33:02 <[lala]> ah, ok Dec 15 03:34:15 <[lala]> but in general: if i switch to another kernel-version - should all kernel-dependent packages get rebuilt automatically? Dec 15 03:34:35 [lala]: It would be nice if they did but they don't Dec 15 03:34:49 <[lala]> that's what i guessed ;-/ Dec 15 03:35:20 <[lala]> so i have to find out which packages are involved and force them to recompile by removing stamps? Dec 15 03:36:13 pxamci_set_ios: clock 312500 power 2 vdd 0.21 Dec 15 03:40:19 <[lala]> RP: kernel update via flash method (updater.sh) is the only way to switch kernel on borzoi? or can i install the kernel via ipkg? Dec 15 03:42:42 [lala]: You can only reflash it - you can flash just the kernel though (just don't provide an initrd.bin/hdimage1.tgz file) Dec 15 03:43:40 <[lala]> RP: yep, the last one i already got told. just wanted to get sure, that ipkg is not an option :-) Dec 15 03:44:50 RP: would it be possible to dd the kernel into place from a running system? Dec 15 03:46:40 9 minutes of dd and still going.. Dec 15 03:48:52 * lrg|home reports that Line capture and Line bypass work on Borzoi, will test a Mic tomorrow :) Dec 15 03:55:08 17 minutes and dd still going Dec 15 03:59:24 22 Dec 15 04:01:00 benchmark and test in one :) Dec 15 04:01:32 koen: mmcspeed set to lowest possible Dec 15 04:02:09 hrw|work: this the mmc card that was giving you grief? Dec 15 04:02:39 XorA: yep - that 1G one Dec 15 04:03:05 ibot: calc 1024/(22*60) Dec 15 04:03:23 hrw|work: is it supposed to be a turbo super speed or anything? Dec 15 04:03:31 0.775757576 Dec 15 04:04:22 XorA: if card will work with lowest speed then I'll try faster. if not then it is fscked too much for Z Dec 15 04:04:39 ~google 1024/(22*60) Dec 15 04:05:52 0.5h Dec 15 04:09:19 anyway it goes better then it was Dec 15 04:10:25 01455000 Dec 15 04:19:51 koen: dd won't work on a Z as we have no driver for the Sharp FTL layer Dec 15 04:20:30 hrw|work: Its still going? Dec 15 04:20:35 RP: yep Dec 15 04:20:58 RP: looks like interface speed is a problem for that card Dec 15 04:21:09 Dec 15 12:20:47 c7x0 user.debug kernel: MMC: starting cmd 12 arg 01b2d000 flags 00000009 Dec 15 04:21:43 hrw|work: It probably means we need to switch to a lower speed and retry the request when we get a command timeout error Dec 15 04:21:55 RP: probably Dec 15 04:22:42 hrw|work: Do you want to try some other values between 0 and 6 to see just how slowly it needs to go? Dec 15 04:23:43 RP: thats the plan Dec 15 04:24:16 RP: at 12:37 I'll break and try 2 instead of 6, then 4 Dec 15 04:25:00 hrw|work: ok, that sounds good. If 2 works, don't bother with 4 though Dec 15 04:25:25 RP: ofcourse Dec 15 04:25:33 RP: then I'll try 1 Dec 15 04:25:51 | ERROR: Cannot satisfy the following dependencies for task-bootstrap: Dec 15 04:25:51 | virutal/kernel-hostap Dec 15 04:26:05 hrw|work: I have a patch ready to push for that Dec 15 04:26:06 virutal sounds nice Dec 15 04:26:18 the typo I mentioned earlier ;-) Dec 15 04:26:22 heh Dec 15 04:26:36 This is being a really nice day for typos Dec 15 04:27:31 someone know x11 calc which work with hexadecimals? I mean put 0x3434+0x453dee5aa3 and get ouput in hex, switch to decimal etc Dec 15 04:28:04 hrw|work: fix pushed Dec 15 04:28:41 * RP has an old fashioned calculator that does hex.... Dec 15 04:29:05 RP: I gave my calculator for sister when finished studies. Dec 15 04:29:30 she had to get some lessions how to use it from her boyfriend anyway Dec 15 04:32:14 FSCK! Dec 15 04:32:18 Dec 15 12:31:50 c7x0 user.err kernel: mmcblk0: error 1 sending read/write command Dec 15 04:33:16 argh........ Dec 15 04:33:38 hrw|work: Same block number as before or not? Dec 15 04:33:44 heh Dec 15 04:33:47 RP: other Dec 15 04:33:49 hostap utils is really broken Dec 15 04:34:20 hrw|work: The pxa can't go any slower than that :-/ Dec 15 04:34:43 http://www.hrw.one.pl/tmp/mmcblk3.txt Dec 15 04:35:07 ~kill company which make 'takeMS' sd cards Dec 15 04:35:09 * ibot shoots a excited fluxgraviton gun at company which make 'takeMS' sd cards Dec 15 04:35:31 hrw|work: Did it continue giving errors after that? Dec 15 04:35:45 yes, of course it did. ignore me Dec 15 04:35:57 RP: it does Dec 15 04:36:23 ~seen greentux Dec 15 04:36:27 greentux was last seen on IRC in channel #oe, 22h 46m 36s ago, saying: 'i reinserted a usb-network adapter sometimes. the udhcpc stops the whole Z...'. Dec 15 04:36:42 Interestingly it was an error 2 followed by an error 1 Dec 15 04:36:54 were the others always error 2 then error 1? Dec 15 04:37:27 03rpurdie 07org.oe.dev * r47525f6e... 10/packages/hostap/hostap-utils_0.4.0.bb: hostap-utils: fix a typo Dec 15 04:37:28 RP: http://www.hrw.one.pl/tmp/mmc-log.txt http://www.hrw.one.pl/tmp/mmclog2.txt Dec 15 04:38:49 Dec 15 12:31:42 c7x0 user.debug kernel: PXAMCI: stat 0000214a Dec 15 04:38:58 hrw|work: So no, this was a new form of error - weird... Dec 15 04:39:10 03koen 07org.oe.dev * r8fb1bc42... 10/packages/hostap/hostap-utils_0.4.0.bb: Dec 15 04:39:10 hostap-utils 0.4.0: fix serious bugs: Dec 15 04:39:10 * RDEPENDS wasn't a subset of DEPENDS Dec 15 04:39:10 * Packages should NOT RDEPEND on kernel-modules, but RRECOMMEND them Dec 15 04:39:10 * ha-u DEPENDS on ha-m and RRECOMMENDS ha-m now Dec 15 04:39:20 lowest bit is 0 so its not command timeout.. Dec 15 04:40:12 hrw|work, rather that slowing down the mmc bitclock, is there any way you put a bigger delay between the commands Dec 15 04:40:31 koen: You'll need to fix hostap-utils 0.3.7 and hostap-daemon Dec 15 04:40:51 RP: and probably tons more Dec 15 04:41:03 but I'll leave that to the respective MAINTAINERS Dec 15 04:41:59 lrg|home: the problem is that I'm total ignorant when it comes to kernel space Dec 15 04:42:02 <_law_> is there a way to disable the zaurus display if i close him? Dec 15 04:42:42 <_law_> so i can save some power in usb-cable mode :-) Dec 15 04:43:11 _law_: under opie screen is off when closed Dec 15 04:43:40 hrw|work, I did some work about 4 yrs ago on cf, mmc, sd, memory sticks and realized they were not all created equal. :( You could always try a udelay() after each command and see if it make any difference. Dec 15 04:43:45 this gcc checkout is taking ages Dec 15 04:44:00 hey gints_ Dec 15 04:44:27 hey koen Dec 15 04:44:41 hrw|work, I remember our software (vxworks based) had to do all sorts of crap to get most cards to work Dec 15 04:46:27 I guess the thing to do might be to add better error handling to the mmc layer... Dec 15 04:47:35 <_law_> hrw|work, so the button is supported Dec 15 04:49:54 _law_: atleast in opie 1.2.1 Dec 15 05:04:34 koen: what do you think about write access for Arjan and his h1940 work? Dec 15 05:04:49 I was just writing an email about that :) Dec 15 05:07:26 ') Dec 15 05:07:56 ArjanS: do you want push access to OE? Dec 15 05:08:31 koen: that would be great! are there any special requirements for that? Dec 15 05:09:04 only the stuff in http://oe.handhelds.org/cgi-bin/moin.cgi/MonotonePhraseBook and http://handhelds.org/hypermail/oe/47/4772.html Dec 15 05:10:27 hrw|work: may be I will get my oe ready... I will report as soon as I can test BT pin Dec 15 05:10:50 gints_: I made ui resizable in cvs Dec 15 05:11:58 ok Dec 15 05:12:00 koen: I'll read them and get back to you Dec 15 05:18:36 koen: this morning I cleaned out my complete OE db and files and started all over with nothing... so I already did a full pull... do I still need to get that clean.db and use that one? Dec 15 05:33:20 a full pull from scratch should have the same effect as using clean.db (or OE.db.bz2) Dec 15 05:33:45 koen: and would the key name arjan@openembedded.org be right? or isn't that used as email address? Dec 15 05:34:20 @openembedded.org would be right Dec 15 05:34:31 if anyone is interested: http://dominion.kabel.utwente.nl/koen/pda/files/angstrom-eabi.diff.bz2 Dec 15 05:34:44 that still seems to have the glibc not found problem, though Dec 15 05:35:07 * koen heads of to buy some groceries Dec 15 05:39:18 koen|away: the mail is out :) Dec 15 05:43:05 RP: test with 4 instead of 6 still going - 1h passed, goes further then was with 6 Dec 15 05:43:36 ~lart ADSL Dec 15 05:43:37 * ibot installs a bad bootloader on ADSL and turns ADSL into a brick Dec 15 05:43:52 yah, that sounds like a bug in matchbox. Dec 15 05:44:55 looking at the code, an XGrabKey() failure ought not to be fatal to it. Dec 15 05:55:38 hrw|work: Perhaps that last error was just unlucky - I suspect the pxa error handling code isn't very efficient... Dec 15 05:56:53 RP: probably. but I would like to get something which will work. Dec 15 05:58:03 RP: in other way I would not trust that card enough to keep data on it during it use in this Zaurus Dec 15 05:58:19 I dont need 1G card for my photocamera Dec 15 05:58:47 hrw|work: I know what you mean... Dec 15 06:01:12 finally error Dec 15 06:01:30 1h18m with clkrt=4 Dec 15 06:05:59 greentux: do not buy those 'takems' cards next time. buy 'kingston', 'saddisk' or other but not totalnonames Dec 15 06:06:03 hi greentux Dec 15 06:06:12 hi hrw|work Dec 15 06:06:17 hi greentux Dec 15 06:06:38 hrw|work: mmmh i took the first match :( Dec 15 06:06:49 hrw|work: so is it useless? Dec 15 06:06:59 greentux: without rewritting mmc support probably yes Dec 15 06:07:12 atleast in pxa powered Dec 15 06:07:34 hrw|work: we can change that card, but that is a 4 week process :) Dec 15 06:07:40 hi RP Dec 15 06:08:01 greentux: and few faxes/mails more Dec 15 06:09:01 ok... lets do this way: i try to get a kingston card... (perhaps you send me a link), then i send you the card, you check that Dec 15 06:09:47 hrw|work: and then you sent the noname back... Dec 15 06:09:50 greentux: noone at your side has c7x0? only cxx00 devices? Dec 15 06:10:04 hrw|work: yes, only 3100 :) why? Dec 15 06:10:34 greentux: do 'dd if=/dev/mmcblk0 of=/dev/null bs=2048' before sending card Dec 15 06:10:56 hrw|work: for testing? Dec 15 06:11:05 greentux: that card which I have now fails it even with weird settings for mmc controler Dec 15 06:11:30 hrw|work: ok, i will use it in my mobile... Dec 15 06:11:48 greentux: and that card pass that test in usb read/writer, in sl-5500 Dec 15 06:12:07 greentux: do it under 2.6 in zaurus if you can Dec 15 06:12:27 hrw|work: I suspect it probably work if we could fix the pxa mmc error handling... Dec 15 06:13:07 RP: I can provide testing power Dec 15 06:14:32 ok, is the card not standard conform or is the mmc driver not 100% ? Dec 15 06:15:10 greentux: We don't know given the SD standard isn't open Dec 15 06:15:56 mmh ok. Dec 15 06:25:09 ArjanS: you can now push to ewi546.ewi.utwente.nl (after pulling and updating of course) Dec 15 06:25:49 ArjanS: you can check with viewmtn (http://ewi546.ewi.utwente.nl/tmp/viewmtn/branch.psp?branch=org.openembedded.dev) which is the latest revision Dec 15 06:25:57 (give or take an hour) Dec 15 06:27:05 koen: thanks! I will try it in a few minutes Dec 15 06:29:52 koen: if I understand it correctly: when you do mt commit, it gets committed to the local oe.db, and when you do mt push it gets pushed to the server? Dec 15 06:30:40 and should I always pull from ewi or is vanille.de good too? Dec 15 06:33:35 ArjanS: Remember to pull/update before commit/push Dec 15 06:33:41 You can use both servers Dec 15 06:38:27 ArjanS: correct, commit=local, push=remote Dec 15 06:38:33 and what RP said :) Dec 15 06:39:01 it might take some time for mickeyl to activate your keys, so pushing will be restricted to ewi for a while Dec 15 06:40:02 that's no problem, thanks :) Dec 15 06:50:00 I get this when issuing 'monotone merge' before pushing: monotone: misuse: branch 'org.openembedded.dev' is merged Dec 15 06:50:27 that means a merge isn't needed :) Dec 15 06:50:36 thanks, just checking :) Dec 15 06:58:12 koen: I have NOTE: package glibc-2.3.5+cvs20051214: started Dec 15 06:58:57 after pH5's hint I have that too Dec 15 06:59:23 03arjan 07org.oe.dev * r4e189300... 10/conf/machine/h1940.conf: h1940.conf: update to make it work with the new kernel and add GUI_MACHINE_CLASS="smallscreen" Dec 15 06:59:28 PREFERRED_PROVIDER_virtual/arm-linux-eabi-libc-for-gcc="glibc" Dec 15 06:59:37 stuff? I set that because bitbake complain Dec 15 07:00:42 I'm updating the angstrom-eabi.diff right now Dec 15 07:02:05 * koen taps foot Dec 15 07:02:11 NOTE: Handling BitBake files: \ (0657/2945) [22 %] Dec 15 07:02:25 on my 'slow' 1800+ Dec 15 07:04:04 athxp 1800? Dec 15 07:04:09 yes Dec 15 07:04:34 reminds me, need to buy new computer Dec 15 07:04:38 uf.. atleast once your buildmachine is slower then mine ;) Dec 15 07:04:41 2200+ here Dec 15 07:04:42 that was the cheapest cpu I could find a few years ago Dec 15 07:04:50 (durons were sold out) Dec 15 07:05:21 @work I have 1700+ but trying to takeover 2000+ which sit in other (not used now) machine Dec 15 07:05:32 Mine's a P3-800! Dec 15 07:05:42 I have a P4 2.8GHz and bitbake cache is still very slow :( Dec 15 07:06:13 caching is 65% user over here, so I think it is IO limited Dec 15 07:06:35 It's all relative - I have 320MB of RAM now, but I did have bitbake (the old memory hogging version) running (v slowly) on a p3-400 w/ 64MB of RAM Dec 15 07:06:52 that reminds me Dec 15 07:06:59 I should replace the db on my slug Dec 15 07:09:58 http://oskuro.net/blog/freesoftware/gnome-2.12-unstable-2005-12-15-14-19 Dec 15 07:10:04 debian has fun with dbus too Dec 15 07:10:38 gnome/experimental use new, gnome/unstable and kde use old Dec 15 07:11:21 does exist a keyword which controls which keyboardlayout is used? Dec 15 07:22:31 koen: http://pastebin.com/465225 is end of eabi here Dec 15 07:22:36 Cobelius: no Dec 15 07:25:12 strange, binutils builds over here Dec 15 07:25:56 will try with clean ccache Dec 15 07:34:05 same error Dec 15 07:36:57 03arjan 07org.oe.dev * rda3df1d1... 10/packages/linux/ (2 files in 2 dirs): linux-h1940: add kernel 2.6.14 for the h1940 series Dec 15 07:39:23 will shutdown now Dec 15 07:44:08 *schnarch* Dec 15 07:45:05 hey zecke Dec 15 07:47:10 hail zecke Dec 15 07:48:30 hi zecke Dec 15 09:10:51 03arjan 07org.oe.dev * r5aa7f8a6... 10/packages/bluez/ (files/hciattach_devlength.patch bluez-utils.inc): bluez-utils-2.21: add very small patch to make it work on the h1940 Dec 15 09:12:07 ArjanS: does that patch was sent upstream? Dec 15 09:25:51 ~lart linuxthreads Dec 15 09:25:52 * ibot takes out a seltzer bottle and sprays linuxthreads in the face. You know, one of those old-school seltzer bottles clowns have? Yeah those. Anyway, consider yourself spritzed Dec 15 09:47:06 sigh Dec 15 09:47:24 I need to get the newer version of bitbake - I wish I could get file rebuild to work Dec 15 09:47:59 newer then r321? Dec 15 09:48:14 bleh Dec 15 09:48:20 glibc is stoooopid Dec 15 09:48:21 I'm using an old version Dec 15 09:48:50 * koen reverts to speak of 7 year olds Dec 15 09:49:19 * koen gets some candy Dec 15 09:49:28 koen: make uclibc speak EABI Dec 15 09:49:41 koen: go and play http://www.geocities.jp/lledoece/nanacatest.html Dec 15 09:51:44 hrw|work: we have the eToken (i told you) up and running :) Dec 15 09:52:11 hrw|work: 239.55m Dec 15 09:52:46 greentux: that auth on usb something? Dec 15 09:52:52 XorA: 3800 Dec 15 09:53:51 hrw|work: we can read the files on the eToken Dec 15 09:53:59 hrw|work: pc/sc working Dec 15 09:55:04 XorA: over 500 and flying... Dec 15 09:55:32 hrw|work: http://ebaumsworld.com/kittycannon.html Dec 15 09:55:52 XorA: 687... Dec 15 09:57:16 ok. time to go home Dec 15 09:58:03 RP: I'm curious what does rmk and Pierre Ossman write about sd card Dec 15 09:59:13 cu Dec 15 10:00:03 cu Dec 15 10:00:23 cya Dec 15 10:04:30 koen...leaving the battery out of my iPAQ for over 12 hours did not help the LCD/touchscreen issue... Dec 15 10:04:49 hmm Dec 15 10:04:57 I'm out of suggestions Dec 15 10:05:04 the LCD backlight is not coming on...what controls that? Dec 15 10:07:28 hrw|gone: that hciattach patch is applied in bluez-utils 2.23, but we have up to 2.21, which don't work yet Dec 15 11:42:47 * Philippe is back (gone 19:59:01) Dec 15 11:42:48 hi Dec 15 11:58:23 the second opcode is just the next line? (0xD, at byte 4) Dec 15 12:08:16 re Dec 15 12:08:43 hrw: wb Dec 15 12:09:36 I;m curious what I have to do to do working serial cable for my Z. other then soldering few wires to that @%$#%# tcx3106 connector on one side and db9 on other.. Dec 15 12:11:11 likeI got in mail (according to sdcard problem): Dec 15 12:11:14 Hmm, it's difficult to tell what's going on from a syslog log - syslog is unfortunately lossy. It looks like some kernel messages were lost around the time when the errors start. Dec 15 12:11:17 Please log using a serial console instead. Dec 15 12:11:56 hopefully I have 2 tcx3106 connectors left Dec 15 12:15:02 do I need max232 to convert zaurus voltage levels to pc ones? or just wires? Dec 15 12:16:29 let's teach the world english and switch to uclibc Dec 15 12:16:36 *&$&)#* glibc Dec 15 12:17:32 lol Dec 15 12:18:27 figuring out this eabi this reminds me of the days I tried to build my own cross toolchain Dec 15 12:18:43 koen: fsck english - Polish has less dialects Dec 15 12:19:04 less time forms (just 3: past, present, future) Dec 15 12:19:21 and hard to spell fonetics ;) Dec 15 12:19:22 yeah, but uclibc already understands english... :) Dec 15 12:21:13 wow.. max232 is 'expensive' part.. 3 pln for normal, 9 pln for smd version Dec 15 12:21:17 ~change 3 pln to eur Dec 15 12:21:22 3.00 Zloty (PLN) makes 0.779082 Euro (EUR) (from http://www.xe.com/) Dec 15 12:23:27 but when I think about soldering 0.6mm raster.. argh.. Dec 15 12:23:39 ~kill sharp for tcx3106 Dec 15 12:23:41 * ibot shoots a inverse pseudophoton gun at sharp for tcx3106 Dec 15 12:25:07 afk Dec 15 12:26:17 hrw, if you play it smart you can get free samples from maxim... Dec 15 12:26:35 yeah Dec 15 12:26:42 and bi-weekly snailmail spam Dec 15 12:27:00 * Philippe never got snailmail spam from them... Dec 15 12:27:08 * koen gets mail from dallas, maxim and ti Dec 15 12:27:28 koen, good stuff to heat the house :- Dec 15 12:28:32 * gremlin[it] got 2 free sampe of SA1110 years ago from Intel ... :) ... Dec 15 12:29:18 useless till i got a course at univ ... so student can see for real how small a processor can be ... :) ... Dec 15 12:31:05 aargh Dec 15 12:31:05 | as: unrecognized option `-meabi=4' Dec 15 12:31:38 :-/ another sip... Dec 15 12:34:05 no, a walk to the closet Dec 15 12:34:13 nfs decided to oops the kernel again Dec 15 12:40:45 stupid linux Dec 15 12:41:00 * koen waits for sun to donate a t2000 server Dec 15 12:41:59 samples rock Dec 15 12:42:12 you can get them direct from internet now for some companies Dec 15 12:42:20 used to have to kiss people's rear ends Dec 15 12:42:34 dear amd, I would like to sample your 8way amd64 barebones Dec 15 12:43:14 dear sun, please let me sample your t2000 series Dec 15 12:43:29 ihihiihi :) Dec 15 12:43:39 dear sgi, please give me an altix cluster (and some space + power) Dec 15 12:43:50 Actually, that one might work Dec 15 12:44:12 dear intel ... give me some pxa-290 sample board :) :) : ) Dec 15 12:44:19 SGI is desperate for some lovin' these days Dec 15 12:44:57 they should fix dual-head support in irix Dec 15 12:45:12 I still can't use my 2 videoboards at the same time in my indigo2 Dec 15 12:46:39 re Dec 15 12:46:45 why the hell here in italy is so difficult to find 'esotis' hardware ?!?!?!?! :( :( :( Dec 15 12:46:52 hey hrw Dec 15 12:46:59 hi hrw Dec 15 12:47:11 Philippe: guess what will be sooner: me going to local shop and pay 3pln for max232 or maxim sending them to me? Dec 15 12:47:58 at least dallas is pretty quick Dec 15 12:48:01 only a few days Dec 15 12:48:38 hrw: ask for samples ;) Dec 15 12:48:53 anyway I still dont know what I need for cable Dec 15 12:49:11 hrw, the first option :-) Dec 15 12:49:31 one place mention max232, max2344, inverters etc.. Dec 15 12:49:40 20:09 < hrw> edit ~/Settings/libxine.conf or something Dec 15 12:49:42 osp Dec 15 12:49:44 ops Dec 15 12:50:11 http://www.dk7in.de/Zaurus.html list inverters but mention also chips Dec 15 12:50:37 Tip von TriSoft = trichloroethylene in babelfish Dec 15 12:50:43 ;) Dec 15 12:50:48 hrw: strictly speaking you do not need any max232 for a serial cable ;) Dec 15 12:51:34 zecke: so does only plain wires will be enough? will serial console work on 3wires or need whole 7? Dec 15 12:52:13 hm. maybe answer can be found in service manual for collie Dec 15 12:52:36 hrw: well not plain (this can destroy the lines on your zaurus..) Dec 15 12:52:50 hrw: the dk7in.de site looks good Dec 15 12:53:47 ok. Dec 15 12:54:06 crap Dec 15 12:54:22 I can't compile gcc to compile glibc to compile gcc Dec 15 12:55:38 ask god to have a gcc :) Dec 15 12:56:12 zecke: so wires + inverters Dec 15 12:56:40 hrw: +resistors or max232 http://axim.narcotize.com/serialcable.html Dec 15 12:56:51 hrw: the one from this site was my first serial cable attempt Dec 15 12:57:09 hrw: my lecturer 'forced' me to build another one with a max232 Dec 15 12:57:28 hrw, or you can take max232cpp's so you do not need to put capacities around it... Dec 15 12:57:40 s/capacities/capacitors/ Dec 15 12:58:04 'hook this up to AC' Dec 15 12:58:21 Philippe: right and save the hazle with the voltage pump Dec 15 12:58:25 when you hear that phrase in the lab, run Dec 15 12:58:44 some 1st year student will connect a DC cap to AC Dec 15 12:58:56 koen: the current should be limited to 1A ;) Dec 15 12:59:20 koen: this is why we use only bipolar caps ;) Dec 15 12:59:29 does any one knows how to restart the g_ether module? (which is compiled into the kernel) Dec 15 12:59:32 * Philippe still remembers the 220V connection through a 8051 to a usb port.... That was some really great fireworks.... Dec 15 13:00:01 YoG: reboot Dec 15 13:00:07 :( Dec 15 13:00:28 koen: no simple way of doing it? Dec 15 13:00:41 hrw|afk: building a lead with a level convertor is a safer bet incase you ever get a sensitive zaurus such as the cxx00 ones... Dec 15 13:00:42 YoG: compile it as module Dec 15 13:00:58 rebooting is simple :) Dec 15 13:01:54 koen: yes, but it takes time, f^&ks up the clock and I have to do it too often... (on my c860) Dec 15 13:01:59 koen: I tried cross tool today and failed on OS X Dec 15 13:02:08 max233a looks nice - nocaps Dec 15 13:02:34 zecke: crosstool sucks, use OE ;) Dec 15 13:02:35 I've had several maxim 23x samples before :) Dec 15 13:02:57 zecke: if i'll compile it as a module, do i have to recompile the kernel also? Dec 15 13:03:18 eabi sucks too Dec 15 13:03:20 adn glibc Dec 15 13:03:24 and gcc Dec 15 13:03:33 and circular deps Dec 15 13:03:35 and nfs Dec 15 13:03:37 koen: recently Linus started to suck Dec 15 13:04:03 * koen doesn't pay attention to linus Dec 15 13:04:27 ever since crazy 2.6 wonderland and 'specs suck' Dec 15 13:04:58 school sucks Dec 15 13:05:03 that too Dec 15 13:05:08 and assignments Dec 15 13:05:33 who of us would be interested in paid OE work? Dec 15 13:06:05 I would, but my parents and girlfriend would kill me Dec 15 13:06:13 * koen wants to finish uni *fast* Dec 15 13:06:16 zecke: Openedhand/me might be Dec 15 13:06:50 hehe Dec 15 13:07:31 mainly they're looking for people but I will keep that in mind! Dec 15 13:07:37 http://mailinator.com/mailinator/index.jsp rule for quick one-time emails Dec 15 13:08:15 zecke: Ah, I understand what you mean now :) Dec 15 13:09:31 but for anyone's reference, OH would probably take on OE projects Dec 15 13:09:49 I'm likely to start to work at a local company and they are looking into growing even further Dec 15 13:09:59 like any current business they look for skilled people Dec 15 13:10:03 RP: pizza fetcher Dec 15 13:11:14 koen: You have nothing to base that comment on ;-) Dec 15 13:11:38 I still want that pizza fetcher in OE kergoth promised Dec 15 13:11:55 * zecke still wants to merge RPs fetcher patches Dec 15 13:12:08 koen: That would be nice Dec 15 13:12:18 zecke: I'd still like them merged as well :) Dec 15 13:12:41 and the other bugfix I submitted... Dec 15 13:12:49 * koen catches up on Lost Dec 15 13:12:54 RP: got a # at hand? Dec 15 13:13:10 merging could be the last thing before sleeping Dec 15 13:14:47 'Ship Account Number' - what it is when I order samples at maxim... Dec 15 13:15:02 zecke: bugs 438 and 436 Dec 15 13:15:44 hrw: Probably just left blank? Dec 15 13:16:29 RP: only usps allow it to be blank. Dec 15 13:17:36 hrw: I don't know then - its been a while since I did that Dec 15 13:19:04 ok. Dec 15 13:19:08 time for tv Dec 15 13:19:16 I Dec 15 13:19:23 I'm all hopped up on something, but I don't know what Dec 15 13:19:27 probably sugar Dec 15 13:19:38 same here Dec 15 13:19:48 * koen|tv ate a bag of sweets Dec 15 13:20:06 RP: #436 is applied?! Dec 15 13:20:45 zecke: I never saw the bug get closed so assumed otherwise, sorry! Dec 15 13:21:26 r321 fixed it already Dec 15 13:21:35 r321 | zecke123 | 2005-11-03 22:54:43 +0100 (Thu, 03 Nov 2005) | 19 lines Dec 15 13:22:32 zecke: If you're applying things, have a look at bug #531 - looks a simple fix? Dec 15 13:29:24 RP: alreay done ;) Dec 15 13:29:40 RP: did that issue lead to any errors? Dec 15 13:30:40 zecke: I don't know. I just saw it and noted it was an obvious typo Dec 15 13:35:00 RP: I've got spitz working well now. Can record from line in and can hear line bypass. Also got mic/hp jact switch and ext spk disable switct (for hp insertion) Dec 15 13:36:01 RP: I'll do mic testing tomorrow and make the same changes to corgi. Dec 15 13:36:13 lrg|home: excellent. I'll be interested to see the patch and if you let me have an alsa config file, I'll add it to OE. Dec 15 13:37:04 lrg|home: Sounds good to me (keep in mind the corgi hardware is slightly different) Dec 15 13:37:08 RP: np, probably looking at tom for patch and Mon for config. I'm now off to retire for the evening :) Dec 15 13:37:32 stop looking at tom1 Dec 15 13:37:35 ! Dec 15 13:38:00 CosmicPenguin, but he's my special fried :) Dec 15 13:38:05 friend Dec 15 13:39:17 CosmicPenguin, would you be able to send me your au1200 audio patch. I'd like to start looking at integrating it soon. I know the ac97 controller is different from the older alchemys, but not sure about the dma or i2s Dec 15 13:39:44 heh Dec 15 13:39:50 sure - I actually just merged it Dec 15 13:40:00 I'll send it right now, if you're not scared of code that might not compile Dec 15 13:40:27 CosmicPenguin, cool, I'm not afraid (honest) Dec 15 13:41:10 Lemme see if I remember your e-mail Dec 15 13:41:38 CosmicPenguin, liam.girdwood at wolfsonmicro.com Dec 15 13:42:02 got it Dec 15 13:42:51 CosmicPenguin, I've only got a pb1500 to test on. Am I likely to have any problems with the dma or i2s ? Dec 15 13:42:57 code Dec 15 13:43:50 CosmicPenguin, I mean wrt to au1200 features that are not present Dec 15 13:44:08 yeah, its a completely different engine.. :) Dec 15 13:44:16 Now, if you had a 1550 on hand, that would work Dec 15 13:44:38 CosmicPenguin, ah. Anyway I could borrow one for 2 months... Dec 15 13:45:09 Let me check Dec 15 13:45:21 CosmicPenguin, Thanks. Dec 15 13:45:27 We had a run on the development platforms, and we ran out, but I think we got more in stock Dec 15 13:46:23 You going to be around tommorrow? Dec 15 13:46:27 yes Dec 15 13:46:32 I'll have an answer for you then Dec 15 13:46:42 But I'll send you the code in the mean time, just for fun Dec 15 13:46:42 CosmicPenguin, many thanks. Dec 15 14:12:02 hey mithro Dec 15 14:13:11 hey koen|tv Dec 15 14:22:51 cu Dec 15 14:42:51 <_law_> whats mmcqd? Dec 15 14:51:40 koen|tv: I'm going to give angstrom patch a try. What do I have to change in my local.conf? Just DISTRO="angstrom-2005.6"? Dec 15 14:51:55 yes Dec 15 14:52:11 I think I 'solved' the binutils problem Dec 15 14:52:19 wrong version? Dec 15 14:52:27 I still had a arm-linux-eabi binutils Dec 15 14:52:53 now I have a arm-none-linux-gnueabi binutils Dec 15 14:53:46 ah, it's always an old version of something Dec 15 14:54:36 gna!! Dec 15 14:54:42 | checking for forced unwind support... no| configure: error: forced unwind support is required Dec 15 14:54:50 ERROR: Build of 'glibc-intermediate' failed Dec 15 14:57:54 ah wait Dec 15 14:58:00 it's just missing crt1.o Dec 15 14:59:46 which should be provided by glibc... Dec 15 15:01:07 so I can't build glibc-intermediate because I need glibc Dec 15 15:01:34 * pH5 puzzled Dec 15 15:01:56 and to build glibc I need to build gcc-cross which needs glibc-intermediate Dec 15 15:02:36 from config.log: http://pastebin.com/465857 Dec 15 15:02:39 * koen wipes tmp Dec 15 15:03:14 <_law_> found another little problem: reboot holds my Z totaly and shutdown -h now doesnt turn off my Z Dec 15 15:03:49 afaik you can't actually shutdown a arm pda Dec 15 15:04:02 koen: you can Dec 15 15:04:20 yeah, but it will use more power than being suspended, right? Dec 15 15:04:31 koen: depends on the hardware Dec 15 15:04:37 but quite possibly, yes Dec 15 15:09:57 i find "removing the battery" works for me ;) Dec 15 15:10:52 yeah Dec 15 15:11:04 that's my trick to do a proper 'halt' :) Dec 15 15:11:36 oh wait Dec 15 15:11:46 my libtool has probably arm-linux-eabi too Dec 15 15:12:19 <_law_> n8 Dec 15 15:13:33 koen: you meen libtool-native? my build just got past that point :/ Dec 15 15:13:51 it's at gettext now Dec 15 15:13:59 gettext-native to be exact Dec 15 15:14:29 same here :) Dec 15 15:34:30 pH5: gcc-c-i Dec 15 15:35:32 I just got stuck again at virtual/arm-none-linux-gnueabi-libc-for-gcc Dec 15 15:35:38 ah yes Dec 15 15:35:52 glibc doesn't provide it and glibc-intermediate now depends on itself or something Dec 15 15:35:54 I jusst do a 'bitbake binutils-cross gcc-cross-initial' first Dec 15 15:36:07 those two are built already Dec 15 15:36:23 next stop is gcc-cross, I think Dec 15 15:36:28 but it needs virtual/... Dec 15 15:36:31 in that case you should be able to bitbake -b glibc-intermediate Dec 15 15:36:42 you need a glibc to build gcc due to ntpl Dec 15 15:36:53 hence the dependencie mess Dec 15 15:38:39 http://pastebin.ca/33826 Dec 15 15:39:17 yep Dec 15 15:39:19 same here Dec 15 15:39:39 could you start a new thread about that on the mailinglist? Dec 15 15:39:49 I'm sure someone will have a comment on how to fix that Dec 15 15:42:11 what is the preferred version for virtual/arm-none-linux-gnueabi-gcc-initial? Dec 15 15:42:14 parsing... Dec 15 15:42:34 2.3.5+cvs20051214 Dec 15 15:42:49 eh sorry. preferred provider Dec 15 15:43:10 gcc-cross-initial (I think) Dec 15 15:43:47 i think you're right. (at least, whatever it is, it's already built) Dec 15 15:44:41 could it be that glibc-i needs to be built without ntpl? Dec 15 15:44:51 * Philippe is away: visual contact - dreamland Dec 15 15:45:47 koen: yes Dec 15 15:45:52 (or no) Dec 15 15:45:59 DEPENDS = "${@['virtual/${TARGET_PREFIX}gcc-initial', 'virtual/${TARGET_PREFIX}gcc']['nptl' in '${GLIBC_ADDONS}']} linux-libc-headers" Dec 15 15:46:00 heh Dec 15 15:46:15 clearly it depends on gcc (which is wrong for glibc-i) Dec 15 15:46:36 the question is: should nptl be disabled or this rule? Dec 15 15:47:03 it depends on gcc-i, which is corect (I think) Dec 15 15:47:16 ah wait Dec 15 15:47:20 see Dec 15 15:47:22 * koen misreads that line Dec 15 15:47:55 so let's add GLIBC_ADDONS = "ports" to glibc-i and try again? Dec 15 15:48:34 | running configure fragment for ports/sysdeps/unix/sysv/linux/arm/eabi/nptl| running configure fragment for ports/sysdeps/unix/sysv/linux/arm/eabi| running configure fragment for nptl/sysdeps/pthread| checking for forced unwind support... no| configure: error: forced unwind support is required Dec 15 15:48:43 looks like it's already building ports Dec 15 15:50:27 ah, I finally see your point Dec 15 15:50:48 | checking for kernel header at least 2.6.14... too old! Dec 15 15:50:58 heh Dec 15 15:51:00 wow - bleeding edge Dec 15 15:52:17 At least it doesn't say 2.6.14-rc5-mm or something like that Dec 15 15:53:01 linux-libcheaders is at 2.6.12.0 atm Dec 15 15:53:52 but http://svn.pld-linux.org/cgi-bin/viewsvn/linux-libc-headers/trunk/ has 2.6.14 Dec 15 15:59:31 pH5: I made that change to glibc-i and am building linux-l-h from svn right now Dec 15 15:59:44 I'll put up r5 of the patch tomorrow morrning Dec 15 15:59:58 I can say tomorrow, because it's 23:59 :) Dec 15 16:00:06 koen: ok, thanks a lot Dec 15 16:00:57 after this I'm going to put 'toolchain monkey' onto my resume Dec 15 16:01:40 :-) Dec 15 16:17:39 * koen cheats by editing include/linux/version.h Dec 15 16:21:11 good night Dec 15 16:28:31 ibot: botmail for pH5: http://dominion.kabel.utwente.nl/koen/pda/files/angstrom-eabi-r5.diff.bz2, probably needs linuxthreads in glibc-i Dec 15 16:28:54 'night all Dec 15 16:28:58 night Dec 15 17:33:31 does anyone know what this error means? Anonymous access over SSH is replaced by the (more efficient) pserver. while trying to retrive quilt? Dec 15 17:38:09 FYI....psmisc-21.8-r0 is failing the do_compile step when building any image from scratch. Dec 15 17:38:30 I can't pretend to understand why 8-( Dec 15 17:47:04 i'm having probs building sdl-image from sctrach too... :( Dec 15 17:48:23 here's my error if anyone cares to look :/ http://pastebin.ca/33839 **** ENDING LOGGING AT Thu Dec 15 18:31:47 2005 **** BEGIN LOGGING AT Thu Dec 15 18:32:14 2005 Dec 16 01:21:12 koen: good morning Dec 16 01:21:29 hey ArjanS Dec 16 01:22:45 would it be okay to put the h1940.conf updates also in .oz354fam083 branch? it wouldn't affect anyone since the h1940 is not a familiar 0.8.3 supported device Dec 16 01:24:47 and perhaps the bluez-utils too, that one is harmless too Dec 16 01:24:56 bluez-utils patch* Dec 16 01:31:55 if so, can/may I do that myself? what's the policy on that? Dec 16 01:32:43 you can't do that yourself, only mickeyl, hrw and I update the branch Dec 16 01:33:00 and yet another (the last) question: if I find .bb's I can fix, but I'm not the maintainer of, should I file a bug for the maintainer to fix or fix it myself? Dec 16 01:33:53 same applies to new versions of packages (f.i. bluez-utils-2.23), should the maintainer add them? Dec 16 01:39:09 if it's a typo you can fix it yourself for bigger changes ask on irc, or use bugzilla Dec 16 01:39:26 ok thanks Dec 16 01:39:31 you can add new packages, but remove the MAINTAINER field if you copy it Dec 16 01:51:26 hi Dec 16 01:55:01 hey oxi Dec 16 01:55:52 should i compile bootstrap-image first to try it on my pda? there seems to be no distro for it yet Dec 16 01:57:21 hi Dec 16 01:57:26 hey hrw|work Dec 16 01:57:39 oxi: bootstrap-image would be a good start Dec 16 01:59:48 thank you :) Dec 16 02:00:51 lo folks Dec 16 02:01:09 hey ade|desk Dec 16 02:01:13 hey XorA Dec 16 02:01:13 morning Dec 16 02:01:29 ade|desk: could you add your diet-x11 patch for 6.2.1 to OE? Dec 16 02:01:36 ZonE-HaKin9|wrk: rm -rf CVSDIR/quilt* Dec 16 02:02:31 oxi: which pda? Dec 16 02:02:57 htc universal Dec 16 02:03:31 http://handhelds.org/moin/moin.cgi/Universal Dec 16 02:03:34 universal.. its mda4, qtek9100 thing? clamshell after wizard? Dec 16 02:04:26 koen: i was going to add it to bug.oe as its not my package, i didn't want to add it until mallum had checked it out to be ok. I am not sure about one of the castings of NULL, but it seems to work ok Dec 16 02:15:18 hrw|work: qtek 9000 i think... i-mate jasjar, t-mobile mda pro Dec 16 02:18:36 yay! Dec 16 02:18:59 r6 solves the glibc dependency loop Dec 16 02:20:39 koen: patch added to bugs.oe Dec 16 02:20:59 ade|desk: thanks Dec 16 02:21:18 koen: want the modified bb too or not ? Dec 16 02:21:33 sure Dec 16 02:35:01 koen: done Dec 16 02:36:02 thanks Dec 16 02:36:31 * koen feels being ver close to a working eabi toolchain Dec 16 02:36:38 very* Dec 16 02:36:55 time for coffee Dec 16 02:37:04 angstrom release before christmas ;) Dec 16 02:37:10 yes Dec 16 02:37:18 christmas next year :p Dec 16 02:37:20 is it ok to have MACHINE = "handheld-common" and DISTRO = "generic" in my local.conf to compile for my pda? Dec 16 02:37:36 I'd choose distro=familiar Dec 16 02:38:08 you can safely remove the bbfatal line Dec 16 02:38:31 distro=familiar-0.8.3 would be even better Dec 16 02:38:54 that's what the other xda people are using too Dec 16 02:39:13 i see. thanks koen :) Dec 16 02:41:01 03pH5 07org.oe.dev * r78bfaa44... 10/packages/psmisc/ (psmisc-21.8/signames-header-fix.patch psmisc_21.8.bb): psmisc-21.8: build fix for #521 Dec 16 02:41:06 03pH5 07org.oe.dev * r83cb1892... 10/packages/glibc/glibc_2.3.5+cvs20050627.bb: glibc_2.3.5+cvs20050627: re-apply arm-no-hwcap.patch Dec 16 02:41:10 03pH5 07org.oe.dev * r30479c63... 10/conf/machine/handheld-common.conf: handheld-common.conf: prepare for pcmcia-cs->pcmciautils transition Dec 16 02:54:23 morning all **** ENDING LOGGING AT Fri Dec 16 02:59:58 2005