**** BEGIN LOGGING AT Tue Nov 01 03:00:03 2005 Nov 01 03:01:17 RP: I might have a sound update for you later today. There are a couple more things that need work. Nov 01 03:02:18 lrg: II'm tempted to say send it anyway :) Nov 01 03:04:07 RP: Hopefully I'll get it out this afternoon. I've got a few minor bugs to clear up first. Nov 01 03:05:11 * XorA fails to solve the mysteries of user id mapping in nfs Nov 01 03:06:14 lrg: Ok, thanks. I'm keen to get dpm working on the wm8750 and make it usable :) Nov 01 03:06:35 lrg: I have this c7x0 bug to fix fairly urgently... Nov 01 03:07:28 RP: we should have wm8750 playback working, however record will have to wait until after I can get the touch stuff submitted Nov 01 03:08:31 lrg: That's fine - I just need the dpm callback to get playback working which gets me into a good position Nov 01 03:17:29 Mmmmm. RP++ # Console I can read while it boots. Nov 01 03:18:04 Noodles: Nice, isn't it :) Nov 01 03:19:03 RP: did mallum finish porting that rotation patch to kdrive? Nov 01 03:19:51 koen: I'll ask Nov 01 03:21:39 Um. Do I need to recalibrate the touch screen with -git3? Nov 01 03:21:49 It doesn't seem to work now. Nov 01 03:22:47 hi Nov 01 03:23:01 Noodles: Ah. This doesn't sound good :-/ Nov 01 03:23:39 gpe-image; it's not frozen (I can ssh in and move around with the cursor pad), but the pointer just sits in the middle of the screen. Nov 01 03:23:57 Noodles: touchscreen hasn't been tested as I don't have a working graphics system on spitz :-/ Nov 01 03:24:34 In fact my spitz is in a bad way. There were probably input changes in -git3... Nov 01 03:24:36 Ok, and trying to recalibrate gives me the "Touch crosshairs..." screen ok, but touching them doesn't move onto the next point. Nov 01 03:25:31 Sounds like a broken ts driver or input system. Have you tslib_tests installed? Nov 01 03:25:47 ts_print and friends can give useful deguf Nov 01 03:27:00 It's not installed, but it's built so easy enough to try out. Nov 01 03:28:52 * Geo_KM is away: Away at the moment Nov 01 03:31:58 Ah. No /dev/input/. That won't help. Nov 01 03:34:15 Noodles: Creating event0/event1 makes it work here Nov 01 03:34:20 Yes. And here. Nov 01 03:34:34 Noodles: At least its not my kernel driver anyway :) Nov 01 03:34:36 :) Nov 01 03:34:42 We probably need a udev update Nov 01 03:35:17 While reading the dmesg I saw "i2c: error: exhausted retries" following "I2C: i2c-0: PXA I2C adapter". Is that something to worry about? Nov 01 03:35:29 Noodles: No, just ignore them Nov 01 03:35:36 I need to patch that error message out Nov 01 03:36:00 On akita, you get screenfuls of them :) Nov 01 03:36:35 :) Nov 01 03:37:33 is there a way to rebuild dependecies on a packages with oe ? ( like -e with portage ) Nov 01 03:37:52 rm -rf tmp/ Nov 01 03:38:51 I tried this, but now it fail in function do_stage because ipkg-compare-versions is missing. Nov 01 03:39:06 I guess I deleted too much things :\ Nov 01 03:44:59 ~lart debian Nov 01 03:45:00 * ibot slaps debian upside and over the head with one freakishly huge killer whale named hugh Nov 01 03:45:20 ppor hugh Nov 01 03:45:24 poor* Nov 01 03:45:29 That'll leave a mark. Nov 01 03:48:35 OT but has anyone used nfs4 with idmapping on debian? Nov 01 03:48:36 i have this error if I rm -rf /tmp then build : http://pastebin.ca/27268 . Nov 01 03:50:44 Not /tmp. Nov 01 03:51:22 I meaned build/tmp Nov 01 03:51:28 Dear pastebin. If you're going to advertise an IPv6 address get some decent IPv6 connectivity and not a piece of wet string. k, thx, bye. Nov 01 03:51:34 sorry. Nov 01 03:52:18 erm Nov 01 03:52:28 Noodles: you noticed that too? ;) Nov 01 03:53:15 did that fetch actually fetch anything? Nov 01 03:58:48 brb Nov 01 04:00:17 I guess no, Nov 01 04:00:26 I already had tarballs downloaded . Nov 01 04:00:44 and are those tarballs ok? Nov 01 04:00:56 I suspect it failed silenty somewhere during fetch Nov 01 04:00:59 yea, they were used on first build. Nov 01 04:01:06 oh. Nov 01 04:01:16 lemme check this Nov 01 04:02:41 all files sounds here, I'll remove them up. Nov 01 04:03:09 it will take me some time for download. Nov 01 04:37:34 * france is back (gone 09:39:56) Nov 01 04:40:13 koen: still the same error. Nov 01 05:24:43 hi ! Nov 01 05:36:17 mickey|TV: dns cleared earlier today for some users. Are you getting e-mail? Nov 01 05:38:44 france: I still can't resolve handhelds.org using host from here. Strengely, dig works... Nov 01 05:40:00 q Nov 01 05:40:07 soz. wrong window :) Nov 01 05:41:12 RP: some folks had their dns clear up this morning. Nov 01 05:41:40 RP: usually after doing a dig or nslookup which appearently cleared some cache. Nov 01 05:41:48 RP: so try it after the dig. Nov 01 05:45:02 france: dig, host and nslookup are all reporting failures here :-( Nov 01 05:45:41 RP: if you have a few minutes /join #toolchain and post the results Nov 01 05:59:00 03koen 07org.oe.dev * re90052e8... 10/: gpe-conf: add 0.1.26 Nov 01 05:59:05 03eFfeM 07org.oe.dev * r35381e2f... 10/: ixp425-eth/ixp400-eth: added pollcontroller patch for ixp400; required for netconsole Nov 01 06:42:17 hi Nov 01 06:42:57 RP: the -rc4 kernel is no longer on the .dev branch? Nov 01 06:43:22 Bernardo: We've upgraded to -git3 Nov 01 06:43:57 ok, so I should once again edit the git3.bb file and set the DEFAULT_PREFERENCE to 1? Nov 01 06:44:04 Bernardo: Should work well on akita although there is a known issue with the /dev/input directory disappearing Nov 01 06:44:06 Bernardo: Yes Nov 01 06:44:26 the /dev/input? Why, a udev problem? Nov 01 06:46:04 * Bernardo was unable to boot his laptop these last couple of days because he had a duplicate udev rules file, so devices ended with the wrong permissions... Nov 01 06:46:34 Bernardo: I suspect its a udev issue, yes Nov 01 07:14:51 * france is away: Away Nov 01 07:16:47 * kergoth yawns Nov 01 07:16:53 hey all Nov 01 07:16:58 Hello Nov 01 07:18:11 hiya Nov 01 07:18:21 hi kergoth Nov 01 07:18:40 * nickv111 whistles. Nov 01 08:05:39 damn ap... Nov 01 08:25:07 03jbowler 07org.oe.dev * rea663319... 10/: nslu2-kernel: split out LE part of drivers/mtd/maps/ixp4xx.c patch in 2.6.14 Nov 01 08:26:55 morning Nov 01 08:58:08 03mickeyl 07org.oe.dev * r348f3f65... 10/: openvpn: add version 2.0.2, remove version 2.0rc1 Nov 01 09:31:56 re Nov 01 09:56:12 meh Nov 01 09:56:22 we need a way to control multilib in oe. Nov 01 09:58:14 kergoth: what do you think of making -Os the default? Nov 01 09:59:29 default for whom? Nov 01 10:01:51 the whole of OE Nov 01 10:03:59 i dunno if thats a good assumption to be making Nov 01 10:04:40 me neither Nov 01 10:04:51 hmm Nov 01 10:05:07 I'm also not sure about that. If anything it should be a distro option rather than OE policy Nov 01 10:06:08 indeed Nov 01 10:06:29 though it seems more bound to the constraints of the machine than the distro, even. distro would involve changing less configuration files though ;) Nov 01 10:07:01 we do have 'embedded' in our name Nov 01 10:08:10 i dunno. its a policy decision, as RP says Nov 01 10:08:23 yeah Nov 01 10:08:34 i like the idea of putting it in the distro, myself Nov 01 10:09:04 We don't have to limit everything in OE to it. It can quite easily be a distro option... Nov 01 10:09:22 I thought it would be nice if oe did stuff like that by default Nov 01 10:09:59 i think -O2 is a more sane default for any arbitrary device than -Os. Nov 01 10:10:07 Despite the name, OE isn't just capable of building for embedded devices... Nov 01 10:10:29 and I agree with what kergoth just said. -Os isn't what most people want Nov 01 10:10:49 isn't Os virtually the same as O2? Nov 01 10:10:57 on /some/ hardware it can be a performance benefit, but if that isnt the case.. i'd still prefer -O2 for my devices. Nov 01 10:10:57 no. Nov 01 10:11:04 speed vs. space. so no Nov 01 10:27:31 Is there any easy way to clean an OE build tree of old packages that no longer exist in the repo? I find I end up with lots of revisions for each package, for example. Nov 01 10:27:53 Noodles: No easy way I know of :-/ Nov 01 10:28:31 * kergoth watches oe trip over itself trying to target x86_64 Nov 01 10:29:06 kergoth: I think Bernardo was having 64 bit problems but couldn't fix them :-/ Nov 01 10:29:29 note: i said /target/ Nov 01 10:29:31 :) Nov 01 10:29:39 Noodles: i use this ugly shell script http://en.pastebin.ca/27318 Nov 01 10:29:40 * kergoth adds the lib64 dirs to his packages Nov 01 10:29:52 * RP goes back to sleep :) Nov 01 10:29:56 hehe Nov 01 10:30:23 * Noodles applies serial-add-support-for-non-standard-xtals-to-16c950-driver.patch from -mm to 2.6.14-git3 in the hope of getting bluetooth working. Nov 01 10:30:37 I'm trying to do too much at once. But I now know c7x0 breakage in 2.6.14-git1 is rmk's fault... Nov 01 10:30:45 heh, fun Nov 01 10:30:49 * kergoth doesnt miss kernel work Nov 01 10:30:53 http://www.gunnarwrobel.de/wiki/rej.html .. thats nifty. Nov 01 10:31:22 emte: have you seen http://www.selenic.com/mercurial/wiki/index.cgi/MqExtension ? quilt integration into an scm tool :) Nov 01 10:31:30 much niftiness. Nov 01 10:31:48 going to link now Nov 01 10:32:01 there have been huge changes lately Nov 01 10:32:47 wow ... Nov 01 10:33:07 that sounds like it could be handy inside BB Nov 01 10:33:50 hmm Nov 01 10:34:04 maybe not Nov 01 10:34:28 taht would tie BB and the SCM together pretty tight, but allow more flexiblity Nov 01 10:35:25 would make possible to use a different version of individual packages than where your global repo is at Nov 01 10:35:25 in other news: glibc cvs has eabi support Nov 01 10:36:40 koen|770: any clue how to enable it in oe? Nov 01 10:37:32 target_os=linux-eabi Nov 01 10:37:33 kergoth, i saw someone complementing the hg<-> snv/svk convertions the other day Nov 01 10:37:41 svn* Nov 01 10:37:55 and make glibc_cvs.bb work Nov 01 10:38:27 i was just going to type "that sounds easy" Nov 01 10:38:46 the latter doesn't, actually (to me at least) Nov 01 10:39:41 I think we need to talk nicely to pb_ :) Nov 01 10:42:44 kergoth, you try the new hgk tool? Nov 01 10:45:00 RP: tried that yesterday without much result Nov 01 10:45:18 i guess we should try harder :) Nov 01 10:47:41 emte: not yet, need to play with hg more Nov 01 10:47:51 me either Nov 01 10:48:03 unfort the win client needs more work imo Nov 01 10:48:14 its not very portable atm Nov 01 10:48:49 so i have been stuck with cvs for most of my school work Nov 01 10:49:16 can anyone think of a short way to describe the opposite of "feature creep" ? Nov 01 10:49:30 i use svk for all sorts of things Nov 01 10:49:34 poor requirements? Nov 01 10:49:54 CosmicPenguin: nah, describing how a vendor keeps trimming features to meet deadlines Nov 01 10:50:10 kergoth, yeah i'd use taht as well but Visual Studio and MPLab dont have support by default Nov 01 10:50:58 ahh Nov 01 10:50:59 Twiun: feature pruning Nov 01 10:51:06 Twiun, value added Nov 01 10:51:10 lardman: thank you! Nov 01 10:51:13 * Twiun hugs lardman Nov 01 10:51:30 didnt he ask for the opposite? Nov 01 10:51:39 * lardman has done lots of phrase replacements lately - submitted thesis two hours ago Nov 01 10:51:48 lol Nov 01 10:51:52 emte: that is the opposite Nov 01 10:51:56 emte: opposite of creep Nov 01 10:52:04 hmm Nov 01 10:52:15 pruning might be too precise though.... :) Nov 01 10:52:17 good thing its not me writting it Nov 01 10:52:28 creep generally implies extra stuff added, pruning is def removal Nov 01 10:52:30 slashing sounds more correct... :) Nov 01 10:52:39 CosmicPenguin: hmmm Nov 01 10:52:49 maybe too bloody for an executive summary Nov 01 10:52:57 lol Nov 01 10:53:08 i liek regression myself Nov 01 10:53:13 like* Nov 01 10:53:18 already covered elsewhere :) Nov 01 10:53:37 "Slipping delivery deadlines, platform bugs, inadequate and sometimes non-existent developer support have severely impacted their time to market and developer's progress when using their platform." is one quote Nov 01 10:53:55 "In [19XX], [Company]'s platform had great promise to change the landscape of operating systems on [certain] devices. [5+] years later, [Company] has yet to deliver. Their technological lead is now being directly threatened by strong competitors aiming at the same market, using the same technology." is another Nov 01 10:54:12 * Twiun twists the knife Nov 01 10:54:43 Twiun: the same company with the stale tinyX sdk? Nov 01 10:54:47 Twiun, you talking about QNX? Nov 01 10:54:53 koen: nah Nov 01 10:54:55 emte: nope Nov 01 10:54:55 Sounded like AMD to me... :) Nov 01 10:55:13 * Twiun names no names Nov 01 10:55:57 RP: I've emailed SoC 0.4, lets hope it solves the dpm problems with the 50 and 31. I off home now. Nov 01 10:57:20 night all Nov 01 11:01:06 lrg: Ok, thanks. Fingers crossed :) Nov 01 11:01:48 RP: Do you have any objections to me adding serial-add-support-for-non-standard-xtals-to-16c950-driver.patch from -mm to 2.6.14-git3 for Zaurus? It makes my bluetooth CF card work. Nov 01 11:42:31 hmm. what's weird.. my irc program told me I was on this channel, but a whois of myself said no : ) Nov 01 11:42:37 in any case.. Nov 01 11:42:42 thanks for the catch XorA|gone : ) Nov 01 11:48:20 I'm trying to get kdepimpi-x11 compiled using bitbake.. and as such have to compile qt-x11-free. However, I'm getting the following error: however, I get the following error: http://oe.pastebin.com/413220 Nov 01 11:48:25 any ideas? Nov 01 11:56:44 Noodles: Sorry, got called away. No, I don't have an objection. Nov 01 12:00:28 is there a guide somewhere on the wiki for writing a bb? Nov 01 12:04:48 there was a 'newbb' page iirc Nov 01 12:05:12 ya...it doesn't provide very much info, unfortunately... Nov 01 12:05:14 thanks though Nov 01 12:06:49 what bbclass should I inherit is what I'm looking for... Nov 01 12:08:59 evening Nov 01 12:09:48 Koen, this is a dumb crosscompiling question...but...i am building something that needs libusb, when I am building, do I link to the arm libusb, or the libusb on my computer? Nov 01 12:10:04 arm Nov 01 12:10:51 ok...thank you Nov 01 12:11:00 hmm. getting a connection refused from monotone.vanille.de... could someone restart its monotone server or something? Nov 01 12:11:31 * Koen|770 looks at mickey|TV Nov 01 12:12:44 done Nov 01 12:13:10 koen: btw., you can still restart it - the server has moved to merlin.tm.informatik.uni-frankfurt.de, login/pw still the same Nov 01 12:13:10 mickey|TV: thanks Nov 01 12:13:13 np Nov 01 12:14:32 mickey|TV: ok Nov 01 12:15:44 orangey: which crossg++ version is that? Nov 01 12:16:22 03rw 07org.oe.dev * rcd6155a0... 10/: gpe-bootsplash: add new familiar bootsplash and bump revision of 1.14 Nov 01 12:17:03 mickey|TV: http://monotone.vanille.de/viewmtn/ doesn't work anymore Nov 01 12:17:25 mickey|TV: how do I check? Nov 01 12:17:53 koen: yep. it's broken since i moved the server to another host. can't repair it atm. Nov 01 12:18:02 orangey: hmm... you should know which version you built Nov 01 12:18:13 if not, /cross/bin/arm-linux-g++ -v Nov 01 12:18:16 (or so) Nov 01 12:18:46 mickey|TV: It claims 4.0.0 Nov 01 12:18:57 -r1 Nov 01 12:20:03 yep Nov 01 12:20:06 what i suspected Nov 01 12:20:19 qt 3.x doesn't build with g++ 4 Nov 01 12:20:32 you need to patch the packed structs away Nov 01 12:20:44 i can't remember where i put the patch Nov 01 12:21:41 try Nov 01 12:21:42 - inline ushort &unicode() { return ucs; } Nov 01 12:21:42 + inline ushort &unicode() { return *(&ucs); } Nov 01 12:21:59 (see gcc34 patch in qte-4 snap) Nov 01 12:27:56 mickey|TV: thank you. Nov 01 12:28:03 mickey|TV: should I just go back to gcc34? Nov 01 12:28:34 yes. 4.0.0 has lots of bugs and 4.0.2 is not in OE yet Nov 01 12:28:55 this is why 3.4.3 is selected as default in most distro .confs Nov 01 12:29:07 we are short on toolchain guys atm. Nov 01 12:29:08 *cough* Nov 01 12:30:20 mickey|TV: glibc_cvs has eabi support Nov 01 12:31:08 yep, read it. good news... if we would knew how to enable it ;) Nov 01 12:31:50 I think the enabling part is already in Nov 01 12:32:14 getting glibc_cvs to compile and make a snapshot is 'all' we need Nov 01 12:43:58 mickey|TV: where do I set it to be 3.4 instead of 4.0.0? Nov 01 12:44:58 ooooh. Nov 01 12:45:03 I see it. Nov 01 12:47:05 OK, I guess I shall have to rebuild everything then. Nov 01 12:49:12 kergoth: poke Nov 01 12:49:15 your umls 2.6.x ? Nov 01 12:49:45 * france is back (gone 05:34:54) Nov 01 12:51:35 hey hardwire Nov 01 12:51:42 long time no see Nov 01 12:53:11 hardwire: yep Nov 01 12:56:03 03noodles 07org.oe.dev * rfc95d462... 10/: Nov 01 12:56:04 linux-openzaurus: Add non-standard xtals patch from -mm to 2.6.14-git3; Nov 01 12:56:04 required for working Billionton Bluetooth CF support. Nov 01 13:01:25 kergoth: nice Nov 01 13:01:34 kergoth: which revision? Nov 01 13:01:46 * hardwire is thinking of updating the uml debian packages to use 2.6.x patches Nov 01 13:01:54 its neat.. they included mppe Nov 01 13:01:59 heh, i'm using 2.6.14 stock. no patches. Nov 01 13:02:01 in the 2.4.26 uml Nov 01 13:02:14 2.6.14 has a uml target? Nov 01 13:02:20 yep Nov 01 13:02:24 make ARCH=um .. as usual Nov 01 13:02:25 * hardwire feels something very wet inbetween his legs. Nov 01 13:02:40 and that supports x86_64 .. which 2.4 never did, afaik Nov 01 13:02:48 but you don't ahve x64 Nov 01 13:02:48 * kergoth is running a uml on his athlon 64 box atm Nov 01 13:02:49 do you? Nov 01 13:02:52 oha Nov 01 13:02:55 oh .. ha Nov 01 13:07:11 kergoth: where did you get the initial config? Nov 01 13:07:48 juse use its defconfig? Nov 01 13:08:19 yay for freebsd :[ Nov 01 13:08:36 what a great way to celebrate halloween: Nov 01 13:08:54 announcing the extirpation of the dreaded Daemon from their logo Nov 01 13:09:47 so they can be just like all the other totally lame genericized sanitized corporate wannabe forks Nov 01 13:10:54 what, no more daemon? wtf? Nov 01 13:11:02 i dont even use freebsd and i'll miss that Nov 01 13:11:03 :( Nov 01 13:12:19 next year openbsd will stop using a fish as a logo Nov 01 13:12:54 koen: :( Nov 01 13:18:15 I guess its more important to entice skittish god-fearing new users, than to worry about alienating pagan, goddess-loving old users Nov 01 13:19:03 I'll have to remember to cast Circle and banishment on any new BSD cd, before installing :[ Nov 01 13:19:07 * raduga bleargh Nov 01 13:21:21 afternoon kergoth Nov 01 13:28:09 http://logo-contest.freebsd.org/result/640-1.png Nov 01 13:28:16 it's still a bit daemony Nov 01 13:29:50 not very Nov 01 13:31:50 It really doesn't seem like a crowd that would bow to such political bs Nov 01 13:32:22 i think the official complaint was that daemons were "unprofessional" Nov 01 13:32:41 Yeah, so? Nov 01 13:32:48 and not providing the "image" that people want to associate with the core OS. Nov 01 13:33:14 raduga: lol Nov 01 13:35:05 We are lucky Linus had a run in with a penguin Nov 01 13:35:31 So cute penguin, professional. Cute demon, unprofessional Nov 01 13:35:33 got it Nov 01 13:35:45 http://www.gwaihir.org/etc/tux.html Nov 01 13:37:24 Lots of people are unhappy with Tux as well. Nov 01 13:37:34 satanic penguin? Nov 01 13:37:47 not satanic, but not really professional either. Nov 01 13:38:09 no less profressional than a numbe of other logos Nov 01 13:38:20 I dunno. :/ tux is my friend Nov 01 13:39:20 If you need your logo to project an appearance of professionalism, then you've already lost Nov 01 13:39:53 Crofton: those are good specifications. Nov 01 13:40:11 Crofton: I'm glad we have that sort of thing exhaustively documented. Nov 01 13:42:36 We might change memory management models every -rc, but damn if our logo isn't well specified... :) Nov 01 13:43:07 got a little bit of a OE problem today. Nov 01 13:43:16 I am trying to build bluez-utils-nodbus Nov 01 13:43:28 and I am having trouble getting through the pcmcia-cs build. Nov 01 13:43:42 | /export/home/build/oe-build/tmp/work/pcmcia-cs-3.2.8-r22/temp/run.do_patchcleancmd.1611: line 415: syntax error near unexpected token `'INHIBIT_PACKAGE_STRIP',' Nov 01 13:43:43 NOTE: Task failed: /export/home/build/oe-build/tmp/work/pcmcia-cs-3.2.8-r22/temp/log.do_patchcleancmd.1611 Nov 01 13:43:43 NOTE: package pcmcia-cs-3.2.8-r22: task do_patch: failed Nov 01 13:43:43 ERROR: TaskFailed event exception, aborting Nov 01 13:43:43 NOTE: package pcmcia-cs-3.2.8: failed Nov 01 13:43:44 ERROR: Build of pcmcia-cs failed Nov 01 13:44:12 uE: check your DISTRO Nov 01 13:44:24 pb_: I have DISTRO set to: Nov 01 13:44:42 nothing! Nov 01 13:44:51 grep DISTRO conf/local.conf Nov 01 13:44:51 #DISTRO = "familiar-0.8.3" Nov 01 13:45:02 uE: that is an unfortunate state of affairs. Nov 01 13:45:03 heh. Nov 01 13:45:48 when I set it to that value I get: Nov 01 13:45:55 ERROR: Unable to parse conf/bitbake.conf (conf/distro/familiar-0.8.3.conf:7: unparsed line: '"${@bb.fatal('This file is unmaintained in the org.oe.dev branch, use the org.openembedded.oz354fam083 branch instead. You can also remove this line, but don't complain if something breaks')}"') Nov 01 13:46:06 Which I assume is the price I pay for being on the bleeding edge. Nov 01 13:46:06 oh dear Nov 01 13:46:14 try just setting it to "familiar". Nov 01 13:46:30 or follow what the error tells you Nov 01 13:46:31 that will give you the same bbfatal Nov 01 13:46:38 ERROR: Remove this line or set a dummy DISTRO_VERSION if you really want to build an unversioned distro Nov 01 13:46:38 NOTE: exceptions.SystemExit:1 while evaluating: Nov 01 13:46:39 ${@bb.data.getVar("DISTRO_VERSION",d,1) or bb.fatal('Remove this line or set a dummy DISTRO_VERSION if you really want to build an unversioned distro')} Nov 01 13:46:44 there might be some other stupid check that objects to "familiar", but you might be able to defeat that by commenting it out. Nov 01 13:46:46 When the computer says to do something, do it! Nov 01 13:46:53 just remove the line from conf/distro/familiar*.conf Nov 01 13:47:10 ...and cross my fingers Nov 01 13:47:17 setting DISTRO='famimiliar' will use gcc 4 Nov 01 13:47:19 and break Nov 01 13:47:24 uE: yeah, that goes without saying Nov 01 13:47:28 0.8.3 should be reasonable safe Nov 01 13:48:39 alright, fingers crossed. Nov 01 13:48:54 very good Nov 01 13:56:12 that's *ahem* better: Nov 01 13:56:22 | ccache arm-linux-gcc -march=armv5te -mtune=xscale -I/export/home/build/oe-build/tmp/staging/arm-linux/include -fexpensive-optimizations -fomit-frame-pointer -frename-registers -O2 -I/export/home/build/oe-build/tmp/work/pcmcia-cs-3.2.8-r22/pcmcia-cs-3.2.8/include -MD -O3 -Wall -Wstrict-prototypes -pipe -I/export/home/build/oe-build/tmp/staging/arm-linux/include -I../modules -c lspnp.c Nov 01 13:56:23 | lspnp.c: In function 'dump_chain': Nov 01 13:56:23 | lspnp.c:499: error: invalid lvalue in assignment Nov 01 13:56:23 | lspnp.c:501: warning: pointer targets in return differ in signedness Nov 01 13:56:24 | lspnp.c: In function 'dump_resources': Nov 01 13:56:26 | lspnp.c:517: warning: pointer targets in assignment differ in signedness Nov 01 13:56:28 | lspnp.c:522: warning: pointer targets in assignment differ in signedness Nov 01 13:56:30 | lspnp.c:525: warning: pointer targets in assignment differ in signedness Nov 01 13:56:32 | make[1]: *** [lspnp.o] Error 1 Nov 01 13:56:34 | make[1]: Leaving directory `/export/home/build/oe-build/tmp/work/pcmcia-cs-3.2.8-r22/pcmcia-cs-3.2.8/debug-tools' Nov 01 13:56:37 | make: *** [all] Error 2 Nov 01 13:56:39 | + die 'oe_runmake failed' Nov 01 13:56:41 | + oefatal 'oe_runmake failed' Nov 01 13:56:43 | + echo FATAL: 'oe_runmake failed' Nov 01 13:56:45 | FATAL: oe_runmake failed Nov 01 13:56:47 | + exit 1 Nov 01 13:56:49 NOTE: Task failed: /export/home/build/oe-build/tmp/work/pcmcia-cs-3.2.8-r22/temp/log.do_compile.15805 Nov 01 13:57:18 03rw 07org.oe.dev * r88d7908c... 10/: gpe-screenshot: add initial public release 0.4. Nov 01 13:57:53 wahahha uE plewase use pastebin Nov 01 13:59:09 sorry newbie mistake Nov 01 14:00:49 everybody gets one Nov 01 14:02:22 has Linus used up his one allowed mistake yet? Nov 01 14:04:57 heh - I doubt this room will ever be darkened by Mr. Torvalds.. :) Nov 01 14:06:46 I've heard he's been sighted before on OPN/freenode but that may be apocryphal Nov 01 14:07:00 and almost certainly infrequent Nov 01 14:07:24 uE: sounds like you either need a better version of pcmcia-cs, or a worse version of gcc Nov 01 14:09:01 pb_: Agreed. Nov 01 14:10:47 uE: you can influence the latter by setting PREFERRED_VERSION_gcc-cross in your local.conf. Nov 01 14:10:56 the former might be more tricky Nov 01 14:12:13 pb_: I am currently trying to cut debug-tools out of my OE build. Hoping that it is the only offender. Nov 01 14:13:34 ah, good plan Nov 01 14:19:07 you might want to edit conf/familiar/familiar-0.8.3.conf since that specifies gcc preferences too Nov 01 14:21:26 kergoth: its odd needing libpcap-dev for a kernel Nov 01 14:21:41 hehe Nov 01 14:21:46 i usually dont enable the pcap networky bits Nov 01 14:21:49 yeh Nov 01 14:21:53 use tun/tap and the uml switch daemon, mostly Nov 01 14:21:56 I am doing this via make-kpkg Nov 01 14:21:59 ah Nov 01 14:21:59 I wonder how it will turn out Nov 01 14:22:09 will it try to install it in /boot :) Nov 01 14:22:20 should I make it use an initrd :) Nov 01 14:22:24 actually I kidna want it to Nov 01 14:22:27 so I can muck with yaird Nov 01 14:24:15 well Nov 01 14:24:28 I think locking down a config at 2.6.14 + asterisk CVS + a few other things should be groovy Nov 01 14:24:44 kergoth: whats your desktop distro nowadays? Nov 01 14:25:25 ubuntu breezy Nov 01 14:25:43 still no OE? Nov 01 14:25:46 breezy my sneezy weezy Nov 01 14:25:55 kergoth: think thats g/f friendly? Nov 01 14:26:27 very Nov 01 14:26:35 its a breeze to use, comes up with happy current gnome Nov 01 14:26:48 naturally i generally dont use that crap, its just a current debian to me Nov 01 14:26:53 I am going to build her a system soon Nov 01 14:27:00 dual boot.. Nov 01 14:27:07 because she wants to start ripping music Nov 01 14:27:12 * kergoth nods Nov 01 14:27:14 and Iw ant her to use sound--juicer + flac Nov 01 14:27:43 heh, i'm picky about my rips. I follow the ubernet standard. Nov 01 14:27:48 well yeh Nov 01 14:27:50 but not me Nov 01 14:28:09 I am goign to store it via flac then transcode to whatever mobile player we decide on layer Nov 01 14:28:18 I have two 200gb drives sitting around Nov 01 14:28:28 I am going to find another one and set up a raid-5 Nov 01 14:28:49 btw.. smartutils rules Nov 01 14:28:51 ubernet supports flac. you just have to use the known excellent ripping tools, which means windows or under wine. Nov 01 14:28:58 * hardwire is now going to know when things want to DIE! Nov 01 14:29:16 kergoth: I really want wine to just work (tm) Nov 01 14:29:22 like Photoshop + Illustrator CS Nov 01 14:29:24 i bought cedega Nov 01 14:29:29 play wow in linux Nov 01 14:29:30 works great Nov 01 14:29:31 :) Nov 01 14:29:33 yeh Nov 01 14:29:41 cedega I used to have debs for Nov 01 14:29:49 I just deleted them a few days ago for space Nov 01 14:29:50 doh Nov 01 14:29:56 heh Nov 01 14:30:03 torrents are your friends Nov 01 14:30:06 :P Nov 01 14:30:06 no vmware? Nov 01 14:30:32 http://www.aclerex.com/ Nov 01 14:30:43 heh.. their web designer reminds me of the terrasoft YDL guy Nov 01 14:30:46 limited. Nov 01 14:31:44 .. Nov 01 14:31:49 I'm an MMORPG fan currently playing World of Warcraft, and that runs perfectly under Cedega in Ubuntu. If only Photoshop CS would run too I'd make the switch completely, as it is I've kept dual boot with a small XP install. Shame as I was hoping to get rid of Windows altogether. Nov 01 14:32:06 hah Nov 01 14:32:47 well Nov 01 14:32:48 ah, removing debug-tools from pcmcia-cs seemed to do the trick. Nov 01 14:32:49 crossover office works Nov 01 14:33:50 its almost worth buying Nov 01 14:35:07 heh.. except illustrator doesn't work Nov 01 14:35:28 hehe.. this uml kernel has squashfs+unionfs+wrr Nov 01 14:37:42 ugh Nov 01 14:37:46 not all oe packages obey base_libdir Nov 01 14:37:54 i'm trying to /target/ x86_64 for my uml setup Nov 01 14:38:02 and the lib vs lib64 crap is melting my brain Nov 01 14:38:06 and oe as well Nov 01 14:38:10 * kergoth sighs Nov 01 14:38:41 doh Nov 01 14:38:56 I want to get more help making bulletdebian Nov 01 14:39:08 uE: very good Nov 01 14:39:11 for embedded devices that need rock-solid storage Nov 01 14:39:19 and can afford 256-2gig of rootfs Nov 01 14:39:23 i love debian, but its tools are nowhere near being suitable for embedded work Nov 01 14:39:27 vs the wittle bitty embedded market Nov 01 14:39:30 just the way it is Nov 01 14:39:34 kergoth: of course. Nov 01 14:39:44 thats why I need to define a word for.. not membedded yet not anything else. Nov 01 14:40:01 because embedded == 16 megs or less to most people Nov 01 14:40:07 and rrunning uclib Nov 01 14:40:38 I think I am just going to mount root via XFS.. freeze it and make a snapshot Nov 01 14:40:47 then if you wanna save the changes.. by all means Nov 01 14:40:51 but no factory default for you Nov 01 14:40:52 MWA HA HA Nov 01 14:41:25 perhaps if I could view the difference of the snapshot and the frozen fs.. I could set aside a list of original files.. Nov 01 14:41:38 and store that on the snapshot merge Nov 01 14:41:44 then I could restore to factory. Nov 01 14:42:05 this is a pain Nov 01 14:42:13 compile faster kernel.. I wanna dick around. Nov 01 14:48:38 Chicken: good stuff. Now don't touch it!! Nov 01 14:50:18 fucking debian Nov 01 14:50:35 * Twiun comforts debian Nov 01 14:50:37 now I have to recomipe this whole damn thing Nov 01 14:50:42 recompile Nov 01 14:50:43 heh Nov 01 14:51:52 what, debian? Nov 01 14:51:56 * Twiun grins Nov 01 14:51:58 holy shit Nov 01 14:52:02 it made the uml kernel correctly Nov 01 14:52:18 dpkg -c ../kernel-uml-2.6.14-1-um_2.6.14-1_i386.deb Nov 01 14:52:18 .. Nov 01 14:52:22 -rwxr-xr-x root/root 5414400 2005-11-01 12:51:48 ./usr/bin/linux-2.6.14-1-um Nov 01 14:52:32 and put the modules in the right place Nov 01 14:52:35 it will do the same for xen Nov 01 14:52:36 amazing Nov 01 14:52:46 make-kpkg --stem kernel --append-to-version=-1-um --revision=2.6.14-1 --arch=um --config menuconfig kernel_image Nov 01 14:52:49 thats my new favorite command Nov 01 14:53:15 heh, anything with "make-kpkg" in is my least favourite command Nov 01 14:54:22 I just want consistancy.. that seemed the easiest way :) Nov 01 14:54:31 kergoth: using a skas like patch at all? Nov 01 14:54:33 on the host Nov 01 14:55:06 03rpurdie 07org.oe.dev * r6ddc006f... 10/: linux-oz-2.6: update dev kernel to work on c7x0 by reverting problematic mainline patch + fix corgi oss sound driver Nov 01 15:11:35 03rpurdie 07org.oe.dev * rdb5d7086... 10/: udev: Add version 071. This fixes input problems seen on the zaurus 2.6 dev kernel with /dev/input/ Nov 01 15:26:49 03rpurdie 07org.oe.dev * rb007529b... 10/: xserver-kdrive_20050207: Add patch to speed up framebuffers which are software rotated in normal use. Applied to spitz, akita and borzoi. Nov 01 15:37:04 * koen waits for the glibc snapshot to finish fetching Nov 01 15:37:22 go koen! Nov 01 15:38:00 It will probably fail on patching, so don't get your hopes up Nov 01 15:38:25 pb_: eabi only has softvfp as softfloat, right? Nov 01 15:39:23 hola Nov 01 15:39:53 alguien en la red Nov 01 15:40:38 'night all Nov 01 15:40:49 kergoth: poke Nov 01 15:41:16 im a new user the opera Nov 01 15:41:30 and another spam attack on the wiki..... Nov 01 15:42:07 ok, Nov 01 15:45:56 arm-audit.patch failed, removing it Nov 01 15:46:26 kergoth: poke of miiighty neeed Nov 01 15:47:21 NOTE: package glibc-2.3.5+cvs20051101-r0: task do_configure: started Nov 01 15:47:24 sweetness Nov 01 15:47:31 only one failing patch Nov 01 15:51:24 koen: What about some more from bugs 426 and 427 :-) Nov 01 15:54:35 RP: doesn't that rotate speedup patch make sense for all platforms? Nov 01 15:58:42 koen: I'm told it is a slight slowdown for platforms that don't rotate Nov 01 16:01:19 looks like h3600, ramses, poodle and collie need it too Nov 01 16:01:55 judging from a look at http://handhelds.org/cgi-bin/cvsweb.cgi/gpe/base/xserver-common/X11/Xserver?rev=1.2&content-type=text/x-cvsweb-markup Nov 01 16:03:13 koen: Agreed. Feel free to update it Nov 01 16:03:44 * koen does so Nov 01 16:04:24 is 'linuxthreads' a usefull feature for glibc? Nov 01 16:05:25 * nickv111 would say, "Yeah, I'm pretty sure it's necessary" except that if he did, everyone would ridicule him because he's probably wrong and should stop giving advice Nov 01 16:06:07 koen: How could I find the correct DPI value for h6300? Nov 01 16:06:29 lamikr: I bet it's 100DPI Nov 01 16:06:29 linuxthreads is useful for anybody who isn't using 2.6, and for UML Nov 01 16:06:57 hmmm Nov 01 16:07:12 that means I have to figure out why it doesn't seem to work Nov 01 16:07:34 koen: so it built? Nov 01 16:07:37 Did it recently break? Nov 01 16:07:48 | configure: error: add-on directory "linuxthreads" does not exist Nov 01 16:07:54 NOTE: package glibc-2.3.5+cvs20051101: failed Nov 01 16:08:07 hmmm.... Nov 01 16:08:47 Koen, ok so no need to update dpi's. Nov 01 16:08:49 I built for linuxthreads back about early Septemberish, but I was using an older glibc Nov 01 16:09:32 CosmicPenguin: far out man Nov 01 16:10:04 * CosmicPenguin gives hardwire the peace sign Nov 01 16:10:07 heh Nov 01 16:10:15 * hardwire attempts to see if uml will boot this initfs Nov 01 16:12:01 CosmicPenguin: yah, linuxthreads has been part of the ports tree for a couple of months now Nov 01 16:16:25 03koen 07org.oe.dev * r57d918c4... 10/: xserver-kdrive 20050207: apply faster-rotate patch for more machines with a rotated framebuffer Nov 01 16:17:50 aha Nov 01 16:18:11 * koen expands do_munge with mv ports/linuxthreads linuxthreads Nov 01 16:18:29 seems to be working Nov 01 16:18:52 I would say get rid of it, if it wasn't for UML Nov 01 16:19:00 or 2.4 Nov 01 16:20:10 How many active 2.4 platforms do we have these days? Nov 01 16:20:31 h3600, h3900, collie, poodle Nov 01 16:20:49 linksys routers Nov 01 16:20:58 hmmmm Nov 01 16:21:02 so there you go... :) Nov 01 16:21:02 nptl Nov 01 16:21:41 is nptl usefull/needed? Nov 01 16:22:04 heh - nptl replaced linuxthreads, didn't it? Nov 01 16:22:21 * koen really has no idea Nov 01 16:22:46 replaced is a bad term - I think its the current preferred threading implementation Nov 01 16:22:58 pb_ could be more clear on the official stance though Nov 01 16:23:12 * koen adds --disable-sanity-checks for now Nov 01 16:24:04 CosmicPenguin: yeah, that's correct. nptl is a superior alternative to linuxthreads. Nov 01 16:24:25 however, nptl doesn't work on non-tls platforms. Nov 01 16:24:38 And UML and flavors of 2.4 don't have tls Nov 01 16:24:51 right Nov 01 16:24:58 I think some newer versions of 2.4 have tls back ported, do they not? Nov 01 16:25:44 not that I know of, though it's possible. Red Hat have been shipping a tls backport in their 2.4 kernels for ages, so the patches certainly exist. Nov 01 16:26:24 Some architectures (m68k, for example) don't have TLS support even in 2.6. Nov 01 16:26:56 so for 2.6 we'd one to use nptl and for 2.4 linuxthreads Nov 01 16:27:18 ARM didn't have it until quite recently either. I don't remember exactly when it was added, but I guess one of the h4x0rs would know. Nov 01 16:27:41 And I know the work is ongoing for mips Nov 01 16:27:54 I saw nico's tls patches flying by a few months ago Nov 01 16:28:10 including a remark from RMK about 'secure browsing' :) Nov 01 16:31:41 it seems to be compiling now, time for bed Nov 01 16:31:43 'night all Nov 01 16:45:56 any arm hackers that might be hanging around - does the current ARM linux tree have a spi bus driver? Nov 01 16:54:39 CosmicPenguin: hi Nov 01 16:54:53 what platform are you OE'ing in Nov 01 17:30:33 wonder if uclibc works with gcc4 Nov 01 18:05:10 hardwire: Geode and Alchemy Nov 01 18:15:46 CosmicPenguin: god writeups on geode work? Nov 01 18:15:51 I use geode gx1's Nov 01 18:15:55 I would love to see OE on them Nov 01 18:16:01 so I can stop trying to put debian on it Nov 01 18:16:07 its fruitfull.. but fruitless Nov 01 18:18:51 boy it is really hard to boot a uml with an initrd and no devfs Nov 01 18:19:42 shuldnt use devfs in 2.6 anyway Nov 01 18:19:54 Iknow Nov 01 18:20:01 but I think the d-i initrds require it Nov 01 18:20:07 so i am going to like.. hack them Nov 01 18:21:01 initrd is old school too. Nov 01 18:21:02 heh Nov 01 18:21:17 kergoth: trying to just bootstrap debian into a uml Nov 01 18:21:21 without using debootstrap Nov 01 18:21:26 or rootstrap Nov 01 18:21:33 or d-i's scripts Nov 01 18:21:38 just to be a fucker Nov 01 18:21:49 right.. good luck. Nov 01 18:21:58 I have no intentions of it working Nov 01 18:22:02 I just want to be a fucker. Nov 01 18:22:12 what dist are you running via uml? Nov 01 18:22:38 right now its a hacked up oe distro. when i was using 2.4, it was striaght out of oe Nov 01 18:22:50 neat Nov 01 18:23:02 hardwire, different from this? http://www.ex-parrot.com/~pdw/debian-uml-install.html Nov 01 18:23:04 well, it wasnt 2.4 vs 2.6.. its just the 32 bit vs 64 bit stuff thats biting me Nov 01 18:23:31 emte: I was just looking at that Nov 01 18:23:33 I may bite Nov 01 18:26:22 emte: trying to find the sarge equivilents Nov 01 18:27:28 there doesnt seem to be much there that looks distro specific Nov 01 18:29:52 heh Nov 01 18:29:58 I got the boot initrd.gz working Nov 01 18:30:02 it wants to find the root initrd now Nov 01 18:34:05 hardwire: I don't think I've put OE on a GX1 Nov 01 18:34:09 but heck, its not like its complicated Nov 01 18:34:22 I have been out of the game sooo long Nov 01 18:34:55 CosmicPenguin: are you using a jffs2 based root? Nov 01 18:35:05 hardwire: no, strictly ATA Nov 01 18:35:12 hmm Nov 01 18:35:22 I should do an i386 install sometime very soon Nov 01 18:35:33 I lied - on the Alchemy we're doing jffs2, but then thats not x86... :) Nov 01 18:35:39 assuming kergoth.. that OE will be long lived ? Nov 01 18:35:57 seems to have a life of its own Nov 01 18:36:01 heh Nov 01 18:36:24 koen and zecke have been fine caretakers of the torch Nov 01 18:37:52 i havent touched it in 6 months to a year Nov 01 18:38:00 definately reached critical mass Nov 01 18:42:26 back Nov 01 18:42:53 hi hardwire CosmicPenguin kergoth Nov 01 18:42:59 hmmm Nov 01 18:43:08 hey Nov 01 18:43:47 hi ggilbert Nov 01 18:44:21 kergoth: I might have a new job soon ... there http://www.gdcanada.com/ Nov 01 18:44:22 so plan a hasnt been working Nov 01 18:44:28 maybe plan b will work Nov 01 18:48:44 I wonder how long till OE switches to something faster than monotone Nov 01 18:51:28 i wanted to set up an svk mirror to encourage people to move to that, but i never got around to fixing the bugs in tailor to move our history over Nov 01 18:51:42 stupid real work! :P Nov 01 18:51:59 always getting in the way of stuff that could be put off due to laziness Nov 01 18:52:25 I'm actually using OE for some stuff internally Nov 01 18:52:57 but I just ripped bits and pieces out and stopped tracking the monotone upstream since it got to be way too slow Nov 01 18:54:19 same Nov 01 18:54:31 its pretty sad that if you get very far behind at all, even days, its faster to just redownload the db Nov 01 18:54:35 theres something seriously wrong there. Nov 01 19:02:00 hehe Nov 01 19:02:05 monotone not working out? Nov 01 19:06:00 somebody forgot deoderant Nov 01 19:06:03 * hardwire runs home hiding his pits Nov 01 19:20:07 kergoth: mercurial is suppose to be very similar to bitkeeper Nov 01 19:22:46 hey mithro Nov 01 19:24:42 i should be in bed Nov 01 19:24:44 gnight all Nov 01 19:58:02 ok Nov 01 19:58:08 I got uml to boot Nov 01 20:00:59 installing from netboot iso Nov 01 20:01:57 oh FAIL! Nov 01 20:01:59 heh Nov 01 20:39:57 hhey all. Nov 01 20:40:30 quick q. I have my distro set as "familiar-0.8.3", but I get: DISTRO = "familiar" when I am compiling Nov 01 21:11:51 03jbowler 07org.oe.dev * r1671361e... 10/: nslu2-kernel: 50-nslu2-arch.patch now matches l-a-k format rules in 2.6.14 Nov 02 00:33:10 moin Nov 02 00:43:25 morning Nov 02 01:28:23 kergoth: ok.. dm_snapshot is 100% what I need Nov 02 01:28:30 but its not something like unionfs.. where I can just squash a root fs and map a different one on top Nov 02 01:29:00 but I expect that with xfs/ext3 set as r/o w/ root.. and a r/w snapshot set up inside of that lvm.. things will work fine Nov 02 01:29:17 factory reset just deletes the snapshot and starts another one. Nov 02 01:29:28 merging isn't 100% an option however Nov 02 02:07:23 ~lart glibc cvs Nov 02 02:07:24 * ibot_ calls glibc cvs on the phone ... the lights are on but nobody's home Nov 02 02:19:40 morning all Nov 02 02:21:06 Noodles: Upgrading udev will fix the /de/input issue Nov 02 02:22:11 Yeah, I saw that changeset. Nov 02 02:22:28 Plus the xdrive one. Nov 02 02:22:35 s/x/k/ Nov 02 02:22:40 m Nov 02 02:22:44 morning all Nov 02 02:23:07 hi Bernardo Nov 02 02:23:08 RP: new udev version? Nov 02 02:23:12 hi RP Nov 02 02:23:29 Bernardo: -git3 kernel needs udev 071 which I added last night Nov 02 02:23:47 071 was the one that nearly borked my laptop... Nov 02 02:24:15 I had two udev rule files with duplicated stuff between them, 071 just ignored both instead of executing them in order Nov 02 02:24:56 anyway, time to rebuild a image... :) Nov 02 02:25:11 Sounds like it might break the devfs.rules then. I don't think any sane person wants to use devfs though... Nov 02 02:25:16 hi Nov 02 02:25:25 I should have some kernel updates coming in shortly for sound... Nov 02 02:25:26 and find out why yesterday I ended up without a /lib/modules/2.6.14-git3... Nov 02 02:25:27 hi hrw|work Nov 02 02:25:28 hi hrw|work Nov 02 02:25:41 RP: git3-r1 still not booting on my c760 ;( Nov 02 02:25:50 RP: and I do need to check your sd code Nov 02 02:25:54 Bernardo: 2.6.14-git3 < 2.6.14-rc4 Nov 02 02:26:13 hrw|work: You recompiled with the revert patch? Nov 02 02:26:29 RP: recompiled an hour ago Nov 02 02:26:45 RP: I know, but I cleaned up all stuff from staging, stamps and work that had a 2.6.xx before rebuilding Nov 02 02:26:54 ${RPSRC}/revert_bootmem-r0.patch;patch=1 \ Nov 02 02:27:07 Bernardo: Did you remove the ipks from deploy? Nov 02 02:27:11 Bernardo: but ipkg wont upgrade your modules on device Nov 02 02:27:28 RP: about the sd, sharp's closed sd driver spits out a lot of crap about timeouts with this 1Gb 60x sd card Nov 02 02:27:53 RP: don't know, maybe that is the problem... :) Nov 02 02:28:00 it was late last night Nov 02 02:28:03 Bernardo: That figures. At least its not just 2.6 that has issues with it... Nov 02 02:28:20 hrw|work: clean flash Nov 02 02:28:25 hrw|work: I'm at a loss to know what's wrong then :-( Nov 02 02:28:43 I'll be trying a c7x0 kernel again shortly after a complete rebuild... Nov 02 02:28:52 RP: but sharp's driver flashes those timeout messages for two secs and then recovers. Do you want the syslog from that? Nov 02 02:29:24 Bernardo: I doubt it would help me much - I think the problem is that timeout code I mentioned... Nov 02 02:29:49 Bernardo: At some point I'll try and fix it but it will have to wait for now Nov 02 02:29:54 ok Nov 02 02:30:12 I'll have to live using a usb card reader for now, then... :) Nov 02 02:31:00 * Bernardo needs some time to start kernel hacking too Nov 02 02:31:51 Bernardo: I had timeout problem months ago (patch was in OE) - maybe more timeout it needed then is in kernel? Nov 02 02:32:54 hrw|work: here the sd card causes kernel panics, and RP thinks it's his timing code Nov 02 02:32:59 could be it Nov 02 02:33:23 I can try tuning it here, if I get a couple hours time today Nov 02 02:34:05 hrw|work: That's sort of the idea but someone's pointed me at code which suggests you need a two level timout for modern code Nov 02 02:34:21 -git3 now fails on spitz with the revert code Nov 02 02:34:24 ah.. Nov 02 02:34:28 * RP cries Nov 02 02:34:54 fsck... I forgot to take card reader... no method to write something on sd or cf at work ;( Nov 02 02:35:26 RP: that means it might also fail on akita, no? Nov 02 02:35:45 Bernardo: its likely, yes Nov 02 02:36:05 RP: can you point me at that two-level timeout code? No promises, but I'd like to take a look Nov 02 02:36:24 Bernardo: No code, just a reference in a manual Nov 02 02:36:51 ok Nov 02 02:37:06 the sd org now allows access to it's specs, right? Nov 02 02:40:35 bbl Nov 02 02:41:45 Bernardo: This timeout issue is well described by the s3c2410 manual. Which you can find here: chapter 19, page 19-16. Nov 02 02:41:45 http://www.samsung.com/Products/Semiconductor/SystemLSI/MobileSolutions/MobileASSP/MobileComputing/S3C2410X/um_s3c2410s_rev12_030428.pdf Nov 02 02:41:59 That's a quote from linux-arm-kernel Nov 02 02:43:55 I have to start using OE to build for x86... Nov 02 02:44:12 want to build <64M rescue distro for pendrive Nov 02 02:49:41 03koen 07org.oe.dev * rfe07d6fa... 10/: h6300: use udev in machine description and kernel, patch courtesy Mika Laitio. closes bug 426 Nov 02 02:49:46 03koen 07org.oe.dev * ra4b55a49... 10/: h6300 kernel: add 2.6.14-rc5 Nov 02 02:52:41 hrw|work: I think I've corrupted that revert patch Nov 02 02:53:15 RP: the worst thing is that I cannot even test it today ;( Nov 02 02:54:42 RP: did the copypage locks go away in 2.6.13 or 2.6.14? Nov 02 02:57:42 koen: I can't remember offhand... Nov 02 02:58:34 let's say 2.6.14 doesn't have them :) Nov 02 02:58:45 koen: That would be a safe comment :) Nov 02 02:59:57 RP: you should teach pavel 'bitbake -b linux-openzaurus.bb -c fetch' **** ENDING LOGGING AT Wed Nov 02 02:59:57 2005