**** BEGIN LOGGING AT Sun Sep 18 02:59:56 2005 Sep 18 03:22:49 good morning :) Sep 18 04:56:14 03ph5 07org.openembedded.dev * r789079a5... 10/packages/udev/ (8 files in 2 dirs): udev: recipe and patches for udev-070 Sep 18 06:11:47 03ccsmart 07org.openembedded.dev * rd57da27a... 10/packages/postfix/postfix_2.0.20.bb: postfix: remove temporary file /var/tmp/main_cf.sh after installation Sep 18 06:11:56 03ccsmart 07org.openembedded.dev * r75f6e838... 10/packages/bogofilter/ (files/bogohelper.sh bogofilter_0.96.0.bb): bogofilter: add bogohelper.sh and master.cf entries Sep 18 06:46:19 03rpurdie 07org.openembedded.dev * rd4dacf3f... 10/packages/hostap/hostap-modules-0.4.4/kernel_updates.patch: hostap: Add fixes so it builds correctly for kernels > 2.6.13 Sep 18 07:33:52 hi all Sep 18 07:34:05 x1205 i2c part done Sep 18 07:34:11 now the rtc part Sep 18 07:35:50 it seems there a common rtc infrastructure into arm Sep 18 07:35:56 but not of widespread user Sep 18 07:35:58 use. Sep 18 07:56:25 03mickeyl 07org.openembedded.dev * rf7245f1f... 10/packages/ubahnnav/ (ubahnnav_0.3.1.bb ubahnnav_0.4.0.bb ubahnnav_0.4.1.bb): ubahnnav: upgrade to 0.4.1 and remove obsolete versions Sep 18 07:56:28 03mickeyl 07org.openembedded.dev * re957dec8... 10/packages/meta/opie-image.bb: opie-image needs to DEPEND on task-bootstrap Sep 18 08:16:18 03mickeyl 07org.openembedded.dev * ra8e888a8... 10/packages/pcmciautils/ (pcmciautils_007.bb pcmciautils_010.bb): pcmciautils: upgrade to 010 Sep 18 08:31:29 03ph5 07org.openembedded.dev * rb9817db6... 10/packages/linux/handhelds-pxa-2.6_2.6.12-hh3.bb: handhelds-pxa-2.6: make 2.6.12-hh3 use the oe defconfig Sep 18 08:46:17 03rpurdie 07org.openembedded.dev * r23ad0768... 10/packages/orinoco/spectrum-fw.bb: spectrum-fw: Fix a path bug in the .bb. Thanks to lardman for pointing it out. Sep 18 09:01:22 03marceln * 10unslung/Makefile: Corrected typo tftp-hpa Sep 18 09:03:34 dwery: hi, your patchset for 2.6.13.1 worked perfectly for me! Sep 18 09:03:56 yvasilev: great Sep 18 09:03:58 ! Sep 18 09:06:07 dwery: have you made any progress with the NIC, how can I help with it where should I start? Sep 18 09:06:45 yvasilev: i've not approched the problem at all. i'm working on x1205 right now Sep 18 09:06:49 * yvasilev has very limited experience with the kernel, but is anxious to learn Sep 18 09:06:54 yvasilev: and i'm half way to it Sep 18 09:07:40 dwery: I'll try to forward port the patch doem 1.4 to 2.0 for the access lib Sep 18 09:08:23 ok! Sep 18 09:13:10 i have compiled dhcpcd and would like to make a package of it, how do i do it? ipkg? Sep 18 10:19:51 yvasilev: You could also take a look at the snapgear patchset - it has something for v2.0 Sep 18 10:20:00 yvasilev: http://ftp.snapgear.org/pub/snapgear/src/snapgear-modules-20050914.sh Sep 18 10:29:15 Jacmet: hello. thats for the usbnet tip... Sep 18 10:29:18 thanks. Sep 18 10:29:20 :) Sep 18 10:29:50 dwery: you're welcome Sep 18 10:36:11 I'm now testing x1205 separation Sep 18 10:49:19 well.. it's working. Sep 18 10:49:32 nice. Sep 18 10:49:37 Jacmet: thanks for the tip, looking :-) Sep 18 10:55:00 <[g2]> Jacmet, thx for the url are you running the 2.0 stuff ? Sep 18 10:55:17 so.. who wants to test this "new" x1205 driver? Sep 18 10:56:13 [g2]: no, I just recently saw the snapshot stuff in a discussion on linux-arm-kernel - I haven't tried it yet Sep 18 10:56:23 dwery: you can send it to me Sep 18 10:56:25 ok Sep 18 10:56:40 <[g2]> Jacmet we're working on setting up a build farm Sep 18 10:57:03 dwery: I'm also interested in cleaning up the kernel patch - the little endian support also needs a bit of cleanup Sep 18 10:57:12 [g2]: for debian/armeb? Sep 18 10:57:17 <[g2]> yes Sep 18 10:57:53 <[g2]> I'd like to support both BE and LE for the Loft Sep 18 10:57:58 <[g2]> it's running BE now Sep 18 10:58:07 [g2]: it'll be obsolete if I get NPE running in little endian ;) Sep 18 10:58:16 [g2]: ok Sep 18 10:58:20 <[g2]> don't think so Sep 18 10:58:37 <[g2]> you're still doing endian swapping on every packet Sep 18 10:58:45 <[g2]> it's a performance issue Sep 18 10:59:15 [g2]: I don't think that's really such a big issue at 10/100mbit Sep 18 10:59:25 <[g2]> CPU time is CPU time Sep 18 10:59:29 Jacmet: well, you're the LE guy :) Sep 18 11:00:13 <[g2]> the NSLU2 doesn't have that much CPU to waste Sep 18 11:00:14 [g2]: yes, do you plan to recode everything in xscale optimized asm as well? ;) Sep 18 11:00:37 <[g2]> personally I'd like to Sep 18 11:00:43 <[g2]> I'd go armv5 Sep 18 11:00:57 <[g2]> and target xscale Sep 18 11:01:08 dwery ;) The endian swap when accessing flash from user space is still a bit messed up - I haven't had time to fix it Sep 18 11:01:14 <[g2]> we'd pickup thumb and some minor other stuff Sep 18 11:01:15 03hrw 07org.openembedded.dev * r1fa7381f... 10/packages/ttf-fonts/ (ttf-dejavu_1.13-20050820.bb ttf-dejavu_1.14.bb): ttf-dejavu: upgraded to 1.14 Sep 18 11:01:44 <[g2]> howevever, I think the currently plan is to build the vanilla ARM in BE mode Sep 18 11:02:05 [g2]: ofcause it matters a bit, and there's room for both LE/BE - but ease of use of LE is more important than 1-2% higher performance for me Sep 18 11:02:29 <[g2]> it's much higher than 1-2% Sep 18 11:02:33 [g2]: going to softfloat would gain more performance I think Sep 18 11:02:50 <[g2]> cause is VFP ? Sep 18 11:02:53 <[g2]> it's Sep 18 11:03:11 <[g2]> or NFP Sep 18 11:03:28 * [g2] can't keep track of them all Sep 18 11:04:22 [g2]: sure? I don't see why it would be any significant performance issue - it's just for a few header fields Sep 18 11:04:54 [g2]: don't remember either - but the exception handling / emulation is quite some overhead compared to softfloat Sep 18 11:05:04 <[g2]> right Sep 18 11:05:14 <[g2]> so the standard ARM isn't softfloat ? Sep 18 11:05:28 [g2]: debian/arm is hardfloat Sep 18 11:05:38 <[g2]> ahh... Sep 18 11:05:52 [g2]: it gives about 10-20% not having fpu exeptions Sep 18 11:05:58 which is a bit silly asm arm-with-float isn't really common Sep 18 11:05:58 <[g2]> so that's the real reason we should probably target v4 soft-float Sep 18 11:06:07 yeah Sep 18 11:06:17 <[g2]> and build for both LE and BE Sep 18 11:06:26 but you'll probably have to stay hardfloat to get accepted into debian Sep 18 11:06:59 but ok, if you want to recompile all 16K packages ;) Sep 18 11:07:08 why care about debian? Sep 18 11:07:36 kolla: because you don't want to maintain those 16K packages yourself? Sep 18 11:07:45 <[g2]> kolla even if gentoo was fully done, native compiling is excessive for most users Sep 18 11:08:12 <[g2]> Jacmet I'm seriously considering having my own build farm Sep 18 11:08:16 jacques: I have no need for those 16k packages Sep 18 11:08:17 [g2]: gentoo works just fine on nslu2, no need for modificatios Sep 18 11:08:18 kolla: that's the big advantage of the current debian/LE dist Sep 18 11:08:19 <[g2]> but then I'm selling hw Sep 18 11:08:50 <[g2]> yvasilev are you running gentoo on an NSLU2 ? Sep 18 11:08:56 yes Sep 18 11:09:02 <[g2]> vapiers port ? Sep 18 11:09:28 jacques: I dont see it as an advantage Sep 18 11:09:28 <[g2]> we gave solar the hw who gave it to vaiper Sep 18 11:09:30 kolla: you may need 25 of thos epackages, another one 351, and a third one even 1K packages.. that's the reasone to use debian. Sep 18 11:09:36 no, my le port, but vapiers be works as well Sep 18 11:09:36 [g2]: ok Sep 18 11:09:40 most of the packages are junk Sep 18 11:09:58 <[g2]> yvasilev so you've built stage1's ? Sep 18 11:10:10 yes Sep 18 11:10:26 <[g2]> native or cross with catalyst ? Sep 18 11:10:32 kolla: what is junk for ou, may be gold for another. you can't know for sure. Sep 18 11:10:46 Jacmet: rtc patch sent Sep 18 11:10:48 naitive Sep 18 11:10:58 [kolla]: others might Sep 18 11:11:12 most likely not :) Sep 18 11:11:28 and there's heaploads of stuff that is missing in debian due to the political issues Sep 18 11:11:29 kolla: see it as an advantage? I beg to differ Sep 18 11:11:30 [g2]: I actualy had build them some time ago on softgun, and now just use them on nslu2 Sep 18 11:11:33 <[g2]> yvasilev glibc or uClibc and hardened or not ? Sep 18 11:11:48 [g2]: uClibc not hardened Sep 18 11:11:58 kolla: given the attention my page got Sep 18 11:12:11 <[g2]> ok, so it's really not much more the uCslugc Sep 18 11:12:18 dwery: ok, thanks Sep 18 11:12:20 <[g2]> but that builds entirely cross Sep 18 11:12:27 kolla: if debian is missing something, you bet YOU are missing something when recompiling the packages you need :D Sep 18 11:12:46 <[g2]> dwery not true Sep 18 11:13:05 <[g2]> OE is quite advanced over debian in terms of patches Sep 18 11:13:06 [g2]: either you know the needs of all your users Sep 18 11:13:12 I'm not missing anything :) Sep 18 11:13:17 [g2]: or you are missing something Sep 18 11:13:32 [g2]: yes, glibc should also work, but there is not much sense in that until nslus come with 512mb+ Sep 18 11:13:38 [g2]: since i think that knowing the requirements "a priori" is quite difficult.... Sep 18 11:14:14 <[g2]> dwery I was commenting that each distro has something to offer Sep 18 11:14:23 [g2]: obvioulsy. Sep 18 11:14:24 yvasilev: heh, I'm doing glibc gentoo on slug :) Sep 18 11:14:26 <[g2]> there is no real king of distros Sep 18 11:14:39 [g2]: the bigget the distro, the lesser chance it has forgot something :) Sep 18 11:14:40 exactly Sep 18 11:14:54 <[g2]> however, I think in two years time there will be 1 master metadata for all Sep 18 11:15:04 dwery: no, the bigger the distro, the slower it adapts Sep 18 11:15:10 kolla: if you also want to try uclibc, the softfloat stages are on mirrors ;-) Sep 18 11:15:18 kolla: true.. but debian adapt to 11 different archs... Sep 18 11:15:56 dwery: yeah, like m68k where they're still paddling 2.2 kernels due to the macs :) Sep 18 11:16:00 <[g2]> well my last comment is on the one-distro for all approach Sep 18 11:16:07 kolla: touche :) Sep 18 11:16:32 dinner's time... se ya later Sep 18 11:16:34 <[g2]> clearly debian ARM is not well tuned for the NSLU2 and modern Xscale arches Sep 18 11:16:40 well, my m68k is not stuck with debian, and hence runs a flashy 2.6.13 kernel and stuff Sep 18 11:16:44 <[g2]> dwery, enjoy! Sep 18 11:16:56 <[g2]> bon ap. Sep 18 11:17:17 kolla: what do you use on your m68k? Sep 18 11:17:20 I also tried depian on mipsel without much success :) Sep 18 11:17:44 yvasilev: gentoo Sep 18 11:17:45 <[g2]> all the openwrt stuff was pushing the mipsel technology Sep 18 11:18:40 also runs gentoo on my zaurus and my gamecube and xbox Sep 18 11:18:48 run Sep 18 11:19:00 kolla: nice! Sep 18 11:19:33 <[g2]> Jacmet so are you building everything native ? Sep 18 11:28:14 [g2]: I'm not building anything, that's the point ;) Sep 18 11:28:26 [g2]: but yeah, I have compiled the kernel cross and native Sep 18 11:48:51 kolla: so you're compiling stuff for days before using it? ;) Sep 18 11:49:56 Jacmet: depends Sep 18 11:51:20 Jacmet: glibc takes forever, that's true, but I dont do that very often, rebuild of apache+php is done in a day or so Sep 18 11:51:32 which is fine by me Sep 18 11:52:08 not like I sit and watch over it anyways Sep 18 11:53:08 kolla: apache+php=1day is on nslu2? Sep 18 11:53:16 heh, exactly - gentoo on embedded/old hw doesn't make much sense to me Sep 18 11:53:28 yvasilev: no, on the m68k Sep 18 11:53:40 my old amiga a1200 that is Sep 18 11:54:11 kolla: ah, that makes much more sense Sep 18 11:54:16 Jacmet: makes perfect sense to me - with gentoo you can keep an oddball system self sustained Sep 18 11:54:20 as long as portage works Sep 18 11:54:52 and some version of gcc :) Sep 18 11:54:57 kolla: you can keep it busy compiling, that's true ;) Sep 18 11:55:03 yesterday we were talking about usb-audio not working thru usb hubs.. where can I add this info to the wiki? Sep 18 11:55:30 kolla: how much ram? Sep 18 11:55:53 dwery: ohh. is that not just an issue with buggy usb2 hubs (audio cards are normally usb1.1)? Sep 18 11:55:56 Jacmet: well, I did run debian on it earlier.. it filled the machine to the brinks with crap I had no need for, it didnt work particularly well, and having changes done in packages took for ever, I still had to compile most of the stuff I needed anyways Sep 18 11:57:02 Jacmet: no, it seems a special support from th ehub driver is required. Sep 18 11:57:16 some packages I cross compile though, like new kernel and kernel-headers Sep 18 11:57:45 dwery: hmm, ok Sep 18 12:00:01 kolla: well, if you have room for gcc + all the devel packages, "all the crap" debian installs shouldn't really be an issue, right? Sep 18 12:00:13 Jacmet: http://article.gmane.org/gmane.linux.alsa.devel/25841 Sep 18 12:01:45 dwery: ahh, I had the same problem - I assumed it was simply my hub that was slightly broken Sep 18 12:03:00 Jacmet: well, I had to have gcc + all anyways, since debian didnt have what I needed Sep 18 12:03:16 btw - how come you cannot post to the nslu2 lists from gmane? is it just yahoo being broken? Sep 18 12:03:48 would anyone be insterested in moving the lists to a proper listserver? Sep 18 12:03:48 but I also got humti revision of various libs and whatnot Sep 18 12:04:06 Jacmet: I hate yahoo... Sep 18 12:04:15 Jacmet: google would be better :) Sep 18 12:04:16 and compiling was a mess - I still find building debian packages a mess Sep 18 12:04:38 dwery: yes, or on a normal mailman/ezmlm-managed list Sep 18 12:05:06 Jacmet: yes, but that one requires an adiministrator :) Sep 18 12:05:31 Jacmet: unless hosted in one of those public mailman servers Sep 18 12:05:42 kolla: what's wrong with apt-get source --compile ? Sep 18 12:06:08 dwery: shouldn't be a problem - I'm sure I could get it hosted on dotsrc.org Sep 18 12:06:32 Jacmet: well, duh - when I can do that I just as well use the binary package provided by debian Sep 18 12:07:04 kolla: then your std ./configure && make Sep 18 12:07:04 much too often I have to create a package from scratch, or tinker around with someone elses ugly build Sep 18 12:07:29 Jacmet: and then when.. make install? Sep 18 12:07:59 dwery: do you know what needs to be enabled in the kernel for the usb audio/hub stuff? Sep 18 12:08:13 I much more prefer to have a system that let me easily make my own builds the way I want them, and keep track of the files are they are installed into the system Sep 18 12:08:13 Jacmet: you mean, to fix that bug ot to test usb audio? Sep 18 12:08:23 dwery: to fix that bug Sep 18 12:08:35 Jacmet: I have absolutely no idea. Sep 18 12:08:36 s/are/as/ Sep 18 12:09:01 Jacmet: but I guess code needs to be added to the usb hub driver. given that that bug is at least one year old... Sep 18 12:09:34 dwery: hmm, but there isn't anything nslu2 specific about the usb stuff (except for the IRQ mappings) Sep 18 12:09:46 Jacmet: it's not an NSLU bug.. is an USB bug Sep 18 12:10:15 <[g2]> Jacmet, did you build cross with crosstool or something else ? Sep 18 12:10:26 Jacmet: i mean, linux usb layer Sep 18 12:10:33 Jacmet: but by all means, keep on using debian if that's your thing :) Sep 18 12:10:34 dwery: ok, but it works on my desktop through the hub, and not on the nslu2? Sep 18 12:10:46 [g2]: for the kernel? yes, the old crosstool setup Sep 18 12:10:55 Jacmet: the problem is when you have USB2 computer, USB2 hub, USB1 audio Sep 18 12:11:02 kolla: and the reverse for you ;) Sep 18 12:11:03 Jacmet: is your computer USB2? Sep 18 12:11:18 dwery: exactly - that's what I have both places Sep 18 12:11:27 Jacmet: oh. Sep 18 12:11:49 Jacmet: my laptop is USB1 and works. Sep 18 12:11:56 dwery: that's the odd thing Sep 18 12:12:02 Jacmet: yeah. Sep 18 12:12:09 Jacmet: I hear rumours that debian might be dropping many of the packages and architectues to get a faster pace on its releases Sep 18 12:12:15 Jacmet: i'm going to try with my other computer Sep 18 12:13:14 kolla: they are discussing what is needed to make a release - E.G. if everything is ready except m68k they might decided to release without it. Once it is ready it will then also be released Sep 18 12:13:41 for all I care they can drop m68k alltogether :) Sep 18 12:14:21 kolla: well, it sounds like for all you care debian can probably stop alltogether ;) Sep 18 12:14:37 sounds good :) Sep 18 12:14:56 ;) Sep 18 12:15:04 just kidding :) Sep 18 12:15:24 we ended up with debian at work, and it sometimes drives me nuts Sep 18 12:15:50 * Jacmet is pretty sure gentoo would drive him nuts as well if he was forced to use it ;) Sep 18 12:16:34 but anyway, there's plenty of room for all the dists (even on nslu2) - so go ahead with your gentoo port if you feel like it Sep 18 12:17:04 well, I wasnt about to break this glibc build just yet anyways :) Sep 18 12:17:33 ;) Sep 18 12:18:13 how's the situation with armel and the ethernet of the slug? Sep 18 12:18:54 kolla: someone is working on it Sep 18 12:19:16 <[g2]> Jacmet your openslug-2.6.11.2 patch is just the aggregated Openslug patches correct ? Sep 18 12:19:25 so it is possible it might work some day? that's nice Sep 18 12:20:00 [g2]: yes, from the 1.2 release Sep 18 12:20:15 <[g2]> 1.2 ? Sep 18 12:20:32 kolla: yes, there's to reason why it couldn't work - someone just has to do the work Sep 18 12:20:34 kolla: it WILL have to work, I need it :) Sep 18 12:20:40 [g2]: openslug 1.2 release Sep 18 12:21:14 <[g2]> ok, that's a little while ago Sep 18 12:21:20 [g2]: it's quite some time ago I did any work on it, that's why it's based on 2.6.11 and that old openslug release Sep 18 12:21:43 Jacmet: you're right. so it must be a slug problem Sep 18 12:21:45 [g2]: but as dwery has just shown, the patches apply fine to latest openslug / 2.6.13.1 Sep 18 12:21:58 Jacmet: ...with minor mods :) Sep 18 12:22:01 dwery: yes, and that's quite odd Sep 18 12:22:05 would be nice if I could have the slug and the zaurus share packages Sep 18 12:22:13 <[g2]> I forget where the serial issue was Sep 18 12:22:14 dwery: yeah ;) Sep 18 12:22:29 <[g2]> 11.8 somewhere to .12 Sep 18 12:22:37 Jacmet: how would you start debugging it? Sep 18 12:22:49 <[g2]> you probably wouldn't notice that :) Sep 18 12:24:05 Jacmet: I think the thing is quite buried in the USB layer.. probably the controller the slug uses has different behaviour that others on most PCs Sep 18 12:24:28 [g2]: do you have USB2 on your board, isn'it? Sep 18 12:24:46 <[g2]> dwery, sure two USB 2.0 ports Sep 18 12:24:55 [g2]: any usb audio dongle? Sep 18 12:25:09 <[g2]> no but there's MiniPCI Sep 18 12:25:25 [g2]: that won't help us to fix the bug :) Sep 18 12:25:33 dwery: yes, apparently - but all the ehci controllers should be compatible Sep 18 12:25:43 Jacmet: ...should.... Sep 18 12:25:50 dwery: "should" is ofcause not the same as they are ;) Sep 18 12:25:51 <[g2]> dwery, the hub bug ? Sep 18 12:26:09 <[g2]> that's an ALSA thing right ? Sep 18 12:26:12 [g2]: yeah. i would like to try on another arm machine Sep 18 12:26:16 yes, alsa. Sep 18 12:26:49 <[g2]> didn't you mention the bug reference or was that someone else Sep 18 12:26:56 kolla: I guess you can Sep 18 12:27:20 [g2]: Tiersten iirc Sep 18 12:27:21 kolla: do you have extra storage on your zaurus? otherwise debian doesn't make much sense Sep 18 12:36:15 03ph5 07org.openembedded.dev * r7ff4c9ec... 10/packages/linux/handhelds-pxa-2.6_2.6.12-hh3.bb: handhelds-pxa-2.6_2.6.12-hh3.bb: add a missing backslash Sep 18 12:39:40 so... 2.6.13.1 done, x1205 splitted.. what to do now.. ixp? Sep 18 12:39:51 4xx eth support? Sep 18 12:40:42 update to 2.6.13.2 ;) Sep 18 12:41:06 dwery: 4xx eth support! I'l lost at it :-( Sep 18 12:41:14 aaaaaaaaaaaargfh Sep 18 12:41:18 2.6.13.2! Sep 18 12:41:25 well... the path looks small :) Sep 18 12:42:16 i guess it shouldn't be a problem. 2.6.14 will. Sep 18 12:42:30 yvasilev: tell me about your experience Sep 18 12:43:32 dwery: can't make intell access lib behave, the patches look reasonably easy to port, but can figure out how ial works Sep 18 12:43:52 yvasilev: did you tried 1.5 or 2.0? Sep 18 12:43:58 2.0 Sep 18 12:44:16 Jacmet: do you have experience with ial? 2.0? Sep 18 12:44:32 now i'm loocking at snapgear path Sep 18 12:49:06 dwery: not really, I gave it a quick hack without luck sometime around may, and I haven't touched it since Sep 18 12:50:01 Jacmet: was it 1.5 or 2.0? Sep 18 12:51:22 dwery: both I think, but I'm not completely sure Sep 18 12:52:04 Jacmet: so.. yvasilev and me are on this now.. anyone else? Sep 18 12:53:33 dwery: not that I'm aware of atleast Sep 18 12:53:49 dwery: but I'm very interested in hearing about it ;) Sep 18 12:53:53 :) Sep 18 12:55:46 <[g2]> Jacmet what would it take for you to try a NSLU2 with an APEX loader that boots the box in LE mode ? Sep 18 12:56:16 <[g2]> and loads the kernel from jffs2 Sep 18 12:57:09 <[g2]> or even a Redboot with 1.5 built in Sep 18 12:57:48 I keep having this issue, that when i flash openslug, it doesnt respond to pings Sep 18 12:58:20 <[g2]> which openslug ? Sep 18 12:58:24 <[g2]> 2.6 ? Sep 18 12:58:28 .5 Sep 18 12:58:53 <[g2]> static or DHCP ? Sep 18 12:58:59 I think the way i got it to work last time was to erase all the settings, and try again Sep 18 12:59:01 yvasilev: plese give me some pointers and i'll start studying ial Sep 18 12:59:02 (via redboot) Sep 18 12:59:03 [g2]: a bit of time ;) Sep 18 12:59:38 <[g2]> Jacmet I'm not sure I know what you mean by that Sep 18 12:59:46 and static ip Sep 18 12:59:52 [g2]: I still need to solder on a jtag connected, but that shouldn't be a problem - I'll try to get around to do it next week Sep 18 13:00:08 [g2]: what would you like to test? that you can boot a LE kernel from it? Sep 18 13:00:36 [g2]: you can grab the kernel image from my homepage if you cannot wait for me to get around to do it Sep 18 13:01:09 dwery: I downloaded the snapgear patches, as they look good and also patch_accesslib_kernel26_20040811.diff. but now cant make ial brake out of dep loop when making anything Sep 18 13:01:15 <[g2]> Jacmet are you planning to change the bootloader out ? Sep 18 13:01:25 <[g2]> and just boot straight LE ? Sep 18 13:02:00 snapgear's apply cleanly, and just looks like you need to redefine IX_XSCALE_SW, LINUX_SRC and IX_TARGET Sep 18 13:02:06 [g2]: not right away, redboot works o - but it could perhaps be interesting Sep 18 13:02:11 the later one should be linuxle Sep 18 13:02:12 yvasilev: from which ial you started? Sep 18 13:02:41 yvasilev: nice - now lets see if it actually works ;) Sep 18 13:02:44 IPL_ixp400AccessLibrary-2_0.zip, IPL_ixp400NpeLibrary-2_0.zip Sep 18 13:03:26 the network patches are: GPL_ixp400LinuxEthernetDriverPatch-1_4.zip GPL_ixp400LinuxIntegrationPatch-1_3.zip, but still not sure they are needed Sep 18 13:03:37 [g2]: the problem about changing to apex is that it is one more step for others to take - and I'mtrying to make it easier to install, not harder Sep 18 13:04:54 <[g2]> Jacmet so you'd like to go to a Redboot with LE built in Sep 18 13:05:16 yvasilev: i'm downloading them Sep 18 13:05:33 [g2]: why le in redboot? Sep 18 13:05:46 <[g2]> dwery because Jacmet runs LE Sep 18 13:05:59 <[g2]> it would just turn the NSLU2 into a LE debian box Sep 18 13:06:01 [g2]: me too. but be redboot works just fine Sep 18 13:06:52 [g2]: no, I would prefer to just stay with the std BE redboot - it works fairly ok Sep 18 13:06:55 <[g2]> My point is if you boot the box in LE mode then there are no LE patches except for the tool chain Sep 18 13:07:32 [g2]: which patches? le patches to the kernel? Sep 18 13:07:46 [g2]: le and be toolchains are identical, just the default changes Sep 18 13:07:49 <[g2]> right now you've got an extra patch to switch to LE and you're byte swapping the kernel Sep 18 13:07:51 [g2]: booting in LE mode only eliminated a oneliner, the rest is still neded Sep 18 13:08:09 [g2]: you still need to access the PCI controller + flash correctly Sep 18 13:08:32 [g2]: byteswapping is only required if you do not flash it Sep 18 13:09:42 <[g2]> yvasilev, the le and be tool chains are not identical, there are patches we've had to do to make BE work Sep 18 13:09:49 to broaden debian adoption there are two things to do: Sep 18 13:09:58 a) built-in ethernet Sep 18 13:10:05 b) installation without serial Sep 18 13:11:13 <[g2]> if you guys created an image (kernel plus rootfs) then you could just web flash ti Sep 18 13:11:26 <[g2]> it's really straight forward Sep 18 13:11:54 <[g2]> that solves b) right ? Sep 18 13:11:59 [g2]: yeah, but it would be nicer to use the std installer - it already supports all of this Sep 18 13:12:27 b) is fairly easy - just a matter of recompiling the installer with the right options Sep 18 13:12:29 [g2]: identical in a sense that they are capable to crossgenerate code (el with eb and eb with el), the patches should have bin to make them generate non faulty eb code, and they should/can be applied to eb as well as el Sep 18 13:13:17 ok, off to bed - see ya Sep 18 13:13:31 <[g2]> yvasilev I'm just stating the difference between theory and practice as I've experienced the nearly 1.5 years with nslu2-linux Sep 18 13:13:39 <[g2]> sweet dreams Sep 18 13:17:22 yvasilev: downloaded the first two Sep 18 13:17:33 dwery: :-) Sep 18 13:17:59 yvasilev: tell me about the two GPLxx patches Sep 18 13:18:53 dwery: the GPL_ixp400LinuxIntegrationPatch-1_3.zip is just to integrate the ial into the kernel build process and should be left out Sep 18 13:20:14 GPL_ixp400LinuxEthernetDriverPatch-1_4.zip is like a little add on to GPL_ixp400LinuxIntegrationPatch-1_3.zip so no use for that too, plus a ixp400_eth.c which we weill probably need Sep 18 13:20:32 mmm... Sep 18 13:20:35 let me see Sep 18 13:20:35 ixp400_eth.c looks like it's the main file for the module Sep 18 13:21:48 and when building ial, I get stuck in the "Building dependency by defining IX_DEPS=1 in make command" part which the main Makefile never leaves Sep 18 13:22:25 [g2]: I realized my problem, I always forget that openslug defaults to using dhcp, even if a static IP was set Sep 18 13:22:42 <[g2]> CompBrain Great! Sep 18 13:24:56 my HD is getting super duper hot Sep 18 13:25:01 and its probably not good. Sep 18 13:28:02 yvasilev: what you did with ixp400_eth.c? copied somewhere? Sep 18 13:28:15 dwery: not jet Sep 18 13:29:44 dwery: I also tried to make a make file for it and add all the deps it needed from ial by hand, but that's clearly not the way to go (too many spaghetti like deps in ial code) Sep 18 13:31:00 yvasilev: i'm starting with a simple make.. immedialty got a hell of errors. :) Sep 18 13:31:31 make for ixp400_eth.c ? Sep 18 13:31:40 no, ial Sep 18 13:31:53 which directory are you uilding in? Sep 18 13:32:29 redefine the vars I mentioned, and comment their original definitions Sep 18 13:32:44 ixp400_xscale_sw Sep 18 13:33:11 have you patched with snapgear's patch? Sep 18 13:33:51 not yet Sep 18 13:34:45 IX_TARGET=linuxle Sep 18 13:34:54 LINUX_SRC=path.... Sep 18 13:35:41 IX_XSCALE_SW := /mnt/misc/kernel/work/b/ixp400_xscale_sw Sep 18 13:35:42 LINUX_SRC := /mnt/misc/kernel/2.6.13.1/linux-2.6.13 Sep 18 13:35:42 IX_TARGET := linuxle Sep 18 13:38:51 mmm.. gcc needs to be redefined too Sep 18 13:39:25 there is LINUX_CROSS_COMPILE like in the kernel Sep 18 13:41:22 compilation started.. it seems they forgot some symbols.. maybe it's time to try the patch you suggested Sep 18 13:42:11 IXP425_TIMER_BASE_PHYS Sep 18 13:42:50 where can I get snapgear's patch? Sep 18 13:43:38 http://ftp.snapgear.org/pub/snapgear/src/snapgear-modules-20050914.sh Sep 18 13:43:57 thx Sep 18 13:51:49 i have dissecte dthe patches Sep 18 13:51:54 and applying one by one Sep 18 13:51:59 when I encounter an error Sep 18 13:52:42 dwery: how do you split them automatically? Sep 18 13:53:23 midnight commander Sep 18 13:53:31 you click enter on th e.path Sep 18 13:53:36 .patch Sep 18 13:53:43 and it will open it for you as a directory Sep 18 13:53:50 extremely useful Sep 18 13:53:59 nice! didn't knew that! Sep 18 13:54:59 those guys at Intel... tsk tsk :) Sep 18 13:55:07 did they ever compiled it? :) Sep 18 13:55:21 LOL Sep 18 13:55:23 :-) Sep 18 13:58:08 dwery: youre trying for the builtin ethernet in little endian mode? (for debian) Sep 18 13:59:22 widrone: yep Sep 18 13:59:59 guys.. ixp_osal compiled... now what? :) Sep 18 14:01:12 i applied: Sep 18 14:01:13 ixp400-2.0/ixp_osal/os/linux/include/platforms/ixp400/IxOsalOsIxp400.h.diff Sep 18 14:01:13 ixp400-2.0/ixp_osal/os/linux/src/core/IxOsalOsThread.c.diff Sep 18 14:01:13 ixp400-2.0/ixp_osal/os/linux/src/core/IxOsalOsMsgQ.c.diff Sep 18 14:01:13 ixp400-2.0/ixp_osal/os/linux/include/platforms/ixp400/ixp425/IxOsalOsIxp425Sys.h.diff Sep 18 14:01:13 ixp400-2.0/ixp_osal/os/linux/src/modules/ioMem/IxOsalOsIoMem.c.diff Sep 18 14:01:21 and did Sep 18 14:01:33 make IX_TARGET=linuxle LINUX_SRC=/usr/src/test IX_XSCALE_SW=/tmp/ixp400_xscale_sw LINUX_CROSS_COMPILE=armv5b-softfloat-linux- libosal Sep 18 14:04:10 yvasilev: can you reach my same point? Sep 18 14:04:32 dwery: compiling Sep 18 14:04:39 dwery: compilled!!! Sep 18 14:05:43 ok. can you confirm you applied the same set of patches? Sep 18 14:06:23 I applied the whole snapgear patch, let me reproduce with just this ones Sep 18 14:06:45 ok Sep 18 14:15:51 Im getting an error running asterisk, undefined symbol, ast_park_call Sep 18 14:16:21 get error about -mshort-load-bytes, after removing this flag it compiles with said patches Sep 18 14:16:42 ok Sep 18 14:16:54 i've not quite well understood what has been built :) Sep 18 14:17:21 [6~ Sep 18 14:17:47 well, we are an unknown distance closer to the goal ;-) Sep 18 14:17:53 :) Sep 18 14:18:54 any idea on what to do now? Sep 18 14:19:09 compile in ixp400_xscale_sw Sep 18 14:20:13 ok, let's try it Sep 18 14:21:17 same procedure, i'll keep track of the applied patches Sep 18 14:21:40 ok, and I'm trying with the complete lot Sep 18 14:25:12 <[g2]> keeping track of applied patches is a good thing (tm) :) Sep 18 14:25:22 dwery: do you get that endless dep generation? Sep 18 14:28:22 hmm Sep 18 14:28:32 slug just slowed down alot, and im flashing orange Sep 18 14:28:41 no.. still applying patches Sep 18 14:28:59 load average of 17 Sep 18 14:31:25 dwery: trying make ixp400.o gives some results Sep 18 14:31:48 will try to make a module with this ixp400.o Sep 18 14:31:52 wow Sep 18 14:32:10 The DebianSlug/DiskImageInstall wiki page is pointing to some file that are pw protected. Anybody have links to these images (kernel/disk) Sep 18 14:32:16 write down the procedure Sep 18 14:32:46 widrone: http://peter.korsgaard.com/articles/debian-nslu2.php Sep 18 14:32:51 widrone: some of them are there Sep 18 14:33:55 yvasilev: i'm building with make IX_TARGET=linuxle LINUX_SRC=/usr/src/test LINUX_CROSS_COMPILE=armv5b-softfloat-linux IX_DEVICE=ixp425 IX_XSCALE_SW=. Sep 18 14:34:08 but i got stuck with this: Sep 18 14:34:18 Makefile:853: lib/linuxle/oslinux/IxLinuxInit.d: No such file or directory Sep 18 14:34:18 cd ./src/npeDl && \ Sep 18 14:34:18 gcc -I./src/include ixNpeDlImageConverter.c -o src/npeDl/ixNpeDlImageConverter Sep 18 14:34:18 In file included from ixNpeDlImageConverter.c:44: Sep 18 14:34:18 IxNpeMicrocode.c:56:28: error: IxNpeMicrocode.h: No such file or directory Sep 18 14:34:37 Microcode is in a separate download Sep 18 14:34:37 unpack IPL_ixp400NpeLibrary-2_0.zip Sep 18 14:34:48 doh! Sep 18 14:35:10 un[acked Sep 18 14:35:16 but the .h is still missing Sep 18 14:35:53 it's there, you need to unpack it over ial Sep 18 14:36:17 i found it,.. probably some problems with include paths Sep 18 14:37:50 got ixp400.ko, load, but no interface with ifconfig -a :-( Sep 18 14:38:04 any debug message? Sep 18 14:38:36 just ixp400: module license 'unspecified' taints kernel. Sep 18 14:38:53 kolla: if you have gentoo working on nslu2, where is the page for others to install it? Sep 18 14:38:58 i;ve got ixp400.o too Sep 18 14:39:25 how did you build the .ko? Sep 18 14:39:40 http://rafb.net/paste/results/DrpwqJ95.html Sep 18 14:40:05 put this make file in ixp400_xscale_sw/lib/linuxle, cd ixp400_xscale_sw/lib/linuxle && make Sep 18 14:40:48 Hmm - looks like nslu2-linux turned into comp.sys.embedded.advocacy overnight :-) Sep 18 14:41:47 yvasilev: thank. i've got it. Sep 18 14:41:54 yvasilev: now is the turn of the ethernet part Sep 18 14:42:33 modules/ixp425/net-2.0/ixp400_eth.c Sep 18 14:42:37 do you have it? Sep 18 14:43:29 the one from GPL_ixp400LinuxEthernetDriverPatch-1_4.zip? Sep 18 14:44:37 i think so. n my case, it has been created by snapgear-20050914.patch Sep 18 14:44:51 just patch ls Sep 18 14:45:02 ups Sep 18 14:46:40 Jacmet: are you working on the debian-installer that doesn't require serial ? Sep 18 14:46:52 yvasilev: there's also a Makefile Sep 18 14:46:54 hmm Sep 18 14:47:02 yvasilev: but i think it have to be adapted Sep 18 14:47:07 There's a lot of stuff going on with patches for the ixp42x? Sep 18 14:47:42 This is gonna be fun Sep 18 14:47:48 NAiL: wanna help? :) Sep 18 14:48:00 heh, I'm not exactly the one to ask, I'm afreid Sep 18 14:48:02 afraid Sep 18 14:48:07 NAiL: :) Sep 18 14:48:08 dwery: got it, was not applaing that patch Sep 18 14:48:12 I'll just have to get it to work on the ds101 as well Sep 18 14:48:24 yvasilev: that patch seems structured for snapgear integration Sep 18 14:48:39 yvasilev: not easy to extract the relevant parts Sep 18 14:49:28 yvasilev: kbuild needs to be involved i think. your opinion? Sep 18 14:50:03 let me give it a good look, i'm not as fast as you ;-) Sep 18 14:50:12 yvasilev: :) Sep 18 14:50:43 dwery, yvasilev: it's great to see you guys tackliing the ixp module Sep 18 14:53:28 dwery: how about if we try to make ixp400_eth.c into the module we already have? Sep 18 14:53:45 mmm Sep 18 14:54:01 you mean ixp400_eth.c inside the other one? Sep 18 14:54:05 rwhitby: :) Sep 18 14:54:48 yes, there is no real xp400.c, so it could take it's place Sep 18 14:55:21 mmm Sep 18 14:56:07 i would prefer keeping the original build system Sep 18 14:56:14 give me a few more minutes on that Makefile Sep 18 14:56:23 look at modules/ixp425/Makefile Sep 18 14:57:46 yes Sep 18 15:00:14 i'm getting some results: Sep 18 15:00:33 tell me, tell me! :-D Sep 18 15:00:54 i'll paste rfb Sep 18 15:01:58 create a file called Kbuild in the same directory as the _eth.c Sep 18 15:02:02 with http://rafb.net/paste/results/QrEjb828.html Sep 18 15:02:16 mpdify CSRDIR Sep 18 15:02:28 with the path of your ixp lib Sep 18 15:02:54 in the same directory, build with something like that make CROSS_COMPILE=armv5b-softfloat-linux- -C /usr/src/test M=`pwd` CONFIG_IXP400_ETH=m Sep 18 15:02:58 where -C ... Sep 18 15:03:02 is the path to your kernel Sep 18 15:03:08 since you have already applied all the patches Sep 18 15:03:10 it should build Sep 18 15:06:30 dwery: get lots of errors about not finding include files, rechecking CSRDIR Sep 18 15:07:47 me too Sep 18 15:07:51 strange, it seem sok Sep 18 15:08:09 yes Sep 18 15:09:44 enabling KBUILD_VERBOSE I can see that includes are ok, but it still can't find the files Sep 18 15:10:32 anyone mind me adding /sbin/buzzer to the rootfs? Sep 18 15:11:02 yvasilev: delete the line with ROOTDIR Sep 18 15:11:59 it was empty and thus generating a wrong invocation to gcc Sep 18 15:12:02 /mnt/misc/kernel/work/b/modules/ixp425/net-2.0/ixp400_eth.c:68:30: modules/autoconf.h: No such file or directory, so we need to include Sep 18 15:13:01 noticed. where's that file located? Sep 18 15:13:37 cant find it, but can it be linux/autoconf.h form the kernel? Sep 18 15:14:44 it seems it got added by snapgear's patch Sep 18 15:15:05 yes, i would try with that one Sep 18 15:15:19 /mnt/misc/kernel/work/b/modules/ixp425/net-2.0/ixp400_eth.c:3483: warning: implicit declaration of function `cancel_db_maintenance' Sep 18 15:15:32 and many *** Warning: "ixNpeDlNpeStopAndReset" [/mnt/misc/kernel/work/b/modules/ixp425/net-2.0/ixp400_eth.ko] undefined! Sep 18 15:15:39 or not... Sep 18 15:15:46 I think there where exported by ial, Sep 18 15:16:20 autoconf.h is created by Sep 18 15:16:22 congi.in Sep 18 15:16:25 config.in Sep 18 15:16:29 in the modules directory Sep 18 15:16:38 ah Sep 18 15:18:09 cancel_db_maintenance(void) is defined in ixp400_eth.c itself Sep 18 15:18:37 so it's only a prototypeproblem. Sep 18 15:19:03 void cancel_db_maintenance(void); Sep 18 15:20:05 have you found how to generate modules/autoconf.h? Sep 18 15:20:43 no, but i think we can substitute it with an epty file for now... Sep 18 15:21:57 i just removed the include, there is still the problems about undefined symbosl, i think we should link against the ial too, not just include its headers ;-) Sep 18 15:23:07 see this: Sep 18 15:23:17 CC [M] /tmp/s/modules/ixp425/net-2.0/ixp400_eth.o Sep 18 15:23:18 /tmp/s/modules/ixp425/net-2.0/ixp400_eth.c: In function `dev_hard_start_xmit': Sep 18 15:23:18 /tmp/s/modules/ixp425/net-2.0/ixp400_eth.c:2318: warning: implicit declaration of function `clean_dcache_range' Sep 18 15:23:18 /tmp/s/modules/ixp425/net-2.0/ixp400_eth.c: In function `dev_change_mtu': Sep 18 15:23:18 /tmp/s/modules/ixp425/net-2.0/ixp400_eth.c:2466: warning: unused variable `priv' Sep 18 15:23:19 /tmp/s/modules/ixp425/net-2.0/ixp400_eth.c: In function `ethacc_uninit': Sep 18 15:23:21 /tmp/s/modules/ixp425/net-2.0/ixp400_eth.c:2690: warning: unused variable `res' Sep 18 15:23:23 /tmp/s/modules/ixp425/net-2.0/ixp400_eth.c: In function `cleanup_module': Sep 18 15:23:25 /tmp/s/modules/ixp425/net-2.0/ixp400_eth.c:3484: warning: implicit declaration of function `cancel_db_maintenance' Sep 18 15:23:28 Building modules, stage 2. Sep 18 15:23:30 pastebin ;-) Sep 18 15:23:30 MODPOST Sep 18 15:23:32 /bin/sh: line 1: 9782 Segmentation fault scripts/mod/modpost -i /usr/src/test/Module.symvers vmlinux /tmp/s/modules/ixp425/net-2.0/ixp400_eth.o Sep 18 15:23:35 make[1]: *** [__modpost] Error 139 Sep 18 15:23:37 make: *** [modules] Error 2 Sep 18 15:23:41 segfault! Sep 18 15:23:43 :) Sep 18 15:24:04 I don't get it Sep 18 15:25:12 i'll try with your makefile Sep 18 15:27:00 dwery: no, yours looks quire correct, we just need to link agains ial Sep 18 15:27:09 http://rafb.net/paste/results/auY3Wp62.html Sep 18 15:28:15 your gcc goes past the point mine breaks Sep 18 15:28:19 also look at ixp400_xscale_sw/src/ethAcc/IxEthAccSymbols.c Sep 18 15:28:22 what suite are you using? Sep 18 15:29:08 gcc 3.4.4 Sep 18 15:29:20 are you building natively? Sep 18 15:29:22 le form gentoo Sep 18 15:29:24 yes Sep 18 15:29:27 oh Sep 18 15:29:27 ok Sep 18 15:29:58 i'm on 3.3.4 not natively Sep 18 15:30:48 so it loock like we need to finish building ial, we don't have objects for that functions jet Sep 18 15:31:35 i think the target is libosial Sep 18 15:31:43 libosal Sep 18 15:31:52 but Sep 18 15:31:55 we already did that Sep 18 15:32:25 or maybe not? anywaye I just relaunched make in that directoy Sep 18 15:32:28 no, this are defined in ixp400_xscale_sw/src/eth* Sep 18 15:33:42 * yvasilev is making the rest of ixp400_xscale_sw Sep 18 15:34:17 which target? :) Sep 18 15:35:08 just make Sep 18 15:35:45 no Sep 18 15:35:51 no Sep 18 15:35:54 need to find a target Sep 18 15:35:57 i thik is a matter of CONFIG_IXP425_COMPONENT_ETHACC Sep 18 15:36:00 and other Sep 18 15:36:04 which are not defined Sep 18 15:37:09 this we can find from GPL_ixp400LinuxIntegrationPatch-1_3.zip, it modifies def-config Sep 18 15:38:39 or you can add the componenbts to BI_ENDIAN_COMPONENTS in Makefile Sep 18 15:40:30 it compiles more object files Sep 18 15:40:36 yes Sep 18 15:40:38 but since my modpost keep segfaulting Sep 18 15:40:45 i can't proceed any further :( Sep 18 15:41:34 this message is from makefile: Sep 18 15:41:36 Makefile: Building Ixp400 library containing components "oslinux" "ethAcc" Sep 18 15:41:37 there is still no lib with that functions defined :-( Sep 18 15:41:59 with libIxp400 target Sep 18 15:42:17 the lib ends up in ixp_osal/lib/ixp425/linux/linuxle/ixp_osal.o Sep 18 15:42:50 where are you making from? Sep 18 15:43:24 _sw Sep 18 15:43:38 make IX_TARGET=linuxle LINUX_SRC=/usr/src/test IX_XSCALE_SW=/tmp/ixp400_xscale_sw LINUX_CROSS_COMPILE=armv5b-softfloat-linux- libIxp400 Sep 18 15:45:53 can not get that Makefile: Building Ixp400 library containing components "oslinux" "ethAcc", and instead i'm getting Assembler messages: FATAL: can't create lib/linuxle/npeDl/IxNpeMicrocode.o: No such file or directory Sep 18 15:46:44 (creating the dir by hand worked, still no message) Sep 18 15:47:43 good evening. does anyone have any info about getting usb audio working on unslung? only solid info I've found seems to be for openslug Sep 18 15:47:44 i have that file Sep 18 15:47:53 i've built a lot of targets randomly :) Sep 18 15:49:23 dwery: have you added -mlittle-endian to CFLAGS and -EL to AS/LD? Sep 18 15:49:32 rwhitby: i;ve read your proposal to debian. nice move! Sep 18 15:49:36 yvasilev: well, not Sep 18 15:49:49 that may be why your toolchain is miss behaving Sep 18 15:49:57 shouldn't IX_TARGET be enough? Sep 18 15:50:18 not, it's expect le toolchain Sep 18 15:50:26 i'll try Sep 18 15:50:42 in which files did you added them? Sep 18 15:51:05 main Makefile Sep 18 15:51:45 bot in _sw and in _osal? Sep 18 15:52:26 you shoul add to all of them Sep 18 15:52:28 you're right!!!! Sep 18 15:52:52 got Makefile: Building Ixp400 library containing components "oslinux" Sep 18 15:53:05 ok. now add the other components by hand in the makefile Sep 18 15:53:06 dwery: thx - you're an important part of that proposal - getting the latest kernel patches upstream for nslu2 :-) Sep 18 15:53:32 rwhitby: i'll do my best :) talked today with i2c maintainer Sep 18 15:53:45 rwhitby: got him to review a preliminary of x1205 Sep 18 15:53:55 excellent - a great step Sep 18 15:54:09 dwery: what do you gave in BI_ENDIAN_COMPONENTS? Sep 18 15:54:28 yvasilev: for now, I only added ethAcc Sep 18 15:54:38 yvasilev: probably need others Sep 18 15:56:07 dwery: btw, do you have an le croos toolchain at hand, it'll be easier with that one for you? Sep 18 15:56:16 Hrm, have the wiki pages for how to build unslung been moved? This page (http://www.nslu2-linux.org/wiki/Unslung/HomePage) says to read the "GettingStarted" page but the link takes me to $Group/$Name which doesn't look like a valid link (and takes me to just http://www.nslu2-linux.com/) Sep 18 15:56:17 back later Sep 18 15:57:20 yvasilev: i only have one Sep 18 15:57:51 yvasilev: can not make it recognize BI_ENDIAN_COMPONENTS Sep 18 15:57:59 stripwax: fixed, thx. Sep 18 15:58:41 oops, fixed now Sep 18 16:01:15 yvasilev: i have BI_ENDIAN_COMPONENTS-y := oslinux ethAcc Sep 18 16:05:14 thanjs Sep 18 16:06:42 src/ethAcc/IxEthAccSymbols.c:54:30: modules/autoconf.h: No such file or directory Sep 18 16:06:57 yvasilev: create an empty file Sep 18 16:07:28 I think we should define it with CONFIG_IXP425_COMPONENT_ETHACC=y or something like that Sep 18 16:07:46 yeah Sep 18 16:08:17 given that it's used only in the Makefile Sep 18 16:08:27 if you include ethAcc in BI_ENDIAN_etc it should work Sep 18 16:08:47 ok Sep 18 16:19:05 yvasilev: re: IxNpeMicrocode message, did you actually get the NPE Microcode file ? Its a separate download. Sep 18 16:20:10 dyoung-zzzz: yes, it compiled fine, you can get the source form intel, the file is IPL_ixp400NpeLibrary-2_0.zip Sep 18 16:20:45 Yes, I know. Sep 18 16:23:56 dwery: how can I make the module link against libIxp425.a with your Kbuild file? Sep 18 16:24:31 yvasilev: i thimk the besy way is to add to obj-y Sep 18 16:24:41 it'wont take a full path probably Sep 18 16:24:44 just do a symlink Sep 18 16:26:19 dyoung-zzzz: sorry, just now understood tour comment, sorry. Yes, all compiled fine, it was that the make file diid not created one dir, so the assembler could not output the file. Sep 18 16:26:48 I have the .ko, but cant manage to have all the symbols Sep 18 16:27:16 that's where libIxp425.a should come into play Sep 18 16:29:13 dwery: no, it does not link against it with obj-y := libIxp425.a Sep 18 16:29:21 yvasilev: just noticed Sep 18 16:29:31 yvasilev: it uses it in built-in.o only Sep 18 16:29:37 yvasilev: where it is not useful Sep 18 16:41:37 yvasilev: found Sep 18 16:41:47 dwery: tellme! Sep 18 16:41:50 add ixp400_eth-y := libIxp425.a Sep 18 16:41:56 and it should link it correclt Sep 18 16:42:01 in Kbuild Sep 18 16:42:30 it builds a .ko for me Sep 18 16:42:50 but i think something is still not working Sep 18 16:43:27 ls Sep 18 16:43:40 did you got the .ko? Sep 18 16:43:51 ixp400_eth: Unknown symbol ixEthAccPortMulticastAddressJoinAll Sep 18 16:43:57 form insmoding it Sep 18 16:44:03 with many others Sep 18 16:44:17 how many bytes is that ko? Sep 18 16:44:26 so it does not get linked with libIxp425.a Sep 18 16:44:38 1257 Sep 18 18:41 ixp400_eth.ko Sep 18 16:45:04 but this is much better, than before Sep 18 16:45:15 maybe we should insert ix400.ko? Sep 18 16:46:36 let me rebuild ix400.ko to see if it has the symbols Sep 18 16:50:26 could someone with be nslu tell us what are the name of the modules they have inserted? Sep 18 16:52:48 lsmod Sep 18 16:52:50 it's easy :) Sep 18 16:55:18 now i got some undefined symbols making ixp400.ko Sep 18 16:55:40 dwery: FYI: http://kernel-handbook.alioth.debian.org/ Sep 18 16:56:28 rwhitby-web: thanks Sep 18 16:57:45 dwery: unfortunately, it seems to just say "send it upstream" Sep 18 16:57:52 insmod: error inserting 'ixp400.ko': -1 Unknown symbol in module Sep 18 16:58:51 we are homing in... my undefined symbols are: Sep 18 16:58:53 *** Warning: "BIT" [/tmp/ixp400_xscale_sw/lib/linuxle/ixp400.ko] undefined! Sep 18 16:58:53 *** Warning: "save_flags" [/tmp/ixp400_xscale_sw/lib/linuxle/ixp400.ko] undefined! Sep 18 16:58:53 *** Warning: "restore_flags" [/tmp/ixp400_xscale_sw/lib/linuxle/ixp400.ko] undefined! Sep 18 16:58:53 *** Warning: "cli" [/tmp/ixp400_xscale_sw/lib/linuxle/ixp400.ko] undefined! Sep 18 16:58:53 *** Warning: "clean_dcache_range" [/tmp/ixp400_xscale_sw/lib/linuxle/ixp400.ko] undefined! Sep 18 16:58:54 *** Warning: "invalidate_dcache_range" [/tmp/ixp400_xscale_sw/lib/linuxle/ixp400.ko] undefined! Sep 18 16:59:28 add ixp400-y := libIxp425.a to my Makefile ;-) Sep 18 17:00:31 same results.... Sep 18 17:01:22 I also added ethDB to BI_ENDIAN_COMPONENTS Sep 18 17:02:53 may I have the complete list in your BI_ENDIAN_COMPONENTS Sep 18 17:02:54 ? Sep 18 17:03:44 oslinux ethAcc and ethDB Sep 18 17:04:23 ok, now with ixp400 and ixp400_eth I get no undefined symbols during compilation, nor with dmesg Sep 18 17:04:28 i'm with oslinux ethAcc qmgr ethMii npeMh ethDB npeDl featureCtrl Sep 18 17:04:31 but still no interface Sep 18 17:05:40 I'll try to include more Sep 18 17:08:22 dwery: I see you solved your toolchain problem, what was it after all? Sep 18 17:09:35 you were right, endiannes Sep 18 17:09:39 from nslu2-general Sep 18 17:09:39 : Sep 18 17:09:41 (02:06:53) ByronT: # lsmod Sep 18 17:09:41 (02:06:53) ByronT: Module Size Used by Tainted: P Sep 18 17:09:41 (02:06:53) ByronT: rbuttons 1304 0 (unused) Sep 18 17:09:41 (02:06:53) ByronT: pbuttons 800 0 (unused) Sep 18 17:09:41 (02:06:53) ByronT: x1226-rtc 3664 0 Sep 18 17:09:43 (02:06:53) ByronT: ixp425_eth 17116 1 Sep 18 17:09:45 (02:06:55) ByronT: ixp400 608216 0 [ixp425_eth] Sep 18 17:09:47 (02:06:57) ByronT: # Sep 18 17:10:36 ok, so we are building the right ones Sep 18 17:12:50 oh well Sep 18 17:12:58 invalidate_dcache_range Sep 18 17:13:01 is no more in the kernel Sep 18 17:13:12 there's a dfine in the _eth.c Sep 18 17:13:16 where is it used? Sep 18 17:13:24 all arounf in the library Sep 18 17:14:26 do you get errors with respect to it? Sep 18 17:15:13 yes.. but i guess the solution is in one of the patches Sep 18 17:15:43 since I have not applied all of them :( Sep 18 17:17:35 I am trying to build with the same list 1.4 was built: BI_ENDIAN_COMPONENTS := qmgr npeMh npeDl ethAcc ethDB ethMii ossl osServices featureCtrl Sep 18 17:21:20 -rw-r--r-- 1 root root 302902 2005-09-19 02:17 ixp400.ko Sep 18 17:21:26 all symbols are ok. Sep 18 17:21:35 :-) Sep 18 17:22:04 half the size of be version Sep 18 17:22:26 found the needed patch? Sep 18 17:22:34 lesser than half.. the size of lsmod is not the size of the file Sep 18 17:22:35 yes Sep 18 17:22:42 the problem is the eth.ko Sep 18 17:22:45 is too small Sep 18 17:23:31 lsmod size > file size or the other way around? Sep 18 17:23:49 (for the same module) Sep 18 17:24:35 > Sep 18 17:24:37 i was wrong. Sep 18 17:24:43 This is my lsmod from the "other debian" version, if that helps Sep 18 17:24:50 ixp425_eth 15236 0 Sep 18 17:24:50 ixp400 615208 1 ixp425_eth Sep 18 17:24:51 yes please Sep 18 17:25:07 widrone: thanks! Sep 18 17:25:13 widrone: which kernel? Sep 18 17:25:55 Linux blug 2.6.12.2 #1 Mon Aug 8 08:52:36 PDT 2005 armv5teb GNU/Linux Sep 18 17:26:45 thanks Sep 18 17:27:09 Its from the DebianSlug/OpenDebianSlug instructions install Sep 18 17:27:41 yvasilev: please do modinfo ixp400_eth.ko Sep 18 17:28:12 depends: Sep 18 17:28:12 vermagic: 2.6.13.1 ARMv5 gcc-3.4 Sep 18 17:28:43 same for ixp400.ko Sep 18 17:28:49 there are no parameters, and there is plenty of them in the source. the module is not building correclt. Sep 18 17:32:00 can't fine an x86 module that will show something about parameters Sep 18 17:32:26 ok, found one Sep 18 17:32:42 now I know how it should look Sep 18 17:34:16 widrone: could you paste the modinfo output of ixp400 and ixp425_eth too please? Sep 18 17:36:59 filename: ixp400.ko Sep 18 17:36:59 depends: Sep 18 17:36:59 vermagic: 2.6.12.2 ARMv5 gcc-3.4 Sep 18 17:38:05 --- Sep 18 17:38:12 filename: /lib/modules/2.6.12.2/kernel/drivers/net/ixp425_eth.ko Sep 18 17:38:12 author: Intel Corporation Sep 18 17:38:12 license: GPL Sep 18 17:38:12 description: IXP425 NPE Ethernet driver Sep 18 17:38:12 depends: Sep 18 17:38:15 vermagic: 2.6.12.2 ARMv5 gcc-3.4 Sep 18 17:38:18 parm: npe_learning:If non-zero, NPE MAC Address Learning & Filtering feature will be enabled (int) Sep 18 17:38:21 parm: log_level:Set log level: 0 - None, 1 - Verbose, 2 - Debug (int) Sep 18 17:38:24 parm: no_csr_init:If non-zero, do not initialise Intel IXP400 Software Release core components (int) Sep 18 17:38:27 parm: no_phy_scan:If non-zero, use hard-coded phy addresses (int) Sep 18 17:38:29 parm: phy_reset:If non-zero, reset the phys (int) Sep 18 17:38:32 parm: netdev_max_backlog:Should be set to the value of /proc/sys/net/core/netdev_max_backlog (perf affecting) (int) Sep 18 17:38:35 Sorry about that, it was a _bit_ bigger ;) Sep 18 17:38:51 wrong Kbuil Sep 18 17:38:52 d Sep 18 17:38:54 my fault Sep 18 17:38:56 widrone: thanks Sep 18 17:39:10 dwery: maybe it the .c file Sep 18 17:43:46 I have some files in /initrd/usr/lib/ipkg/info/ namely ixp425-eth.control and ixp4xx-csr.control that contain urls to diffs and patches Sep 18 17:44:08 That wont contain any further pointers in your work? Guessing not, but asking anyway Sep 18 17:44:56 * yvasilev has now idea about what control files are :-| Sep 18 17:56:15 WHOOOOOOOOOOOOOOOOAaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaAAAAAAAAAaaaaaaaaaaAAAaaaaaaaaaAAAAaaAAAaaAA Sep 18 17:56:18 Yo Yo Yo Sep 18 17:56:26 worked??? Sep 18 17:56:37 ???????????? Sep 18 17:56:41 not very well but worked: Sep 18 17:56:43 206cc:/tmp# modprobe --force ixp400_eth Sep 18 17:56:43 ixp400_eth: no version magic, tainting kernel. Sep 18 17:56:43 ixp400_eth: Initializing IXP400 NPE Ethernet driver software v. 1.4A Sep 18 17:56:43 ixp400_eth: CPU clock speed (approx) = 263 MHz Sep 18 17:56:43 ixp400.o: ERROR, no Microcode found in memory Sep 18 17:56:44 ixp400_eth: Error starting NPE for Ethernet port 0! Sep 18 17:56:46 FATAL: Error inserting ixp400_eth (/lib/modules/2.6.13.1/ixp400_eth.ko): Operation not permitted Sep 18 17:57:04 close Sep 18 17:57:08 in Kbuild Sep 18 17:57:13 you only need to have Sep 18 17:57:14 obj-m := ixp400_eth.o Sep 18 17:57:33 plus, there's a problem with ntohs and its inverse, which were not defined Sep 18 17:57:42 paste your complete kbuild to the web please Sep 18 17:58:34 ok Sep 18 17:58:52 this line "ixp400_eth: Error starting NPE for Ethernet port 0!" sounds so close!!! :-) Sep 18 18:00:13 http://rafb.net/paste/results/kNnxbL54.html Sep 18 18:00:19 also check the lastest include pat Sep 18 18:00:21 path Sep 18 18:00:27 ok Sep 18 18:00:39 in eth.c Sep 18 18:00:40 add Sep 18 18:00:51 #include Sep 18 18:00:59 # define ntohl(x) __bswap_32 (x) Sep 18 18:00:59 # define ntohs(x) __bswap_16 (x) Sep 18 18:00:59 # define htonl(x) __bswap_32 (x) Sep 18 18:00:59 # define htons(x) __bswap_16 (x) Sep 18 18:01:31 also comment out the call to cancel_db_maintenance(); Sep 18 18:02:06 eth.c = ixp400_eth.c? Sep 18 18:02:30 yes Sep 18 18:02:38 k Sep 18 18:02:46 the microcode is related to ixp400.o Sep 18 18:03:34 I'm olmost finished rebuilding libIxp425.a will all the components (it is so slow because of the stupid loop I get) Sep 18 18:06:31 why do we need to comment cancel_db_maintenance? Sep 18 18:09:42 byteswap.h: No such file or directory Sep 18 18:10:11 yvasilev: because the required component is not included Sep 18 18:10:25 yvasilev: for byteswap you have to search where you have it Sep 18 18:10:31 in your toolchain Sep 18 18:10:37 ok Sep 18 18:12:27 ixp400_eth: no version magic, tainting kernel. Sep 18 18:12:27 ixp400_eth: Initializing IXP400 NPE Ethernet driver software v. 1.4A Sep 18 18:12:27 ixp400_eth: CPU clock speed (approx) = 263 MHz Sep 18 18:12:27 [error] ixEthMiiPhyScan : unexpected Mii PHY ID 00008201 Sep 18 18:12:27 ixp400_eth: ixp0 is using NPEB and the PHY at address 0 Sep 18 18:12:28 ixp400_eth: ixp1 is using NPEC and the PHY at address 1 Sep 18 18:12:30 ixp400_eth: Use default MAC address 00:02:b3:01:01:01 for port 0 Sep 18 18:12:32 ixp400_eth: Use default MAC address 00:02:b3:02:02:02 for port 1 Sep 18 18:12:34 Unable to handle kernel paging request at virtual address a305000e Sep 18 18:12:36 pgd = c0f68000 Sep 18 18:12:38 [a305000e] *pgd=00000000 Sep 18 18:13:04 for the microcode, you need to define IX_INCLUDE_MICROCODE=1 while compiling ix400.o Sep 18 18:13:51 yvasilev: do you have a defconfig of an old driver? Sep 18 18:14:45 lock at the patch in GPL_ixp400LinuxIntegrationPatch-1_3.zip Sep 18 18:15:13 nothing more nslu2 related? Sep 18 18:15:52 all cancel_db_maintenance does is that it calls cancel_delayed_work which is provided by the kernel Sep 18 18:16:39 there should be a openslug's .config in monotone Sep 18 18:17:13 where should I define IX_INCLUDE_MICROCODE=1 ? Sep 18 18:17:16 i had an unresolved symbol on that one... Sep 18 18:17:31 you define it in you command line or main Makefile of _sw Sep 18 18:17:52 CFLAGS? Sep 18 18:18:49 yes Sep 18 18:18:55 not Sep 18 18:18:57 sorry Sep 18 18:19:00 just ine the Makefile Sep 18 18:19:34 ok, rebuilding Sep 18 18:22:35 is it working for you now? Sep 18 18:24:29 it crashes the kernel Sep 18 18:24:44 that's not so good Sep 18 18:25:21 http://lists.debian.org/debian-devel/2005/09/msg00860.html Sep 18 18:27:32 we should know which NPE should be used, since the ixp has more than one Sep 18 18:29:02 yvasilev: did you built ok? Sep 18 18:29:10 building Sep 18 18:30:34 i'll do one more try and then go to be... Sep 18 18:30:35 bed Sep 18 18:30:48 we reached a point i didn't thought it was possible anyhow :) Sep 18 18:30:55 :-) Sep 18 18:32:19 dwery: I'm also unsure about the byteswap.h, why do you think we need it? Sep 18 18:32:52 shouldn't it be part of the microcode to swap packets and so? Sep 18 18:34:27 *** Warning: "IxNpeMicrocode_array" [/mnt/misc/kernel/work/b/ixp400_xscale_sw/lib/linuxle/ixp400.ko] undefined! Sep 18 18:34:51 I'm building with IX_INCLUDE_MICROCODE=1 Sep 18 18:35:42 yvasilev: search in _eth.c, you will fine a couple of invocations Sep 18 18:35:47 to ntohs Sep 18 18:35:59 widrone: still there? Sep 18 18:36:07 I'd need an ifconfig Sep 18 18:36:38 Yes im here Sep 18 18:37:03 please di Sep 18 18:37:04 do Sep 18 18:37:07 ifconfig -a Sep 18 18:37:28 and paste to http://rafb.net/paste/ ;-) Sep 18 18:38:40 http://rafb.net/paste/results/utjEX645.html Sep 18 18:38:54 also dmesg please Sep 18 18:40:13 Ill try to get the relevant out of it, its full of stuff from me inserting all sorts of USB stuff =) Sep 18 18:40:38 :) i just neet the network related part Sep 18 18:40:55 widrone: you probablu can remobe the modules, note where dmesg ended, and then reincert them and copy that part Sep 18 18:42:04 dwery: how have you got around IxNpeMicrocode_array not being defined? Sep 18 18:42:40 no Sep 18 18:42:52 http://rafb.net/paste/results/THoTHJ20.html Sep 18 18:43:05 try defining that IX_ in the cmdline Sep 18 18:43:14 yvasilev: which timezone are you in? Sep 18 18:43:28 -6 Sep 18 18:43:45 you? Sep 18 18:43:52 +2 Sep 18 18:44:01 widrone: is that everything? Sep 18 18:44:07 IX_INCLUDE_MICROCODE=1 is defined on command line and Makefile Sep 18 18:44:26 yvasilev: strange Sep 18 18:45:07 Nope sorry, there is more there. Ill reboot it and get you a clean one. Sep 18 18:45:24 thanks Sep 18 18:45:28 dwery: just remembered, weren't you in italy? Sep 18 18:45:31 Its lost within rfcomm and other noise. Sep 18 18:45:35 yvasilev: yeah\ Sep 18 18:45:49 should it not be +1? Sep 18 18:45:55 DST Sep 18 18:46:04 CEST, actually. Sep 18 18:46:06 ah Sep 18 18:46:48 this summer time changes it a little bit (about one hour ;-) ) Sep 18 18:46:56 :) Sep 18 18:47:31 the kernel keeps crashing Sep 18 18:47:38 .badly... Sep 18 18:48:06 lucky of you, I can't even compile :-| Sep 18 18:48:28 we will now need an intregation/kbuild specialist Sep 18 19:07:57 I get http://rafb.net/paste/results/8OjEJF90.html from readelf -r ixp400_npeDl.o | less Sep 18 19:08:26 dwery: could you please tell me what section is it in yours? Sep 18 19:08:48 ok Sep 18 19:10:46 Relocation section '.rel.data' at offset 0x4f8c contains 1 entries: Sep 18 19:10:46 Offset Info Type Sym.Value Sym. Name Sep 18 19:10:46 00000024 0000b102 R_ARM_ABS32 00000000 IxNpeMicrocode_array Sep 18 19:10:57 sorry for not pasting it but i'm quite a zombie right now Sep 18 19:11:12 same Sep 18 19:12:03 trying to add to EXPORT_SYMBOL Sep 18 19:14:58 i'm going to benow Sep 18 19:15:00 bed Sep 18 19:15:10 i think there's also an issue with the PMU timer Sep 18 19:15:18 there's a note in _eth.c Sep 18 19:15:28 so the kernel must be recompiled without it Sep 18 19:15:41 but the build system keeps re-enabling it. i'll check it tomorrow. Sep 18 19:16:29 dwery: ok, I'll keep trying for a little bit longer Sep 18 19:16:51 yvasilev: at least 8 hours... -6 + 2 :D Sep 18 19:16:58 :-D Sep 18 19:17:05 good night and congrats mate Sep 18 19:17:17 night Sep 18 19:20:09 this is the last oops, for reference: http://rafb.net/paste/results/f5j9Jk83.html Sep 18 19:25:09 dwery-zzzz: thanks, this is will be very useful! Sep 18 19:27:28 dwery-zzzz: did the ixp400 printed anything before ixp400_eth messages? Sep 18 19:34:26 Sorry about not getting back with the dmsg, the reboot landed me in a strange place (probably since I stuck a usb hub in there while testing) Sep 18 19:36:42 widrone: no problem, you helped us a lot so it's more like thank you :-) Sep 18 19:37:49 widrone: and if you have that dmesg, I'm quite interested in it ;-) Sep 18 19:37:59 http://rafb.net/paste/results/omV9U010.html Sep 18 19:38:21 There you go, got all the usbdisks reordered so it booted =) Sep 18 19:38:36 thanks!!!! Sep 18 19:39:18 This is my favorite line: ixp425_eth: CPU clock speed (approx) = 0 MHz Sep 18 19:39:57 your slug is fast! ;-) Sep 19 00:05:09 rwhitby-away: I will put it online once I have stages ready, which I'm working on Sep 19 00:33:10 rwhitby-away: how I did it.. install openslug, mess around to get portage working under openslug, no luck, install the fedora core 3 into a chroot, get portage working under that.. build packages, install into a seed... which is what I am doing now :) Sep 19 00:33:55 glibc takes days to get in.. if I didnt know better I'd say it ignored my USE="userlocales" Sep 19 00:35:22 it's on day 4 with testing iconv, hopefully done tomorrow :) Sep 19 01:41:12 im trying to port a webcam driver to the nslu2 - when trying to compile the kernel module, i get "ld: unrecognised emulation mode: elf_i386" my toolchain is the first thing in my path, and it's definitely invoking the arm linker. can anyone help? Sep 19 02:01:13 03hrw 07org.openembedded.dev * rfa314921... 10/packages/meta/meta-opie.bb: Sep 19 02:01:13 meta-opie: added opie-freetype into task-opie-base Sep 19 02:01:13 - that was the problem of "no fonts in fresh opie image" Sep 19 02:01:14 kolla: rofl? Day 4? Sep 19 02:07:32 NAiLzZz: just wait till its done and a new security issue is found in glibc so he has to recompile ;) Sep 19 02:12:29 haha Sep 19 02:33:01 yawn Sep 19 02:33:08 g'morning Sep 19 02:33:19 morning dwery Sep 19 02:33:40 dwery: http://lists.debian.org/debian-embedded/2005/09/msg00012.html Sep 19 02:34:53 rwhitby: goot start. Sep 19 02:37:03 yvasilev: still there? Sep 19 02:40:09 probably not :) Sep 19 02:42:19 for those who weren't there, we had a shot at IAL 2.0 Sep 19 02:42:49 dwery: status is that it compiles but crashes, right? Sep 19 02:44:58 dwery: that's basically as far as I got as well Sep 19 02:45:08 Jacmet: compiles is not the right word.. let's say we managed to extract some binaries from that source. Sep 19 02:45:33 Jacmet: same crash? Sep 19 02:55:59 dwery: I think so - some invalid memory access right Sep 19 02:56:15 03hrw 07org.openembedded.dev * r427cbc86... 10/packages/qpf-fonts/ (10 files): qpf-dejavu: updated to 1.14 Sep 19 02:56:16 dwery: sorry, I don't really remember the details anymore Sep 19 02:57:14 Jacmet: where should we go now? Sep 19 02:57:31 Jacmet: try the same on BE? abort? cancel? **** ENDING LOGGING AT Mon Sep 19 02:59:56 2005