**** BEGIN LOGGING AT Thu Jan 03 02:59:56 2008 Jan 03 03:20:56 paste the whole oops, please Jan 03 03:34:43 i found in forum people with same problem Jan 03 03:34:49 im building Jan 03 03:35:03 the php4 version in svn surver dont build Jan 03 03:35:10 have a problem in patch 002 Jan 03 12:03:42 morning Jan 03 12:04:19 anybody a idea if somebody is working on thecus nas device support for openwrt? Jan 03 12:05:15 its a N4100+ what interrsts me. has a intel 80219 io cpu (arm) Jan 03 12:17:59 roh: i don't think anybody's working on that yet Jan 03 12:21:24 ah.. morning nbd Jan 03 12:21:35 morning, roh ;) Jan 03 12:21:41 harald said he rund a debian on it, si support seems not to be the problem Jan 03 12:22:00 it has 600mhz it seems and 4x sata and 2x gigabit with real nics Jan 03 12:22:11 sounds like a nice device Jan 03 12:22:16 around 400bucks Jan 03 12:22:21 :/ Jan 03 12:22:41 http://www.thecus.com/products_over.php?cid=11&pid=28 Jan 03 12:25:31 only 30% power savings compared to a full pc? Jan 03 14:27:26 The 2100 is cheaper than that Jan 03 14:59:09 hey guys Jan 03 14:59:18 hey Nix Jan 03 14:59:33 oh.. just the guy I wanted to see Jan 03 14:59:37 :) Jan 03 14:59:58 is it possible at present to give a svn source instead of ftp/http in an owrt Makefile? Jan 03 15:00:15 yes Jan 03 15:00:18 madwifi does it Jan 03 15:00:20 in trunk Jan 03 15:00:38 it actually checks out the repo, generates a tarball from that Jan 03 15:00:49 and it can then use the tarball for mirrors Jan 03 15:01:14 ooh.. cool Jan 03 15:01:19 thanks Jan 03 15:01:32 it actually tries to grab a tarball before doing the svn checkout itself Jan 03 15:01:36 because that's usually faster Jan 03 15:01:37 hmm Jan 03 15:02:11 right, well I want to work on some packages locally which will never be made public so I dont want to go abusing the openwrt mirrors every time I build :-D Jan 03 15:02:43 the servers don't care about a few more 404 requests Jan 03 15:02:48 ok :-) Jan 03 15:04:38 thanks Jan 03 15:04:55 you're welcome Jan 03 15:05:34 * Nix wonders where blogic is... he broke the ifxmips target around xmas day and hasn't been heard from since.. Jan 03 15:06:48 what specifically is broken? Jan 03 15:08:51 simply doesnt build.. he is aware of it.. Jan 03 15:09:03 something about being too hungover to fix it :-D Jan 03 15:14:08 <_trine> blogic seems to be generating a track record for broken code its about time he fixed his mmc code Jan 03 15:14:36 <_trine> perhaps he should be called brokenlogic Jan 03 15:16:56 :-D Jan 03 15:17:21 well, his older code is fine.. its just his latest commit which doesnt even compile :-D Jan 03 15:17:50 https://dev.openwrt.org/changeset?new=trunk%2Ftarget%2Flinux%2Fifxmips%2Ffiles%2Farch%2Fmips%2Fifxmips%2Fprom.c%409936&old=trunk%2Ftarget%2Flinux%2Fifxmips%2Ffiles%2Farch%2Fmips%2Fifxmips%2Fprom.c%409929 Jan 03 16:05:16 Nix: what error message do you get? Jan 03 17:07:05 nbd * r10091 /trunk/package/busybox/patches/523-conffiles_fix.patch: add updated ipkg conffiles patch by cdtdaddy Jan 03 17:07:56 nbd * r10092 /trunk/package/busybox/patches/522-memleak.patch: restore old memleak fix Jan 03 17:08:20 I nearly read "restor old memleak" Jan 03 17:08:22 +e Jan 03 17:08:32 hehe Jan 03 17:27:55 florian * r10093 /packages/utils/sane-backends/Makefile: Upgrade sane-backeds to 1.0.18, also install headers and libraries (#2960) Jan 03 17:32:21 florian * r10094 /packages/libs/cgilib/ (Makefile patches/300-cpp_bugfix.patch): Add C++ bufgix patch from #2965, bump release number Jan 03 19:33:53 florian * r10095 /trunk/package/broadcom-wl/Makefile: Force broadcom-compat-ldso to be 2.4 only (#2963) Jan 03 19:47:11 pavlov * r10096 /packages/net/tftp-hpa/Makefile: bump up tftp-hpa to 0.48 Jan 03 21:28:59 kaloz * r10097 /trunk/target/linux/brcm-2.4/image/Makefile: bump version numbers of Linksys images Jan 03 21:58:56 florian * r10098 /trunk/target/linux/ar7/config-2.6.23: Resync ar7 kernel config (#2966) Jan 03 22:14:59 Florian, how hard do you think this would be? https://dev.openwrt.org/ticket/2864 I've attempted this myself, but I've not had much joy so far. Jan 03 22:17:53 florian * r10099 /trunk/target/imagebuilder/files/Makefile: Fix imagebuilder for 2.6 kernels (#2900) Jan 03 22:19:42 <[florian]> nabcore: no idea yet, but we certainly need to sort this out Jan 03 22:20:30 it's not the most idea solution, but at least it's a stop gap since Master mode works very well with the older driver Jan 03 22:20:37 *ideal Jan 03 22:21:07 <[florian]> nabcore: well, looks like master mode does not really work with mac80211 Jan 03 22:21:25 that's what I'm finding Jan 03 22:21:52 but I think master mode working only with WEP is better than no master mode at all Jan 03 22:22:25 <[florian]> absolutely Jan 03 22:23:25 AR7 is so so nearly there... there's been some nice work from NRForbes here; https://dev.openwrt.org/ticket/2817 now the leds work Jan 03 22:23:59 * Bartman007 waits for danube devices to hit the market Jan 03 22:24:30 <[florian]> yes, though I wonder what we are missing to support wpa Jan 03 22:25:13 that, I think, is in the realm of the mac80211 stack Jan 03 22:26:17 <[florian]> when did we get non working acx driver ? Jan 03 22:26:50 When the vlynq-pci support was dropped in r9143 Jan 03 22:27:01 the newer acx-mac80211 driver was updated accordingly to reflect this in r9144. As far as I see, the older acx driver was not updated in the same way. Jan 03 22:27:52 In addition to to master mode not working on the acx-mac80211, there are some other bugs as well: #2559,#2555,#2547 Jan 03 22:28:10 Look at ticket https://dev.openwrt.org/ticket/2864 for the links Jan 03 22:30:13 Besides this, I think the only other outstanding issue is a possible issue with the DSL driver: https://dev.openwrt.org/ticket/2377 Jan 03 22:30:59 nbd offered a suggestion that I tried the other night, but it did not work :( Jan 03 22:31:49 IIRC, this issue started to appear in the transition from 2.6.19 to 2.6.23 Jan 03 22:34:30 <[florian]> humm ok, this is pretty tricky to fix Jan 03 22:34:44 <[florian]> can you reproduce the bug reliably ? Jan 03 22:34:55 https://dev.openwrt.org/ticket/2377 ? Jan 03 22:35:03 <[florian]> yes that particular one Jan 03 22:35:24 yes... I can always reproduce it Jan 03 22:35:52 <[florian]> yes, but how long after you did setup dsl carrier ? Jan 03 22:36:38 Can you make your question clearer? Jan 03 22:37:15 I do have this appending to my /etc/init.d/rcS : "sleep 35 ; /sbin/ifup wan &" Jan 03 22:37:28 <[florian]> well, how long does it take for you to reproduce the bug, days, or minutes after dsl is running ? Jan 03 22:37:46 minutes Jan 03 22:37:51 There are times on this: http://pastebin.com/d40058005 Jan 03 22:38:50 it's basically after running pppd having gotten a DSL sync Jan 03 22:40:07 as oliver says, this could actually be happening all the time, but since it's WARN_ON_ONC, I'm only seeing it once ( https://dev.openwrt.org/ticket/2377#comment:12 ) Jan 03 22:41:01 <[florian]> yes, absolutely Jan 03 22:41:26 <[florian]> will do some more ar7 hacking this weekend on it Jan 03 22:41:34 thank you. Jan 03 22:41:51 <[florian]> ideally fixing acx-mac80211 would be better Jan 03 22:42:02 Generally I'm at a loss. My knowledge + coding is not of a sufficient level to resolve this one Jan 03 22:42:26 <[florian]> but I think for now, we could probbaly stick with putting back vlynq-pci compatible code in acx Jan 03 22:43:35 You mean modify the acx code to work with the newer vlynq-pci? Jan 03 22:43:52 <[florian]> yes Jan 03 22:46:09 Cool; I think ejka dropped the old vlynq-pci for a reason in ( https://dev.openwrt.org/changeset/9143 ), I think it may have been to fix some vlynq timing related issues, but I cannot be sure. Jan 03 22:48:20 The newer method to access the vlynq as used in acx-mac80211 seems to be more efficient in the number of lines of code and I think ejka had always viewed the older vlynq-pci bridging as a bit of a hack. Jan 03 22:49:51 so.. I think it's a case of appling this patch here; https://dev.openwrt.org/browser/trunk/package/acx-mac80211/patches/002-acx_vlynq.diff?rev=9144 in an analogous to the older acx driver. Jan 03 22:50:06 <[florian]> that's also my idea Jan 03 22:51:21 I've attempted this but failed :( Jan 03 22:51:35 <[florian]> it won't apply that easily Jan 03 22:53:44 ok.. I've got to go for a bit. Thanks again florian for the chat and advice. Jan 03 22:55:18 nabcore: if you have suggestion about acx i have write access on its SVN on sourceforge.net Jan 03 22:59:33 <[florian]> nabcore: you are welcome Jan 03 23:09:01 florian * r10100 /packages/lang/jamvm/ (Makefile patches/): Upgrade jamvm to 1.5.0, also fixes its compilation issues (#2935) Jan 03 23:10:23 agb * r10101 /packages/libs/speex/ (Makefile patches/100-compile_fix.patch): Jan 03 23:10:23 packages: Update libspeex to 1.2beta3 Jan 03 23:10:23 Signed-off-by: Alfred E. Heggestad Jan 03 23:20:19 hi matteo; thanks for that as well Jan 03 23:20:25 nabcore? Jan 03 23:20:37 Hi people ! Jan 03 23:20:43 hey Jan 03 23:20:57 And hi Kaloz ! Jan 03 23:49:10 can't resolve symbol '__uClibc_start_main' Jan 03 23:49:22 anyone have this error with nas in 2.4 kernel trunk? Jan 04 00:22:31 matteo; the original acx driver makes use of a function called ndev2adev(). This operates on net_device structure and returns a pointer from netdev_priv(net_device). I'm trying to write one that operates on a vlynq_device structure. Jan 04 00:24:08 good luck Jan 04 00:47:23 i'm happy to see ar7 discussion :) thank you for working on it! i'm going to give it a go again, but last time i tried it kept locking up at 12 hours into use. Jan 04 00:48:10 (approx.) Jan 04 00:51:12 orkid; have you got serial access? Jan 04 00:54:57 i do finally, the modem is in the basement , and i need to setup a computer there to hook up the console (and space is limited). hopefully i'm going to do it tomorrow. Jan 04 00:55:21 will the errors most likely show to serial by default, or should i run logread -f ?? Jan 04 00:56:07 they should do Jan 04 00:56:24 also, should i worry about loading devel snapshots all the time on flash and using up it's erase/write cycles ? :S Jan 04 00:56:37 (its) Jan 04 00:57:22 no; someone did the maths on this channel once, and it's not going to wear out your router Jan 04 00:58:54 the more testing something gets, the fewer bugs it will have, hence get involved Jan 04 01:01:15 i would really like to, but it's hard for me to figure out the structure of things in the dev directory, what patches get applied to what, etc. i'll get the serial set up, and at least provide some bug info i hope. it would be cool if someone could explain how things are connected. i've read BuildRoot, but maybe it's time to do it again, more thoroughly. and i'm going through KN King's C Programming A Modern Approach again to remind myself of C. Jan 04 01:01:24 long road ahead :) Jan 04 01:01:31 yup Jan 04 01:02:20 nas: can't resolve symbol 'bcopy' Jan 04 01:02:36 I'm no pro either... I'm still learning. But perhaps in the beginning, the most use one can be is to provide testing and accurate bug reports Jan 04 01:03:23 i have still yet to figure out how to get a 'trace' (or is that just the kernel's dump into the log), i'm thinking of reading ldd3 (linux device drivers 3rd ed) from oreilly also. Jan 04 01:07:43 GazeteiRo; have you filed a bug? Jan 04 01:08:14 GazeteiRo: check your .config for CONFIG_PACKAGE_brcm-compat-ldso Jan 04 01:08:59 nbd has been working on this one, a ticket was posted earlier, might want to confirm that one Jan 04 01:21:06 Ok Jan 04 01:21:41 <[florian]> nabcore: I almost done with a compile tested patch for acx to use the new vlynq code Jan 04 01:22:36 netprince, CONFIG_PACKAGE_brcm-compat-ldso=y Jan 04 01:23:13 hmm, thats not good, sounds like its broken, :) Jan 04 01:24:21 I am compiling trunk too, wont be done till tomorrow though Jan 04 01:25:29 xii =) Jan 04 01:25:57 this may be a trunk problem no? Jan 04 01:26:07 looks like it Jan 04 01:26:39 nas is not quite compatible with the new uclib if I understand correctly Jan 04 01:26:45 florian, ace! I've just been doing some testing as well. Did you have to write a ieee2adev like function for the vlynq device? Jan 04 01:27:15 <[florian]> nabcore: nope, because we do not use mac80211, it will appear as a classical netdev interface Jan 04 01:27:22 <[florian]> this is the kind of things I am sorting out now Jan 04 01:27:25 ok Jan 04 01:28:06 netprince, have you used opensource brcm drivers Jan 04 01:28:07 ? Jan 04 01:28:36 GazeteiRo: no, been too busy with work stuff Jan 04 01:29:51 Florian, did oyu start off with the acx-mac80211 code and work backwards to the acx to enable the newer vlynq access or did you start with the acx code and try to add the newer vlynq ? Jan 04 01:30:52 <[florian]> nabcore: I used the new vlynq patch and remove everything that is ieee80211 specific Jan 04 01:31:02 <[florian]> nabcore: and put back all the netdev registration Jan 04 01:32:37 that's the route I was attempted... I came a little stuck when I encountered such assignments as adev = ieee2adev(ieee); Jan 04 01:32:46 *attempting Jan 04 01:33:33 but that function just returned a pointer to the priv in struct ieee80211_hw Jan 04 01:33:46 wow [florian] that fast? Jan 04 01:34:04 <[florian]> orkid: I can't find sleep, so let's do useful things :p Jan 04 01:34:33 :) Jan 04 01:34:51 your brain doesn't always function as well, but tedious tasks done slowly, it can be pretty good at. Jan 04 01:36:16 <[florian]> that's right, though right now I am feeling better than before in the day Jan 04 01:36:57 I think I was trying to get the correct pointer from the vlynq_device->dev and then pont acx_device_t *adev at it Jan 04 01:38:03 yeah. i've had that happen too. stay up too late, sleep too much, then feel like crap, until i stay up way too late again Jan 04 01:38:47 how do you guys develop? do you simply hack on a trunk directory with vim, and then keep doing 'make' ? Jan 04 01:39:03 <[florian]> using quilt is much easier Jan 04 01:39:20 <[florian]> but this was my previous way of working Jan 04 01:39:24 orkid; that's what I'm doing at the moment Jan 04 01:39:50 * Bartman007 uses git Jan 04 01:40:21 dev in branches, the squash the branches into functional patches and dcommit them to svn. Jan 04 01:49:11 orkid; Have you got the DSL working on your router? Jan 04 02:09:22 <[florian]> nabcore: do you have any build tree you could recompile acx, or should I provide you with an ipkg ? Jan 04 02:10:07 just building r10101, got a build error in asterisk Jan 04 02:10:12 florian * r10102 /trunk/package/acx/patches/003-native_vlynq.patch: Add native vlynq support to the old acx driver (#2864) Jan 04 02:10:40 "configure: *** The Speex installation on this system appears to be broken." Jan 04 02:10:49 <[florian]> russell_: will check that Jan 04 02:11:13 "Either correct the installation, or run configure without explicitly specifying --with-speex" Jan 04 02:11:26 [florian]: cool Jan 04 02:12:15 <[florian]> too many broken packages :/ Jan 04 02:12:26 florian; I have r9763 I can build against Jan 04 02:12:52 <[florian]> nabcore: ok, can I paste you the link to an ipkg ? Jan 04 02:13:34 it was asterisk-1.4.x btw Jan 04 02:14:20 a build-all for wgt634u Jan 04 02:14:31 florian, I just grabbed the patch from svn Jan 04 02:14:40 (003-native_vlynq.patch) Jan 04 02:14:58 building now Jan 04 02:15:20 <[florian]> nabcore: ok, here is the ipkg : http://f.fainelli.free.fr/kmod-acx_2.6.23.1+20071003-ar7-2_mipsel.ipk Jan 04 02:15:28 <[florian]> I just bumped the release number Jan 04 02:15:35 ok; I'll try that ipk now Jan 04 02:16:55 ok; one sec just need a reboot, I'd been messing with my own acx builds and the router may be in a bad state Jan 04 02:19:46 <[florian]> nabcore: what does it looks like :p ? Jan 04 02:20:22 looking good :) Jan 04 02:20:23 http://pastebin.ca/840885 Jan 04 02:20:52 <[florian]> at least, it did not crash Jan 04 02:21:07 just doing some testing with the actual client now Jan 04 02:21:11 <[florian]> sure Jan 04 02:22:10 hmmm /proc/driver/* seems to have some weirdness. All names begin with acx_wlan%d Jan 04 02:22:39 <[florian]> oh, yes, I think I see the problem with it Jan 04 02:24:18 <[florian]> does your wifi client work ? Jan 04 02:24:39 just checking now Jan 04 02:25:45 one sec Jan 04 02:27:22 florian; yes it does :) Jan 04 02:27:50 <[florian]> great, I would not suppose this patch will ever work tonight :p Jan 04 02:28:23 that's some great work there.. thanks again! Jan 04 02:28:35 <[florian]> you are welcome Jan 04 02:34:09 <[florian]> russell_: will fix this issue with speex and asterisk-1.4 Jan 04 02:34:58 florian; all I have to do now is to remember the correct /etc/config/wireless settings... it's been so long since I used them ;) Jan 04 02:36:17 <[florian]> nabcore: you were running without wifi all this time ? Jan 04 02:37:14 yup Jan 04 02:37:20 <[florian]> wohh Jan 04 02:37:29 and this is my only router :) Jan 04 02:38:09 it was ok... my desktop (which I use all the time) was plugged in via ethernet instead of using its wireless card Jan 04 02:38:45 florian * r10103 /packages/net/portsentry/ (Makefile files/ files/portsentry.init): Add basic init script for portsentry #2962 Jan 04 02:39:06 it was just that the g/f got a bit annoyed with the fact that she had to use a lead when she used her laptop Jan 04 02:44:31 florian * r10104 /packages/ipv6/babel/Makefile: Upgrade babel to 0.7 Jan 04 02:49:31 florian, did you also see at http://pastebin.ca/840885 that the leds were working :) Jan 04 02:51:39 <[florian]> nabcore: ah no, I did not see that, that's good news ! Jan 04 02:52:26 yes; it all works! https://dev.openwrt.org/ticket/2817 thanks to NRForbes Jan 04 02:53:43 <[florian]> will merge his patch right now Jan 04 02:54:36 cool; I'm not sure if he wanted to add some more bits to it or not: https://dev.openwrt.org/ticket/2817#comment:2 (last line) Jan 04 02:55:44 but.. I've tested his patch and it works fine on the DG834G(v2) Jan 04 02:59:11 florian * r10105 /trunk/target/linux/ar7/files/arch/mips/ar7/platform.c: Add dg834 leds support from #2817 Jan 04 02:59:53 :D **** ENDING LOGGING AT Fri Jan 04 02:59:56 2008