**** BEGIN LOGGING AT Wed Aug 27 02:59:56 2008 Aug 27 03:50:35 blogic: dead? Aug 27 03:58:21 Weedy_lappy: all the developers appear to be, no talk in the past 18 hours Aug 27 03:58:57 until now Aug 27 04:00:09 sn9: would you believe 18 hour latency? Aug 27 04:00:24 heh Aug 27 04:24:39 Bartman007: do you know which commit broke the old firewall? Aug 27 04:24:49 i need to use my firewall.user Aug 27 04:26:20 Weedy_lappy: try this patch, if it works I'll commit it. http://openwrt.pastebin.com/d25c7becc Aug 27 04:26:52 it readds the old firewall and allows you to select it (but defaults to the uci based firewall) Aug 27 04:27:39 we can yank it again after the uci firewall is bug-free and documented but until that point I would like to see the old firewall stick around Aug 27 06:12:08 Weedy_lappy: no, why ? Aug 27 06:12:42 Bartman007: name the bugs Aug 27 06:13:02 Bartman007: which part of the config file is missing as an example ? Aug 27 06:15:06 examples can't be considered full documentation Aug 27 06:18:21 unless there are enough examples to detail the allowed values for each option; I'm more concerned w/ the ppp issue though Aug 27 06:22:06 Bartman007: so, you are saying that piece of shit old firewall with no features and no doc is good ? :) Aug 27 06:23:48 I'm saying it is what people are used to, I think the new UCI firewall is better but would like to see a fall back option Aug 27 06:24:01 why ? Aug 27 06:24:11 so the users never learn how to use the new firewall ? Aug 27 06:24:19 having 2 is bloat Aug 27 06:24:26 so, we can remove the uci foo Aug 27 06:24:32 if that makes everyone happy Aug 27 06:24:41 oh wait ? where is the doc for it ? Aug 27 06:24:50 lol Aug 27 06:25:11 Bartman007: btw, name the bugs Aug 27 06:25:15 blogic: having options is bloat? this coming from one of the xorg guys? Aug 27 06:25:47 I'm not trying to blame you, or insult you in any way, anything that is coming off harsh is not intended that way Aug 27 06:26:50 Bartman007: btw, name the bugs Aug 27 06:26:52 I'm not saying xorg should be tossed out, or calling it bloat, but I don't see the harm in providing the old firewall as an option (that has to be selected at compile time) for those who want it Aug 27 06:27:04 Bartman007: btw, name the bugs Aug 27 06:29:27 moonwatcher has run into that issue where the ppp interface numbers keep incrementing, I know you said you are working/testing on it and I will welcome the fix when it comes Aug 27 06:38:23 Bartman007: that bug seems to be fixed, i have not seen it in 3 days Aug 27 06:40:10 ok, I'll ask moonwatcher to retest with a newer build Aug 27 09:03:22 Bartman007: about the old firewall: i'm also in favor of leaving it out completely. if there's something wrong with the new one, i want people to bug us about it instead of silently falling back to the crappy one Aug 27 09:15:20 nbd: yeah, after talking with blogic a bit more I agree Aug 27 09:15:30 ok Aug 27 09:31:20 moin jow Aug 27 09:31:42 moin Aug 27 09:31:52 brb smoking Aug 27 10:03:53 hi there! just a quick question... which is the default dhcp client installed in kamikaze? Aug 27 10:03:59 hey Aug 27 10:04:21 (I have to perform dhcp command manually...) Aug 27 10:08:30 try udhcpc -i interface Aug 27 10:08:54 thanks! works perfectly! Aug 27 12:04:03 blogic * r12395 /trunk/package/firewall/files/uci_firewall.sh: adds 5 new chains to the uci firewall that can be used to hook custom rules Aug 27 12:23:23 blogic: blogic oh you added the old names back? Aug 27 12:24:26 Weedy_lappy: got several requests for it Aug 27 12:24:37 there are a few less rules, but it should be sufficiant Aug 27 12:24:42 i hope anyways :) Aug 27 12:24:57 if users need it, we would be the last to not add it :) Aug 27 12:25:44 blogic: well i just needed my rules to work again Aug 27 12:26:17 unless you can do it with the new firewall http://openwrt.pastebin.com/m142d5fe1 Aug 27 12:26:19 blogic: We had a feature request for the firewall. The users want to specify "tcp+udp" as protocol... the normal solution would be to create two rules for each udp and tcp protocol. Would it be possible that the uci firewall layer handles this transparently? Aug 27 12:26:59 xMff: hmmmz Aug 27 12:27:25 xMff: why not, we can make rule generation an extra function and if proto is tcpudp we call it twice Aug 27 12:27:32 xMff: i will add it during the day Aug 27 12:27:39 Weedy_lappy: lemme have a look Aug 27 12:28:09 blogic: thanks Aug 27 12:28:12 blogic: basicaly i use snat cause i need my forwards to apply to the lan and wan Aug 27 12:28:33 Weedy_lappy: looks a rather advanced firewall :) Aug 27 12:28:48 i guess you will have to use custom chains and wait for v2.0 Aug 27 12:28:51 soz Aug 27 12:29:09 but i am trying to get all basic features working first before moving to the more advanced stuff Aug 27 12:29:25 can the chains i added handle your needs ? Aug 27 12:29:33 or would extra chains be usefull ? Aug 27 12:29:46 i mean you can hook anything you want into the chains Aug 27 12:29:56 well if you added the "old" chains my rules should work again Aug 27 12:30:14 the old firewall is considered "bad" :) Aug 27 12:30:27 let me see Aug 27 12:30:40 we could add an option that will add the old chains if users need them Aug 27 12:30:47 i will ponder about a good solution Aug 27 12:30:53 and let you know Aug 27 12:31:00 i have your mail add anyways i think Aug 27 12:31:15 like i said i just need the lan to get fwded too Aug 27 12:31:40 its nice to put domain.com in and not internalIP:80 Aug 27 12:32:14 so if you make a fwd_lan = yes option i wouldnt need anything else Aug 27 12:34:41 i was thinking "option legacy_chains 1" Aug 27 12:34:49 hi glp Aug 27 12:34:56 blogic: huh? Aug 27 12:40:30 Bartman007: i have a setup now, where the recursive creation of chain rules is triggered ... Aug 27 12:40:40 lemme fix it for real now :) Aug 27 13:58:29 blogic: if you like: http://dev.luci.freifunk-halle.net/firewall-portrange-support-patch ;-) Aug 27 13:59:01 lemme have a look Aug 27 13:59:44 CyrusFF: tested and working ? Aug 27 13:59:51 you have svn ... feel free ... Aug 27 13:59:53 :) Aug 27 14:00:45 i have created 2 rules one with ranges and one without range, both appear correctly in iptables -vL Aug 27 14:00:59 go for it then :) Aug 27 14:01:04 ok ;-) Aug 27 14:05:11 cyrus * r12396 /trunk/package/firewall/files/uci_firewall.sh: firewall: Added support for port-ranges as firstPort-lastPort to redirect sections Aug 27 14:10:15 florian * r12397 /trunk/package/mac80211/ (Makefile patches/001-disable_drivers.patch): Aug 27 14:10:15 This patch enables the new ath5k driver. Works for me with X86 Alix Aug 27 14:10:15 board an AR2413 minipci card. iw and iwconfig works. Aug 27 14:10:15 Signed-off-by: Dirk Nehring Aug 27 14:13:42 nbd * r12398 /trunk/tools/ (Makefile quilt/ quilt/Makefile): build and install a current quilt version in tools/ Aug 27 14:13:56 nbd * r12399 /trunk/scripts/env: fix ./scripts/env delete Aug 27 14:14:09 nbd * r12400 /trunk/include/ (kernel-build.mk kernel-defaults.mk quilt.mk): enable quilt by default for the kernel tree Aug 27 15:24:27 blogic * r12401 /trunk/include/download.mk: fixes cvs download, signed-off by Bruno Randolf Aug 27 15:56:03 blogic * r12402 /trunk/package/firewall/files/uci_firewall.sh: fixes uci firewall init order, Signed-off-by: Roberto Riggio Aug 27 16:48:58 blogic * r12403 /trunk/package/firewall/files/uci_firewall.sh: make sure uci firewall reverts its states when stopped Aug 27 18:07:29 blogic: thanks for those commits, kicking off new snapshot builds. Aug 27 18:09:22 the ppp0 bug is testing right now Aug 27 18:14:02 i will commit it in a sec Aug 27 18:31:36 blogic * r12404 /trunk/package/firewall/files/uci_firewall.sh: fix device duplication in firewall if the balancing of ifup and ifdown is broken Aug 27 18:33:07 juhosg * r12405 /trunk/target/linux/adm5120/ (image/router_le.mk router_le/profiles/Infineon.mk): [adm5120] generate images for the EASY 5120P-ATA board Aug 27 18:48:33 blogic * r12406 /trunk/target/linux/ifxmips/Makefile: ifxmips should not overwrite DEFAULT_PACKAGES Aug 27 18:54:54 blogic * r12407 /trunk/package/firewall/files/uci_firewall.sh: add proto tcpudp to firewall Aug 27 19:17:01 blogic * r12408 /trunk/package/firewall/files/uci_firewall.sh: make uci firewall backwards compatible to the old firewall.user Aug 27 19:17:52 Bartman007: should be all the problems fixed now Aug 27 19:18:38 blogic: thanks a lot for tcpudp, now I can go ahead and squash one further ticket :) Aug 27 19:18:45 3hehe Aug 27 19:18:47 :) Aug 27 19:35:59 nbd * r12409 /trunk/package/busybox/patches/002-darwin_compile_fix.patch: fix busybox compile on mac os x Aug 27 20:18:55 nbd * r12410 /trunk/tools/ (Makefile autoconf/ autoconf/Makefile): add autoconf (fixes opkg build issues) Aug 27 20:33:18 blogic * r12411 /trunk/package/base-files/files/etc/init.d/led: fixes broken if-then in led init script Aug 27 21:02:04 nbd * r12412 /trunk/package/uboot-ifxmips/patches/100-ifx.patch: fix uboot-ifxmips compile on mac os x Aug 27 21:16:27 nbd * r12413 /trunk/package/uci/Makefile: upgrade to uci 0.6.1 - lots of cli and lua binding fixes Aug 27 21:24:22 Anyone familiar with the ks8995 switch chip? Aug 27 21:26:49 MateIn4: so what kind of info do you need? Aug 27 21:27:15 MateIn4: what i meant was a driver that hooks into my soon-to-be-committed switch config layer Aug 27 21:27:28 the spi-ks8995 package no longer builds on latest trunk. this got built a few weeks ago. Aug 27 21:28:00 let me try Aug 27 21:28:17 I'm working on a new port to the mi424-wr and it has one of these chips. I need to initialize it to enable switching Aug 27 21:28:48 My understanding is that one of the avila boards has one of these things also. Aug 27 21:28:55 i think so Aug 27 21:28:58 don't know for sure Aug 27 21:30:01 ok, i think i found the bug Aug 27 21:30:04 testing Aug 27 21:34:07 nbd * r12414 /trunk/include/kernel.mk: fix kconfig for packages that only define =y, =m or =n rules Aug 27 21:34:11 MateIn4: fixed now Aug 27 21:34:33 nbd: thanks! I'll give it a try Aug 27 21:41:31 I've submitted a patch for supporting a new device mi424-wr here: http://lists.openwrt.org/pipermail/openwrt-devel/2008-August/003107.html. However, I've gotten no feedback. What is the process for inclusion into the trunk? Aug 27 21:42:13 sometimes patches get lost because we have much work to do Aug 27 21:42:18 please contact Kaloz directly Aug 27 21:42:21 he's maintaining that target Aug 27 21:42:37 the patch looks good to me Aug 27 21:42:40 nbd: thanks for the pointer Aug 27 21:50:58 MateIn4: actually you should probably contact me in the short term Aug 27 21:51:18 I have an mi424wr for testing. Aug 27 21:53:08 err, contact Kaloz too, his knowledge of ixp far eclipses mine. Aug 27 21:54:32 Bartman007: well, he is upstream maintainer for ixp so that would be expected :) Aug 27 21:54:55 blogic: oh yeah, I forgot Kaloz is a bigshot now :-) Aug 27 21:55:53 he is as the say a "made man" kinda like "part of the family" :) Aug 27 21:56:00 Bartman007: I've got the mi424-wr booting. I'm able to connect via the wan port. the lan ports are not accessible because the switch chip is not enabled. Aug 27 21:57:12 Another issue is that I ported redboot to the mi424wr. I've got the image and I also make an ecos .epk file with my changes. But where should I put it? Aug 27 21:58:32 MateIn4: please email me, I'll create a dir in agb.openwrt.org/ (also can be reached at downloads.openwrt.org/people/agb/) Aug 27 21:58:34 nbd: my build now fails. seems that I get compiler errors related to Luci Aug 27 21:59:00 show me Aug 27 21:59:33 nbd * r12415 /packages/net/ddns-scripts/files/etc/ (hotplug.d/iface/25-ddns ppp/ip-up.d/25-ddns): ddns-scripts: remove wrong comment and proto specific behavior Aug 27 22:07:51 MateIn4: please do a scripts/feeds update luci, there was a recent Makefile change which conflicts with older revisions Aug 27 22:08:23 CyrusFF: do you want me to start included luci in snapshot builds? Aug 27 22:08:39 including* Aug 27 22:08:46 Bartman007: yeah would be nice Aug 27 22:09:11 i just fixed the -hopefully- last fallout from the UCI api change Aug 27 22:10:07 CyrusFF: olsrd init.d-Script und BusyBox httpd Patch so okay fuer Dich? Aug 27 22:12:06 x-alina: yeah fine Aug 27 22:12:28 :) Aug 27 22:13:48 x-alina: xm has forwarded the init-script to aaron Aug 27 22:18:26 Muss mich jetzt erstmal auf meinen Praktikum (hoffentlich bald Job) konzentrieren. Werd dann an LuCI naechste oder uebernaechste Woche weiterarbeiten. (Das init.d-Script werd ich die naechsten Tage noch etwas flotter machen...) Aug 27 22:19:22 super, danke Aug 27 22:19:45 nbd * r12416 /trunk/package/ppp/files/ppp.sh: fix ppp interface enumeration - don't enumerate already processed interfaces Aug 27 22:30:06 nbd: know anything about cpmac? Aug 27 22:30:38 not much, just a bit Aug 27 22:31:25 i am seeing something strange in the bootlog Aug 27 22:31:51 it complains about duplicate sysfs nodes Aug 27 22:32:29 interesting Aug 27 22:32:35 i don't know anything about that Aug 27 22:33:43 anybody else i can think of who might is marked away Aug 27 22:36:55 Just a quick note: after upgrading and I'm able to rebuild without errors and the spi-ks8995 has been built; thank you all. Aug 27 22:37:02 you're welcome Aug 27 22:37:49 Just a quick note: after upgrading and I'm able to rebuild without errors and the spi-ks8995 has been built; thank you all. Aug 27 23:43:14 hi... anyone know how to enable large file samba support using uci? Aug 27 23:43:19 I would need to use something like option 'large readwrite' 'yes' in /etc/config/samba but I get an error about an invalid char when I try a uci show samba afterwards... Aug 27 23:43:44 try to replace the space with underscore Aug 27 23:43:48 just a wild guess Aug 27 23:43:59 but spaces in option names are illegal Aug 27 23:45:17 can I escape a space? Aug 27 23:45:33 not in option names Aug 27 23:45:48 spaces are illegal in option names, they can be used for the value though Aug 27 23:45:52 thay have to be [a-zA-Z0-9_]+ Aug 27 23:46:12 so how do openwrt propose to enable the option? Aug 27 23:46:31 hmm normally the init script takes care of that Aug 27 23:48:24 well with a mostly default build... i can't copy more than 2GB across to a drive hooked up... to the machine... via usb... and 2 GB is getting pretty small these days. Aug 27 23:49:14 johnrw: modify the init script to add an option that enables it. large_readwrite or large_rw seems like a fitting option name. Aug 27 23:49:48 johnrw: hi Aug 27 23:50:03 hi Aug 27 23:50:10 johnrw: huh? Aug 27 23:50:18 that feature is not supported yet ;( Aug 27 23:50:27 i'll put it on the list Aug 27 23:50:36 is the rest of samba3 working for you ? Aug 27 23:51:35 well I am not sure which one I am running... I got kinda frustrated and all but there is only an smbd daemon running... Aug 27 23:51:49 would that make it samba3? Aug 27 23:52:06 I am used to seeing nmbd as well Aug 27 23:52:27 no nmbd Aug 27 23:52:27 but the throughput was going along nicely... Aug 27 23:52:32 and yes, its samba3 Aug 27 23:53:12 ok... well I threw in some other packages... smbclient smbcommon i think... Aug 27 23:53:19 bootlog of trunk on a wag54gp2: http://openwrt.pastebin.com/f72ff522 Aug 27 23:53:31 ok Aug 27 23:55:29 no idea what's up with cpmac Aug 27 23:55:38 but as long as we are talking improvements... it would be great to have a query system to mount/umount dynamic attached drives... but I guess you know about that... I am glad webif offers a button to umount... but so far I haven't found a way to mount an available samba mount... that is unmounted. Aug 27 23:56:20 and is there anyone besides farnz and ejka that knows anything about vlynq? Aug 27 23:56:30 or even issue a mount -a command... via webif... Aug 27 23:56:47 sn9: not me Aug 27 23:57:34 blogic... how long do you think that take for you? Aug 27 23:58:47 dunno Aug 27 23:58:50 maybe next week Aug 27 23:59:45 ok... well I am not trying to be impatient... or anything but unless I keep track of cvs commits... I'd never find out when you make that change. Aug 28 00:01:10 send a mail to blogic@openwrt.org Aug 28 00:01:21 and i'll remeber to ml you when its fixes Aug 28 00:01:23 and i'll remeber to ml you when its fixed Aug 28 00:01:29 It looks like I can edit /etc/samba/smb.conf manually... and the changes do survive reboot... Aug 28 00:01:59 ok, great. Aug 28 00:02:06 what about implementing some kind of smb.conf.local ? Aug 28 00:02:18 which would be included if it exists Aug 28 00:02:27 yep Aug 28 00:02:45 k Aug 28 00:02:49 johnrw: /etc/samba/smb.conf is used as a template Aug 28 00:05:10 does that mean changes I make there will persist? Aug 28 00:05:19 sounds like it Aug 28 00:05:32 which works for me... Aug 28 00:09:25 As far as speed... I have a drive attached (usb2) to an x86 box and a 4.14GB file copy seems to take just over 10 minutes... is that about right? Aug 28 00:10:26 yep Aug 28 00:10:31 well Aug 28 00:10:35 could be faster Aug 28 00:10:40 what cpu speed ? Aug 28 00:11:42 well its an old pIII 866mhz... Aug 28 00:12:36 i think that speed sounds about right Aug 28 00:13:08 kernel module spi-ks8995, I've tested it; if loaded it initializes the hw. But it seems to me that there needs to be an entry added to /etc/modules.d to auto load it. Aug 28 00:13:33 MateIn4: will be fixed after my next series of commits Aug 28 00:14:12 OK, will retest after the commits. Thanks! Aug 28 00:15:15 nbd * r12417 /trunk/package/ppp/files/ppp.sh: ppp: derive interface unit numbers from previous enumeration Aug 28 00:15:24 nbd * r12418 /trunk/package/mtd/src/jffs2.c: mtd: add jffs2 cleanmarker for big endian systems Aug 28 00:15:37 nbd * r12419 /trunk/package/spi-ks8995/Makefile: add autoload for spi-ks8995 Aug 28 00:15:53 MateIn4: there, should be fixed now Aug 28 00:16:11 I'll retest. Aug 28 00:40:32 nbd: I can report your patches spi-ks8995 work. Aug 28 00:40:33 MateIn4: great Aug 28 00:42:35 Is there a way to specify that certain boards depend on the spi-ks8995? Of course, the mi424-wr falls in this category but there are others (e.g. some avila models) Aug 28 00:44:50 you could create a target profile for it Aug 28 00:46:54 I'll have to investigate how to do that. Aug 28 00:48:34 blogic: Do you know if the samba3 server is compiled without large file support by default on x86 generic? Aug 28 00:51:30 yes Aug 28 00:51:30 It seems that 'large readwrite' 'large_readwrite' have no effect. Aug 28 00:52:07 so can i enable it in make menuconfig? Aug 28 00:54:51 no Aug 28 00:54:51 its not working Aug 28 00:54:51 be patient Aug 28 00:54:51 ok... Aug 28 01:00:42 ok... well I gotta hit the savck early tonight... so I'll email you... if I can't figure out how to monitor patches. Thanks blogic. Aug 28 02:17:21 hi there Aug 28 02:17:28 benoar: what? Aug 28 02:18:05 wrt: do you have some openwrt router handy ? Aug 28 02:19:16 is wrt some kind of bot ? Aug 28 02:19:22 ) Aug 28 02:19:24 ;) Aug 28 02:19:30 ok, thanks :-) Aug 28 02:20:00 I am hunting bug #3946 Aug 28 02:20:08 could someone confirm that running : Aug 28 02:20:09 ipcalc.sh 192.168.1.1 255.255.255.0 100 151 Aug 28 02:20:16 gives some meaningfull result ? Aug 28 02:20:44 I am pretty sure, it is just to be sure it only happens on IXP4xx Aug 28 02:56:01 benoar: more than just ixp4xx Aug 28 02:56:49 I've seen in on atheros also, so I would assume it is big endian systems Aug 28 02:57:20 (regardless of what platforms it occurs on, it isn't just ixp4xx specific) Aug 28 02:59:08 Bartman007: thanks for the tip Aug 28 02:59:22 this is a big problem, has it been noticed earlier ? **** ENDING LOGGING AT Thu Aug 28 02:59:56 2008